Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

United States
~\ Department of
} Agriculture

Forest Service

Pacific Northwest
Research Station

Research Note
PNW-RN-457

January 1987

ig

Abstract

Introduction

Pecky Rot in Incense-Cedar:
Evaluation of Five Scaling
Methods

James M. Cahill, W.Y. Pong, and
D.L. Weyermann

A sample of 58 logs was used to evaluate five methods of making scale deduc-
tions for pecky rot in incense-cedar (Libocedrus decurrens Torr.) logs. Bias and ac-
curacy were computed for three Scribner and two cubic scaling methods. The
lumber yield of sound incense-cedar logs, as measured in a product recovery
study, was used as the basis for comparison. Results showed that the most biased
and least accurate scaling systems deducted the greatest percentage of log
volume.

Keywords: Log scaling, defect deduction (-merchantable volume, pecky rot, lumber
recovery, incense-cedar, Libocedrus decurrens.

Incense-cedar (Libocedrus decurrens Torr.) is an important commercial softwood
species in the Western United States. Currently, there are about 14 billion board
feet of commercial size incense-cedar in the West; a majority of this volume is in
northern California and southern Oregon (USDA Forest Service 1973b). Incense-
cedar is manufactured into a variety of products including lumber, pencil stock,
fenceposts, and shakes.

Although wood products manufactured from incense-cedar are recognized as ex-
tremely resistant to decay, standing sawtimber is highly defective. Pecky or dry rot
(Polyporus amarus Hedgc.) causes extensive heart rot in trees throughout the
range of this species (Boyce 1920, Wagener and others 1958). In the early stages
of decay, pecky rot occurs in small scattered pockets confined to the heartwood
(fig. 1a). As the decay intensifies, the number and size of the decay pockets in-
crease and eventually coalesce (fig. 1b). The varying amount and scattered occur-
rence of pecky rot create problems for scalers trying to estimate the net volume of
wood available for lumber production. Scalers often ask whether the scaling rules
used to deduct for pecky rot are accurate: Do specific scaling rules deduct too
much or too little for the presence of peck?

The purpose of this paper is to report the results of an analysis of five ways to
make deductions for pecky rot. Three methods of Scribner deductions and two
methods of cubic log scale deductions were analyzed for their ability to-estimate
the volume of wood available for lumber production. The scaling systems were
compared by ranking the estimates of bias and accuracy.

JAMES M. CAHILL is a research forester, WY. PONG is a
research forest products technologist, and D.L. WEYERMANN is
a computer programmer, Pacific Northwest Research Station,
Forestry Sciences Laboratory, PRO. Box 3890, Portland, Oregon
97208.

Methods

Data Base

Figure 1—Incense-cedar log ends: (A) Scattered pockets typical
of the early stages of decay. (B) Large, numerous decay pockets
typical of advanced decay.

The general approach used to analyze the scaling options was to compare the
lumber yield of logs containing pecky rot to the lumber yield of sound logs. Lumber
yield and log scale data measured at a product recovery study were used. Fahey
and others (1981) and Cahill (1983) used the same approach to analyze other scal-
ing problems. Log scale deductions are considered accurate if the lumber yield
based on the net volume of a log with peck in it is the same as the lumber yield of
a sound log of equivalent volume.

The data used to conduct this analysis were collected during a product recovery
study in 1982. A total of 130 incense-cedar trees were selected from six areas in
the Eldorado National Forest in California. Trees were selected to capture the range
of quality and sizes that exists in incense-cedar. Diameter at breast height (d.b.h.)
for the sample ranged from 16 to 54 inches. The pecky rot included in the sample
represented the range of severity that would be seen in normal scaling operations.

Alternative Scaling
Techniques

Certified USDA Forest Service scalers recorded Scribner (USDA Forest Service
1973a) and cubic (USDA Forest Service 1978) scales on the woods-length logs prior
to sawing. Deductions by scaling defect were recorded for each log and were based
on all scaling methods. The logs were processed in a sawmill where moulding, shop,
and board items were cut. Study lumber was tallied for individual logs after drying
and surfacing. The percentage of lumber manufactured by grade group is shown:

Lumber grade Percent
Moulding 10
Shop grades 36
3 & Better Common 29
4 Common 15
5 Common 10

Deductions for pecky rot were made using the following Scribner and cubic scaling
rules:

1. Scribner standard scale. Scale was taken using the standard National Forest log
scaling handbook rules for Scribner scale (USDA Forest Service 1973a). All peck,
regardless of the arrangement on the log ends, is deducted. Since this study was
completed, Forest Service scaling practices have changed. Handbook rules now
state that only pockets of peck less than 5 inches apart are considered as a scal-
ing defect.

2. Scribner 13-inch rule. A deduction for pecky rot is made only when there is less
than 134 inches of solid wood between pockets of rot as the pockets appear on the
log end. For example, deductions would be made on a log where the pockets of
peck were 1 inch apart. No deductions would be made if the peck holes were 2
inches apart.

3. Scribner 4-inch rule. A deduction for pecky rot is made only when there is less
than 4 inches of solid wood between pockets of rot.

4. Cubic, total deduction. The cubic volume of rot is estimated using Smalian’s for-
mula. Diameters of the column containing the scattered peck are measured on
both ends if possible. If peck shows on only one end, length of the rot column is
estimated by the scaler.

Example: If a rot column of 10 inches shows only on the large end of a 33-foot log,
the scaler would estimate that the rot extends 8 feet into the log and the peck
deduction would be:

Peck volume = (10? + 107) x 8 x (0.0027274) = 4.4 cubic feet (CF) ;

where:

10 and 10 = the diameters (in inches) of the rot column,

8 = the estimated length of the rot column (in feet), and
0.0027274 = aconstant used to convert to cubic feet.

Analysis Procedure

5. Cubic, soft deduction. Under this method the scaler estimates the volume of
wood fiber that is decayed and not available for lumber production. If the scaler
estimates that only 50 percent of the rot column of the log in the example above is
in soft, unusable wood, then the cubic soft deduction would be 2.2 cubic feet (4.4
x 0150)?

The first step in analyzing the scaling options was to sort the logs into two groups:
Sound logs and logs with only peck as a scalable defect (pecky logs). Logs with
defects other than pecky rot or logs with multiple defects were not included in the
analysis. The following tabulation shows the number of logs, the average scaling
diameter, and the range in diameters for the two groups of logs:

Number Average Range in

of logs diameter diameters

(Inches) (Inches)
Sound logs 185 12 6-31
Pecky logs 58 19 8-33

The second step in the analysis established the yield of lumber from sound
incense-cedar logs. This provided a basis for comparing the different scaling op-
tions. Lumber yield of the sound logs is represented by the regression of rough
green lumber tally in cubic feet over gross Scribner and gross cubic scales (figs. 2
and 3). The relation for gross cubic scale volume was simple linear function (fig. 3);
a segmented regression was needed to adequately describe the relation between
Scribner scale and cubic lumber tally (fig. 2). The segmented regression was neces-
sary because Scribner scale underestimates the volume in small logs; hence, the
rate of lumber recovery varies depending on log size. The regression was seg-
mented at 240 board feet after we examined summary statistics (R@ and standard
error of the regression) for several models with different break points.

The final step in the procedure compared the actual lumber tally of a pecky log,
after scale deductions, to the recovery of a sound log of the same net scale. The
closer the lumber recovery of a defective log is to a sound log of equivalent volume,
the better. Lumber yield for sound logs was estimated from the regression of lumber
tally over scale volume (figs. 2 and 3).

150

125

100

~“N
oa

Lumber tally (cubic feet)
oO
(e)

NO
oOo

O 500 1000 1500 2000
Gross Scribner scale

Figure 2—Regression line and data scatter of lumber tally (cubic

feet) versus gross Scribner scale for sound incense-cedar logs.

The regression equation is:

LT = 8.19 + 0.0807 * S - 0.0282 * | * (240 - S) ;

where

LT = lumber tally,

S = Scribner log scale; and

| = an indicator variable (| = 1 when S < 240 and! = O whens > 240).
R? = 098; N = 185.

150

~ 125/-

t)

(ch)

Ol

fe}

= |

(3)

> 75

a SO)

2

=

= |

ay BS
0

0) 50 100 150 200 250

Gross cubic scale

Figure 3—Regression line and data scatter of lumber tally (cubic
feet) versus gross cubic log scale for sound incense-cedar logs.
The regression equation is:

LT = -1.526 + 0.635 * CS ;

where:

LT = the lumber tally, and
CS = the cubic log scale.
R? = 0.99; N = 185.

300

An objective evaluation of the scaling options was made by comparing the average
bias and accuracy for the 58 pecky logs and by using the lumber recovery of the
sound logs as a base for comparison. Bias is defined as the average deviation of
the actual lumber tally of a pecky log from the predicted tally of a sound log of the
same net volume.’ A system with little bias (bias close to zero) indicates that, on
the average, the deduction is placing pecky logs on the same lumber yield basis
as a sound log. Accuracy is defined as the square root of the average squared
deviations2 and is a measure of the variability of the yield of lumber from the
pecky logs around the sound regression line. The larger the number associated
with accuracy, the greater the variability. When the scaling options are evaluated,
both bias and accuracy should be considered because a system that is both un-
biased (a number close to zero) and accurate (a small number) is desirable.

Table 1 shows an example of the calculation of bias for a hypothetical incense-
cedar log. It was assumed that pecky rot is the only defect in the log. Only the
three Scribner scaling options are used in this example. The net log scales are dif-
ferent because each scaling option used different rules to estimate the amount of
defect. The yield of lumber from a sound log for each net scale estimate was taken
from the regression line in figure 2. Actual lumber tally would be the cubic feet of
lumber measured at a product recovery study. Bias is the actual yield of lumber
subtracted from the estimated yield of a sound log. In this example, the Scribner
4-inch rule worked best because that bias was closest to zero (+2.6). Accuracy was
not computed in this example. The average bias and accuracy are used for large
samples of defective logs to make comparisons among scaling systems.

Estimate of bias is the average deviation where the deviation
is actual recovery from a defective log (y;) minus the predicted
recovery of a sound log of the same volume (y;); that is,

L(y; - yi Mn.
2/Estimate of accuracy is the square root of the average squared
deviation; that is, (L(y;-y, )?/n)’.

Table 1—Example of bias calculation for a hypothetical incense-cedar log

Lumber yield—/

Scaling Net from Actual
option scale a sound log lumber tally 2/ Bias 2/
Board

feet ------- Cubic feet - - - - - - -
Net 134-inch rule 350 36.4 33.0 33.0-36.4 = -3.4
Net 4-inch rule 275 30.4 33.0 33.0-30.4 = +26
Standard Forest
Service scale 200 23.2 33.0 33.0-23.2 = +98

“YLumber yield for a sound log was estimated from the regression equations in figures 2 and 3.
2/Actual lumber tally is the cubic feet of lumber measured at a product recovery study.

¥/Bias is the actual lumber tally, in cubic feet, minus the predicted lumber tally from a sound log.

Results

Table 2 shows the average bias and accuracy and the percentage of log volume
deducted for the Scribner and cubic scaling options.

The 134-inch rule deducted the smallest amount of volume (28 percent) for the
Scribner options and had the least bias and greatest accuracy. The positive sign
associated with the bias (+4.95) indicated that, on the average, the deductions
were still excessive. The standard Forest Service scale deducted the greatest
volume (46 percent) and was the most biased and least accurate. A better estimate
of net scale can be made by making no deductions for peck than by using the cur-
rent, standard Forest Service scale.

Bias and accuracy for the cubic scales were best for the deduction that adjusted
the volume of rot by the amount of wood actually decayed (cubic soft). This option
also deducted the least amount of log volume. Making no deductions would have
provided a better estimate of usable wood than the cubic total deduction.

The 58 pecky logs were then divided into two groups: (1) Logs with rot showing on
One end (ROT1 logs, N = 25) and (2) logs with rot showing on both ends (ROT2
logs, N = 33). Rot occurred along the entire length of the woods-length, ROT2 log;
the decay in the ROT1 logs ended at some intermediate point. The average bias,
accuracy, and percentage of log scale deduction are shown for both groups in
table 3.

We expected the logs in the ROT2 group to have a greater lumber loss than the
ROT1 logs because they had decay scattered throughout the entire log length. This
was confirmed by the data in table 3 that compares the average bias for gross
Scribner and gross cubic scales. ROT2 logs, on the average, were more negatively
biased than were ROT1 logs.

Table 3 also shows that all the scaling systems were more biased and less ac-
curate for ROT2 logs than for ROT1 logs. The Scribner systems, for instance, tend-
ed to overdeduct when pecky rot occurred on both log ends, as compared with
logs with rot showing on one end.

Greater bias and less accuracy were also present in the ROT2 logs when the cubic
systems were used. The cubic soft method had a small bias on the ROT1 logs
(bias = +0.79) but underdeducted (bias = -3.60) on the ROT2 logs.

Table 2—Average bias, accuracy, and percentage of scaling defect for incense-
cedar logs with pecky rot as the only scaling defect

Average
amount of
Scaling option Bias Accuracy defect
— - -Cubic feet- - - Percent
Scribner:
Scribner gross - 5.73 13.81 —
134-inch rule + 4.95 11.42 28
4-inch rule + 8.88 14.20 36
Standard Forest
Service scale + 13.99 18.25 46
Cubic:
Cubic gross — 4.37 13:63 -
Cubic soft Seal 137 5
Cubic all + 9.02 13.87 28

“Number of logs equals 58.

Table 3—Average bias, accuracy, and percentage of defect for logs with
pecky rot showing on one end (ROT1) and logs with pecky rot showing on

both ends (ROT2)

Scaling option

Scribner:
Gross Scribner
134-inch rule
4-inch rule
Standard Forest
Service scale

Cubic:
Gross cubic
Cubic soft
Cubic all

ROT1 ROT2

Average Average

amount amount
Bias Accuracy of defect Bias Accuracy of defect
-— Cubic feet —- Percent -— Cubic feet -- Percent
279 6.54 —_— — 7.96 17.40 —
+2.67 7.79 19 + 6.68 13.53 34
+ 5.34 8.92 28 +11.56 lien 43
+9.14 12.91 36 17.66 21.43 54
—1.10 5.06 _— — 6.85 17.53 —
+ .79 4.75 4 — 3.60 14.50 6
+5.86 11.96 18 +11.41 WSK 7A 36

Conclusions

Acknowledgment

Metric Equivalents

Literature Cited

All Scribner scaling options overdeducted for the presence of pecky rot. The
magnitude of bias was greater for logs with pecky rot showing on both log ends
versus logs with rot showing on one end. The 134-inch rule performed best of the
Scribner systems tested; it had the least amount of bias and the greatest degree of
accuracy. Future efforts to revise the Scribner scaling rules for pecky rot should
note that the scaling options that performed the worst were those that deducted
the greatest percentage of log volume.

Of the cubic systems, the cubic soft method was superior to the cubic total method.
The cubic soft method slightly overdeducted for logs with rot showing on one end
and underdeducted for logs with rot showing on both ends.

We thank Larry Knecht, regional check scaler, USDA Forest Service, Pacific
Southwest Region, for his assistance in defining the scaling systems and for his
leadership during the scaling of the study logs.

1 inch = 2.540 centimeters
1 foot = 0.305 meter
1 cubic foot = 0.028 cubic meter

Boyce, J.S. The dry-rot of incense cedar. Bull. 871. Washington, DC: U.S. Depart-
ment of Agriculture; 1920. 58 p.

Cahill, J.M. Cubic log scale deductions in dead timber: an empirical approach.
Forest Products Journal. 33(4): 21-24; 1983.

Fahey, T.D.; Snellgrove, T.A.; Cahill, J.M.; Max, T.A. Evaluating scaling systems.
Journal of Forestry. 79(11): 745-748; 1981.

U.S. Department of Agriculture, Forest Service. National Forest log scaling hand-
book. FSH 2409.11. Amend 4. Washington, DC: 1973a. 247 p.

U.S. Department of Agriculture, Forest Service. The outlook for timber in the
United States, FRR-20. Washington, DC: 1973b. 367 p.

U.S. Department of Agriculture, Forest Service. Cubic scaling handbook (a
review draft). Washington, DC: 1978. 200 p.

Wagener, Willis W.; Bega, Robert V. Heart rots of incense-cedar. For. Pest Leaf.
30. [Washington, DC]: U.S. Department of Agriculture, Forest Service; 1958. 7 p.

*GPO 791-012 (1987)

The Forest Service of the U.S. Department of
Agriculture is dedicated to the principle of multipie
use management of the Nation’s forest resources
for sustained yields of wood, water, forage, wildlife,
and recreation. Through forestry research,
cooperation with the States and private forest
owners, and management of the National Forests
and National Grasslands, it strives — as directed by
Congress — to provide increasingly greater service
to a growing Nation.

The U.S. Department of Agriculture is an Equal
Opportunity Employer. Applicants for all Department
programs will be given equal consideration without
regard to age, race, color, sex, religion, or national
origin.

Pacific Northwest Research Station
319 SW. Pine St.

PO. Box 3890

Portland, Oregon 97208

»J.S. Department of Agriculture

»?acific Northwest Research Station
319 SW. Pine Street

-2.0. Box 3890

)?ortland, Oregon 97208

»

Dfficial Business
?enalty for Private Use, $300

do NOT detach label

BULK RA
POSTAGE
FEES PA
USDA-F:
PERMIT No. |