Historic, archived document
Do not assume content reflects current
scientific knowledge, policies, or practices.
rea
—
AERIAL PHOTO
TECHNIQUES
FOR A RECREATION INVENTORY
OF MOUNTAIN LAKES AND STREAMS
RS ve?
ans
tad oa,
co as
— H0He2
FOREST SERVICE/U. S. DEPARTMENT OF AGRICULTURE
U.S. Forest Service
Research Paper INT -37
1967
AERIAL PHOTO TECHNIQUES FOR A RECREATION INVENTORY
OF MOUNTAIN LAKES AND STREAMS
by
Roscoe B. Herrington
and
S. Ross Tocher
INTERMOUNTAIN FOREST AND RANGE EXPERIMENT STATION
Forest Service
U.S. Department of Agriculture
Ogden, Utah 84401
Joseph F. Pechanec, Director
in cooperation with
UTAH STATE UNIVERSITY
College of Forest, Range, and Wildlife Management
Logan, Utah 84321
rE
THE AUTHORS
ROSCOE B. HERRINGTON is Recreation Resource Analyst
at the Intermountain Forest and Range Experiment Sta-
tion. Prior to his present appointment, his experience
included assignments in timber resource inventory,
timber marketing research, and production economics
research in Montana, Idaho, Colorado, and Utah.
. ROSS TOCHER was Assistant Professor of Forestry at
Utah State University, Logan, Utah, during this study;
he taught courses in recreation, aerial photo use, and
fire control. At present, Mr. Tocher is completing the
requirements for the Ph.D. degree at the University of
Michigan.
THE WATER RESOURCE INVENTORY IN RECREATIONAL PLANNING
The importance of water in comprehensive resource planning is clear to all. The man-
agement and development of the water resource is a difficult and complex task which must be
carried out according to the multiple use concept of public land management.
In planning for recreational water use, a complete inventory of lakes and streams ina
given area is a necessity. Such an inventory ideally provides information about the location and
size of the water body; about water volume fluctuations and water characteristics such as clarity,
temperature, and pH; and about insect life and the presence of plankton and underwater vegetation.
To date, any attempts at making such inventories have been confined to the larger lakes
and streams. There are many obstacles to the collection of water resource information, par -
ticularly in mountainous areas, which are of major interest for recreation. The field season for
exploration is often short. Roads may be few, and even trails may be lacking to many of the
smaller lakes and streams. Travel on the ground is blocked by snow and ice during much of the
year and is strenuous at any time. Also, great fluctuations in water volume leave only a short
time in late summer when comparative measurements can be made. The roaring spring torrent
may dwindle to a trickle by fall.
Thus, a field survey of the smaller lakes and streams is generally too time consuming
and costly to be practical. Other possibilities, such as individual air eurvey2 using planes or
helicopters, are both expensive and hazardous in rugged mountains.
A survey made from maps alone is unsatisfactory. The scale of the aerial photography
from which maps are usually made is not large enough to allow for identification of some of the
smaller water bodies, and much desired detail is not normally included on maps. Meadows are
sometimes mislabeled as lakes, and dry ravines are not consistently distinguished from year -
round watercourses.
A step toward the solution of the problem is the use of aerial photographs of 1:20, 000
scale. Photogrammetric techniques may be used to make rapid and reasonably accurate inven -
tories of mountain lakes and streams. By this means, existing maps may be corrected and sup-
plemented, and additional information useful for recreational planning may be gathered.
A TEST OF PHOTOGRAPHIC SURVEY TECHNIQUES
To test the procedures in a photographic survey, a study was conducted on the North
Slope of the Uinta Mountains in northeastern Utah (fig. 1). An attempt was made to collect as
much information as possible from the photographs, for use in an initial reconnaissance inven -
tory of recreational water resources. The data gained from interpretation of photos was checked
for accuracy by ground surveys. ‘The test showed that although some of the information desirable
for such a survey can be collected only in the field, a substantial amount of general descriptive
detail can be obtained rapidly and at reasonable cost from aerial photographs.
The study area was fairly typical of mountainous areas offering recreational possibilities.
Several hundred miles of streams and several hundred lakes lie within the 677,000 acres studied,
most of which is administered by the U.S. Forest Service, although scattered tracts are pri-
vately owned. The mountains are composed primarily of quartzites and other sedimentary rocks
and have been heavily glaciated, although no glaciers are now present. Elevations range from
about 6,200 to over 13,000 feet, and
much of the area is above timberline ;
(11,000 feet). The predominant tree NORTH SLOPE
cover is lodgepole pine and there are STUDY AREA
lesser amounts of Engelmann spruce,
Douglas-fir, and subalpine fir. Many Gace
aspen stands grow among these conif- River
erous forests. Alpine meadows are
numerous and give way to tundra above
timberline. Low willows are common
in the meadows and along streams.
At lower elevations, ponderosa pine,
pinyon pine, juniper, and sagebrush
are found.
Few roads currently penetrate
the North Slope. A much enlarged
road network is planned outside the
roadless High Uintas Primitive Area.
Thus many remote streams and lakes
may suddenly become accessible to
large numbers of people, creating a
need for an inventory of available
recreation resources and plans for
their development.
This paper describes the tech-
niques used in the test survey and
makes general recommendations.
Wherever possible, evaluations of
accuracy of--and estimates of time
consumed by--the various operations
are given.
Figure 1, --Location of study area.
PRELIMINARY PROCEDURES
To be successful, a water resource inventory using aerial photographs must be planned
and organized to make the best use of available funds and manpower. The amount and kind of
detail to be collected must be decided upon. Selection and training of the photo interpreter re-
quire a varying amount of advance preparation, depending on the level of intensity of the study.
THE DRAINAGE BASIN UNIT
The drainage patterns of mountains greatly influence the development of timber and other
resources, as well as water. Water resource data are most useful to administrators when
mapped and presented by units that correspond to drainage basins. The acreage included in any
particular drainage unit is not of great importance, but the unit boundaries should coincide as
nearly as possible with the boundaries of working units established for timber, forage, and other
resources. This organization allows comparison of resource values common to a specific area.
Fishery biologists, foresters, water engineers, highway planners, and land managers can adjust
their specific development plans to avoid damaging other resources.
Ridgelines, divides, and saddles form the natural separation between drainage units and
are logical choices for boundaries. They have the advantage that they can be easily identified
without costly surveying on maps and photos as well as in the field. The boundaries of drainage
units should be marked on a map of adequate scale; one-fourth inch to the mile is usually suf-
ficient. The same boundaries should be marked on a series of maps of larger scale, such as
1 to 2 inches per mile. These are used to show the position of lakes and streams as the
inventory progresses.
On both the large and small scale maps, the drainage basin units should be systematically
numbered for identification and reference of inventory data.
PURCHASE AND PREPARATION OF AERIAL PHOTOS
When the boundaries of the area to be studied have been determined, the photos may be
ordered. Median scale (1:20,000) aerial photographs are readily available at moderate cost
and cover most of the United States. Such photos are usually taken with panchromatic film and
a filter to remove blue light, permitting moderate penetration of haze. They can be purchased
from the U.S. Agricultural Stabilization and Conservation Service or from the U.S. Forest
Service. Index sheets show the photographic coverage available over any specific area, and
the required prints should be ordered from these sheets by individual photo number.
In aerial photography, images of all terrain features visible on one photo are duplicated on
one or more adjacent photos; that is, there is a large amount of overlap between adjacent photos.
Therefore, it is particularly important that the effective area of each photo be marked before
the inventory is begun, so that no water areas are either omitted or inventoried more than once.
The effective area is an approximately rectangular area in the central portion of each photo.
Procedures for delineating effective area are given in appendix A.
Because the scale on an aerial photo varies with the height of the camera above ground
level, it not only changes from one photo to another, but from one portion of a photo to another,
even within the effective area. Changes in scale are particularly great in mountainous regions
where ground elevation changes abruptly and where conditions require frequent changes in flying
height. Unless the photo interpreter recognizes scale variations, he will get erroneous impres-
sions or measurements. For this reason, care must be taken to determine the scale of each
aerial photo to be used in the inventory. This is not necessarily a difficult or time-consuming
procedure, as convenient tables may be worked out to simplify the calculation. Detailed dis-
cussion of procedures for scale determination may be found in "Manual of Photographic Inter -
pretation.""* In general, scale determination requires the use of accurate contour maps from
which the elevation above sea level can be read for specific points on a given photo.
* American Society of Photogrammetry. Manual of photographic interpretation. 868 pp.,
Washington, D.C., 1960. See chapter 3, Fundamentals of photo interpretation.
QUALIFICATIONS AND TRAINING OF THE PHOTO INTERPRETER
The efficient use of aerial photography for a water resource inventory requires that per -
sons assigned the duty of interpretation have some technical knowledge of photogrammetry and
some training and experience in actual interpretation. In any practical situation, the success
of the inventory will depend largely on the skill of the interpreter. Every effort should be made
to secure a qualified person or to give adequate training to those who are assigned to the work.
Those without general aerial photogrammetric training may refer to the publications on this sub-
ject by Moessner~ and Moessner and Choate. Most colleges and universities offer courses in
photogrammetric techniques.
A primary requisite for the photo interpreter is the ability to use the stereoscope. Normal
vision in both eyes is therefore required. Almost all the work with the photographs is done with
the three-dimensional image provided by stereoscopic viewing. In this, as in all skills, experi-
ence is valuable in that it allows the user to produce results faster and more accurately. Time
spent in practice with the stereoscope is a necessary part of training.
In addition to specific training and experience with the photos themselves, the interpreter
should have field experience. He should have the opportunity to make a few comparative ground
measurements of features typical of those he will be expected to recognize or measure on photos.
He should travel through the inventory area with the aerial photos and make direct comparisons
between the photo image and the ground situation. In a water inventory, for example, he would
want to study the ground situation in terms of lake depth, stream width, slope of bank, vegetation
type, and similar features.
INVENTORY OF LAKES
The inventory of lakes has two aspects: (1) the determination of the total number of lakes
and their location, and (2) the physical description of each lake. Because this information is
valuable even for small lakes and streams, it is desirable that the inventory be as complete
as possible.
RECOGNITION OF LAKES ON PHOTOS
For the most part, lakes are easily recognized on aerial photos. The smooth texture of
the lake surface is generally so different from the surrounding vegetation or ground surface
that the interpreter can instantly pick out the lake images. However, photo images of lakes
vary considerably in tone from deep black to light gray. When light is reflected from a lake
surface, suspended sediment, light-colored vegetation, or lake bottom material, the image on
panchromatic photos appears light gray. But when light is absorbed by deep water, dark materi-
als in the water, or the lake bottom, the resultant tone of the photo image is dark.
=Moessner, Karl E. A simple test for stereoscopic perception. U.S. Forest Serv.,
Central States Forest Exp. Sta. Tech. Pap. 144, 14pp., illus. 1954.
Moessner, Karl E., and Grover A. Choate. Estimating slope percent for land manage-
ment from aerial photos. U.S. Forest Serv. Res. Note INT-26, 8 pp., illus. 1964.
Lake images that are light gray or white seem to blend with the image tones of adjacent
rock slopes, and may easily be overlooked (fig. 2). Stereoscopic examination reveals them,
however, and field investigation shows that such lakes are almost always heavily silted; their
water is opaque, and all underwater detail is obscured.
Occasionally alpine meadows and dry lakebeds have photographic tones similar to those
of opaque lakes. Stereoscopic study of the photo images, however, reveals textural differences
between the water surface of a lake and the vegetation of a meadow. When meadows surround
marshlike lakes, the exact position of the shoreline is difficult to delineate on photos. In such
cases, ground inspection often shows that the area is subject to spring flooding followed by a
lowering of water that exposes meadows during the late summer. Dry lakebeds, on the other
hand, can be recognized by their basinlike conformation.
RECORDING THE LAKES BY DRAINAGE UNITS
The photo interpreter can quickly identify and count all the lakes within each drainage unit.
Any convenient method may be used to identify them if they are not already named. It was help-
ful in this study to number each lake initially by both drainage unit and the number of the aerial
photo on which it appears. For example, lake 3-10-EBW-9-198 would be the third lake in drain-
age unit 10 on photo EBW-9-198. The center of each lake was pinpricked on the photo and the
lake number was written on the back of the photo next to the pinprick. All the lakes in one drain-
age unit were later numbered consecutively on the map and cross-referenced to the photo numbers
for rapid retrieval of the photo on which
a particular lake was shown. These
numbers were used on the drainage unit
map to show the location of the lake.
When large numbers of small lakes less
than an acre each are clustered on the
photo, it may be unnecessary or difficult
to indicate the precise position of each
on the map. In such instances, the total
number of lakes less than 1 acre and of
a single type can be simply recorded for
each drainage unit, and no specific
locations need be shown.
. Lae
Sea
Figure 2,--An opaque lake image, compared
with a clear lake image. Lakes that ap-
pear cloudy on the ground may produce
light images on panchromatic photographs;
such images may blend with the tones of
rock slopes. (Enlarged X2)
DESCRIBING LAKE CHARACTERISTICS
By measurements made entirely on aerial photos, lakes can be described in terms of
(1) type, (2) size, (3) elevation, (4) depth, (5) clarity of water, and (6) shore cover. Descrip-
tion of bottom material, water temperature, forms of aquatic vegetation, and fish life is not
possible from aerial photographs.
The extent to which lakes are studied and classified depends primarily on the intensity of
the survey to be made. In this study, lakes larger than 1 acre warrant description according to
type, acreage, depth, clarity, and shore cover. Smaller lakes usually require type and size
classification only.
Lake Type
Five lake types were recognized in the study area. In other geographic regions, additional
or different lake types may be appropriate. The following type classes were used:
1. Cirque lakes--those in the cirque basins formed at the head of former glaciers.
2. Moraine lakes--those formed behind the terminal or lateral moraines left by former
glaciers.
3. Reservoirs--lakes formed or altered in level by human activity. Reservoirs can be
recognized by the amount of shoreline exposed by fluctuating water levels.
4, Beaver ponds--small impoundments created when watercourses are dammed by
beaver.
5. Potholes--small depressions formed by disintegrating glaciers and retaining less than
1 acre of water.
The last two lake types are the most difficult to identify because they are frequently simi-
lar in size. However, it is desirable to distinguish between them, because beaver ponds are
not so permanent as other lakes (spring runoff frequently breaks beaver dams), and because
they have different recreational values. Beaver ponds provide fishpools, and the higher water
temperatures in the ponds usually allow fish to multiply more rapidly than they generally do in
the cold mountain streams. On the other hand, potholes may not support game fish, but they
occur in significant numbers and often serve as reflection pools, thus helping to create attract-
ive surroundings for hiking and camping. Some pothole lakes are occupied by beaver, and
beaver houses may be visible on photos. Because such impoundments are not created by beaver,
they should be classified as potholes rather than beaver ponds.
In general, beaver-formed ponds are associated with flowing water of small streams
(fig. 3). They tend to photograph in medium or light tones of gray, and can generally be recog-
nized by the characteristic fan-shaped pool behind the dam. A ripple of "white water" flowing
over the crest of the dam is sometimes visible on the photo. Recognizable beaver houses may
provide an additional clue. In contrast to beaver ponds, potholes usually occur in groups in the
glacial dumps of terminal and lateral moraines or in glacier-scarred basins. These depres -
sions often contain substantial amounts of organic matter, and the ponds thus tend to photograph
in darker tones than do beaver ponds.
“
as
eae
S ‘
%
os
4
a
es
Figure 3. --Beaver
ponds on aerial
photo. White in
upper pond is a
beaver house.
re
2s:
Sta
Lake Size
Description by class. --Lakes of different sizes have different possibilities for recreational
use. Therefore classification according to surface area is recommended. The following five
classes have been found useful for mountain lakes:
Class 1--less than 1 acre
Class 2--1.00 to 4.99 acres
Class 3--5.00 to 9.99 acres
Class 4--10.00 to 19.99 acres
Class 5--20 acres or more
The size class of a lake can be estimated directly on a photo by means of a lake size guide,
a transparent overlay printed with concentric circles that correspond to the lake size classes at
the average scale of the photos being examined. By placing the guide over a lake image on a
photo, the interpreter can rapidly select the size class most nearly matching the lake surface
area. It should be remembered that division of lakes into general size classes with the guide is
merely an approximation of acreage. The acreage of a long, narrow lake, for example, cannot
be accurately measured with a circular guide, but a reasonable estimate can be quickly made.
Description by acreage. --When it is deemed necessary to describe lake size more pre -
cisely than by classes, each lake image can be measured with a dot grid. This grid is a trans-
parent overlay covered with equally spaced dots, each of which represents the center of a small
square. For most lake measurements, a dot grid having 256 dots per square inch is adequate.
Where added precision is desired, a microdot grid having 1,024 dots per square inch can be used.
The photo interpreter can rapidly compute the lake area by the following procedures:
Step 1.--Determine the exact scale of the photo at the lake surface elevation.
Step 2.--Determine the acreage per dot. For example, 1 square inch of photo surface
at a scale of 1:19, 000 is:
19,000 X 19,000
361,000,000 square inches
361,000,000 + 144 = 2,506, 944 square feet
2,506, 944 + 43,560 = 57.55 acres
Consequently, for a dot grid with 256 dots per square inch, the acreage per
dot for a scale of 1:19,000 is 57.55 + 256 = 0.225 acre.
Step 3.--Count the dots superimposed on the photo image of the lake. This should be
done while the photos are being viewed stereoscopically.
Step 4.--Multiply the number of dots counted times the acreage per dot.
Lake Elevation
The approximate elevation of a lake can be determined once the lake position is indicated
on the contour map. Elevation is critical in estimating the approximate number of snow- and
ice-free days that may be expected during the year, an important influence on general recrea-
tional use. The duration of the ice-free period also influences the amount of biological activity
(such as fish growth) in the lake.
Depth
In this study, determination of maximum lake depth from photos could not be done accu-
rately. Instead, photo interpretation was used to determine whether lakes were less or more
than 15 feet deep, and if more, to estimate the percentage of each lake area that was under 15
feet deep. The procedures used were essentially those reported by Moessner ae
Any appropriate depth contour may be used. The 15-foot depth contour was judged to be
a significant lake measurement in the study area for three reasons: (1) It is about the minimum
depth that allows fish in high mountain lakes to escape suffocation and mortality during severe
winters; (2) it approximately defines the "littoral zone" in which light penetration and water
temperature permit the production of food for fish; and (3) it is the maximum depth at which light
penetration allows recognition of underwater detail on panchromatic photographs of clear lakes.
The techniques used for estimating depth on aerial photos are based on the assumption
that the immediate bank slope surrounding a lake continues more or less unchanged for some
distance under water. Thus, lakes with steep banks on all sides are considered more likely to
be deep than are lakes with nearly level banks.
Estimations of lake depth may be based mainly on the interpreter's judgment of steepness
of bank slope or may be calculated from bank slope measurements on the photo. Measurement
is, of course, more accurate, and it allows the interpreter to predict depth at specific distances
from shore.
Moessner, Karl E. Estimating depth of small mountain lakes by photo measurement
techniques. Photogram. Eng. XXIX(4): 580-588, illus. 1963.
8
Depth classification without measurement. --iIn this method, the photos are examined
stereoscopically and a judgment is made on uiuneasured estimates of bank slopes. Other clues
to the slope of the lake bottom are considered, including the general configuration of the lake,
the expected characteristics of lakes of known origin, images of underwater detail, such as
vegetation or rocks, and the lighter or darker tones of different parts of the lake.
In the test study, two experienced interpreters worked independently to estimate the po-
sition of the 15-foot depth line for each lake and to draw it on the photo. Their estimates were
inaccurate for the shallower lakes. None of the four lakes determined from field measurements
to be less than 15 feet deep was correctly classified by either interpreter. Only one lake was
judged to be less than 15 feet deep, and field measurements showed this lake to be actually
deeper than 15 feet. Although depth judgments were often wrong, lakes with steep banks were
seldom misclassified. Training of photo interpreters should include practice in comparing
estimates of bank slopes with parallax measurements of the same slopes.
Depth classification using parallax measurement. --In this method, selected bank slopes
around each lake image are measured, using parallax wedges, as described by Moessner and
Choate.> Accuracy may be improved if the banks most likely to characterize lake depth are
chosen according to indications such as the clues mentioned above. For example, if underwater
vegetation is evident near a steeply sloping bank, that bank is probably not a good choice for
parallax measurement.
The underwater extension of each measured bank is calculated, and the distance from
shore at which a water depth of 15 feet would be expected is calculated. The resulting points
are then plotted on the photo. For example, a 20-percent bank slope (20/100) would result in
a 15-foot lake depth at 75 feet from shore if it continued under water without change:
15 feet © 20
Distance from shore 100
Distance from shore = ee)
= 75 feet
In the test study, three trained interpreters measured bank slopes of 21 lakes on the
photos. Their measurements gave dependable depth information for steep-sided lakes. The
17 lakes that were deeper than 15 feet, as determined from depth measurements made from a
boat, were all correctly identified by parallax measurements. However, this procedure proved
undependable for lakes with the shallowest water and the gentlest bank slopes. Of the 12 depth
estimates made for the four lakes found to be less than 15 feet deep, only three were correct
(25 percent).
Lake size as an indicator of lake depth. --To cut down the time required for depth classifi-
cation using parallax measurements, the relation of lake size to the steepness of general terrain
and lake banks may be used as an indicator of depth. The bank slope of a small lake must be
quite steep if the lake depth is to reach 15 feet or more. Conversely, large lakes are likely to
be 15 feet deep even when their bank slopes are relatively gentle. This general relation between
lake size and minimum average bank slopes may be tabulated for convenience, using general
estimates taken from the photos and based on field experience. For lakes on the North Slope,
the relationships between size and slope were as follows:
See footnote 3.
Lake size Slope+
(Acres) (Percent)
Less than 5 28
5-10 5)
10-15 22
15-20 19
20 -40 15
More than 40 10
+ Average of the two steepest measured slopes.
Such a table may be used to classify lakes into two groups as more or less than 15 feet
deep, on the basis of average bank slope measurement, without actual plotting of the 15-foot
depth contour. Eliminating the plotting of the underwater contour considerably reduces the work
time required per lake. Although the time saving does result in somewhat reduced accuracy,
80 percent of the lakes in the study were classified correctly by this method. Lakes that fail to
meet these minimum criteria of slope and lake size must be measured with additional slope
readings and the 15-foot depth contour must be plotted.
Estimating Shallow Area of Lakes
In a recreation inventory, it is often desirable to know what proportion of a deep lake is
relatively shallow. As mentioned earlier, fish are not likely to survive over winter in lakes
less than 15 feet deep but are largely dependent on food produced in the shallow zone of less than
this depth.
Two methods can be used to determine the proportion of shallow zone area. The first,
contour plotting, requires that slopes be measured on photos, and that the 15-foot depth line be
plotted. The shallow area is then determined by the dot-counting technique described earlier.
The second method, formula computation, still requires bank slope measurements but elimi-
nates the plotting and dot-counting chore by using a formula to compute the proportion of shallow
area. The proportion can easily be converted into acreage if a dot count of the totai lake area
has been made. ‘
Shallow area estimation by plotting contour. --Once the 15-foot depth line has been drawn
on a photo, either from unmeasured estimates or from projection of 5 to 10 measured bank
slopes per lake, it is easy to obtain the shallow zone area of each lake by the dot-count pro-
cedures described earlier.
Contour plotting with parallax measurements was found to be the most accurate of the two
methods when results were compared with field measurements. Without parallax measurements,
only 47 percent (16 out of 34) of the photo estimates of the shallow zone area were within +25 per -
cent of field measurements. However, estimates based on parallax measurements and projec -
tions of bank slopes were within +25 percent of field measurements 80 percent of the time (41
out of 51).
Formula computation of shallow area. --The formula method of estimating shallow zone
area uses 5 to 10 bank slope measurements and projections. The equation is as follows:
10
Proportion of lake in shallow zone = 2D(L_ + W_~ 2D)
LW
where _
D = the average distance from shore to the 15-foot depth line as projected from measured
bank slopes
L = length of lake
W = width of lake.
The formula is derived on the assumption that each lake is a perfect ellipse. Formula esti-
mates of shallow zone area were within +25 percent of field measurements 69 percent of the
time (35 out of 51, table 1, page 21).
Formula -derived estimates of shallow area averaged somewhat less than field estimates.
An adjustment factor may be used to improve accuracy if field measurements are available for
a few sample lakes. The factor is the result of dividing the average field-measured proportion
of total lake area to shallow area by the average formula-estimated proportion. Each formula
estimate is then multiplied by this factor to obtain an adjusted formula estimate. In the test
study, an average factor for the three interpreters was used (1.16). The adjusted estimates
were within +25 percent of field measurements 75 percent of the time (38 out of 51). Because
estimates tend to differ from one photo interpreter to another, slightly better accuracy can be
obtained by using a separate adjustment factor for each photo interpreter.
The accuracy of the four methods may be summarized according to percent of estimates
coming within +25 percent of values based on field measurements as follows: (1) 80 percent
accurate (41/51) --plotting of 15-foot depth contour with parallax measurements; (2) 75 percent
accurate (38/51)--adjusted formula estimation; (3) 69 percent accurate (35/51) --unadjusted
formula estimation; and (4) 47 percent accurate (16/34) --contour plotting without parallax meas-
urement. (The estimates on which these percentages are based are given in table 1, page 21.)
“If the lake is an ellipse, then the shallow zone area equals the total lake area minus the
deep zone area. Thus,
Total lake area = ee
Deep area = gE aes = 2p) wo 2)
Shallow area = total area — deep area
_xLW_ _ x(L — 2D) (W - 2D)
4 4
which reduces to x 4
2xD(L + W — 2D)
4
This must be divided by the total area of the lake to convert it to the proportion of lake in
shallow zone.
_ 2nD(L + W —- 2D)/4 2D(L + W - 2D)
Shallow area aLW/4 Lw
NOTE: In order to convert a proportion obtained by the formula method to acreage of shallow
area, when no dot count of total lake area has been made, the total area can be calculated by
the formula xLW/4.
ll
Clarity of Lake Water
The relative clarity of the water contained in lakes can be judged on the basis of the color
tone of the photo image. Lakes which appear light gray or white (opaque) on panchromatic photos
generally contain suspended solids. It is more difficult to predict the water clarity of a lake
which appears dark on the photo. The water may be clear or may contain varying amounts of
suspended material.
Shore Cover
The type of vegetation surrounding a lake influences its recreational utility. Wood for
campfires is difficult to obtain at lakes without tree cover. Dense brush cover may require
construction of trails. Boggy or marshy areas surrounding a lake make access difficult. For
these and other reasons, it is desirable to indicate the approximate percentage of the lakeshore
that falls in the following cover classes:
1, Forest--stands of trees averaging 20 feet or more in height.
2. Brush--woody vegetation less than 20 feet in height.
3. Bog--wet areas of grasses and sedges that usually appear in dark tones on photos.
4, Meadow--areas of grasses and sedges that are drier than bog and usually appear in
lighter tones on photos.
5. Barren--no vegetation present.
INVENTORY OF STREAMS
Aerial photographs provide far less information for streams than for lakes. Streams less
than 20 feet wide can be easily obscured by tree or brush cover, so that any measurement is
difficult. Depth measurements such as those made on lakes are impossible, even when the
stream is plainly visible. Nonwater features are often mistaken for streams on photos. Live
streams that are not hidden by overtopping vegetation appear as thin and meandering dark lines
(fig. 4). Shadows of steep drainage banks, as well as the outcrops of dark rock strata, can
appear as similar lines.
In spite of these problems, photo interpretation and measurement can be used to obtain
estimates of stream length, width, and gradient, and a general description of streambank cover.
The principal value of the photos in a stream inventory is to correct and supplement the informa -
tion given on available maps of the area. In addition, photos are a logical tool for field location
of sample points if on-the-ground measurement of additional stream characteristics is planned.
If no field measurements are to be taken, aerial photos should be used to improve the
basic description of streams given on the best maps of the area. Most modern maps are made
from aerial photos and show the location of streams fairly accurately. However, because in
general these maps are based on small-scale photos, much detail visible on 1:20,000 scale
photos is not apparent to the mapmakers.
12
aL VE rs
€
Figure 4.--A typical live stream appears on panchromatic photography as a
meandering, thin, dark line.
RECOGNITION OF STREAMS ON AERIAL PHOTOS
Larger streams, which are easily recognized on the photos, are usually correctly mapped.
Many smaller streams, however, particularly in mountainous areas, cannot be identified from a
map alone as flowing, dry, or intermittent. A number of these can be correctly classified from
photos, and substantial savings over field inventory costs can be made.
Where live streams are obscured by vegetation, their presence may sometimes be inferred
from related photo features. A meandering pattern of dark blotches through a forested area may
be made by marshes, springs, seeps, or other riparian vegetation and water. On the other hand,
dry drainage channels usually appear on panchromatic photography in light-colored image tones,
because the deposits of clay, silt, sand, and rock tend to reflect light. Such indirect evidence
as exposed boulders, the absence of riparian vegetation, and lack of springs or marsh areas all
suggest that a channel is dry.
13
Intermittent streams are the most difficult to identify on aerial photos, especially in the
upper reaches where flow is strongly influenced by periodic rainfall. A small channel may con-
tain a live stream on the date of photography and be dry immediately thereafter. Similarly,
field surveys may find water where none was present at the time of photography. Nevertheless,
experience indicates that if the photo interpreter "sees" water in a channel when he is viewing
photos stereoscopically, a field check will usually confirm his judgment. This probability is
raised by the fact that most resource photos are taken during late summer and autumn, when
stream levels are lowest.
When streams are identified, obviously dry channels can be eliminated from the inventory.
Any later fieldwork can be concentrated on streams that remain questionable after the photo
interpretation has been completed.
RECORDING STREAMS BY DRAINAGE UNITS
Most drainage units contain one main channel and one or more channels tributary to it.
Although most of the main stream channels are named, the smaller tributaries are frequently
nameless. For convenience in reference, it is desirable to use some system of numerical and
alphabetical identification. In this study, the main channel was given the same number as that
of the drainage unit in which it was located. Tributary streams were then identified with the
letter ''T' following the unit designation number. The first tributary was labeled TA, the second
TB, and so on, until each flowing tributary was identified. In the example shown in figure 5, the
third tributary in drainage unit 10 is identified as 1OTC.
Figure 5.--Typical drainage unit, showing
pattern and identification of streams.
Tributary stream
AvAnan, Main stream
Unit boundary
14
DESCRIBING STREAM CHARACTERISTICS
Stream Length
Aerial photos are particularly effective for editing maps for stream length. If available,
maps with a scale as large as 1 or 2 inches to the mile are best for use with the photos. End
points of each stream should be located first on the photos. The end points are transferred to
the map as accurately as possible and the apparent length of the stream and its tributaries can
then be measured on the corrected map. The stream's lower end is defined as the point where
it flows into another stream. The upper end is defined as the highest point in the drainage that
gives evidence of permanent flow of water ne Dry or intermittent channels can be eliminated
so that only the essential portions of any stream under consideration will be included. Dividers,
set to the scale of the map, are used to step off the length at quarter-mile intervals to give the
unadjusted length of the entire stream or any portion of the stream. This unadjusted length
underestimates actual stream length, because small bends and other deviations in stream di-
rection cannot be accurately mapped. However, the mapped stream length can be multiplied by
a "meander factor" to give an adjusted stream length.
The meander factor is the ratio between the meander distance as visible in the photo and
the straight-line distance for a given stream segment.
The procedure for determining meander distance on the photo is to rule a thin straight
line on a strip of transparent paper and then match this line to short segments of the stream.
Two points approximately a quarter of a mile apart in straight-line ground distance are selected
on the stream image. The paper is placed over the photo so that one end of the ruled line is
over one of the selected points, and a pin is pushed through the paper and into the photo at this
point. ‘The paper is then pivoted so that the line lies over a short section of the stream course.
A second pin is then placed at the point where the stream image diverges from the ruled line.
The first pin is removed and the paper pivoted to coincide with the next section of stream. This
process is repeated until the point a quarter of a mile away is reached. The amount of line used
is then measured, and the meander factor is calculated:
Photo meander distance
Meander factor = ————_———__—————
Photo straight-line distance
Because the value sought is a ratio, any units of measurement can be used on the photo; conver -
sion to ground distances is not required. Several measurements may be made at different points
along the stream photo image and an average meander factor established. Thus a stream with an
average meander factor of 1.470 and an unadjusted mapped length of 8.6 miles would have an
adjusted length of 8.6 X 1.470 = 12.64 miles.
In some instances, a general average meander factor for a particular drainage basin may
be used for all streams in the basin. This saves time without any considerable loss of accuracy.
’ If work on photos is followed by on-the-ground sampling, the upper end of the stream may
be redefined by other criteria. During the fieldwork on the North Slope study the upper end of
each stream was arbitrarily defined as the point where stream width dwindled to 4 feet.
15
Stream Width
Stream width is difficult to measure accurately from photos. Perhaps the most practical
course is to group the stream segments on photos into width classes. This method produces
reliable estimates. Appropriate class limits for small mountain streams are:
1-20 feet
21-50 feet
51-100 feet
101-200 feet
Limits can be established as needed, depending on conditions in the area under study. However,
for evaluation of recreational possibilities of streams, the above classification is generally ade-
quate. If estimates of stream surface area are needed, the adjusted stream length and the
midpoint value for the width class can be multiplied together. Length of segments may be de-
termined in any convenient manner, depending on the general amount of variation in stream width.
Stream Gradient
Gradient can be determined most easily directly from contour maps. If contour maps are
not available, elevation differences may be determined by parallax wedge measurements on
aerial photos. Once the adjusted length for the entire stream has been determined, the gradient,
in percent, can be found by dividing the elevation difference between two points by the adjusted
stream length between these points and multiplying the result by 100.
Streambank Cover
Streambank cover can be recognized by an experienced photo interpreter and can be
classified for stream segments as:
1. Forest--stands of trees averaging 20 feet or more in height.
2. Brush--woody vegetation up to 20 feet in height.
3. Open--no shrub or tree cover recognizable on photos.
If vegetation cover maps have been prepared for other purposes, they may provide ade-
quate cover information without further photo work.
LEVELS OF INTENSITY OF AN INVENTORY
The procedures in a water resource inventory may be combined in several ways to make
up inventories of varying degrees of intensity.
LAKE INVENTORY
In the early planning stage of the inventory, some decision will probably be made as to
the amount of information desired; however, in the course of the study, it may become advis-
able to step up or step down in intensity. Particularly promising lake areas discovered in the
course of a low-intensity study may warrant closer examination and measurement. In a high-
intensity study, a group of lakes of similar nature may require only sampling.
As a guide in the choice of level, information is needed on the time required for each
procedure. In the present study the following evaluation has been made:
16
Approximate time
Procedure required per lake
Lake size determination:
By class, with lake size guide 1 minute
By area, dot count + hour
Determination of elevation, type,
clarity, shore cover 5 minutes
Depth determination:
By plotting without measurement 5 minutes
By plotting with parallax measurement + hour
Determination of shallow area:
By plotting without measurement,
including dot count + hour
By plotting with parallax measure -
ment, including dot count 14 hours
By use of formula 4 hour or less
Three possible combinations of procedures, with time approximations for photo interpre -
tation, are as follows:
1. Low-intensity photo inventory (average of 10 minutes per lake).
a. Inventory all lakes and ponds within each drainage basin unit by lake type, size
class, elevation, water clarity, and shore cover.
b. Use unmeasured estimates of bank slope, the lake size guide, and a table of mini-
mum bank slopes to classify all lakes larger than 1 acre as either "probably deep" (more than
15 feet deep) or "probably shallow" (less than 15 feet deep).
2. Medium-intensity photo inventory (average of 45 minutes per lake).
a. Inventory all lakes and ponds within each drainage basin unit by lake type, size
class, elevation, water clarity, and shore cover.
b. Use a combination of unmeasured and measured estimates of bank slopes, the
lake size guide, and the table of minimum bank slopes to classify all lakes larger than | acre
as either "deep" or "shallow."" Lakes having bank slopes that are judged to exceed the mini-
mum required by the table need not be checked with parallax measurements and may be classi-
fied as ''deep."’ For all other lakes, take parallax measurements of the one or two bank slopes
considered most indicative of lake depth. Classify lakes as "deep" if slopes equal or exceed
minimum values in table, "shallow" if slopes are less than these table values.
c. Use parallax measurements and the formula for calculating shallow area of lakes
that seem to justify further evaluation because of size, location, or depth.
17
3. Full-intensity photo inventory (1 hour and 30 minutes per lake).
a. Inventory all lakes and ponds within each drainage basin by lake type, size class,
elevation, water clarity, and shore cover. Measure area of all lakes over 1 acre in size by dot-
count procedures.
'
b. Classify all lakes over 1 acre in size as ''deep," or "shallow." Classify as "deep"
if measured bank slopes exceed the table values, "shallow" if they do not.
c. For all lakes classified as ''deep,"’ plot the 15-foot depth line from 5 to 10 meas-
ured bank slopes and compute the shallow zone area by dot-count procedures.
Any further details unobtainable by a full-intensity inventory must be gathered on the
ground. In the Uinta Mountains, field measurements cost approximately seven times as much
per lake as full-intensity photo measurements. However, they may often be justified after photo
techniques have been used to identify the lakes most likely to offer outstanding opportunities for
recreation.
STREAM INVENTORY
The primary objective of a stream inventory from aerial photos is to evaluate the stream
resource of each drainage unit in terms that allowcomparison with the stream resource in other
units. If a higher intensity inventory is desired, however, the same techniques can be used to
describe individual segments of a stream, so as to allow comparison of different parts of the
same stream. For example, streams can be described in 1-mile segments, permitting the
identification of unique or separate sections of a given stream.
The time requirements for stream inventories cannot be estimated as easily as those for
lake inventories, because streams vary greatly in length. However, as a rough guide, most
streams can be adequately classified in 4 hours. The description of a particular segment of a
stream will take between 2 and 3 hours.
18
APPENDIX A
DETERMINATION OF EFFECTIVE AREAS ON AERIAL PHOTOS
Delineation of effective areas on aerial photos requires marking off the boundaries of the
area on each photo that appears on that photo alone. Overlap occurs perpendicular to the flight
line and between lines of flight. The procedure is as follows:
Marking boundaries perpendicular to the flight line--endlap.
Step 1. --Taking the first two photos in a flight line, place photo 1 over photo 2 so that
images common to the two photos are approximately superimposed. Then, rule a straight line
(in ink or crayon pencil) on photo 1 so it approximately bisects the area of endlap and is approx-
imately perpendicular to the line of flight. This line should pass through two easily recognized
image points, several inches apart on the photos, and representing the highest points of topog-
raphy in a position suitable for the ruled line.
Step 2.--Duplicate this ruled line on photo 2 by drawing it through the images of the same
points on this second photo. This duplication is simpler if the photos are examined in stereo-
vision. If both lines pass through image points representing high points in the topography, then
no photo detail will be excluded. Some detail at low elevations may be duplicated within the
effective area of the two photos, but if ruled lines pass through high points, effective areas can
easily be separated and adjustments made under stereovision.
Step 3.--Repeat this procedure for all photos in a single flight line, pairing photo 2 with
photo 3, 3 with 4, etc.
Marking boundaries between lines of flight--sidelap.
Step 1.--Place the first photo in one flight line over the first photo in an adjacent flight
line so that images common to the two photos are approximately superimposed. Then, rule a
straight line on the top photo so it is approximately parallel to the flight line and approximately
bisects the area of sidelap. If possible, this line should pass through two prominent image
points. Draw this sidelap boundary only until it meets the endlap boundaries drawn previously.
Step 2.--Duplicate this ruled line on the bottom photo. Since stereovision of images from
adjacent flight lines is difficult or impossible, duplication of the ruled line must be guided by lo-
cation of image points crossed by the line on the top photo. It is convenient to hold the two photos
on a desk top in overlapped position with one hand. The other hand can be used to pull up one
edge of the top photo so that images common to the two photos can be identified and marked.
Step 3.--Repeat the procedure for all other sidelapping photo pairs. The boundaries
ruled for each endlap and sidelap area will define the effective area on each photo.
9
APPENDIX B
The guide is prepared by drafting concentric circles to represent areas of 1, 5, 10, and
20 acres at the average scale of the photos being used. Radii for the circles are easily calcu-
lated from (1) the formula for the area of a circle (A = xr), and (2) the average photo scale.
For example, the radius of a l-acre circle is calculated thus:
Avi=ene= = 43,560 square feet
Is
tl
43,560 + 3.1416 = 13,866 square feet
118 feet
RK
ll
This can be converted to distance on the photo itself by the relationship:
Radius on photo
>———_——— = Photo scale
Radius on ground
For a l-acre circle at a photo scale of 1:20,000 this becomes
Radius on photo ll
118 feet ~ 20,000
Radius on photo = 118 feet + 20,000 = .0059 foot
Radii for a lake size guide applicable to 1:20,000-scale photos are as follows:
Area Radius on photo
l acre 0.0059 foot
5 acres -0132 foot
10 acres .0186 foot
20 acres .0264 foot
A simple method of putting the circles on transparent material is to photograph them and pre-
pare a positive transparency.
20
APPENDIX C
Evaluations of the accuracy of the
photo interpreters’ conclusions as to
depth and shallow area have been re-
ferred to in the text. Table 1 gives
detailed data from which the percentages
were drawn, covering the estimation of
shallow area by four different methods.
Note that somewhat consistent variation
is evident in the conclusions arrived at
by the three interpreters. Any given
interpreter may have a tendency to under -
estimate or overestimate; in preliminary
field training this might well be noticed
and thus be taken into account in actual
survey work. Also, because of their par -
ticular characteristics, certain lakes
were overestimated or underestimated by
all interpreters.
The methods are as follows:
Method A: 15-foot depth contour
plotted on basis of vis-
ual estimates; shallow
area acreage obtained
from dot count.
Method B: 15-foot depth contour
plotted on basis of par -
allax measurements of
bank slopes; shallow
area acreage obtained
from dot count.
Method C: Average distance from
shore to 15-foot depth
contour calculated from
parallax measurements;
proportion of shallow
area computed by ellipse
formula; acreage cal-
culated on basis of total
lake acreage obtained
from dot count.
Method D: Shallow area computed
as in method C; result
adjusted by correction
factor of 1.16.
Table 1,--Estimates of shallow area of lakes made by three photo
interpreters, using four methods, as compared with shallow
area measured in the field (North Slope, Uinta Mountains)
Lake no. : Measured :
Shallow area estimated by:
and total : shallow = :—WW—W——*____
seneage agcat : Method A : Method B : Method C : Method D
----- ee ee eee Acres- ---------------
INTERPRETER 1
Le 20! l.79 117 138% 1.56* 1.817
25 3207 1375 178% 2<L4* 1.85* 2.15*
3. 4.51 2.83 1.61 3.13" 2..76* 3,20*
4. 6.64 5.16 4.09* 4.80* 3.57 4,14*
5. 7.44 6.04 712" 4,83* 4.80* 5.57%
6. 8.78 6.00 3.10 58% 5. 15* 3897"
7s 9.83 8.06 6.54* 5.02 5.57 6. 46*
8. 17.61 8.98 11.04* 8.51* 9, 26* 10.74*
9. 20.30 18.67 16.88* 13.19 ib at 13.58
10° 22.3 6,68 11.00 5.84* 5.69* 6.60*
11. 29.05 20.74 15.19 17.94% 11.59 13.44
12... 32.36 17.74 13.67* 16.70* 13.01 15.09*
13. 32.44 15.63 10.53 16.43* 15.12% 17.54*
14, 40.30 26.81 29.40* 26 .44* 18,30 Zi 23"
15. 40.80 18.89 20.28* 25.56 19.50* 22.62"
16. 41.12 16.29 24.20 17 L1* 11.23 13..03*
17, 63.91 16.18 87.91 21.96 18.34* 24327
INTERPRETER 2
Le. 2.01 iD 2.41 1.46* 1.70% 1,97*
2, 3567 1:75 2.16* 1.81* Loe 2.0)"
3. 4.51 2.83 4,26 3.10% 3.20* ey fl
4. 6.64 5.16 4,72* 5.28* 505% 5.86*
5. 7.44 6.04 6.62* 4..82* 4.62* 5.30"
6. 8.78 6.00 692* 7.90 8.56 9.93
7. 9.83 8.06 9221 6.29% 6.46* 7.49*
8. 17.61 8.98 12.67 8.81* 10. 14* 11.76
9. 20.30 18.67 19.76* 14°9/* 15.27" 17 571%
WO. 22531 6.68 17.93 6.42* 6.02* 6.98*
11. 29.05 20.74 Zoie2o™ 22.96% 21327 24.73*
12. 32.36 17.74 29.13 16 .20* 11.20 12.99
13. 32.44 15.63 25.71 13.99* 12. 36* 14, 34*
14, 40.30 26.81 40.30 29. 30* 20, 39* 23.65*
15. 40.80 18.89 39.41 19:,.31* 15.95* 18:50*
16. 41.12 16.29 34.04 17533" 14. 10* 16. 36*
175-6369: 16.18 46.68 2211 18.53* 21,49
INTERPRETER 3
iy 2.51 1.75 obs 1.40* 1,22 1,42*
2. 3.67 L725 -- IEPA 2.07* 2.40
3. 4.51 2.83 -- 2. 70* 2.02 2.34*
4, 6.64 5.16 -- 4,91* 3.76 4,36*
5. 7.44 6.04 -- 6 .69* 4,97* o.77*
6. 8.78 6.00 -- 6.29% 5.64* 6.54*
7s, 9283 8.06 -- 6.60* 6.81* 7.90*
8. 17,61 8.98 -- 10.,56* 9.76% 11,32
9, 20.30 18.67 -- 16 .28* 13.44 15.59*
10, 22.31 6.68 -- Dsao* 6.31* Tae.
11, 29.05 20.74 -- 20.39% 16.18* 18.77*
12... 32.636 17.74 -- 16 .93* 11.65 LZisou*
13. 32.44 15.63 -- 19.02* 19.07* 22 512
14, 40.30 26.81 -- 27.60* 12.78 14.82
15. 40.80 18.89 -- 25.87 16.48* 19,12*
16, 41.12 16.29 -- 20.77 16,78* 19, 46*
17, 63.91 16.18 -- 28.62 22.50 26.10
Total number of
estimates (*) falling
within +25 percent 16/34 41/51 35/51 38/51
of measured shal-
low area
Obtained from photos by dot-grid method.
? Based on soundings taken in field and dot count.
° Method A was used by two interpreters only.
21
AFPS, Ogden, Utah
sent a aige GS.
} = J ier a
, ¥
os ty .
ae a) Pee i ep
= { hr ee : 5
s * ll aid i nl >,
Veg RNa ATO EAT it whip
i 2 RET ey ROR
ot
, ©
&
‘a
Wier
; a
i
z > : K
be i
*
*poute[dxs ore
AJOWWSAUT WIedTIS pue OYe]T JNpUOd 0} papsasu sJUSWIOINSedUI
ojoyd [Te IOJ soernpasozg ‘yadep oye, Jo sqewasnseoul pray
ywM yidop aye] jo uolyeurwazajep ojoyd jo Aoeinooe sazeduioy
‘suleotis pue SOY] ULeWMNOU Jo SOTISTTOIOeATeYO oY} sINseoU
0} YeINQ WI peso? sanbtuysa} oJoyd Telree Jo s}fNsea saqirosaq
(LE-. LNI ‘deg ‘soy ‘Aras jsatoy *S*n) “snip ‘dd 1Z
"yeig ‘uepso ‘*eis “dxq oasuey pue isoTOy utTenoul
-ZTawy] ‘°ATaS IsatTOWY °S*Q “sweets pue soyYey] ureqnou!
Jo AZOJUSAUT UOTIVaTIEI & TOF Sonbruysa} ojoyd [eIIey *LO6T
*ZayIO |, SSOY °S pue ‘*g s0dSsOy ‘uOISUTTIOY
*poute[dxo oie
AZOJWSAUL WiedTIS pue dsYeT JONpUOD 0} popsou sjUaWIeINseoUI
oloyd [[e@ Toy sainpasorg “yidep aye, Jo sywowornseoul pTery
yim yidop oye JO uoTIeuluITajap ojoyd Jo Aoeainooe sazreduioyg
*‘suleolis pue soye, uleUNOW Jo SOI]STIOJOeAeYO oY) sAinseoauI
01 YeIQ UI poise} sonbruysea} ojoyd Telree Jo si{nser saqiiosaq
(LE- LNI “deg ‘soy ‘Artes ysetoy *S'n) ‘snqt ‘°dd 1z
‘ye1Q ‘uepso ‘*e1Sg *dxq o8uey pue isotOJ uleqnow
-ioaqyuy ‘*ATaSg isatoy °S*°Q *swesarjs pue soyey urequnoul
Jo ALOJWOAAUT UOTIVATIEI & IOJ Sonbruysoay ojoyd [ertey °/96T
*Zeyoo0 |, SSOYy *S pue ‘*g so0dSOY ‘uo sUTTIOY
*poute[dxe o1e
AZOWWSAUT WiedITIS pue dye] JONpuod 02 papsou sj wauUloInseoUr
ojoyd {[@ Joy soinpsdsoig *wdep oye] Jo sjusuloIMseau pyaty
{WIM yap seye, jo uoljeurwizajaep ooyd jo Aoesmmooe sazreduiog
*suleot}]s pue Soye]T UTeWNOUW Jo SOTsIX9NJDeTeYO oY] sINsPoUI
0} YeIQ UI paisa sanbruyse3 ojJ0yd [elroe Jo si~nsaez saqt1rosaq
(Z€-. LNI ‘deg ‘soy *Atas isaztoy °S*n) ‘snit ‘dd 1z
"yeIn ‘uspsQ ‘*e1g ‘dxq osuey pue isolOy uTeInoW
-Zawy] ‘*ArTaS isetOy °S°Q “swears pue soyeyT uTeqnou
Jo ALOWOAU UOTIVdIIEI & TOJ sonbruyse ooyd [eIey */96T
*Zaydo |, ssoy *S pue ‘*g s00soy ‘uoISUTIZAaH
*poutetdxs ore
ALOWWSAU WiedTIS pue sye] JONpuoD oO} papsau sIUSWIDINSeoUI
ojoyd [Te Joy sainpasoig *ydep oye] Jo syuowornsesew pyaty
yim yydep oye] jo uoTyeulwxza3ep ojoyd jo Aoeinoose sareduiog
“sweats pue SOXeT UleJWNOW Jo SoOMsTIaOeITeYO oy oAnseoUI
0} YIN UI poise? sonbruysei oJ0yd J[elae Jo si~nser saqti9seq
(LE- LNI ‘deg ‘soy *Ateg jsotog *S*n) ‘snip ‘°dd 1z
"yeIn ‘uepsQ ‘*e1S “‘dxq osuey pue jsoTO4 UTeWnoU
-Zo Wy] ‘*ATOS saIOY °S°Q «*SsuieaI}s pue soyey] uUTeIMoU!
Jo ATOWOAUT UOTJGIIIOI & TOF Sonbruyse3 ooyd Tetzey °/96T
‘Tayo |, SSOyY °S pue ‘°g s0DSOyY ‘u0IsUTTIDH
‘ ton
0
a SA a a ey at
Headquarters for the Intermountain
Forest and Range Experiment Station
are in Ogden, Utah. Project headquar-
ters are also at:
Boise, Idaho
Bozeman, Montana (in cooperation
with Montana State University)
Logan, Utah (in cooperation with
Utah State University)
Missoula, Montana (in cooperation
with University of Montana)
Moscow, Idaho (in cooperation with
the University of Idaho)
Provo, Utah (in cooperation with
Brigham Young University)