s

363.739
H2TRIUC
1995

STATE OF MONTANA
NATURAL RESOURCE DAMAGE PROGRAM

TERRESTRIAL RESOURCES INJURY ASSESSMENT REPORT
UPPER CLARK FORK RIVER NPL SITES

JANUARY 1995

sr'^-T^'f^J^

STATE DOCUMENTS COLLECTION

MAY 1 2 1995

MONTANA STATE LIBRARY

1515 E. 6th AVE.
HELENA, MONTANA 59620

"mOn'tANAs'tATE LIBRARY

3 0864 0015 3854 8

•L-s:

thi hi*

M 2 9 20Q2

TERRESTRIAL RESOURCES
INJURY ASSESSMENT REPORT

UPPER CLARK FORK RIVER BASIN

TERRESTRIAL RESOURCES
INJURY ASSESSMENT REPORT

UPPER CLARK FORK RIVER BASIN

STATE OF MONTANA
NATURAL RESOURCE DAMAGE LITIGATION PROGRAM

Prepared by:

RCG/Hagler Bailly

P.O. Drawer O

Boulder, CO 80306-1906

(303) 449-5515

Testifying Experts:

Joshua Ltpton, Ph.D.

Hector Galbr

aitii, Ph.D

#

KatherlneXeJeiine

<f

JANUARY 1995

TERRESTRIAL RESOURCES
INJURY ASSESSMENT REPORT

UPPER CLARK FORK RIVER BASIN

STATE OF MONTANA
NATURAL RESOURCE DAMAGE LITIGATION PROGRAM

Prepared by:

RCG/Hagler Bailly

P.O. Drawer O

Boulder, CO 80306-1906

(303) 449-5515

Testifying Expert:

Harold BergmanrTh.D.
University of Wyoming

TERRESTRIAL RESOURCES
INJURY ASSESSMENT REPORT

UPPER CLARK FORK RIVER BASIN

STATE OF MONTANA
NATURAL RESOURCE DAMAGE LITIGATION PROGRAM

Prepared by:

RCG/Hagler Bailly

P.O. Drawer O

Boulder, CO 80306-1906

(303) 449-5515

Testifying Expert:

S/Uc^tAc

Lyman McDonald, Ph.D.
Western Ecosystems Technology, Inc.

Digitized by the Internet Archive

in 2011 with funding from

Montana State Library

http://www.archive.org/details/terrestrialresou1995lipt

CONTENTS

TABLES vi

FIGURES x

1.0 INTRODUCTION AND SUMMARY 1-1

1.1 INJURY ASSESSMENT 1-4

1.2 REPORT ORGANIZATION AND SUMMARY OF FINDINGS 1-6

1.3 REFERENCE 1-11

2.0 SOURCES AND PATHWAYS OF HAZARDOUS SUBSTANCES TO

UPLAND RESOURCES 2-1

2.1 SOURCE/PATHWAY SUMMARY 2-1

2.2 AIR PATHWAY: EXPOSURE OF AIR TO HAZARDOUS
SUBSTANCES 2-2

2.2.1 Smelter Stack Emissions 2-2

2.2.2 Fugitive Emissions 2-4

2.3 AIR PATHWAY: AREAL EXTENT OF EXPOSURE 2-7

2.3.1 Historical Accounts 2-7

2.3.2 Areal Patterns Indicative of a Point Source 2-9

2.3.3 Soil Profile Concentrations Indicative of a Surficial Source .... 2-18

2.3.4 Fugitive Emissions 2-18

2.3.5 Areal Extent - Summary 2-25

2.4 MOBILITY AND RATES OF TRANSPORT OF HAZARDOUS
SUBSTANCES IN AIR 2-25

2.4.1 Main Stack Emissions 2-25

2.4.2 Fugitive Emissions 2-26

2.5 OTHER PATHWAYS 2-28

2.5.1 Leaching/Erosion 2-28

2.5.2 Biological Uptake 2-28

2.6 SUMMARY 2-29

2.7 REFERENCES 2-30

3.0 GEOLOGIC RESOURCE — UPLAND SOILS 3-1

3.1 DESCRIPTION OF UPLAND SOIL RESOURCES 3-1

3.1.1 Stucky Ridge 3-2

3.1.2 Smelter Hill 3-2

3.1.3 Mount Haggin 3-4

3.2 INJURY DEFINITION 3-4

3.2.1 Phytotoxicity 3-4

3.2.2 Assessment of Soils Phytotoxicity 3-6

RCG/Hagler Bailly i

CONTENTS

3.2.2.1 Laboratory Testing 3-6

3.2.2.2 Vegetation Communities 3-8

3.3 INJURY DETERMINATION AND QUANTIFICATION 3-8

3.3.1 Baseline Conditions 3-8

3.3.2 Soil Concentrations of Hazardous Substances 3-11

3.3.2.1 NRDA Investigations 3-11

3.3.2.2 Previous Investigations 3-15

3.3.3 Category of Injury: Phytotoxicity - Laboratory Evidence 3-20

3.3.3.1 NRDA Phytotoxicity Investigation 3-20

3.3.3.2 Statistical Relationships Between Phytotoxicity

and Hazardous Substances 3-24

3.3.3.3 Previous Phytotoxicity Studies 3-29

3.3.3.4 Implications of Phytotoxicity 3-31

3.3.4 Extent of Injury 3-31

3.3.4.1 Areal Extent of Injury 3-31

3.3.4.2 Volume of Injured Soils 3-31

3.3.5 Ability of Soil Resource to Recover 3-32

3.4 SUMMARY 3-34

3.5 REFERENCES 3-34

4.0 UPLAND BIOLOGICAL RESOURCES — VEGETATION, WILDLIFE, AND

WILDLIFE HABITAT 4-1

4. 1 DESCRIPTION OF UPLAND BIOLOGICAL RESOURCES 4-2

4.2 INJURY DEFINITION 4-2

4.3 ASSESSMENT METHODOLOGIES 4-3

4.3.1 Vegetation Assessment 4-3

4.3.2 Wildlife Habitat Assessment 4-4

4.3.2.1 Forest Habitat: Indicator Species = Marten 4-5

4.3.2.2 Forest Edge and Grassland Habitat: Indicator

Species = Elk 4-6

4.3.2.3 Habitat Layers 4-6

4.3.3 Baseline Conditions 4-7

4.3.3.1 Historical Information 4-7

4.3.3.2 Control Areas 4-7

4.4 INJURY DETERMINATION AND QUANTIFICATION 4-9

4.4.1 Previous Investigations 4-9

4.4.2 NRDA Investigations 4-18

4.4.2.1 Vegetation Characteristics 4-19

4.4.2.2 Relationships between Observed Vegetation and
Phytotoxicity 4-36

RCG/Hagler Bailly

CONTENTS

4.4.3 Habitat Suitability 4-38

4.4.4 Causality Evaluation: Vegetation Losses in the Injured Area .... 4-41

4.4.4.1 Fire 4-42

4.4.4.2 Logging 4-46

4.4.4.3 Grazing 4-47

4.4.4.4 Sulfur Dioxide 4-49

4.4.4.5 Evidence of Hazardous Substance Causality 4-51

4.4.5 Extent of Injury 4-53

4.5 ABILITY OF RESOURCE TO RECOVER 4-54

4.6 SUMMARY 4-55

4.7 REFERENCES 4-55

5.0 SOURCES AND PATHWAYS OF HAZARDOUS SUBSTANCES TO

RIPARIAN RESOURCES 5-1

5.1 SILVER BOW CREEK 5-1

5.1.1 Identification of Sources of Hazardous Substances — Silver Bow
Creek 5-1

5.1.2 Identification of Transport Pathway — Silver Bow Creek 5-1

5.1.3 Extent of Pathway Contamination — Silver Bow Creek 5-3

5.2 CLARK FORK RIVER 5-5

5.2.1 Identification of Sources of Hazardous Substances — Clark Fork
River 5-5

5.2.2 Identification of Transport Pathways — Clark Fork River 5-5

5.2.3 Extent of Pathway Contamination — Clark Fork River 5-6

5.3 OPPORTUNITY PONDS 5-7

5.4 REFERENCES 5-7

6.0 GEOLOGIC RESOURCE — RIPARIAN SOILS 6-1

6.1 DESCRIPTION OF RIPARIAN SOIL RESOURCE 6-1

6.1.1 Silver Bow Creek 6-1

6.1.2 Clark Fork River 6-2

6.1.3 Opportunity Ponds 6-2

6.2 INJURY DEFINITION 6-3

6.2.1 Assessment of Phytotoxicity 6-3

6.2.2 Changes in pH 6-5

6.3 INJURY DETERMINATION AND QUANTIFICATION 6-5

6.3.1 Baseline Conditions 6-5

6.3.2 Injury Determination 6-7

6.3.2.1 NRDA Riparian Soil Sampling 6-7

6.3.2.2 NRDA Phytotoxicity Tests 6-10

RCG/Hagler Bailly /;;

CONTENTS

6.3.2.3 STARS Phytotoxicity Tests 6-23

6.3.2.4 Field Evidence 6-23

6.3.2.5 Reduction of Soil pH 6-27

6.3.3 Injury Quantification 6-27

6.3.3.1 Areal Extent of Unvegetated Slickens 6-27

6.3.3.2 Volume of Contaminated Material 6-28

6.3.4 Ability of Resource to Recover 6-31

6.4 REFERENCES 6-33

7.0 BIOLOGICAL RESOURCES — RIPARIAN VEGETATION, WILDLIFE,

AND WILDLIFE HABITAT 7-1

7.1 DESCRIPTION OF RIPARIAN BIOLOGICAL RESOURCES 7-1

7.2 INJURY DEFINITION 7-2

7.2.1 Methods of Measuring Injury to Wildlife Habitat Quantity and
Quality 7-3

7.2.1.1 HEP Models 7-3

7.2.1.2 Wildlife Population Surveys 7-4

7.2.1.3 Vegetation Assessment 7-4

7.2.2 Baseline Conditions 7-5

7.3 INJURY DETERMINATION AND QUANTIFICATION 7-6

7.3.1 NRDA Investigations 7-6

7.3.1.1 Vegetation Characteristics 7-8

7.3.1.2 Habitat Suitability 7-17

7.3.2 Extent of Injury 7-26

7.3.3 Ability of the Resource to Recover 7-26

7.4 REFERENCES 7-27

8.0 BIOLOGICAL RESOURCES — OTTER, MINK, RACCOON 8-1

8.1 INTRODUCTION 8-2

8.2 PATHWAY DETERMINATION 8-4

8.2.1 Mink and Otter Dietary Exposure to Hazardous Substances 8-4

8.2.2 Mnk Tissue Concentrations of Hazardous Substances 8-8

8.3 INJURY DEFINITION 8-12

8.4 INJURY DETERMINATION: RESULTS OF STUDIES 8-13

8.5 INJURY QUANTIFICATION 8-16

8.5.1 Population Reductions among Mink, Otter, and Raccoons on the

Upper Clark Fork 8-16

8.6 CONCLUSIONS 8-18

8.7 REFERENCES ... 8-20

RCG/Hagler Bailly fv

CONTENTS

APPENDIX A:

APPENDIX B:

Assessment of Injury to Soils: Concentrations of Hazardous Substances
in Exposed Soils and Comparison with Baseline Conditions

Evaluation of Phytotoxicity of Upland and Riparian Soils, Clark Fork
River Basin, Montana

APPENDIX C:

Assessment of Injury to Vegetation, Wildlife and Wildlife Habitat
1992 Data

APPENDK D:

Results of 1994 Vegetation Sampling in Anaconda Uplands, Southwest
Montana

APPENDDC E: Assessment of Injury to Semi-Aquatic Mammals

RCG/Hagler Bailly

TABLES

Table 1-1 Injury and Pathway Summary for Terrestrial Resources 1-4

Table 2-1 Record of Main Stack Emissions for the Period Between October 1 and

October 31, 1976 2-3

Table 2-2 Anaconda Smelter History 2-5

Table 2-3 Summary of Air Quality Data for Four Stations in the Anaconda

Smelter Site Area — High-Volume Sampler Data 2-6

Table 2-4 Summary of Trace Metal Concentrations in Dustfall Samples Collected

by 13 Dustfall Samplers 2-8

Table 2-5 Comparison of 0-2 Inch and 0-6 Inch Soil Concentrations:

Upland Areas Near Anaconda 2-24

Table 3-1 A Simple Classification of the Effects of Hazardous Substances on

Plants 3-6

Table 3-2 Background Concentrations Reported for Uncontaminated Soils of the

United States 3-9

Table 3-3 Baseline Concentrations of Hazardous Substances in Anaconda Area

Soils 3-11

Table 3-4 Summary Statistics for Anaconda Upland Impact and German Gulch

Control 2-Inch Soil Samples 3-13

Table 3-5 Comparisons of Mean Soil Nutrient Concentrations, pH CEC, %

Organic Matter, and % Particle Size in Impact and Reference Soils .... 3-15

Table 3-6 Summary Statistics for Smelter Hill OU RI/FS Phase I Soils Data 3-18

Table 3-7 Summary Statistics for Smelter Hill OU RI/FS Phase II Soil Samples,

Disturbed and Unreclaimed Soils 3-19

Table 3-8 Statistical Comparison of Upland Impact Areas to Pooled Control

Means 3-22

Table 3-9 Summary of Phytotoxicity as Determined by Overall Site Scores 3-23

Table 3-10 Soil Concentrations Reported to Cause Phytotoxicity and Ratios of

Mean Impact Area Concentrations to Phytotoxic Levels 3-25

Table 3-11 Kendall-t Correlations Between Phytotoxicity Score and [ETI-Adjusted

Metal Concentration 3-28

Table 3-12 Kendall-t Partial Correlation Coefficients Between Phytotoxicity Score

and Bioavailable Metal Concentration, with Partialing Variable [H+] . . . 3-28
Table 3-13 Kendall-t Correlations Between Phytotoxicity Score and [H'J-Adjusted

Metal Concentration 3-29

Table 3-14 Kendall-t Partial Correlation Coefficients Between Phytotoxicity Score

and Bioavailable Metal Concentration, with Partialing Variable [TT] . . . 3-30

RCG/Hagler Bailly vz

TABLES

Table 4-1 Comparison of Elevation, Aspect, Slope, and Distance to Winter Road

for Paired Impact and Control Sites 4-10

Table 4-2 Birds Characteristic of Southwest Montana Lodgepole Pine and Douglas Fir

Forests and their Feeding and Nesting Vegetation Layers 4-14

Table 4-3 Mammals Characteristic of Southwest Montana Lodgepole Pine and

Douglas Fir Forests and their Feeding and Cover Vegetation Layers ... 4-16
Table 4-4 Comparison of the Proportional Representation of Cover Types in

Impact and Control Areas 4-24

Table 4-5 Comparison of Grasslands in 21 Upland Impact and Control Areas .... 4-25
Table 4-6 Comparison of the Proportional Representation of Habitat Layers in

Impact Areas and Control Areas 4-25

Table 4-7 Comparison of the Mean Number of Habitat Layers Present in Impact

and Control Sites 4-32

Table 4-8 Results of t-tests Comparing Cover Types Observed at Impact Sites in

1992 and 1994 4-32

Table 4-9 Proportional Representations of Dominant Species in Grassland

Cover Type 4-34

Table 4-10 Comparison of Marten Habitat Suitability in Impact and Control Areas . 4-38

Table 4-11 Comparison of Elk Habitat Suitability in Impact and Control Areas .... 4-39

Table 4-12 Comparison of Habitat Complexity in Impact and Control Areas 4-40

Table 4-13 Bird and Mammal Populations that are Likely to have been Lost or Suffered

Reduced Viability Due to Removal of Tree Canopy, Bole, Shrub Midstory,

Understory, and Terrestrial Subsurface Habitat Layers in

Upland Impact Areas 4-41

Table 4-14 Comparisons of Mean Potassium Concentrations, pH and CEC in Impact and

Reference Soils 4-51

Table 4-15 Evaluation of Causality 4-54

Table 5-1 Sources of Hazardous Substances to Silver Bow Creek 5-2

Table 5-2 Average Total Metals and Arsenic Concentrations in Tailings and

Floodplain Sediments of Silver Bow Creek 5-4

Table 5-3 Concentrations in Tailings and Floodplain Sediments of the Clark Fork

River 5-6

Table 6-1 Baseline and Background Concentrations in Riparian Soils and Baseline
Conditions for this NRDA are the Mean of Divide Creek and Little
Blackfoot River Samples 6-6

Table 6-2 Two-Sample Randomization Test of the Mean Difference Between

Impact and Control Soil Concentrations 6-12

RCG/Hagler Bailly w7

TABLES

Table 6-3 Results of the Two-Sample Randomization Test Comparing

Concentrations in Opportunity Ponds Soils and Riparian Control Reach

Soils 6-12

Table 6-4 Summary Statistics for Riparian Impact and Control Reach 2-Inch Soils

Samples 6-13

Table 6-5 Streambank Tailings and Revegetation Study Data Collected from

Unvegetated Streamside Tailings Along Silver Bow Creek 6-15

Table 6-6 Tailings Data Collected for the Silver Bow Creek RI from Ramsay Flats

Tailings Material 6-16

Table 6-7 Tailings Samples Collected from Streamside Tailings Along Silver Bow

Creek for the Tailings Investigation 6-17

Table 6-8 Data Collected from Three Bare Tailings Sites in the Clark Fork River

Floodplain for the Upper Clark Fork River Screening Study 6-20

Table 6-9 Screening Tests: Total Number of Germinating/Emergent Seedlings .... 6-22

Table 7-1 Birds Characteristic of Southwest Montana Riparian Shrub/Forest and

their Feeding and Nesting Vegetation Layers 7-7

Table 7-2 Mammals Characteristic of Southwest Montana Riparian Shrub/Forest

and their Feeding and Cover Vegetation Layers 7-8

Table 7-3 Comparison of the Proportional Representation of Cover Types in

Impact and Control Reaches 7-17

Table 7-4 Comparison of the Proportional Representation of Habitat Layers on

Impact and Control Streams 7-20

Table 7-5 Comparison of the Mean Number of Cover Types Observed on Impact

and Control Reaches 7-24

Table 7-6 Comparison of Vertical Diversity and Habitat Complexity in Impact and

Control Reaches 7-24

Table 7-7 Bird and Mammal Populations that are Likely to have Suffered Local

Extirpation or Reduced Viability because of Reductions of Tree Canopy,

Bole, Shrub Mdstory, Understory, and Terrestrial Subsurface Habitat

Layers in Riparian Impact Areas 7-25

Table 7-8 Mean Numbers and Densities of Birds Seen during Two Avian Surveys
on Silver Bow Creek and the Little Blackfoot River between Elliston
and Avon 7-26

Table 8-1 Frequency of Occurrence of Food in Intestine and Stomach of Mink

Collected from All Clark Fork River and Reference Sites during 1991-

1992 Trapping Season 8-5

RCG/Hagler Bailly W/7

TABLES

Table 8-2

Table 8-3

Table 8-4

Table 8-5

Table 8-6

Table 8-7
Table 8-8

Whitefish Mean Body Weight and Mean Whole Body Metal
Concentrations for Fish Collected from the Clark Fork River and

Reference Sites in 1991 and 1992 8-6

Sucker Mean Body Weight and Mean Whole Body Metal
Concentrations for Fish Collected from the Clark Fork River and

Reference Sites in 1992 8-7

Mink Mean Liver Metal Concentrations by Trapping Location for Mink
Collected on the Clark Fork River and Reference Sites During the 1991-

92 Trapping Season 8-9

Mink Mean Concentrations by Trapping Location for Mink Collected on
the Clark Fork River and Reference Sites During the 1991-92 Trapping

Season 8-10

Mink Mean Brain Metal Concentrations by Trapping Location for Mink
Collected on the Clark Fork River and Reference Sites During the 1991-

92 Trapping Season 8-11

TRC for Pb and Cd in the Diet of Mink and Otter 8-14

TRC Values and Dietary Exposure Concentrations for Cd and Pb for

Mink and Otter Consuming Fish Collected from below Warm Springs

Ponds and Reference Sites 8-15

RCG/Hagler Bailly

IX

FIGURES

Figure 1-1 Upper Clark Fork River Basin Assessment Area: Terrestrial Resources .. 1-3
Figure 1-2 Ecological Relationships of Terrestrial Resources in the Upper Clark Fork

River Basin 1-7

Figure 2-1 Arsenic Emitted from the Anaconda Smelter Main Stack 1974 - 1980 ... 2-3
Figure 2-2 Distribution of Cadmium in Soil with Distance from the Main Smelter

Stack 2-10

Figure 2-3 Cadmium Concentrations in Vegetation and Grasshoppers in 1970 2-11

Figure 2-4 Location of Transects in Figures 2-5 through 2-9 2-12

Figure 2-5 Northwest Transect 2-13

Figure 2-6 Northeast Transect 2-14

Figure 2-7 Southeast Transect 2-15

Figure 2-8 Mill Creek Transect 2-16

Figure 2-9 West Valley Transect 2-17

Figure 2-10 Arsenic Concentration in Soils with Distance from the Main Smelter

Stack 2-19

Figure 2-11 Cadmium Concentration in Soils with Distance from the Main Smelter

Stack 2-20

Figure 2-12 Copper Concentration in Soils with Distance from the Main Smelter

Stack 2-21

Figure 2-13 Lead Concentration in Soils with Distance from the Main Smelter Stack 2-22

Figure 2-14 Zinc Concentration in Soils with Distance from the Main Smelter Stack . 2-23
Figure 2-15 Smoothed Patterns for Copper Concentrations in Soils as a Function of

Downwind Distance from a Copper Smelter 2-24

Figure 2-16 Effects of Rugged Terrain on Dispersion 2-27

Figure 3-1 Upland Impact and Control Soil Sampling Sites 3-3

Figure 3-2 Mean Concentrations of Hazardous Substances in Upland Impact and

Control Soils 3-14

Figure 3-3 Smelter Hill Operable Unit 3-16

Figure 3-4 Scatter Plot of Total Arsenic and Total Copper, and pH at Sites Tested

for Phytotoxicity 3-26

RCG/Hagler Bailly

FIGURES

Figure 4-1 Upland Impact and Control Vegetation Sampling Sites 4-8

Figure 4-2 Proportional Representation of the Cover Types of Stucky Ridge and

Corresponding Control Sites 4-11

Figure 4-3 Proportional Representation of the Cover Types of Smelter Hill and

Corresponding Control Sites 4-12

Figure 4-4 Proportional Representation of the Cover Types of Mount Haggin and

Corresponding Control Sites 4-13

Figure 4-5a Upland Sample Site A6 on Stucky Ridge Showing Devegetation and

Loss of Topsoil 4-21

Figure 4-5b Upland Sample Site B13 on Smelter Hill Showing Devegetation and

Dead and Moribund Vegetation 4-21

Figure 4-5c Upland Sample Site C8 on Mt. Haggin Showing Devegetation and

Evidence of Previous Afforestation 4-22

Figure 4-5d Upland Sample Site C9 on Mt. Haggin Showing Devegetation and

Evidence of Previous Afforestation 4-22

Figure 4-5e German Gulch Upland Control Area Showing Douglas Fir

Forest Cover 4-23

Figure 4-5f German Gulch Upland Control Area Showing Lodgepole Pine Forest

Cover 4-23

Figure 4-5g Upland Control Sample Sites B13 and C7 4-24

Figure 4-6 Proportional Representation of the Habitat Layers of Stucky Ridge and

Corresponding Control Sites 4-26

Figure 4-7 Proportional Representation of the Habitat Layers of Smelter Hill and

Corresponding Control Sites 4-27

Figure 4-8 Proportional Representation of the Habitat Layers of Mount Haggin and

Corresponding Control Sites 4-28

Figure 4-9 Proportional Representation of the Number of Habitat Layers of Stucky

Ridge and Corresponding Control Sites 4-29

Figure 4-10 Proportional Representation of the Number of Habitat Layers of Smelter

Hill and Corresponding Control Sites 4-30

Figure 4-1 1 Proportional Representation of the Number of Habitat Layers of Mount

Haggin and Corresponding Control Sites 4-31

Figure 4-12 Comparisons of Proportional Representations of Cover Types in 1992

and 1994 4-33

Figure 4-13 Mean Percent Cover of Bare Ground, Vegetation, and Plant Litter in

Grassland in the Impact Area, as Determined by Line Intercept

Measurements 4-35

Figure 4-14 Sample Sites at which Evidence of Previous Afforestation was Found . . 4-37
Figure 4-15 Generalized Forest Succession in Cool, Dry Douglas Fir Habitat Types . 4-44

RCG/Hagler Bailly xi

FIGURES

Figure 4-16 Expected Fingerprint of Fire in the Anaconda Area 4-45

Figure 4-17a Alder Gulch in 1871 4-48

Figure 4-17b The Same View of Alder Gulch in 1981, 110 Years Later 4-48

Figure 6-1 Trends in the Mobility of Metals as Influenced by Soil pH; Data for

Light Mineral Soil 6-6

Figure 6-2 Concentrations of Hazardous Substances in the Silver Bow Creek

Floodplain 6-8

Figure 6-3 Location of Riparian Soil, Vegetation, and Wildlife Sampling Reaches:

Silver Bow Creek and the Clark Fork River 6-11

Figure 6-4 Mean Concentrations of Hazardous Substances at Impact and Control

Reaches 6-14

Figure 6-5 Hybrid Poplar Plants After 28 Days Exposure to Riparian Soil 6-24

Figure 6-6 Representative Hybrid Poplar Plants Excavated from Riparian Soil after

28 Days Exposure 6-25

Figure 6-7 Concentrations of Each Element at Sample Points Collected Within

Each Impact Reach 6-32

Figure 7-1 Location of Riparian Sampling Sites: Silver Bow Creek and the Clark

Fork River 7-9

Figure 7-2 Proportional Representation of the Dominant Cover Types Encountered

on Riparian Control and Impact Reaches 7-10

Figure 7-3 Proportional Representation of the Cover Types Encountered on the

Opportunity Ponds 7-11

Figure 7-4 Comparison of the Proportional Representation of Riparian Shrub/Forest

on Control and Impact Reaches 7-12

Figure 7-5a Riparian Impact Sample Site R7 Showing Slickens, Devegetation, and

Dead Willows 7-14

Figure 7-5b Riparian Impact Sample Site R8 Showing Slickens, Devegetation, and

Dead Vegetation 7-14

Figure 7-5c Riparian Impact Sample Site RIO Showing Slickens, Devegetation, and

Dead Willow Shrub 7-15

Figure 7-5d Riparian Control Sample Site on Little Blackfoot River Showing

Riparian Forest and Dense Shrub and Herb Layers 7-16

Figure 7-5e Riparian Control Sample Site on Little Blackfoot River Showing

Riparian Forest and Dense Shrub and Herb Layers 7-16

Figure 7-6 Proportional Representation of Habitat Layers on Riparian Control and

Impact Reaches 7-18

Figure 7-7 Proportional Representation of Habitat Layers on the Opportunity Ponds 7-19

RCG/Hagler Bailly xii

FIGURES

Figure 7-8 Proportional Representation of the Mean Number of Habitat Layers on

Riparian Control and Impact Reaches 7-21

Figure 7-9 Proportional Representation of the Mean Number of Habitat Layers on

the Opportunity Ponds 7-22

Figure 7-10 Maximum Number of Habitat Layers Observed on Control and Impact

Riparian Reaches 7-23

Figure 8-1 Map of Western Montana Showing the Numbers and Locations of Otter
Trapped in each Montana Department of Fish, Wildlife, and Parks
Administrative Region during the 1977-78 through 1989-90 Trapping
Seasons 8-17

Figure 8-2 Abundance of Otter, Mink, Raccoon, and Beaver Sign on the Six Paired

Clark Fork River and Reference Sites During the Fall of 1992 8-19

RCG/Hagler Bailly xiii

ACRONYMS

AMC Anaconda Minerals Company

ASI Anaconda Soil Investigation

ASTM American Society for Testing and Materials

ARCO Atlantic Richfield Company

BWC Bottom of Water Column

CEC Cation Exchange Capacity

DOI United States Department of the Interior

DTPA Diethylenetriaminepentaacetic Acid

GPS Global Positioning System

HEP Habitat Evaluation Procedures

HPS Handling/Process/Storage Area

HSI Habitat Suitability Index

MBMG Montana Bureau of Mines and Geology

MDFWP Montana Department of Fish, Wildlife, and Parks

MSU Montana State University

NPL National Priorities List

NRDA Natural Resource Damage Assessment

NRDLP Natural Resource Damage Litigation Program

OM Organic Matter

OU Operable Unit

PSCI Preliminary Site Characterization Information

QA Quality Assurance

QA/QC Quality Assurance/Quality Control

QC Quality Control

RI/FS Remedial Investigation/Feasibility Study

SBC Silver Bow Creek

SI Suitability index

SM Shrub Midstory

SOP Standard Operating Procedure

STARS Streamside Tailings and Revegetation Studies

SW Surface Water Layer

TB Tree Bole

TC Tree Canopy

TS Terrestrial Subsurface

US Understory

U.S. BLM United States Bureau of Land Management

U.S. EPA United States Environmental Protection Agency

U.S. FWS United States Fish and Wildlife Service

USGS United States Geological Survey

U.S. SCS United States Soil Conservation Service

RCG/Hagler Bailly

X7V

ACRONYMS

UTM Universal Transverse Mercator Grid

WC Water Column Layer

XRF X-Ray Flourescence

RCG/Hagler Bailly *v

1-1

1.0 INTRODUCTION AND SUMMARY

Terrestrial resources of the upper Clark Fork River Basin have been injured as a result of
historical and ongoing releases of hazardous substances from mining and mineral-processing
operations. This report describes the results of injury determination and quantification for
terrestrial resources. The resources addressed are soils, vegetation, wildlife, and wildlife
habitat. The geographic scope of the injury assessment includes upland areas near Anaconda,
and riparian areas along Silver Bow Creek (from upstream of the Colorado Tailings in Butte
to the Warm Springs Ponds) and the Clark Fork River (from its headwaters just below the
Warm Springs Ponds to the Milltown Reservoir) (Figure 1-1).

Injuries to soils, vegetation, wildlife, and wildlife habitat have been caused by historical and
ongoing releases of the hazardous substances arsenic, cadmium, copper, lead, and zinc from
mining and mineral-processing operations in Butte and Anaconda. The following information
on terrestrial resources is presented in the various chapters of the report:

Soils

►

Riparian (floodplain) soils have been injured throughout the length of Silver
Bow Creek, at least the upper 17 miles of the Clark Fork River, and
approximately 3,400 acres of the Opportunity Ponds. These soils were found
to be severely phytotoxic in controlled laboratory tests, little to no vegetation
exists on these injured soils.

Riparian soils are exposed to hazardous substances primarily through exposure
to surface water pathways.

Riparian soils serve as an exposure pathway to riparian vegetation, wildlife,
and wildlife habitat. In addition, riparian soils, through surface runoff, serve as
a pathway to surface water (and hence to both aquatic resources and to
downstream riparian soils).

Upland soils have been injured throughout approximately 17.8 square miles
near Anaconda. Soils from these upland areas, including portions of Stucky
Ridge, Smelter Hill, and the Mt. Haggin area, were found to be phytotoxic in
controlled laboratory tests. These soils have caused injury to vegetation
communities in these upland areas.

Upland soils have been exposed to hazardous substances by smelter emissions
and ongoing fugitive releases from contaminated soils. Hence, both air and
soils act as pathways to soil resources.

Upland soils serve as exposure pathways to upland vegetation, wildlife, and
wildlife habitat resources.

RCG/Hagler Bailly

1-2
Vegetation

► Riparian vegetation has been injured throughout the length of Silver Bow
Creek, the upper 17 miles of the Clark Fork River, and the approximately
3,400 acres in the Opportunity Ponds. These soils were found to be severely
phytotoxic in controlled laboratory tests; little to no vegetation exists in these
areas.

*■ Injuries to riparian vegetation include reduced cover, increased bare ground,

reduced forest/shrub community, and reduced habitat complexity (number of
habitat layers).

»• Upland vegetation has been injured throughout approximately 17.8 square miles

near Anaconda. Soils from these upland areas, including portions of Stucky
Ridge, Smelter Hill, and the Mt. Haggin area, were found to be phytotoxic in
controlled laboratory tests, vegetation communities have been injured in these
upland areas.

► Injuries to upland vegetation include reduced cover, increased bare ground,
reduced forest/shrub/grassland communities, and reduced habitat complexity
(number of habitat layers).

Wildlife and Wildlife Habitat

► Riparian wildlife and wildlife habitat have been injured throughout the length
of Silver Bow Creek, the upper 17 miles of the Clark Fork River, and the
approximately 3,400 acres in the Opportunity Ponds. These injured areas no
longer provide sufficient habitat to support viable populations of wildlife
species typical of Montana riparian habitat.

*■ Semi-aquatic furbearers, including otter, mink, and raccoon, have been injured

as a result of exposure to hazardous substances in their food. Mink and
raccoon populations have been reduced throughout the length of Silver Bow
Creek and the Clark Fork River (from Warm Springs Ponds to Milltown).
Otter have been entirely eliminated from this area.

► Upland wildlife and wildlife habitat have been injured throughout
approximately 17.8 square miles near Anaconda. These areas no longer
provide sufficient habitat to support viable populations of wildlife species
typical of Montana upland habitat.

Injury to these individual upland terrestrial resources has disrupted fundamental ecological
processes and cycles, and has resulted in ecosystem-level injury. Table 1-1 provides a
summary of injuries and pathways to terrestrial resources.

RCG/Hagler Bailly

<u

-
S
O
in
<U

(A

"5

4)

<

<

4)

>

2

_

o

'-

'_
<u
a.
a.
P

•-

3
OX)

•s

CQ

00

CO

a

1-4

Table 1-1
Injury and Pathway Summary for Terrestrial Resources

Resource

Location

Injuries

Pathways to
Injured Resource

Soils

► Silver Bow Creek

► Clark Fork River1

► Anaconda uplands

► Concentrations of
hazardous substances
cause phytotoxicity

»• Soil pH < 4.0

► Concentrations of
hazardous substances
cause injury to
vegetation

► Air

*■ Surface water

*■ Other soils

Vegetation

► Silver Bow Creek
►• Clark Fork River1

► Anaconda uplands

»• Phytotoxicity (death)
► Phytotoxicity
(reduced growth)

»• Soils

Wildlife — riparian

»■ Silver Bow Creek
► Clark Fork River1

► Reduced habitat

► Reduced population
viability

»■ Soils

Wildlife — upland

► Anaconda uplands

►■ Reduced habitat
► Reduced population
viability

► Soils

Wildlife —
furbearers

► Silver Bow Creek

► Clark Fork River

► Reduced viability,
including population
reductions

► Fish

1 Injun' to soils, vegetation, and riparian/terrestrial wildlife and wildlife habitat was assessed in
the Clark Fork River from Warm Springs Ponds to Deer Lodge.

2 Injury to semi-aquatic furbearers was assessed in the Clark Fork River from Warm Springs Ponds
to Milltown.

1.1 INJURY ASSESSMENT

The U.S. Department of the Interior (DOI) has promulgated regulations for the performance
of natural resource damage assessments [43 CFR Part 11]. This assessment was performed in
accordance with these regulations.

The term injury is defined as a measurable adverse change, either long- or short-term, in the
chemical or physical quality or the viability of a natural resource resulting either directly or
indirectly from exposure to a release of a hazardous substance, or exposure to a product of
reactions resulting from the release of a hazardous substance. [43 CFR § 1 1.14 (v)].

RCG/Hagler Bailly

1-5

The assessment of injury to terrestrial resources of the Clark Fork River Basin included three
steps. The first two steps constituted the injury determination phase, and the final step was
the injury quantification phase:

► Injury Definition. In the injury definition phase, those injuries that were found
to meet the definitions of injury in 43 CFR § 1 1.62 were evaluated.

*■ Pathway Determination. In the pathway determination phase, exposure

pathways of hazardous substances to injured natural resources were identified
[43 CFR § 11.63]. The DOI notes that pathway determination may be
accomplished by the "demonstration of sufficient concentrations in the pathway
for it to have carried the substance to the injured resources." [51 FR 27684].
In this assessment, "sufficient concentrations" of hazardous substances in
pathway resources have been demonstrated in air (through known smelter
emissions), soils, and fish.

► Injury Quantification. The effects of the releases of hazardous substances were
quantified in terms of changes from "baseline conditions" [43 CFR § 11.70
(a)].

Baseline conditions are the conditions that "would have existed at the assessment area had the
. . . release of the hazardous substance ... not occurred" [43 CFR § 11.14 (e)] and are the
conditions to which injured natural resources should be restored [43 CFR § 11.14 (11)].
Baseline conditions should take into account both natural processes and human activities, and
should include the normal range of physical, chemical, or biological conditions for the
assessment area or injured resource [43 CFR § 11.72 (b)]. In addition, baseline data
collection "shall be restricted to those data necessary for a reasonable cost assessment" [43
CFR § 11.72 (b)(4)]. Where historical baseline data are not available, "baseline data should
be collected from control areas" [43 CFR § 11.72 (d)]. "Control area" is defined as "an area
or resource unaffected by the discharge ... or release of the hazardous substance under
investigation. A control area or resource is selected for its comparability to the assessment
area or resource and may be used for establishing the baseline condition and for comparison
to injured resources" [43 CFR § 11.14 (i)]. Control area selection is based on criteria set
forth at 43 CFR § 11.72 (d)(l-7):

► One or more control areas shall be selected based upon their similarity to the
assessment area and lack of exposure to the . . . release.

► Where the . . . release occurs in a medium flowing in a single direction, such
as a river or stream, at least one control area upstream or upcurrent of the
assessment area shall be included, unless local conditions indicate such an area
is inapplicable as a control area.

RCG/Hagler Bailly

U6

► The comparability of each control area to the assessment area shall be
demonstrated, to the extent technically feasible, as that phrase is used in this
part.

► Data shall be collected from the control area over a period sufficient to
estimate normal variability in the characteristics being measured and should
represent at least one full cycle normally expected in that resource.

► Methods used to collect data at the control area shall be comparable to those
used at the assessment area, and shall be subject to the quality assurance
provisions of the Assessment Plan.

► Data collected at the control area should be compared to values reported in the
scientific or management literature for similar resources to demonstrate that the
data represent a normal range of conditions.

► A control area may be used for determining the baseline for more than one
kind of resource, if sampling and data collection for each resource do not
interfere with sampling and data collection for the other resources

Control areas were identified for this assessment. These control areas are described in greater
detail in the injury chapters and accompanying appendices.

1.2 REPORT ORGANIZATION AND SUMMARY OF FINDINGS

Soil, vegetation, and wildlife resources are ecologically inter-dependent (Figure 1-2) For
example, hazardous substances released from sources are transported in the air to soils.
Accumulation of these substances in the soil causes phytotoxic responses, including plant
mortality and reduced seed germination. This, in turn, eliminates vegetation required as
habitat by wildlife species and reduces the viability of wildlife populations. Because of these
relationships, the assessment of injury to these components has been conducted and reported
collectively.

Chapter 2.0 demonstrates that the Anaconda smelter was the primary source of hazardous
substances in the upland areas near Anaconda. Additional sources include waste piles and
ongoing fugitive emissions from waste piles and contaminated soils. Chapter 2.0 also
describes leaching, erosion, and biological uptake pathways by which soil and vegetation
resources have been and continue to be contaminated by hazardous substances.

RCG/Hagler Bailly

C/5

o
o

I—

3
O

CO

u

ei

c
.2

"s_
ea
Ou

re

H5

•c

JD

sj»

CO

X

,4) —

iS

§2

to ;£

^

«J

£

O.

5

C9

_

O

U

_

sj

c

in

i-

o

O
u

c/3

<D
O

>—

O

<D

*o

c

a.

D

c

o

re

re

JD

U

re

00

"O

--a

c

"O

(0

~

a.
D

£

"3

(J

"3d

o

"o
u
W

«s

-

OX)

re
CQ

re

O

u

Chapter 3.0, together with two related appendices, Assessment of Injury to Soils
(Appendix A) and Assessment of Phytotoxicity (Appendix B), presents the determination and
quantification of injury to upland soils. The results of the assessment of injury to these soils
demonstrate that:

► 17.8 square miles of surface soils in the uplands surrounding Anaconda contain
concentrations of the hazardous substances arsenic, cadmium, copper, lead, and
zinc that are significantly elevated relative to baseline.

►• Concentrations of hazardous substances decrease with depth, indicating a recent

and common origin of the hazardous substances.

► Concentrations of hazardous substances decrease with distance from the
Anaconda smelter, confirming that it was the principal source of hazardous
substance releases to upland soils; fugitive emissions from exposed soils
confirm that re-releases are ongoing.

► Controlled laboratory studies of phytotoxicity using standard laboratory test
species grown in exposed soils confirmed that they are phytotoxic and hence
are injured. Phytotoxic effects included reduced seed germination rates (death),
and reduced growth of roots and shoots.

► Concentrations of hazardous substances in impacted soils are correlated with
the degree of phytotoxicity determined in laboratory phytotoxicity tests.

► Injury to soils is further confirmed by injuries to vegetation communities, as
quantified in field studies.

Chapter 4.0, together with Appendices C and D, describe the degree and extent of injury to
upland vegetation and wildlife (through the loss or degradation of wildlife habitat). The
results of the assessment of injury to vegetation and wildlife habitat indicate that:

► 17.8 square miles, including portions of Stucky Ridge, Smelter Hill, and the
Mount Haggin area, are extensively devegetated or support impoverished
vegetation communities.

► Throughout the 17.8 square mile injured area, a community-level phytotoxic
response has occurred. This is characterized by reductions in or complete loss
of vegetative cover; decreased cover of native plants and increased cover of
invasive weed species; and shifts in plant community types from
predominantly forests with open grassland to predominantly sparse grassland or
bare ground.

RCG/Hagler Bailly

1-9

► The observed community-level response is consistent with the effects of
accumulated metals and arsenic in surface soil horizons. The response is
inconsistent with known effects of fire, logging, grazing, and S02 deposition on
vegetation communities.

► Vegetation reduction and loss (measured as percent bare ground) was positively
(and statistically significantly) correlated with degree of phytotoxicity and
concentrations of hazardous substances.

► Vegetation community complexity and habitat availability and suitability
(measured as number of habitat layers) were negatively (and statistically
significantly) correlated with degree of phytotoxicity and concentrations of
hazardous substances.

»• The structural complexity normally provided by indigenous upland vegetation

communities in southwest Montana has been significantly reduced in the
Anaconda uplands. This structural simplification has reduced the available
habitat for wildlife species.

► Habitat for wildlife species that depend on the two main indigenous habitat
types (conifer forest or forest/grassland edge), is significantly reduced in
suitability and availability relative to baseline.

► The loss of suitable habitat (including protective cover and forage) has resulted
in decreased viability of wildlife populations.

► The ability of the injured upland areas to revegetate and provide wildlife
habitat is precluded by phytotoxic levels of hazardous substances in soils.
Throughout much of the injured upland area, current injuries are likely to
persist indefinitely without restoration.

Chapter 5.0 briefly describes sources of hazardous substances to riparian resources. These
sources are described in detail in the Aquatic Resources Injury Assessment Report (Lipton et
al., 1995). Upstream sources associated with mining and mineral-processing operations in
Butte and Anaconda are identified as the sources of hazardous substance releases to riparian
resources.

Chapter 6.0, together with Appendices A and B, presents the determination and
quantification of injury to riparian soils. The results of the assessment of injury to soils
indicate that:

► Some 800 acres of floodplain soils along Silver Bow Creek, 215 acres on the
upper Clark Fork River, as well as the approximately 3,400 acres of the
Opportunity Ponds, contain the hazardous substances arsenic, cadmium, copper,

RCG/Hagler Bailly

MO

lead, and zinc at concentrations that are significantly elevated relative to
baseline.

► Controlled laboratory studies of phytotoxicity using standard laboratory test
species and hybrid poplars (a surrogate for native riparian willows and
cottonwoods) grown in exposed soils confirmed that these soils are severely
phytotoxic and hence are injured. Phytotoxic effects included reduced seed
germination rates (death), and reduced growth of roots and shoots. In addition,
riparian soils throughout this injured area demonstrate pH values less than 4.0,
and hence are injured.

► Injury to riparian soils is further confirmed by field studies demonstrating the
almost complete absence of vegetation on impacted floodplains.

Chapter 7.0, together with Appendix C, describes the degree and extent of injury to riparian
vegetauon and wildlife (through the loss or degradation of wildlife habitat). The results of
the assessment of injury to vegetauon and wildlife habitat indicate that:

► Some 1,000 acres of floodplain soils along Silver Bow Creek and the upper
Clark Fork River, as well as the approximately 3,400 acres of the Opportunity
Ponds, are almost completely devegetated or support impoverished vegetation
communities.

► Throughout these areas, a community-level phytotoxic response has occurred.
This is characterized by reductions in or complete loss of vegetative cover;
decreased cover of native plants; shifts in plant community types from
predominantly riparian forest/shrub-communities and agricultural use to
predominantly grassland or bare ground; and loss of habitat layers.

► The structural complexity normally provided by riparian vegetation
communities has been significantly reduced. The result of this structural
simplification is a reduction in the available habitat for wildlife species.

► Habitat for wildlife species that depend on riparian forest/shrub communities is
significantly reduced in suitability and availability relative to baseline.

► The loss of suitable habitat (including protective cover and forage) has resulted
in decreased viability of wildlife populations.

► The ability of the injured riparian areas to revegetate and provide wildlife
habitat is precluded by phytotoxic levels of hazardous substances in soils.
Throughout much of the injured upland area, current injuries are likely to be
persist indefinitely without restoration.

RCG/Hagler Bailly

1-11

Chapter 8.0, together with Appendix E, determines and quantifies injuries to piscivorous
furbearers, specifically otter, mink, and raccoon. The results of this analysis demonstrate that:

► Piscivorous furbearers have been, and continue to be, exposed to elevated
concentrations of hazardous substances in their diets — as demonstrated by
tissue analysis of representative prey and by significantly elevated tissue
concentrauons in mink trapped from the Clark Fork River relative to those in
minks trapped at control sites.

► Dietary exposure concentrauons of lead exceed a safe ussue residue criterion.

»• Populations of otter, mink, and raccoon are significantly reduced relative to

baseline conditions. Otter are completely absent from Silver Bow Creek and
the Clark Fork River from Warm Springs Ponds to Milltown. These population
reductions are not caused by human disturbance, trapping, land use, or other
human impacts.

1.3 REFERENCE

Lipton, J., H Bergman, D Chapman, T. Hillman, M. Kerr, J. Moore, and D. Woodward.
1995. Aquatic Resources Injury Assessment Report, Upper Clark Fork River Basin. Prepared
by RCG/Hagler Bailly for the State of Montana, Natural Resource Damage Litigation
Program. January.

RCG/Hagler Bailly

2-]_

2.0 SOURCES AND PATHWAYS OF HAZARDOUS SUBSTANCES TO
UPLAND RESOURCES

This chapter describes the sources of hazardous substances to upland terrestrial resources.
Sources of hazardous substances to riparian resources are described briefly in Chapter 5.0 and
in greater detail in Lipton et al. (1995).

The data presented in this chapter support the following conclusions:

► Air has been exposed to hazardous substances through smelter emissions and
ongoing releases from exposed soils and waste piles.

► Air has served as an ongoing pathway to terrestrial resources.

*■ Areal patterns of soil contamination are indicative of a point source located in

the Anaconda area

*■ The vertical distribution of soil contamination demonstrates that hazardous

substances have been released by a surface source rather than reflecting
geologic parent material.

*■ The areal extent of exposure exceeds 100 square miles.

2.1 SOURCE/PATHWAY SUMMARY

Smelting and ore refining began in the Anaconda area in 1884 (Tetra Tech, 1987). From that
time until the closure of the Anaconda smelter in September 1980, the smelters owned by
ARCO and its predecessors operated nearly continuously. Smelter emissions, transported by
air pathways, are a primary source of hazardous substances for hundreds of square miles of
surface soils. Mining and mineral-processing wastes disposed of in the Anaconda area are
additional sources of hazardous substances These waste piles and contaminated soils are
sources of continuous and ongoing releases of hazardous substances to this day. The
principal pathway by which hazardous substances have been transported from the smelter
stack and waste deposit sources to upland geologic and biologic resources is the air (either
directly from the smelter stack or through aerial re-entrainment of hazardous substances from
unconfined waste piles and contaminated soils).

RCG/Hagler Bailly

'

?.?

2.2 AIR PATHWAY: EXPOSURE OF AIR TO HAZARDOUS SUBSTANCES

To determine that air has served as a pathway, it was determined that the air resource has
been exposed to the release of hazardous substances [43 CFR § 11.63 (d)].

2.2.1 Smelter Stack Emissions

Air has been exposed to hazardous substances by direct contact with smelter stack emissions.
Stack emissions included compounds of arsenic, copper, cadmium, lead, and zinc (Tetra Tech,
1987). Emissions content varied over the 96 years of operation, changing with the rate of
copper production, smelting technology, and efforts at emissions control (Tetra Tech, 1987).
The smelting of these ores in Anaconda resulted in oxidation and release of oxides of arsenic,
cadmium, copper, lead, and zinc, as well as sulphurous compounds. Smelting thus released
hazardous substances to the air. Estimates of emissions are summarized below:

► A study conducted in 1907 measured the average daily release from the main

chimney in Anaconda at 59,270 pounds (29.65 tons) arsenic trioxide, 4,340
pounds (2.17 tons) copper, 4,775 pounds (2.39 tons) lead, and 6,090 pounds
(3.05 tons) zinc (Harkins and Swain, 1907).

»■ Between 1911 and 1916, the average arsenic discharge from the smelter ranged

from 40 to 62 tons per day (Anaconda Smelter Smoke Commission, as cited in
Taskey, 1972).

*■ Between 1914 and 1918, arsenic emissions were estimated at 75 tons per day

(Wells, 1920, as cited in Taskey, 1972).

►• In 1962, the arsenic content of air in the town of Anaconda averaged 0.45

ug/m3, and was among the highest concentrations in the country (Montana
State Board of Health, 1962, as cited in Taskey, 1972).

The Anaconda Minerals Company kept monthly records of main stack emissions during the
1970s (B. Raisch, Montana Air Quality Bureau, pers. comm). The data from one of the
reports (October 1976) are presented in Table 2-1. These data show that in October 1976,
some 190 tons of arsenic and 160 tons of copper were released to the air. Information
regarding main stack emissions is available for all months during 1974, eight months in 1975,
one month in 1976, three months in 1979, and two months in 1980 (see Figure 2-1 and
Appendix A). Overall, these data demonstrate that air was exposed to hazardous substances.
In addition, the U.S. Environmental Protection Agency (U.S. EPA) collected emission samples
downwind from the baghouse, installed in 1975, and found that elemental arsenic was 60% of
the mass of fine particulates released (U.S. EPA, 1983; Tetra Tech, 1987).

RCG/Hagler Bailly

Table 2-1

Record of Main Stack Emissions for the Period Between October 1 and October 31, 1976

Constituent

Assay

Quantity (total tons)

Copper

14.1%

160.022

Lead

4.88%

55.383

Zinc

5.91%

67.073

Arsenic

16.74%

189.983

Sulfur

9.0%

102.141

Bismuth

0.59%

6.695

Antimony

0.36%

4.426

Cadmium

0.46%

5.220

Tellurium

0.29%

3.291

Insoluble

14.4%

163.427

Silver

9.44 (oz/ton)

10,713.550 (total oz)

Gold

0.0575 (oz/ton)

65.257 (total oz)

Source: Report signed by Morris W. Bowman, Chief Environmental Engineer, The Anaconda

Company, Montana Mining Division, December 21, 1976 (Bowman, 1976).

500

400

x

c
o

^ 300
•a

E

tc
o
E

V
w

t-

<

200 -

100 -

/ •/•

M*

rt*

d3

<&

#

<$>

tf

/"

N41

NQ

V

^

s*

v?'

<>

.S*

X>

■c-

x»

<v

"*>

tf>

■*■

*

*•

*!?>

Figure 2-1. Arsenic Emitted from the Anaconda Smelter Main Stack, 1974-1980.

RCG/Hagler Bailh

2-4

Engineering controls to reduce emissions were added in 1971, 1975, and 1976 (Table 2-2),
and together with reduced production beginning in 1977, resulted in an exponential decline in
the release of arsenic (and other hazardous substances) (Figure 2-1). Since no major
emissions control modifications appear to have been made between 1924 and 1971, releases
of hazardous substances during the 96 years of operation were most likely substantially
greater than those indicated by 1970s data.

2.2.2 Fugitive Emissions

Air continues to serve as a pathway through re-entrainment of previously deposited hazardous
substances. Exposure of air to these fugitive emissions from unconfined waste sources has
been confirmed at numerous sites associated with the Anaconda smelting operations A U.S.
EPA Record of Decision (ROD) issued in 1987 ordered the removal and relocation of the
town of Mill Creek and its inhabitants, in part because of fugitive emissions of flue dust near
the smelter and from adjacent highly contaminated soils (U.S. EPA, 1987).

The disposal of large volumes of waste from mineral processing in Anaconda in tailings
ponds, waste piles, and structural fill areas has resulted in large source areas for fugitive
emissions releases. In the Old Works/East Anaconda Development Area Operable Unit,
waste sources include the Old Works structural areas, flues 1-6, waste piles 1-8, heap roast
slag piles, Red Sands waste, (former) Old Works Waste Ponds, Warm Springs Creek
floodplain, railroad beds, soils in disturbed areas contaminated with mixed deposits of
processing wastes, flue dust piles, and physically undisturbed soils that have been impacted
by smelter emissions (ARCO, 1992; PTI, 1992a).

In the Smelter Hill Operable Unit, waste sources include soils in the primary Handling/
Process/Storage areas to depths exceeding 48 inches, and soils in the East Anaconda Yard
(location of the main entrance, a sulfuric acid plant, a crushing plant, a brick plant, and a
railroad yard during the period of operation, and the Bradley Ponds, which received sludge
from the flue gas scrubber during smelter operation) (PTI, 1991a). Hazardous substance
concentrations increase with depth in the East Anaconda Yard, reaching maximum
concentrations measured at depths greater than 48 inches. Where reclamation has taken place
(covering with 18 inches of imported soils), metals concentrations are lowest. The coal pile
tracks, railroad beds, and soils in the main stack area also contain waste materials and exhibit
extreme concentrations to depths of 48 inches or more (PTI, 1991a). Reclaimed soils, wastes
underlying reclaimed soils, and all unreclaimed soils impacted by smelter emissions during
the period of operation also serve as sources of hazardous substances to upland resources.
Evidence of exposure via fugitive emissions includes data from 24-hour average
concentrations of total suspended particulates, arsenic, cadmium, and lead in Anaconda air.
Four stations were monitored between April 1984 and March 1986; concentrations ranged
over two or more orders of magnitude (compare maximum and minimum columns in
Table 2-3) (Tetra Tech, 1987).

RCG/Hagler Bailly

2-5

Table 2-2
Anaconda Smelter History

1 884 Upper Works begins operations, 26 reverberatory furnaces with individual stacks,

capacity = 500 tons of ore/day.

1886 Upper Old Works capacity doubled (1,000 tons ore/day). All furnaces connected to a

200 ft stack.

1888 Lower Old Works begins operating. 29 reverberatory furnaces connected by flues to 3
stacks. 800 tons ore/day capacity.

1889 Lower Old Works destroyed by fire, rebuilt. Total capacity of Old Works : 4,000 tons
ore/day.

1890 Converter plant built between Upper and Lower Works.

1 894 Second converter built east of Lower Works, connected to a single stack.

1902 Washoe Smelter begins operations, 14 furnaces, 12,000 tons ore/day capacity.
Roasters, blast furnaces, and converters connected by flues to four separate stacks.

1903 Upper and Lower Works closed. New flue system installed and connected to single
300 ft. stack.

1918 Cottrell electrostatic precipitators installed on main flue. Precipitator dust sent to
arsenic recovery plant.

1919 New 585 ft stack completed on main flue.

1924 Selective flotation process begins. Reduced volume and temperature of flue gases

entering Cottrells allows increased recover.' of flue dust and reduces SO: emissions.

1942-1969 Three H:S04 (sulfuric acid) production plants built.

1961 Concentrate brought from Weed concentrator in Butte. Washoe concentrator use

becomes intermittent.

1964 Arsenic recovery plant closed. 1,000 tons flue dust/month shipped to ASARCO in

Tacoma, WA.

1971 Steel balloon flue built, gas and dust collection improved. Spray chambers built,
temperature control of flue gas improved.

1972 Flue dust shipments to ASARCO cease.

1973 Acid plant built. S02 emissions reduced.
1974-1977 Arbiter plant operates.

1975 Baghouse dust collector built, particulates emissions reduced. Converter hoods built,
fugitive gas emissions reduced.

1976 New flue from electric furnace to baghouse built.

1977 Reverberatory furnaces closed. Two maintained on standby status. Cottrells closed.
1980 Washoe Smelter closed.

Source: Tetra Tech, 1987.

RCG/Hagler Bailly

sD
i

E ^

3 n

||

ee 3.

so

rs

©

CS

os

©

o

Os

©

w-1
CN

oc

(N

fN

OO

o

rs

c

o

rs

©'

fl-
rs

ro

s w

«

«

Q

u

E M

"5.

3 j

^^

o
o

o
o

©

— M

o

O

o
o

o
o

©

rs

©

£

©

©'

©

O

o

o

©

©

03
«5

s~

U

E

^

u

01

e

CS

>
■

u

DC
X

1

• - ci
W 3.

E w

o

rS

o
o
©

Os

©

o

©

rs

©
©

ro

oe

O

o

o
o

o
©

©

fl-

©
©

•fl-

©

C3

j=

u

«*

1_

<

C/3

■B

EL
E

"S3

■5

E

W2

VO

Im

fl-

ee

00

oo

SO

r~

in

o

©

o

O

OS

e

m

—

—

•—

~—

t^

m

rs

rs

rs

■fl-

SO

a

^™

—~

^M

^™

rs

^«

^—

■»■

^M

^m

u

i o

u

E

s

H «

*rf

H

_s

|_

fl-

sc

so

so

SO

«i

O

VO

vo

VO

VO

e

oe

oc

oe

oc

oc

oc

OC

oc

OC

oc

v~t

u

u

at

OC

OC

«9

r~

t~~

r-»

t^-

^"

Vm

p^

F^

r^

f^

^»

^*—

c
e

rS

<N

rs

(N

SO

'*■

fS

rs

rs

rs

vS

■A

>—

m

r-i

m

ro

1

i

m

c>

m

<-•">

53

e

i

■

i

i

•fl-

TJ-

■

■

i

■

£

■*

■fl-

■fl-

fl-

ee

OC

i-

•fl-

fl-

A-

■fl-

fl-

oc

oe

X

oc

OC

i

^*>

oc

oe

oe

oc

oe

i_

fl-

*!r

u.

"«

so

so

so

—

*mf

S

^B

SO

^™

^~

3

w

Efa

~

fl-

fl-

fl-

r^

r^

~~

A-

A^

fl^

i_

£

-

—

u

E

o

E

2

■g

Q_

'c

-

-a

a.

•o

a.

'c

3

-a

Cu

•a

a

CS

>

1/2

o

Vi

'e

re

oo

re
u

CO

u

(A

<

E

re

o

in

re
u

>>

r-

<

-a
re

U

J

f-

-J

H

-a
re
U

-J

H

75
s

™—

a

<

U

<

re

U

^

r--

e
>>

■a

e

CS

00

c

2
<

re

oo

Cs

es

E

b

u

3

o

■ —

o

"o

o

5T

o

o

o

u

E

u

c

3

-c
o

u

f-

s

E

trt

—>

tyj

oo

re

OO

_©

"c

>>

c

n

o

**

o

in

re

CS

oo

o

o

r-

OO

c

-=
o

o

c

u

c

u

u

u

3

O

oo

re
CO

a

2-7

The highly variable air quality is due in part to wind-induced fugitive dust levels. These data
confirm that air has continued to be exposed to ongoing releases of hazardous substances.

Ongoing exposure of air to hazardous substances through re-entrainment of wastes on Smelter
Hill is also confirmed by concentrations of arsenic, cadmium, copper, lead, and zinc
determined in dustfall samples. Thirteen dustfall samplers were placed downwind of source
areas at or near Smelter Hill property boundary lines in July 1989 (McVehil-Monnet, 1992a,b;
PTL 1991a). Maximum monthly concentrations of hazardous substances in dustfall through
March 1991, as reported in the Preliminary Site Characterization Information Report for the
Smelter Hill PJ/FS (PTI, 1991a), were 115,333 ppm arsenic, 10,800 ppm cadmium, 390,000
ppm copper, 51,333 ppm lead, and 199,677 ppm zinc. Maximum concentrations reported for
dustfall sampling in 1992 were substantially lower (Table 2-4) (ARCO, 1992; McVehil-
Monnet, 1992a,b). All concentrations reported (PTI, 1991a; ARCO, 1992; McVehil-Monnet,
1992a,b) demonstrate a continuing release of hazardous substances from Smelter Hill, and a
potential for exposure and transport of hazardous substances away from Smelter Hill.

The Anaconda Soil Investigation (ASI) has identified patterns of high surface soil
concentrations of arsenic (exceeding 600 ppm) and copper in soils north and northwest of the
Opportunity Ponds (PTI, 1992b). These high soils concentrations have been attributed to
fugitive emissions from the Opportunity Ponds based on the prevailing wind directions in the
Deer Lodge Valley and proximity of the contaminated soils to the Opportunity Ponds (PTI,
1992b). In addition, in the Old Works area, Red Sands (a waste deposit) material has been
detected in soil samples collected west and southwest of the sewage treatment ponds, and in
soils adjacent to the Red Sands waste pile, suggesting wind or water erosion as pathways of
transport (PTI, 1992a). Arsenic concentrations exceeding 300 ppm in the extreme eastern
section of Anaconda have been attributed to prevailing wind patterns and close proximity to
the Old Works and Washoe Smelter sites (PTI, 1992b).

2.3 AIR PATHWAY: AREAL EXTENT OF EXPOSURE

2.3.1 Historical Accounts

The areal extent of exposure is defined as the geographical surface area or space where
emissions from the source of discharge or release are found or otherwise determined to be
present for such duration and frequency as to potentially result in injury to resources present
within the area or space [43 CFR § 11.63 (d)(3)(ii)]. The areal extent of air exposed to
concentrations and duration of emissions sufficient to have caused injury can be defined by
the areal extent of injured soils and vegetation. Historical accounts document the visible
reach of the smelter plume extending as far north as Garrison, south to Butte, and southwest
up Mill Valley to the Continental Divide (Swain and Harkins, 1908; Peirce et al., 1913):

RCG/Hagler Bailly

2-8

Table 2-4

Summary of Trace Metal Concentrations (ppm) in Dustfall Samples

Collected by 13 Dustfall Samplers

Month
1992

Arsenic

Mean Mai Mm

Cadmium

Mean Mai Min

Copper

Mean Max Min

Mean

Lead

Max

Min

Mean

Zinc

Max Min

Jan

2,423 10.986 43

72 183 3

7,911 12,700 413

1,144

4,564

23

2,772

3,686 388

Feb

2,516 12,100 520

62 210 20

6,975 18,800 1,460

1,408

3,015

380

7,315

16,100 2,470

Mar

672 2,350 9

25 85 1

2,016 4,890 49

388

1,150

9

4,011

8,099 236

Apr*

1,242 4,633 117

46 189 3

2,933 7,950 786

607

1,621

118

3,176

8,950 1,292

May

244 1,261 35

8 22 2

821 3,595 128

122

550

5

1,018

4,405 1 88

Note: Samplers were maintained as part of the Anaconda Smelter RI/FS Air Resources S

Program (McVehil-Monnet, 1992a,b).
* Data from 10 samplers only.

ampling

... the smoke stream can be traced as far as the eye can reach . . . trailing
down the valley for 30 miles toward Garrison, or often eastward in the
direction of Butte, or sweeping over into the Mill Valley and filling the narrow
ravines which lead down from the Continental Divide, 14 miles to the south.

The Mill Valley district southwest of the smelter is the one toward which the
smoke blows most during the early summer, while in late August the air
currents begin to go northward down the Deer Lodge Valley, and from this
time until the snow covers the ground the greater part of the smoke blows in
this direction (Swain and Harkins, 1908).

Smelter fumes could be detected as far west as Georgetown Lake, and 25 miles southwest,
across the Continental Divide (Peirce et al., 1913). Farmers and ranchers in the Deer Lodge
Valley reported heavy dust accumulations on floors and rafters of barns. The dust was
subsequently determined to have a range of AsjOj from 410 to 9,190 ppm (310 to 6,957 ppm
As) (Swain and Harkins, 1908). Peirce et al. (1913) reported a pattern of visible injury in
forests, including poor needle retention, slow growth, and absence of regeneration, extending
19 miles north, 12 miles east, 10 miles south, 25 miles southwest, and 19 miles west of
Smelter Hill.

RCG/Hagler Bailly

2-9

2.3.2 Areal Patterns Indicative of a Point Source

Measured levels of arsenic, cadmium, copper, lead, and zinc in soils and plants in the Deer
Lodge Valley serve as proxy measures of past aerial loadings, and as indicators of dispersal
distances of stack emissions. As receptors of hazardous substances, soils and biota also
define the extent of injury caused by hazardous substances transported in air.

Previous studies have shown that, in general, concentrations of arsenic, cadmium, copper,
lead, and zinc decrease with increasing distance from the smelter (Taskey, 1972; Munshower,
1972; Munshower, 1977). For example:

► Harkins and Swain (1907) measured arsenic concentrations in plant tissue and
found arsenic concentrations in grass and hay grown as far as 35 miles from
the stack were two to four orders of magnitude (i.e., 100 to 10,000 times)
higher than in hay and grass grown 100 miles northwest of the smelter

► Taskey (1972) found that arsenic, copper, lead, and zinc in the soils near
Anaconda were concentrated within the upper 8 inches of the soil profile, and
that the highest concentrations were generally within 5 miles of the Old Works
and Washoe Smelter sites.

However, because of the topographic complexity of the terrain surrounding the smelter,
concentrations of hazardous substances do not decline monotonically with distance:

► Haywood (1910, as cited in Taskey, 1972) found a range of copper
concentrations in soils increasing from 485 ppm 9.5 miles northeast of the
smelter to 2,791 ppm 12 miles northeast of the smelter.

► Taskey (1972) demonstrated that soils 24 miles north of the smelter possessed
greater trace metal content than soils half that distance from the stack, in the
same direction, and he identified soils on prominent terrain with unusually high
concentrations of hazardous substances. The nonlinear distribution pattern was
similar to that reported by Haywood (1910, as cited in Taskey, 1972) for
copper.

Munshower (1972) analyzed Deer Lodge Valley grassland soils for cadmium at 53 sites along
a 27-mile-long transect extending northeast from the smelter, and two east-west transects 9
and 21 miles northeast from the smelter. Cadmium content in surface soils increased
gradually from 2 ppm to the south of Deer Lodge to approximately 3 ppm within 7 miles of
the smelter, then rapidly to 30 ppm within 1.5 miles of the smelter (Munshower, 1972)
(Figure 2-2). The remaining variation was attributed to fugitive emissions from dry tailings
ponds and slag heaps. Cadmium was shown to be elevated in soils relative to background
levels (0.2-0.6 ppm) as far as 37 miles from the smelter (Munshower, 1977).

RCG/Hagler Bailly

■10

Cd cone.
IPS") |6

Y«.S/X'+1.9

» 12 15 18 21 2* 27

Dtstonc* from Sm«llc* |nulei)

Figure 2-2. Distribution of Cadmium (ppm) in Soil with Distance from the Main
Smelter Stack. Source: Munshower, 1972.

Within a 5 ppm cadmium isopol, low soil pH values and high concentrations of heavy metals
in the soil were determined to restrict plant community development and germination
(Munshower, 1972). Observation of vegetation grown in soils 2 miles from the smelter
revealed reduced seed germination rates and restricted root development in plants that
survived the seedling stage (Munshower, 1972) In addition, cadmium was determined to be
accumulating in forb species and herbivore species to concentrations exceeding the soil
concentrations (Figure 2-3). The accumulation of cadmium in herbivores was attributed to
both ingestion of contaminated plants and respiratory intake (Munshower, 1972).

Recent sampling efforts have determined patterns of soil concentrations of hazardous
substances in proximity to the smelter, most samples have been collected from the Anaconda
Smelter site, the Old Works, Anaconda, Opportunity, and Mill Creek (e.g., Appendix A, this
report, Taskey, 1972; AMC, 1983; Tetra Tech, 1985a,b, Tetra Tech, 1986, Tetra Tech, 1987,
CDM, 1987, CDM, 1988; PTI, 1990, PTI, 1991a,b; PTI, 1992a,b). Within 3 miles of the
Anaconda smelter, copper concentrations exceed 1,000 ppm, and arsenic concentrations
exceed 1,000 ppm within a 2-mile radius of the stack (Tetra Tech, 1987). The most
comprehensive study performed to date is the Anaconda Soil Investigation (ASI) (PTI,
1992b). The ASI covers a 120 square mile area, but does not determine the spatial extent of

RCG/Hagler Bailh

2-11

4

Cd
coot- e

Wiatd C'o-,i c*d (orb Cd

i if

{

B 10 1J U 16 18 20 22 2* 26 100

Dmc-vff from SmeJre* (mtlvi)

Figure 2-3. Cadmium Concentrations in Vegetation and Grasshoppers in 1970. The

control site (far right on the x-axis) was 100 miles from the main smelter stack.
Source: Munshower, 1972.

elevated metals concentration. Contoured1 metals concentrations (PTI, 1992b) have shown
that the dispersal of metals from smelting in Anaconda is approximately radial with slight
elongation to the northeast, possibly reflecting the dominant downwind direction. However,
no samples were collected as part of the ASI from the uplands north, west, and southwest of
the smelter to identify possible elongation in other directions.

Figure 2-4 shows the location of transects used to describe soil exposure to hazardous
substances based on existing data; plots of metal concentration with distance from the smelter,
based on data from Taskey (1972), Tetra Tech (1987), and PTI (1992b), are shown in Figures
2-5 through 2-9. Note that the baseline values presented in Figures 2-5 through 2-9 (Brick
and Moore, 1992) are mean values of baseline concentrations for U.S. and world soils
reported in the scientific literature (Kabata-Pendias and Pendias, 1984; Adriano, 1986;
Alloway, 1990; CH2M Hill et al., 1991). Where a wide range of values was reported, an
average value was used, or, in the case of lead, 15 ppm was used because that number was
cited by two different authors (Brick and Moore, 1992).

Contouring of metals concentrations is similar to contouring land surfaces by elevation. Lines of
equal concentration (isopleths) are drawn to demarcate the patterns of soil contamination.

RCG/Hagler Bailly

2-12

/

/

/

/

/

/

West Vallcv Transect

\

\

\

Six Mile Radius C^

\

/

\

/

Twenty Mile Radius

X

Figure 2-4. Location of Transects in Figures 2-5 through 2-9.

RCG/Hagler Bailly

2-13

E

4

Q.

Q.

1000-

i o

C

o

£

ro

4

©

G

D

O

c
o

100-s

*

4

*

*

1 ,

O

>^

d

o

,

i

i.

Oi

10-

G

▲

c

i i i

) ;

> ^

l (

1 I 1

5 i

1 1 1

3 1

0 1

2 1

4 1

6 1

8 2(

<— Pb Baseline
<— As Baseline

Distance from Smelter (miles)

10000 —

A CU
G Zn

4
4

£

D

E

Q.

S iooo-

c
.2
to "

c

CD
O

c

o 1 oo _

A
G

1

A
G <

4

(

-\

&

G

q IUU -:

i

L

<— Zn Baseline

■ O— Gii Baseline

o :

A

10-

1 1 i

-rrr

1 1 I

1 1 !

1 1 i

^TT-

1 1 1

1 1 1

i 1 1

0 2 4 6 8 10 12 14 16 18 20
Distance from Smelter (miles)

Figure 2-5. Northwest Transect. Source: Brick and Moore, 1992

RCG/Hagler Bailly

2-14

10000-

E

Q.

— 1000

c
o

CO

c
a>
o
c
o
O

'6

100

10

£

V

.

i

L

-

C4

6

A

<^

;

k

k

G

▲

C

D

p

«

w

A *

S?-

i

,

•

▲

▲

-

▲

■

O

1

0 '<

> ^

I (

3 t

I 1

1 » i "^T™

0 12 14 16 18 2(

Distance from Smelter (miles)

<— Pb Baseline
<— As Baseline

E
a.

C

o
ro

"E

CD

o

c
o
O

"6
c/>

uuuu —
1000-

c

AA

TO,

▲

c

▲

1

A

)Q

5

ft

A

100 -

M

10-

i i i

TTT-

i 1 1

■ i i

i i i

<— Zn Baseline
<3— Cu Baseline

0 2 4 6 8 10 12 14 16 18 20
Distance from Smetter (miles)

Figure 2-6. Northeast Transect. Source: Brick and Moore, 1992

RCG/Hagler Bailly

2-15

10000^,

E

Q.

s iooo-

o

CD

-
o

o

"6

100

10.

a-

c

1

G

▲

c

3

3

rrrr

▲

£

iii iii iii

Pb Baseline
• As Baseline

23456789 10
Distance from Smelter (miles)

E

Q.
Q.

C

o

to

c

o
c
o
O

'6

;

000-

2

i

J
v

' ft

a

▲

100

C

t

▲
G

10-

i i i i

i i i i

|i in

!""

i i i i

[iii

<— Zn Baseline
<— Cu Baseline

0123456789 10
Distance from Smelter (miles)

Figure 2-7. Southeast Transect. Source: Brick and Moore, 1992

RCG/Hagler Baillv

2-16

10000^

E

Q.

& 1000^

c
o

CO

i—

c
a>
o
c
o
O

"6

CO

■*

Po

100

10

«.

1 1 1 1

Ik

n

▲

D .ajS

ioG

CD

3 A

P

A_

G

'i i 1 1

1 1 1 1 1 1 1 i
2 3 4 5 6 7

Distance from Smetter (miles)

TT-

<— Pb Baseline
<— As Baseline

10000

E

Q.

— 1000

c
o

CO

c

0)

o
c
o
O

100

o
CO

:

Ao

c%

i

▲

i

C i

i

A

:

" c

* c

3

o

^c&
A c

4

■ — '

A

r

▲

o
o

a

(

I II II

) '

1 1 1 1

1 1 1 1

1 1 1 1

III!

£

Till

Mil

TTTT")

r 8

<— Zn Baseline
<— Cu Baseline

Distance from Smelter (miles)

Figure 2-8. Mill Creek Transect. Source: Brick and Moore. 1992

RCG/Hagler Bailly

2-17

1000-

E
a.
a.

c
o

75

^ 100

a
o

c
o
O

'5
CO

Qk

Q

▲
o

10.

o
▲

o

8 10 12 14 16
Distance from Smelter (miles)

<— Pb Baseline
■m-J <— As Baseline

10000

E

D.
Q.

~ 1000-

c
o

ro

k_

c
<x>
o
c
o
O

"6

CO

100

10

k

3

9;.o

o

▲

<— Zn Baseline
<— Cu Baseline

i i i
8 10 12 14 16

0 2 4 6

Distance from Smelter (miles)

Figure 2-9. West Valley Transect. Source: Brick and Moore, 1992

RCG/Hagler Bailly

2-18

Concentrations of arsenic, cadmium, copper, lead, and zinc in surface soils (0-2 inches)
collected as part of this assessment (see Chapter 3.0 and Appendix A) were analyzed to assess
the relationship between concentration and distance from the stack. Soils were collected from
upland, mountainous terrain north (Stucky Ridge), southwest (Smelter Hill and Weather Hill
area), and south (Mount Haggin) of the smelter, as well as from matching control areas
located south of the smelter in German Gulch. The exponential decline in concentration with
distance from the stack for all five elements is illustrated in Figures 2-10 through 2-14.
Statistical analyses confirmed that all five relationships exhibit a significantly negative slope
(p = 0.0002), indicating that soils concentrations decrease with distance from the stack. Such
a pattern is typical of contamination emanating from a point source.

The exponential decline with distance illustrated in Figures 2-10 to 2-14 is commonly
associated with main point sources of industrial pollution, including metal smelters
(Figure 2-15). For example, "halos" of contaminated soils, typically pronounced in a
downwind direction, have been observed surrounding a copper smelter in Tacoma,
Washington, a gold smelter at Yellowknife, Canada, a zinc smelter in Pennsylvania, and the
copper-nickel smelting complex at Sudbury, Ontario, Canada (Kabata-Pendias and Pendias,
1992; Baker, 1990, O'Neill, 1990). At Sudbury, soils enriched with copper extend some 25
miles from the point of release, and within 4 miles of the smelter, soil copper concentrations
in excess of 1,000 ppm are common (O'Neill, 1990).

2.3.3 Soil Profile Concentrations Indicative of a Surficial Source

Soils collected from the 0-2 inch and the 0-6 inch layers from exposed upland areas near
Anaconda were compared statistically to test for differences in concentrations of hazardous
substances (see Chapter 3.0 and Appendix A). The results confirm that concentrations of
arsenic, cadmium, copper, lead, and zinc are significantly greater in the upper stratum than in
the lower stratum (Table 2-5), indicating that the source of the hazardous substances is
surficial rather than attributable to the underlying parent (geologic) material. That even
cadmium and zinc (the most mobile of the elements analyzed, and typically rapidly leached
from surface soil horizons) are elevated in surface horizons is evidence suggesting a recent
and common origin of these hazardous substances.

2.3.4 Fugitive Emissions

Exposure of the air pathway continues through re-releases of hazardous substances via
windblown dust. Currently, unvegetated areas in the Anaconda area subject to wind erosion
include Smelter Hill, Mount Haggin Wildlife Preserve, the Old Works area, portions of
Stucky Ridge, the Anaconda Ponds, the Opportunity Ponds, and the entire flatland between
Willow Creek and Lost Creek east of Smelter Hill. The Anaconda Ponds cover
approximately 560 acres. The Opportunity Ponds are virtually barren, but in recent years

RCG/Hagler Bailly

0>

I

CO

^M

w

0

s-

0

CO

— >

u

0
u

c
0

C

E

U

0

c

CO

<i>

u

0

QO

«

3

W

~~

u

-o

£

c

E

-O

s

c

£

L.

M

X

£

u

BO

>% a>

CUJ

fO

>, co

BO

a;

3:

2

0

V

CO

3:

3

E

->_>

C/3

3

E

—J

cowSDmcoS

• o+

•

0

-t-

o

+

f

CO

cu
-i t- —

» - CD

+

+

+

+

• •

po

o

o

o o
o
o

o

0

- CO

in

CO

t\»

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

CO

CO

"*

OJ

0

co

CO

Tf

CM

cv

(uidd) oiuasjy

co T3

re re '—

re co

*Z re
w cu

CD o

CD rv

— —

— E
<u .r

fN

re c

o

c

re E

Q 6
. 2

■2-2

5 re

on .

c/3 13

c -

c
o
cu

■a

re
^ o

«o .5

an -c
c =:
re u

u a,

— re

C«_ -O

C re

•a £

CD E.
Q. D.

£ 73 => °

0

(0

CO

W "O

C

£ 2

CO

E ^

-i

0 0

E
0

<^ 3

v C/3

£ E

es 0

0)

0

V5 U-

c

Q ~°

CC

CD

— CD

.— CD

t/5 O

CD

re

u o

1 ^

~ ^

_ T3

^ CD

CU ~

re D

CD

CD _C

CO "-1

cu ~

•r » .- O

O CD co co

O cu

re

C

— CO

O —

"•5 '0

es co

- <->

<v c

u • —

5 CM

0 A

U °

cu

V.

E

Q.

< 3

o
0

_c re

_H C

3 o

o g

c =>

re i_

££

o c1-

,«- ""

1

~l

IS

m

o
u

o

(N
i

in

~^

t/i

n

— >

0

1/1
0

c

— »

o

u

-i->

E

o

0

c

0)

u

u

0

00

— *

3

M

~~

u

-o

5

5^

So'x

-o

e
"5b

>> oj

Qfl

(0

&^ CU

QO

CO

X

S

CO

3

E

«J

CO

3

E

«j

CO CO

.«=.

b co co

-i

• o +

•

o

+

+ '

• i -

+;

-H

+

•H

+

+

++:

+ !

+ :

i

+ •

+ «-

o

o

o

o

o

o

CT>

(D

r-

co

in

o

CO

—>
CO

C

'co

E
o

Si

OJ
CJ

c

(0

CO •-

CM

o

CO

o
in

o

o

CO

o

02

(uidd) uiniuipeo

—

o

~

O

re

c

■ 3

•■*

3

J;

—

• =

-a

o

s

c

D

o>

co

CD

=b

o

^

=n

"5

c

re

«

o

cu

>>

u.

•— i

SO

re

CO

—

c

5

c

re

o

C

3

i_

'5

O

cu

<

c

2

~*T

re

"3

C

•_

^■^

re

o

P

Q

re

-3

5.

v>

c

— ■

J£

r— ;

cu

^r

o

1—4

u

(N

K

i—

B.

o

cu

CO

3
CU

L.

"a3
E

in

o
o

mm

00

~^

ew

CO

o

cu
re

=

CO

CD

M

C/l

CU

"O

o

c

2

-3

re

"es

C

oo

s

>>

cu

T3

OJ

o

3

B

CU

JZ

~~

"S

*■*

00

•a

D

*■

E

cu

cu

©

o

c

,u.

re

—

U-

■q

_

cu

T3

cu

o

cu

CO

CO

C

Tj

H

re

CO

cu

'5

CO

cu

CO

5

CU

T3
CU

CO

CD

CO

re

-—

^__

C

'£

o

I—

g

SO

re

CO

c
o

re

c

"S

_^

o

o

CO

"5

CO

JS

CU

c

3

•™

.C

3

JJ

u

O

_

o

o

c

cZ.

• "~

c

re

■

o

re
E

-a
cu

«■

O

o

CU

cu

c

a

Cl
CU

E

L_

o

E

CO

U

CL '—

c
o

w

-D

re

_3

E

3

C

re

re.

E

en*
re
cu

cu
o

c

u

re

CU

re

o
o

„_■

—

1

ew

s

CM

-

K

—

(Q

O

u

~o

00

O

C

■~

c

<— 1

.*->

E

u

o

C

cy

cu

u

o

bC

"

3

to

~

U

XJ

£

c

E

-o

5

c

£

u

So

X

£

L,

'3d

>> CJ

ClO

0

>* o>

Of)

u

to
X

s

CJ

0J

to

3

s

—J

t/i

-

c

■»J

[flWSDWW

2

• o+

•

0

+

'.^

+

+_

+

+

+

-H-
+

+

• i

•

+

o

•

°+

o

o

•

o

o

o

OD

- OS

- CD

cu

CD

m

o
u

o

c

(0

CO — «

cu

o

o

o

o

o

o

o

o

o

o

o

■ o

o

in

o

m

o

lO

CO

CU

CM

«-^

«-H

(uidd) jaddoo

§■1 3

O re ■-

TO c/3
O CJ

cj r^j

c o

c —
u

— an
.= SO

O -

re =

Q o

- o re

— C d>

aj re
so

.2 J

cj F

.= CJ

-2 _

C/3 -

O S

re

-g j3

- u \^

._

- >-

E

s °

< — *<

CO CJ

x

— c

CJ

.r co

(0

c;

CO

5T U

c

9J ~U

— S

(0

*" ' '

S

I-&

<D

o o

tS 3 .E

CO

T3
CJ

CJ

c 'P

C E

3 C

CJ Q"

-= o

w o

c/i —
4)

13

o —

2 in

•- £

aj 2

E co

_ "3

T23 CJ

CJ — ■

re _)

T3

CJ

CJ J3

^ = o

c
u

o CJ

1/3 a.

s E

*— re

= w

o —

— '5

e; «

i .c

r <->

u c

(j • —

= rsi

U °

4J

H

U 3

■

u

u

3
M

en J2

aj ._

r: o

Or, M

o cj

II

° E

j3 re
3 5

a £

c =
re i_

E^

u -a

a «

Bi

O CL
U. CJ

CQ

00

n

a
o

oi

CN

CN

i

CN

ir.

—

</)

c

i-

0

V)

— -

_

o

3

— '

0

3

.4-J

£

U

O

3

CD

o

CJ

O

M

— |

3

OX

~

u

-o

X

c

£

-o

X

3

oS

s_

'So

x

£

S-

'5d

>> V

M

a

>,

a>

oo

£6

u

CJ

c
X

2

— >

3

E

_,

w

3

3

*->

CO

do s r> co to

S

• o +

i

•

O

+

o

+

+

+

+

+

+

+ • •

+ •

%

o

o

o
o

o

o

as

CO

CO

in 3

CO —

C\i

o
o

CO

o
o
in

o
o

o
o

CO

o
o

C\}

o
o

(uidd) pe8q

«^»

CD

<~s

CO

-o

CC

o
^

Q

u

B/a

cd

"S

o

o

^

[—

hj

»—

_

•o

^J

EJ

CJ

u

^~-

Cs!

co

cu

E

lo

o

EL

,_

—

~™

0)

*■

o

k_

SO

CJ

c/f

co

=0

c

re

E

re

o

■6

O

re

o

ti-

c

w

c_

re

o

E

CJ

T3

C

re

C3

2

CJ

5

CA

T3

u

re

CC

'5

CO

T3
C
CD

M

GO

^

co
CJ

1

re

CD

u

_

bC

re

,H

cu

30

_o

s

~

*a3

re

CJ

X

t/3

t/2

r-

^^

o

M

r™

ca

— >

^

CO~
30

CO

CD

a.
a.

o

CO

3

s

re

o

■"■

■3

t~-

S

>>

■ — '

0)

a

o

cu

CO

SJ

re

£

c/5

—

c/5

*_>

(a>

T3

E

01
u

5

CJ

T3
CU

o

—

en

J

•o

re
■a

cu

Efl

CO

CJ

3

5

o

CD

—

cu

CO

^z

zz.

'_

— >
00

5

co

o
cj

en
ca

co

CD

CO

O

CO

'5

■■*

c

__

CO

O

O

c/>

E

_

c

co

C

o

o

-o

©

'5

co

re
cu

o

30

^

c

a

o

3

i_

c

a

u

CN
i

o

c
re

E

3
CJ

o
O

©
U

c

CU

>^

■ ~~

o

~~*

re

g

E

_

o

13

c

ri

^*

r!|

0)

i.

M

Es.

cn

c

c

CO

~™

CO

o

U

0

CO

-_J

!_

o

c
o

c

E

o

o

c

<D

0)

o

0

us

3

M

'

u

"0 ;=

C

E

T3

X.

c

2 u

'5b

"S

£

u

'5b

>> 0)

Qfi

(0

>% 0)

00

CO

3C

S

o

— *

CO

3B

— !

CO

3

E

-*->

CO CO 2 ZD CO CO S

• o +

•

o

+

+

+

+

o ,

t

+

- t>»

+

+ •
+ •

o
• o

o

o

o

o

a>

CD

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

CO

t

cv

o

CO

CD

"<*•

CM

re

T3

re
C

cni

o
o

o

Cw

cu

„

CJ

-

cn

CS

CU

u.
C8

o

c/T

re

53

CO

CU

o
cj

•5

l_

—

c

cu

Q.

CD

CJ

C3

'n

Q.

L.

a.

c

T3

C/5

P

re

c

•*

. ~

CU

re

C

c

E

CO

o

re

Q.

"5o

c_

• —

00

O

"3

re

re

C

ffl

Q

•T*

CU
CO

c

CU

a-

e

re

—

re

•

a

re

-^

o

cu

J3

w

-a

JC

re

CO

cu

CO

o

^

c

^— V

E

CU

re

c
cu

*^^

"ftj

•^

C

a

2

X

—

cu

V)

o

CO

_

c_

(N

c0

CL)

c_

—

—J

s

4— *

3

CD W

C

'5

E

« 2

o

I*

CI)

o

C

(0

co .2
Q

C\!

<£ rn *"'

co
cu

- cu

— £0

_ -a
§2

*£

cu cj
U 3

% £>
5 p

CO

CU en

— CU

re —

cu re

'■B co

E "5

— 5

<u cu

re t»«

T3 O

— -a
% 2

"* cj

'© "S

co "

.5 ""

'Z s.

5 E

•— re

■_ —

- o

— en

cu

CJ JZ.

- CJ

O C

H o

on

cu *J

."£ re

CO c

1

re

c
o

CJ

c

3

_ 3

CN

= 6

5 .£

c *a

re «j

E c

CU C-

a 2

re

(ujdd) oui2

cu

-
=
esc

il

a
u

2-24

Distance from smelter (km)

Figure 2-15. Smoothed Patterns for Copper Concentrations in Soils as a Function of
Downwind Distance from a Copper Smelter (mean values of 3 years).

t = total concentration, s = soluble concentration (Kabata-Pendias and Pendias,
1992)

Table 2-5

Comparison of (1-2 Inch and 0-6 Inch Soil Concentrations:

Upland Areas Near Anaconda (n=12)

Element

p-value1

Arsenic

0.0049*

Cadmium

000073*

Copper

0.00049*

Lead

0.00024*

Zinc

0.00073*

l

Compared using Fisher's Paired Permutation Test (Manly, 1991).

*

Indicates significantly greater concentrations in the 0-2 inch soil laver at a =

- 1"/

- I /o.

RCG/Hagler Bailly

2-25

erosion has been reduced by application of a surfactant. The Opportunity Ponds cover
approximately 3,400 acres. There are also 173 acres of slag disposal, 107 acres of
nonvegetated rocky barrens, and 498 acres of "industrial disturbance" including former
locations of the smelter buildings on Smelter Hill (PTI/MNRIS, 1992).

2.3.5 Areal Extent - Summary

The data described above demonstrates that air in the Deer Lodge Valley has been exposed to
smelter and fugitive emissions and has served, and continues to serve, as a pathway of
hazardous substances to geologic and biologic resources. Soils exposed by aerial deposition
during the 96 years of smelting now are a reservoir of accumulated hazardous substances
(Munshower, 1972). The extent of airfall contamination is highest around the Anaconda
smelter, as evidenced by soil concentrations within a 6-mile radius (approximately 110 square
miles) (Brick and Moore, 1992). In this area, concentrations of metals are approximately 10
times background concentrations, on average. The total area of exposed and contaminated
soils is probably closer to 1,260 square miles, which is approximately the area of a 20 mile
radius circle around the Anaconda Smelter (Brick and Moore, 1992).

The areal extent of exposed air that historically served as a pathway of hazardous substances,
based on elevated concentrations of hazardous substances in surface soils, included air in the
Anaconda area, Smelter Hill and the Old Works, Stucky Ridge and the Lost Creek drainage,
Mount Haggin Wildlife Refuge and the Mill Creek drainage, and the Deer Lodge Valley to
Garrison. Since the emissions and deposition of hazardous substances were prolonged and
ubiquitous in the Deer Lodge Valley and uplands to the west, all areas with sparse or no
vegetation or other soil covering are potential sources of continuing releases of hazardous
substances via fugitive dust.

Surface level winds within at least 1,260 square miles have been exposed to hazardous
substances; this represents the extent of transects that have been established to determine soils
concentrations, not the extent of contamination. Within a radius of 20 miles, air has been a
pathway of smelter emissions to soil, surface water, and vegetation resources.

2.4 MOBILITY AND RATES OF TRANSPORT OF HAZARDOUS SUBSTANCES
IN AIR

2.4.1 Main Stack Emissions

Historically, particles emitted in gas streams from flue systems venting smelters and other
facilities associated with ore refining in the Anaconda complex were transported by local
winds and dispersed. Rates of dispersion of airborne particles depend on the size of the
particles and the meteorological conditions. The gravitational settling velocity of a particle

RCG/Hagler Bailly

2-26

50 um in diameter is 67.5 cm per second. A particle of this size would fall 133 feet in 1
minute, or the entire 585-foot height of the Washoe Smelter stack in 4.4 minutes. The
average Anaconda wind speed at stack height was approximately 5 meters per second (about
1 1 miles per hour) (Gelhaus et al., 1979), at this windspeed, a particle of this size would be
transported 1.3 kilometers (0.6 miles) before it hit the ground. A smaller particle, 10 urn in
diameter, has a gravitational settling velocity of 2.7 cm per second, and under the same
meteorological conditions would fall to the ground from the top of the stack in 1.8 hours, or
approximately 33 kilometers (20 miles) away. Small particles travel with the wind, at the
speed of the wind.

The direction of transport of particles from the smelter stack depended on the winds at the top
of the stack and at the higher level to which the hot flue gases rose. The rugged terrain
surrounding the Anaconda smelter has a dramatic effect on plume transport and dispersion,
rugged, complex terrain can cause pooling in valleys, stagnation, drainage channeling along
valleys, and plume impingement on elevated terrain (Figure 2-16). Plume impingement
occurs when emissions are transported directly toward the side of a mountain or hill so that
the high plume concentrations intersect the ground. High concentrations and particle
deposition rates would occur under such conditions.

Emissions from the early years of smelting in Anaconda released from shorter stacks, fugitive
emissions from waste piles (including flue dust) throughout the years of smelter operation,
wind-blown dust from tailings piles, large particles spewn from the stack, and all particulate
emissions during precipitation events would have been deposited at greater rates closer to the
source than particles emitted from the 585 foot stack in fair weather.

2.4.2 Fugitive Emissions

A model was constructed in 1986 before the release of the Mill Creek ROD to estimate the
potential for the continuing contamination of Mill Creek soils by natural wind erosion
processes involving resuspension, transport, and deposition of hazardous substances in
exposed contaminated soils and wastes (McVehil-Monnet, 1986). Emission rates derived for
specific areas within the Smelter Hill Operable Unit site ranged (conservatively) from 1.5 to
5.1 pounds of arsenic per year (McVehil-Monnet, 1986). At that rate, the estimated annual
arsenic deposition due to airborne recontamination at Mill Creek is in the range of 0.0022 to
0.061 g/m2, or an increase of 0.05 to 1.5 ppm arsenic per year in the upper 1 inch of soil
(McVehil-Monnet, 1986). In modeling potential ongoing contamination at Mill Creek,
McVehil-Monnet (1986) assumed that the deposited arsenic is mixed into the upper 1 inch of
soil (density 1.6 g/cm3). Over a 20-year period, expected increase in soil arsenic where the
town of Mil Creek used to stand ranged from 1 to 30 mg/kg (ppm) (McVehil-Monnet, 1986).
Predictions of average airborne concentrations of arsenic in Mil Creek were shown to be
consistent with observed concentrations in the air at Mil Creek and elsewhere near the
smelter (McVehil-Monnet, 1986).

RCG/Hagler Bailly

2-27

Side View

Side View

A. Plume Impingement on
High Terrain

B. Pooling in Valleys

C. Persistence Due to
Channeling

Top View

Figure 2-16. Effects of Rugged Terrain on Dispersion. Source: Hanna and Strimaitis,
1990.

RCG/Hagler Bailly

2-28

2.5 OTHER PATHWAYS

2.5.1 Leaching/Erosion

Waste deposits in the Old Works, Smelter Hill, and tailings ponds areas that contain
hazardous substances are sources of hazardous substances to underlying and adjacent soils by
leaching and wind and water erosion pathways. Elevated concentrations in native soils
beneath the waste/alluvium interface were identified during RI/FS work in the Old Works
Operable Unit (PTI, 1991b) and in the Smelter Hill Operable Unit (PTL 1991a).

Leaching and vertical migration of hazardous substances from Anaconda area waste deposits
into the underlying geologic and groundwater resources have been documented in numerous
investigations (e.g., Tetra Tech, 1987; PTI, 1991a,b; PTI, 1992a). Evidence of erosion from
waste deposits in the Old Works Area, the Opportunity Ponds, and Smelter Hill area has also
been documented (PTI, 1991a; Tetra Tech, 1987). Contamination of shallow groundwater is
evident on Smelter Hill in the area near the stack; migration of dissolved contaminants is
occurring to depths of up to 200 feet (PTI, 1991a). In addition, analyses of drainage channel
sediments in the Smelter Hill area have shown that metals concentrations in the
handling/process/storage area, and in Geyser Gulch and Weather Hill drainages, are elevated.
The concentrations of metals in sediments are similar to concentrations found in surface soils
(PTI, 1991a), indicating that runoff erosion is an ongoing pathway of transport of hazardous
substances.

These examples are evidence of pathway mechanisms from waste deposits to geologic and
hydrologic resources. Evidence of vertical migration is visible as well; vertical sections
exposed in backhoe trenches in the Old Works area showed blue-green zones of copper
enrichment in the alluvium below the waste (Terra Tech, 1987). Trace metal data indicate
that leaching and re-precipitation in native alluvium have occurred (Tetra Tech, 1987).

2.5.2 Biological Uptake

Upland vegetation and wildlife resources are exposed to sources of hazardous substances in
upland soils in the Anaconda area. Investigations conducted for the Smelter Hill RI/FS
determined concentrations of hazardous substances in soils and in plant tissue. Total metals
concentrations in surface soils (0-2 inches) were significantly correlated with metals
concentrations in plant tissue for all metals tested (PTI, 1991a). Additional studies have
demonstrated that:

► Arsenic concentrations in plant tissue collected across Smelter Hill averaged

24.6 ± 34.0 ppm. Arsenic content in a wide variety of plants from
uncontaminated regions typically averages between 0.02 and 7.0 ppm (Alloway,
1990).

RCG/Hagler Bailly

2-29

► Plant tissue concentrations of cadmium at the Smelter Hill site averaged 2.35 ±
2.02 ppm. Nationwide levels of cadmium in plants are generally less than 1
ppm (Kabata-Pendias and Pendias, 1984 as cited in PTI, 1991a). Cadmium is
generally less toxic to plants than to consumers, thus cadmium accumulation in
plants represents a potential pathway to herbivorous wildlife.

► Copper plant tissue concentrations at the Smelter Hill site averaged 91.87 ±
166.5 ppm (ranging as high as 467 ppm in horsebrush). Levels in unimpacted
plants generally range between 1.5 and 30 ppm (CH2M Hill, 1987, as cited in
PTI, 1991a).

► Lead plant tissue concentrations across the Smelter Hill site averaged 15.43 ±
29.5 ppm. Nationwide lead levels in a variety of plant tissues generally range
from 0.5 to 4 ppm (Kabata-Pendias and Pendias, 1984, as cited in PTI, 1991a).

»• Zinc concentrations in plant tissue across the Smelter Hill site averaged 208 ±
173.5 ppm. Most agricultural crops exhibit toxicity when zinc tissue levels
reach 200-300 ppm (MacNichol and Beckett, 1985), but the range for reported
toxicity is much greater. Toxicity has been reported for tissue levels as low as
60 ppm.

It should be noted that concentrations of arsenic, cadmium, copper, lead, and zinc in upland
soils in the Anaconda area are, in places, so elevated that analysis of uptake residues is hardly
relevant: rather, the soils are phytotoxic. Chapter 3.0 and Appendix B describe the toxic
effects on Smelter Hill soils and vegetation.

2.6 SUMMARY

The data presented in this chapter support the following conclusions:

►■ Air has been exposed to hazardous substances through smelter emissions and
ongoing releases from exposed soils and waste piles.

► Air has served as an ongoing pathway to terrestrial resources.

►

Aerial patterns of soil contamination are indicative of a point source located in
the Anaconda area.

The vertical distribution of soil contamination demonstrates that hazardous
substances have been released by a surface source rather than reflecting
geologic parent material.

The area! extent of exposure exceeds over 100 square miles.

RCG/Hagler Bailly

2-30
2.7 REFERENCES

Adriano, D.C. 1986. Trace Elements in the Terrestrial Environment. Springer- Verlag, Inc.
NY.

Alloway, B.J. (ed.). 1990. Heavy Metals in Soils. NY: John Wiley and Sons, Inc. 339 pp.

AMC. 1983. Screening Study, Anaconda Smelter Site, Anaconda, Montana. Anaconda
Minerals Company, Anaconda, MT. Report prepared for U.S. Environmental Protection
Agency and Montana Department of Health and Environmental Sciences by Anaconda
Minerals Company.

ARCO. 1992. Preliminary Site Characterization Information Report: Anaconda Smelter
NPL Site, Old Works/East Anaconda Development Area Operable Unit Remedial
Investigation/Feasibility Study. Anaconda, MT.

Baker, D.E. 1990. Copper. In B.J. Alloway (ed), Heavy Metals in Soils. NY: John Wiley
and Sons, Inc. pp. 151-176.

Bowman, M.W. 1976. Main Stack Emanations October 1976. (Monthly Report.)
Anaconda, MT.

Brick, CM. and J.N. Moore. 1992. Clark Fork Damage Assessment Soil Mapping Project
Report. Prepared for Montana Natural Resource Damage Litigation Program.

CDM. 1987. Final Data Report for Solid Matrix Screening Study, Anaconda Smelter Site,
Anaconda, MT. Performance of remedial response activities at uncontrolled hazardous waste
sites. Prepared by Camp Dresser & McKee for U.S. Environmental Protection Agency.
Denver, CO.

CDM. 1988. Final Data Report for Community Soils Screening Study: Anaconda Smelter
Site. Prepared by Camp Dresser & McKee for U.S. Environmental Protection Agency.
Denver, CO.

CH2M Hill, Inc., Chen-Northern, and Montana State University Reclamation Research Unit.
1991. Draft Final: Upper Clark Fork River Screening Study. Volume J. Prepared for Montana
Department of Helath and Environmental Sciences.

Gelhaus, J.W., S.R. Harner, and M.D. Roach. 1979. Deer Lodge Valley Dispersion Study.
Air Quality Bureau, Environmental Sciences Division, Department of Health and
Environmental Sciences. Helena, MT.

RCG/Hagler Bailly

2-31

Hanna, S.R and D.G. Strimaitis. 1990. Rugged Terrain Effects on Diffusions. In W.
Blumen, ed., Atmospheric Processes Over Complex Terrain. Meteorological Monographs.
Vol. 23, No. 45. American Meteorological Society, Boston, MA.

Harkins, W.D. and RE. Swain. 1907. Papers on Smelter Smoke: The Determination of
Arsenic and Other Constituents of Smelter Smoke, with a Study of the Effects of High Stacks
and Large Condensing Flues. J. Amer. Chem. Soc. 29:970.

Kabata-Pendias, A. and H. Pendias. 1984. Trace Elements in Soils and Plants. 2nd ed.
CRC Press, Ann Arbor, MI.

Kabata-Pendias, A. and H. Pendias. 1992. Trace Elements in Soils and Plants. 2nd ed.
CRC Press, Ann Arbor, MI.

Lipton, J., H. Bergman, D. Chapman, T. Hillman, M. Kerr, J. Moore, and D. Woodward.
1995. Aquatic Resources Injury Assessment Report, Upper Clark Fork River Basin. Prepared
by RCG/Hagler Bailly for the State of Montana, Natural Resource Damage Litigation
Program. January.

MacNichol, R.D. and P.H.T. Beckett. 1985. Critical Tissue Concentrations of Potentially
Toxic Elements. Plant and Soil 85: 1 07- 1 29.

McVehil-Monnet. 1986. An Analysis of the Potential for Airborne Recontami nation of Mill
Creek, Montana from the Resuspension, Transport and Deposition of Surface Soil Arsenic
from the Anaconda Smelter Site. Prepared by McVehil-Monnet & Associates for Anaconda
Minerals Company. Denver, CO.

McVehil-Monnet. 1992a. Aerometric Monitoring Report for the Anaconda Smelter Remedial
Investigation/Feasibility Study - First Quarter 1992. Volume I. Prepared by McVehil-Monnet
& Associates, Inc. for PTI Environmental Services, Inc. and ARCO. Englewood, CO.

McVehil-Monnet. 1992b. Aerometric Monitoring Report for the Anaconda Smelter Remedial
Investigation/Feasibility Study - Second Quarter 1992. Volume I. Prepared by McVehil-
Monnet & Associates, Inc. for PTI Environmental Services, Inc. and ARCO. Englewood,
CO.

Munshower, F.F. 1972. Cadmium Compartmentation and Cycling in a Grassland Ecosystem
in the Deer Lodge Valley, MT.

Munshower, F.F. 1977. Cadmium Accumulation in Plants and Animals of Polluted and
Nonpolluted Grasslands. J. Environ. Qual. 6(4):411.

RCG/Hagler Bailly

2-32

O'Neill, P. 1990. Arsenic. In B.J. Alloway (ed), Heavy Metals in Soils. John Wiley and
Sons, Inc., NY, pp. 83-99.

Peirce, G.J., RE. Swain and J.P. Mitchell. 1913. The Effects of Smelter Smoke on
Vegetation and the Condition of the National Forests in the Vicinity of Anaconda, Montana.
Report Presented to the Department of Justice, L. Johnson, Special Assistant to the Attorney
General. February 5, 1913, Stanford University, CA.

PTI. 1990. Old Works Engineering Evaluation Cost Analysis: Data Summary, Validation,
Usability Report, Volume I. Prepared by PTI Environmental Services for ARCO. Bellevue,
WA

PTI. 1991a. Smelter Hill Remedial Investigation and Feasibility Study: Preliminary Site
Characterization Information. Prepared by PTI Environmental Services for ARCO. Bellevue,
WA.

PTI. 1991b. Old Works Engineering Evaluation/Cost Analysis: Preliminary Site
Characterization Information. Prepared by PTI Environmental Services for ARCO, Anaconda,
MT.

PTI. 1992a. Anaconda Soil Investigation: Data Summary/Data Validation/Data Usability
Report, (Draft). Prepared by PTI Environmental Services for ARCO. Bellevue, WA.

PTI. 1992b. Anaconda Soil Investigation: Preliminary Site Characterization Information
Report. Prepared for ARCO by PTI Environmental Services, Bellevue, WA.

PTI/MNRIS. 1992. Map of Smelter Hill Vegetation, with Corresponding Areal Coverage of
Each Vegetation Map Unit. Prepared by Montana Natural Resource Information System,
Helena, MT, Using Data Provided by PTI Environmental Services.

Swain, RE. and W.D. Harkins. 1908. Papers on Smelter Smoke: Arsenic in Vegetation
Exposed to Smelter Smoke. J. Amer. Chem. Soc. 30(6): 915.

Taskey, RD. 1972. Soil Contamination in Anaconda, Montana: History and Influence on
Plant Growth. M.S. Thesis, University of Montana, Missoula, MT.

Terra Tech. 1985a. Health Effects Soils Investigation Data Report, Anaconda Smelter RI/FS.
Terra Tech, Inc., Bellevue, WA.

Terra Tech. 1985b. Summary of Available Soils Data Mill Creek, Montana. Tetra Tech,
Inc., Bellevue, WA.

RCG/Hagler Bailly

2-33

Tetra Tech. 1986. Anaconda Smelter Remedial Investigation and Feasibility Study. Phase I
Data Compilation. Tetra Tech, Inc., Bellevue, WA.

Tetra Tech. 1987. Anaconda Smelter Remedial Investigation/Feasibility Study (Draft
Report). Prepared by Tetra Tech, Inc., Bellevue, WA for Anaconda Minerals Company.

U.S. EPA. 1983. Inorganic Emissions from High Arsenic Primary Copper Smelters:
Background Information for Proposed Standards. EPA-450/33-33-009A U.S. Environmental
Protection Agency Office of Air Quality Planning and Standards, Research Triangle Park, NC.

U.S. EPA. 1987. Record of Decision: Mill Creek, Montana. Anaconda Smelter Superfund
Site, First Operable Unit. Prepared by U.S. EPA Region Vm, Montana Office.

RCG/Hagler Bailly

3.0 GEOLOGIC RESOURCE — UPLAND SOILS

Geologic resources are solid components of the Earth's crust, including soils, sediments,
rocks, and minerals [43 CFR § 11.14]. In this chapter, injuries to upland soil resources near
Anaconda are determined and quantified.

Overall, the data presented in this chapter demonstrate that:

► Concentrations of hazardous substances in exposed soils are significantly
greater than baseline concentrations.

► Exposed soils are phytotoxic relative to control soils, and hence are injured.

►• The degree of phototoxicity observed in controlled laboratory experiments was

correlated with the concentrations of hazardous substances in the soil

► Concentrations of nitrate-nitrogen, potassium, and phosphorous, cation
exchange capacity, pH, and percent sand, silt, and clay in control and impact
soils were not significantly different.

»■ The observed effects on plant growth were consistent with the phytotoxic

effects of heavy metals.

»• Concentrations of hazardous substances in exposed soils exceed phytotoxic

thresholds identified in the literature; the observed phytotoxic response
therefore is consistent with scientific literature.

* The areal extent of severe injury to soils is estimated to be approximately 17.8

square miles.

In Chapter 4.0, it is demonstrated that upland soils that have been exposed to elevated
concentrations of hazardous substances and that were found in laboratory studies to be
phytotoxic, do indeed manifest severe injury to vegetation. Hence, evidence of phytotoxic
injury is supported by both controlled laboratory experiments and by field measurements.

3.1 DESCRIPTION OF UPLAND SOIL RESOURCES

Upland soils that have been injured by releases of hazardous substances from smelting and
related ore refining operations in Anaconda include the soils of:

RCG/Hagler Bailly

3-2

► Stucky Ridge

► Smelter Hill, including the A and C hills, Weather Hill, and other unnamed
slopes between Mill Creek and Warm Springs Creek to the west of the main
smelter stack

► Mt. Haggin Wildlife Management Area, south and east of Mill Creek,
extending south as far as the Continental Divide.

These upland areas exhibit extensive loss of native vegetative cover and topsoil. These three
areas are referred to as impact areas for simplicity of discussion (Figure 3-1). In reality, they
constitute a single, contiguous, injured landscape.

3.1.1 Stuckv Ridge

Stucky Ridge (Area A, Figure 3-1) is a glacial moraine that borders the town of Anaconda to
the north. Historical stacks and connecting flues were built along the southern flank of the
ridge (the Old Works area, Anaconda Smelter NPL Site). Historical photographs of the Old
Works (circa 1886) indicate that Stucky Ridge was formerly vegetated by arid grassland and
open steppe communities on exposed slopes and forest communities in the moister drainages.
Currently, most of Stucky Ridge is either completely devegetated or supports species-poor
communities. Extensive soil loss on southern slopes and hilltops has exposed a cobbled
surface, or the rocky core of the moraine, and is evident as deep erosion gullies in the
northern slopes and eastern end of the ridge,,

3.1.2 Smelter Hill

Smelter Hill (Area B, Figure 3-1) refers to the general locale of the former Washoe Smelter
(also called the Anaconda Smelter) facilities. The Smelter Hill Operable Unit of the
Anaconda Smelter NPL site (located within the boundaries of Area B, Figure 3-1) occupies
approximately 4 square miles adjacent to and south of the town of Anaconda (PTI, 1991).
The main flue to the main stack was on the north side of Smelter Hill. A brick smoke stack
is on the top of Smelter Hill (elevation 5,799 feet above mean sea level). The stack, an
equipment garage, and two guard shacks are the only remaining structures of the former
industrial facility (PTI, 1991). Weather Hill (elevation 6,372 feet above mean sea level), to
the south of the stack, the A and C hills bordering the town of Anaconda to the south, and
hillsides to the west of the stack are currently devegetated or support species-poor
communities. Extensive soil loss and cobbled, wind-eroded surfaces are prevalent. The site
where the main structures of the industrial facility once stood is called the handling/
process/storage (HPS) area. The HPS area is on the northern and eastern flanks of Smelter
Hill and extends to the large black slag pile and the Anaconda Tailings Ponds that lie just

RCG/Hagler Bailly

3-3

Upland Impact and Control
Soil Sampling Sites

Anaconda, Montana

Dashed lines Indicate areas del'ieated as grossly injured.
0 12 3 Miles

German Gulch

Figure 3-1. Upland Impact and Control Soil Sampling Sites.

RCG/Hagler Bailly

3-4

outside the operable unit. The soils of the HPS area have been heavily disturbed, both by
contamination with hazardous substances and by past construction and current reclamation
activities.

Two intermittent streams drain the area. Geyser Gulch flows northeast through the valley
between Weather Hill and Smelter Hill. Walker Gulch flows through the northwestern part of
the site and drains under the main site entrance road to a ditch north of the black slag pile.
Mill Creek, a perennial stream that drains to Silver Bow Creek, borders the NPL site to the
southeast (Figure 3-1).

3.1.3 Mount Haggin

The Mount Haggin Wildlife Management Area (23,000 hectares) is owned and managed by
the Montana Department of Fish, Wildlife, and Parks. The land was purchased in 1976 from
the Mount Haggin Livestock Company through The Nature Conservancy. The Mount Haggin
area is characterized by steep mountainous slopes. Portions of the management area (Area C,
Figure 3-1), including slopes to the south and west of Mill Creek, at least as distant as the
Continental Divide, are barren of vegetation and exhibit signs of extensive soil loss. The
Mount Haggin area was logged historically, as evidenced by the remains of old logging
camps amidst acres of un-decomposed logs, stumps, and exposed root systems. In addition,
extensive areas littered with standing dead or fallen dead trees are visible on slopes within the
Mount Haggin area.

3.2 INJURY DEFINITION

For this assessment, injury to soil is defined as:

►

Concentrations in the soil of substances sufficient to cause a phytotoxic
response such as retardation of plant growth [43 CFR § 11.62 (e) 10]

Concentrations of substances sufficient to have caused injury as defined to
surface water, groundwater, air or biological resources when exposed to the
substances [43 CFR § 11.62 (e) 11].

3.2.1 Phvtotoxicitv

All green plants require the same basic set of mineral nutrients and use them for similar
purposes (Fitter and Hay, 1987). Trace elements essential to plants at very low
concentrations are referred to as plant micronutrients. Copper and zinc are essential to plants
at very low concentrations, but become toxic at higher concentrations (Alloway, 1990).

RCG/Hagler Bailly

3-5

Arsenic, cadmium, and lead have no role in plant metabolism, and their accumulation in
plants is toxic (Kabata-Pendias and Pendias, 1984).

For most hazardous substances, there is a concentration range at which no injurious effects
are suffered by any species, and one at which all species are vulnerable (Fitter and Hay,
1987). Phytotoxic response stems from an inhibition of enzymatic reactions, metabolic
processes, or growth (Fitter and Hay, 1987). Plant toxins generally inhibit either the
acquisition or the utilization of nutrients and water by the plants (Alloway, 1990) (see
Table 3-1). Stunted growth, yellowing or blanching (chlorosis), localized death of tissue
(necrosis), leaf curling or withering (epinasty), and reddish-brown discoloration are visible
symptoms of severe metal phytotoxicity (Heale and Olmrod, 1982, Woolhouse, 1983; Van
Assche and Clijsters, 1990).

Other characteristic symptoms of metal toxicity include stunted root growth and either
browning or death of the root meristem. Plant root growth is controlled by plant hormones;
disruption of the concentration gradient by elevated metals concentrations in the root-soil
interface will affect hormonal activity and may inhibit root elongation and water and nutrient
uptake (Fitter and Hay, 1987; U.S. DOI, 1987). Poorly developed root systems can cause
water stress, growth reduction, or early plant aging and death (U.S. DOI, 1987). Phytotoxic
response may also be evidenced as an inability of a mature plant to produce viable seeds, or a
failure of seeds to germinate (U.S. DOI, 1987). Without annual production of seeds and
successful regeneration, most plant species cannot persist in a local environment (U.S. DOI,
1987).

Injury to vegetation grown on soils or sediments contaminated with hazardous substances can
be direct or indirect. Direct phytotoxicity is generally observed within days of exposure and
results from exposure of the plant to compounds that disrupt ionic balance within the plant or
inhibit gas and ion exchange in the plant root zone (U.S. DOI, 1987). Indirect effects
resulting from the disruption of biotic processes important to long-term plant growth, survival,
and reproduction can occur over an extended period, i.e., months or years (U.S. DOI, 1987).
Indirect effects include residual impacts of hazardous substances in the rooting zone and
impacts to the soil organisms that contribute to nutrient availability in the plant root system
The response of plants grown on soils containing elevated concentrations of hazardous
substances varies widely by species; some may tolerate toxic concentrations by exclusion or
excretion of hazardous substances, but more sensitive species are eliminated either by death,
indirectly through competitive interactions with resistant species, or by additional stressors
such as drought and disease (Newman, 1983).

RCG/Hagler Bailly

3-6

Table 3-1
A Simple Classification of the Effects of Hazardous Substances on Plants

Effects on the Ability of Plants to Acquire Resources

1. Acquisition of water:

a. Osmotic effects arising from excess solute concentrations

b. Inhibition of cell division, reducing root growth

2. Acquisition of nutrients:

a. Competition between ions

b. Damage to membranes

c. Effects on symbionts

d. Inhibition of cell division

3. Acquisition of carbon dioxide and light energy:

a. Stomatal malfunction caused by toxic gases

b. Chlorophyll bleaching

Effects on the Ability to Utilize Resources

1. Inhibition of enzyme action

2. Inhibition of cell division

3. Loss of respiratory substrates; oxygen deficiency

Source: Fitter and Hay, 1987.

3.2.2 Assessment of Soils Phytotoxicity

Soil phytotoxicity was assessed through controlled laboratory experiments using soils
collected from impact and control areas, and by the assessment of vegetation community
integrity in impact and control areas.

3.2.2.1 Laboratory Testing

Phytotoxicity is an adverse response of plants to chemical stressors. Excessive soil
concentrations of both essential and nonessential metals result in phytotoxic responses that
range from decreased productivity or vigor to death. However, no regulatory criteria exist for
soil metals concentrations sufficient to cause phytotoxic responses. In this assessment,
phytotoxicity of hazardous substances in soils was determined by laboratory tests of seed
germination, seedling growth, and root elongation [43 CFR § 11.64 (e) (6)]. Standard
Operating Procedures consistent with the ASTM protocol for early seedling growth studies

RCG/Hagler Bailly

3-7

(ASTM, 19941) were adapted to evaluate whether site soils cause phytotoxicity (see Appendix
B). Screening tests examined the toxicity of 100% site soils to the standard laboratory test
species alfalfa, lettuce, and wheat seedlings over a two-week exposure period. Performance
of test species in impact site soils was compared to performance of control species grown in
control soils. Control soils were collected from control sites previously designated and
sampled as representative of baseline conditions (Section 3.3.2.1 and Appendix A).

Determination of phytotoxicity was based on a suite of species-endpoint responses, including:

Germination (number of seedlings emerged/20 seeds; five replicates of 20 seeds
each)

Shoot height (mm, for each surviving seedling; mean of plants measured)

Maximum root length (mm, for each surviving seedling; mean of plants
measured)

Shoot mass (g oven dry weight for each seedling; mean of plants measured)

Root mass (g oven dry weight for each surviving seedling; mean of plants
measured)

Total mass (g oven dry weight for each surviving seedling; mean of plants
measured).

Differences between impact and control samples were assessed statistically using randomized
t-tests. A soil sample was deemed phytotoxic when plants grown in it displayed significantly
lower performance in the endpoints identified above, relative to the control soil. The degree
of phytotoxicity was quantified based on (1) the magnitude of the difference from control
performance, (2) the number of endpoints exhibiting a phytotoxic response, and (3) the
number of species exhibiting a phytotoxic response. Quantitative phytotoxicity scores were
then assigned to each soil sample based on the degree of phytotoxic response. These toxicity
scores, in turn, were used to classify soils as either:

*■ Non phytotoxic

► Phytotoxic:

At the time the laboratory phytotoxicity tests were performed, this standard was available as a
working draft. The standard tests incorporated the methods outlined in Greene et al., 1989 and in Gorsuch
et al, 1990.

RCG/Hagler Bailly

3-8

Mildly toxic
Moderately toxic
Highly toxic
Severely toxic.

3.2.2.2 Vegetation Communities

Soil phytotoxicity was also evaluated by comparing the structure and composition of
vegetation communities in impact and control areas. Injury to vegetation communities is
defined in terms of community integrity, i.e., as changes in the species composition, diversity,
dispersion, or percent cover [43 CFR §11.71 (1)(6)], and was determined by reference to the
established baseline area [43 CFR §11.72]. Injury to vegetation is addressed in Chapter 4.0,
however, obvious injury to vegetation is indicative of injury to soil, and will be discussed as
it pertains to soils in the injury determination and quantification section of this chapter.

3.3 INJURY DETERMINATION AND QUANTIFICATION

This section presents evidence that soils in the Anaconda area contain elevated concentrations
of hazardous substances attributable to smelter emissions and fugitive dust (see Chapter 2.0),
that these concentrations of hazardous substances are sufficient to cause a phytotoxic
response, and, hence, that soils are injured.

3.3.1 Baseline Conditions

Baseline is defined as the condition or conditions that would have existed at the assessment
area had the release of hazardous substances not occurred [43 CFR § 11.14 (e)]. Baseline
concentrations are not necessarily naturally occurring "background" concentrations; other
anthropogenic impacts may have contributed to pre-release conditions, such as releases of
lead from leaded gasoline-burning vehicles. However, the naturally occurring "normal"
concentrations of trace elements in soils are of interest as background values pertinent to the
assessment of the degree of soil contamination in both impact and control areas Natural
background concentrations of arsenic, cadmium, copper, lead, and zinc are primarily
controlled by the parent material. Chemical and physical weathering breaks down bedrock,
releasing constituent minerals, which become a part of the soil. The natural soil composition
is further affected by mobility and activity of the various ions. Thus, the range of naturally
occurring concentrations throughout the United States (Table 3-2) is due principally to parent
material variability and soil weathering processes and rates.

RCG/Hagler Bailly

3-9

Table 3-2
Background Concentrations Reported for Uncontaminated Soils of the United States

(total metals in ppm)

Arsenic

Cadmium

Copper

Lead

Zinc

Range
Mean

<0.1 -93
10

0.005 - 2.4
0.53

1 - 100

25

10- 70

32

17- 125
64

The range of arsenic in U.S. soils is broad, < 0. 1 - 93 ppm, but most soils are in the less
than 10 ppm range (Kabata-Pendias and Pendias, 1992). Baseline values for arsenic in soils
overlying granitic rock typically range from 0.7 to 15 ppm (Kabata-Pendias and Pendias,
1984, as cited in Morrison Knudsen, 1991). The lowest concentrations of arsenic are found in
sandy soils, particularly sandy soils derived from granites (means in the range of 2-6 ppm).

Higher As concentrations are found in alluvial soils and soils rich in organic matter (means in
the range of 5-25 ppm). Acid sulfate soils are reported to accumulate a high proportion of
native As, up to 30 to 50 ppm. An overall mean value for arsenic in U.S. soils is reported to
be between 5.8 and 10 ppm (Kabata-Pendias and Pendias, 1992; O'Neill, 1990).

Average soil concentrations of cadmium in the United States range between 0.005 and 2.4
ppm (Alloway, 1990). The highest average cadmium content is found in organic soils (0.78
ppm), while the lowest mean is found in highly leached spodosols (0.37 ppm). The
calculated worldwide mean of cadmium in surface soils is 0.53 ppm (Kabata-Pendias and
Pendias, 1992).

Based on recent compilations of estimates, the average natural copper concentration in soil
worldwide ranges from 20 to 30 ppm (Baker, 1990). Generally, copper concentrations
reported for various soil types (sandy soil and podzols, silty soils, loamy and clay soils,
rendzinas, chernozems, histosols, and other organic soils) range from 1 to 100 ppm (Kabata-
Pendias and Pendias, 1992). Copper concentrations typically are lowest in soils formed by
granitic, igneous, or carbonate parent materials. Kabata-Pendias and Pendias (1992) cite 13-
24 ppm as the average copper concentration in soils of the United States, and Adriano (1986)
cites 25 ppm as an average value for naturally occurring copper in U.S. soils.

Background values for lead in soils of the world rock typically range between 10 and 67 ppm
and average 32 ppm (Kabata-Pendias and Pendias, 1992). Over all soil types, a mean value is
approximately 32 ppm. An upper limit for the lead content of normal soils has been
suggested as 70 ppm, higher values reflect anthropogenic inputs (Kabata-Pendias and Pendias
1992).

RCG/Hagler Bailly

3-10

Mean zinc content in surface soils ranges from 17 to 125 ppm (Kabata-Pendias and Pendias,
1992). The mean calculated for worldwide soils is 64 ppm (Kabata-Pendias and Pendias,
1992). Kiekens (1990) cites an average content of 50 ppm. Highest naturally occurring mean
values are found in alluvial soils; the lowest naturally occurring mean values are in light
mineral and organic soils.

Anaconda Baseline

Historical data regarding the baseline conditions specific to Anaconda area soils before
releases of hazardous substances are not available since releases have occurred continuously
from the late 1800s through the present. In the absence of pre-release data, concentrations in
soils in the German Gulch area, approximately 6 to 10 miles south of the Anaconda smelter,
have been defined as baseline (see Appendix A).

The German Gulch area was selected as an appropriate control area because of its proximity
to the impact area, and its similarity in elevation and overall aspect. Surface soils collected in
German Gulch were analyzed for total arsenic, cadmium, copper, lead, and zinc (sampling
locations shown in Figure 3-1). Means and ranges for the 20 control soils sampled were in
excess of U.S. means, and copper and cadmium concentrations were in excess of reported
ranges for uncontaminated U.S. soils. These results, coupled with soils distribution profiles
(Chapter 2.0, Section 2.4.3), indicate that soils in the German Gulch area most likely have
been exposed to aerial deposition of emissions from the Anaconda smelter. For example:

*■ Arsenic in German Gulch samples ranged from 47.2 to 136.4 ppm and

averaged 89.5 ppm. Overall mean values for arsenic in U.S. soils are reported
to be between 5.8 and 10 ppm (range < 0.1-93 ppm).

► Cadmium in German Gulch samples ranged from 1.3 to 5.1 ppm and averaged
3.0 ppm. The calculated worldwide mean value for cadmium in soils is 0.53
ppm (range 0.005-2.4 ppm).

* Copper in German Gulch samples ranged from 58.6 to 236.5 ppm and averaged

148.4 ppm. An average value for copper in U.S. soils is 25 ppm (range 1-100
ppm).

► Lead in German Gulch samples ranged from 31.1 to 119.7 ppm and averaged
68.1 ppm. Mean soil concentration of lead in U.S. soils is approximately 32
ppm (range 10-70 ppm).

*■ Zinc in German Gulch samples ranged from 78.4 to 196.3 ppm and averaged

135.9 ppm. Baseline zinc concentrations in U.S. soils average 50 ppm (range
17-125 ppm).

RCG/Hagler Bailly

3-11

Maximum concentrations from the control sites are clearly elevated above what can be
considered naturally occurring levels. German Gulch mean values thus represent conservative
baseline values (Table 3-3).

Table 3-3

Baseline Concentrations (ppm) of Hazardous Substances in Anaconda Area Soils

(measured in German Gulch area 0-2 inch soils)

Arsenic

Cadmium

Copper

Lead

Zinc

Range
Mean

47.2-136.4
89.5

1.3-5.1

3.0

58.6-236.5
148.4

31.1-119.7
68.1

78.4-196.3
135.9

3.3.2 Soil Concentrations of Hazardous Substances

3.3.2.1

NRDA Investigations

As a component of the soil resource investigation of this NRDA, a soil sampling study was
designed to sample soils from areas identified as grossly impacted. Grossly impacted areas
were identified in the field based on the following factors:

Complete or virtual elimination of the indigenous major plant associations
Little or no regeneration of the indigenous major plant associations

► Extensive topsoil exposure and erosion associated with vegetation loss.

The following areas were delineated for sampling using a combination of aerial photographs
and field observations:

+ 3.8 square miles on and immediately adjacent to Stucky Ridge (Area A,

Figure 3-1)

► 7.3 square miles on Smelter Hill, including Weather Hill and other slopes to
the south and west of the stack (Area B, Figure 3-1)

► 6.7 square miles within the Mount Haggin Wildlife Preserve, south and west of
Mill Creek, extending toward the Continental Divide (Area C, Figure 3-1).

Forty surface soil samples were collected from these three impacted areas. Sample site
locations were determined based on the Universal Transverse Mercator grid (Figure 3-1). Soil
samples were composited along transects centered on the sample site. The sample design
allowed statistical comparisons of each impact area, as well as comparison among individual

RCG/Hagler Bailly

3-12

sample locations. For 20 of the impact area sample sites, a paired control site was identified
and sampled (German Gulch; Figure 3-1). For a full description of the soils investigations,
see Appendix A.

Soils collected from the 0-2 inch and 0-6 inch layers were sent to a laboratory for analysis of
arsenic, cadmium, copper, lead, and zinc. A subset of the samples collected was analyzed for
pH, texture (percent sand, silt, clay), soil nutrients including nitrate, phosphorous, and
potassium, organic matter, and cation exchange capacity (CEC). Field vegetation and wildlife
habitat measurements were also recorded at impact and control sites for analysis of injury to
biological resources (see Chapter 4.0 and Appendices C and D).

Statistical comparison of the two soil data sets revealed that total concentrations of arsenic,
cadmium, copper, lead, and zinc in the impact areas are significantly greater than
concentrations of the same elements in paired control sites. A summary of the concentration
data is presented in Table 3-4; mean value comparisons between impact and control soils are
illustrated in Figure 3-2; raw data are presented as an attachment to Appendix A. Statistical
comparisons of ancillary parameters indicated that nutrient concentration, pH, CEC, and soil
texture do not differ significantly between control and impact soils, but that control soils have
a significantly greater percent organic matter than impact soils (Table 3-5). Specifically, the
following statistically significant differences were found:

►■ Arsenic, cadmium, copper, lead, and zinc occur in significantly greater

concentrations in upland impact area soils than in German Gulch control soils
(p = 0.0002).

»■ Arsenic, copper, and zinc occur in significantly greater concentrations in

Stuck>' Ridge soils than in paired control soils (p = 0.03).

*■ Arsenic, cadmium, copper, lead, and zinc occur in significantly greater

concentrations in Smelter Hill area soils than in paired control soils
(p = 0.00391).

► Arsenic, cadmium, copper, and lead occur in significantly greater
concentrations in Mount Haggin area soils than in paired control soils
(p = 0.008 (As); p = 0.015 (Cd, Cu, Pb)).

► Percent organic matter is significantly greater in control soils than in impact
soils (p = 0.0078). No differences in pH, CEC, percent sand, silt, or clay, nor
concentrations of nitrate, potassium, or phosphorous were detected (p > 0.05).

RCG/Hagler Bailly

3-13

Table 3-4
Summary Statistics for Anaconda Upland Impact and German Gulch Control 2-Inch Soil

Samples (total metals in ppm)

Sample Identification

As

Cd

Cu

Pb

Zn

Stucky Ridge n = 10

Arithmetic mean
Maximum

Minimum
Standard deviation

303.50
624.30
142.70
144.84

3.83
7.30
2.30
1.39

1,324.23

2,856.00

513.40

677.17

117.00
196.00
68.70
36.97

300.74
580.50
167.40
129.56

Stucky Ridge controls n = 5

Arithmetic mean
Maximum
Minimum
Standard deviation

78.16
119.6
47.20
24.78

3.50
4.10
2.20
0.69

183.16
220.40
135.50
31.42

89.36
112.70
60.80
17.73

15088
182.60

121.00
21.20

Smelter Hill n = 14

Arithmetic mean
Maximum
Minimum
Standard deviation

605.29

1,846.70

183.30

410.42

17.64
51.20
7.40
12.16

1,229.48

2,547.00

404.90

627.76

275.59
548.80
156.80
108.96

562.27
1,515.00
205.10
363.18

Smelter Hill controls n = 8

Arithmetic mean
Maximum
Minimum
Standard deviation

90.91
132.30
47.30
22.82

2.86
4.30
1.70
0.94

143.41
236.50
96.40
43.96

58.38
88.00
41.00
15.07

140.96
196.30
80.20
36.68

Mount Haggin n = 16

Arithmetic mean
Maximum
Minimum
Standard deviation

295.83
630.00
107.60
144.59

7.80
14.20
2.30
3.70

356.38
678.70
139.10

155.55

153.26
229.20
78.10
49.64

215.78

379.30

76.60

93.69

Mount Haggin controls n = 7

Arithmetic mean
Maximum
Minimum
Standard deviation

99.37
136.40
53.00
27.17

2.64
5.10
1.30

1.30

118.73
201.50
58.60
49.99

56.54
119.70
31.10
33.10

115.80
165.40
78.40
27.95

RCG/Hagler Bailly

3-14

700

600

E. 600

c
f

£ 300

<

e 200

100

Control Soil
Impact Soil

Stucky Ridge Smelter Hill Mount Hoggin

1400

Stucky Ridge Smelter Hill Mount Hoggin

Stucky Ridge Smelter Hill Mount Hoggin

Figure 3-2. Mean Concentrations of Hazardous Substances in Upland Impact and
Control Soils.

RCG/Hagler Bailly

3-15

Table 3-5

Comparisons of Mean Soil Nutrient Concentrations (ppm), pH, CEC (meq/lOOg),

% Organic Matter, and % Particle Size in Impact and Reference Soils (0 to 5 cm)

Variable

Reference Samples

Impact Samples

p-value

Potassium

609.7

501.6

0.4538

N03-nitrogen

1.8

4.7

0.0904

Phosphorous

137.1

139.4

0.9358

pH

5.7

5.4

0.2908

CEC

16.4

16.0

0.8888

% Organic matter

6.4

3.9

0.0078*

% Sand

59.6

56.5

0.5124

% Silt

25.2

24.3

0.7562

% Clav

16.1

18.8

0.2310

* indicates a significantly greater amount in reference soils (2-tailed tests).

3.3.2.2

Previous Investigations

The results of this study are consistent with previous studies performed as part of the RI/FS.
The soils of the Smelter Hill Operable Unit have been sampled extensively as part of the
remedial investigation. The Preliminary Site Characterization Information Report (PSCI)
(PTI, 1991) was released in November 1991. In January 1993, the Smelter Hill RI/FS Phase
II soils data, collected during summer 1992 in response to data gaps identified in the PSCI,
were released (PTI, 1993). The Smelter Hill Operable Unit (Figure 3-3) and the Smelter Hill
study area sampled for this NRDA partially overlap. For this reason, soils concentrations of
hazardous substances are presented here for comparison to the data presented above.

Hazardous substances are elevated in soils throughout the Smelter Hill Operable Unit (PTI,
1991). Sources of hazardous substances in the Operable Unit (OU) include soils in disturbed
areas contaminated with mixed deposits of processing wastes, flue dust piles, soils that
received deposition of smelter emissions and have accumulated a reservoir of hazardous
substances, surface drainages that transport contaminated sediments, and shallow groundwater
beneath the site (PTI, 1991). For the Phase I investigations, the OU was divided for sampling
into three areas: (1) the Handling/Process/Storage (HPS) area, where the main structure of the
Anaconda Smelter facilities once stood; (2) disturbed soils, where regrading or surface
restructuring has occurred, exclusive of HPS areas; and (3) unreclaimed soils — those that
have not undergone regrading or major surface restructuring, and where the primary source of

RCG/Hagler Bailly

u

J2

—

u

C

o

B

l

-

a

o
u

3-17

contamination was smelter emissions. Phase II sampling was targeted toward areas that had
not been sampled previously and towards "hotspot" sites, where more information was needed
to characterize the contamination.

HPS Area

Areas within the HPS area that are of concern are the primary HPS area, the East Anaconda
Yard, the coal pile tracks, the railroad beds, and the stack area. In the primary HPS area,
concentrations of hazardous substances in disturbed soils are highly variable, but remain
extremely elevated to depths exceeding 48 inches. Concentrations of arsenic as high as
567,000 ppm have been detected in buried soils in the HPS area (Table 3-6) (PTI, 1991). In
the East Anaconda Yard, hazardous substance concentrations also increase with depth,
reaching maximum concentrations at depths > 48 inches. Where reclamation has occurred
(covering with 18 inches of imported soils), metals concentrations are lowest.

Soils in the coal pile tracks area contain extremely elevated concentrauons of hazardous
substances, and the railroad beds exhibit the highest median metals concentrations of all areas
sampled on Smelter Hill aside from flue dust (PTI, 1991). Concentrations of hazardous
substances in surface and sub-surface soils surrounding the Washoe Smelter stack
(approximately 10 acres) are extremely high (PTI, 1991). In the past, smelter emissions were
the primary source of hazardous substances to this area, but arsenic concentrations vary with
soil matrix type, suggesting a history of deposition of waste material in piles.

Mean concentrations of arsenic reported in HPS area soils are 6 to nearly 60 times greater
than baseline, and range between 57 and 6,000 times mean background concentrations
(Table 3-6). Mean copper concentrauons in the HPS area are 15 to nearly 40 times baseline
and 90 to 230 times mean background concentrations. Zinc and lead concentrauons reported
for HPS area soils exhibit similar exceedences of both background and baseline.

Disturbed Soils

Areas designated as disturbed soils include reclaimed areas on Smelter Hill, reclaimed areas
in the East Anaconda Yard, areas around the East Anaconda entrance to Smelter Hill,
bulldozed slopes below the stack, and other areas east of the HPS area (PTI, 1991).
Additional disturbed soil samples were collected during the Phase II sampling (PTI, 1993),
including samples in the vicinity of the drainage area near the southeastern corner of the
Anaconda Tailings Pond, in areas adjacent to the hydro-separator, and near two sites sampled
previously that had extremely high concentrations of metals.

Concentrations of arsenic, cadmium, copper, lead, and zinc in disturbed area soils (Table 3-6
and 3-7) vary widely, but in general are grossly elevated and exceed concentrations reported
for NRDA samples. Mean arsenic concentrations in 0-2 inch disturbed soils exceed baseline
by 6 to 30 times, and background by 60 to nearly 300 times. Mean copper 0-2 inch

RCG/Hagler Bailly

3-18

Table 3-6

Summary Statistics for Smelter Hill OU RI/FS Phase I Soils Data

(all concentrations in ppm; all depths in inches)

Location Statistic

As

Cd

Cu

Pb

Zn

Primary HPS

Mean (all depths, 0- > 48)

Maximum
East Anaconda yard

Mean (all depths, 0-48)

Maximum
Coal pile tracks

Mean (all depths, 0- > 48)

Maximum
Railroad beds

Mean (all depths, 0-24)

Maximum
Stack area

Mean (all depths, 0-7 > 48)

Maximum

5,115
567,000

578
6460

833
65,800

2048
19,800

4548
79.700

NA

5683
67,800

2340
65,900

5193
63,100

3933
34,900

2252
34,900

970
35,100

1352
60,000

608
15,700

793

5520

1297
28.600

982

42,000

1911
18,300

3895
88,100

4722
23,300

1756
15,550

Disturbed soils

Mean (all depths, 0- > 48)

Mean 0-2

Maximum

807

1064

29.300

22
33
584

2877

4379

160.000

489
714

22.400

2337

3320

61.600

Unreclaimed soils

Mean (all depths. 0-24)
Mean 0-2
Maximum

753

1631

27,200

21
53
964

1561

4143

72.400

353
910
6430

954

2233
30.400

Background U.S. concentrations*
Mean
Range

10
< 0.1-93

0.53
0.005-2.4

25
1-100

32
10-70

64
17-125

Baseline concentrations*
Mean
Range

89.5

47.2-
136.4

3.0
1.3-5.1

148.4
58.6-236.5

68.1

31.1-119.7

135.9
78.4-
196.3

* Background and baseline concentrations are included for comparison.
Source: PTI, 1991.

RCG/Hagler Bailly

3-19

Table 3-7

Summary Statistics for I
Distoi

Smelter Hill OU RI/FS Phase II Soil Samples,

bed and Unreclaimed Soils

(all

concentrations in ppm)

Location Statistic

As

Cu

Pb

Zn

Unreclaimed Soils

Drainage area below the loop

Mean (0-2")

724

1,051

376

816

Mean (all depths, 0-10")

505

683

233

627

Maximum

971

1,240

462

1,250

Northwest corner of OU

Mean (0-2")

605

1,547

313

593

Mean (all depths, 0-10")

400

1,044

180

442

Maximum

1,750

5,310

1,150

1,640

Stack area

Mean (0-2")

2,315

3,335

930

1.190

Mean (all depths, 0-10")

1,434

2,181

533

874

Maximum

3,020

3,490

1,370

1,250

Weather Hill

Mean (0-2")

936

1,034

464

810

Mean (all depths, 0-10")

700

668

293

648

Maximum

1,890

3.630

2.040

3,520

Disturbed Soils

Stations D-8 and D-50

Mean (0-2")

1,836

7,718

980

4,244

Mean (all depths, 0-120")

8,351

18,040

1.396

7,688

Maximum

37,000

93,500

4.870

24,500

Anaconda Ponds drainage area

Mean (0-2")

590

1,085

291

1.678

Mean (all depths, 0-24")

347

490

152

815

Maximum

1,340

2,070

590

4,710

Hydro separator area

Mean (0-2")

2,962

11,913

2,336

12,246

Mean (all depths, 0-24")

1,133

4,317

833

4,596

Maximum

5.510

20,700

4,610

22,500

Stack area

Mean (0-2")

1,669

1,224

406

738

Mean (0-10")

1,157

.. 883

274

639

Maximum

4,330

3,610

1,150

1,440

Source: PT1, 1993.

RCG/Hagler Bailly

3-20

concentrations in disturbed soils exceed mean baseline by 7 to over 80 times, and background
by over 40 to nearly 500 times. Lead and zinc exceedences of baseline and background are
similar (Tables 3-6 and 3-7).

Unreclaimed Soils

The other source of concern in the operable unit is unreclaimed soils, where the primary
source of contamination was smelter emissions, and continues to be ongoing fugitive
emissions. Unreclaimed soils have not undergone regrading or major surface restructuring
(PTI, 1991) and, within the operable unit, are west and south of Smelter Hill. Additional
samples of unreclaimed soils were collected during the Phase II sampling to refine
characterization of the western slope of Weather Hill, drainages to the west of Weather Hill,
and localized areas near the stack and below the loop track. Unreclaimed surface soils
contain elevated concentrations of hazardous substances, and migration of surface deposited
metals has extended to the upper 10-24 inches of soil (PTI, 1991). Mean 0-2 inch
concentrations reported for Phase I unreclaimed soils (Table 3-6), and Northwest Corner and
Weather Hill 0-2 inch Phase II samples (Table 3-7) are similar to those reported measured in
the impact areas for the NRDA investigations (Table 3-4).

3.3.3 Category of Injury: Phototoxicity - Laboratory Evidence

3.3.3.1 NRDA Phytotoxicity Investigation

Summary of Test Results

To test whether hazardous substances in exposed soils cause a phytotoxic response relative to
control soils, laboratory germination and growth tests were conducted Standard operating
procedures consistent with the ASTM protocol for early seedling growth studies (ASTM,
1994) were adapted to compare seed germination and seedling performance of standard
laboratory test species (alfalfa, lettuce, and wheat) in impact and control area soils.

Laboratory tests of soils from the upland impact areas confirmed that upland soils are
phytotoxic (Appendix B). Eighteen of the 20 soil samples (90%) collected from the impact
area exhibited phytotoxicity, including 4 of 5 soils (80%) from Stucky Ridge, 8 of 8 soils
from Smelter Hill (100%), and 6 of 7 of soils (86%) from Mount Haggin. These results are
consistent with the observed loss of vegetation in the three impact areas (see Chapter 4.0).

Further, it should be noted that these determinations of phytotoxicity were based solely on
comparisons to control site soils. As mentioned previously, these control soils are most likely
somewhat enriched in hazardous substances from smelter deposition and thus represent a
conservative baseline. In addition, the phytotoxicity screening tests were relatively short (2

RCG/Hagler Bailly

3-21

weeks); the results of such short term tests probably understate the magnitude of phytotoxic
responses.

Description of Phytotoxicity Tests

Soils collected from the impact and control sites (0-2 inch layer) were used in upland
phytotoxicity tests. Alfalfa, lettuce, or wheat seeds were placed in approximately 100 g of
site soil (impact or control). Five replicate tests of 100% impact soil or 100% control soil
were conducted. Seeds that germinated were observed over a 14-day post-germination period
and, upon harvesting, were measured to determine whether impact and control plants differed
in shoot length, shoot mass, root length, root mass, and overall mass (see Appendix B).

The following endpoints were used to evaluate phytotoxicity:

*■ Germination (number of seedlings emerged/20 seeds; five replicates of 20 seed

each)

»■ Shoot height (mm, for each surviving seedling, mean of plants measured)

►•

Maximum root length (mm, for each surviving seedling; mean of plants
measured)

»■ Shoot mass (g oven dry weight for each seedling; mean of plants measured)

*■ Root mass (g oven dry weight for each surviving seedling; mean of plants

measured)

* Total mass (g oven dry weight for each surviving seedling; mean of plants

measured).

A total of 18 species-endpoint measures were tabulated for all plants that germinated (six
endpoints; three species). Plants that did not germinate were (conservatively) not considered
in the subsequent analyses of growth parameters. The performance of test species was
compared statistically to that of control species. A response was deemed phytotoxic when
any of the plants in impact soils exhibited any endpoints that were statistically less than the
endpoints exhibited by control plants. Species endpoints were compared for each site, and
were pooled and tested by area. T-tests comparing impact response with control response
demonstrated statistical differences for each of the three upland areas (Table 3-8).

Mean root length of alfalfa, wheat, and lettuce, mean shoot height and root mass of alfalfa
and wheat, and mean shoot mass, total mass, and percent germination of alfalfa were all
significantly greater in German Gulch control soils than in Stucky Ridge soils (Table 3-8).
Mean root length, root mass, shoot height, and total mass of alfalfa, wheat and lettuce, mean

RCG/Hagler Bailly

3-22

Table 3-8

Statistical Comparison of Upland Impact Areas to Pooled Control Means

I"

Mount

German

Stucky Ridge

Smelter Hill

Haggin

Endpoint

Gulch Mean

Mean

Mean

Mean

Alfalfa shoot height

36.2

21.0*

17.6*

20.1*

root length

79.5

23.0*

15.1*

24.0*

shoot mass

0.326

0.175*

0.146*

0.144*

root mass

0.245

0.090*

0.061*

0.100*

total mass

0.571

0.265*

0.207*

0.244*

% germination

90.2

57.4*

70.6*

59.1*

Lettuce shoot height

32.9

18.8

18.0*

14.7*

root length

39.9

16.6*

18.9*

12.1*

shoot mass

0.064

0.051

0.030*

0.044

root mass

0.048

0.027

0.012*

0.021*

total mass

0.111

0.079

0.042*

0.065

% germination

29.2

36.4

24.8

43 1

Wheat shoot height

156.0

107.4*

95.4*

113.3*

root length

162.1

50.6*

48.4*

51.4*

shoot mass

1.380

1.076

1.026

1.128

root mass

1.628

0.964*

0.909*

1.183*

total mass

3.008

1.040

1.935*

2.311*

% germination

87.6

90.8

904

92.0

* indicates that the impact value is

significantly less than the control '

^alue at a = 5%.

shoot mass of alfalfa and wheat, and mean percent germination of alfalfa were significantly
greater in German Gulch control soils than in Smelter Hill soils. Similarly, mean root length,
root mass, and shoot height of alfalfa, wheat and lettuce, mean total mass of alfalfa and
wheat, and mean shoot mass and percent germination of alfalfa were significantly greater in
German Gulch control soils than in Mount Haggin soils. Thus in all three impact areas,
growth inhibition of alfalfa was quantified as a significant reduction of all six endpoints.
Growth inhibiuon of wheat was quantified as a significant reduction of up to four of six
endpoints (Smelter Hill and Mount Haggin soils), and growth inhibiuon of lettuce was
quantified as a significant reduction of up to five of six endpoints (Smelter Hill soils). The
tests demonstrated statistically significant differences for each impact area, for all endpoints,
and for all species.

The degree of phytotoxic response was quantified based on (1) the magnitude of the
difference between impact and controls, (2) the number of endpoints exhibiting a phytotoxic

RCG/Hagler Bailly

3-23

response, and (3) the number of species exhibiting a phytotoxic response (Appendix B)
Sample sites were described as being either:

»■

Nonphytotoxic

Mildly phytotoxic

Moderately phytotoxic
*■ Highly phytotoxic

> Severely phytotoxic.

Of the 20 upland sites, 18 (90%) were determined to be phytotoxic (Table 3-9). The majority
(1 1 of 20) were found to be highly phytotoxic.

Table 3-9

Summary of Phytotoxicity as Determined by Overall Site Scores

(see Figure 3-1 for site locations)

Stucky Ridge
Site

Smelter Hill
Site

Mount Haggin
Site

Severely phytotoxic

A-10

Highly phytotoxic

A-03
A-06

B-02
B-04
B-09
B-11
B-13

C-01
C-05
C-07
C-09

Moderately phytotoxic

B-16

C-12
C-14

Mildly phytotoxic

A-02

B-03
B-05

Nonphytotoxic

A-08

C-03

Proportion phytotoxic

80%

100%

86%

Root growth was consistently the most affected endpoint (18 of 20 site soils); reduction in
both root length and mass was observed. This result is typical of responses to toxic metals
(Berry, 1977; Clark et al., 1981; Smith and Brenna, 1983, Atkinson, 1985; Fairley, 1985;
Fitter, 1985; Krawczyk et al., 1988; ICF, 1989; Alloway, 1990). Nutrient deficiency, in
contrast, typically causes increased root length and equivalent or slightly reduced root mass
relative to plants grown in nutrient sufficient conditions. The comparison of soil nutrient
concentrations indicated no significant differences between nitrate, potassium, or phosphorous
concentrations in impact and control soils (Table 3-5); therefore, neither the analytical data

RCG/Hagler Bailly

3-24

nor the observed phytotoxic response is consistent with a conclusion that the response was
caused by lack of nutrients in the site soils.

Reduction in root viability, however, does have implications regarding soil restoration and the
establishment and maintenance of a complex plant community. Legume and actinorrhizal
species (e.g., alfalfa, peas, conifers) form complex associations with soil biota, resulting in
root nodule formation (Curtis, 1983; Fitter and Hay, 1987). Root nodules, the loci of nitrogen
fixation, can facilitate development of nutrient- rich soil (Ludden and Burris, 1985, Evans et
al., 1985; Kapustka, 1987). Nodules (and symbiotic plant-bacterial associations) only form on
healthy root systems. The response of test species and the observed absence of cyanobacteria
(mosses, lichens, and crusts common to soil surfaces) in the impact areas indicate long-term
impairment of the capacity to support nitrogen fixation.

3.3.3.2 Statistical Relationships Between Phytotoxicity and Hazardous

Substances

Preceding sections of this chapter have demonstrated that soils collected from impacted areas
surrounding the Anaconda Smelter (Smelter Hill, Stucky Ridge, Mt. Haggin) contain
significantly elevated concentrations of hazardous substances relative to control soils, and are
significantly phytotoxic relative to control soils. Further, these phytotoxic impacted areas
have significantly reduced vegetative cover, including percent cover, number of vegetative
habitat layers, and presence of dominance types (see Chapter 4.0). The conclusion that
hazardous substances in upland soils have caused the observed injury is supported by the
following observations:

1. Metals concentrations exceed phytotoxic thresholds observed in other
studies.

Mean concentrations of arsenic, copper, and zinc from the impact areas are
substantially higher than those known to cause phytotoxic responses in controlled
laboratory experiments (Table 3-10). For example, the mean concentration of arsenic
in impact areas was 7-24 times higher than known phytotoxic concentrations, and the
mean copper concentration was 6-14 times higher than known phytotoxic
concentrations. Mean concentrations of cadmium and lead were higher than
phytotoxic concentrations reported in some studies.

2. Phytotoxicity is consistent with hazardous substance concentration and pH.

Figure 3-4 presents a scatter plot of copper and arsenic concentrations (the hazardous
substances with the greatest ratio of concentration to phytotoxic threshold, Table 3-9),
and soil pH. For each sample site, the degree of phytotoxicity is indicated with a
symbol (e.g., highly toxic sites appear as solid squares, nontoxic impact sites appear as

RCG/Hagler Bailly

3-25

Table 3-10
Soil Concentrations Reported to Cause Phytotoiicity and Ratios of Mean Impact Area

Concentrations to Phytotoxic Levels

Element

Minimum Concentrations
Causing Phytotoiicity

Ratio of Mean Impact Area Concentrations to
Phytotoxic Threshold1

Arsenic

Copper

Cadmium

Lead

Zinc

15 -50
60 - 125

3 -5

100 - 400

70 - 400

7.3 - 24.6
6.7- 14.1
2.3 - 3.0
0.4 - 1.8
0.9 - 5.0

1 Values > 1 indicate average soils concentrations greater than phytotoxic threshold.
Summarized by Kabata-Pendias and Pendias, 1992.

open triangles; control sites appear as open circles). Control sites have lower
concentrations of both arsenic and copper, while highly and severely toxic sites tend to
have more elevated concentrations of both hazardous substances. The figure also
demonstrates the well-documented mediating influence of soil pH on metal availability
and toxicity (Kabata-Pendias and Pendias, 1992): as pH increases (e.g., pH > 7), sites
are either mildly toxic or nontoxic (except for a single highly toxic site with extremely
elevated copper concentration).

3. Phytotoxicity is correlated with metals concentrations.

Statistically significant relationships were found to exist between concentrations of
hazardous substances and the degree of phototoxicity measured in impacted soils
samples. This dose-response relationship (i.e., degree of phytotoxicity positively
correlated with the concentration of hazardous substances) provides additional
evidence of the causal linkage between the hazardous substances in impacted soils
areas, and the observed injury to vegetation. In addition, as described in Chapter 4,
injuries to vegetation in upland areas near Anaconda were positively (and statistically
significantly) correlated with both the degree of phytotoxicity and concentrations of
arsenic and copper. These correlations provide further evidence relating observed
injuries in laboratory testing to conditions observed in the field.

Relationships between phytotoxicity scores and metal concentrations were examined
using nonparametric correlation (see Appendix B). This approach is not sensitive to
departures from normality and does not attempt to measure the degree of linearity of
the relationship between two variables. Correlations between phytotoxicity and

RCG/Hagler Bailly

c
o
u

X

o

c
o

z.

u

X

o
E-
>->

~ o

2 2 =

.c

o

X

o

t-
>>

0)

u

<D

>

CD
CO

(}<]►<«■♦

(UlS/S^0^

isddoo

o
to

^ >

c ed

c ~

^ —

51

o
in

u

o

<— en
U

"O —

o O
H is

» o

S °
53 "a

Efl

E 2

_ to

"2 o

— rs

c
u
o

c c

E °

ex .-
" o

1 8

CO
Q.
Q.

o
o

: c -o

O rz
C °

c c

u ■-

2 ?

E -

w

= c E

K 1) o

.Si O Ji

C . , CD

<w ^ ' ~

i/: <L> —

<< Kl CO

_ o ■£
£ x '53

St.

u 2

O "° X

E -S «

•" — "Z
H ° o

es E c

o
H

CO

D*

C3

3-27

biologically available metals (M^2 in impacted and control soils for all five metals
were highly significant (p < 0.005 in all cases; Table 3-1 l).3 The correlation between
pH and each metal was negative and highly significant (p < 0.001 in all cases).

Since the range of pH measured in impact samples was within the range that is
generally tolerable for plants, it is unlikely that pH had a direct influence on
phytotoxicity. Rather, the significant negative correlation between pH and
phytotoxicity is a by-product of the high degree of correlation between metals
concentrations and pH. To be conservative, partial correlations between phytotoxicity
and Mb were examined to account for any possible direct effect of pH (Table 3-12).
The partial correlations reflect the degree of relationship between metal concentrations
and phytotoxicity if pH were held constant. Significant partial correlations (p < 0.05)
were found between phytotoxicity and four of the five metals (arsenic, copper, lead,
and zinc). The partial correlation between phytotoxicity and cadmium was significant
at p < 0.06.

When impact soil samples only were included in the correlation analysis, correlations
between phytotoxicity and pH-adjusted metals concentrations remained positive and
highly significant (p < 0.01 in all cases except cadmium; Table 3-13). The highest
degree of correlation with phytotoxicity was for copper and arsenic; the lowest was for

A simplified transformation based on solubility relationships (Stumm and Morgan, 1980, Drever,
1982) was adopted to approximate the relationship of the bioavailable fraction of total metal, total metals,
and acidity for correlation analysis. Metals concentrations were adjusted for soil pH for each individual
site to derive a "surrogate" for bioavailability using the following transformation:

[Mb] = [MJ*[rT]*1012

where Mb represents the metal concentration adjusted for hydrogen ion concentration (pH), or the
bioavailability surrogate, and M, represents the actual measured total metals concentration. Multiplication
by the constant 10,: is an aid for representational purposes and has no effect on correlation coefficients or
significance levels.

A sensitivity analysis was performed on these correlations to determine whether they were robust to
variations in the toxicity scoring protocol used. Five alternative scoring protocols were evaluated: (1) use
of a binary scoring protocol where non-significant species-endpoint responses received a "0," and all
significant responses received a "1", (2) a linear scoring approach (scores equal to 0, 1, 2, 3, 4 depending
on the degree of response); and (3) three scoring systems which used a subset of the 18 species-endpoints
(total mass for all species; shoot height and root length for all species; and all six endpoints for wheat
only). As with the primary scoring system, correlations using the alternative scoring systems were positive
and highly significant (p < 0.01 for all elements and all scoring systems).

The results of this sensitivity analysis demonstrate that the significant correlations are robust to variations
in the scoring protocol, and are not a function of the specific scoring protocol that was employed.

RCG/Hagler Bailly

5-28

Table 3-11
Kendall-t Correlations Between Phvtotoiicity Score and [H+]-Adjusted Metal Concentration

(Mb)

Phytotcxicity
Score

Asb

Cub

Cdb

Pbb

Znb

Asb

048

Cub

0.52

0.90

Cdb

0.41

0.84

0.83

Pbb

0.48

0.90

0.85

0.88

Znb

0.44

0.92

0.87

0.88

0.90

[W]

0.39

0.80

0.71

0.73

0.81

078

p < 0.005 in all cases, under the hypothesis H0: x < 0 vs. Ha: t > 0, n = 28.

Kendall

Table 3-12
-t Partial Correlation Coefficients Between Phytotoxicity Score
Concentration (Mb), with Partialing Variable [H+], r

and Bioavailable Metal
=28

Variable

Partial t

Asb

0.31**

Cub
Cdb

Pbb
Znb

0.38**
0.20
0.30*
0.24*

* and **

1988).

indicate p

< 0.05 and p < 0.01, respectively (quantiles approximate,

Siegel and Castellan,

RCG/Hagler Bailly

3-29

Table 3-13

Kendall- 1 Correlations Between Phvtotoxicity Score and [H+J-Adjnsted Metal Concentration

<M„>

Phytotoxicity
Score

As„

Cub

Cdb

Pbb Znb

Asb

0.51**

Cub

0.52**

0.92**

Cdb

0.37*

0.79**

0.75**

Pbb

0.47**

0.87**

0.79**

0.83**

Znb

0.46**

0.88**

0.86**

0.84**

0.86**

[H+]

0.50**

0.79**

0.70**

0.70**

0.84** 0.73**

* and ** represent p < 0.05 and p <
n = 20.

0.01,

respectively,

under the hypothesis H0

: t < 0 vs. H,: t > 0,

cadmium. Correlations among the metals were again all positive and highly
significant (p < 0.001 in all cases), which is indicative of their common source (the
smelter at Anaconda). The partial correlation analysis indicated a significant positive
relationship between phytotoxicity and copper (p < 0.05) and a slightly weaker
relationship with arsenic (p < 0.10, Table 3-14).

Thus, even in the more conservative analysis (impact samples only), concentrations of
hazardous substances were found to be positively correlated with phytotoxicity, indicating the
causal connection between concentration of hazardous substances and injury to vegetation.

3.3.3.3

Previous Phytotoxicity Studies

Earlier greenhouse studies (Taskey, 1972) using soils collected in the Anaconda area
determined that growth of both lodgepole pine and Douglas fir were greatly reduced when the
combined concentrations of arsenic, copper, lead, and zinc were approximately 1,000 ppm or
greater. Taskey demonstrated that the phytotoxicity observed in the greenhouse could be
expected in soils within 5 miles of the smelter (Taskey, 1972).

Hartman (1976) sampled soils near the Anaconda smelter, on Weather Hill and near the Old
Works to examine the effects of cadmium, copper, lead, and zinc soil contamination on
fungal flora. He observed a consistent negative correlation between both fungal species
diversity and numbers of fungal propagules (a population count) and soil heavy metal

RCG/Hagler Bailly

3-30

Table 3-14

Kendall- 1 Partial Correlation Coefficients Between Phytotoxicity Score and Bioavailable Metal

Concentration (Mb), with Partialing Variable [H*), n = 20

Variable Partial t

Asb 0.21*

Cut, 0.27**

Cdb 0.03

Pbb 0.09

Znk 0.15

** and * indicate p < 0.05 and p < 0.10, respectively (quantiles approximate, Siegel and Castellan,
1988).

concentrations, as metals concentrations increased, the diversity of microflora] species and the
number of propagules decreased. Variation in other abiotic growth factors that influence
fungal growth and reproduction were assessed and failed to correlate with the observed
diversity and population data. In addition, Hartman noticed that conifer roots in the
Anaconda area lacked essential symbiotic micorrhizae, and attributed the failure of the
Anaconda Company's revegetation attempts (possibly referring to reforestation and sporadic
seeding attempts by the Anaconda Minerals Company in the 1960s and 1970s) to reduced
fungal diversity and propagule density.

The Hartman (1976) study is additional confirmation that soils in the impact areas were
phytotoxic because of elevated concentrations of hazardous substances. Reduced fungal
diversity and propagule density have implications for biochemical cycling in soils. Soil
microorganisms are ecologically important in that they function in production, consumption,
and transportation within the soil ecosystem, and are involved in the flow of energy and in
chemical cycling in the soil (Kabata-Pendias and Pendias, 1992). Decomposition of soil
organic matter involves numerous enzymatic stages, in which many species of fungi may be
required to perform in sequence for optimum organic matter decomposition. The study
concluded that high concentrations of cadmium, copper, lead, and zinc caused impoverished
microfungal communities and population reductions, which in turn resulted in depressed
nutrient cycling and incomplete decomposition of plant and animal residues in Anaconda area
soils (Hartman, 1976).

RCG/Hagler Bailly

3-31

3.3.3.4 Implications of Phototoxicity

In an ecological context, phytotoxic soils have serious implications regarding species
diversity, vegetative structural composition, and the nutritional quality of vegetation
(Appendix B). The accumulation of phytotoxic substances in surface soils will continue to
affect plant establishment and growth, particularly of native species, indefinitely. Unlike the
transient effects of fire, logging, or transient and ephemeral air toxins, metal contamination in
soils is virtually permanent (Kabata-Pendias and Pendias, 1992) In addition, preclusion of
plant growth and normal decomposition processes further prevents the process of soil
formation, which might aid in recovery of soils. Chapter 4.0, Vegetation and Wildlife
Resources, discusses vegetation population and community-level effects of phytotoxic soils.

3.3.4 Extent of Injury

3.3.4.1 Areal Extent of Injury

The areal extent of injury to upland soils encompasses 17.8 square miles (46 square km) and
includes the following lands:

*> 3.8 square miles on and immediately adjacent to Stucky Ridge, a de-vegetated

ridge north of the Old Works

► 7.3 square miles on Smelter Hill, including Weather Hill and slopes to the
south and west of the stack

► 6.7 square miles in the Mount Haggin Wildlife Preserve, south and west of
Mill Creek, extending toward the Continental Divide.

Soils within these upland impact areas were determined to be phytotoxic in comparison to
control area soils, and thus injured.

3.3.4.2 Volume of Injured Soils

The volume of injured soil can be estimated using the aerial extent of the impact area (17.8
square miles) and the depth of contamination measured. Based on the results of the sampling
described in Section 3.3.1.1 and Appendix A, surface soils (0-2 inches) exhibit the greatest
contamination in the Anaconda uplands. Thus the volume of contaminated 0-2 inch soil in
the impact areas defined exceeds 3 million cubic yards (2.3 million cubic meters). This
volume of soil has been shown to cause injury to biologic resources; this number most likely
under-represents the amount of injured soil in the uplands because of the convoluted terrain,

RCG/Hagler Bailly

3- j 2

the difficulty in determining the true areal extent of the delineated impact areas, and the fact
that the study boundaries do not include all injured soils.

If the estimate is extended to soils within a 6-mile radius of the stack that exhibit, on averace,
concentrations elevated 10 times above baseline, then it may be estimated that 1 9.5 million
cubic yards (14.9 million cubic meters) of 0-2 inch soils have been contaminated with
hazardous substances emitted from the Anaconda smelter. If all soils exposed to smelter
deposition that are known to exhibit elevated concentrations of hazardous substances are
included in the calculation (0-2 inch soils within a 20-mile radius), the volume of
contaminated soil becomes 217 million cubic yards (166 million cubic meters).

These estimates fail to account for concomitant injury caused by soil material lost in the past
90 or more years to accelerated erosion (see Section 3.3.4), or the injury to additional areas
and resources caused by the redistribution of contaminated eroded soils.

As part of the Anaconda Smelter RI/FS, Tetra Tech (1987) attempted to estimate the mass of
soil arsenic attributable to stack emissions using depth-averaged arsenic concentrations. The
estimate was based on a 12-inch mixing zone, a dry density of 1.5 g/cm2, and data from 22
sampling stations where arsenic concentrations had been measured at more that one depth
within the top 12 inches of soil. Based on the results, the mass of arsenic in soils attributable
to smelter stack emissions was estimated to be within an order of magnitude of 20 million
kilograms (22,000 tons) within an affected land area of approximately 68,000 acres (106
square miles). The estimate was limited to the areas sampled, and does not take into account
areas where sunficial soil arsenic concentrations are above background (about 10 mg/kg, in
Tetra Tech, 1987) but below 100 ppm (Tetra Tech, 1987).

3.3.5 Ability of Soil Resource to Recover

Natural recovery of upland soils in the mountainous terrain north, west, and southwest of the
smelter will require many hundreds, if not thousands, of years. Estimations of trace metal
persistence in soils have indicated that complete removal (to baseline levels) of contaminants
from soils through natural recovery processes is nearly impossible (Bowen, 1979, Kabata-
Pendias and Pendias, 1992). In addition, though soils are a sink for trace elements,
remobilization processes, induced by changes in soil pH or the moisture regime, for example,
will ultimately transfer elements into biogeochemical cycles, and consequently disrupt the
flow of elements in the soil-plant-animal ecosystem.

Processes by which trace elements are removed from soils include:

*■ Leaching to subsurface soil horizons

RCG/Hagler Bailly

► Erosion (which merely re-distributes the contamination and constitutes a re-
release of hazardous substances)

*■ Plant uptake or adsorption in soil organic matter (a temporary immobilization)

► Dilution of soil through natural chemical weathering and biological soil
formation processes.

Leaching of hazardous substances to subsurface layers in the semi-arid montane climate of
the Anaconda area is apparently very slow. Taskey (1972) observed that most of the
contamination was concentrated in the top 6 to 8 inches of the soil profile. Considering that
the loadings had occurred at that time for 88 years, the amount of leaching is relatively small
(Taskey, 1972). Similarly, 0-2 inch samples collected in June 1992 exhibited significantly
greater concentrations of hazardous substances than 0-6 inch soils (Appendix C), which also
suggests that leaching is proceeding slowly. Leaching of contaminants from Anaconda area
soils cannot be expected to be a viable mechanism of soil recovery.

Erosion of unvegetated slopes, however, is occurring at tremendous rates. Entire root
systems of dead conifers are exposed on hillsides in the Mount Haggin area; erosion gullies 6
to 20 feet deep scar the north and east sides of Stucky Ridge; a cobbled pavement is exposed
at the surface of wind-scoured slopes on the south side of Stucky Ridge and on much of
Weather Hill. The cobbled surfaces stabilize erosion somewhat, but are not conducive to
development of an organic layer, and despite the extensive erosion that has obviously
occurred, metals levels in soils underlying the cobble pavement remain significantly elevated
relative to baseline.

On the A and C hills adjacent to the smelter stack, the estimated amount of soil eroded in the
past 75 years has been calculated as approximately 16 tons/acre/year (Gariglio, 1984, SCS,
unpublished letter). The 1984 calculated rate of erosion for the steeper areas of the hills,
using the Universal Soil Loss Equation, was 60 tons/acre/year (Gariglio, 1984, SCS,
unpublished letter). Soil is normally replaced in similar arid environments by natural
weathering and biotic processes at 3 tons/acre/year. In the absence of vegetation
establishment to stabilize these soils, erosion will continue to re-distribute contaminated soils
and degrade the upland landscape. Therefore, erosion of contaminated soils has not served as
a recovery process in the Anaconda uplands.

The final two natural processes by which soils may recover, uptake by plants soil organic
matter, and dilution, will be precluded by the current phytotoxic concentrations. In the more
severely contaminated areas, natural re-establishment of vegetation has been inhibited or
precluded. Vegetation is slowly re-establishing in parts of the impact area, but at a very slow
rate, and community development is impoverished (see Chapter 4.0). Plant species that are

RCG/Hagler Bailly

3-34

recolonizing the area are predominantly noxious weed species, species that have evolved local
tolerance to elevated metals concentrations, and species that do not depend on seed
germination for regeneration.

In conclusion, natural recovery of the upland soil resource would be highly variable,
depending on topography, degree of contamination, the composition of the parent material,
and rates of future soil formation, but it is likely that over the enure impact area, recovery
periods would range from hundreds to many thousands of years.

3.4 SUMMARY

Overall, the data presented in this chapter demonstrate that:

»■ Concentrations of hazardous substances in exposed soils are significantly

greater than baseline concentrations.

»> Concentrations of soil nutrients (nitrate-nitrogen, potassium, and phosphorous),

CEC, pH, and percent sand, silt, and clay in control and impact soils are not
significantly different.

► Exposed soils are phytotoxic relative to control soils, and hence are injured.

► The degree of phototoxicity observed in controlled laboratory experiments is
correlated with the concentrations of hazardous substances in the soil.

► The observed effects on plant growth are consistent with the phytotoxic effects
of heavy metals.

► Concentrations of hazardous substances in exposed soils exceed phytotoxic
thresholds identified in the literature; the observed phytotoxic response
therefore is consistent with scientific literature.

»• The areal extent of severe injury to soils is estimated to be approximately 17.8

square miles.

3.5 REFERENCES

Adriano, DC. 1986. Trace Elements in the Terrestrial Environment. Springer- Verlag, Inc.,

NY.

Alloway, B.J. (ed.) 1990. Heavy Metals in Soils. John Wiley and Sons, Inc., NY 339 pp.
RCG/Hagler Bailly

3-35

ASTM, 1994. Practice for Conducting Early Seedling Growth Test, El 598.

Atkinson, D. 1985. Spatial and temporal aspects of root distribution indicated by the use of
a root observation laboratory. In A.H. Fitter, D. Atkinson, D.J. Read, and MB. Usher,
Ecological Interactions in Soil: Plants, Microbes, and Animals, pp. 43-65. Special
Publication Series of the British Ecological Society No. 4. Blackwell Scientific Publications.

Baker, D.E. 1990. Copper. In B.J. Alloway (ed), Heavy Metals in Soils. John Wiley and
Sons, Inc., NY, pp. 151-176.

Berry, W.L. 1977. Dose response curves for lettuce subjected to acute toxic levels of copper
and zinc. In H. Drucker and RE Wildung, Biological Implications of Metal in the
Environment. ERDA Symposium Series 42:358-369. Published by Technical Information
Center ERDA.

Bowen, H.M.J. 1979. Environmental Chemistry of the Elements. Academic Press, London.

Clark, R.B., PA. Pier, D. Knudsen, and J.W. Maranville. 1981. Effect of trace element
deficiencies and excesses on mineral nutrients in sorghum. J. Plant Nutrition 3: 357-374.

Curtis, H. 1983. Biology. Worth Publishers, Inc., NY.

Drever, J.I. 1982. The Geochemistry of Natural Waters. Prentice-Hall, NJ.

Evans, H.J., P.J. Bottomley, and W.E. Newton. 1985. Nitrogen Fixation Research Progress.
Martinus Nijh off Publishers.

Fairley, R.I. 1985. Grass root production in restored soil following opencast mining. In
A.H. Fitter, D. Atkinson, D.J. Read, and M.B. Usher, Ecological Interactions in Soil: Plants,
Microbes, and Animals, pp. 81-87. Special Publication Series of the British Ecological
Society No. 4. Blackwell Scientific Publications.

Fitter, A.H. 1985. Functional significance of root morphology and root system architecture.
In A.H. Fitter, D. Atkinson, D.J. Read, and MB. Usher, Ecological Interactions in Soil:
Plants, Microbes, and Animals, pp. 87-106. Special Publication Series of the British
Ecological Society No. 4. Blackwell Scientific Publications.

Fitter, A.H, and R.K.M. Hay. 1987. Environmental Physiology of Plants. Academic Press,
NY.

Gariglio, F. Unpublished letter to F. Steadier, RC&D Forester, Anaconda, MT. Dated April
26, 1984.

RCG/Hagler Bailly

3-36

Gorsuch, J.W., R.O. Kringle, and K.A. Robillard. 1990. Chemical effects on the germination
and early growth of terrestrial plants. In W. Wang, J. Gorsuch, and W. Lower (eds), Plants
for Toxicity Assessment. ASTM STP 1091. American Society for Testing Materials,
Philadelphia, PA.

Green, J.C., C.L. Bartels, W.J. Warren-Hicks, BR. Parkhurst, GL. Linder, S.A. Peterson, and
WE. Miller. 1989. Protocols for Short Term Toxicity Screening of Hazardous Waste Sites.
U.S. EPA Environmental Research Laboratory, Corvallis, OR. EPA/600/3-88/029.

Hartman, L.M. 1976. Fungal Flora of the Soil as Conditioned by Varying Concentrations of
Heavy Metals. University of Montana, Ph.D. Dissertation, Univ. Montana, Missoula. 147 pp.

Heale, E.L. and D.P. Olmrod. 1982. Effects of nickel and copper on Acer rubrum, Cornus
stolenifera, Lonicera tatarica, and Pinus resinosa. Can. J. Bot. 60: 2674.

ICF Incorporated. 1989. Scoping Study of the Effects of Soil Contamination on Terrestrial
Biota. Final report prepared for the Office of Toxic Substances, U.S. EPA.

Kabata-Pendias, A. and H. Pendias. 1992. Trace Elements in Soils and Plants. 2nd ed.
CRC Press, Ann Arbor, MI.

Kapustka, L.A. 1987. Interactions of plants and nonpathogenic soil micro-organisms. In
D.W. Newman and K.G Wilson, Models in Plant Physiology and Biochemistry. Vol JH, pp.
49-55. CRC Press.

Kiekens, L. 1990. Zinc. In B.J. Alloway (ed), Heavy Metals in Soils. John Wiley and
Sons, Inc., NY, pp. 261-279.

Krawczyk, D.F., S.F. Fletcher, ML. Robideaux, and CM. Wise. 1988. Plant growth
experiments in Bunker Hill and zeolite-amended soil. U.S. EPA Environmental Research
Laboratory, Corvallis, OR.

Ludden, P.W. and J.E. Bums. 1985. Nitrogen fixation and CO 2 metabolism. Elsevier Press.

Morrison-Knudsen. 1991. Engineering Evaluation/Cost Analysis: Silver Bow Creek NPL Site,
Butte Priority Soils Operable Unit (Final Draft). Prepared by Morrison-Knudsen Corporation
for ARCO to be submitted to U.S. EPA.

Newman, E.I. 1983. Interactions between plants. In OL. Lange, P. Nobel, C.B. Osborne and
H. Ziegler (eds), Encyclopedia of Plant Physiology Vol 12C: Physiological Plant Ecology III.
Springer-Verlag, Berlin, 679-710.

O'Neill, P. 1990. Arsenic. In B.J. Alloway (ed.), Heavy Metals in Soils. John Wiley and
Sons, Inc., NY, pp. 83-99.

RCG/Hagler Bailly

3-37

PTI. 1991. Smelter Hill Remedial Investigation and Feasibility Study: Preliminary Site
Characterization Information. Prepared by PTI Environmental Services for ARCO, Anaconda,
MT. 121 pp.

PTI. 1993. Smelter Hill Remedial Investigation and Feasibility Study. Phase II Data
Summary/Data Validation/Data Usability Report. Prepared for ARCO by PTI Environmental
Services, Butte, MT.

Siegel, S. and N.J. Castellan, Jr. 1988. Nonparametric Statistics for the Behavioral Sciences.
McGraw-Hill, New York.

Smith, G.C. and E.G. Brenna. 1983. Cadmium-zinc interrelationships in tomato plants.
Phytopathology 73: 879-882.

Stumm, W. and J.J. Morgan. 1980. Aquatic Chemistry: An Introduction Emphasizing
Chemical Equilibria in Natural Waters. Wiley-Interscience, NY.

Taskey, R.D. 1972. Soil contamination at Anaconda Montana: History and influence on
plant growth. M.S. Thesis, University of Montana, Missoula, MT.

Tetra Tech. 1987. Anaconda Smelter Remedial Investigation/Feasibility Study (Draft
Report). Prepared for Anaconda Minerals Company. Document Control No. TTB 173 Dl.
217 pp.

U.S. DOI. 1987. Approaches to the Assessment of Injury to Soil Arising from Discharges of
Hazardous Substances and Oil. Type B Technical Information Document. Prepared by
Pacific Northwest Laboratories, Richland, WA.

Van Assche, F. and H. Clijsters. 1990. Effects of metals on enzyme activity in plants:
Commissioned review. Plant Cell Environ. 13: 195-206.

Woolhouse, H.W. 1983. Toxicity and tolerance in the response of plants to metals. In O.L.
Lange, P. Nobel, C.B. Osborne and H. Ziegler (eds), Encyclopedia of Plant Physiology Vol
12C: Physiological Plant Ecology III. Springer- Verlag, Berlin, pp. 245-300.

RCG/Hagler Bailly

4-1^

4.0 UPLAND BIOLOGICAL RESOURCES — VEGETATION,
WILDLIFE, AND WILDLIFE HABITAT

Upland biological resources include vegetation, wildlife, and wildlife habitat. This chapter
describes and quantifies injuries to these biological resources in the upland areas adjacent to
the town of Anaconda.

Chapters 2.0 and 3.0 demonstrated that the soils of these upland areas are contaminated with
hazardous substances at concentrations that are toxic to plants. The results presented in this
chapter demonstrate that vegetation, wildlife, and wildlife habitat have been injured in the
upland impact areas surrounding Anaconda as a result of exposure to these contaminated
soils. Specifically:

►■ Exposure to phytotoxic soils has resulted in gross modifications to, or the

elimination of, vegetation communities at impacted sites. The extent of forest
and grassland has been significantly reduced at impact sites relative to
matching controls; the extent of bare ground is significantly greater at impacted
sites.

► The number of habitat layers has been significantly reduced in the impacted
areas.

► Habitat has been significantly reduced in extent and quality for marten (an
indicator species for forest habitat) and elk (an indicator species for forest
edge/grassland habitat). The viability of wildlife populations that depend on
these habitat types therefore has been reduced in the injured areas.

► As a result of habitat loss and degradation, a large number of wildlife species
that depend on this habitat are likely to have been lost from the injured area or
have undergone population reductions.

► Loss of vegetation and wildlife habitat is caused by exposure to hazardous
substances. Lack of vegetative cover is positively correlated with hazardous
substance concentrations and the degree of phytotoxicity observed in controlled
laboratory tests.

No causal factor other than hazardous substance exposure consistently and plausibly explains
conditions in the injured area.

RCG/Hagler Bailly

4-2

4.1 DESCRIPTION OF UPLAND BIOLOGICAL RESOURCES

Upland biological resources include vegetation, wildlife, and the wildlife habitat provided by
vegetation communities in the upland impact areas surrounding Anaconda (i.e., Stucky Ridge,
Smelter Hill, and Mount Haggjn) that were described in preceding chapters. The native
vegetation in the uplands surrounding Anaconda comprises species adapted to semi-arid
montane conditions; typical vegetation includes mesic or xeric species of trees, shrubs, forbs,
and bunchgrasses (Pfister et al., 1983; Mueggler and Stewart, 1980). Likewise, wildlife
species associated with these upland habitats are those adapted to semi-arid montane
conditions and the vegetation communities typical of such areas.

4.2 INJURY DEFINITION

Injury to upland biological resources is defined as:

*■ ... adverse changes in viability: death, . . . physiological malfunctions

(including malfunctions in reproduction), or physiological deformations [43
CFR§ 11.62(f)(i)].

In the Clark Fork Basin, injury to vegetation that resulted from releases of hazardous
substances is expressed in the complete elimination of vegetation, or changes in the
composition, structure, and/or distribution of vegetation communities. At the level of the
individual plants, these community changes have been caused by death and physical
deformations (i.e., reduced growth leading to a loss in viability).

Plant death and reduced growth have been found in this assessment to satisfy the four
acceptance criteria for biological responses [43 CFR § 11.62 (f) (2) (i-iv)]. Specifically, plant
death and reduced growth:

* Are often the result of exposure to hazardous substances, as shown in various

scientific studies, including the phytotoxicity studies described in Chapter 3.0
and Appendix B.

►• Have been shown to occur in controlled laboratory experiments (see

Chapter 3.0 and Appendix B). Death and reduced growth occurred in plants
grown on soils collected from the impacted upland areas.

*■ Are often the result of exposure to hazardous substances among free-ranging

organisms, as documented in the literature (see Chapter 3.0), and in the field
studies described in this chapter and in Appendix C.

RCG/Hagler Bailly

4-3

► Are routine measurements that are practical to perform and produce

scientifically valid results.

Overall, injury (death, reduced growth relative to controls) to vegetation has been confirmed
by the results of phytotoxicity studies and by the observed loss of vegetation in exposed areas
(see Section 4.2.3).

The viability of upland wildlife resources in the upper Clark Fork Basin has been reduced by:

*■ Reductions in habitat quantity and quality for selected indicator species [43

CFR § 11.63 (f)(4)(ii)(A)] relative to uncontaminated control areas [43 CFR §
11.72(d)(1)].

Since the quality and quantity of upland wildlife habitat are defined largely by vegetative
community characteristics, which, in turn, are a function of the relative viability of individual
plants, the first three acceptance criteria listed are also pertinent to wildlife injury. The final
criterion has also been met: routine measurements of wildlife habitat quality and quantity can
be performed using habitat evaluation procedure (HEP) models developed by U.S. Fish and
Wildlife Service and identified as appropriate models for use in NRDAs [43 CFR §11.71
(1)(8); U.S. DOI, 1987].

The following sections describe the relationship between vegetation and wildlife habitat, and
methods of assessment of injury to wildlife habitat.

4.3 ASSESSMENT METHODOLOGIES

4.3.1 Vegetation Assessment

The soils of the Stucky Ridge, Smelter Hill, and Mount Haggin impact areas have been
shown to be phytotoxic (Chapter 3.0); the cumulative result of phytotoxic responses among
individual plants has been expressed as changes in community composition and structure (e.g.,
species composition, dispersion, or percent cover) [43 CFR § 11.71 (1)(6)] of the indigenous
upland vegetation communities.

Injury to vegetation communities was assessed by statistically comparing vegetation
measurements made at study sites in the area impacted by smelter emissions and at paired
control sites. The following vegetation variables were measured:

•• Proportional representation of cover types (e.g., coniferous forest, deciduous

forest, grasslands)

RCG/Hagler Bailly

4-4

► Proportional representation of habitat layers (e.g., tree canopy, tree bole, shrub
layer, terrestrial subsurface)

► Percent cover of vegetation and individual plant species.

Vegetative cover then was related to the observed phytotoxicity of site soils as well as
concentrations of hazardous substances to confirm the causal relationship.

4.3.2 Wildlife Habitat Assessment

The distribution of wildlife species reflects differing habitat requirements, including
preferences for thermal and hiding cover, and breeding and foraging sites. Certain species
have very restrictive habitat requirements (habitat specialists), while others are less specific
(habitat generalists). The difference between habitat specialists and generalists is one of
degree, because even the generalist species have well-defined, though less restrictive,
preferences for food, water, cover, reproduction, and other needs.

Plant community structure and distribution are generally the primary determinants of whether
an area provides habitat suitable for terrestrial wildlife species (Cooperrider et al., 1986).
Most species have habitat requirements that specify comparatively narrow ranges of
vegetational parameters (e.g., many bird and mammal species require mature conifer forest
types). While the vegetational characteristics of an area may not capture all of the
environmental parameters important to a particular animal species, vegetation normally
defines the baseline quality of the habitat.

At the species level, the value of habitat can be quantified using a set of measurable habitat
variables. Measurable information concerning the quality, or suitability, of habitat for a
species is often derived from vegetational cover types used by the species (U.S. FWS, 1981).
Aspects of a species' habitat preferences that are difficult to measure or quantify can often be
approximated using a related vegetation parameter; for example, if food availability for a
species is determined by small mammal abundance, and it is known that small mammal
abundance can be approximated by the percent herbaceous cover, then herbaceous cover can
be measured as an index of food availability (U.S. FWS, 1981).

When comparing habitat integrity and suitability for wildlife at selected study sites, a useful
approach is to select indicator species whose habitat requirements are shared by a wide range
of species. For example, many avian species are dependent on mature conifer canopy.
Evaluating the habitat in an area for one such species gives an indication of its more general
suitability for all canopy-dependent species.

Habitat Evaluation Procedure (HEP) models were used to assess injury to upland wildlife
habitat in the upper Clark Fork Basin as recommended in 43 CFR § 11.71 (1)(8) and

RCG/Hagler Bailly

4-5

U.S. DOI (1987). HEP models are based on the relationships between the ecological features
of an area and its ability to provide habitat for wildlife. In HEP models, the important
features of an area of habitat that influence its suitability for the model species are identified,
and quantitative relationships that describe the availability of those features and the suitability
of the habitat are established. HEP models provide indices of the potential carrying capacity
of an area of habitat for the model species.

The selection of HEP models for this injury assessment was based on the types of habitats
that were initially identified as potentially injured. Initial field observations and discussions
with local ecologists and wildlife managers indicated that upland plant associations potentially
impacted by hazardous substances included coniferous forests [primarily Douglas fir
(Pseudotsuga menziesii) or lodgepole pine (Pinus contorla)], and montane meadow and forest
edge (i.e., interspersed grassland and forest patches).

4.3.2.1 Forest Habitat: Indicator Species = Marten

Injury to the forest habitat types was determined using marten (Martes americana) as an
indicator species. Marten are primarily arboreal mammals and in Montana are confined to
forested areas, particularly spruce-fir forests (Koehler and Hornocker, 1977; Hawley and
Newby, 1957; Weckwerth and Hawley, 1962) or Douglas fir and lodgepole pine stands
(Fager, 1992). A HEP model for marten winter habitat published by the U.S. Fish and
Wildlife Service (Allen, 1984) for the boreal forest biome and Rocky Mountain forests of the
western United States (Allen pers. comm.) was used in this assessment. Much of the data
used in the development of the model were obtained during habitat studies in Montana,
making it particularly suitable for use in the assessment area

Variables measured to evaluate winter habitat quality for marten included:

► Percent tree canopy closure. Only forested areas are suitable marten habitat.

► Percent of the canopy that is fir or spruce. Coniferous forests, including
Douglas fir and lodgepole pine types, are preferred over deciduous forests.

► Successional stage of the forest stand. Mature or old growth stands are
preferred over earlier successional stages (Koehler and Hornocker, 1977).

► Percent of the ground surface covered by fallen timber greater than 3 inches in
diameter. This parameter serves as a proxy measure for the availability of the
marten's rodent prey (Allen, 1984).

The suitability of marten habitat was assessed in impact and control areas, and resulting
values were compared statistically to determine whether significant differences exist.

RCG/Hagler Bailly

4-6

4.3.2.2 Forest Edge and Grassland Habitat: Indicator Species = Elk

Injury to forest edge and grassland habitat was determined using elk (Cervus elaphus) as an
indicator species. In southwest Montana, elk depend on forests for security, and grassland
vegetation provides the mainstay of the elk diet. An elk HEP model describing winter habitat
suitability was developed for this assessment (see Appendix C).

Variables measured to evaluate winter habitat suitability for elk included:

► Availability and quality of forest cover

► Availability and quality of food, measured as percent cover of palatable forage
species.

4.3.2.3 Habitat Layers

In addition to the individual species models, a third HEP model, the layers of habitat model
(Short, 1984), was used to address the relationship between the vertical complexity of
vegetation communities and their capacity to provide habitat for a diversity of wildlife.
Habitats that are structurally complex (i.e., that have many habitat layers) generally support a
more diverse fauna than structurally simple habitats. This relationship has been shown for
birds (MacArthur and MacArthur, 1961; Cody, 1975), reptiles (Pianka, 1967), fish (Tonn and
Magnuson, 1982), and molluscs (Harman, 1972). The layers of habitat model quantifies this
relationship, and habitats that are structurally complex (i.e., able to support diverse wildlife
populations) score higher than less structurally complex habitats.

In this assessment, the layers of habitat model was used with information about the habitat
preferences of individual species to classify wildlife species by vegetation layer "guilds," or
groups of species whose niches occur in the same vegetation layers or groups of vegetation
layers (Short, 1984). This, together with vegetation structure measurements made in impact
and control areas, allowed the identification of species that are likely to have been become
locally extinct or have reduced viability in the impacted areas because of the demonstrated
changes in habitat structure.

Using the HEP models described above, measurements of habitat suitability at study sites
within the area impacted by smelter emissions were compared with measurements made at
paired control sites in German Gulch.

RCG/Hagler Bailly

4-7

4.3.3 Baseline Conditions

Injury to wildlife habitat was defined relative to control areas [43 CFR § 11.72 (d)(1)].
Baseline conditions are those that would be expected to be present in the Stucky Ridge,
Smelter Hill, and Mount Haggin impact areas in the absence of releases of hazardous
substances. Baseline conditions are not necessarily pristine, since other nonmining
anthropogenic activities (e.g., timber harvesting and livestock grazing) are likely to have
occurred in the impacted areas in the absence of mining-related injury. Baseline conditions
for the injured areas were evaluated using two methods: historical evidence, and observations
made at the German Gulch control areas described in Chapter 3.0.

4.3.3.1 Historical Information

Historical documents describe the high-quality wildlife habitat previously provided by the
now-injured uplands adjacent to Anaconda. The mountains south of Anaconda were "thickly
timbered with pine and fir trees, the meadows covered with knee-deep bunchgrass . . . deer,
elk, mountain goats and mountain lions were so abundant that ranchers did not need to hunt
for smaller animals" (in Wolle, 1963); hill "C" (overlooking Anaconda to the south) was
"heavily timbered at that time with pine trees some three to four feet in diameter" (Deer
Lodge County Historical Group, 1975). A later document states, "Prior to the late 1800s the
hills locally referred to as the A and C hills were vegetated by a mosaic of Douglas fir and
aspen stands intermixed with grasslands." (USDA, 1986). Historical photographs (USDA,
1986) show that in the 1880s the upper elevations of Smelter Hill and the adjacent hills to the
west were covered in conifer forest, which extended as "stringers" down the northern aspects
of these hills to the valley floor at 5,000 feet. Stumps and dead standing and fallen trees are
evidence of previous forest cover in the Anaconda area uplands. Therefore, the historical
documentation indicates that the injured area was vegetated with a mixture of coniferous
forest and grasslands.

4.3.3.2 Control Areas

The German Gulch area was selected as the overall control area based on its similarity in
elevation and aspect, and its proximity to the impacted area (Figure 4-1). Historical accounts
and photographs were consulted to confirm that German Gulch and the Anaconda uplands
(before mining-related injury) supported similar vegetation. Like the injured upland areas, the
control areas of German Gulch were clear-felled of timber in the early part of this century,
and timber harvesting continued until 1991 (M. Frisina, Montana Department of Fish,
Wildlife, and Parks, pers. comm). Grazing by livestock is a continuing land use in the
control area, and the presence of fire scars attests that burning has occurred in the past.
Based on current land use patterns in the German Gulch area, logging, grazing by livestock,
and fires would be expected in the impact area in the absence of mining-related injury.

RCG/Hagler Bailly

4-8

Stucky Ridge

□ Sampled in 1992
■ Sampled in 7 99 J
ZlControl Sites

*' Mount Hassm

Kilometers

0 l 2 3

AA3.AA8

BB9 ".BBS/BBIS /
.4.4 6a A ABB 16 'f

ccr^/- -BBS

BB7 ~^ AA2 A

BB13 . AA10

CC9,

~BB5^ CC7 ^3

CCS

a

BB11

CC1

/

ItoC/tfagter Bully

j*

Figure 4-1. Upland Impact and Control Vegetation Sampling Sites.

RCG/Hagler Bailly

4-9

Within the overall German Gulch area, individual sampling sites were selected as a paired
control for each 1992 sampling site in the injured upland area. Each control site was chosen
so that it resembled its respective impact sampling site in the environmental features that
influence vegetation community structure and composition, and thus habitat (Table 4-1).

Vegetation measurements made at the control sites are consistent with the historical
documentation cited previously. Specifically, sites selected as controls for the Stucky Ridge
area are dominated by grasslands, intermixed with evergreen forest, shrubland, and deciduous
forest (Figure 4-2). Sites selected as controls for the Smelter Hill area are dominated by
coniferous forests interspersed with grasslands (Figure 4-3). Sites selected as controls for the
Mount Haggin area are dominated by coniferous forests (Figure 4-4).

Vegetation and habitat measurements made in the control area also confirm that the baseline
conifer forests and open grasslands continue to provide high quality habitat for a rich
diversity of animal species. At least 32 avian species and 10 mammal species are
characteristic of conifer forests in southwest Montana (Tables 4-2 and 4-3). The avian
species include game birds such as blue and spruce grouse, and birds of prey; the mammal
species include elk, marten, black bear, and bobcat. Over 70% of the native species
characteristic of the area were observed in the control area during fieldwork conducted in
1992, indicating that current or recent land use practices (logging and livestock grazing) in
the control area do not eliminate wildlife habitat quality or availability.

4.4 INJURY DETERMINATION AND QUANTIFICATION

This section provides evidence confirming and quantifying reductions in wildlife habitat
quantity and quality in the injured upland areas in the vicinity of Anaconda.

4.4.1 Previous Investigations

Early (1886) photographs of the Smelter Hill area show that, prior to injury, the northern
slopes of Smelter Hill were vegetated with forests and grasslands. Sporadic attempts have
been made to revegetate limited portions of the impacted area to reduce erosion on the hills
south of the town of Anaconda (USD A 1986; Headwaters RC&D Forester, 1992; NPI, date
unknown). Between 1960 and 1967, AMC planted more than 60,000 lodgepole pine and
Douglas fir near Mill Creek, on the A Hill, and in Sheep Gulch (NPI, date unknown).
Between 1971 and 1977, AMC planted 16,500 lodgepole pine, ponderosa pine {Pirtus
ponderosa), and unspecified "trees" on the A and C Hills and in Sheep Gulch (NPI, date
unknown). All locations recorded are vague. Apparently, the attempts by AMC to reforest
met with marginal success; subsequent efforts to revegetate the same general areas were
initiated in 1985 (Headwaters RC&D Forester, 1992).

RCG/Hagler Bailly

o

I

I

*o

v*

ro

«n

1

OS

C>)

t-.

ro

so

CN

SO

<N

so

so

<n

rv)

w-i

so

©\

CJ

s
e

m m

O

©

mm

©'

—

— ;

O

— -

©

O

o

O

©'

—

©

Csl

88

U

e

OS

u

c

£

e

«*

■— »

w
0

u

88
Q.

E

<n

w>

cs

o

v>

(N

OS

rvl

SO

vq

^~

SO

f»i

^^

<N

so

■*

<N

i

a

88

n

O

©'

©'

o

©

©'

©'

©

— '

©'

(N

•9

o

m

e

Ctf

L.

CJ

e

£

u

e

e
e
U

2

2

O

o

2

o

2

00

oo

a

a

OO

2

2

2

2

2

o

oo

2

2

8 5

u

«=

§00

e.
e

.Bo

(A

a i

w

88
C

2

2

a

o

2

2

2

2

00

o

a

OO

S

a

00

oo

2

a

oo

2

O

*? .-^

E

•1 w e

■""

■■■ m *-

.aoo o

88 ► 08

L- — ' C

e

L.

uj

>

UJ
OO

^

UJ

J^

>

£

UJ

2

<"S

E
C

oo

UJ

z

UJ

UJ

OO

2

UJ

2

j|

00

2

il

If

£

2

2

UJ

• -

—

u

*

*~

^

£"*

w

-«~ da

u

c.

<

♦j

of Eleva

83
C

E

UJ
OO

00

UJ

2

oo

UJ

1/3

£

UJ

UJ

&o

£

ai

1

UJ

2

UJ

i

£

|

%

^

UJ

2

UJ

2

UJ

_u

e
e

e

o
00

o

O

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

o

©

©

©

II

Lb

88

e.

^
t.

o

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

©

o

o

v>

t

(N

m

©

<n

—-

©

r^

3

©

«/■>

r»

so

©

so

©

C\

«r>

—

O

SO

w-i

<o

w-i

v>

so

«r>

so

so

w>

o

Wl

«n

so

so

SO

SO

so

so

so

E

^.^.

e

. r.

e

c

e

2

U

e
©

U

(A

en

83

u
O

•o
o

88

>

"p

o

©

©

©

©

l/i

©

©

©

©

©

©

©

©

©

%

©

©

©

©

©

©

©

S

W

**

is

o

©

©

©

©

O

©

©

©

©

©

©

©

©

©

c

©

©

©

©

©

©

©

li

fa.

S3

E

c

oo

**

T

m

(N

fi

o

SO

Os

©

«r>

c*.

©

SO

so

OO

(N

eo

FN

©

t^

OS

o
CJ

>r>

w-i

w-i

«o

«-)

e

c3

so

so

«/">

so

V)

SO

SO

W!

w>

so

so

so

SO

r~

so

«-)

2

■^

5
.S

d

§

•a

5

u

o

c/l

CO

=

II

CI

2

rsj

ro

SO

eo

©

u
W

CM

CO

W)

l-»

OS

_

m

<o

SO

33

r<->

W-l

r-

o

<N

^l-

O0

oo

<

<

<

<

<

03

as

m

03

03

03

m

03

ffi

§

U

U

u

U

U

u

U

1

3
on

e

OO

o

-

t- CO tO *fr CNJ

■ ■ a a

o o o o
SNOIlVAd3SaO dO NOIldOdOUd

V)

cu

</l

_ >*

o 73

Cfl

QC CU

s <-

"5 N?

— ox

n
cu -a

Q
Z

e „

W =>

_ o

ff O

T3 —

< er

s

m 0

ee, II

< e

z

III
III

1 m

° 3

CO

cu ■£
M «

> x

UJ

"O to

IU £0

CO

•— CD

•* a.

</>

-1

o

£2!

3
O

LU

c
1-
z

= 1

E§

Q.

o

o

oo .2

o oc

>■

y

e o

uj X

r

.. o.

O CO

h-

CO
Ul

8.1

CE

CO

Ps

IU

III

LU

<

>

0.

s

> o

UJ u.

o
o

■

cu c

— «->

o

'Sfc

z

C "°

<

—J

(0

t" «

co

05 o

<
B

e £

a

cu c

i- '55

u

co

=>

OS «

T3

c

O

u

cu

is

— UJ

o c

uj o

C5 -O

c c
e

a.
o.

<

O u.

© r3

4)

a

Im

3

U ro

O

>-,

- +

00

'5
CO

IN

1—
D

*»

00

cu

5

3
OX

il

o
u

i

T

cu

"©

LU
CL

LU

>

O

o

en

O

E
»-
Z

o
u

s

CO
UJ

u

<

a

5

CO <D ** 04

■ ■ ■ •

o o o o

SNOiiVAdasao do NOiiaodOdd

Jo

u x

1 S

u

■_ . .
u eu

«

m °:
<1

W CO

.■5 c
— .2

»- s

« C

— <U
O CO

E s

V'J 1)

O

■ c

S..2
U .5

CU CD

■C u

** C

*S "

S£

.2 =5

S c

= «

q. ao

cu '175

OS co

.2 o

© c

o #

<fc *

I

re

3

a
u

Hi

i_ 0s-

u S

u _-
.£ £

c m

o .

C \i>

S s

u «

* s

r «

w Q.

S I-

C/)
UJ

Q.

>

oc

LU

>

O

o

ex

■
=

o) e

K 3
t o

w §

_J

c O
z »

° s,

uj «j
t >

O

c
o

s

o
o.
o

H U

8 •

Q. —

2 —

C

-

C
-~

c u

op

as v

O Q.

c <

-a

3
O

C0h-(Din^-C0CN4T-

oooooooo
SNOIIVAUBSSO JO NOIlHOdOHd

B (U

C o c_

e- en 2

CD

---
re

a
u

i

M

la

tv

a*.

09

-J

U

U

u

u

00

ffl

CO

U

u

U

u

o

CQ

CQ

CQ

U

CQ

ex
a

H

f-

H

H

H

H

H

H

H

H

H

(-

H

H

H

H

r-

V9

la

ci

CO

O

O

09

2

s

s

s

s

p

^

^

s

U

S

co

M

to

CO

^

CO

CO

?£

*5

OQ

OQ

*5

*5

CJ

CO

CQ

CO

U

^J

0)

la

D

U

o

0

u

—

CO

t-

H

co

CO

H

U

co

t-

^

e
to

■5
o

H

H

H

H

H

CO

CO

3

H

l-

D

CO

la

CI

to

to

an

OS

jj3n

9

O a.

V

-o >.

c «

00 J

v e

.E »

C- —

— CI

O DC

c ci

V >

l^

■o be

Table 4-2

ithwest Montana Lo

Feeding and Nestin

c

5

£

Q
©

■Si

a

.2
C

■a

Si

•S-
(J
u

C
1

Q
tj

■a

■Si

§«

o

.<n

5
u

la

•Si
5«
o

a

©

"J

1

1

c

*0

c
■*

1
I

c
%i

1

1

5
o

?°

■n

C
<u

o

«n
|

s

a
5

,<n

-^

1

C

,0

I

c

.jn

1
c
a
u

s

£
c

41

C

-Si

2

*a.

o

I

J

5

1

a?

i

C

1

s:

o

■8

<3
<n

CD

Q

■8

c
-Si

13
|

a

?°
%i

1<3

s

u

B

8

o
"3

o±

BQ

T

^c

"^

Q

Q

T

«c

o.

U

U

a.

G

^

-V

CO

K

a;

CO

co a

EA

. -C

CI

"— -a^

O

"5

. -o

CI

.i B

c

— OB

CO

la

tu

«a

U

S3

la

09

X

U

M

■C

La

BQ

*

1

o

•

•
w

•

*

•

*

■s

to
-C

j*

S3

+

u

o

It

1

is

-X

o
u

Q.

*

a.

u

u
o

u

o

u

1

•
U

u

u

to

La

o

u

■o

to
_«:
o

■s

3
C

c

T3

+

*

■a

-b

X
o

«o

o

S

V)

•

<n

•s

"O

a.

o

*

o

.S

2
c

3
O

■ .

T3

*

CJ

_3

•a

'5
■

u

a:

1

J3

if

1/5

J=

V3

~aa
O
Q.
O
O

G

O

E

o

$

o

2

u

CO

3
O

u.

eo
o

3

03

3
O

u-

oo

CJ

u

5

O.
CO

to
en

E

o

o
2

o
o

o

a

T3

o
o

-o
o

'in
i

U

_>

o

o

E
1)

In

o

•

a*>

0>

6

cs
en

"a.
_CJ

e

CO

—

3
C
en

I*
So

D

O

fi

Xi

i

•a
o
Qi

e

V-

o
tn

3
O

X

O
u.

V

>^

x>

3

X

c
'S

—

c

3

o

es

CQ

aa

"ob

CO

a
o

I

V3

L.

«v

>>

S3
•J

OO

2

2
oo

U

u

U

00

00

oo

U

O

U

U

_B

D

00

U

H

H

H

D

3

D

H

H

H

H

H

■ft

V

Z

an

L.

oo

2

93

2

00

P

s

2

->

s

m

oo

^

^

oo

U

oo

oo

oo

oo

00

u

00

U

S

B

U

2

00

D

H

O

ID

D

D

U

H

D

H

£

■5

f-

00

H

H

H

H

la

V

s

Li.

s

M

3

O

Q

e

*9

E
V

1

(ft

•c

*>

x>

B

93

3

93

-J

(ft

u

e

"0

S

_©

i

a.

a

CO

1

JU

8

~S

M

43

Q.

— m

B -3

>

11
00

Table 4-2 (co
of Southwest Montana Lo

0

z

"O

e

99

ex
s

■5
9
0

u.

I_

"3

j=

(ft

fly

U

4v

"J

sc

a

(ft

■Si

1
t

So
c

-s:
O

a

s

1

fc.

5

e

©

o
.2

J

a

03

»■■>

Q

s:
a

a
o

>3

.u

a

•fc

c

Q

Q

5

C

1

Q

60

c

<3

.5

C

<u

a.

.2

(u
N

§■

o
a

-Si

u

c

a
s:

r5

.5

"5

t

tu
0

C

O

-2i
0

c

a

S

5

bq

a
0

1

s
c

5.
.(5

1

c5

k,

1

It

1

c

c

a

Q

c3

OS

H S n

0 — '

"5 ^_>.

OJ ^ 3

111

BjJ

b b° g>

tft 3 3
■O T3 .§

B | |

w

e

C.

ON

■a

99

1/3

11 S i

-0"

la

u

60

(ft

h
93

C

2 p a

00 ±1 0

. > E Q.

u

O

«1

«>T

O C 3

— 0 H

TJ

O

O —

u

Ov

•°.s.s

09

*

»

"3

£ I |

II « 8

CD J J

H O 0

*

L_

T3
U

c

o

O

E
O

—

to
"5

en

*
.c
(A

5

■S

*
E

x>
o

U.
5

u

C
o

E

*
c

X

-
■o

U

*
O

OO

•

o

s

Q.
(ft

-a
o

•
O
0

E
3

•
■iS

•

*

■a

X)
(ft

u.
O
00

u
0

CD

£■ y 0

e S fi

93 (ft uT

(ft

C
1)
(A

e
o

(ft

"c
o

(ft

B

'5

E
E

I

_o

I

E

u

(ft

u

o.

(A

60

c
E.
a.

o

—

■•aj

O

O
1

u

X.
(ft

0

—
60

O

c

-E
(ft
(ft

0

k.
u

■a

U

•

a

—
(ft

0
B

0

Ua

60
00

.5

E
V

>

(ft

0
0

E
0
u-

0

O

(" (ft

O O 0

£ « a
11 'g ^

1

f-

oo

<

u

>■

U

^

D

CL

ai

'dl

LU

Di

H * +

(0
CQ

60

93

u

I

s 1

CO

U

V

co

co

CO

CO

CO

2

CO

2

CO

£ ■

D

H

H

D

u

CO

H

H

co

co

U J

H

D

H

r-

M

«ri

tn

U

u

e

St,

■a

m

u

s

ai

e

o

a

Of w

e u

s

CO

s

co

co

CO

co

co

CO

1

ine an
Layei

CO

?5

5

CO

|

5

CO

1

|

|

|

JO

3

M

D

(-

H

CO

t-

co

co

co

co

"e3

a. a

■£

u •-

CO

— ■— •

•D

e os

t

s

le 4-3
tana Lod;
Cover Ve

ii

CO

H SJ «**

£ —

•5 s -m

1) J5 3

-o ^ •->

HS §

S3 .S

3

ffl

sll

South'
ir Fee

3
C

5

o

3

5

3

to

3

C

s:

Us

t_ u

<3

£

Q

O

i.

3

o "c «S

ristic ol
and th

c

5

<o

e

■8

C

<u

!

.2

c

Q

t

I

w 3 5

^ "° -o

w

<«3

5.

5

•5

3

■Si

1

•Si

.<2

.<o

.£

|

3 " '(O
<N «" 3 "5

s ■ * 1

03

u

■

ST

£

2

£

^

1

1

1

3

U

u

c

M

CO

« .. C Q.

8

3 i) o a

E

E

€ -0 -S .s

S3

1 ii b b

3 OQ £ £

g H o o

S K § u

*

11 "8 8

*

(J

c

e
o

u | S-S-

o
O

t

"5

O"

.5
o.

o

CI

B

o

.S

u

c
'3

TO
O

JD

o

8 8 | a

o
c

vt

u

3
U

o

CO

.5

c

3
O

2

+
•

8 ii •« ^

U CJ "" —

1/3

o«

a.

CQ

a.

UJ

OQ

-j

UJ

K H * +

eg
CQ

a
u
a:

4-T7

The vegetation and phytotoxicity investigation for the Smelter Hill RI/FS Phase I was
designed to evaluate the various vegetation types that grow within the physically undisturbed
areas (no regrading, restructuring, or reclamation) of the Smelter Hill OU and to evaluate
vegetation growing on the reclaimed areas (including the Anaconda and Opportunity Ponds
berms and parts of the HPS area) (PTI, 1990a). Unreclaimed areas were those portions of the
study area where the original (or remaining) soils have been injured by airborne
contamination, but the substrate has not been graded or altered during construction or
reclamation activities (see Chapter 3.0). Ninety-three of the 198 sites samples for the RI/FS
were unreclaimed soil sites on Smelter Hill (PTI, 1990b).

Smelter Hill was mapped by vegetation type, and types were delineated based on aerial
photographs and field verification. The "natural" plant types identified and sampled were
Great Basin wild rye (Elymus cinereus) grasslands, aspen (Populus tremuloides)
woodlands/scrub forest, wet meadow, rocky barrens, horsebrush (Jetradymia canescens)
shrubland, and willow (PTI, 1991). Sample sites consisted of 10 x 10 meter plots (PTI,
1990a). Percent cover, density, and production were measured, and vegetation biomass
samples were collected for analysis of acid-extractable metals (PTI, 1990a,b).

All mean values for vegetative percent cover by vegetation type were less than 35%, except
for wet meadow vegetation (9% of the total area sampled, located near Mill Creek), which
had a mean cover of 65% (PTI, 1991). Nearly 50% of the study area was classified as Great
Basin wild rye grasslands. Great Basin wild rye is an invasive weedy species, and
monocultures are characteristic of disturbed sites (J. Joy, USDA/Forest Service, pers. comm.).
Of the 58 Great Basin wild rye sites sampled, vegetative cover averaged only 27% (range 5-
46%) (PTI, 1991). Bare soil cover ranged from 3 to 71% and averaged 26%, and litter and
bare rock cover ranged from 23 to 66% and averaged 47% (PTI, 1991). Much of the cover
type designated as litter and bare rock is probably the cobbled, eroded surface now
characteristic of much of Smelter and Weather Hills and vicinity.

The only vegetation type with trees was the aspen woodland type (PTI, 1990b). Aspen
woodlands and scrub forests were less than 2% of the area sampled and were confined to
moist drainages or filled-in beaver ponds (PTI, 1991). The clonal nature of aspen and its role
as an early successional colonist contribute to its localized presence on Smelter Hill.

Vegetative productivity was negatively correlated with some or all metals in all area soils
sampled, except the horsebrush area (PTI, 1991). Horsebrush consistently had the highest
maximum and average tissue concentrations of all analytes, indicating its tolerance of metal
contamination (PTI, 1991). Total metals concentrations in 0-2 inch unreclaimed soils were
significantly correlated with metals concentrations in plant tissue (PTI, 1991) (see
Chapter 2.0). In Great Basin wild rye communities, percent vegetative cover was positively
correlated with plant tissue concentrations and plant-available metals, which is also indicative
of a species with high tolerance to metals contamination. In wet meadow sites, percent cover
was negatively correlated with all plant-available metals concentrations (PTI, 1991).

RCG/Hagler Bailly

4-18

4.4.2 NRDA Investigations

Field studies of vegetation community structure and composition and the wildlife habitat
provided by those communities were performed in 1992 and 1994 (see Appendices C and D).
Reductions in habitat quality and quantity were assessed within areas delineated as grossly
injured based on aerial photographs and preliminary field observations. Grossly injured areas
were identified and delineated using the following criteria:

► Complete or virtual elimination of the indigenous major plant associations

► Little or no regeneration of the indigenous major plant associations

► Extensive top-soil exposure and erosion associated with vegetation loss.
The grossly injured areas (Figure 4-1) are:

► The eastern portion of Stucky Ridge and the hills north of Lost Creek
(3.8 square miles)

►• Smelter Hill, including slopes south and west of the main smelter stack

(7.3 square miles)

*• The Mount Haggin uplands south and east of Mill Creek (6.7 square miles).

At five sites1 in the Stucky Ridge area, nine sites in the Smelter FEU area, and seven sites in
the Mount Haggin area, and at 21 paired control sites in German Gulch, the following
measurements were made in 1992:

► Dominant vegetative cover type

► Number and type of habitat layers present

► Approximate height of each habitat layer in the dominant cover type

► Percent canopy closure of each species in the tree canopy, shrub midstory, and
understory layers

Each site comprised two intersecting 500m transects oriented north-south and east-west.
Measurements were made at 10 points at 100m intervals along the transects. Vegetation measurements at
each of the 10 points were made using 10m line intercept and visual observation methods.

RCG/Hagler Bailly

4-19

»• Percent canopy closure and percent of ground covered by downfall (within

evergreen forest cover type only)

* The presence or absence of evidence of previous afforestation (tree stumps,

downed timber, standing dead timber)

► Density of elk pellet groups.

In 1994, the following measurements were made at an additional 13 sites in the Stucky Ridge
area, 22 sites in the Smelter Hill area, and 22 sites in the Mount Haggin area:2

► Dominant vegetative cover type

► Percent canopy closure of plants, vegetative litter, bare ground, rock, and
downed timber in the tree canopy, shrub midstory, and understory layers

*■ The presence or absence of evidence of previous afforestation (tree stumps,

downed timber, standing dead timber).

As in 1992, data collected at each of these sites represent a compilation of 10 individual
sampling points per site (see Appendix D).

4.4.2.1 Vegetation Characteristics

During 1992 and 1994, 78 sites were sampled in the Stucky Ridge, Smelter Hill, and Mount
Haggin impact areas, and 21 sites were sampled in the German Gulch control area. Each site
comprised 10 subsites spaced at 100m intervals along intersecting transects.

1992 Data

Cover Types. Analysis of cover type data revealed that the Stucky Ridge area is
predominantly bare ground with some grassland and deciduous shrub; the Smelter Hill area is
predominantly bare ground or grassland, with some deciduous and evergreen shrub; and the
Mount Haggin area is predominantly bare ground and grassland, with some deciduous and
evergreen shrub (see Figures 4-2, 4-3, and 4-4). In contrast, bare ground was rarely observed
in the control areas, which were predominantly evergreen forest or grassland. Representative
views of control and impact upland sample sites showing these gross differences in vegetative
communities are shown in Figures 4-5a through 4-5g.

No control sites were sampled in 1994.

RCG/Hagler Bailly

4-20

Cover types were compared statistically between impact and paired control sites. The three
impact areas had a significantly greater proportion of bare ground and a significantly smaller
proportion of evergreen forest than control areas (Table 4-4).

Grasslands. Further analysis indicated that impact sites dominated by the grassland cover
type were significantly sparser (i.e., greater mean percent cover of bare ground) than control
area grasslands. Moreover, the mean percent cover of native grassland species in control sites
was significantly greater than the mean percent cover of native grassland species in the
impact areas (which were dominated to a greater extent by invasive weed species)
(Table 4-5).

Habitat Layers. Analysis of habitat layer data revealed that the impact areas lack two of the
five vegetation layers (tree canopy and tree bole) characteristic of the control sites (Figures
4-6, 4-7, and 4-8). Furthermore, the control areas support a significantly greater development
of the remaining three layers (shrub midstory, understory, and soil) than the impact areas
(Table 4-6). The number of habitat layers in control areas ranged between two and five; at
most of the control sites, five layers were present. In contrast, the number of layers present
at the impact sites ranged from zero to three (Figures 4-9, 4-10, and 4-11). The mean
number of layers at each impact site was significantly less than the mean number of layers at
the respective control site (Table 4-7).

Thus, injury to upland vegetation in the Stucky Ridge, Smelter Hill, and Mount Haggin
impact areas is evident as widespread elimination of indigenous conifer forests, a decrease in
the cover of native grassland species, and an increase in bare, devegetated ground. The types
of habitat layers lost, primarily tree bole and tree canopy, are indicative of the predominant
loss of forest vegetation.

1994 Data

The data obtained during fieldwork in 1994 are presented in Attachments A and B to
Appendix D.

Cover Types. There were no significant differences in the proportional representations of
cover types between the two years (Table 4-8). As in 1992, the most prevalent cover types in
all three impact areas were bare ground and grassland. Whereas bare ground comprised
approximately 40% of the cover types on Smelter Hill and Mount Haggin, it comprised
almost 68% on Stucky Ridge (Figure 4-12). This greater proportional representation of bare
ground on Stucky Ridge is evident in the 1992 data also (Figure 4-12). Conversely, both the
1992 and 1994 data show that grassland had greater proportional representation on Smelter
Hill and Mount Haggin than on Stucky Ridge (see below). At all sites combined, bare
ground comprised 46% of the cover types and grassland almost 37%. Deciduous shrubland
(mainly aspen stands) had only limited cover on all three impact areas, particularly on Stucky
Ridge where it had a proportional representation of only about 6% (compared with 14% and
19% on Mount Haggin and Smelter Hill, respectively).

RCG/Hagler Bailly

(S

■/■■■- ?■ ?<fv$:u

.» ' >., . • •# i V J? ™ c

_ a e
<u .2 e

5« s a

1 i 't'i i r* 7 J3HHBHHH ^ M ^

I • / / ' ( ' ', ' . a.<I "« a a

.'J ■ ? ' fT v-l i| •

w

3
OH

\. V •' ■';.

■u ©

CUD «

H

' . ' ' '. •

■3
CO

4-22

Figure 4-Sc. Upland Sample Site C8 on Mt. Haggin Showing Devegetation and Evidence
of Previous Afforestation.

Figure 4-5d. Upland Sample Site C9 on Mt. Haggin Showing Devegetation and Evidence
of Previous Afforestation.

RCG/Hagler Bailly

i

< g

s ~

o v

•o to

S.fc

o.^

» 8

| |

3 O

= oil
5 .£
E J

1/1

u
3
SI

to
CD

CO

a

4-24

f

Figure 4-5g. Upland Control Sample Sites B13 and C7. Sites located on forested slope
facing camera.

Table 4-4
Comparison of the Proportional Representation of Cover Types in Impact and Control Areas

Area

Cover Type

p-valuc

All impact areas vs. all controls

Evergreen forest
Grassland
Bare ground

< 0.002***

< 0.23

< 0.0002***

Stucky Ridge vs controls

Evergreen forest
Grassland
Bare ground

< 0.06+
<0.93

< 0 031**

Smelter Hill vs. controls

Evergreen forest
Grassland
Bare ground

< 0.002***

< 0.31

< 0.002***

Mount Haggin vs. controls

Evergreen forest
Grassland
Bare ground

< 0.008***

< 0.04*

< 0.031*

+, *, **, and *** indicate significantly greater proportional representation in the control site for a =
10%; 5%; 3.3%, and 1%, respectively.

RCG/Hagler Bailly

4-25

Table 4-5
Comparison of Grasslands in 21 Upland Impact and Control Areas

% Bare Ground

% Cover of Native Species

Impact
Control

61.7
11.0

10.9
34.5

Table 4-6
Comparison of the Proportional Representation of Habitat Layers in Impact Areas and Control

Areas

Habitat Layer

p-value

All impact areas vs. all controls

Tree canopy
Tree bole
Shrub midstory
Understory
Soil

< 0.0002***

< 0.0002***

< 0.0004***

< 0.0002***

< 0.0002***

Stucky Ridge vs. controls
(Area A)

Tree canopy
Tree bole
Shrub midstory

< 0.03125**

< 0.03125**

< 0.03125**

Smelter Hill vs. controls

(Area B)

Tree canopy
Tree bole
Shrub midstory
Understory
Soil

< 0.00195***

< 0.00195***

< 0.00977***

< 0.00195***

< 0.00195***

Mount Haggin vs. controls
(Area C)

Tree canopy
Tree bole
Shrub midstory

< 0.00781***

< 0.00781***

< 0.03125**

** and *** indicate significantly greater proportional representation in the control area at a = 3.3%
and 1%, respectively.

RCG/Hagler Bailly

CN

©

u

*ri

o U

U *

M"2

■a c

e °r

o <

c

en . .

<U (L>

«- O

U i_

O 3

Ul

U o

u

w 00

<

"O

u.

c _:

cr

s^

m

/— v m

CO
CO

< w

es II

u

?: b

< -

w W

>

■o°.i

cr

CO

n

UJ

•■■ —

CO

cr

CO

OS 5
. c

UJ

_l

jfi v>

Q

o

o u

z

3

CO

cc

rstu

1 repr

UJ

z
o

a.

o

c

m

=>
cr

X

>
CC

o

t-
co

a

>

CO
Ul

CO

►-
u

tat Layer

l proporti<

to

2

UJ

<
a

2

i

Habi

:nce ii

o
o

of the

differe

Ul
UJ

cr

i-

UJ

_i
o

CD

>-
a
O

z

nal Representation

Indicates significant

UJ

UJ

.2 *

IT

<

H *

t-

o

SNOIlVAd3SaO dO NOIlHOdOdd

ft. on

■-

3
01)

CN

C/j

O
<
li-
CC
3
to

a

3
CO

>-
cr
O

K
CO

a

UJ

o

z

3

O

u

B
O

U

M)

.Su

o -a
a. c

CW Q.

o <

— u

a £
Si

CO

0)
UJ

a.

LU

Si

o

cr
t-
z
o
o

s

<

c

s

o

UJ UJ

cr O
t- CD

a
uj O

uj Z

cr <
»- u

SNOI1VAH3S80 dO NOIlUOdObd

11

O L_

2 §■

-J O

s!

I °

a -
— a.

II

= c

ev co

a g

« 00

PS -s;

5 «

U T3

O C
c •
0. •

i

T

a>

k.

oO

a

u

C*

CO

LU

hi
>
o
o

eo

O

a
t-
z
o
o

<
a

s

o o

i- >

= o

O u

M «

B

"S*

O

Q.T3

K C

«J 03

1_

y P

m

1 II

■ B

.^■V.

u «

es c

4- O

lm ' —

■<J 03

w e

c a

Ml U

a >-

** 03

E c

3 O

11

-si-

« 0.

b e

►»

« ?3

— . 0

— B

93 °

J3 fc

SB •-

k2 -O

MH

V c

J= ??

-w O

<- tg

© B

E .2?

O «

— <U

es *i

_: rs

E £

8 S

0

~

0.*

x

8> *

-n

PJ *

C

— -a

cu

es c

E 3
.2 *

<

t» *

la

0

0 «j

3

I- .t.

O

a- en

U«3

SNOIlVAd3SaO dO NOIldOdOdd

3

E

c
o

-

-

o

U

<

es

<

■o

w

3

55

(0

(ft

k

— o

■ >>

_, a

o —

E

£ 2

o • -

U A

< =

Q- 3

I z

to ir> *fr co cnj t-

■ ■ • ■ ■ ■

o o © o © o
SNOIlVAd3SaO dO NOIldOdOdd

4* (J
— X

° c
e w

I a

n O

D. °

■ 5

c • -

.2 M

"E "S

o -

°- =

° a
A. U

CO

-

3

00

o

ex

■a
c
o

D.

<»

cu
i_

Li
O

U

CO

K

09

UJ*

ee

><

CO

cu
<

cc

5-

UJ

E

ff Ui

>

u
cu

U. _l

<

CO

"5

E

C/5

l_

UJ

f-
co

s—

e

CO

<

o

CU

UJ UJ

H

cr

X <

■■H

z

J

h- _J

m

u

JD

<

Id

CO

CO

X

UJ

ce
—

cc

co

c_

2>

Li.

t-

O

£5

o

<
a
5

cu

a

DC
UJ

i

E

=9

Z

eu rj

UJ

£ <

O -J

CD

° c

3

1 &

CO

Z

e - ■

ex

aj u

UJ

<

h» CO lO ** CO C\J t-

■ ■■••■•

o o o o o o o

SNOiivAuasao do NOiiaodOdd

E I

& o

— en

« «j

c . -

.2 M

L. O

o u

°i

I

•V
cu

ex

(0

CO

(9

a
u

</)

DC
LU

c
o
—

o

L.

u

85

s

©

M M

LU u

t «

CO >">

o -^

cr -w

b 08

CD
<

X

O

oc

LU
CD

o
o

Ql

LU O

CO o

< i

I z

i « u

° c
- o

I &

c

h- co ir> *r co cm i-

• ■■■■■ ■

o o o o o o o
SNOLLVAH3S90 dO NOLLHOdOHd

D. O
0J CO

— ""-
03 »

B SS
.2 M

u "3

o u

§■§

u o
C U

3
W)

fa

CQ

00
ea

o
o

4-32

Table 4-7
Comparison of the Mean Number of Habitat Layers Present in Impact and Control Sites

Area

p-value

All impact vs. all controls
Stucky Ridge (Area A) vs. controls
Smelter Hill (Area B) vs. controls
Mount Haggin (Area C) vs. controls

< 0.0002***

< 0.03125**

< 0.00195***

< 0.00781***

** and *** indicate a significantly greater mean number of layers in the control area at a = 3.3% and
1%, respectively.

Table 4-8

Results of t-tests Comparing Cover Types Observed

at Impact Sites in 1992 and 1994

Cover Types

Stucky Ridge
p-Value

Smelter Hill
p-Value

Mount Haggin
p-Value

Combined
p-Value

Evergreen forest

p = 0.3370

•

p = 0.3287

p = 0.1591

Deciduous forest

•

p = 0.8684

•

p = 0.8017

Evergreen
shrubland

*

p = 0.4566

p = 0.7348

p = 0.7319

Deciduous
shrubland

p = 0.9821

p = 0.3134

p - 0.6528

p = 0.6023

Grassland

p = 0.3664

p = 0.3931

p = 0.8611

p = 0.3388

Bare ground

p = 0.3621

p = 0.2319

p = 0.7004

p = 0.3179

* not observed at either 1992 or 1994 impact sites.

RCG/Hagler Bailly

•-0000000000
S3dAl fcFEAOO JO N0U.VlN3S3UcB« TVNOUHOdCWd

°» s 3

SBdAl HdACO dO NOUVlN3S3bcBH TVNOuaCtetOBd

05 O)

c
o

C
O

o.
o

c

©\

•a
e
es

*s

o\

OS

a

H

u

>
o

U

SUA! H3A00 dO NOUVJX3S3UcBtl TVNOUUOdOUd

S3dA_l U3AO0 dO NOLLV1W3S3UJ3H TVUOUbOdOHd

cd

£ P

<u

c -a
o c

•- to

Is
i j

e> a.

B O
o -^
.S2 ra

a.

E 2

o n.

u a

«s

-

3

m

-
as

00

CO

o

U

4-34

Grasslands. As in 1992, the 1994 measurements showed that grasslands were only sparsely
vegetated (Figure 4-13). When all sites are combined, only 36% of the line intercepts were
actually covered by vegetation, the remainder was mainly vegetative litter (27%), and bare
ground (36%). Litter in the impact grassland areas studied did not resemble the deep
vegetative litter that is normal for unimpacted grasslands or in forested areas elsewhere in
southwest Montana, but was typically a shallow cover of dead plant material overlying bare
ground. The extent of bare ground within the grassland community was most marked on
Stucky Ridge (46% bare ground), followed by Smelter Hill (37% bare ground), and Mount
Haggin (30% bare ground). In addition, measurements made in 1994 showed that the impact
area grasslands were dominated by invasive weed species (i.e., early colonists of disturbed or
stressed areas), rather than plant species that are indigenous to grasslands in surrounding
unimpacted areas (Table 4-9). Only 2.8% of grassland sites in the entire impact area were
dominated by indigenous grass species. The remainder were dominated by invasive
nongraminoid weeds (26%), or invasive graminoid species (69%), particularly redtop
(Agrostis stolonifera), or Great Basin wild rye. Most of those invasive species, particularly
spotted knapweed (Centaurea maculosa), thistle (Cirsium spp.), and Great Basin wild rye are
relatively unpalatable to grazing animals, hence, of low forage value to wildlife.

Table 4-9
Proportional Representations (% of sites) of Dominant Species in Grassland Cover Type

Invasive Species

Native
Species*

Spotted
Knapweed

Thistle

Redtop

Great

Basin

Wild Rye

Whitetop

Total
Invasive
Species

Stucky
Ridge

36.6

30.1

21.9

5.2

6.2

100

0

Smelter
Hill

26.8

0

42.1

21.9

0

90.8

5.9

Mount
Haggin

0

0

84.9

14.3

0

99.2

0.8

Combined

18.8

6.35

53.5

15.6

1.3

95.6

2.8

* Native species include mainly bluebunch wheatgrass, Idaho fescue, and rough fescue.

RCG/Hagler Bailly

co ,
o ,

CO .
CO.

< I

C i
O i

111 I

CO

1/1

-
-

o

L.

-

=

es

o o o

U3A031N33U3d

H3A00 lN33H3d

U3AC3J-N3DU3d

U3AO0 1N30a3d

— «

= =

.. *<

.2 5

*- es

> D-

— u

§ =

u S

S3

ca j

o "a

£ !

- —

■S 5

=

2 S

o m

2 eg

5 2

> w
O

— -

c -

S 5

I

ra

CQ

S

on

4-36

Evidence of Previous Afforestation

Evidence of previous afforestation (tree stumps and dead timber) was found at 37 (65%) of
the 57 sites sampled in 1994, indicating that, at least on the higher ground, much of the area
designated as grossly injured once supported extensive conifer forest. No previously
afforested sites were found on Stucky Ridge, however, one was found on the injured area to
the north of Lost Creek. Figure 4-14 displays all sites (1992 and 1994) at which evidence of
previous afforestation was found.

4.4.2.2 Relationships between Observed Vegetation and Fhytotoxicity

As described in Chapter 3.0, the results of the phytotoxicity laboratory tests were used to
classify impact sites by degree of phytotoxicity using toxicity scores. These toxicity scores
were then related to observed vegetation conditions to evaluate whether more toxic sites
supported, on average, less vegetation than less toxic sites. Kendall-tau correlations between
the percent bare ground measured at sites and the toxicity scores at the same sites
were positive3 (correlation of 0.58) and highly significant (p = 0.0001) demonstrating the
relationship between phytotoxicity and loss of vegetation.4

As further evidence of the causal relationship, percent bare ground was correlated with
measured concentrations of hazardous substances. Arsenic (Kendall-tau = 0.59, p < 0.001),
copper (Kendall-tau = 0.57; p < 0.001), cadmium (Kendall-tau = 0.33, p = 0.018), lead
(Kendall-tau = 0.49, p < 0.001), and zinc (Kendall-tau = 0.40, p < 0.01) were all significantly
correlated with percent bare ground.

Similarly, the mean number of habitat layers at each site was correlated with toxicity score
and hazardous substance concentrations. Again, these correlations were statistically
significant: the mean number of habitat layers was negatively correlated5 with toxicity score
(Kendall-tau = -0.56, p < 0.001), as well as with concentrations of arsenic (Kendall-tau =
-0.59, p < 0.001), cadmium (Kendall-tau = -0.45, p < 0.001), copper (Kendall-tau = -0.51,
p < 0.001), lead (Kendall-tau = -0.52, p < 0.001), and zinc (Kendall-tau, -0.43, p < 0.005).

i.e., as toxicity increases, percent bare ground increases.

4 As noted in Chapter 3.0, sensitivity analysis was performed on the toxicity scoring system. The
results of this analysis demonstrated that the significant positive correlations were robust to variations in
the scoring protocol; alternative scoring systems resulted in positive correlations (0.57-0.63) that were all
significant (p < 0.001).

In this case, the negative correlation indicates that as the toxicity score increases, the number of
habitat layers decreases.

RCG/Hagler Bailly

4-37

m m mi
\ m m b .

/■ B ■ *

i ■ ■ m/

V Mount Haggin

Kilometers

iRCC/Hagler Baxlly

Figure 4-14. Sample Sites at which Evidence of Previous Afforestation was Found (1992
and 1994). (Closed symbols are those with evidence of previous afforestation,
while open are those without.)

RCG/Hagler Bailly

4-38

The results of these correlations (bare ground vs. toxicity, bare ground vs. hazardous
substances, number of habitat layers vs. toxicity, number of habitat layers vs. hazardous
substances) present clear evidence relating the results of the laboratory toxicity studies and the
field vegetation measurements.

4.4.3 Habitat Suitability

The capacities of the injured and control areas to provide wildlife habitat were compared
using HEP models.

Marten

Statistical comparisons of marten habitat suitability index (HSI) values calculated for impact
and control area sites revealed that Smelter Hill and Mount Haggin provide significantly
poorer habitat for martens than respective control areas (Table 4-10). The results confirm
injury to marten in the Smelter Hill and Mount Haggin impact areas by demonstrating a
significant reduction relative to baseline habitat suitability.

Table 4-10
Comparison of Marten Habitat Suitability in Impact and Control Areas

Area

p-value

All impact areas vs. all controls
Stucky Ridge (Area A) vs. controls
Smelter Hill (Area B) vs. controls
Mount Haggin (Area C) vs. controls

< 0.0002*
<0.06

< 0.004*

< 0.008*

* Indicates significantly greater marten habitat suitability in the control area at a = 1%.

Optimal marten habitat is provided by coniferous forests composed predominantly of mature
fir, spruce, Douglas fir, or lodgepole pine, with z 50% canopy closure and z. 25-50% ground
cover of deadfall (Allen, 1984). The loss of conifer forests in the Smelter Hill and Mount
Haggin areas represents a fundamental reduction in habitat suitability for martens. Stucky
Ridge is at a lower elevation and was most likely vegetated primarily by grasslands before
injury (approximately 30% of Stucky Ridge would have been forested compared to more than
70% of Smelter Hill and Mount Haggin; see Appendix C). Thus, the results of the marten
HEP analysis are consistent with what would be expected based on the probable pre-impact
patterns of forestation.

RCG/Hagler Bailly

4-39

Elk

Statistical comparisons of elk HSI values calculated for impact and control area sites revealed
that Stucky Ridge, Smelter Hill, and Mount Haggin, and all impact areas combined, provide
significantly poorer habitat for elk than respective control areas (Table 4-11). Significantly
lower densities of elk pellet groups at impact sites, relative to controls, confirm the lower use
of the injured areas by elk (Appendix C).

Table 4-11
Comparison of Elk Habitat Suitability in Impact and Control Areas

Elk Habitat Effectiveness

Area

Cover

Forage

Overall

All impact vs. all control sites
Stucky Ridge (Area A) vs. controls
Smelter Hill (Area B) vs. controls
Mount Haggin (Area C) vs. controls

< 0.0002***

< 0.03**

< 0.002***

< 0.0008***

< 0.0002***

< 0.03**

< 0.04*

< 0.03**

< 0.0002***

< 0.03**

< 0.002***

< 0.008***

*, **, and *** indicate significantly greater suitability in control areas at a = 5%, 3.3%, and 1%,
respectively.

Optimal elk habitat is provided in part by coniferous forests of z. 50% canopy closure
interspersed with grasslands consisting of preferred forage species such as bunchgrasses and
bitterbrush. Deforestation and the modification of the grassland community composition have
resulted in the virtual elimination of cover and a reduction in forage for elk over all impact
areas sampled.

Habitat Layers

The layers of habitat model showed that the loss of forest, shrub, herbaceous, and topsoil
cover has resulted in reduced structural complexity for all impact sites. Concomitantly,
habitat availability has been reduced significantly relative to baseline. Habitats with five
habitat layers (tree canopy, tree bole, shrub midstory, understory, and terrestrial subsurface)
define the baseline; the impact areas have been reduced to habitats dominated by bare ground
(lacking even a terrestrial subsurface) or grassland. Table 4-12 presents the results of tests
comparing the habitat complexity of impact and control sites. Stucky Ridge, Smelter Hill,
Mount Haggin, and all impact sites combined exhibited significantly reduced vertical habitat
complexity relative to control areas.

RCG/Hagler Bailly

4-40

Table 4-12
Comparison of Habitat Complexity in Impact and Control Areas

Area

p-value

All impact areas and controls
Stucky Ridge (Area A) vs. controls
Smelter Hill (Area B) vs. controls
Mount Haggin (Area C) vs. controls

< 0.0002***

< 0.03*

< 0.002***

< 0.008***

* and *** indicate significantly greater habitat complexity in the control area at a = 3.3% and 1%
respectively.

Populations of organisms dependent on tree canopy and tree bole are likely to have suffered
the greatest loss of viability because of the loss of upland conifer forests. Species that are
likely to have been lost from the injured area because of this habitat destruction are listed in
Table 4-13. Other organisms less dependent on tree canopy and bole are likely to have
suffered reduced population viability due to the reduced extent of shrub, herbaceous, and soil
layers in the impact areas.

Twenty-seven of the 32 bird species (84%) characteristic of southwest Montana lodgepole
pine and Douglas fir forests are dependent on tree canopy or tree bole layers for nesting or
feeding (Table 4-2). These 27 (including birds of prey, woodpeckers, and songbirds) are
likely to have been lost from the impacted areas as a result of the loss of these layers. The
remaining five species have suffered reduced population viability.

Forty percent of the mammal species listed characteristic of southwest Montana lodgepole
pine and Douglas fir forests are strictly arboreal or dependent on dense conifer cover. These
species (including squirrels, porcupine, and marten) are likely to have been lost from the
impact areas as a result of tree canopy loss (Table 4-3). The remaining six species (including
black bear and elk) are likely to have suffered reduced population viability.

Thus, the elimination or modification of upland vegetation communities on Stucky Ridge,
Smelter Hill and Mount Haggin has resulted in a severe reduction in the quantity and quality
of wildlife habitat. Injured areas that could have supported plant communities, providing
important habitat for a diversity of wildlife species, are largely devegetated and extremely
limited in their ability to support wildlife populations.

RCG/Hagler Bailly

4-41

1

Table 4-13

Bird and Mammal Populations

that are Likely to have been Losi

. or Suffered Reduced Viability

Due to Removal of Tree Canopy, Bole, Shrub Midstory, Understory, and Terrestrial Subsurface

Habitat Layers in Upland Impact Areas

Reduced Viability

Species Lost

Swainsons thrush

Red-tailed hawk

Red-breasted nuthatch

American robin

Sharp-shinned hawk

House wren

Chipping sparrow

Cooper's hawk

Ruby-crowned kinglet

White-crowned sparrow

Northern goshawk

Mountain bluebird

Dark-eyed junco

Blue grouse

Townsend's solitaire

Black bear

Spruce grouse

Cedar waxwing

Ermine

Northern saw-whet owl

Yellow-rumped warbler

Snowshoe hare

Downy woodpecker

Western tanager

Mountain Uon

Hairy woodpecker

Pine grosbeak

Bobcat

Olive-sided flycatcher

Red crossbill

Elk

Western wood-pewee

Pine siskin

Grey jay

Evening grosbeak

Steller's jay

Red squirrel

Clark's nutcracker

Porcupine

Mountain chickadee

Pine marten
Lynx

4.4.4 Causality Evaluation: Vegetation Losses in the Injured Area

The comparisons between control and injured impact areas indicate significant and substantial
differences in vegetation community structure and composition. The coniferous forests
formerly present in the impact area largely have been replaced by bare ground or sparsely
vegetated grasslands. Grasslands in the impact area are vegetatively impoverished and
dominated by weed species relative to control area grasslands. Processes other than
contamination of soils by hazardous substances can contribute to such effects on vegetation.
Potential causes of vegetation loss near Anaconda include fire, logging, and grazing.
Evidence indicates that the German Gulch control area has been subject to fires, logging, and
grazing pressure; therefore, control sites account for effects of these stressors on vegetation.
In addition, while the smelter was in operation, the surrounding landscape was exposed to
deposition of sulfur oxides and sulfuric acid, as well as metals and arsenic. This section
further evaluates the likelihood, based on characteristic vegetation responses to stress, that
factors other than hazardous substances caused the observed injuries in the impact area.

RCG/Haglcr Bailly

4-42

In general, plant species are found where their ecological requirements, growth patterns, and
regeneration cycles are compatible with the frequency and regularity of disturbances (Oliver
and Larson, 1990). For example, it is possible to predict shifts in species composition and
community structure after a fire (e.g., Arno et al., 1985), after clear cutting or selective
logging (Arno et al., 1985), and in response to intensive grazing (e.g., "increasers" and
"decreasers") (U.S. SCS, 1986). Each of these stressors results in a vegetation community
"fingerprint," i.e., a predictable floristic composition and structure. The "fingerprints" of each
of fire, logging, grazing, and even acid deposition, are broadly recognizable, and are distinctly
inconsistent with the conditions observed currently in the impact area. None of these
disturbances precludes regeneration and succession of indigenous plant species. In contrast,
the "fingerprint" observed in the impact area is consistent with effects of metal phytotoxicity
observed in the laboratory and at other mining/smelting sites, and reported in the literature.
In the following paragraphs, we discuss the ecological consequences of fire, logging, grazing,
and acid deposition with respect to current conditions observed in the impact and control
areas, and the consistency of these expected patterns with empirical data demonstrating injury.

4.4.4.1 Fire

Fire has historically been the primary initiator of succession in northern Rocky Mountain
forests and is a regular feature of southwestern Montana forests (Arno and Fischer, 1989).
Natural recurrence intervals of fires of varying intensity in Northern Rocky Mountain pine
and Douglas fir forests before European settlement ranged from 6 to 120 years (Lemon, 1937;
Arno, 1980; Huston, 1973; Davis, 1980; Laven et al, 1980; Martin, 1982; Pyne, 1984; as cited
in Oliver and Larson, 1990), and from 100 to 300 years in sub-alpine forests (Romme, 1980;
Hawkes, 1980; Pyne, 1984; as cited in Oliver and Larson, 1990).

Post-fire succession in Northern Rocky Mountain forests follows a predictable, sequential
development to pre-fire conditions (Arno and Fischer, 1989; Fischer and Bradley, 1987). The
composition of serai stages and time required to return to pre-fire conditions may vary, but
successional development of community structure and composition to pre-fire conditions is
relatively consistent (Lyon and Stickney, 1976). In the absence of fire or other overstory-
removing disturbances, shade tolerant conifers will dominate and deciduous forest (aspens)
will either revert to shrub and grassland or be replaced by conifers (Brown and DeByle, 1989;
Gruell et al., 1986).

Indigenous vegetation communities in the Anaconda uplands (approximately 5,000 to
7,000 feet) comprise a mosaic of cool, dry, Douglas fir habitat types (Pfister et al., 1977) and
grassland communities on southwest slopes. Douglas fir habitat types are associated with
well-drained mountain slopes and valleys, and extend from lower timberline up to about 7,500
feet on warm aspects in southwestern Montana (Pfister et al., 1977). Undergrowth is variable,
but in the Anaconda area, may include huckleberry (Vaccinium spp), snowbush ceanothus
(Ceanothus velutinus), pinegrass {Calamagrostis rubescens), beargrass (Xerophyllum tenax),

RCG/Hagler Bailly

4-43

bluebunch wheatgrass (Agropyron spicatum), Idaho fescue (Festuca idahoensis), ninebark
(Physocarpus malvaceus), and kinnikinnick (Arctostaphylos uva-ursi), among others (Pfister et
al., 1977). Historically, fire maintained the Douglas fir habitat types; estimates of the mean
presettlement fire interval range from 35 to 40 years (Arno and Gruell, 1983).

Figure 4-15 depicts generalized forest succession in cool, dry Douglas fir habitat types
(Fischer and Clayton, 1983). The serai stages described in this figure are potential vegetation
types and successional trajectories expected in the Anaconda area, and observed in the
German Gulch area. Frequent fire in cool-dry Douglas fir habitat types can maintain
grassland communities. Early successional communities are derived primarily from the pre-
disturbance community, and typically establish in the first growing year following the
disturbance (Lyon and Stickney, 1976). In the first few years following a stand replacing fire
in the Anaconda/German Gulch area, grassland/forb/shrublands dominated by huckleberry-
pinegrass, beargrass-pinegrass, or snowbush ceanothus are expected colonizers (Arno et al.,
1985). With favorable seedbed conditions, an even-aged stand of lodgepole and Douglas fir
usually develops. Any fire in either the seedling or sapling stage reverts the site to grassland.
Cool fires during the pole sized stage thin the susceptible stems; a severe fire again reverts
the site to grassland. A cool fire in a mature stand will thin the undergrowth and create a
park-like forest. At successional climax, Douglas fir will dominate the overstory.

Early successional species that are shade intolerant and adapted either to survive fire or to
colonize rapidly following fire, can establish immediately. Species that reproduce
vegetatively, such as aspen, are typically stimulated by the removal of overstory vegetation.
Rhizomatous species, such as huckleberry, snowberry (Symphoricarpos albus) and
thimbleberry (Rubus parviflora), and bulbs and corms of some herbaceous species can survive
all but the hottest fires and develop rapidly post-disturbance. Seeds present before the fire
represent another general adaptation to onsite survival; long-term seed storage occurs in the
seedbed on the forest floor, and in serotinous cones attached to the crown (lodgepole pine).
Northern Rocky Mountain plants with dormant, ground-stored, fire-activated seed comprise
predominately shrub and herbaceous growth forms, such as snowbush ceanothus (Lyon and
Stickney, 1976). In addition, off-site species demonstrate fire-adaptedness in the ability to
introduce disseminules from areas distant from the burn. A review of regional literature
reveals that in general, the majority of plant species in Douglas fir habitat types prior to a
burn will survive or reestablish on the burn (Lyon and Stickney, 1974; Arno et al., 1985;
Fischer and Clayton, 1983), and that in the first decade post-fire, a rapid re-growth of
lodgepole pine understory is expected.

Thus, the expected fingerprint of fire on the Anaconda uplands landscape includes patches of
fire-maintained grassland/forb/shrubland, patches of even-aged sapling to pole-sized lodgepole
pine stands with an understory of beargrass, huckleberry, and/or pinegrass, and patches of
mature forest strongly dominated by Douglas fir in the overstory, with pinegrass and
huckleberry in the understory (Figure 4-16). In fact, this is what was observed in the German

RCG/Hagler Bailly

il

52

as

a •

2 =

N .c

Si °

o "7

oc.o

I?

O *?

\r. O

DC

U C
a _r

■a

.E o

"5b8.
5 00

c1-

8.1

"8°

8-1

-ll

O s

ts.s

b E

6r5

ri

g>-s

Q.3

B .

•?£■*

2 o §

s o i

■«&■

«s=e*

ff

— c
, . .'d\. . . .

§ 5

55

t»

4) CI

Q. 00

£»Os

C «

o c

•a o

a >,

V CO

SPo

r

>^

41

CO *

o
o

6

s J

8j C/3

I

t« •£

E

£ «

s

o

o

rt

<» w

8
7

3

s «

E

s

._ T3

-C C

es o

o
6

c

4>

ft*

u

<. «>

-

=

<n

E

s

o

E
P

M c

3 —

o

T

o -a

o c

8

CO

.►>

C <U

w->

•■■ I—

<N

c<3

1

s

O J3

•s «

So

l|

ES

■a ?

CU C
N —

— -a
es cu

cu C

c a

cu w

O o
■

cw
u

3

in

■lized forest
sion

Ire

s fire

-D u « O 5

!^1

53 O

s -s

• — CO
O <L>

3 'C

£: o

«5 ■£

« E

o g
■S 8

E M

iS

e .2
o —

.E o.

|s

i_

TO

e
c

es

e

<

u

>^ oo

** On

c — :

| <s

Ǥ &

c o E

— <u <J

•- "5 <N

. E on

o ° ~

= 2 15

"2 5 ~

Q. o u
fli CO o

S> « «

e <u -o

;— i- tC

U. re u-

~* CD

"9 -3

« O 8

v o £

C. « ^

* 5 o
U < oo

re
CO

-

3

tZ

o

4-46

Gulch area, and is reflected in the number and type of habitat layers and in the species
composition observed at the control sites. Vegetation communities in the German Gulch area
sampled included mature Douglas fir stands, forest stands representative of the 39-90 years
post-disturbance stage, with 6-10 inch dbh Douglas firs and an understory of pinegrass, even-
aged sapling-sized lodgepole pine stands, and grassland/shrubland on south facing slopes and
in recently burned areas. Evidence of fire in the German Gulch area is apparent in fire scars
both on living trees and in ring scars in stumps.

In areas where the potential climax stage is grassland, such as at lower elevations in the
control and impact areas and on south facing slopes, the fingerprint of fire would be evident
as grassland/shrubland dominated by fire-adapted species. Grasses recover quickly from fire,
and while herbage production may be less than pre-burn level during the first year, it
typically increases threefold or more by the fifth growing season (Gruell et al., 1986).

In contrast, the fingerprint observed in the vegetation communities in the Anaconda upland
comprises abundant bare ground, sparse grasslands, no regeneration of forest communities,
and no evidence of the successional stages expected. The fingerprint of fire is inconsistent
with that observed, and thus we conclude that it is unlikely that fire, at any frequency or
intensity, is the causal agent of the observed injuries in the impact area.

4.4.4.2 Logging

Many turn of the century mining and smelting operations were dependent on a large supply of
timber for fuel, shaft supports, and construction. In the arid west climates, denudation of
hillsides surrounding mining towns occurred rapidly. Historical photos indicate that logging
and land-clearing was typically complete (Gruell, 1983). Likewise, when smelting began in
the Anaconda area in the 1890s, the surrounding hillsides were cleared to construct and fuel
the smelter. Historical photographs confirm that the German Gulch reference area was also
logged extensively in the first half of this century, yet it now supports forest patches of
varying age.

Historical photographs of cool-dry Douglas fir habitat types in southwestern Montana
compared with more recent photographs indicate that formerly denuded hillsides surrounding
many old mining towns have been completely revegetated by now-mature Douglas fir forests
(Gruell, 1983; Platts et al., 1987). Historical photographs indicate that even sites where
placer mining and logging caused tremendous losses of topsoil from large areas of riparian
habitat, Douglas fir, aspen, cottonwood (Populus trichocarpa), willow, chokecherry (Prunus
virginiana), and other understory shrubs have regenerated successfully (Platts et al., 1987).

Disturbance by logging and land clearing can be similar to successional initiation by fire,
except that harvest does not favor regeneration of fire-activated seeds, does not return
minerals previously stored in biomass to the soils, and may be selective of dominant vigorous

RCG/Hagler Bailly

4-47

trees, leaving the more susceptible and stressed stems. Logging/land clearing can increase
erosion rates, but overstory clearing typically releases and stimulates the rapid growth of
shade intolerant early successional species (Gruell, 1983). Evidence of recent logging, within
the past ten years, was observed during the 1992 field sampling in German Gulch. Removal
of the overstory apparently stimulated the growth of shade intolerant understory grasses,
forbs, and shrubs.

In the absence of continual physical or chemical disturbance, we expect that the Anaconda
uplands would have been rapidly revegetated as shown in Figure 4-17a,b (Gruell, 1983). The
expected fingerprint resulting from logging would include evidence of logging as stumps and
slash, favored regeneration of Douglas fir over lodgepole pine, and essentially no effects on
grasslands. In fact, these characteristics are apparent in German Gulch.

In contrast, the fingerprint observed in the vegetation communities in the Anaconda uplands
comprises evidence of stumps and logging camps, but no regeneration of shade intolerant
understory species, and no regeneration of Douglas fir. Moreover, large areas of the injured
Anaconda uplands have never been logged as attested by the fact that there are no stumps and
the trees that previously flourished there are now present as standing- and fallen-dead timber.
The vegetation community characteristics observed are inconsistent with the fingerprint
expected for logging; thus we conclude that it is unlikely that logging is the causal agent of
the observed injuries in the impact area.

4.4.4.3 Grazing

Grasslands in the Anaconda area prior to impact by mining were likely composed of some
mix of bluebunch wheatgrass, needlegrass (Stipa spp), rough fescue (Festuca scabrella),
Idaho fescue, and scattered shrubs and forbs (U.S. SCS, 1986). In response to excessive
grazing by cattle, species expected to decrease in abundance, "decreasers", include bluebunch
wheatgrass, rough fescue, tufted hairgrass (Deschampsia cespitosa), bearded wheatgrass
(Agropyron caninum), and serviceberry {Amelanchier alnifolia). Species expected to increase
in abundance in response to grazing, "increasers", include Idaho fescue, needleandthread
(Stipa comatd), prairie junegrass (Koeleria macrantha), lupine (Lupinus spp), paintbrushes
(Castilleja spp.), snowberry, timber danthonia (Danthonia unispicata), fringed sagewort
(Artemisia frigida), phlox (Phlox spp.), and other forbs. Noxious/weedy species that invade
in response to grazing pressure include spotted knapweed, thistles, dandelion (Taraxacum
officinale), houndstounge (Cyroglossum officinale), annual and biennial forbs, and cheatgrass
(Bromus tectorum) (U.S. SCS, 1986).

RCG/Hagler Bailly

4-48

Figure 4-17a. Alder Gulch (a cool, dry Douglas fir habitat type near Virginia City,

Montana) in 1871. Gold was discovered in the drainage in 1863 and ensuing
placer mining removed much of the native vegetation. In 1871, the dominant
herbaceous layer was bluebunch wheatgrass, a few Rocky Mountain juniper
(Juniperus scopulorum) were evident on the south facing slope at left, and
Douglas fir regeneration was present on the north-facing slope at right.
Stumps in the area and fire-scars in the stumps indicate that logging occurred
and fires were frequent. (Source: Gruell, 1983).

Figure 4-17b. The Same View of Alder Gulch in 1981, 110 Years Later. Douglas fir and
Rocky Mountain juniper regeneration has covered the formerly devegetated
slopes; the north slope is now densely covered by Douglas fir. Aspen,
narrowleaf cottonwood (Populus angustifolia), willow, and chokecherry have
colonized the canyon bottom. Physical devastation by placer mining has not
precluded regeneration of the native riparian and upland communities.
(Source: Gruell, 1983).

RCG/Hagler Bailly

4-49

Vegetation surveys in 1992 and 1994 indicate that the most widespread grassland species in
the injured area are Great Basin wild rye and redtop. Both are known to be successful
invaders of tailings-contaminated sites (MSU/RRU, 1993a; U.S. SCS, 1993; U.S. BLM,
1992), and neither is an expected increaser or invader resulting from grazing pressure.
Moreover, neither Stucky Ridge, Smelter Hill, nor the part of Mount Haggin sampled is
currently grazed by domestic livestock. In contrast, grazing by domestic livestock is common
and ongoing in German Gulch. Dominant grass species at the control sites included
bluebunch wheatgrass, Idaho fescue, tufted hairgrass, and various forbs. In addition, despite
grazing pressure, forests stands representative of all age classes persist, vertical structure of
the vegetation communities present remains diverse (all habitat layers are represented), and
recognizable successional stages are present.

The expected fingerprint, greater abundance of decreaser species and invasive species, with
minimal effects on forest community development, is not evident in the impact area. The
composition of the grasslands in the impact area, the pervasive absence of forests, and the
simple fact that the area is not heavily grazed, are evidence that the cause of the observed
injury is not grazing pressure.

4.4.4.4 Sulfur Dioxide

During the period of smelter operation (1884-1980), combustion of sulfide ores likely released
sulfur dioxide to the atmosphere, which may have resulted in wet or dry acid deposition.
Sulfur oxides released to the atmosphere may be oxidized to sulfate (S042") and deposited as
sulfate or sulfuric acid (H2S04) on plant surfaces and soils. This acid deposition is known to
reduce plant productivity by leaching nutrients from leaves and interfering with
photosynthesis (U.S. EPA, 1984; Fitter and Hay, 1987). Over extended periods, acid
deposition on soils can cause accelerated leaching of soil nutrients and acidification of soils,
with effects dependent on soil type, climate, and to a lesser degree, vegetation cover type
(Reuss and Jonson, 1986; U.S. EPA, 1984). However, the effect of acid deposition on plants
is transient, and on soils, recognizable in the chemistry of the soil. As the following
paragraphs describe, neither effect persists in the impact area currently.

Reports prepared for the Anaconda Company in the 1970s indicate that at that time, sulfur
dioxide (S02) effects on the landscape were minimal (Treshow, 1972). Field investigations
conducted to identify the extent of plant injury attributable to sulfur dioxide determined that
the vegetation in the vicinity of the smelter was limited to a few species, but the existing
plants were free of abnormal markings indicative of S02 damage. Further from the stack,
aspen, lodgepole pine, limber pine (Pinus flexilis), and Douglas fir were reported to be normal
and free from markings attributable to S02 (Treshow, 1972; Treshow, 1974). Some other
investigators, however (e.g., Carlson (1974), Walsh and Bissel (1979), and Bissel (1982)),
reported evidence to the contrary: abnormal markings associated with acid deposition, shifts

RCG/Hagler Bailly

4-50

in plant community composition, and reductions in total vegetative cover downwind of the
smelter.

Regardless of whether plants were injured by acid deposition during the period of smelter
operation, the source has ceased, and acid deposition at the leaf surface is no longer a factor
contributing to the poor regeneration and succession in the impact area.

Since the period of releases of sulfur oxides lasted 90 years, there is the potential that sulfate
anions deposited on soils may have acidified the soils. Acidification of soils is a natural
process, but the deposition of acid anions on soils can potentially accelerate the natural
processes by expediting leaching of base cations (Ca2+, Mg2*, K+, or Na+) from the soil.
Accelerated leaching of base cations from soil has two primary effects: it reduces the
available supply of plant nutrients, and it results in an increase in the concentration of
hydrogen ions in the soil solution (reduces the pH).

In general, acid deposition does not produce acid soils unless it is accompanied by slow
leaching through the soil column. Runoff water will minimally impact soil chemically, due to
its brief contact with soil particles (U.S. EPA 1984). Acid deposition on dry soils, and even
some moist forest soils, may cause no lasting ill effects (U.S. EPA 1984; Stuanes et al.,
1992). However, the literature regarding the effects of acid deposition in the northern Rocky
Mountains is sparse, either because the air is relatively clean compared to other parts of the
world, or the soils and thus streams are not prone to acidification.

The data collected as part of this assessment indicate no significant acidification of soils
either as an increase in the H+ concentration in the soil solution (a decrease in pH), or as a
reduction of exchangeable cations, relative to the control area. No significant differences in
the cation exchange capacity of control and impact soils, the pH of the control and impact
soils, nor the potassium concentration (an exchangeable base cation) of control and impact
soils were observed (Table 4-14). Soils with a cation exchange capacity exceeding 15
meq/100 g are generally considered insensitive to acid deposition (U.S. EPA, 1984). The
mean CEC values for impact and control soils (16.6 meq/lOOg and 16.4 meq/lOOg) are both
within the range of soils considered to be insensitive to acid deposition (U.S. EPA 1984), and
are comparable to CECs of other soils of the world. Although most soils do not exceed 30
meq/100 g, CEC measured in the impact soils ranged from 35.9 to 7 meq/100 g, and in the
control soils, 26.6 to 8.5 meq/100 g. Moreover, it should be noted that, unlike hazardous
substance concentrations, pH was not correlated with either bare ground (Kendall-tau = -0.08,
p = 0.65) or with the number of habitat layers (Kendall-tau = 0.08, p = 0.57).

It is possible that acidification of the soil solution did occur in the past and that the sulfate
anions leached smelter-deposited metals through soils. This process would have made the
hazardous substances in the soils more available for plant uptake, as metal solubilization and
biological toxicity are pH dependent (Kabata-Pendias and Pendias, 1992; U.S. EPA, 1984).

RCG/Hagler Bailly

4-51

Table 4-14

Comparisons of Mean Potassium Concentrations (ppm), pH, and CEC (meq/lOOg)

in Impact and Reference Soils (0 to 5 cm)

Variable

Reference Samples
n -8

Impact Samples
n-23

p-value

Potassium

pH

CEC

609.7

5.7

16.4

501.6

5.4

16.0

0.4538
0.2908
0.8888

| * p-values (> 0.05) indicate no significant differences between control and impact soils (2-tailed
tests).

Thus acidity could have actually increased the phytotoxicity of the metals deposited
concurrently (U.S. EPA, 1984). The distribution of metals in the soils currently however,
suggests that metals have not leached from the upper soil horizon to lower horizons
(Appendix A).

The evidence overall suggests that acid deposition had no discernible lasting effects on soil
chemistry in the Anaconda uplands. Despite the facts that soils were not acidified and acid
deposition ceased fourteen years ago, vegetation communities show no signs that are
indicative of a release from a transient stress. The fingerprint of injury caused by acid
deposition is not consistent with the pattern observed, and therefore is unlikely to be the
cause of tne persistent injury in the Anaconda uplands today.

4.4.4.5 Evidence of Hazardous Substance Causality

The soil chemistry data, the vegetation community measurements, the phytotoxicity tests, and
the correlations between phytotoxicity scores and hazardous substance concentrations in soils,
phytotoxicity scores and percent bare ground, phytotoxicity scores and the number of habitat
layers, concentrations of hazardous substances in soils and percent bare ground,
concentrations of hazardous substances and the number of habitat layers, as well as numerous
other studies conducted for the RI/FS process, consistently support the conclusion that it is
elevated concentrations of hazardous substances in the soils of the impact area that has
injured the vegetation communities.

Other vegetation surveys conducted within the Superfund boundaries have also identified
metals concentrations in soils as the primary factor limiting plant growth (MSU/RRU, 1993a,
1993b). Vegetation survey results indicate that species present near the former smelter are
metals tolerant and drought resistant. These include Great Basin wild rye (seeded in

RCG/Hagler Bailly

4-52

revegetation attempts on Smelter Hill, MSU/RRU, 1993a), a metals-tolerant and drought
tolerant grass, spotted knapweed, an invasive metals and drought tolerant noxious weed, and
tufted hairgrass and redtop bentgrass, also known to be metals tolerant though less drought
tolerant (MSU/RRU, 1993a). Shrubs and trees that have persisted in isolated patches on
Smelter Hill and Mount Haggin typically have rooted below the upper layers of contaminated
soil; fine roots are observed only in deeper soil horizons (MSU/RRU, 1993a). The aspen
groves that persist on parts of Smelter Hill and Mount Haggin assessment areas are
reproducing asexually. These colonies likely have survived since pre-mining days. Their
persistence is attributable to the fact that the genet (genetic individual) already had roots
established below the contamination and could continue to rootsprout despite conditions at the
soil surface inhospitable for seedling growth (MSU/RRU, 1993a).

The evidence of hazardous substance causality is supported by the "consistency criteria" of
Hill (1965). The results of the upland soils, vegetation, and wildlife habitat investigations
demonstrated that:

► The impact soils are contaminated with hazardous substances (Chapter 3.0,
Appendix A).

*• The concentrations of hazardous substances measured in impact soils exceed

phytotoxic thresholds (Chapter 3.0, Appendix B).

► The impact soils are phytotoxic in controlled laboratory studies (Chapter 3.0,
Appendix B).

► Degree of phytotoxicity is correlated with concentrations of hazardous
substances (Chapter 3.0, Appendix B).

► Degree of phytotoxicity scores is correlated with percent bare ground and
with the mean number of habitat layers.

► Concentrations of hazardous substances are correlated with percent bare
ground and with the mean number of habitat layers.

* No significant differences between impact and control soil nutrient

concentrations, pH, CEC or soil texture are observed (Chapter 3.0,
Appendix A).

► Impact area vegetation communities are characterized by an increased
percentage of bare ground, absence of forest cover, and reduced indigenous
species diversity.

RCG/Hagler Bailly

4-53

► Impact area wildlife habitat is characterized by a reduced number of habitat

layers, particularly tree canopy, tree bole, and terrestrial subsurface, and thus
reduced habitat availability and suitability.

All of this evidence is consistent with the conclusion that hazardous substances caused the
injury. Metals and arsenic contamination is toxic to vegetation (MSU/RRU, 1993a; Kabata-
Pendias and Pendias, 1992; Fitter and Hay, 1987), soil flora (Kabata-Pendias and Pendias,
1992), soil fauna (Kabata-Pendias and Pendias, 1992), is virtually permanent (Kabata-Pendias
and Pendias, 1992), and thus is consistent with the conditions observed in the impact area.
Metals and arsenic contamination explains the loss of vegetation, the persistence of the injury,
the differences observed between the impact and control sites, the laboratory toxicity, the
impacts to forest, shrub, and grassland communities, and the uniformity of the impact
throughout the assessment area (Table 4-15).

The effects of fire, grazing, logging, and acid deposition are clearly inconsistent with what is
observed in the impact areas investigated as part of this NRDA (Table 4-15). Fire and
logging can cause initial loss of vegetation, but do not explain impacts across all community
types throughout the assessment area, do not explain differences observed between the control
and impact sites do not explain the impacts in areas that have not been logged, and do not
explain the toxicity of the soils. Grazing can result in loss of vegetation in grasslands and
shrublands and understory components of forest communities, but does not explain absence of
trees, soil toxicity, or the differences observed between impact and control sites. Soil nutrient
depletion can cause long-term loss of vegetation and impacts across all vegetation types, but
was not observed in impact area soils relative to control area soils. Acid deposition can cause
initial loss of vegetation, but does not explain the persistence of the observed injury,
particularly in rapidly regenerating grasslands.

The accumulation of hazardous substances in the soil is the only consistent and plausible
cause of the injury observed in the Anaconda uplands.

4.4.5 Eitent of Injury

The State's delineation of injured areas showed that 17.8 square miles (11,366 acres) of
upland vegetation and wildlife habitat have been grossly injured. Although the degree of
injury varies throughout the injured area, significant reductions in habitat quantity and quality
relative to uncontaminated control sites has been confirmed over the 17.8 square mile area.
This injury extends north from the Washoe Smelter stack to include Stucky Ridge and
hillsides north of Lost Creek, and south to Mount Haggin as far as the Continental Divide.

RCG/Hagler Bailly

4-54

Table 4-15
Evaluation of Causality

Fire

Logging

Grazing

Nutrients

SO,

Hazardous
Substances

Can cause loss of
vegetation?

yes

yes

yes

yes

yes

yes

Long-term impact?

no

no

no

yes

no

yes

Impact site different from
controls?

no

no

no

no

?

yes

Can explain laboratory
toxicity?

no

no

no

no

no

yes

Explains impacts to forest,
shrub, grassland?

no

no

yes

yes

no

yes

Explains impacts to entire
assessment area?

no

no

no

no

no

yes

4.5 ABILITY OF RESOURCE TO RECOVER

Plant communities have the capability to recolonize disturbed land and, eventually, re-
establish wildlife habitat. The extent to which this may occur and the direction of the
recovery (i.e., the resulting composition and structure of the recolonist communities, and the
degree to which they resemble the pre-impact communities) are determined by whether
residual stress persists, and whether the original stress has not irrevocably altered the abiotic
conditions of the site.

Phytotoxicity studies performed by the state during the assessment of injury (see Chapter 3.0)
have shown that the ability of the injured upland areas to revegetate and provide wildlife
habitat is severely constrained by the current high to severe phytotoxicity of soils caused by
elevated concentrations of hazardous substances. Furthermore, the extensive loss of topsoil
(including the associated seed bank and nutrients) limits potential recovery. In areas exposed
to wind erosion, it is likely that the current vegetative conditions will be virtually permanent.
In more sheltered areas where some topsoil persists, limited recovery may take place;
however, the natural recovery process will be extremely slow. Furthermore, the
concentrations of hazardous substances in the soil and their demonstrated phytotoxic effects
render the outcome of this revegetation process uncertain; it is by no means certain that
without restoration, the eventual secondary climax community will resemble that which was
lost as a result of exposure to hazardous substances.

RCG/Hagler Bailly

4-55

4.6 SUMMARY

The data presented in this chapter demonstrate that:

► Upland vegetation and wildlife habitat (and, in turn, wildlife population
viability) have been injured within a 17.8 mi2 area in the vicinity of
Anaconda, including areas of Stucky Ridge, Smelter Hill, and Mount Haggin

+ Vegetation injuries have resulted in statistically significant reductions in

evergreen forest in impact areas, with concomitant significant increases in
bare ground.

► Grassland communities in injured areas have significantly reduced vegetative
cover and increases in invasive weed species that are less palatable to grazing
wildlife.

*■ Injured areas demonstrated significant reductions in habitat layers, including

tree canopy, tree bole, shrub midstory, understory, and soil subsurfaces.

*■ Wildlife habitat quality was reduced in injured areas, including forests

(representative species = marten), forest/grassland (representative species =
elk), and lost habitat layers. These reductions in wildlife habitat reduce the
viability of wildlife populations in injured areas.

► Vegetation reduction and loss (measured as percent bare ground and the
number of habitat layers) were statistically significantly correlated with degree
of phytotoxicity and hazardous substance concentrations.

*■ Exposure to hazardous substances in soils provides the only consistent and

plausible explanation for the injuries observed in the impact area.

4.7 REFERENCES

Allen, AW. 1984. Habitat Suitability Index Models: Marten. U.S. Fish and Wildlife
Service FWS/OBS-82/10.11.

Arno, S. and W.C. Fischer. 1989. Using vegetation classifications to guide fire management.
In Proceedings - Land classifications based on vegetation. Applications for Resource
Management. U.S. Department of Agriculture, Forest Service Intermountain Research Station.
Gen. Tech. Report INT-257. pp. 81-86.

RCG/Hagler Bailly

4-56

Amo, S.F. and G.E. Gruell. 1983. Fire history at the forest grassland ecotones in
southwestern Montana. Journal of Range Management. 36(3):332-336.

Arno, S.F., D.G. Simmerman, and RE. Keane. 1985. Forest succession on four habitat types
in western Montana. U.S. Department of Agriculture, Forest Service. Intermountain Forest
and Range Experiment Station, Ogden UT. Gen. Tech. Report INT-177.

Bergeron, D., C. Jones, D.L. Genter, and D. Sullivan. 1992. P.D. Skaar's Montana Bird
Distribution. Montana Natural Heritage Program Special Publication No. 2.

Bissel, G. 1982. A survey of trace metals pollution in a forest ecosystem southwest of the
Anaconda copper smelter, Montana. M.S. Thesis, University of Montana, Missoula, MT.

Bradley, A.F., N.V. Noste, W.C. Fischer. 1992. Fire Ecology of Forests and Woodlands in
Utah. U.S. Department of Agriculture, Forest Service. Intermountain Research Station,
Ogden UT. Gen. Tech. Report INT-287.

Brown, J.K. and N.V. DeByle. 1989. Effects of prescribed fire on biomass and plant
succession in western aspen. U.S. Department of Agriculture, Forest Service Intermountain
Research Station. Research Paper INT-412.

Carlson, C.E. 1974. Evaluation of sulfur dioxide injury to vegetation on federal lands near
the Anaconda Copper Smelter at Anaconda, Montana. Insect and Disease Rep. 74-15.
Missoula, MT. U.S. Department of Agriculture, Forest Service, Northern Region, State and
Private Forestry.

Chapman, J. A. and G.A. Feldhamer. 1982. Wild Mammals of North America. The Johns
Hopkins Press, Baltimore and London.

Cody, M.L. 1975. Towards a Theory of Continental Species Diversities. In Ecology and
Evolution of Communities (ML. Cody and J.M. Diamond, eds), Belknap, Cambridge, MA.

Cooperrider, A.Y., R.J. Boyd, and R.S. Hanson. 1986. Inventory and Monitoring of Wildlife
Habitat. U.S. Department of the Interior, Bureau of Land Management Service Center,
Denver, CO.

Deer Lodge County Historical Group. 1975. In the Shadow of Mount Haggin. DLCHG
Heritage Project Report.

Fager, C.W. 1992. Unpublished M.Sc. Thesis, Montana State University.

RCG/Hagler Bailly

4-57

Fischer, W.C. and A.F. Bradley. 1987. Fire ecology of western Montana habitat types. U.S.
Department of Agriculture, Forest Service, Intermountain Research Station, Ogden, UT.
General Technical Report INT-223.

Fischer, W.C. and B.D. Clayton. 1983. Fire ecology of Montana forest habitat types east of
the Continental Divide. U.S. Department of Agriculture, Forest Service. Intermountain
Forest and Range Experiment Station, Ogden UT. Gen. Tech. Report INT-141.

Fitter, AH., and RK.M. Hay. 1987. Environmental Physiology of Plants. Academic Press,
NY.

Gruell, G.E. 1983. Fire and vegetative trends in the Northern Rockies: Interpretations from
1871-1982 photographs. U.S. Department of Agriculture, Forest Service. Intermountain
Forest and Range Experiment Station, Ogden UT. Gen. Tech. Report INT- 158.

Gruell, G.E., J.K Brown, C.L. Bushey. 1986. Prescribed fire opportunities in grasslands
invaded by Douglas fir: state of the art guidelines. U.S. Department of Agriculture, Forest
Service Intermountain Research Station. General Technical Report INT- 198.

Harman, W.N. 1972. Benthic Substrates: Their Effect on Freshwater Molluscs. Ecology
53:271-272.

Hawley, V.D., and F.E. Newby. 1957. Marten Home Ranges and Population Fluctuations.
J. Mammals 38:174-184.

Headwaters RC&D Forester. 1992. A&C Hill Survival Study for 1985 through 1991.

Hill, A.B. 1965. The Environment and Disease: Association or Causation? Proceedings of the
Royal Society of Medicine. 58: 295-300.

Johnsgard, P. A. 1992. Birds of the Rocky Mountains. University of Nebraska Press.,
London.

Kabata-Pendias, A. and H. Pendias. 1992. Trace Elements in Soils and Plants. 2nd ed.
CRC Press, Ann Arbor, MI.

Koehler, G.M., and M.G. Hornocker. 1977. Fire Effects on Marten Habitat in the Selway-
Bitterroot Wilderness. J. Wildlife Management 41:500-505.

Lyon, L.J. and P.F. Stickney. 1976. Early vegetal succession following large northern Rocky
Mountain wildfires. In Proceedings, Tall Timbers Fire Ecology Conference No. 14; 1974
October 8-10; Missoula, MT. Tallahassee, FL; Tall Timbers Research Station and the
Intermountain Fire Research Council, pp. 355-375.

RCG/Hagler Bailly

4-58

MacArthur, R.H. and J. MacArthur. 1961. On Bird Species Diversity. Ecology 42:594-598.

MSU/RRU. 1993a. Anaconda revegetation treatability studies. Phase I: Literature review,
reclamation assessments, and demonstration site selection, Anaconda Smelter Superfund Site.
Prepared for ARCO, U.S. EPA, Montana DHES, and Research and Development Institute,
Inc. by Montana State University Reclamation Research Unit. Bozeman, MT.

MSU/RRU. 1993b. Anaconda revegetation treatability studies. Draft Phase II Work Plan:
Greenhouse Investigation, Anaconda Smelter Superfund Site. Prepared for ARCO, U.S.
EPA, Montana DHES, and Research and Development Institute, Inc. by Montana State
University Reclamation Research Unit. Bozeman, MT.

Mueggler, W.F. and W.L. Stewart. 1980. Grassland and Shrubland Habitat Types of
Western Montana. U.S. Forest Service General Technical Report INT-66. Ogden, UT.

NPI. date unknown. Revegetation Planning Assistance for the Anaconda Smelting Facility.
Prepared for Ecology and Environment, Inc. by Native Plants, Inc., Salt Lake City, UT.

Oliver, CD. and B.C. Larson. 1990. Forest Stand Dynamics. Biological Resource
Management Series. McGraw-Hill, New York. 467 pp.

Pfister, R.D., B.L. Kovalchik, S.F. Arno, and R.C. Presby. 1977. Forest Habitat Types of
Montana. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range
Experiment Station, Ogden, UT. General Technical Report INT-34.

Pfister, R.D., B.L. Kovalchik, S.F. Arno, and RC. Presby. 1983. Forest Habitat Types of
Montana. Intermountain Forest and Range Experiment Station. USDA Forest Service, Ogden,
UT.

Pianka, E.R. 1967. On Lizard Species Diversity: North American Flatland Deserts. Ecology
48:333-351.

Platts, W.S., C. Armour, GD. Booth, M. Bryant, J.L. Bufford, P. Culpin, S. Jensen, G.W.
Lienkaemper, G.W. Minshall, S.B. Monsen, RL. Nelson, J.R Sedell, J.S.Tuhy. 1987.
Methods for Evaluating Riparian Habitats with Applications to Management. U.S.
Department of Agriculture, Forest Service, Intermountain Research Station. General
Technical Report INT-221.

PTI. 1990a. Smelter Hill Remedial Investigation/Feasibility Study: Vegetation and
Phytotoxicity Investigation, Sampling and Analysis Plan. Prepared for ARCO by PTI
Environmental Services, Bellevue, WA.

RCG/Hagler Bailly

4-59

PTI. 1990b. Smelter Hill Remedial Investigation/Feasibility Study: Phytotoxicity, Surface
Water, and Groundwater Investigations Data Summary/Validation/Utilization Report. Volume
I. Prepared for ARCO by PTI Environmental Services, Bellevue, WA.

PTI. 1991. Smelter Hill Remedial Investigation/Feasibility Study. Preliminary Site
Characterization Information. Prepared for ARCO by PTI Environmental Services, Bellevue,
WA.

Reuss, JO. and D.W. Johnson. 1986. Acid Deposition and the Acidification of Soils and
Waters. Springer Verlag, New York. 119 pp.

Short, H.L. 1984. Habitat Suitability Index Models: The Arizona Guild and Layers of
Habitat Models. U.S. Fish and Wildlife Service FWS/OBS-82/10.70.

Stuanes, A.O., H. Van Miegroet, D.W. Cole, and G. Abrahamsen. 1992. Recovery from
acidification. In D.W. Johnson and S.E. Linberg, (eds), Atmospheric Deposition and Forest
Nutrient Cycling. Springer- Verlag, New York, pp 467-494.

Tonn, W.M. and J.J. Magnuson. 1982. Patterns in the Species Composition and Richness of
Fish Assemblages in Northern Wisconsin Lakes. Ecology 63:1 149-1 166.

Treshow, M. 1972. Observations of vegetation in the vicinity of the Anaconda Co. Prepared
for the Anaconda Company.

Treshow, M. 1974. Observation of plant species in the Anaconda, Montana area, Aug. 1-2,
1974. Prepared for the Anaconda Company.

U.S. BLM. 1992. Smelterville Flats rehabilitation project progress report. Prepared by the
U.S. Department of the Interior, Bureau of Land Management.

U.S. DA. 1986. Anaconda Erosion Control and Stabilization: Critical Area Treatment RC&D
Measure Plan and Environmental Assessment. Deer Lodge County, Montana. Sponsored by
Anaconda-Deer Lodge City-County Government, Deer Lodge Valley Conservation District,
and Headwaters Resource Conservation and Development Area. Prepared with assistance
from U.S. Department of Agriculture, Soil Conservation Service, Bozeman, MT.

U.S. DOI. 1987. Guidance on Use of Habitat Evaluation Procedures and Suitability Index
Models for CERCLA Application. Type B Technical Information Document. Prepared by
the National Ecology Research Center, U.S. Fish and Wildlife Service, U.S. Department of
the Interior. Fort Collins, CO.

U.S. EPA. 1984. The acidic deposition phenomenon and its effects. Critical assessment review
papers: Volume II Effects Sciences. U.S. Environmental Protection Agency, Office of
Research and Development. EPA-600/8-83-016BF.

RCG/Hagler Bailly

4-60

U.S. FWS. 1981. Standards for the Development of Habitat Suitability Index Models.
Ecological Services Manual 103. Division of Ecological Services, U.S. Fish and Wildlife
Service, U.S. Department of the Interior. Washington, D.C.

U.S. SCS. 1986. Anaconda erosion control and stabilization project: soils report. U.S. Soil
Conservation Service, Anaconda, MT.

U.S. SCS. 1993. BLM-Smelterville: SCS grass variety seeding trial 1990-1993. Presented by
R. Peyton, U.S. Department of Agriculture Soil Conservation Service Area Forester, Moscow,
at West Region Soils Conference, Coeur d'Alene, Idaho.

Walsh, N. and G. Bissel. 1979. Impact of copper smelter emissions on two subalpine
ecosystems near Anaconda, Montana. Rep. 79-7. Missoula, MT. Missoula, MT. U.S.
Department of Agriculture, Forest Service, Northern Region, State and Private Forestry.

Weckwerth, R.P., and V.D. Hawley. 1962. Marten Food Habits and Population Fluctuations
in Montana. J. Wildlife Management 26:55-74.

Wolle, M.S. 1963. Montana Pay Dirt, a Guide to the Mining Camps of the Treasure State.
Ohio University Press.

RCG/Hagler Bailly

5-1

5.0 SOURCES AND PATHWAYS OF HAZARDOUS SUBSTANCES TO
RIPARIAN RESOURCES

Pathway determination establishes the route or media by which hazardous substances have
been transported from their sources to riparian soils and floodplain sediments of Silver Bow
Creek, the upper Clark Fork River, and the Opportunity Ponds area [43 CFR § 1 1.63 (b)]
This chapter identifies sources of hazardous substances, pathways by which hazardous
substances have been and continue to be transported from sources to riparian resources of the
upper Clark Fork Basin, and the areal extent of the exposed pathways.

5.1 SILVER BOW CREEK

5.1.1 Identification of Sources of Hazardous Substances — Silver Bow Creek

Surface water resources have historically served as the primary pathway by which riparian
soils have been exposed to hazardous substances. Numerous sources of hazardous substances
have historically contributed to contamination of Silver Bow Creek. Although direct
discharge of mining and mineral-processing wastes to Silver Bow Creek has ceased, re-
releases of hazardous substances into the creek are ongoing, resulting in continued exposure
of floodplain soils and sediments.

Historical and current sources of hazardous substances to Silver Bow Creek are described in
detail in Lipton et al (1995) These sources are summarized briefly in Table 5-1

5.1.2 Identification of Transport Pathway — Silver Bow Creek

Pathways of transport of hazardous substances from sources to the floodplains of Silver Bow
Creek include direct disposal and surface water pathways. Contaminated sediments
transported in Silver Bow Creek have been deposited on Silver Bow Creek floodplains as a
result of increased sediment loading and channel aggradation1 during seasonal high water,
and as a result of downstream erosion and redeposition of contaminated material

The addition of large quantities of sediment in the wastes discharged to Silver Bow Creek
resulted in aggradation of the river channel. The decreased riverbed volume initially caused
increased flooding frequency (GCM Services, Inc., 1983, as cited in MultiTech, 1987a) and
accelerated river meandering as the increased sediment load clogged the channel. As a result
of clogged channels, a braided stream pattern developed and progressed downstream. Former

Modification of the channel bed in the direction of uniformity of grade by deposition, in this case,
filling in of the channel.

RCG/Hagler Bailly

5-2

Table 5-1
Sources of Hazardous Substances to Silver Bow Creek

Historical Discharge of Raw Mining and Mineral-Processing Wastes Directly into Silver
Bow Creek

Discharges occurred from the inception of mining in the Butte area in approximately 1878,
until 1976, when the Weed Concentrator treatment process was expanded to include mill
wastewater (MultiTech. 1987a).

Tailmgs, waste rock, smelter slag, process water, and mmewater were discharged directly to
Silver Bow Creek (MultiTech. 1987a).

Smelting Waste Deposits

At least six major smelters were built along Silver Bow Creek between 1879 and 1885

between Meaderville and Williamsburg (Freeman, 1900, Meinzer, 1914, Smith, 1952,

Historical Research Associates, 1983; all as cited in MultiTech, 1987a).

Smelters operated continuously until 1920 (except for the Pittsmont Smelter, which operated

until 1930).

Tailings and slag waste products were deposited on the Silver Bow Creek floodplain or

sluiced to tributaries of Silver Bow Creek (Flynn. 1937. as cited in MultiTech. 1987a. CH:M

Hill and Chen-Northern. 1990).

Waste Rock Deposits

Waste rock dumps identified in the Butte area include the Belmont and other inactive mines:
waste rock dumps in the Warren Avenue drainage basin; numerous inactive mines and
associated waste dumps outside the MRJ boundary- fence, eroded mme wastes in the Anaconda
Road-Butte Brewery Basm; heavily eroded waste rock dumps in Buffalo Gulch basin; and
waste rock dumps in Missoula Gulch (Camp Dresser and McK.ee. 1991).

Tailings Deposits

Tailings deposits that have been identified as significant sources of hazardous substances

include the Parrot Smelter Tailmgs. the Butte Reduction Works tailmgs, and the Colorado

Tailmgs.

Historically, tailings were deposited on the Silver Bow Creek floodplain. Photographs (1955)

show extensive exposed tailmgs and slag deposits m the upper Metro Storm Drain area

(CH:M Hill and Chen-Northem, 1990).

An estimated 3.7-7.8 million cubic yards of contaminated tailings and waste material have

been deposited along the Silver Bow Creek floodplain and serve as sources of re-released

hazardous substances (Canonie, 1992).

RCG/Hagler Bailly

5-3

stream channels are evident in aerial photographs as deposits of fine-grained sediment devoid
of vegetation.

Exposed tailings, waste rock piles, contaminated soils, and the large expanses of exposed
streamside tailings along Silver Bow Creek are subject to erosion and entrainment during high
water, snowmelt, and precipitation-induced runoff (CH2M Hill and Chen-Northern, 1990).
Streams at high water stage carry increased suspended sediment loads, and as high waters
recede, the sediment load is deposited on the floodplains. Tremendous amounts of material
containing hazardous substances may be moved during seasonal high water events Currently,
the entire Silver Bow Creek floodplain is contaminated with fluvially deposited pure and
mixed tailings (Hydrometrics, 1983; MultiTech, 1987b; PTI, 1989a; CH2M Hill and Chen-
Northern, 1990). Re-entrainment and transport of streamside tailings constitute a continuing
release and transport of hazardous substances; hazardous substances in floodplain sediments
will be gradually transported downstream.

During warm dry periods, moisture in the tailings moves upward and evaporates, leaving a
metal-salts precipitate on the tailings surface that forms a crust up to several centimeters thick
(MultiTech, 1987b; Nimick, 1990). The crusts dissolve easily in rainwater, and runoff
becomes very acidic and metal-rich (see Lipton et al., 1995) Since the tailings have low
permeability, very little precipitation infiltrates below the surface; most of the runoff
discharges into the stream (Nimick, 1990). These pathways contribute to additional metals
loadings in surface water resources of Silver Bow Creek.

Riparian soils and floodplain sediments of Silver Bow Creek themselves serve as a pathway
of contamination to surface water resources of Silver Bow Creek. Silver Bow Creek
redistributes streambank and floodplain sediments by entraining and redepositing them in bed,
bank, and overbank deposits. Through mass wasting, bank erosion, and slumping, and by
way of surface runoff over easily dissolved metals salts that form on tailings deposits,
hazardous substances are re-released to surface water resources (MultiTech, 1987b)

5.1.3 Extent of Pathway Contamination — Silver Bow Creek

As a result of hazardous substance releases to and transport within Silver Bow Creek,
suspended, bank, and bed sediments throughout the creek are contaminated with hazardous
substances relative to baseline (see Chapter 6.0). The entire floodplain of Silver Bow Creek
is contaminated with hazardous substances and is a continuing source of hazardous substances
to Silver Bow Creek surface water resources. Therefore, the surface water pathway to
riparian soils and sediments extends the length of Silver Bow Creek.

Chemical analysis of floodplain materials from Silver Bow Creek indicates that tailings
materials contain elevated concentrations of hazardous substances and low pH. The presence
of hazardous substances in slickens and mixed tailings deposits and riparian soils has been

RCG/Hagler Bailly

5-4

confirmed by many studies through chemical analysis of deposited tailings and mixed tailings
and soil material (e.g., Peckham, 1979; Hydrometrics, 1983, 1987, Rice and Ray, 1984, 1985,
MultiTech, 1987a,b; MSU et al., 1989, PTI, 1989a,b, CH2M Hill and Chen-Northern, 1989,
1990, CH2M Hill et al., 1991; Appendix A). All data indicate that the tailings deposits and
soils impacted by tailings have concentrations of arsenic, cadmium, copper, lead, and zinc
well in excess of expected (baseline) concentrations in native rjparian soils.

Average concentrations in tailings and floodplain sediments for the length of Silver Bow
Creek and the Opportunity Ponds compiled from previous investigations are presented in
Table 5-2. Additional data are presented in Chapter 6.0.

Table 5-2

Average Total Metals and Arsenic Concentrations in Tailings and Floodplain Sediments

of Silver Bow Creek (in ppm)

Location on Silver Bow Creek

As

Cd

Cu

Pb

Zn

(n)

Montana Street to the Upper Metro Storm Drain1

+ Exposed tailings

601

9

1523

644

2854

14

►• Flue dust

761

15

477

2645

4510

2

*■ Floodplain sediments/mixed alluvium and tailings

306

10

1936

773

4217

26

►■ Waste rock - raihvav roadbed fill

298

3

1389

421

857

8

►• Waste rock - transported fill and alluvium

113

4

552

265

1571

11

►• Soil

97

1

570

565

1101

4

Below Colorado Tailings to Miles Crossing2

485

6.5

220

1280

3970

21

Below Colorado Tailings to Miles Crossing3

854

4.6

1230

370

842

41

Miles Crossing to Warm Springs Ponds2

371

6.2

2030

1040

2920

28

Butte to the Mill-Willow Bypass4

678

17.3

2520

1480

3790

35

Opportunity Ponds5

210

4.9

2030

474

1200

42

Baseline6

27.8

1.2

34.2

35.9

102.2

12

1 CH2M Hill and Chen-Northern, 1990. 4 MSU et al.,

1989.

2 Canonie, 1992. 5 Tetra Tech,

1987.

3 MultiTech, 1987b. 6 Chapter 6.0,

this repc

>rt.

RCG/Hagler Bailly

5-5

5.2 CLARK FORK RIVER

5.2.1 Identification of Sources of Hazardous Substances — Clark Fork River

The historical sources of contamination in the Clark Fork River pathways and resources are
the same as those described for Silver Bow Creek (see Lipton et al., 1995). Before 1918,
discharged wastes were neither treated nor impounded, and raw mining and mineral-
processing wastes were transported unobstructed downstream to the Clark Fork River. The
three Warm Springs Ponds, constructed in 1918, 1919, and 1959, were built to settle out
mining wastes, but it is estimated that by 1959, more than 1 10 million tons of metals-
contaminated wastes had been released directly to the Clark Fork River from Warm Springs
Ponds and Silver Bow Creek (Andrews, 1987). In addition, though the Warm Springs Ponds
were effective in settling a fraction of the suspended sediments transported by Silver Bow
Creek, large amounts of tailings and other mine waste material continued to be discharged to
the Clark Fork River, particularly during high flows, which by-passed the Warm Spring
Ponds, and during labor strikes when the ponds were not in full operation (Phillips, 1985;
Johnson and Schmidt, 1988).

Additional sources of hazardous substances to the Clark Fork River include historical
discharges from the Anaconda Smelter via outflow from the Opportunity Ponds, erosion and
transport via Mill Creek of soils contaminated by stack emissions, and erosion and transport
of tailings dumps or other waste deposits along Warm Springs Creek (see Lipton et al., 1995).
A continuing source of hazardous substances to the surface water pathway in the Clark Fork
River is erosion and re-entrainment of contaminated riverbanks and floodplains. Through
mass wasting, bank slumping, and surface runoff over exposed tailings deposits, hazardous
substances are re-released to the Clark Fork River.

5.2.2 Identification of Transport Pathways — Clark Fork River

Hazardous substances have been transported from sources to the floodplains of the Clark Fork
River via Silver Bow Creek surface water resources and, to a lesser extent, Mill Creek and
Warm Springs Creek. The uppermost sediment layer along much of the Clark Fork River lies
above the native, pre-mining, floodplain, and consists of tailings and mixed alluvium and
tailings. Old, barren meander channels within the upper Clark Fork floodplain are evidence
of the fluvial transport of hazardous substances.

The thick overbank deposits along the Clark Fork River resulting from large sediment releases
of mining wastes ensure persistent contamination of water and bed sediments of the Clark
Fork River through continual release of contaminated floodplain soils and sediments caused
by surface water runoff following snowmelt or precipitation, mass movement during floods,
and riverbank wasting and slumping (Moore, 1985; Rice and Ray, 1985, Andrews, 1987;
Moore et al., 1989, Axtmann and Luoma, 1991). The mass of contaminated floodplain

RCG/Hagler Bailly

5-6

sediments and soils will continue to serve as a storage-and -release pathway of hazardous
substances to surface water and bed sediments, and through river flooding cycles and under
the influence of gravity, hazardous substances will continue to move downstream at a very
slow rate (Nimick, 1990).

5.2.3 Extent of Pathway Contamination — Clark Fork River

As a result of hazardous substance releases to and transport within Silver Bow Creek and the
Clark Fork River, particularly before the construction of the Warm Springs Ponds, bed
sediments throughout the river are contaminated with hazardous substances relative to
baseline conditions (see Lipton et al., 1995).

The original, pre-mining floodplain of at least the first 10 kilometers (6 miles) of the Clark
Fork River is buried by tailings and mixed tailings and alluvium (Nimick, 1990). Mixtures of
cleaner fluvial sediments have been deposited on top of tailings, and tailings are continually
cycled between the channel and the floodplain. Average metals and arsenic concentrations in
floodplain sediments are listed in Table 5-3 and illustrate the extent of soil resources exposed
by the surface water pathway.

Table 5-3

Concentrations in Tailings and Floodplain Sediments of the Clark Fork River (ppm)

Location on the Clark Fork River

As

Cd

Cu

Pb

Zn

(n)

Below Warm Springs Ponds extending 10 km north1

769

3.64

4532

712

1839

83

North of Warm Springs Ponds:

600

5.7

3662

547.3

2206

67

Warm Springs Ponds to Deer Lodge3

634

8.8

1760

461

1160

8

Warm Springs ponds to Drummond4

292

NA

2183

262

1298

240

Warm Springs Ponds to Turah5

NA

9.3

1147

164

2529

17

North of Deer Lodge6

176

5.0

1630

NA

NA

40

Deer Lodge to Drummond3

610

8.4

1090

398

1120

9

Drummond to Milltown3

116

10

783

87

2660

9

Baseline7

27.8

1.2

34.2

35.9

102.2

12

1 Nimick, 1990. 5 Axtmann and Luoma, 1991

2 Brooks, 1988. 6 Rice and Ray, 1984.

3 Moore, 1985. 7 Chapter 6.0, this report.

4 CH:M Hill et al., 1991. NA = Not analyzed.

RCG/Hagler Bailly

5-7

5.3 OPPORTUNITY PONDS

Historical sources of hazardous substances to the riparian soils in the Opportunity Ponds area
included Anaconda mill tailings, plant effluents, discharges from smaller ponds, and, between
1932 and 1947, Silver Bow Creek water. The ponds were constructed to receive smelting-
related wastes, and to clarify process wastewater before discharge to the Warm Springs Ponds
or the Clark Fork River.

According to a report prepared by the Anaconda Minerals Company in the mid-1970s (AMC,
date unknown), Pond A (approximately 215 acres) was constructed in 1914, and by 1919 was
filled and abandoned. Ponds B and C were constructed concurrentiy, and discharges to the B
Ponds (approximately 925 acres) began in 1918. During the 1950s, Pond B was dredged, and
the wastes transported to the Butte mines for fill. Pond B-2 later received fine-grained
tailings from the electric furnace slag project. The C Ponds (approximately 1 180 acres)
received effluent from the B Ponds, and discharged to the D ponds. The D Ponds
(approximately 810 acres) were filled with tailings deposited in the 1960s and 1970s The D-
2 Pond was built on top of 100 acres of slime ponds built in 1910 and abandoned in 1918
(TetraTech, 1987).

5.4 REFERENCES

AMC. Date unknown. Tailings Ponds. Section of Anaconda Minerals Co. document (1970s)
that describes tailings ponds and related area used by the Company.

Andrews, E.D. 1987. Longitudinal dispersion of trace metals in the Clark Fork River,
Montana. In R.C. Everett and DM. McKnight (eds), Chemical Quality of Water in the
Hydrologic Cycle. Lewis Publishers, Chelsea, MI.

Axtmann, E.V. and S.N. Luoma. 1991. Large-scale distribution of metal contamination in
the fine-grained sediments of the Clark Fork River, MT. Appl. Geochem. 6:75-88.

Brooks, R. 1988. Distribution and concentration of metals in sediments and water of the
Clark Fork River floodplain, Montana. M.S. Thesis, University of Montana, Missoula, MT.

Camp Dresser and McKee. 1991. Statement of Work: Remedial Investigation/Feasibility
Study, Priority Soils Operable Unit - Silver Bow Creek/Butte Area NPL Site, Butte, MT
(Final Draft). Prepared by CDM Federal Programs Corp. for the U.S. EPA - Region VIII,
Helena, MT.

Canonic 1992. 1991 Remedial Investigation Activities, Data Summary Report: Silver Bow
Creek/Butte Area NPL Site, Streamside Tailings Operable Unit RI/FS. Prepared for ARCO
by Canonie Environmental Service Corp., Bozeman, MT.

RCG/Hagler Bailly

5-8

CH2M Hill and Chen-Northern. 1989. Final: Public Health and Environmental Assessment
Data Summary Report, Rocker and Ramsay Areas, Silver Bow Creek CERCLA Site,
Montana. Prepared for Montana Department of Health and Environmental Sciences, Helena,
MT.

CH2M Hill and Chen-Northern. 1990. Silver Bow Creek CERCLA Phase II Remedial
Investigation Data Summary, Area I Operable Unit - Volume I and Volume n (Draft Final).

CH2M Hill, Inc, Chen-Northern, and Montana State University Reclamation Research Unit.
1991. Draft Final: Upper Clark Fork River Screening Study. Volume 1. Prepared for
Montana Department of Health and Environmental Sciences.

Hydrometrics, Inc. 1983. Summit and Deer Lodge Valleys long term environmental
rehabilitation study, Butte-Anaconda, MT. Prepared for the Anaconda Minerals Company.

Hydrometrics. 1987. Investigation of Potential Resources Contamination Near Rocker,
Montana. Prepared for Anaconda Minerals Co., Anaconda, MT.

Johnson, H.E. and C.L. Schmidt. 1988. Clark Fork Basin Project: Status Report and Action
Plan. Prepared for the Office of the Governor, Helena, MT.

Lipton, J., H. Bergman, D. Chapman, T. Hillman, M. Kerr, J. Moore, and D. Woodward.
1995. Aquatic Resources Injury Assessment Report, Upper Clark Fork River Basin. Prepared
by RCG/Hagler Bailly for the State of Montana, Natural Resource Damage Litigation
Program. January.

Moore, J.N. 1985. Source of Metal Contamination in Milltown Reservoir, Montana: An
Interpretation Based on Clark Fork River Bank Sediment. Prepared for U.S. EPA.
September. 60 pp.

Moore, J.N., E.J. Brook, and C. Johns. 1989. Grain size partitioning in contaminated, coarse
grained river floodplain sediment, Clark Fork River, MT. Environ. Geol. Water Sci. 14: 107-
115.

MSU/RRU, Schafer and Associates, and CH2M Hill. 1989. Silver Bow Creek RI/FS
Streambank Tailings and Revegetation Studies: STARS Phase I Bench-Scale Soil Column and
Greenhouse Treatability Studies, and Tailings Ranking System. Final Summary Report:
Volume I. Prepared for Montana Department of Health and Environmental Sciences by
Montana State University/Reclamation Research Unit, Schafer and Associates, and CH2M
Hill, Inc.

MultiTech. 1987a. Silver Bow Creek Remedial Investigation Final Report Summary.
Prepared for Montana Department of Health and Environmental Sciences, Helena, MT.

RCG/Hagler Bailly

5-9

MultiTech. 1987b. Silver Bow Creek Remedial Investigation Final Report Appendix B, Part
1, Report: Groundwater and Tailings Investigation and Part 4, Attachment VLChemical Data.
Prepared for Montana Department of Health and Environmental Sciences, Helena, MT.

Nimick, DA. 1990. Stratigraphy and Chemistry of Metal-Contaminated Floodplain
Sediments, Upper Clark Fork River valley, Montana. M.S. Thesis, University of Montana,
Missoula, MT. 118 pp.

Peckham, A.E. 1979. Metals Assessment of Silver Bow Creek Between Butte and Gregson,
Montana. Prepared by National Enforcement Investigations Center, Denver, CO, June 1979.
32 pp.

Phillips, G.R. 1985. Relationships among fish populations, metals concentrations, and stream
discharge in the upper Clark Fork River. In C.E. Carlson and L.L. Bahls, eds., Proceedings
of the Clark Fork River Symposium. Butte, MT: Montana College of Mineral Science and
Technology, pp. 57-73.

PTI. 1989a. Silver Bow Creek Tailings Investigation, Draft Report. Prepared by PTI
Environmental Services for Parcel, Mauro, Hultin & Spaanstra, Denver, CO. 23 pp.

PTI. 1989b. Draft Report: Silver Bow Creek Water Quality and Sediment Sampling Data.
Prepared by PTI Environmental Services for Parcel, Mauro, Hultin & Spaanstra, Denver, CO.
Rice, P.M. and G.J. Ray. 1984. Floral and Faunal Survey and Toxic Metal Contamination
Study of the Grant-Kohrs National Historic Site. Project for the Gordon Environmental
Studies Laboratory.

Rice, P.M. and G.J. Ray. 1984. Floral and faunal survey and toxic metal contamination
study of the Grant-Kohrs National Historic Site. Project for the Gordon Environmental Studies
Laboratory, University of Montana, Missoula, MT.

Rice, P.M. and G.J. Ray. 1985. Heavy metals in flood plain deposits along the upper Clark
Fork River. In C.E. Carlson and L.L. Bahls (eds), Proceedings of the Clark Fork
Symposium. Montana Academy of Sciences, Montana College of Mineral Science and
Technology, Butte, MT.

Tetra Tech. 1987. Anaconda Smelter Remedial Investigation/Feasibility Study: Master
Investigation Draft Remedial Investigation Report. Prepared for Anaconda Minerals Company
by Tetra Tech, Inc. Bellevue, WA.

RCG/Hagler Bailly

6J_

6.0 GEOLOGIC RESOURCE — RIPARIAN SOILS

This chapter described injuries to riparian soils along Silver Bow Creek, along the upper
Clark Fork River (from Warm Springs Ponds to Deer Lodge), and in the Opportunity Ponds
Overall, the data presented in this chapter demonstrate that:

► Concentrations of hazardous substances in riparian soils of Silver Bow Creek,
the upper Clark Fork River, and the Opportunity Ponds are significantly greater
than baseline conditions.

► Riparian soils of Silver Bow Creek, the upper Clark Fork River, and the
Opportunity Ponds are severely phytotoxic, as demonstrated in controlled
laboratory tests, and hence are injured.

► Riparian soils of Silver Bow Creek, the upper Clark Fork River, and the
Opportunity Ponds have pH values less than 4.0, and hence are injured

► Injury to riparian soils of Silver Bow Creek, the upper Clark Fork River, and
the Opportunity Ponds is further confirmed by field studies quantifying a
significant reduction in vegetation.

► The areal extent of injury, delineated by the area of riparian soils effectively
devoid of vegetation, covers approximately 800 acres along Silver Bow Creek,
215 acres along the upper Clark Fork River, and some 3,400 acres in the
Opportunity Ponds area.

6.1 DESCRIPTION OF RIPARIAN SOIL RESOURCE

Geologic resources that have been injured by releases of hazardous substances from mining
and mineral processing operations in Butte and Anaconda include the riparian soils and
floodplain sediments adjacent to Silver Bow Creek and the upper Clark Fork River from
Warm Springs Ponds to Deer Lodge, and former low-lying soils associated with Mill and
Willow Creeks and minor drainages and springs in the Opportunity Ponds area.

6.1.1 Silver Bow Creek

Silver Bow Creek is a perennial stream that begins at the confluence of the Metro Storm
Drain and Blacktail Creek in Butte, and flows approximately 24 miles (40 kilometers) west
and north to discharge into the Warm Springs Ponds (Canonie, 1992a). Approximately 22
miles of Silver Bow Creek, from below the Colorado Tailings to the Warm Springs Ponds,

RCG/Hagler Bailly

6-2

constitute the Streamside Tailings Operable Unit of the Silver Bow Creek/Butte Area NPL
Site.

The uppermost sediment layer along Silver Bow Creek lies above the naturally formed (pre-
mining) floodplain (MultiTech, 1987a). The natural floodplain soils contain dead vegetation
partially buried by deposited tailings and tailings-contaminated alluvium, indicating the recent
and common origin of the sediment layer (MultiTech, 1987a). Evidence compiled during the
Lower Area I Remedial Investigation (CH2M Hill and Chen-Northern, 1990) indicates that
soils within the operable unit were originally developed on upland slopes under coniferous
forests, or in valley-fill sediments under grassland vegetation. Historical maps show that
Silver Bow Creek and Blacktail Creek riparian areas supported extensive marshy areas and
lowland swamps (CH2M Hill and Chen-Northern, 1990).

6.1.2 Clark Fork River

The headwaters of the Clark Fork River are formed by the discharge of the Warm Springs
Ponds, and the confluences of the Mill-Willow Bypass and Warm Springs Creek. The upper
Clark Fork flows north and west to the Milltown Reservoir, a distance of approximately 120
miles (195 kilometers). Between the Warm Springs Ponds discharge and Deer Lodge,
approximately 17 river miles, concentrations of hazardous substances are elevated and
hundreds of acres of unvegetated tailings deposits occur intermittently.

The uppermost sediment layer along much of the Clark Fork River lies above the naturally
formed floodplain (Moore, 1985) and it is estimated that visible slickens areas constitute
approximately 18% of the tailings material stored in the floodplain (Nimick, 1990).

6.1.3 Opportunity Ponds

The Opportunity Ponds are a series of tailings ponds constructed incrementally between 1914
and the late 1950s to treat mill tailings and wastes from the Anaconda smelter (MultiTech,
1983). Current drainage patterns suggest that portions of the approximately 3,400 acres now
covered by the Opportunity Ponds in the past supported wetland communities associated with
Warm Springs Creek, Mill Creek, Willow Creek, and additional minor drainages and springs.
Approximately 50 acres in Pond D-l of the Opportunity Pond complex remains damp most of
the year because of groundwater seepage (AMC, date unknown); the natural groundwater
seepage most likely supported a wetland community before the deposition of tailings.

Inflows to the Opportunity Ponds included Anaconda mill tailings, plant effluents, and
discharges from smaller ponds. Between 1937 and 1942, a dam near Gregson diverted water
from Silver Bow Creek to the Opportunity Ponds (MultiTech, 1987a). Estimates of the
smelter waste stream volume vary, but approximately 10-30 cubic feet per second (6-19

RCG/Hagler Bailly

6-3

million gallons per day) of wastes containing tailings slurry were discharged to the ponds
(Tetra Tech, 1987).

6.2 INJURY DEFINITION

In this assessment, injury to riparian soil is defined as:

*• Concentrations in the soil of substances sufficient to cause a phytotoxic

response such as retardation of plant growth [43 CFR § 11.62 (e) 10]

► Concentrations of substances sufficient to raise the . . . soil pH to above 8.5 or
to reduce it to below 4.0 [43 CFR § 11.62 (e) 2]

► Concentrations of substances sufficient to have caused injury as defined to
surface water, groundwater, air or biological resources when exposed to the
substances [43 CFR § 11.62 (e) 11].

These environmental responses are described in Chapter 3.0 and in Section 6.2.2, respectively.

6.2.1 Assessment of Phvtotoxicity

As described in Chapter 3.0, controlled laboratory tests were used to evaluate phvtotoxicity
based on standard operating procedures consistent with the ASTM protocol for early seedling
growth studies (ASTM, 1994; Appendix B of this report). Screening tests examined the
toxicity of 100% impact site soils (Silver Bow Creek, the Clark Fork River, and Opportunity
Ponds) to the standard laboratory test species alfalfa, lettuce, and wheat over a two-week
exposure period. Extended tests examined the performance of hybrid poplar cuttings (a
representative taxonomic surrogate for willow and Cottonwood species) exposed to 100%
impact site soils for four weeks. Performance of test species in impact site soils was
compared to performance of control species grown in control soils. Control soils were
collected from control sites previously designated and sampled as representative of baseline
conditions (see Appendix A).

As described in Chapter 3.0, determination of phvtotoxicity was based on a suite of species-
endpoint responses, including:

► Germination (number of seedlings emerged/20 seeds; five replicates of 20 seed
each)

► Shoot height (mm, for each surviving seedling; mean of plants measured)

RCG/Hagler Bailly

^ 6-4

► Maximum root length (mm, for each surviving seedling; mean of plants
measured)

► Shoot mass (g oven dry weight for each seedling, mean of plants measured)

► Root mass (g oven dry weight for each surviving seedling; mean of plants
measured)

► Total mass (g oven dry weight for each surviving seedling; mean of plants
measured).

A response was deemed phytotoxic when plants in impact soil exhibited any of the above
endpoints that was statistically less than the corresponding response of plants grown in control
soils. The degree of phytotoxicity was quantified based on (1) the magnitude of the
difference from control performance, (2) the number of endpoints exhibiting a phytotoxic
response, and (3) the number of species exhibiting a phytotoxic response. Quantitative
phytotoxicity scores were then assigned to each soil sample based on the degree of phytotoxic
response. These toxicity "scores," in turn, were used to classify soils as either:

► Nontoxic

► Phytotoxic:

■ Mildly toxic

■ Moderately toxic

■ Highly toxic

■ Severely toxic.

Differences between impact and control samples were assessed statistically using randomized
t-tests.

Secondary Effects of Phytotoxiciry/Implications

The presence of phytotoxic concentrations in soils results in injury to vegetation communities.
Phytotoxic injury to vegetation communities is manifested as a change in the plant population
density, species composition, diversity, dispersion, or percent cover [43 CFR §11.71 (6)], and
can be determined by reference to an established baseline area. Injuries to riparian vegetation
resources are addressed in Chapter 7.0, however, obvious phytotoxic injuries in the form of
devegetated riparian zones are indicative of injury to soil and will be discussed in the injury
determination and quantification section of this report. ••

RCG/Hagler Bailly

6-5

6.2.2 Changes in pH

Soil pH, the negative logarithm of the hydrogen ion concentration, is a measure of the acidity
of a soil. Soil pH reflects the mineral content of the parent material and the degree of soil
weathering, and is typically an indicator of soil contamination (Hassett, 1992). Soil pH
values below 4.0 suggest that oxidation of reduced sulfur compounds has occurred or is
occurring in the soil (Hassett, 1992). The effect of low pH in soils contaminated with metal
and metalloid hazardous substances, in general, is increased mobility and plant availability of
soil contaminants (Figure 6-1). Soil pH of less than 4.0 resulting from the release of
hazardous substances constitutes an injury to the soil resource [43 CFR § 11.62 (e)(2)].

6.3 INJURY DETERMINATION AND QUANTD7ICATION

6.3.1 Baseline Conditions

Baseline is defined as the condition or conditions that would have existed absent the releases
of hazardous substances [43 CFR § 11.14 (e)]. Historical data regarding the baseline
conditions of Silver Bow Creek and the Clark Fork River riparian soils and floodplain
sediments before releases of hazardous substances are not available because releases have
occurred since the late 1800s. Nor are adequate upstream control areas available, because
Silver Bow Creek has been drastically disturbed by mining nearly to its headwaters
Therefore, baseline conditions were quantified using comparable streams in southwestern
Montana that have not been exposed to hazardous substances.

Control areas against which Silver Bow Creek and the upper Clark Fork River riparian soils
and floodplain sediments have been compared to determine injury include Divide Creek, Little
Blackfoot River, and Flint Creek (see Appendices A and C). Reaches of each of these
streams were identified as comparable in geomorphology and hydrology to specific reaches of
Silver Bow Creek and the Clark Fork River, and each was sampled to determine
representative concentrations of arsenic, cadmium, copper, lead, and zinc in riparian soils and
floodplain sediments (see Appendix A). Arsenic, lead, and zinc concentrations measured in
Flint Creek floodplain soils confirm that mining operations and associated releases to Flint
Creek have contaminated the floodplains with elevated concentrations of hazardous substances
(see Appendices A and C). Therefore, Flint Creek does not reflect baseline conditions for
floodplain soils and is not used in calculating the baseline for Silver Bow Creek, the Clark
Fork River, and Opportunity Ponds soils. Baseline conditions (arsenic = 27.8 ppm,
cadmium = 1.2 ppm, copper = 34.2 ppm, lead = 35.9 ppm, and zinc = 102.2 ppm) were
established as the averaged concentrations of six samples from Divide Creek and six samples
from Little Blackfoot River, each a composite of five sub-samples (see Table 6-1 and
Appendix A).

RCG/Hagler Bailly

6-6

Figure 6-1.

[-?

o

\NNv

>s

Vn Nv\

s

\ \\r

o>

\\ \ >%*.

c

'••> \ \

s- \ \

V)

V0> c \ \

-Mobility
eok Medium

*

\ \ \

1

"... ^ ^

1

2 3 4 5 6 7

pH of soil

Trends in the Mobility of Metals as Influenced by Soil pH; Data for Light
Mineral Soil (Kabata-Pendias and Pendias, 1992).

Table 6-1

Baseline and Background Concentrations (ppm) in Riparian Soils and Baseline Conditions for

this NRDA (bold) are the Mean of Divide Creek and Little Blackfoot River Samples

Location

Arsenic

Cadmium

Copper

Lead

Zinc

Divide Creek

Little Blackfoot River

27.45
28.13

1.7
0.7

43.4
25.0

26.2
45.6

92.3
112.0

Silver Bow Creek1
(native floodplain)

11.1

NA

26.6

19.5

98.5

Tin Cup Joe Creek2
Blackfoot River

26
4

1.3
<0.03

53
13

NA
NA

NA
NA

Clark Fork
tributaries3

26.5

<2.5

27.0

24.0

94.0

Baseline

27.8

1.2

34.2

35.9

102.2

1 Canonie, 1992a 2 Rice and Ray, 1984. 3 Moore et al., 1989.

RCG/Hagler Bailly

6-7

Previous investigations have identified background concentrations for specific sections of
Silver Bow Creek and the Clark Fork River. Table 6-1 presents results of three studies that
agree with the baseline concentrations of hazardous substances in floodplain soils determined
for this NRDA.

Silver Bow Creek RI/FS boreholes (Canonie, 1992a) drilled to sample the depth of
contamination along Silver Bow Creek floodplains indicate that concentrations stabilized at
minimum background values more than 10 feet below the surface. Figure 6-2 illustrates the
exponential decrease in concentration with depth for arsenic, copper, lead, and zinc These
background values presumably provide pre-release concentrations for comparison.

As part of a study of arsenic, cadmium, and copper contamination of the Grant-Kohrs Ranch
near Deer Lodge, Rice and Ray (1984) compared observed concentrations in the Clark Fork
River floodplain soils to concentrations in soils collected from Tin Cup Joe Creek, 5 miles
southwest of the Grant-Kohrs ranch, and in soils from another control site on the Blackfoot
River, 60 miles northwest of Grant-Kohrs. Neither of the two sampling sites had been
exposed to contaminated over-bank sediments; background concentrations determined from
them are presented in Table 6-1. In addition, background concentrations of total metals in 24
floodplain sediment samples taken from the Clark Fork tributaries are presented in Table 6-1
as average concentration in ppm (Moore et al., 1989).

Overall, the results of these studies confirm the baseline concentrations shown in Table 6-1.
Indeed, these baseline conditions are somewhat conservative when compared to the Silver
Bow Creek native floodplain.

6.3.2 Injury Determination

6.3.2.1 NRDA Riparian Soil Sampling

As a component of the soil resource investigation of this NRDA, soil samples were collected
from three riparian areas in the upper Clark Fork River Basin: Silver Bow Creek from
downstream of the Colorado Tailings to the Warm Springs Ponds, the upper Clark Fork River
from the Warm Springs Ponds discharge to Deer Lodge, and the Opportunity Ponds. Silver
Bow Creek and the upper Clark Fork River had been previously subdivided into four reaches
during the work to determine injury to fish habitat and populations (Lipton et al., 1995).
Three reaches, defined by geomorphology and hydrology, were also used for the riparian
soils, vegetation, and wildlife sampling (see Appendix A). These reaches, referred to as
impact reaches, were:

*■ Silver Bow Creek from below the Colorado Tailings in Butte to the upstream

end of the Durant Canyon (Upper Silver Bow Creek)

RCG/Hagler Bailly

6-8

E
a
a

c

v

1800

1600

1400

1200

1000

800

600

400

200

0

r—

- •

—

_ •

—

E

"

■

c

—

a.

* •

u

—

o

- • •

a

•\ •

a

0

J" •

•
•

u

-i :

•

. • .

t •

-? *

M

»«

t • •«•

! . j

• T

1

10 20 30

Sample Depth (ft)

14000

12000

10000

8000

6000

4000

2000

0

40

— • L» j^r +J^mm *• • • • «* • • *■• *• • • • •

10 20 30

Sample Depth (ft)

40

10000

9000

_•

8000

7000

E
a
a

6000
5000

•

■c

0

4000

•

0)

3000

2000

•

1000
0

r •

i • • • m • • m* ••

E
a
a

o

c

10 20 30

Sample Depth (ft)

14000 -
12000 -
10000 -

8000

6000

4000

2000
0

40

* «

_L l- L_

* "T

10 20 30
Sample Depth (ft)

_J
40

Figure 6-2. Concentrations of Hazardous Substances in the Silver Bow Creek

Flood plain. Concentrations decrease with depth to approximately 10 feet
below the surface The asymptotic region of each graph indicates approximate
pre-mining concentrations in floodplain soils and sediments. Source: Canonie,
1992a.

RCG/Hagler Bailly

6-9

► Silver Bow Creek from the downstream end of the Durant Canyon to the Warm
Springs Ponds (Lower Silver Bow Creek)

► The Clark Fork River from the Warm Springs Ponds discharge to Deer Lodge
(the upper Clark Fork River).

The location of sampling transects is shown in Figure 6-3.

The Opportunity Ponds were also sampled as an impacted riparian area.

Sample sites on Silver Bow Creek and the Clark Fork River floodplains were located
randomly within unvegetated slickens to obtain a representative sample of slickens chemical
composition. Since there is no comparable reach upstream of the impacts, control streams
were selected for comparison of soils and riparian vegetation. Criteria used to select control
reaches were based on vegetation characteristics and the factors that control potential riparian
vegetation composition and adjacent upland vegetation composition: valley bottom type,
stream and valley gradient, channel sinuosity, elevation, and stream basin orientation (see
Appendix A).

Three stream reaches were identified as potential control sites for the three impact reaches.
Each was ground-truthed to ensure comparability. Divide Creek from the confluence of the
North and East Forks was selected as the control for upper Silver Bow Creek. The Little
Blackfoot River from Elliston to Avon was selected as the control reach for lower Silver Bow
Creek, and Flint Creek from Sheryl to the Clark Fork River confluence was selected as the
control for the upper Clark Fork River. The Opportunity Ponds were compared to the mean
of the sites sampled as controls for the other riparian impact reaches.

Seventeen composited impact soil samples were collected from unvegetated slickens on Silver
Bow Creek and the Clark Fork River, and 18 soil samples were composited from control
sites. Samples were dried, mixed, and split for analyses. All samples were analyzed for total
metals. Measured values of hazardous substances (total metals) in riparian impact and control
soils were compared by reach. Two sample randomization tests (5,000 randomizations)
(Manly, 1991) comparing impact and control reaches showed that concentrations of arsenic,
cadmium, copper, lead, and zinc were significantly higher in the riparian impact soils than in
their respective controls (Table 6-2). Concentrations of arsenic, copper, lead, and zinc in
Opportunity Ponds samples were significantly greater than in riparian control samples
(Table 6-3).

Summary statistics for riparian impact and control reaches are presented in Table 6-4 and
Figure 6-4. Means for arsenic in each impact reach exceeded means in the control reaches by
4.5 to 13.7 times. Cadmium means in the control streams were exceeded by 3.8 to 7.2 times
in the impact reaches, copper by 10 to 40 times, lead by 2.2 to 22.5 times, and zinc by 3.5 to
30.5. The comparisons confirm that samples collected from unvegetated slickens along Silver

RCG/Hagler Bailly

__ 6-10

Bow Creek, the Clark Fork River, and the Opportunity Ponds contain extremely elevated
concentrations of arsenic, cadmium, copper, lead, and zinc relative to baseline concentrations
(Table 6-4). In addition, the impact data are comparable to data collected in previous
investigations of slickens (Tables 6-5 to 6-8).

6.3.2.2 NRDA Phytotoxicity Tests

To assess phytotoxic responses, controlled laboratory experiments were conducted to test
seedling germination and plant growth performance in impact soils relative to control soils.
Phytotoxicity was defined as a statistically significant decrement in plant growth (germination,
shoot length, root length, shoot mass, root mass, and total mass) relative to control plants.
The results of phytotoxicity tests of riparian soils and floodplain sediments of Silver Bow
Creek and the Clark Fork River, and soil material in the Opportunity Ponds, indicate that
these soils are extremely phytotoxic (Appendix B).

Four soil samples collected from slickens along Silver Bow Creek and the Clark Fork River
and one from the Opportunity Ponds were used as substrate for the screening tests. Standard
reference plants (alfalfa, lettuce, and wheat) were used in the screening tests.
Germination/emergence, the first toxicity endpoint, was assessed as the number of seedlings
visible above the soil surface after 7 days and at the end of the 14-day exposure period. For
each riparian impact site, the mean emergence of five trials (20 seeds each) per species was
compared to the emergence determined in the control sample using a t-test. Alfalfa and
wheat emergence were significantly less than controls for sites R-04, R-08, and R-14
(Table 6-9). Lettuce emergence was significantly less than controls for all impact samples
(R-04, R-08, R- 13, and R-14).

Alfalfa, lettuce, and wheat shoot height and root length for all six impact sites tested,
including the Opportunity Ponds, were significantly less than those measured for the control
soils. Shoot and root mass of alfalfa and lettuce were significantly less than masses
determined for the control samples in all six impact site soils tested, and wheat shoot and root
mass were significantly less than the control in five of the site soils tested. Based on the
results of seedling emergence and growth of alfalfa, lettuce, and wheat, Silver Bow Creek, the
Clark Fork River, and Opportunity Ponds soils were classified as severely phytotoxic.

Four soil samples collected from slickens along Silver Bow Creek and the Clark Fork River
were used as substrate for the extended impact tests (R-04, R-08, R-13, and R-14). Five
hybrid poplar stems were planted in each of the four impact soil samples, for an initial total
of 20 test plants. A composite of soil collected from the Little Blackfoot River and Flint
Creek was used as control soil for the extended tests. Five hybrid poplar stems were planted
in the control soil.

RCG/Hagler Bailly

6-11

Figure 6-3. Location of Riparian Soil, Vegetation, and Wildlife Sampling Reaches:
Silver Bow Creek and the Clark Fork River.

RCG/Hagler Bailly

6-12

1

Table 6-2
Two-Sample Randomization Test of the Mean Difference Between Impact and Control Soil

Concentrations (ppm)

Test

Upper Silver Bow Creek

vs. Divide Creek

n - 6 pairs

Lower Silver Bow Creek
vs. Little Blackfoot

n ■= 6 pairs

Upper Clark Fork

River vs. Flint Creek

n = 5 pairs

Hazardous
Substance

Mean
Difference

p-value <

Mean
Difference

p-value <

Mean
Difference

p-value <

Arsenic

Cadmium

Copper

Lead

Zinc

347.68

7.18

1317.28

564.87

2725.25

0.001*

0.0042*
0.0014*
0.0002*
0.001*

235.97

4.32

880.43

431.88

1392.33

0.0012*
0.0012*
0.0014*
0.0012*
0.0006*

303.66
3.42
1962.7
166.6
880.6

0.0034*
0.0178*
0.0038*
0.0046*
0.001*

* indicates sigi

uficantiy greater concentration

in the impact samples at a =

3.3%.

Table 6-3

Results of the Two-Sample Randomization Test Comparing Concentrations

in Opportunity Ponds Soils and Riparian Control Reach Soils

Hazardous Substance

Mean Difference

p-value <

Arsenic

Cadmium

Copper

Lead

Zinc

351.02

13.53

2314.55

238.95

563.39

0.0014*
0.056 **
0.0002*
0.0064*
0.0358*

* and ** indicate greater concentration in the Opportunity Ponds at a = 5% and a = 6%, respectively.

RCG/Hagler Bailly

6-13

Table 6-4

Summary Statistics for Ripar

ian Impact and Control Reach 2-Inch Soils Samples

(total metals in ppm)

Sample Identification

As

Cd

Cu

Pb

Zn

Upper Silver Bow Creek n = 6

Arithmetic mean

374.6

8.8

1361.0

591.1

2816.6

Maximum

509.0

17.8

4014.0

885.8

5108.0

Minimum

274.3

3.7

478.0

392.0

1460.9

Standard deviation

81.1

5.1

1252.3

178.5

1399.3

Divide Creek (control) n = 6

Arithmetic mean

27.5

1.7

43.4

26.2

92.3

Maximum

76.4

5.8

56.4

40.0

136.0

Minimum

10.0

0.5

28.7

17.3

67.3

Standard deviation

22.2

1.9

12.3

8.2

25.8

Lower Silver Bow Creek n = 6

Arithmetic mean

264.1

5.0

905.5

477.5

1504.3

Maximum

425.6

8.3

1414.5

836.6

2152.0

Minimum

163.6

1.7

407.5

307.3

720.5

Standard deviation

110.5

2.5

408.3

195.6

590.5

Little Blackfoot (control) n = 5

Arithmetic mean

28.1

0.7

25.0

45.6

111.9

Maximum

44.1

1.4

43.6

100.1

169.7

Minimum

12.5

0.3

13.8

19.1

62.4

Standard deviation

12.9

0.4

10.5

26.6

39.3

Upper Clark Fork River n = 5

Arithmetic mean

391.3

4.7

2012.6

292.6

1230.3

Maximum

525.5

8.5

3644.0

360.0

1607.2

Minimum

291.8

1.1

566.5

236.8

549.5

Standard deviation

92.1

2.4

987.8

49.3

359.9

Flint Creek (control) n = 5

Arithmetic mean

87.6

1.2

49.9

127.8

349.7

Maximum

145.4

1.9

72.8

205.7

523.8

Minimum

36.6

0.6

28.2

51.2

194.4

Standard deviation

39.7

0.6

17.3

60.3

1346

Opportunity Ponds n = 5

Arithmetic mean

396.4

14.7

2353.4

301.9

738.3

Maximum

1398.1

68.1

9980.0

722.6

2676.0

Minimum

18.5

0.7

301.6

54.2

68.0

Standard deviation

506.5

26.7

3816.3

257.2

977.7

RCG/Hagler Bailly

6-14

400

a- 300

B 200

100

**

Upper SBC Upper CFR

Lower SBC Opportun. Pnds.

Impact Soil
Control Soil

***

2500

**

Upper SBC Upper CFR

Lower SBC Opportun. Pods.

Upper SBC Upper CFR

Lower SBC Opportun. Pnds.

3000

Upper SBC Upper CFR

Lower SBC Opportun. Pndi.

Upper SBC Upper CFR

Lower SBC Opportun. Pndi.

Figure 6-4. Mean Concentrations of Hazardous Substances at Impact and Control
Reaches. * indicates significant difference at o = 3.3%; ** indicates
significant difference at a = 5%, *** indicates significant difference at o = 6%.

RCG/Hagler Bailly

6-15

Table 6-5
Streambank Tailings and Rcvegetation Study (STARS)
Collected from Unvegetated Strcamsidc Tailings
along Silver Bow Creek

Data

Sample

As

CM

Cu

Pb

Zn

PH

Depth(ui)

1

1003.0

13.2

2130.0

1490.0

3940.0

4.2

0-15

(TEST) 2

1745.0

108.0

11200.0

1960.0

220000

3.8

5-15

3

354.0

26.9

208.0

9O8.0

5580.0

2.7

0-15

4

146.0

4.5

2150.0

398.0

2970.0

43

0-15

5

975.0

322

8620.0

3100.0

74904

4.4

0-13

(TEST) 7

83.6

13.7

461.0

341.0

3840.0

3J

0-15

6

2124

15.7

6160.0

4920.0

78604

6.6

13-28

8

491.0

8.4

456 J)

11 60 J)

2480.0

4.1

0-15

(TEST) 9

928.0

19.5

4040.0

6477.0

102000

5.8

20-35

10

51.2

3*

811.0

3124

814.0

22

0-8

11

2544.0

38.6

1600.0

1672.0

20404

3.4

0-6

12

38.9

15.0

3920.0

264.0

39004

U

2-17

13

285.0

1£

374.0

578.0

1380.0

3£

0-15

14

3140.0

36.4

5530.0

33804

5290.0

3j6

0-15

15

3510

9.4

391.0

504.0

2670.0

3.8

0-15

16

1040.0

32.1

7430.0

27204

49.4

6.1

0-13

17

204.0

4.7

260.0

273.0

1190.0

3.1

2-12

18

1755.0

23.1

4070.0

1986.0

4460.0

32

0-15

19

1530.0

22.6

2430.0

1560.0

4120.0

33

0-15

20

341.0

as

685.0

682.0

5020.0

3.6

0-15

JTEST) 21

2560.0

37.0

3900.0

2980.0

74304

3.7

0-15

22

1050.0

21.9

4650.0

2730.0

4150.0

4.6

34-49

23

651.0

2£L5

5080.0

1840.0

5500.0

4.5

0 -12

24

414.0

16.6

1550.0

1350.0

5580.0

3.7

0-15

25

93.5

4.1

390.0

442.0

937.0

23

0-8

26

261.0

6.4

868.0

635.0

1760.0

33

0-15

fTEST) 27

235.0

12.6

1980.0

2520.0

19.2

4.4

0-6

28

203.0

14.4

779.0

2180.0

25O0.0

4.6

0-8

29

83.9

5.7

329.0

125.0

476.0

53

0-8

30

216.0

2.6

470.0

563.0

938.0

3A

0-15

31

114.0

4.1

1290.0

3674

5594

Z&

0-15

(TEST) 32 i

19.3

u

1930.0

82.6

2240.0

7.4

12-27

33

209.0

M

871.0

4674

1110.0

3.4

0

34

195.0

5.6

761.0

375.0

1440.0

Z7

0

35

779.0

196.0

3.6

534.0

456.0

NR

0

Concentrations reported are total metals in ppm. TEST indicates that the sample
greenhouse phytotoxiciry tests; in all six sous tested, plants were unable to establisl
indicating severe phytotoxicify. Shading indicates pH below 4.0, a criterion for inji
determination in soil.
Source: MSU et al, 1989a1b

was used in
iry

RCG/Hagler Bailly

6-16

Table 6-6

Tailings Data Collected for the Silver Bow Creek RI

from Ramsay Flats Tailings Material

Site #

Horizon

Soil Type

PH

As

Cd

Cn

Pb Zn

si

Cl

T

3.7

610.0

a

T

42

488.0

C3

T

6.0

153.0

S2

c

T

3.2

430.0

C3

T

IS

392.0

C*

T

4.4

592

S3

a

T

3.1

2950.0

C2

T

3.8

497.0

o

T

5.0

5i6

C4

T

7.0

55.2

11.7

1330.0

371.0

41600

S4

Cl

T

2.6

109.0

C2

T

2.4

88.6

a

T

2-8

297.0

C4

T

4.4

255.0

S5

Cl

T

3.6

1800.0

C2

T

3.7

1910.0

a

T

6.0

350.0

C4

T

12

56.2

St

Cl

T

3.7

1020.0

13.7

2410.0

1530.0

5000 0

a

T

34

2240.0

a

T

3.8

1830.0

an

T

3J

1510.0

S7

Cl

T

43 \ 418.0

a

T

3.7

544.0

S8

Cl

T

3.6

1860.0

C2

T

53

754.0

C3

T

422.0

123

4280.0

684.0

2710.0

C4

T

6.1

-88.6

S9 CI

T

4.0

1260.0

C2

T

62

922.0

C3

T

5.8

175.0

C4 T

63

155.0

S10

Cl T

12

1940.0

C2 T

1.1

39.7

sn

ci t

3.8

592.0

200

1220.0

13OO0

7550.0

C2

T

59

564.0

17.0

5150.0

3250.0

8220.0

S18

Cl

T

44 1 1210.0

15.8

1910.0

1550.0

3780.0

S20

a

T 5.1 1 176.0

2-5

3M.0

904

198.0

S21

Cl

T

4.7

522.0

7.4

43800

315.0

2320.0

S22

a

T

4.7

290.0

14.0

4940.0

1120

20300

S23

Cl

T

7.7

424.0

12.0

2290.0

356.0

24400

S25

a

T

43

2520.0

26.6

2930.0

2070.0

7170.0

Cl

T

3.6

2680.0

21.8

33300

2800.0

6890.0

S24

SURFACE

X

2£

0.0

86.5

65500.0

81.5

31200.0

S27

SURFACE

X

33

0.0

483

98500.0

149.0

22200.0

S28

a

T

U

2130.0

214

3460.0

4840.0

8350.0

a

T

88.0

21.6

6000.0

7350.0

16000.0

a

T

7.6

182.0

9.0

600.0

1380.0

2240X)

S29

SURFACE

X

3.9

0J0

120.0

73000.0

25.8

30900.0

Forty-six samples were analyzed for pH and arsenic (ppm); 15 samples were analyzed for total
cadmium, copper, lead, and zinc (ppm). Negative numbers may represent samples below detection; no
explanation was provided in the original. Shading indicates pH below 4.0, a criterion for injury
determination in soil.
Source: MultiTech, 1987b.

RCG/Hagler Bailly

6-17

Table 6-7

Tailings Samples Collected from Streamside Tailings along

Silver Bow Creek for the Tailings Investigation

Sample

Depth (in)

Borehole

PH

As

Cn

Pb

Zn

Surface Vegetation

SSI

0-2

1850-5

4.7

291

3006

630

3896

no vegetation

SS2

6-10

1850-5

5

674

1967

3785

4769

SS3

20- 28

1850-5

6.9

1460

999

9999

2984

SS4

28-30

1850-5

73

444

207

2912

4743

SS5

34

1850-5

5.6

0

0

30

1159

SS6

6-8

1850-6

6.8

847

4828

1604

7245

25% cover

SS7

0-9

1810-5

5.2

91

1010

687

1261

buried willows near

SS8

9-16

1810-5

4.9

169

4044

2782

4630

SS9

16-40

1810-5

3.2

1091

160

1604

5114

SS10

40

1810-5

3-2

555

136

837

1332

SS11

0-1

1750-1

4.1

4434

1287

2513

2521

sparte grass clumps

SS12

3-8

1750-1

53

322

2305

480

1879

SS13

8-16

1750-1

53

472

2500

347

1353

SS14

16-40

1750-1

6

0

0

187

1235

SS15

0-10

1750-2

4

353

370

1450

3561

do vegetation

SS16

10-24

1750-2

5.1

140

6884

1221

7929

SS17

24-30

1750-2

3^

223

60

170

1264

SS18

30-

1750-2

6.1

313

855

279

3430

SS19

0-2

1750-0

2.6

163

186

379

773

NR

SS20

0-2

1750-0

2.7

124

494

935

0

SS21

0-4

1640-5

Z7

157

0

370

0

oo vegetation

SS22

4-7

1640-5

3.2

110

579

378

100

SS23

7-17

1640-5

3.1

4621

106

0

533

SS24

0-2

1630-2

43

1638

1711

1407

2914

no vegetation

SS25

2-5

1630-2

4.1

704

437

1112

2326

SS26

5-24

1630-2

4

362

319

1029

2162

SS27

0-1

1630-3

23

77

1615

1276

0

dead grass clumps

SS28

0-3

1560-2

3.1

169

0

118

343

no vegetation

SS29

3-8

1560-2

3.1

123

0

498

0

SS30

8-12

1560-2

6.2

0

0

268

3350

SS31

0-2

1540-2

2.4

183

935

806

708

dead streambank vegetation

SS32

9-10

1540-2

4.6

212

9999

41

1380

SS33

10-18

1540-2

4.8

251

9999

0

1761

SS34

18-24

1540-2

62

0

0

271

7825

SS35

24-28

1540-2

5.7

0

77

2046

9999

SS36

0-1

1460-1

2.1

61

416

623

0

do vegetation

SS37

1-5

1460-1

2

108

401

454

0

SS38

5-10

1460-1

Z7

436

241

1501

182

SS39

10-23

1460-1

23

0

182

3858

0

SS40

23-31

1460-1

3.7

0

0

441

7958

SS41

0-2

1440-2

4.1

816

1416

2992

8884

do vegetatioD

SS42

2-6

1440-2

3.9

619

535

1126

2226

SS43

36

1440-2

73

0

0

85

0

SS44

23-28

1440-2

53

4

1449

105

2441

SS45

2-8

1370-2

4.1

1826

1808

4248

3698

no vegetation

SS46

25-30

1370-2

73

0

0

225

4016

SS47

44-50

1370-2

7.9

0

0

0

0

RCG/Hagler Bailly

6-18

Table 6-7

(cont.)

Tailings Samples Collected from Stream side Tailings

along

Silver Bow Creek for the

Tailings Investigation

Sample

Interval (in)

Borehole

PH

As

Cu

Pb

Zn

Surface Vegetation

SS48

0-7

1350-2

4.4

1522

2792

3022

5164

no vegetation

SS49

7-14

1350-2

4.8

245

1154

551

2116

SS50

14-22

1350-2

52

75

7551

28

2605

SS51

22-26

1350-2

4.9

29

3414

3414

2542

SS52

0-14

1330-3

3.9

945

4031

4031

999

no vegetation

SS53

14-17

1330-3

3.6

341

1033

1033

3035

SS54

35

1330-3

3

488

1385

1385

3189

SS55

41-45

1330-3

5.6

211

9540

9542

9999

SS56

0-1

1330-1

3.7

952

7004

7004

9158

no vegetation

SS57

2-5

1330-1

3.7

1755

4324

4324

5890

SS58

8-11

1330-1

3*

1057

4471

4471

2208

SS59

0-2

1300-2

3.7

536

2823

2823

9999

no vegeuiion

SS60

4-6

1300-2

3*

145

15

15

4546

SS61

18-24

1300-2

3.7

338

727

727

3599

SS62

50

1300-2

55

276

3343

3343

6690

SS63

75

1300-2

4.8

254

1639

1639

5003

SS64

84

1300-2

6.5

0

0

0

9999

SS65A

102

1300-2

6.7

0

0

0

441

SS65B

0-2

1280-2

3.6

1220

5354

5354

9594

no vegetation

SS66

5-8

1280-2

3.4

2114

3394

3394

7302

SS67

29-31

1280-2

4.6

1565

9999

9999

7491

SS68

41-44

1280-2

6

453

2292

2292

9999

SS69

59-63

1280-2

6.5

0

0

0

426

SS70

0-4

1260-3

3.7

1793

3773

3773

6819

willow in vicinity

SS71

8-10

1260-3

3.7

1044

2698

2698

4960

SS72

10-16

1260-3

3J

595

1527

1527

2957

SS73

18-24

1260-3

3.7

1550

5377

5377

8236

SS74

24-30

1260-3

S3

1123

9999

9999

9999

SS75

30-34

1260-3

5.9

633

9290

9290

9999

SS76

36-40

1260-3

5.9

665

1829

1829

9999

SS77

49-54

1260-3

6.5

191

1604

1604

9999

SS78

57-64

1260-3

5.9

56

0

0

1229

SS79

70-74

1260-3

5.7

0

0

0

91

SS6U

0-4

1220-3

3.9

1753

1646

1646

8631

next to willow

SS81

12-17

1220-3

5.7

941

9999

9999

9999

SS82

20-23

1220-3

6

876

8064

8064

9999

SS83

36-40

1220-3

7.5

0

585

585

5243

SS84

41-45

1220-3

12

0

0

0

0

SS85

46-50

1220-3

6.8

0

0

0

253

SS86

61-65

1220-3

15

0

0

0

232

SS87

0-9

1160-3

4.6

1404

5356

5356

5579

willows in vicinity

SS88

9-16

1160-3

8.1

249

2524

2524

9999

SS89

0-5

130-1

4.9

464

1786

1786

3427

willow in vicinity

SS90

5-6

130-1

4.7

556

2400

2400

2788

SS91

6-8

130-1

5

302

930

930

2449

SS92

8-15

130-1

53

1166

2497

2497

6402

SS93

15-20

130-1

6.4

68

4868

4868

6233

SS94

27-31

130-1

8

0

0

0

1353

RCG/Hagler Bailly

6-19

Table 6-7

(cont.)

Tailings Samples Collected from Stream side Tailings

along

Silver Bow Creek for the Tailings Investigation

Sample

Interval (in)

Borehole

pH

As

Cn

Pb

Zn

surface vegetation

SS95

0-5

170-1

4

123

9

9

573

dead willows

SS96

5-7

170-1

4

210

249

249

0

SS97

7-8

170-1

4

379

336

336

0

SS98

8-12

170-1

4

0

3611

3611

4253

SS99

12-22

170-1

62

4

408

408

1169

SS100

0-2

220-2

7

240

774

774

587

dead willows

SS101

2-3

220-2

4.1

613

1966

1966

0

SS102

3-19

220-2

4

187

1022

1022

1901

SS103

19-25

220-2

4j6

132

2974

1087

3719

SS104

25-30

220-2

5.6

35

2699

0

3480

SS105

0-5

320-2

5.5

422

369

1265

2321 grasset, dead willows in vicinity

SS106

5-16

320-2

5.1

294

0

253

444

SS107

16-22

320-2

53

317

3254

7738

3190

SS108

0-2

320-3

4.8

467

1896

2653

9999

dead willows

SS109

2-12

320-3

43

1419

1260

2424

6686

SS110

12-20

320-3

43

220

196

623

2096

SS111

25-29

320-3

43

383

5221

5485

2406

SS112

37-41

320-3

4.4

0

1647

0

1903

SS113

0-5

430-1

4.7

427

715

918

3180 dense willows in vicinity

SS114

5-9

430-1

4.4

73

903

555

1962

SS115

0-7

1080-1

3,6

198

1201

1346

1999

no vegetation

SSI 16

7-12

1080-1

3

399

0

685

3847

SS117

21-27

1080-1

3.2

289

0

296

758

SSI 18

0-1

1040-1

33

80

154

367

1159

no vegetation

SSI 19

1-11

1040-1

33

34

0

79

0

SS120

11-16

1040-1

2.8

42

0

88

0

SS121

0-10

510-1

43

324

1826

1708

3930

dead willows

SS122

10-14

510-1

43

344

1438

1662

3675

SS123

21-25

510-1

6.1

429

2777

1195

7059

SS124

34-38

510-1

6.2

358

3042

961

2758

Concentr

at ions reported are total metals determined by x-ray

luorescence (XRF) in ppm.

Values reported as 9999 ppm exceeded the instrument detection limit. Shaded pH values are

less than >

X.O and constitute

i an injury to soil.

Source:

PTI, 1989.

RCG/Hagler Bailly

6-20

Table 6-8

Data Collected from Three Bare Tailings Sites in the Clark Fork River

Floodplain for the Upper Clark Fork River Screening Study

Station | Depth (in)

Material

PH

As

Cn Pb Zn

She 10

175S-175E

0-2

TVS

6.9

420

2370

354

2670

175S-120E

03-1.5

OM

6.8

470

2230

338

1810

175S -175E

2-4

TVS

4.9

270

2650

280

2640

0N-60E

1-4

OM

5

550

5920

443

2360

60S-60W

0-1.5

T

3.9

320

1490

258

541

60S-60E

0-13

T

3.8

210

1500

217

813

60S - 175E

0-1 J

T

3.7

340

3830

304

1360

175N60E

0-1 J

T/S

73

560

3490

520

3160

175N - 60W

0-1 J

T/5

43

563

1370

377

695

420N - 75E

0-1.5

T/S

5

350

710

389

506

420N - 75E

0-13

T

5.4

1340

7720

991

2330

300N - 175W

12-16

T

63

440

1410

367

886

175N - 175W

0-1.5

T

IB

1310

2060

1030

433

60N-420E

0-1 J

T/S

5.7

500

3810

461

2070

120S - 175E

0-1 J

T/S

12

590

2930

485

2436

12OS-0E

9-11

T

33

1770

1950

637

1740

120S-60W

0-13

T/S

4.6

460

1020

320

400

ON -175W

0-1 J

T

4.8

640

2760

381

1030

60N -175W

0-1.5

T/S

5.9

570

1670

359

702

120N - 175W

0-13

NR

NR

320

6140

337

2920

Site 12

SOON - 100W

16-20

T

6

23.6

96

18

627

800N - 100E

0-13

T

4.7

780

7940

473

5540

700N - 100E

0-13

T

4.6

370

3050

310

1390

700N - 200E

0-1 3

T

4.9

280

4550

241

4630

700N - 300E

0-1 3

OM

7.6

370

1330

384

1440

6O0N-0E

0-1.5

T

5.4

740

2250

730

975

600N - 100W

0-13

T

43

480

2670

449

1040

400N-200W

0-13

T/S

7.1

610

8160

1170

2490

400N-OE

5-8

T/S

6.6

180

183

178

515

4O0N - 100E

0-1 5

T

4.9

370

4090

292

1860

5O0N-2O0E

0-1.5

T/S

5.8

330

1490

251

802

300N-200E

0-13

T

62

500

5600

457

2720

400N-200E

0-13

T

4.7

1100

2530

684

2350

300N - 100E

0-15

T

4.8

290

1970

245

1790

3O0N-0E

0-1 J

T/S

65

870

6730

1330

2670

200N - now

0-1 5

T

5.1

366

1820

390

827

2O0N-2O0E

0-1 J

T/S

53

320

2140

400

815

100N - 0E

0-1 5

T

4.7

320

2260

341

973

ON - 100E

2-3

T

53

115

87100

188

13300

100N - 190W

0-li

TIS

5.4

230

383

393

1360

RCG/Hagler Bailly

6-21

Table 6-8 (cont.)

Data Collected from Three Bare Tailings Sites in the Clark Fork River

Floodplain for the Upper Clark Fork River Screening Study

Station Depth (in) Material pH As Cu Pb Zn

Site 24

OS- 200W

0-2

T

*2

1620

2070

870

2270

OS- 100W

0-2

T/S

6

320

1630

241

1440

65S-0W

0-2

T

4

290

916

293

426

150S-0W

0-2

T

43

520

1320

417

1160

150S - OW

9-11

T

43

900

2980

1380

1780

150S-0W

11-13

T

5.7

38

6660

85

1270

200S- 100W

0-2

T

4

520

863

373

755

2O0S- 100W

10-13

T

33

1950

1010

840

968

100S-200W

0-2

S

73

120

1510

247

1120

1O0S-30OW

0-2

T/S

5.1

470

897

306

868

200S-300W

0-2

S

7.1

140

1020

161

1600

10OS-4O0W

0-4

T

4.6

110

2540

85

713

0S-4O0W

0-2

S

5.7

290

2010

266

834

0S-400W

7-9

S

7.7

63

10900

102

5400

0S-400W

11-14

S

8

15

138

16

168

OS-600W

5-10

S

33

920

1270

688

680

150S - 600W

0-2

T/S

73

160

887

165

1360

150S - 500W

12-14

T/S

33

1280

664

640

1710

150S - 500W

14-17

T/S

52

40

10200

58

2420

200S - 195W

0-3

T/S

3

1030

19400

540

8150

Concentrations reported are total metals (XRF) in ppm. Material types are designated as: T =
tailings; T/S = mixed tailings and soil; OM = organic matter; and NR = not recorded.
Shaded pH values are less than 4.0 and constitute an injury to soil.
Source: CH,M et al., 1991.

RCG/Hagler Bailly

6-22

Table 6-9
Screening Tests: Total Number of Germinating/Emergent Seedlings

(out of 20)

Sample

Alfalfa

Lettuce

Wheat

Riparian control

Riparian impact
R-04
R-08
R-13
R-14

Opportunity Ponds

20

0
2± 1
8±2

0

1 ± 1

9

0
0±0
1 ± 1

0

0± 1

19

0
4±3
16±2

0

3± 1

Note: Values are mean ± standard deviation of five replicates. Each replicate had 20 seeds.

Ten of the initial 20 poplars in 100% impact soil (five each from sites R-04 and R-14) wilted
and died within 24 hours of transplanting; two of the five poplars in control soils died As a
precaution against misinterpretation of transplanting stress versus toxicity, these three
treatments were repeated (five more poplars were planted in each of R-04 soil, R-14 soil and
control soil). Again, all 10 impact soil plants in the repeat trials died (all five in each of R-
04 and R-14 soils). All five plants in the repeat control soil tests remained normal in
appearance (Figures 6-5 and 6-6). No plants died during the remainder of the test in either
control soils or 50% mixed soils. All five plants in impact soil R-08 were dead at the
conclusion of the 4-week test.

Thus, mortality was 100% in three of the four impact soils (R-04,, R-042, R-08, R-14,, and
R-142) and 40% in the fourth (R-13). Roots of plants grown in the impact soils were poorly
developed, and in many instances, were black and brittle. Growth of the poplars in impact
samples was substantially less than in control soil. Using new shoot height growth as the
baseline, hybrid poplar growth inhibition was 72% in R-08 soil and 76% in R-13 soils, and
can be considered 100% (due to mortality) in the remaining impact soils. Shoot height, root
length, shoot mass, and root mass for all 100% site-soil impact samples (R-04, R-08, R-13,
and R-14) were significantly less than corresponding values determined for the control sample
(RR-07). Blending of impact soil with control soil (50% site-soil, 50% control soil) alleviated
much of the inhibition; assessment endpoints (shoot length and mass, and root length and
mass) of mixed soils were not statistically different from those of the control soils

RCG/Hagler Bailly

6-2:

In both screening tests (involving alfalfa, lettuce, and wheat) and extended tests (hybrid
poplars) in 100% impact soil, phytotoxic responses relative to controls were consistently
observed. This phytotoxic response is confirmation of injury to these soils.

6.3.2.3 STARS Phytotoxicity Tests

The only other greenhouse phytotoxicity study that has been conducted using Silver Bow
Creek soils reached similar conclusions. As a component of the Silver Bow Creek RI/FS, the
Streambank Tailings and Revegetation Study (STARS) was initiated to develop remedies for
in situ treatment of tailings deposited along Silver Bow Creek (MSU et al., 1989a). The
STARS project consisted of three phases; the first phase involved collection of bulk samples
from Silver Bow Creek floodplains, characterization of the soil material, and greenhouse
growth studies to identify suitable amendments to neutralize acidity and reduce contaminant
mobility. Thirty-five bulk samples were collected along Silver Bow Creek, all from
unvegetated tailings deposits. The soils were analyzed for chemical and physical parameters,
which included total, soluble, and extractable arsenic, cadmium, copper, lead, and zinc, and
pH. Total metals and pH results were presented in Table 6-5. The mean total concentrations
(ppm) determined were: arsenic — 694.35 (25 times baseline); cadmium — 22.78 (19 times
baseline); copper — 2,508.5 (73 times baseline), lead — 1,482.1 (41 times baseline), and zinc
— 3,782.5 (37 times baseline). Mean pH was 4.09.

Six of the 35 soils (indicated as TEST in Table 6-5) were selected for greenhouse growth
trials. Species chosen for the growth studies were selected based on tolerance to acid
conditions and high concentrations of metals (MSU et al., 1989a). In all repetitions, control
plants (plants grown in 100% impact soil, unamended with lime or other treatments) failed to
germinate or died shortly after germination. No statistical comparisons between plant
response in 100% impact soil and plant response in amended soil were made, and the controls
were not discussed further in the report (MSU et al., 1989a). The inability of even the most
tolerant plant species to survive in streamside tailings material is further confirmation of
injury to soils in the Silver Bow Creek floodplain.

6.3.2.4 Field Evidence

Gross evidence of phytotoxicity is prevalent throughout the floodplains of Silver Bow Creek,
and in patches along the Clark Fork River between its headwaters and Deer Lodge. The
Silver Bow Creek RI Groundwater and Tailings Investigation (MultiTech, 1987b) sampled
tailings and underlying soil in the Ramsay Flats area. The Ramsay Flats tailings deposit
covers approximately 160 acres, most of it completely devoid of vegetation, and has been
determined to be a source of hazardous substances to Silver Bow Creek (MultiTech, 1987b).
Total arsenic and pH were determined for 46 tailings samples from 18 boreholes (Table 6-6).
Total cadmium, copper, lead, and zinc were determined for 15 of the 46 tailings samples.

RCG/Hagler Bailly

VO

J=

♦-*

<4-

o

o

x>

4— '

u

DO

s.

o

C

—

QJ

J=

0)

c

o

>

c/i

*-<

o.

c

CO

O Q-

U 'c

r- ^

** c

£ c
c —

O
C/3

u

>

° £

*" c

•■ C_

O _=

[jj c

« —

o

i_ ^ <L>

K I/-. *i

•< _re *T

= ^

i o Si

Sf

re =fi

•c 2
>» a, 2

>
DC

9

5fl

r3

o
o

in

4

IU

0)

c/3 —
O ~

U 3

Ol

<N c

— O

< l

1)

o

>

S..9

e
c c

n —

£ g

u

C/5 "^

o E

o
c
x

c _

— J^

si

a

cc

a
u

6-26

Mean values for the analytes listed in Table 6-6 are as follows: pH — 4.78; arsenic — 806.72
ppm (29 times baseline); cadmium — 15.2 ppm (12 times baseline); copper — 2974.27 ppm
(87 times baseline); lead — 1866.56 ppm (52 times baseline); and zinc 5270.53 ppm (52
times baseline). Three samples of metal salts, which form as a crust on streamside tailings
deposits, were analyzed for metals. The following mean concentrations were determined:
cadmium — 84.93 ppm; copper — 79,000 ppm; lead — 85.4 ppm; and zinc — 28,100 ppm.
The extremely elevated concentrations in the readily soluble salts constitute a major pathway
of hazardous substance transport.

The Silver Bow Creek Tailings Investigation (PTI, 1989) was conducted to quantify the areal
and volumetric extent of contamination between the Colorado Tailings and the Warm Springs
Ponds. Concentrations of As, Cu, Pb, and Zn, determined using an x-ray fluorescence (XRF)
spectrometer, and pH for 124 samples from 31 boreholes are presented in Table 6-7. The
surface vegetation was noted in the field. One borehole was sampled "next to [a] willow";
the remainder were collected from areas devoid of vegetation or in the vicinity of willows.
No definition of "willow in vicinity" was recorded. Average total metals concentrations
determined in the 0-6 inch soil layer were: arsenic — 560 ppm (20 times baseline), copper —
1,550 ppm (45 times baseline); lead — 1,420 ppm (40 times baseline); and zinc — 3,110 (30
times baseline). Concentrations increased with depth, but the depth of greatest mean-
concentration varied by substance. Mean arsenic was greatest in the 6-12 inch interval (780
ppm); mean copper was greatest in the 18-24 and 24-36 inch intervals (3,540 and 3,660 ppm,
respectively); mean lead was greatest in the 18-24 and 36-48 inch intervals (3,460 and 3,630
ppm, respectively); and mean zinc was elevated relatively consistently in all depth intervals
below 6 inches (Table 6-7). Concentrations reported are substantially greater than baseline
(Table 6-4), but may not be truly representative because numerous samples were recorded at
the maximum detection limit of 9,999 ppm.

The Upper Clark Fork River Screening Study (CH2M Hill et al., 1991) was designed to
determine the magnitude of contamination at sites representative of the degree of
contamination present between Warm Springs and Drummond. A total of 12 sites were
sampled, sites were categorized as bare tailings, buried tailings, historically irrigated, and
good plant growth sites. Each site ranged from 5 to 10 acres, and was described in terms of
geomorphology, dominant soils, landform, aspect, degree of contamination, noticeable metals
effects on vegetation, proximity to the Clark Fork River channel, land use history, and
vegetation type and productivity. The data presented in Table 6-8 include concentrations of
As, Cu, Pb, and Zn determined in samples from the three bare tailings sites sampled along the
Clark Fork River (10, 12, and 24). Site 10 was 2.5 miles north of the Warm Springs Ponds;
Site 12 was 1 mile south of Deer Lodge; and Site 24 was approximately 2 miles southeast of
Garrison. Average concentrations in Sites 10, 12 and 24 were, respectively: arsenic — 600,
432, and 540 ppm; copper — 2852, 7317, and 3444 ppm; lead — 442, 446, and 389 ppm;
and zinc — 1575, 2406, and 1912 ppm. Mean values over the three sites exceeded baseline
as follows: arsenic, 19 times; copper, 133 times, lead, 12 times, and zinc, 19 times. Average

RCG/Hagler Bailly

6-27

pH for each site was 5.2, 5.5, and 5.2. Percent bare ground at each of the bare tailings sites
was estimated as 62%, 86.2%, and 38.6%.

6.3.2.5 Reduction of Soil pH

Soil pH of less than 4.0 resulting from the release of hazardous substances constitutes an
injury to the soil resource [43 CFR § 1 1.62 (e)(2)]. Tables 6-5 to 6-8 present soil pH
measured in unvegetated floodplain soil samples from Silver Bow Creek and the Clark Fork
River. Soil pH values less than 4.0 are indicated by shading.

Of the 35 samples collected for the STARS project, 22 (63%) had a pH of less than 4 0
(MSU et al., 1989a,b). Of the 31 boreholes sampled for the Silver Bow Creek Tailings
Investigation (PTI, 1989), 17 (55%) had pH values less than 4.0, and of the 29 boreholes and
surface samples analyzed for the SBC/RI Groundwater and Tailings Investigation (MultiTech,
1987b), 13 (62%) had a pH of less than 4.0.

Of the nine samples used in the NRDA phytotoxicity testing, four (44%) had a pH less than
4.0 (3.5-3.8), one sample had exactly 4.0, and the remaining four ranged between 4 4 and 6.2.
No control soils had pH values less than 4.0. All of these studies confirm injury to riparian
soils.

6.3.3 Injury Quantification

6.3.3.1 Areal Extent of Unvegetated Slickens

Slickens locations have been mapped through inspection of aerial photographs (MultiTech,
1987c; Hydrometrics, 1983; MultiTech, 1986, Nimick, 1990; U.S. EPA, 1992). Fluvially
deposited tailings and tailings waste piles dominate the 23-mile floodplain of Silver Bow
Creek (Canonie, 1992a). The Colorado Tailings cover approximately 40 unvegetated acres
(CH2M Hill and Chen-Northern, 1990), and the slickens in the Ramsay Flats vicinity cover
approximately 160 unvegetated acres (MultiTech, 1987b). In total, along Silver Bow Creek
from the upstream end of the Colorado Tailings to the Warm Springs Ponds, there are
approximately 1,270 acres of unvegetated fluvially deposited tailings (Hydrometrics, Inc.,
1983).

In addition to areas of unvegetated tailings deposits determined from maps and aerial
photographs, there are hundreds of acres in the Silver Bow Creek floodplain where
historically deposited tailings are partially covered with un-decomposed willows that died
when the tailings were deposited. MultiTech (1987a) estimated 1,133 acres of unvegetated
tailings deposits between the Metro Storm Drain confluence in Butte and the Kohrs Bridge
near Deer Lodge, and considered that estimate of impacted soils to be conservative because

RCG/Hagler Bailly

6-28

extensive stands of dead willow were grouped as living vegetation. Hydrometrics, Inc (1983)
estimated a total areal coverage of tailings from below the Colorado Tailings to the Warm
Springs Ponds of 1,270 acres. It is likely that the 1,270 acres identified by Hydrometrics
included some areas covered by dead, undecomposed willows, such as the extensive stands
between the Gregson Bridge and Opportunity. The inclusion of areas of dead willow stands
alone would dramatically increase the areal extent of measurable injury from surface tailings.

The total areal extent of floodplain contamination in the Clark Fork riparian corridor is not
well defined. The extent of injured soils is difficult to determine because injured soils may
be buried, and therefore would not exhibit visible signs of injury such as devegetation. In
addition, as Silver Bow Creek and the Clark Fork River meander periodically and erode parts
of the mining terrace of tailings sediments, new mixtures of eroded tailings, original bank
material, and channel sediment are created and deposited, creating a changing mosaic of
slickens (Nimick, 1990). Currently, estimates of extent of contamination are bounded by
limits of studies rather than actual limits of contamination, however, buried and mixed
tailings deposits enriched with hazardous substances by as much as four to five orders of
magnitude above baseline do occur along the Clark Fork River (Nimick, 1990). Soils
containing concentrations of that magnitude do not support vegetation (see Section 6.3.2.2).

In the upper 10 kilometers (6.2 miles) of the Clark Fork River, there are approximately 98.5
acres of unvegetated fluvially deposited tailings (MultiTech, 1987c). Nimick (1990) mapped
tailings distribution, including buried tailings, in the upper 10 kilometers (6.2 miles) of the
Clark Fork River, and identified 678 acres. Nimick's estimate includes vegetated mixed
tailings/alluvium, re-worked tailings, and buried tailings, thus the discrepancy between the
MultiTech and Nimick estimates. Below Garrison, slickens deposits exist but constitute a
lesser percentage of the floodplain area. No reliable maps of slickens locations exist for the
section of river below Deer Lodge to Milltown.

The width of the flood-contaminated land in the upper 10 kilometers (6.2 miles) of the Clark
Fork River is generally between 180 and 490 meters (590-1,600 feet), but ranges between 90
and 900 meters (295-2,950 feet) (Nimick, 1990). The thickest tailings deposits sampled are
either near the river or near the course of the late 1800s channel (Nimick, 1990). The band
of thick deposits usually lies within the meander belt of the current channel. Tailings 10 to
30 centimeters (4-12 inches) thick are extensive in areas where the floodplain was wide.

6.3.3.2 Volume of Contaminated Material

The total volume of tailings and soil material in Lower Area One that is enriched with
hazardous substances is approximately 2.2 million cubic yards (0.6 million in the Colorado
Tailings area, and 1.6 million cubic yards of wastes associated with the former Butte
Reduction Works site) (Camp Dresser and McKee, 1991). Large quantities of fill material
have been deposited in the upper Metro Storm Drain area covering extensive, previously

RCG/Hagler Bailly

6-29

exposed tailings deposits generated by the Parrott Smelter. As much as 20 feet of fill
material overlies tailings deposits as thick as 14 feet in the upper Metro Storm Drain area
(CH2M Hill and Chen-Northern, 1990). Estimated volumes of 190,000 cubic yards of tailings
and mixed alluvium and tailings, 300,000 cubic yards of slag, slag-sand, and gravel, 525,000
cubic yards of waste rock, and 840,000 cubic yards of fill material overly native sand and
gravel in the upper Metro Storm Drain area (CH2M Hill and Chen-Northern, 1990).

Native silts and clays in the lower Metro Storm Drain area are covered by mixed alluvium
and tailings averaging approximately 2 feet deep and landfill material ranging from 4 to 1 1
feet deep. The area south of the drain area was apparently used as a landfill/dump, on top of
mixed tailings and alluvium. Approximately 0.2 million cubic yards of tailings and mixed
tailings and alluvium, and 0.57 million cubic yards of demolition debris and landfill debris,
cover nauve silts and clays in the lower Metro Storm Drain area (CH2M Hill and Chen-
Northern, 1990).

Between Montana Street and the Colorado Tailings, slag walls presumably built to contain
tailings waste generated by the Butte Reduction Works were constructed on the Silver Bow
Creek floodplain. Materials now overlying the former floodplain consist of approximately
0.43 million cubic yards of tailings and mixed tailings and alluvium, and 1.63 million cubic
yards of various types of waste, including manganese flue dust, railroad bed fill, and
transported fill (CH2M Hill and Chen-Northern, 1990). Source materials in the Butte
Reduction Works area extend to a depth of 10 to 15 feet, and contamination has been
detected to a depth of 2 feet in overlain nauve soils (Camp Dresser and McKee, 1991).

The Colorado Tailings resulted from the deposition of tailings and mine and mill waste from
the Colorado Smelter. The extensive tailings deposit, approximately 40 acres, consists of
relatively continuous tailings material, up to 4.5 feet deep, and additional transported fill
material. The approximate thickness of all material units overlying native sediment is 18 feet.
Approximately 230,000 cubic yards of tailings and mixed tailings and alluvium, and 580,000
cubic yards of additional fill material, are present in the deposit (CH2M Hill and Chen-
Northern, 1990). In the Colorado Tailings, source areas of hazardous substances include the
tailings and the underlying peat layer.

The Rocker Operable Unit of the Silver Bow Creek/Butte Area NPL site is approximately 7
miles west of Butte and is bordered to the north by Silver Bow Creek (Keystone, 1992). The
former Rocker Timber Framing and Treating Plant, which treated mine timbers with a
preservative containing arsenic, was operated by the Anaconda Company until 1957, when
operations ceased. Extensive quantities of waste material from the treatment plant were
dumped on the banks of Silver Bow Creek; in 1989, approximately 1,021 cubic yards of
arsenic-contaminated wood chips and soils were removed from areas in the site where arsenic
concentrations exceeded 10,000 ppm (Keystone, 1992).

RCG/Hagler Bailly

6-30

Extensive deposits of mill tailings, mine waste material, and precipitates, sometimes mixed
with native sediments, occur within the entire floodplain from Butte to the Warm Springs
Ponds. Within the Streamside Tailings Operable Unit, which extends from downstream of
Colorado Tailings to the Warm Springs Ponds and excludes the Rocker area, the estimated
volume of tailings and tailings-impacted material is 3.7 to 7.8 million cubic yards (Canonie,
1992a).

Between the downstream end of the Colorado Tailings and the upstream end of the Durant
Canyon, approximately 1.7 to 4.1 million cubic yards of tailings and mixed tailings and
alluvium overlie native fluvially deposited sands, silty sands, and gravels. Among the most
prominent features of this reach of river are the 160-acre tailings deposit at Ramsay Flats and
extensive tailings deposits downstream of Miles Crossing. Downstream of Miles Crossing to
Finlen, Silver Bow Creek cuts through a volcanic canyon. In the canyon, approximately
0.733 million to 0.965 million cubic yards of tailings and mixed tailings and alluvium overlie
native sands, silts, and clay. From Finlen to the Warm Springs Ponds, estimates of the
amount of overlying tailings and mixed tailings and alluvium range from 1.27 to 2.8 million
cubic yards (Canonie, 1992b).

Downstream of Warm Springs Ponds, visible deposits are sporadic, occurring along inside
bends of the Clark Fork River (MultiTech, 1987c). Stratigraphic evidence of the Clark Fork
River floodplains indicates that sediments consisting of essentially pure tailings were
deposited on the pre-mining floodplain to a depth of 1 to 2 meters above the current channel
elevation (Nimick, 1990). Numerous studies have identified extensive deposits of metal-
enriched sediment deposits on the banks and floodplains of the Clark Fork River (Moore,
1985, Hydrometrics, Inc., 1983). The high metal concentrations, physical attributes, and
evidence of recent sedimentation indicate that the sediments contain mill tailings along with
variable quantities of mine waste rock, flocculated metals, and natural sediment

In the floodplain of the Clark Fork River, an estimated 2 million cubic meters (2.6 million
cubic yards) of contaminated sediment have been deposited (Moore and Luoma, 1990;
Axtmann and Luoma, 1991). Nimick (1990) identified flood-deposited tailings up to 120 cm
thick overlying the pre-mining floodplain. He estimated 0.92 million cubic yards (0.704
million cubic meters) of tailings cover 678 acres (275 hectares). The highest metal
concentrations are no longer in near-surface tailings because of downward migration of metals
(Nimick, 1990). Studies that have quantified extent of contamination of floodplain soils by
visual examination alone have grossly underestimated the true extent.

In the Opportunity Ponds, there are an estimated 435 million cubic yards (333 million cubic
meters) of wastes (Tetra Tech, 1987). Borehole investigation indicate that the average depth
of tailings in the Ponds ranges from 14.3 to 35.0 feet (Tetra Tech, 1987).

RCG/Hagler Bailly

6-3J_

6.3.4 Ability of Resource to Recover

Natural recovery of the riparian soils and floodplain sediments of Silver Bow Creek and the
Clark Fork River will require many hundreds, if not thousands of years, for the following
reasons:

► The total volume of tailings and soil material in Lower Area One that is
enriched with hazardous substances is approximately 2.2 million cubic yards
(0.6 million in the Colorado Tailings area, and 1.6 million cubic yards of
wastes associated with the former Butte Reduction Works site) (Camp Dresser
and McKee, 1991). Within the approximately 23 miles of the Streamside
Tailings Operable Unit (downstream of the Colorado Tailings to the Warm
Springs Ponds), the estimated volume of tailings and impacted soils is 3.7 to
7.8 million cubic yards (Canonie, 1992a). In the floodplain of the Clark Fork
River, an estimated 2 million cubic meters (2.6 million cubic yards) of
contaminated sediment have been deposited (Moore and Luoma, 1 990;
Axtmann and Luoma, 1991). In the Opportunity Ponds, there are an estimated
435 million cubic yards (333 million cubic meters) of wastes (Tetra Tech,
1987). The sheer volume of hazardous materials insures that biological,
surface water, and groundwater resources will be exposed to hazardous
substances for many years to come.

► The persistence of contamination in soils, particularly by heavy metals, is
virtually permanent (Kabata-Pendias and Pendias, 1992). Metals accumulated
in soils are depleted slowly by leaching, plant uptake, erosion, or deflation.
Estimations of removal rates indicate that the complete removal of metallic
contaminants from soils is nearly impossible (Kabata-Pendias and Pendias,
1992).

► Studies of downstream metal trends in bank sediments (Moore, 1985; Moore et
al., 1989, Axtmann and Luoma, 1987) found considerable variability in metals
concentrations and little evidence of downstream decreases in metals levels,
particularly between Warm Springs and Garrison (Nimiclc, 1990). Data
collected during the 1992 NRDA field sampling (Figure 6-7) identify similar
variability and absence of trends of decreasing metals and arsenic
concentrations downstream. Therefore, dilution of slickens and floodplain
sediments is not occurring at a rate sufficient to anticipate recovery of Silver
Bow Creek and the Clark Fork River riparian zones.

RCG/Hagler Bailly

6-32

1000

E

Q.

a

c
o

c

0)

u

c
o
o

800

600-

400-

200

123456789 1011121314151617

Upper SVvw Bow Out low* S«y«. Bow Cint Uppw Q>rk Fath Hract

Sample Site Number

6000

5000-

E
a

c
o

c
a>
o
c
o
O

1 2 S 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Upp«i ttoar Sow CiMk LP— Mnr Bow CrMk Upow Curt Fork Mow

Sample Site Number

Figure 6-7. Concentrations of Each Element at Sample Points Collected Within Each
Impact Reach. Sample point concentrations are plotted in a downstream
direction; the axis is not scaled to show linear distance. The figures illustrate
the absence of trend in concentration of hazardous substances with distance

RCG/Hagler Bailly

6-33

6.4 REFERENCES

AMC. Date unknown. Tailings Ponds. Section of Anaconda Minerals Co. document (1970s)
that describes tailings ponds and related area used by the Company.

ASTM. 1994. Practice for Conducting Early Seedling Growth Test, E1598.

Axtmann, E.V. and S.N. Luoma. 1987. Trace metal distributions in floodplain and fine
grained bed sediments of the Clark Fork River, Montana: Proceedings, 6th Annual Conference
on Heavy Metals in the Environment. New Orleans, Louisiana. September, 1987. CEP
Consultants, Edinburgh, pp. 494-496.

Axtmann, E.V. and S.N. Luoma. 1991. Large-scale distribution of metal contamination in
the fine-grained sediments of the Clark Fork River, Montana, U.S.A. Appl. Geochem.
6:75-88.

Camp Dresser and McKee. 1991. Statement of Work: Remedial Investigation/Feasibility
Study, Priority Soils Operable Unit - Silver Bow Creek/Butte Area NPL Site, Butte, MT
(Final Draft). Prepared by CDM Federal Programs Corp. for the U.S. EPA - Region VTH,
Helena, MT.

Canonie. 1992a. Preliminary Site Characterization Information Report: Silver Bow
Creek/Butte Area NPL Site, Streamside Tailings Operable Unit RI/FS. Prepared for ARCO
by Canonie Environmental Service Corp., Bozeman, MT.

Canonie. 1992b. 1991 Remedial Investigation Activities, Data Summary Report: Silver Bow
Creek/Butte Area NPL Site, Streamside Tailings Operable Unit RI/FS, Appendix D. Prepared
for ARCO by Canonie Environmental Service Corp., Bozeman, MT.

CH2M Hill and Chen-Northern. 1990. Silver Bow Creek CERCLA Phase II Remedial
Investigation Data Summary, Area I Operable Unit - Volume I and Volume II (Draft Final).

CH2M Hill, Chen-Northern, and Montana State University Reclamation Research Unit. 1991.
Draft Final: Upper Clark Fork River Screening Study. Volume 1. Prepared for Montana
Department of Health and Environmental Sciences.

Hassett, J.J. 1992. Soils and Their Environment. Prentice Hall, Englewood Cliffs, New
Jersey. 424 pp.

Hydrometrics, Inc. 1983. Summit and Deer Lodge Valleys long term environmental
rehabilitation study, Butte-Anaconda, Montana. Prepared for the Anaconda Minerals
Company.

RCG/Hagler Bailly

6-34

Kabata-Pendias, A. and H. Pendias. 1992. Trace Elements in Soils and Plants. 2nd ed.
CRC Press, Ann Arbor, MI.

Keystone. 1992. Preliminary Site Characterization Information Report, Rocker Timber
Framing and Treating Plant Operable unit, Rocker, Montana. Prepared for ARCO by
Keystone Environmental Resources, Inc. Monroeville, PA. July 1992.

Lipton, J., H. Bergman, D. Chapman, T. Hillman, M. Kerr, J. Moore, and D. Woodward.
1995. Aquatic Resources Injury Assessment Report, Upper Clark Fork River Basin. Prepared
by RCG/Hagler Bailly for the State of Montana, Natural Resource Damage Litigation
Program. January.

Manly, B.F.J. 1991. Randomization and Monte Carlo Methods in Biology. New York:
Chapman and Hall. 281 pp.

Moore, J.N. 1985. Source of Metal Contamination in Milltown Reservoir, Montana: An
Interpretation Based on Clark Fork River Bank Sediment. Prepared for U.S. EPA.
September. 60 pp.

Moore, J.N. and S.N. Luoma. 1990. Hazardous wastes from large-scale metal extraction. A
case study. Environ. Sci. Technol. 24(9): 1278-1285.

Moore, J.N., E.J. Brook, and C. Johns. 1989. Grain size partitioning in contaminated, coarse
grained river floodplain sediment, Clark Fork River, Montana, U.S.A. Environ. Geol. Water
Sci. 14: 107-115.

MSU/RRU, Schafer and Associates, and CH2M Hill. 1989a. Silver Bow Creek RI/FS
Streambank Tailings and Revegetation Studies: STARS Phase I Bench-Scale Soil Column and
Greenhouse Treatability Studies, and Tailings Ranking System. Final Summary Report:
Volume I. Prepared for Montana Department of Health and Environmental Sciences.

MSU/RRU, Schafer and Associates, and CH2M Hill. 1989b. Silver Bow Creek RI/FS
Streambank Tailings and Revegetation Studies: STARS Phase I Bench-Scale Soil Column and
Greenhouse Treatability Studies, and Tailings Ranking System. Final Summary Report:
Volume U. Prepared for Montana Department of Health and Environmental Sciences.

MultiTech. 1983. Anaconda Smelter Soils/Cropland Sampling and Analysis Study Plan.
Review Draft. Prepared for U.S. EPA, Helena, Montana, and Ecology and Environment, Inc.

MultiTech. 1986. Riparian vegetation of Silver Bow Creek and the upper Clark Fork River.
Prepared for Montana Department of Health and Environmental Sciences, Solid and
Hazardous Waste Bureau.

RCG/Hagler Bailly

6-35

MultiTech. 1987a. Silver Bow Creek Remedial Investigation Final Report Summary.
Prepared for Montana Department of Health and Environmental Sciences, Helena, MT.

MultiTech. 1987b. Silver Bow Creek Remedial Investigation Final Report Appendix B, Part

1, Report: Groundwater and Tailings Investigation and Part 4, Attachment VLChemical Data.
Prepared for Montana Department of Health and Environmental Sciences, Helena, MT.

MultiTech. 1987c. Silver Bow Creek Remedial Investigation Final Report Appendix D, Part

2, Report: Vegetation Mapping. Prepared for Montana Department of Health and
Environmental Sciences, Helena, MT.

Nimick, D.A. 1990. Stratigraphy and Chemistry of Metal-Contaminated Floodplain
Sediments, Upper Clark Fork River valley, Montana. M.S. Thesis, University of Montana,
Missoula, MT. 118 pp.

PTI. 1989. Silver Bow Creek Tailings Investigation, Draft Report. Prepared by PTI
Environmental Services for Parcel, Mauro, Hultin & Spaanstra, Denver, CO. 23 pp.

Rice, P. M. and G. J. Ray. 1984. Floral and Faunal Survey and Toxic Metal Contamination
Study of the Grant-Kohrs National Historic Site. Project for the Gordon Environmental
Studies Laboratory.

Tetra Tech. 1987. Anaconda Smelter Remedial Investigation/Feasibility Study: Master
Investigation Draft Remedial Investigation Report. Prepared for Anaconda Minerals
Company.

U.S. EPA. 1992. Enforcement /Action Memorandum Lower Area One Operable Unit of the
Silver Bow Creek/Butte Area (original portion) Superfund Site, Butte, MT.

RCG/Hagler Bailly

7-1

7.0 BIOLOGICAL RESOURCES — RIPARIAN VEGETATION,
WILDLIFE, AND WILDLIFE HABITAT

Riparian biological resources include vegetation, wildlife, and wildlife habitat. This chapter
describes and quantifies injuries to these resources in the riparian corridor along Silver Bow
Creek, from the Colorado Tailings to the Warm Springs Ponds, along the upper Clark Fork
River, from the Warm Springs Ponds to Deer Lodge, and in the Opportunity Ponds area.
Overall, the results presented in the chapter demonstrate that vegetation, wildlife, and wildlife
habitat have been injured in riparian assessment areas. Specifically:

► Riparian slickens areas along Silver Bow Creek and the upper Clark Fork River
— shown to be phytotoxic in Chapter 6.0 — are virtually devoid of vegetation.
Matching control sites contain a mixture of riparian forest/shrub communities
and agricultural land uses.

►• The number of habitat layers has been significantly reduced in impacted areas.

► Habitat has been significantly reduced for white-tailed deer (an indicator
species for riparian forest/shrub ecosystems). The viability of wildlife species
that depend on this habitat type has been reduced.

► As a result of habitat loss, a large number of wildlife species are likely to have
been lost to the injured areas or have undergone reduced population viability.

7.1 DESCRIPTION OF RIPARIAN BIOLOGICAL RESOURCES

Riparian biological resources include terrestrial wildlife populations and the terrestrial wildlife
habitat provided by vegetation communities associated with Silver Bow Creek, the upper
Clark Fork River, and the Opportunity Ponds. Riparian vegetation in southwest Montana
generally comprises associations of species adapted to hydric or semi-hydric conditions and
regular disturbance, particularly cottonwoods, willows, rushes, and sedges (Hansen et al.,
1989). In southwest Montana, many wildlife species are characteristic of riparian areas and
are dependent on the existence of riparian vegetation communities.

The riparian soils of Silver Bow Creek and the upper Clark Fork River from the Warm
Springs Ponds to Deer Lodge, and the Opportunity Ponds areas have been injured by releases
of hazardous substances (see Chapter 6.0). The preceding chapters have traced the pathways
of hazardous substances from the sources in Butte to surface water, sediments, aquatic biota,
and floodplain soils of Silver Bow Creek and the Clark Fork River. The banks and
floodplains, including native riparian soils, have been injured by the deposition of tailings
material transported in surface water (Chapter 6.0). Floodplain soils contain elevated
concentrations of hazardous substances relative to baseline, and are severely phytotoxic.

RCG/Hagler Bailly

7-2

Vegetation establishment and growth are precluded along much of Silver Bow Creek and
along the upper Clark Fork River by the phytotoxic condition of the soils; wildlife
populations dependent on the habitat provided by riparian vegetation have been concomitantly
reduced or lost within the injured areas.

The Opportunity Ponds area was formerly a low-lying area that most likely supported wetland
communities associated with Mill and Willow Creeks, and other minor drainages and springs.
Current hydrological and geomorphological patterns, historical anecdotes, and the rationale for
locating the settling ponds where they are suggest that the Opportunity Ponds area was
naturally wet and, in the absence of mining-related impacts, would probably have supported
diverse riparian and possibly mixed-meadow vegetation community types. Currently, the
Opportunity Ponds area is devoid of vegetation and supports no wildlife habitat.

7.2 INJURY DEFINITION

Injury to riparian biological resources is defined as:

► ... adverse changes in viability: death, . . . physiological malfunctions
(including malfunctions in reproduction), or physiological deformations [43
CFR§ 11.62(f)(1)].

In the Clark Fork Basin, injury to vegetation that resulted from releases of hazardous
substances is expressed in the complete eradication of vegetation, or changes in the
composition, structure, and/or distribution of vegetation communities. At the level of the
individual plant, these community changes have been caused by death and physical
deformations (i.e., reduced growth leading to a loss in viability).

As described in Chapter 4.0, plant death and reduced growth satisfy the four acceptance
criteria for biological responses [43 CFR § 11.62 (f) (2) (i-iv)].

Overall, injury to vegetation has been confirmed by the results of the phytotoxicity studies
(death, reduced growth relative to controls), and by the observed loss of vegetation in exposed
areas (see Section 7.3).

The viability of riparian wildlife populations in the upper Clark Fork Basin has been reduced

by:

► Reductions in habitat quantity and quality for selected indicator species [43
CFR § 11.63 (f)(4)(ii)(A)] relative to uncontaminated control areas [43 CFR §
11.72(d)(1)].

RCG/Hagler Bailly

7-3

The following sections describe the relationship between vegetation and wildlife habitat, and
methods of assessment of injury to wildlife habitat (described in detail in Chapter 4.0).

7.2.1 Methods of Measuring Injury to Wildlife Habitat Quantity and Quality

Injury to wildlife and wildlife habitat in the impacted riparian areas was determined using a
combination of wildlife habitat models, surveys of selected organisms, and field investigations
of vegetation.

7.2.1.1 HEP Models

Assessment of injury to vegetation, as a supporting means of quantifying the extent of both
soil injury and wildlife habitat injury, is discussed in Chapter 4.0. HEP models were used to
assess injury to riparian wildlife habitat along Silver Bow Creek, along the upper Clark Fork
River, and in the Opportunity Ponds.

The selection of HEP models for use in this injury assessment was based on the types of
habitats initially identified as potentially injured. Initial field observations and studies
performed by the Montana Riparian Association (MRA, 1992) indicated that riparian plant
associations impacted by the deposition of hazardous substances include riparian shrub and
riparian forest communities. Two HEP models were selected to determine injury to these
riparian wildlife habitats.

Riparian Forest/Shrub Habitat: Indicator Species = White-Tailed Deer

Determination of injury to riparian forest and shrub habitat types was performed using white-
tailed deer as an indicator species. An existing HEP model for assessing white-tailed deer
winter habitat (Short, 1986) was modified for use in Montana by incorporating northern
Rocky Mountain regional browse species as a nutritional parameter.

Variables measured to evaluate winter habitat quality for white-tailed deer included:

► Water availability. Free water must be available within 1.6 km (1 mile) of the

habitat block being evaluated.

►• Habitat block area. The area of evaluation must be at least 40 hectares (100

acres).

►

Cover availability. Overstory or midstory cover must provide protective or
thermal cover on at least 20% of the area being evaluated.

RCG/Hagler Bailly

7-4

► Forage availability. A sample site is categorized by whether or not it provides

suitable white-tailed deer forage.

Since the habitat evaluated was riparian, water was close to all impact and control areas, and
thus the first parameter was satisfied for all sample sites. Likewise, all areas evaluated
exceeded the 40 ha (100 acre) area requirement. Habitat suitability indices were calculated
for impacted areas along Silver Bow Creek, the upper Clark Fork River, and in the
Opportunity Ponds and statistically compared to baseline habitat values determined for control
sites (Divide Creek, Little Blackfoot River, Flint Creek).

Habitat Layers HEP Model

In addition to the white-tailed deer model, the layers of habitat (Short, 1984) HEP model was
used to address the relationship between the vertical complexity of vegetation communities
and their capacity to provide habitat for a diversity of wildlife. The layers of habitat model is
described in detail in Chapter 4.0.

7.2.1.2 Wildlife Population Surveys

Riparian avian populations were surveyed to determine whether differences exist in the
population densities on Silver Bow Creek relative to a control stream. Two surveys of
riparian bird species characteristic of southwestern Montana (common merganser, belted
kingfisher, great blue heron, dipper, and spotted sandpiper), chosen because of their
conspicuousness and the relative ease with which they can be counted, were conducted on
Silver Bow Creek between the downstream end of Durant Canyon and Warm Springs Ponds
in late June 1992, and on a paired control reach on the Little Blackfoot River. On each
survey, an observer experienced in surveying riparian bird populations walked the study reach
of the river and recorded the numbers of individual birds encountered.

7.2.1.3 Vegetation Assessment

The floodplain soils and sediments of Silver Bow Creek, the upper Clark Fork River, and the
Opportunity Ponds have been shown to be phytotoxic (Appendix B and Chapter 6.0); a
phytotoxic response is reflected in the vegetation community as a change in the plant
population density, species composition, dispersion, or percent cover [43 CFR § 11.71 (1)(6)].

Injury to vegetation communities was assessed by comparing the proportional representation
of native cover types (e.g., deciduous forest and deciduous shrubland) and agricultural cover
types (e.g., hayfields, pasture), proportional representation of numbers and types of habitat
layers (e.g., tree canopy, tree bole, shrub layer, understory, soil), and percent cover of
individual forage species and bare ground or slickens in impact and control areas. Statistical

RCG/Hagler Bailly

7-5

comparisons were made between impact and control areas to identify and quantify reductions
in vegetation abundance and compositional quality (Appendix C).

7.2.2 Baseline Conditions

No historical records describe the composition or structure of the wildlife habitat provided by
the riparian vegetation communities associated with Silver Bow Creek, the Clark Fork River,
and the Opportunity Ponds before mining-related impacts. However, baseline conditions have
been determined from information contained in USGS maps, from aerial photographs, and by
measurements of vegetation communities and wildlife populations made on control rivers.
These procedures are in accordance with the methods recommended for establishing baseline
conditions in the DOI regulations [43 CFR § 11.72 (d)].

Divide Creek, the Little Blackfoot River, and Flint Creek were selected as controls for the
injured riparian reaches of Silver Bow Creek, Opportunity Ponds and the upper Clark Fork
River because they are comparable in the important topographical, hydrogeological, and
climatic factors that influence the distribution and composition of plant communities (see
Appendix A). The control reaches are not undisturbed by human activities because
agriculture is a major land use on all three.

Vegetation measurements made on the control reaches showed that baseline for the injured
reaches of Silver Bow Creek and the upper Clark Fork River and for Opportunity Ponds
consists of a mixture of natural and agricultural plant communities. Baseline natural plant
communities include (1) riparian forest, with an overstory of tall cottonwoods, a shrub
midstory of willow and dogwoods, and an understory rich in grasses and herbs; and (2)
riparian shrub communities that resemble the riparian forest but lack an arboreal layer
(Appendix C). Baseline agricultural plant communities comprise pasture grazed by livestock
and hay meadows (Appendix C). Although these communities lack an overstory of shrubs or
forest trees, they support a diverse understory of grasses and herbs.

These baseline riparian plant communities provide many ecological niches for wildlife
species. In general, the more complex the horizontal and vertical structure of the habitat, the
more potential niches in that habitat. Baseline riparian forest is a complex habitat.
Measurements made on the control streams showed that riparian forest displayed up to five
distinct vertical habitat layers (tree canopy, tree bole, shrub midstory, herbaceous layer, and
soil; Appendix C). Furthermore, because the riparian zone is a linear habitat near other
habitat types (e.g., agricultural land and/or upland community types), it also has horizontal
diversity in the form of extensive "edge" habitat.

Riparian forests of Montana are important islands of biodiversity; despite their relatively small
percentage of the landscape (approximately 1%), riparian zones in Montana provide habitat
for many mammalian and avian species (Ohmart and Anderson, 1986). Eighty-nine percent

RCG/Hagler Bailly

7-6

of the terrestrial bird species in Montana exploit riparian habitats during breeding season, and
36% breed only in riparian areas (Mosconi and Hutto, 1982). At least 38 bird and mammal
species are characteristic of southwestern Montana riparian habitats, including waterfowl,
songbirds, birds of prey, white-tailed deer, fox, and mink (Tables 7-1 and 7-2).

7.3 INJURY DETERMINATION AND QUANTIFICATION

This section confirms and quantifies reductions in habitat quantity and quality (suitability) in
riparian assessment areas.

7.3.1 NRDA Investigations

Reductions in habitat quality and quantity were assessed within areas delineated as grossly
injured based on aerial photographs, existing maps of tailings deposits, and preliminary field
observations. Grossly injured areas, or slickens, were identified and delineated using the
following criteria:

► Complete or virtual elimination of the indigenous major plant associations

► Little or no regeneration of the indigenous major plant association.

Riparian areas expected to support riparian vegetation and wildlife habitat in the absence of
mining-related impacts were identified and sampled on the following stream reaches:

► Silver Bow Creek from downstream of the Colorado Tailings in Butte to the
upstream end of the Durant Canyon (upper Silver Bow Creek)

► Silver Bow Creek from the downstream end of the Durant Canyon to the Warm
Springs Ponds (lower Silver Bow Creek)

►• The Clark Fork River from the Warm Springs Ponds discharge to Deer Lodge

The Opportunity Ponds were also sampled as an impacted riparian area. The Durant Canyon
of Silver Bow Creek was not sampled because of time constraints; however, since the Durant
Canyon lies between the upper and lower Silver Bow Creek reaches sampled, pathways and
sources to the canyon are identical, existing slickens composition are expected to be identical,
and thus injury to vegetation and wildlife habitat is mechanistically similar to those found
upstream and downstream.

Measurements of the following were recorded along transects at 17 sites within slickens along
Silver Bow Creek and the upper Clark Fork River (Figure 7-1):

RCG/Hagler Bailly

7-7

Table 7-1

Birds Characteristic of Southwest Montana Riparian Shrub/Forest

and their Feeding and Nesting Vegetation Layers

Species

Feeding Layers

Nesting Layers

Great blue heron

Ardea herodias

-

TC

Mallard

Anas platyrhynchos

-

US

Common merganser
Wood duck

Mergus merganser
Aix sponsa

-

TB/US
TB

Osprey

Pandion hahaetus

-

TC

Bald eagle
American kestrel

Haliaeetus leucocephalus
Falco sparverius

US

TC
TC

Great-horned owl

Bubo virginianus

US/SM

TC/TB

Belted kingfisher
Northern flicker

Ceryle alcyon
Colaptes auratus

TS/US

TS
TB

Mourning dove

Zenaida macroura

US

TC/SM

Western wood-pewee

Contopus sordidulus

TC

TC

Willow flycatcher
Eastern kingbird
Tree swallow

Empidonax traillii
Tyr annus tyr annus
Tachycineta bicolor

SM
TC

SM
TC
TB

Bank swallow
Black-billed magpie

Riparia npana
Pica pica

US

TS
TC/SM

Black-capped chickadee

Parus atricapillus

TC/US

TB

House wTen

Troglodytes aedon

SM/US

TB

American robin

Turdus migratorius

US

TC/SM

Veery

Caiharus fusee sc ens

SM

SM

European starling

Sturnus vulgaris

US

TB

Warbling vireo

Vireo gilvus

SM

SM

Yellow warbler

Dendrowa petechia

SM

SM

American redstart

Setophaga ruttcilla

SM/TC

SM/TC

Common yellowthroat

Geothlypis trichas

SM/US

SM/US

Song sparrow
Red-winged blackbird

Melospiza melodia
Agelaius phoeniceus

SM/US
SM/US

SM/US
SM

Brewer's blackbird
Black-headed grosbeak
Brown-headed cowbird

Euphagus cyanocephalus
Pheucticus melanocephalus
Molothrus ater

US
TC
US

SM

SM

TC/SM/US

Northern oriole

Icterus galbula

TC

TC

TC = tree canopy; TB = tree bole; SM = shrub midstory; US
References: Bergeron et al., 1992; Johnsgard. 1992.

= understory; TS = te

rrestnal subsurface.

RCG/Hagler Bailly

7-8

Table 7-2

Mammals Characteristic of Southwest Montana Riparian Shrub/Forest

and their Feeding and Cover Vegetation Layers

Species

Feeding
Layers

Cover
Layers

Little brown bat Myotis lucifragus
Red fox Vulpes vulpes
Raccoon Procyon lotor
Mink Mustela vison
White-tailed deer Odocoileus virginicmus
Moose Alces alces

US
US

us

SM/US

SMAJS

TB

us/rs

TB/TC
US
SM

SM

TC = tree canopy; TB = tree bole; SM = shrub midstory; US = understory; TS =
Reference: Chapman and Feldhamer, 1982.

= terrestrial subsurface.

► Dominant vegetative cover type

► Number and type of habitat layers present

► Approximate height of each habitat layer in the dominant cover type

► Percent canopy closure of each species in the tree canopy, shrub midstory, and
understory layers

► Presence of white-tailed deer browse.

Data collected at each of the 17 sites represent a compilation of multiple sampling points; the
number of sampling points per site was determined by the width of the slickens. Sample
points were spaced at 10-meter intervals (see Appendix C).

7.3.1.1

Vegetation Characteristics

Analysis of cover type data revealed that the impact transects were dominated by one main
cover type, slickens, which constituted approximately 80% of observations (Figures 7-2 and
7-3). Control sites were predominantly shrub/forest (shrub and forest combined), hayfields, or
pasture. The one cover type observed on both impact and control reaches, riparian
shrub/forest, was more prevalent on the control sites (Figure 7-4). Representative views of
riparian impact and control sample sites are shown in Figures 7-5a through 7-5e.

RCG/Hagler Bailly

7-9

XT •

Locations of

Deer

Loi^r Impact Transects

R187K

MR16

LAl5
R14iy

1(3

//R13

4

Warm Springs IJ

North

^-^s~^ Opgortunity 1 J>

'Warm
' Springs
Ponds

\v^-^4*11

0

Scale of Miles
12 3 4 5

^f 2 R9

Smelter J \_ . _

Hiii 4IR °

1 R^

R8?

XV

_^B4

R1 4fe« 3

Prepared by Montana State Library
Natural Resource Information System

Map *93nrdc21 • 8/5/93

R6^bf

Creek «2 ^**^A /

Butte?

Figure 7-1. Location of Riparian Sampling Sites: Silver Bow Creek and the Clark
Fork River.

RCG/Hagler Bailly

o

I

<u

o
u.
\

3

CD

en

C

u C

3 a> t-

—» k w

>i » U £

CO CD C — *

W — C CQ O

»:-»>»»»x-»»»

^M

S o

uon^iusssjday jBUoi^Jodojj

a>

ff>

u

X

T3

C

TJ

o

s

c

«

c

^

<

u

o

u

O

-1

_

o

2

CO

o

U

1/5

c

«

(1)

SB

C

c

a:

t—

o

s

o

c

E

■a

<u

o

_

—

a-

^

«

UJ o

r-

tA

—j

tv

C

c

E

«*»

o

co

o
a.

H
l_

>

0)

0)

o

a.

©

re

c

u

CTj

in

to

~

"3

CO

re

£

■<t

o
C

0)

c

c/l

c

n

o
e

O
O.

a.

OJ

s

la

E

re

GO

in

u

C
O

DJ

"re
s
o

~
©
c
o

es

X

u

re

a.

'_

M

>^

—

'in
CD

1

u

r-

00
C3

u

3=

3
DC

a
u

u.

k:

o
o

D
O
Vi

in

"3

g

O

- 0>

00

CD

- 10

CO
CM

o

o —
z —

o 5

<

o
-

=

C

Q.
UJ

Q.

<
(0

< u

0

e
e

s

-

c
U

< —

s

D

CM i- CO CO *■ CM

■ ■ • ■ ■

i- o o o o

SNOUVAHBSaO dO NOUHOdOUd

c
e

M

u
i_
&
u
OS

lu

.2 *

u "5
o c

© 9r

■- —

c <

ox

to

—

00
(S

U
a:

i

V)

W

ee

0)

00

«
a.

E

■o
s

«

"o
1-

c

u

_

e

T £

i en

CO

in

Of ■

LLI

o

o

!Q

cc ■

I £

CO

o

s l

CO

- CM

co in ^t co cn i-

• ••••■

o o o o o o

SN0I1VAU3S90 dO NOIldOdOUd

s

Cfl

u

c

c
o

c

ee
s
to

—

o
a.
o
i_
C

U (J

-.*

o c
m w

5 °-

O Q.
•- <

5 «

|S

O o

ea
CD

-
3

Si

■ —

00

n

a
u

7-13

Impact site cover types were compared statistically to control sites. The proportional
representation of barren slickens was significantly greater on impact streams than on control
streams; the proportional representation of hay, pasture, and riparian forest was significantly
higher on control streams than on impact streams. Results of comparisons of all impact and
control streams, and of comparisons of paired control and impact streams, are presented in
Table 7-3. Each impact reach exhibited significantly greater proportional representation of
slickens than its paired control reach. The Little Blackfoot control reach exhibited
significantly greater proportion of riparian forest, hay, and pasture than the lower Silver Bow
Creek, and Flint Creek exhibited a significantly greater proportional representation of hay and
riparian shrub than the upper Clark Fork River.

Analysis of habitat layers data revealed that the impact reaches were lacking in three of the
five habitat layers observed consistently on control reaches (Figures 7-6 and 7-7). Tree
canopy and terrestrial subsurface layers were represented only sporadically on impact reaches,
and the tree bole layer was entirely absent. Absence of all habitat layers was recorded at
more than 60% of the impact sample points, and most layers that were recorded were present
at less than 30% of the impact sample points. The control reaches exhibited significantly
greater proportional representation of tree canopy, tree bole, understory, and terrestrial
subsurface layers than did impact reaches (p = 0.12, 0.03, 0.0002, and 0.0002, respectively).

The proportional representation of habitat layers was compared individually by paired impact
and control reaches (Table 7-4). All three control reaches had significantly higher proportions
of understory and terrestrial subsurface than corresponding impact reaches. The Little
Blackfoot had a significantly higher proportion of sites at which shrub midstory, tree canopy,
and tree bole were present than did the corresponding impact reach, lower Silver Bow Creek.
None of the impact reaches had significantly higher proportions of any habitat layer than the
control reaches.

The number of habitat layers on control reaches ranged between two and five (Figures 7-8
and 7-9). The range was a function of distance from the river, the near-bank riparian
shrub/forest is a vertically diverse habitat, whereas hayfields and pastures, which dominated at
distances greater than 60 meters from the bank, are less diverse, exhibiting only understory
and terrestrial subsurface layers (Figure 7-10). In contrast, for the impact areas, the most
common observation was absence of all habitat layers, and the greatest number of habitat
layers recorded at any point was three (Table 7-5).

Within the Opportunity Ponds, the most prevalent cover type, and almost the only cover type,
was bare ground (Figure 7-4). No site had more than two habitat layers, and approximately
80% of the sites sampled had no habitat layers. No tree canopy, tree bole, or terrestrial
subsurface, and limited representation of shrub midstory, and understory was observed in the
Opportunity Ponds. The observations were compared statistically to all of the control
reaches; the control areas exhibited significantly greater proportions of shrub/forest (p < 0.05),

RCG/Hagler Bailly

7-14

Figure 7-5a. Riparian Impact Sample Site R7 Showing Slickens, Devegetation, and Dead
Willows.

i—

•- -• — £ ^ art • k •»»

— — _■-

Figure 7-5b. Riparian Impact Sample Site R8 Showing Slickens, Devegetation, and Dead
Vegetation.

RCG/Haglcr Bailly

7-15

Figure 7-5c. Riparian Impact Sample Site RIO Showing Slickens, Devegetation, and
Dead Willow Shrub.

RCG/Hagler Bailly

i

u

•_

—

c

—

-

_

^m

Gff

o

eg

'u

■-.

c.

_•

y.

ci

-
—

Of)

u

~

&

—

~

o

^

s

^™

OS

o

c/:

a

_

Z.

'•_

c

>

jC

o

5

C/3

^-

i.

y.

B

o

=

~

o

c*

•-

—
-SL

a

u

2^

«

—

B

~;

CQ

r;

"6

tr.

r~

4.

la

=

:j

DO

C3

a
a:

7-17

Table 7-3
Comparison of the Proportional Representation of Cover Types in Impact and Control Reaches

Area

Cover Type

p-value <

All impact reaches vs. all control reaches

Barren slickens
Hay
Pasture
Riparian forest

0.0002***
0.002***
0.0002***
0009***

Upper Silver Bow Creek vs. Divide Creek

Barren slickens
Pasture

0.0014***
0.031*

Lower Silver Bow Creek vs. Little Blackfoot

Barren slickens
Hay
Pasture
Riparian forest

0.0014***
0.036*
0.046*
0.028**

Upper Clark Fork River vs. Flint Creek

Barren slickens

Hay

Riparian shrub

0.002***

0.018**

0.03**

*, **, and *** indicate significant differences in the proportional representation at a = 5%, 3.3%, and
1%, respectively. Slickens representation was significantly greater on all impact streams: the
remaining cover types were significantly more abundant on control streams

hay (p < 0.08), and pasture (p < 0.06), and mean number of habitat layers (p < 0.0002) The
Opportunity Ponds showed a significantly greater proportion of bare tailings (p < 0.0002)

Thus injury to the riparian vegetation of Silver Bow Creek, the upper Clark Fork River, and
the Opportunity Ponds is evident as a virtual elimination of the riparian forest, shrub,
agricultural grassland cover types, and the tree canopy, tree bole, and soil layers, in areas of
slickens or tailings deposition. In areas where mixed agricultural and native plant
communities consisting of trees, shrubs, forbs, and grasses are expected, slickens and tailings
are bare, unvegetated landscapes.

7.3.1.2

Habitat Suitability

The capacities of the injured and control areas to provide wildlife habitat were compared
using HEP models.

RCG/Hagler Bailly

■ 08

CI CI CI CI CI CJ CI CI CI CI
1X3468701 10

Sample Point

a. Tree Canopy

7-18

£ 0.8

CI CI CI CI CI CI CI CI CI CI
] 23488789 10

Sample point

b. Tree Bole

Sample Point

c. Shrub Midstory

I

ci ci ci ci ci c: c i

12 3 4 5 8 7

Sample Point

d. Understory

c I
10

Sample Point

e Subsurface

Figure 7-6. Proportional Representation of the Habitat Layers on Riparian

Control (C) and Impact (I) Reaches. Sample points represent distance from
the river at 10 meter intervals Source: Appendix C.

RCG/Hagler Bailly

On
i

<D LO ^" CO CM i-

■ ■••■•■

o o o o o o o

SNO!lVAU3SaO dO NOIlHOdOHd

O

Q.
LJJ

Q.
<

s
o

CO

S
IU

O

IU

o

<
u.

<r
3
10
m
3

CO

E2

>■

E

o

«0

a

n

3

s
x

CO

□

IU

-J

o

CD

>

a.
O
z
<
u

U

X

'"&

c

c
c

<

u

L.

3
O

on

</■!

©
c

>>
"5

c
C

a

tz

o
a

4*
-

C

OS

D _

e

H

o
—
o
-

—

-

3

u

00
a

u
u

a!

-::

Table "-4

Comparison of the Proportional Representation of Habitat Layers

on Impact and Control Streams

Area Habitat Layer

p-value <

All impact reaches vs all control reaches

Tree canopy

: 12

Tree bole

0 0?"

-

Understorv

0.0002*"

Terr es~i! subsurface

o fwv)"** * *

Upper Silver Bow Creek vs Divide Creek

Understorv

c o:::*-*

Terrestrial subsurface

:::***

Lc^er Silver 3rv. Creek vs L:r„e Elackfco:

Tree car. re-

0 03"

?"_-. er

Tree bole

0.03"

Shrub rr:ds:rr.

0.038'

Understorv

0.0012*"

Terrestrial subsurface

0.0018*"

| Upper Clark Fork River vs. Flint Creek

Understorv

o.oc:*"

Terrestrial subsurface

""""**•£

*. **. and "* indicate siEniiicantlv zreaie: rerresentation on the control streams

a: q = 5°o. 33%.

ar.c '. : :. resr-ecnveh.

White-Tailed Deer

Statistical comparison of white-tailed deer habitat suitability index (HSI) values calculated for
impact reaches with HSI values calculated for control reaches revealed that slickens greater
than 100 meters wide provide significantly reduced habitat suitability (presence of cover and
browse) relative to the control reaches (Appendix C). In this assessment, white-tailed deer is
an indicator species in that it represents a broader component of the ecosystem, or all those
organisms dependent upon the integrity of the riparian forest and shrub.

Habitat Layers

The layers of habitat models showed that slickens are significantly less vertically diverse than
control areas, both when all control areas were compared with all impact areas and when
paired impact and control reaches were compared (Table 7-6). The results of this assessment
which addressed the relationship between habitat vertical complexity and the capacitv to
provide a diversity of wildlife habitat showed that slickens and tailings provide significantly
!ess diverse wildlife habitat relative to the control areas

r.ZT riacier Ba:'.>.

CN

c

■

U ■ B

■ o) «• ff

n u u >, u S

- C GJ c >, *

S >. >. _J S3 £

J5 « •£ -a

i o £ 3 5

O C * Ji O --

2 O H E- U. U.

S3

\ /
s /

sjaXeq ;o jaquinfj ;o uoncjussajday leuonJodojj

C

u u

c .

= £

o oj

_ ■ —

— p

* o

= 4J

re u

a- C

S =5

o m
_ o

S a.

c •_
« -p

* I

©

— w5

- o
c ^

a. £

00
I

r-
-

3

iZ

CO

re

a
u

_ o

o

CL
LU

o
in

e
o
-

-
o

O

■~

uj -

o -3

r _

□ 2

<c IE

O u

J

0) z

c r

Q.

UJ —

O R

2

<

CO

u „

o o o o

SNOI1VAH3S90 dO NOIldOdOUd

e
v

v.

V

i.
a.
u

a
lu

.2 x

u '•&
© c
c U

£ a.
ft. <

©\

■

-
-
s

m

BO

a

u

55»»>>>>>S->>S-Si>SS

wmmmmmmmmmmmmmm

wmmmmmmmmmmmmmmm

wmmmmmmmmmmmmm®

mmmmmmmmmmmmmmmmm

in

n

CM

sjsAeq -^B^iqBH jo jsquin^j uimuixej\

W5

-o

«£

a. >>

E J

C/3 l-

V

V)

• k.

W) U

i >

■S ■-

u -o

4^ p>\

*=£

c «

«8 C

'Z »

ee >

C CO

—

o

l)R

rest

u

ct (

b/fo

•^

K 3

c c
E "5

u

CB

_

.id
e re

c

o

CO

C)ai

ear-b

re

X)

cu

_

ITZ

r-

o

P

u

4-1

£ «

o

c

S "re

ct

'o

CO

co

CU

u
cu

u

o —

E

^^

e

"3 cu

o

m

id

VC

u

w

bscr

10 n

c
re

—

O -

i-

■*

re

eu

o

ayers

river

re
cu

k_

=0

CO

J cu

to

cu

o

•- E

c
re

*-•

-= o

t?

w

* £

■5

CJ

"~ cu

O =

re
cu

■^

>_ 2

re

**

QJ CO

C

o

m Numb

present di

'e

o

to
CU

3

3 U

p. —

To

.= co

re

Q.

* C

eg .::

"O

»- o

e

^

<£ q.

re

'a

©'

CD

(H^

i-

1

r-

—

21.

cu

a

u

5

3
U

a
o

eZ

0i

7-24

Table 7-5
Comparison of the Mean Number of Cover Types Observed on Impact and Control Reaches

Reach

p-value

All impact reaches vs. all control reaches
Upper Silver Bow Creek vs. Divide Creek
Lower Silver Bow Creek vs. Little Blackfoot River
Upper Clark Fork River vs. Flint Creek

0.0002***
0.0012***
0.001***
0.0028***

*** Indicates significantly greater number of habitat layers in the control reach at a = 1%.

Table 7-6
Comparison of Vertical Diversity and Habitat Complexity in Impact and Control Reaches

Reach

p-value <

All impact reaches vs. all control reaches
Upper Silver Bow Creek vs. Divide Creek
Lower Silver Bow Creek vs. Little Blackfoot River
Upper Clark Fork River vs. Flint Creek
Opportunity Ponds vs. all controls

0.0002***
0.001***
0.001***
0.001***
0.0002***

*** indicates significantly greater diversify in the control area at a = 1%.

Wildlife Populations

Populations of organisms dependent on the tree canopy and tree bole are likely to have
suffered the greatest loss of viability from the almost complete removal of riparian forest that
has occurred in the impact area. Species likely to have been lost to the injured areas as a
result of habitat loss are listed in Table 7-7. Other populations of riparian shrub and forest
organisms less dependent on tree canopy and bole are likely to have suffered reductions in
population viability because of the diminished representation of riparian shrub in the impact
areas.

The two surveys of bird populations showed that the densities of mergansers, great blue
herons, belted kingfishers, and dippers on the unimpacted Little Blackfoot River were
approximately 14 times greater than on Silver Bow Creek (Table 7-8). The densities of
spotted sandpipers were approximately similar on both reaches. Great blue herons,
mergansers, and belted kingfishers are piscivores and dippers are benthivores (they obtain

RCG/Hagler Bailly

7-25

Table 7-7

Bird and Mammal Populations that are Likely to have been Lost or Suffered Reduced Viability

because of Reductions of Tree Canopy, Bole, Shrub Midstory, Understory, and Terrestrial

Subsurface Habitat Layers in Riparian Impact Areas

Reduced Viability

Lost Species

Mallard

Great blue heron

Common merganser

Wood duck

Willow flycatcher

American kestrel

Black-billed magpie

Osprey

American robin

Bald eagle

Warbling vireo

Tree swallow

Yellow warbler

Great homed owl

Veery

Belted kingfisher

Song sparrow

Northern flicker

Red-winged blackbird

Bank swallow

Brewer's blackbird

Western wood-pewee

Brown-headed cowbird

Eastern kingbird

Red fox

Black-capped chickadee

Raccoon

House wren

White-tailed deer

European starling

Mink

American redstart

Moose

Northern oriole

Black-headed grosbeak

Little brown bat

their food from either the water column or the river bottom) Spotted sandpipers, however,
obtain much of their invertebrate prey from the river bank. This dietary difference explains
why spotted sandpipers are able to maintain population viability along a river reach that
(because of the impacts of hazardous substances to the fish and benthos) is unable to support
healthy populations of the other species surveyed.

Implications

The habitat degradation described above reduces the viability of wildlife populations
dependent on the tree canopy and tree bole layers of riparian forest (which has been entirely
lost from the injured areas). Of the 38 bird and mammal species characteristic of riparian
forest in southwest Montana, 50% are dependent on one or more of these layers and are likely
to have been lost within the injured areas. Thus, the deposition of slickens and tailings along

RCG/Hagler Bailly

7-26

Table 7-8

Mean Numbers and Densities (Birds/River Kilometer) of Birds Seen during Two Avian Surveys

on Silver Bow Creek and the Little Blackfoot River between Elliston and Avon

Species

Little Blackfoot

Silver Bow Creek

Mean Number

Mean
Density

Mean Number

Mean Density

Common merganser
Belted kmgfisher
Great blue heron
Spotted sandpiper
Dipper

1*

1

1.5

7.5
4.5

0.2
0.2
0.2
1.5
0.6

0

0
1

7.5
0

0

0

0.1

1.1

0

All species

15.5

2.7

8.5

1.2

* The merganser observation was of an adult female bird with a brood of seven ducklings.

Silver Bow Creek, along the upper Clark Fork River, and on Opportunity Ponds has resulted
in a significant reduction in wildlife habitat and habitat viability. Injured areas, that could
have supported natural and agricultural plant communities providing important habitat for a
diversity of wildlife species, are currently largely devegetated and extremely limited in their
ability to support viable wildlife populations.

7.3.2 Extent of Injury

Riparian vegetation and wildlife habitat have been injured by releases of hazardous substances
from mining-related activities along approximately 34 miles of river from Colorado Tailings
on Silver Bow Creek downstream to Deer Lodge on the Clark Fork River. Eight hundred
acres of riparian habitat along Silver Bow Creek, 215 acres along the upper Clark Fork River,
and some 5 square miles (3,400 acres) at Opportunity Ponds have been injured by the
deposition of slickens and tailings.

7.3.3 Ability of the Resource to Recover

In the absence of stress, plant communities have the capability to recolonize disturbed land.
The extent to which this may occur and the direction of the recovery (i.e., the resulting
composition and structure of the recolonist communities, and the degree to which they
resemble the pre-impact communities) are determined by whether residual stress persists, and
whether the original stress has not irrevocably altered the abiotic conditions of the site.

RCG/Hagler Bailly

7-27

Phytotoxicity studies performed during the assessment of injury (see Appendix B) have shown
that the ability of the riparian areas to revegetate and provide wildlife habitat is severely
constrained by the current high to severe phytotoxicity of soils caused by elevated
concentrations of hazardous substances. The existing contamination has prevented much of
the slickens from undergoing any substantial revegetation. Redistribution of tailings material
during high water events further spreads tailings material and hazardous substances to
additional areas. Evidence of the redistribution of phytotoxic substances can be seen where
previously vegetated floodplains now support only dead, undecomposed willow stands. The
only areas where revegetation has occurred are where a slickens deposit has been removed or
has been overlain with uncontaminated soil material.

Without restoration, these constraints on revegetation will persist in the future, and it is
extremely unlikely that substantial revegetation would occur except over a time scale of
centuries. In addition, it is highly unlikely that natural restoration would result in riparian
species assemblages that resemble those lost as a result of exposure to hazardous substances,
because none of the pre-impact species have evolved tolerances and or persisted despite
contamination.

7.4 REFERENCES

Bergeron, D., C. Jones, D.L. Genter, and D. Sullivan. 1992. P.D. Skaar's Montana Bird
Distribution. Montana Natural Heritage Program Special Publication No. 2.

Chapman, J. A. and G.A. Feldhamer. 1982. Wild Mammals of North America. The Johns
Hopkins Press, Baltimore and London.

Hansen, P., R. Pfister, J. Joy, D. Svoboda, K. Boggs, L. Myers, S. Chadde, and J. Pierce.
1989. Classification and Management of Riparian Sites in Southwestern Montana. Montana
Riparian Association, School of Forestry, University of Montana, Missoula, MT.

Johnsgard, P. A. 1992. Birds of the Rocky Mountains. University of Nebraska Press,
London.

Mosconi, S.L. and R.L. Hutto. 1982. The Effects of Grazing on Land Birds of a Western
Montana Riparian Habitat. In Wildlife-Livestock Relationships Symposium: Proc. 10, Univ.
Idaho For. Wildl. Range Exp. Sta., Moscow, ID. pp. 133-140.

MRA. 1992. Riparian Vegetation Damage Assessment on the Clark Fork River and Silver
Bow Creek Floodplains. Prepared for Montana Department of Health and Environmental
Sciences by the Montana Riparian Association, Montana Forest and Conservation Experiment
Station, School of Forestry, University of Montana. Missoula, MT.

RCG/Hagler Bailly

7-28

Ohmart, R.D. and B.W. Anderson. 1986. Riparian Habitats. In A.Y. Cooperrider, R.J.
Boyd, and H.R. Stuart, (eds.), Inventory and Monitoring of Wildlife Habitat. U.S. Dept
Inter., Bur. Land. Manage Service Center, Denver, CO. pp. 169-200.

Short, H.L. 1984. Habitat Suitability Index Models: The Arizona Guild and Layers of
Habitat Models. U.S. Fish and Wildlife Service. FWS/OBS-82/10.70.

Short, H.L. 1986. Habitat Suitability Index Models: White-Tailed Deer in the Gulf of
Mexico and South Atlantic Coastal Plains. U.S. Fish and Wildlife Service. Biol. Rep.
82(10.123).

RCG/Hagler Bailly

JM

8.0 BIOLOGICAL RESOURCES — OTTER, MINK, RACCOON

This chapter summarizes the results of injury determination and quantification studies for
semi-aquatic furbearers, specifically river otter (Lutra canadensis), mink (Mustela visori) and
raccoon (Procyon lotor)1 Otter and mink rely heavily (entirely, for otter) on a diet of fish
and aquatic macroinvertebrates (Melquist and Dronkert, 1987; Eagle and Whitman, 1987);
raccoons (Procyon lotor) feed primarily on aquatic macroinvertebrates (Sanderson, 1987). It
was reported previously (Lipton et al., 1995) that Silver Bow Creek is almost entirely devoid
of fish. Hence, Silver Bow Creek cannot support viable populations of species that rely on
fish in their diets. This, alone, is sufficient to conclude that otter, mink, and other fish-eating
wildlife are injured throughout Silver Bow Creek. These injures are caused by exposure to
hazardous substances released from multiple sources in the Butte area.

Overall, the results of the injury determination and quantification studies demonstrate that
otter, mink, and raccoon have been injured throughout the lengths of Silver Bow Creek and
the Clark Fork River (from Warm Springs Ponds to Milltown). Specifically:

► Silver Bow Creek is almost entirely devoid offish (see Lipton et al., 1995). It
cannot, therefore, support viable populations of otter or other fish-eating
wildlife.

► Populations of otters, mink, and raccoons are significantly reduced relative to
baseline conditions. Otter are completely absent from Silver Bow Creek and
the Clark Fork River from Warm Springs Ponds to Milltown as shown in field
surveys, trapping records, and anecdotal accounts. In contrast, otter were found
at control sites along the Big Hole River. Population densities of mink,
raccoon, and otter were found to be higher at the control sites than at sites
along the Clark Fork River. These population reductions (elimination, in the
case of otter) are not caused by human disturbance, trapping, land-use, or other
anthropogenic impacts.

► Hazardous metals and arsenic are known to cause adverse effects on mink and
otter, including death and local population reductions, as shown in the scientific
literature. Dietary exposure concentrations of lead measured in fish sampled
from the Clark Fork River were found to exceed a safe tissue residue criterion.

*■ These three species have been, and continue to be, exposed to elevated

concentrations of hazardous substances in their diets — as demonstrated by

These studies are described in Appendix E: "Exposure to and injury from environmental metal
contamination on semi-aquatic mammals in the Upper Clark Fork River, Montana" by H.L. Bergman and
M.J. Szumski.

RCG/Hagler Bailly

8-2

tissue analysis of representative prey items and by significantly elevated tissue
concentrations in mink trapped from the Clark Fork River relative to control
sites.

8.1 INTRODUCTION

River otter, mink, and raccoon collected from areas exposed to anthropogenic inputs of metals
have been shown to contain elevated concentrations of lead (Blus and Henny, 1990; Valentine
et al., 1988; Wren et al., 1988), cadmium (Wren et al., 1988; Blus et al., 1987), and copper
(Wren et al., 1988; Everett and Anthony, 1976) compared to animals from uncontaminated
areas. Adverse effects on wildlife, including death (Wren, 1985; Diters and Nielsen, 1978;
Benson et al., 1976; Wobeser, 1976) and reduced local population (Blus and Henny, 1990;
Eisler, 1988; Eisler, 1985), have been attributed to metal contamination from anthropogenic
sources. Otter and mink are now considered by United States, Canadian, and European
experts the two mammalian species most sensitive to aquatic pollutants in streams and lakes
because of their position at the top of the aquatic food chain, and because of their known
sensitivities to contaminants (Addison et al., 1991).

As reported in Lipton et al. (1995), Silver Bow Creek is almost entirely devoid offish
because of elevated concentrations of the hazardous substances arsenic, cadmium, copper, lead
and zinc. Therefore, Silver Bow Creek cannot now (nor has it been able to in the past
decades) support a viable population of otters because of hazardous substance exposures.
Similarly, the Clark Fork River was devoid of fish until the 1970s because of hazardous
substances released from Butte and Anaconda (Lipton et al., 1995); the Clark Fork River
could not have supported viable otter populations when devoid of fish.

Based on recent trapping records, and confirmed by field studies of population abundance (see
Section 8.5.1), river otter currently are absent from the upper Clark Fork River. Since there
are effectively no otter on the upper Clark Fork River, determining the effects of hazardous
substances on this species was based on a review of the scientific literature and on
comparisons of population densities with a nearby control river not exposed to elevated
concentrations of hazardous substances (the Big Hole River).

Evaluation of hazardous substance exposures to otter was based on use of a surrogate species
that is found in the Clark Fork River. Mink were selected as the surrogate species because
they occupy a similar piscivorous, upper trophic position as otter, and are still present on the
upper Clark Fork River. Mink, however, because of their dietary preferences (they consume
less fish in their diets than otter) are less exposed to hazardous substances than otter.
Therefore, they represent a lower-bound estimate of contaminant exposure and should not be
construed as being fully reflective of injury to otter.

RCG/Hagler Bailly

8-3

A study performed as part of the terrestrial resources injury assessment (Appendix E)
quantified the extent of exposure to hazardous substances in prey of mink, otter, and raccoons
in the upper Clark Fork River, and evaluated whether food-chain exposures to these hazardous
substances are sufficient to result in injury. This was accomplished by:

► Identifying the principal aquatic foods of mink inhabiting the upper Clark Fork
River and control sites

* Determining metal concentrations in the principal aquatic foods of mink
collected from the Clark Fork River and control sites

► Comparing metal concentrations in the Clark Fork River mink and mink from
control sites

► Determining whether the Clark Fork River fish tissue metal concentrations
exceed the safe limit for consumption by mink and otter

*■ Determining whether morphological and histological injuries associated with

metal toxicity occur in mink collected from the upper Clark Fork River and
control sites.

In addition, the study evaluated whether populations of piscivorous mammals have been
reduced in the Clark Fork River relative to control sites. This quantification was
accomplished by:

*■ Determining the relative abundance of otter, mink, and raccoon on the upper

Clark Fork River and control sites

► Comparing the relative quality of streamside mink and otter habitat on the
upper Clark Fork River and control sites to determine whether potential
differences in animal abundance are related to differences in streamside habitat
quality

► Comparing the relative trapping pressure, numbers of humans and livestock,
irrigation demands, and average distance to the nearest highway along the
upper Clark Fork River and control sites to determine whether potential
differences in animal abundance are related to anthropogenic disturbance

► Examining trapping records to quantify the number of otter trapped from the
Clark Fork River and from control sites.

RCG/Haglcr Bailly

8-4

8.2 PATHWAY DETERMINATION

There are two main exposure pathways for mink, otter, and raccoons:

►■ The principal exposure pathway to mink and otter is via ingestion of fish

exposed to hazardous substances from water and from benthic
macroinvertebrates (see Lipton et al., 1995). Aquatic biota make up the entire
diet of otter and a substantial component of mink diet.

► The principal exposure pathway to raccoons is via ingestion of aquatic

macroinvertebrates exposed to hazardous substances from water and streambed
sediments (see Lipton et al., 1995). Aquatic macroinvertebrates are a
substantial component of the diet of raccoons (Sanderson, 1987).

Confirmation that these resources have bioaccumulated metals (and, therefore, act as an
exposure pathway to raccoons, mink, and otter) is presented below for fish and in Lipton et
al. (1995) for benthic macroinvertebrates and fish.

In addition, as described previously, injuries caused by hazardous substances to fish and
invertebrates that serve as prey to semi-aquatic mammals represent an indirect cause of injury.

8.2.1 Mink and Otter Dietary Exposure to Hazardous Substances

As described in Appendix E, stomach and intestinal analyses carried out on 26 of 27 mink
trapped on the Clark Fork and reference sites during the 1991-1992 trapping season showed
that fish, including mountain whitefish, sucker, and trout species, was the greater part of mink
diet (Table 8-1).

Whole-body metal concentrations in brown trout collected from the Clark Fork River and
from control sites were reported in Lipton et al. (1995). Trout from the upper Clark Fork
River were found to have significantly higher levels of copper, cadmium, lead, and arsenic
than trout from control areas.

Whole-body metal concentration means and standard errors from whitefish and sucker
collected on the Clark Fork and reference sites are given in Tables 8-2 and 8-3, respectively.
Whitefish caught downstream from Warm Springs Ponds in 1991, and downstream from
Warm Springs Ponds and upstream from Turah Bridge in 1992, had significantly higher
whole-body concentrations of Cd and Pb compared to fish from reference sites (Table 8-2).
Whole-body Cu concentrations were significantly greater in 1992 fish from Warm Springs
and Turah Bridge; As concentrations were significantly greater in 1991 fish downstream from
Warm Springs as compared to fish from reference sites.

RCG/Hagler Bailly

8-5

Table 8-1

Frequency of Occurrence of Food in Intestine and Stomach

of Mink Collected from All Clark Fork River and Reference Sites

during 1991-1992 Trapping Season

Intestine Samples (N = 26)*

Food Item Percent Frequency of

Occurrence in Samples

Unidentified Fish 30.8

Sucker 15.4

Trout 11.5

Whitefish 3.8

Overall Fish 61. 5b

Small Rodent 11.5

Muskrat 7.7

Aquatic Invertebrates 26.9

Empty 2 6.9

Stomach Samples (N = 26)*

Food Item Percent Frequency of

Occurrence in Samples

Unidentified Fish 7.7

Trout 3 . 8

Overall Fish 11. 5C

Small Rodent 3.8

Muskrat 3.8

Empty 80.8

* Intestinal and stomach contents were not determined in one of
the 27 mink trapped; thus N = 26 for this analysis.

b 84 percent of intestines with food present

c 60 percent of stomachs with food present
Source: Appendix E.

RCG/Hagler Bailly

8-6

Table 8-2

Whitefish Mean Body Weight and Mean Whole Body Metal Concentrations

(and Standard Errors of the Means) for Fish Collected from

the Clark Fork River and Reference Sites in 1991 and 1992

Mean

(SEM)

Year

Fish

Cadmium

Lead

Copper

Zinc

Arsenic

Location

N

Wt .(g)

ng/g dw

ng/g dw

ug/g dw

ug/g dw

ng/g dw

1991

Warm Springs

4

316

87*

2031

4.0

64.6

930"

(67)

(27)

(49)

(0.2)

(1.0)

(137)

Reference

5

313

34

64

3.2

88.3

546

Sites

(23)

(10)

(10)

(0.6)

(9.3)

(110)

1992

Warm Springs

4

270

2341

468'

6.1*

66.1

386

(37)

(24)

(90)

(0.6)

(3.6)

(40)

Turah Bridge

4

324

183*

558*

8.0*

71.5

338

(23)

(37)

(251)

(1.8)

(3.9)

(193)

Reference

7

296

43

106

3.0

75.7

537

Sites

(28)

(7)

(23)

(0.1)

(4.1)

(111)

* Significantly higher than control at a = 0.05

Source: Appendix E.

RCG/Hagler Bailly

8-7

Table 8-3

Sucker Mean Body Weight and Mean Whole Body Metal Concentrations

(and Standard Errors of the Means) for Fish Collected

from the Clark Fork River and Reference Sites in 1992

Year,
Location

Mean (SEM)

Fish Cadmium Lead Copper Zinc Arsenic
N wt (g) ng/g dw ng/g dw ug/g dw ug/g dw ng/g dw

1991

Inadequate number of fish for statistical comparisons

1992
Warm Springs 5

Turah Bridge 4

Reference 4
Sites

883
(61)

588
(66)

607
(112)

345*
(42)

166'
(10)

72
(28)

357*
(46)

408'
(77)

172
(84)

8.2'

65.8

584

(0.7)

(4.8)

(82)

10.6*

77.5

277

(0.8)

(3.9)

(53)

4.5

72.6

1807

(0.1)

(1.6)

(1252)

* Significantly higher than control at a = 0.05

Source: Appendix E.

RCG/Hagler Bailly

8-8

Suckers collected downstream from Warm Springs Ponds and upstream from Turah Bridge in
1992 had significantly higher whole-body concentrations of Cd, Pb, and Cu compared to
suckers collected from reference sites (Table 8-3).

Overall, the results of these studies demonstrate that three main fish prey of mink and otter
had elevated tissue levels of hazardous substances at the Clark Fork River sites, as compared
with fish from control locations. Fish, therefore, represent a pathway of these hazardous
substances to these furbearers.

8.2.2 Mink Tissue Concentrations of Hazardous Substances

In addition to confirming the presence of hazardous substances in pathway resources,
exposure to furbearers was confirmed by analysis of tissue samples for mink, as described
below.

Cadmium

Adult liver and kidney Cd concentrations were found to be significantly higher in mink from
Warm Springs Ponds and in pooled adult samples from the Clark Fork sites, compared to
reference sites (Tables 8-4 and 8-5). Liver Cd concentrations in juveniles from the pooled
Clark Fork sites are higher than from reference sites. Juvenile mink brain Cd concentrations
from Warm Springs Ponds were elevated compared to mink from reference sites (Table 8-6).

Lead

The Clark Fork mink liver and kidney Pb concentrations are significantly greater than those
from reference sites in both age classes, and across all the Clark Fork site comparisons
(Tables 8-4 and 8-5). Brain Pb is significantly higher in adult mink from Warm Springs
Ponds and in adult mink from the pooled Clark Fork sites than in adult mink from reference
sites (Table 8-6).

Copper

The Clark Fork mink liver Cu concentrations are significantly greater than reference sites in
both age classes and across all site comparisons (Table 8-4). Kidney Cu is significantly
higher in the Warm Springs Ponds juveniles than in reference juveniles (Table 8-5), and
significantly higher in the comparison between the pooled Clark Fork sites and reference sites
for juveniles at the 10% level (p = 0.0649). Brain Cu is significandy higher in juvenile mink
from Warm Springs Ponds than from reference sites, and this difference is significant at the
10% level (p = 0.0608) in the comparison of juveniles from pooled Clark Fork sites
(Table 8-6).

RCG/Hagler Bailly

8-9

Table 8-4

Mink Mean (and Standard Error of the Mean) Liver Metal Concentrations

by Trapping Location for Mink Collected on the Clark Fork River and Reference Sites

During the 1991-92 Trapping Season

Mean (SEM)

Cadmium

Lead

Copper

Zinc

Arsenic

Location

N

ng/g dw

ng/g dw

ug/g dw

ug/g dw

ng/g dw

Juveniles

Warm Springs

3

676

506'

54*

117*

221*

Ponds

(379)

(80)

(12)

(12)

(91)

Deer Lodgeb

2

646

1093

56

111

308

(326)

(237)

(15)

(16)

(137)

Clinton

5

359

600*

431

88

406*

(37)

(63)

(2)

(3)

(137)

Combined CFR

10

512

670*

49'

101

3 31*

Sites

(122)

(88)

(4)

(6)

(76)

Combined

5

275

212

28

89

24

Reference

(68)

(74)

(3)

(9)

(29)

Adults

Warm Springs

5

1014*

994*

35'

103'

1092

Ponds

(187)

(132)

(4)

(6)

(846)

Deer Lodgeb

1

1394

826

47

72

189

Clinton

0

NO ADULT

MINK

CAPTURED

Combined CFR

6

1078*

966*

37*

98

942

Sites

(166)

(112)

(4)

(7)

(706)

Combined

6

434

187

19

85

269

Reference

(78)

(104)

(2)

(5)

(178)

' Significantly higher than control at a = 0.05

No statistical comparison made due to small sample size.

Source: Appendix E.

RCG/Hagler Bailly

8-10

Table 8-5

Mink Mean (and Standard Error of the Mean) Kidney Metal Concentrations by

Trapping Location for Mink Collected on the Clark Fork River and Reference Sites

During the 1991-92 Trapping Season

Mean (SEM)

Cadmium

Lead

Copper

Zinc

Arsenic

Location

N

ng/g dw

ng/g dw

ug/g dw

ug/g dw

ng/g dw

Juveniles

Warm Springs

3

1735

463*

23"

86

341

Ponds

(1040)

(45)

(2)

(10)

(91)

Deer Lodgeb

2

875

944

24

89

1263

(357)

(111)

(1)

(10)

(886)

Clinton

5

605

613"

18

75

196

(145)

(70)

(1)

(3)

(63)

Combined CFR

10

998

634"

21

81

453

Sites

(326)

(68)

(1)

(4)

(194)

Combined

5

566

252

17

92

274

Reference

(131)

(83)

(1)

(13)

(168)

Adults

Warm Springs

5

2604*

756*

18

78

1122

Ponds

(718)

(105)

(2)

(4)

(792)

Deer Lodgeb

1

3491

996

19

73

413

Clinton

NO ADULT MINK CAPTURED

Combined CFR

6

2752'

796*

18

78

1004

Sites

(604)

(95)

(1)

(3)

(657)

Combined

6

1114

211

15

69

180

Reference

(326)

(113)

(1)

(3)

(107)

* Significantly higher than control at a = 0.05

No statistical comparison made due to small sample size.

Source: Appendix E.

RCG/Hagler Bailly

8-11

Table 8-6

Mink Mean (and Standard Error of the Mean) Brain Metal Concentrations by Trapping

Location for Mink Collected on the Clark Fork River and Reference Sites

During the 1991-92 Trapping Season

Mean (SEM)

Cadmium

Lead

Copper

Zinc

Arsenic

Location

N

ng/g dw

ng/g dw

ug/g dw

ug/g dw

ng/g dw

Juveniles

Warm Springs

3

194

111

2 0*

67

267c

Ponds

(118)

(96)

(1)

(10)

(161)

Deer Lodgeb

2

18

268

15

58

145

(22)

(126)

(2)

(0)

(31)

Clinton

5

1

28

15

50

164

(3)

(24)

(1)

(2)

(31)

Combined CFR

10

62

28

16

57"

183

Sites

(42)

(45)

(1)

(4)

(36)

Combined

5

4

19

14

48

118

Reference

(5)

(13)

(1)

(2)

(58)

Adults

Warm Springs

5

7

159*

12

46

589

Ponds

(3)

(26)

(1)

(2)

(316)

Deer Lodgeb

1

8

350

10

47

124

Clinton

0

NO ADULT MINK CAPTURED

Combined CFR

6

7

191'

11

47

511

Sites

(3)

(38)

(1)

(1)

(270)

Combined

6

41

58

10

45

138

Reference

(19)

(20)

(0)

(4)

(44)

* Significantly higher than control at a = 0.05

b No statistical comparison made due to small sample size.

c N = 2

Source: Appendix E.

RCG/Hagler Bailly

8-12

Zinc

Liver Zn concentrations are significantly higher in both adult and juvenile mink from Warm
Springs compared to reference sites (Table 8-4). Brain Zn concentrations in juveniles from
the pooled Clark Fork sites were significantly elevated compared to reference sites
(Table 8-6).

Arsenic

Liver As concentrations are significantly elevated in juvenile mink from all the Clark Fork
sites compared to reference sites (Table 8-4).

Thus, there is a clear pattern of elevated levels of metals and arsenic in the tissues of mink
trapped on the Clark Fork, as compared to mink from control rivers. This further confirms
exposure to furbearers.

8.3 INJURY DEFINITION

An injury to mink, otter, or raccoons has occurred if they or their offspring "have undergone
one of the following adverse changes in viability: death . . . physiological; malfunctions
(including malfunctions in reproduction), or physical deformations" [43 CFR § 11.62 (0(1)1

In the Clark Fork Basin, injury to mink, otter, and raccoons is expressed in population
reductions relative to control sites. At the level of the individual animal, these population
changes have been caused by death and reproductive impairment. Injuries to mink, otter, and
raccoons have been found in this assessment to satisfy the four acceptance criteria for
biological responses [43 CFR § 11.62 (f) (2) (i-iv)]. Specifically, these injuries:

*■ Are often the result of exposure to hazardous substances, as shown in various

scientific studies, including the studies described in Appendix E

► Have been shown to occur in controlled laboratory experiments (see
Appendix E)

► Are often the result of exposure to hazardous substances among free-ranging
organisms, as documented in the literature (see Appendix E), and in the field
studies described in this chapter and in Appendix E

► Are routine measurements that are practical to perform and produce
scientifically valid results.

RCG/Hagler Bailly

8-13

8.4 INJURY DETERMINATION: RESULTS OF STUDIES

As described previously, Silver Bow Creek cannot support viable populations of otter because
it is devoid of fish. As outlined in Lipton et al. (1995), fish are not found in Silver Bow
Creek as a result of releases of hazardous substances. Therefore, otter have been injured
throughout Silver Bow Creek. Determination of injury to otter, mink, and raccoon in the
Clark Fork River was based on a review of the scientific literature (see Sections 8.4.1 and
8.4.2) and on comparisons of population densities (Section 8.5.1).

Evaluation of Whether Dietary Exposure Poses Risks to Mink and Otter on the Clark
Fork River

Models have been developed to estimate the maximum safe dietary exposure, or Tissue
Residue Criterion (TRC), of hazardous substances to piscivorous mammals. The model
employed by the New York State Department of Environmental Conservation (Newell et al.,
1987) was chosen for this assessment because it has been widely accepted by the scientific
community, and because it serves as the basis for other similar models such as the Canadian
Tissue Residue Guidelines (U.S. EPA 1992) (Appendix E). The model generates a TRC,
which is the maximum acceptable exposure concentration for each metal (in ug of metal/kg
diet) of a contaminant (Pb or Cd) in the diet.

Table 8-7 shows the Cd and Pb TRC values obtained from the model. TRC values were
slightly higher for otter than for mink for both Pb (89 and 67 ug/kg diet for otter and mink,
respectively) and Cd (533 and 400 ug/kg diet for otter and mink, respectively).

These criteria were then compared to the mean dietary exposure concentration in fish
downstream from Warm Springs Ponds and from reference sites in Table 8-8. The Pb dietary
exposure to mink consuming 50% trophic level three fish from the Warm Springs Ponds site
(51 ug Pb/kg diet) approaches the Pb criterion (67 ug Pb/kg diet) (Table 8-8). Lead exposure
to mink consuming a diet of 100% trophic level three fish from the Warm Springs Ponds site
(102 ug Pb/kg diet), and to otter consuming a diet of 50% trophic level three and 50%
trophic level four fish from the Warm Springs Ponds site (157 ug Pb/kg diet), exceeds the
respective Pb criterion. Hence, this model predicts adverse impacts on otter and mink given
these exposures. Dietary cadmium exposures to Warm Springs Ponds and reference site fish
did not exceed the respective criteria for either mink or otter (Table 8-8).

Based on the results of this comparison, otter and, to a lesser extent, mink are exposed to
concentrations of hazardous substances sufficient to cause injury.

RCG/Hagler Bailly

i
00

re
v

U ju

.12 .*

V

3

ON
00

CI

i-
U

o

■o

a

m

u

i—

u.

o

3

o

re

U.

01

?-

u.

C?

c

b

re

r

u

o

u

c

u.

3

3

VO

_ o

oo *+

H -S

•a

u

s

03

-3

Oh

u

OS
H

OS

Q

X

o

< -*

O u>

-J 3

O

c

c

oo

CO

o

1

1)

u

00

o

oo

On

C

o

S
u
u

-£ D

Q 03

u

c

E
•o

CO

U

0)

"8

c
u

W

5

o

c

3

•o

>>

c

^J

u

S

ex

Q.

CQ

<

ej

00

re

P

3

a

o

u

00

Bi

m

'j

1

tf_*

3

—

oo

e

•™

CO

o

re

1)

u

o

<_

j;

c

U

u

"3

'J

la

CU

C -^

1>

CJ

u

CU

U .2

"3

0.

©

1)

L. OO

U

s £

to

C to

to

pa

c =

C

M

>< ~— '

"u

LO

»■

b

;>-.

CO

M

c

S

Lead
Dictar

E

L_

3

V)

s

o

es

U

L.

"C

V

u —

«a

«J

— CM

S"-3

O £

-D 00

■O (73

"35 -*

- w

«S «

- *>

t/;

«S cu

P

I- PS

«£ -3

c

u

oeB. R

o

"to

80 "O "O

c

c

., s e

u

CJ

£ « o

o

aj

-a _ 2-

H w 6X

o S"
U ~

•_ s

i- to

£ "S

3 ^

to ■ —

to

Q.

o to

to

2 «va

C- -

C

X >— '

"C

© f-

tu

C-

— s

co

S «

E 2?

3 TO

E

ncenti
ow \V

Cadmi
Diet

u

CO

51

TO

"u.

tu

1)

L.

'La

s

u ~

VI

1)

o
c

3 •«

"O oo

"5 -*

£

(2 §"

R

3

«a

to

u

to

5

p

"O

c

X>

C8

*s

VI

tu

_3

ss

"ee

'0.2

>

C J=
CO D

U

O

w

OS

H

H

r-i

no

o

r-

ND

NO

on
oo

CM

ON

rr

On

ON

o
r-

>

CO

c
o

to £J

i_ ON

«J ON

^ —

5 <=

■P E
<— oo

£ «

O 60

'•a >»
2 -a

u _—

P!2

u

o

IE X!

p E

= ^

— o V

-* ^

IIS

2 "g Lc .<2

x
U

"8

u

i> C ,-C

c
u

c
o
o

1

oo ^

? 5

p e

"S =

S TO

ON
ON

<

c

CO

cr^ 1

3

TO

la

O

3
U

C

o

TO

?

-o

O

o

c

la

TO

;—

JO

±£

£

a

'e

i.

en

C

c

TO C

~, u

^r

^ 1

-o

TO

la ■

m

c-x

X oo
u

>

oo >

u

etal/
dry

E E

o

00 £

<—

3 *-

o

c/j "O

n
_o

"to

ted a
3 an

C o

o

u -S

D.

C/3 aO

E

p

« .£

>-.

t—

la

Cl c

X U

TO

:ary e
) pres

C

o

"8

4J 02

WJ

.— to

C3

Q tC

CD

H

X

"*5

_>»

c

pa

cu

CO

a.

CQ

a.

<

ka

u

O0

TO

cu

o

5

ka

3

a

O

o

tvn

Da

8-16

8.5 INJURY QUANTIFICATION

Because of their secretive natures, mink, otter, and raccoon populations cannot be readily
determined by direct counts of animals. The most reliable ways of assessing their population
status is by carrying out sign surveys (surveys of tracks, scat), or by analyzing trapping
records. These approaches were used to compare mink, otter, and raccoon populations on the
Clark Fork River and at control rivers (Appendix E).

8.5.1 Population Reductions among Mink. Otter, and Raccoons on the Upper Clark
Fork

The populations of mink, otters, and raccoons on the Clark Fork and control rivers were
assessed using sign surveys as indices of population levels, and by analyzing trapping records.

Trapping Records

Analysis of trapping records kept by Montana Department of Fish, Wildlife, and Parks
(MDFWP) indicates that there appears to be a "hole" in the map showing otter trapping
records for western Montana in the area corresponding to the upper Clark Fork and its
tributaries (Appendix E; Figure 8-1). This could be explained either by reduced otter
populations on the Clark Fork, or by disparities in the amount of trapping effort between the
upper Clark Fork and the surrounding drainages.

As described in Appendix E, in virtually all cases, otters are not actively pursued by trappers
in Montana, but are trapped incidentally in beaver sets. The number of otter trapped in a
given area, therefore, is a reflection of effort in beaver trapping in that area. One would
therefore expect two areas with similar numbers of beaver trapped to also produce a similar
number of incidental otter captures, if there were no differences in otter numbers between the
two areas.

Analysis of the trapping records for administrative regions two (which contains the Clark
Fork) and three (which lies immediately south of the Clark Fork valley) showed that the two
areas have similar beaver harvests, based on the number of animals caught over the 5 years
examined. Thus, the beaver trapping effort in the region that includes the Clark Fork was not
less than in the area that contains the Bitterroot.

This review of MDFWP trapping records was further focused to examine the number of otter
captured per river mile between the upper Clark Fork and two adjacent river systems: the
Bitterroot River immediately west of the Clark Fork, and the Big Hole River drainage
immediately south. Significantly fewer otter were trapped on the upper Clark Fork than on
the Big Hole and Bitterroot rivers. No statistical difference was seen between the numbers of

RCG/Hagler Bailly

8-17

Figure 8-1. Map of Western Montana Showing the Numbers and Locations of Otter
Trapped in each Montana Department of Fish, Wildlife, and Parks
Administrative Region (numbers) during the 1977-78 through 1989-90
Trapping Seasons. (Adapted from Zackheim 1982); small circles indicated
single otter, large circles indicate five otter. Note that only two otters were
trapped on the upper Clark Fork River during this time period. Source:
Appendix E.

RCG/Hagler Bailly

otter trapped on the Bitterroot and Big Hole rivers. Thus, the reason fewer otter are trapped
on the upper Clark Fork is most likely a reduced or eliminated river otter population.

Sign Surveys

Field studies were used to quantify otter sign at the Clark Fork River and a control river (the
Big Hole River). Significant differences were found in the amount of furbearer sign found
between the paired Clark Fork River and control sites (Figure 8-2). In the fall 1992 sign
survey, no otter sign was found on any of the Clark Fork River sites, whereas sign was found
on four of the six reference sites. Otter, mink, and raccoon sign were significantly more
common on reference sites than on the paired Clark Fork River sites (Figure 8-2).

Thus, the sign surveys demonstrated that mink, otter, and raccoon population densities are
lower on the Clark Fork than on control rivers. Otter have apparently been entirely
eliminated from the Clark Fork River between Butte and Missoula.

The differences in mink and otter population levels between the Clark Fork and control rivers
are not due to habitat differences or human disturbance (grazing pressure, proximity of
highways, trapping pressure, irrigation demands, or human population density); there were no
significant differences between the Clark Fork and control rivers in the habitat representation
or in human use patterns (Appendix E).

8.6 CONCLUSIONS

As described in Appendix E, the following conclusions can be supported by the injury
determination and quantification studies:

► Otter are injured throughout Silver Bow Creek.

► Whole-body concentrations of Pb, Cd, and Cu are significantly elevated in the
prey of otters and mink from the upper Clark Fork River, as compared to fish
from control sites.

*■ Metal residue concentrations for Pb, as well as Cd, Cu, Zn, and As, are

significantly elevated in liver, kidney, and brain tissues from mink trapped on
the upper Clark Fork River as compared to tissues from mink trapped on
control sites.

*■ For mink on the Clark Fork River, dietary exposure concentrations of Pb

exceed the Pb tissue residue criterion (a safe dietary intake concentration of
Pb), assuming that more than about 65% of their diet is suckers from below
Warm Springs Ponds.

RCG/Hagler Bailly

i
00

V©

»D

TT f^i ^

udi/uSiS

«

w

B

v

la

V

€■

V

0.

■o

E

u

■ UJ

>

Sf.s

C3

U g

C.J

w Q.

«

ark Fork River

.05. Source: Ap

$■!

U°

•—

V

O

"3 _,

O

o

s*

u

&«
■^ .«-»

C3

« s

g

Beaver Sign on S

significantly differ

►V

■■*■-«

■o «;

^*

B S

• w

■—

^

ter, Mink, Raccoon
Fall of 1992. ** ind

«■■

o

-*-*

<— — •

° M

-*-»

o

dance
Durin

bun
tes

_>,

5

< 1/3

•

o

c*

■

ox)

00

W

V

5

3

o

Ml

o

a!

8-20

► Dietary exposure concentrations of Pb to otter consuming fish from the upper
Clark Fork River exceed the safe Pb tissue residue criterion, based on an
estimate of probable exposure from consumption of 50% suckers and 50% trout
downstream from Warm Springs Ponds.

► Dietary exposure concentrations of Cd to mink and otter, though below the
respective safe tissue residue criteria for Cd, may contribute to metal toxicity in
mink and otter through synergistic interactions with Pb.

► Otter, mink, and raccoon population abundance on the upper Clark Fork River
is lower than on paired reference sites on the Big Hole River.

► Otter are absent from the upper Clark Fork River, based on complete absence
of sign information from MDFWP and Bureau of Land Management personnel
as well as from fur trappers in the area and historical trapping records for otter,
whereas they are present on all comparable rivers in western Montana.

► Habitat structural features that are important for otter and other piscivorous
mammals in the riparian zone along the upper Clark Fork River are similar to
the habitat structural features on paired reference sites on the Big Hole River,
where otter, mink, and raccoons are more abundant.

► The levels of human disturbance to otter and other piscivorous mammals
caused by trapping, grazing, irrigation demands, and proximity to highways and
human populations appear to be similar or less severe in the upper Clark Fork
valley as compared to similar river valleys in western Montana.

All of the above findings from this study and related studies, taken together in a weight-of-
evidence approach, indicate that the observed absence of otter on the upper Clark Fork River
is caused by metal contamination (particularly lead, with possible contributions from cadmium
and other metals) in fish and other aquatic organisms. Furthermore, the findings from this
study also support a conclusion that significantly reduced populations of mink and raccoon on
the upper Clark Fork River are also caused by metal contamination in the diet.

8.7 REFERENCES

Addison, E.M., G.A. Fox, and M.G. Gilbertson. 1991. Proceedings of the Expert
Consultation Meeting on Mink and Otter. Great Lakes Science Advisory Board's Ecological
Committee Report to the International Joint Commission. Windsor, Ontario. 30p.

Benson, WW., D.W. Brock, J. Gabica, and M. Loomis. 1976. Swan mortality due to certain
heavy metals in the Mission Lake area, ID. Bull. Environ. Contam. Toxicol. 15:171-174.

RCG/Hagler Bailly

8-21

Blus, L.J., and C.J. Henny. 1990. Lead and cadmium concentrations in mink from northern
Idaho. Northwest Sci. 64 : 2 1 9-223 .

Blus, L.J., C.J. Henny, and B.M. Mulhern. 1987. Concentrations of metals in mink and other
mammals from Washington and Idaho. Environ. Poll. 44:307-318.

Diters, R, and S. Nielsen. 1978. Lead poisoning of raccoons in Connecticut. J.Wildl.
Disease. 14:187-192.

Eagle, T.C., and J. Whitman. 1987. Mink (Chapter 46). pp 615-624. In Novak et al. (eds).
Wild Furbearer Management and Conservation in North America. Ontario Trappers Assoc.
Ontario, Canada.

Eisler, R. 1985. Cadmium Hazards to Fish, Wildlife and Invertebrates: A Synoptic Review.
U.S. Fish Wildl. Ser. Biological Report 85(1.2) 46p.

Eisler, R. 1988. Lead Hazards to Fish, Wildlife, and Invertebrates: A Synoptic Review.
U.S. Fish Wildl. Ser. Biological Report 85(1.14) 134p.

Everett, J., and R Anthony. 1976. Heavy metal accumulation in muskrats in relation to
water quality. Trans. N.E. Sect. Wildl. Soc. 33:105-118.

Ferm, V.H. and W.M. Layton, Jr. 1981. Teratogenic and mutagenic effects of cadmium. In
J.O. Nriagu (ed), Cadmium in the Environment. Part 2, Health Effects. John Wiley, New
York. pp. 743-756.

Melquist, W.E., and A. Dronkert. 1987. River Otter (Chapter 47). pp 627-641. In Novak et
al. (eds). Wild Furbearer Management and Conservation in North America. Ontario Trappers
Assoc. Ontario, Canada.

Newell, A. J., D.W. Johnson, and L.K. Allen. 1987. Niagara River Biota Contamination
Project: Fish Flesh Criteria for Piscivorous Wildlife. Technical Report 87-3. New York
State Department of Environmental Conservation. Albany, NY. 182pp.

Lipton, J., H. Bergman, D. Chapman, T. Hillman, M. Kerr, J. Moore, and D. Woodward.
1995. Aquatic Resources Injury Assessment Report, Upper Clark Fork River Basin. Prepared
by RCG/Hagler Bailly for the State of Montana, Natural Resource Damage Litigation
Program. January.

Rice, D.C. 1985. Chronic low-lead exposure from birth produces deficits in discrimination
reversal in monkeys. Toxicol. Appl. Pharmacol. 76:201-210.

RCG/Hagler Bailly

8-22

Sanderson, G.C. 1987. Raccoon (Chapter 38, pp 487-499). In Novak et al. (eds). Wild
Furbearer Management and Conservation in North America. Ontario Trappers Assoc.
Ontario, Canada.

U.S. EPA. 1992. Proceedings of the National Wildlife Criteria Methodologies Meeting.
U.S. EPA Office of Water and Office of Science and Technology. Charlottesville, VA. 40p.

U.S. EPA. 1993. Wildlife Criteria Portions of the Proposed Water Quality Guidelines for the
Great Lakes System. U.S. EPA Office of Water and Office of Science and Technology.
Washington D.C. 20460.

Valentine, R.L., C.A. Bache, W.H. Gutenmann, and D.J. Lisk. 1988. Tissue concentrations
of heavy metals and polychlorinated biphenyls in raccoons in central New York. Bull.
Environ. Contam. Toxicol. 40:711-716.

Wobeser, G. 1976. Mercury poisoning in a wild mink. J. Wild!. Disease. 12:335-340.

Wren, CD. 1985. Probable case of mercury poisoning in a wild otter, L. canadensis, in
northwestern Ontario. Can. Field Nat. 99(1): 112-1 14.

Wren, CD., K.L. Fischer, and P.M. Stokes. 1988. Levels of lead, cadmium and other
elements in mink and otter from Ontario, Canada. Environ. Poll. 52:193-202.

Zackheim, H.S. 1982. Ecology and Population Status of the River Otter in Southwestern
Montana. M.S. thesis, University of Montana. Missoula, MT. 100pp.

RCG/Hagler Bailly