Title Evaluation of the section modulus for tree-stem cross sections of irregular shape

Author(s) Koizumi, Akio; Hirai, Takuro

Citation Journal of Wood Science, 52(3): 213-219

Issue Date 2006-06

Doc URL http://hdl.handle.net/2115/14459

Rights The original publication is available at www.springerlink.com

Type article (author version)

File Information Evalution.pdf

Instructions for use

Hokkaido University Collection of Scholarly and Academic Papers : HUSCAP

Title: Evaluation of the section modulus for tree-stem cross sections of irregular shape

Authors: Akio Koizumi and Takuro Hirai

Corresponding address:
Akio Koizumi
Graduate school of Agriculture, Hokkaido University, N9 W9 Sapporo 060-8589
Phone: 011-706-3340, Fax: 011-706-3636
Email: [email protected]

Key words:
Windthrow resistance, Section modulus, Bitmap image, Hollow trunk, Bending strength
Abstract

The effect of section modulus as estimated for tree stems of irregular cross section with hollow trunks on

windthrow resistance was discussed. The sample trees were 12 aged poplar trees growing along the roadside in

Sapporo. Binarized bitmap images of photographs of the crosscut surface of the sample tree stems were used to

calculate numerical solutions for section moduli. The error for image resolution was simulated to less than 1%

under the condition in which the image was divided into more than 400 pixels. The coefficient of variation in

section modulus concerning six neutral axis directions in a horizontal plane was 11-14%. The reduction in section

modulus caused by decay or hollow trunk was 36-56% of the full cross section. The effect of section modulus on

critical wind velocity was found to be considerable. The critical wind velocity calculated for some of the sample

trees was less than 40 m/s, and these trees were considered to be in danger of bending failure of the stems.
Introduction

Aged trees of large size having decayed hollow cross-sections growing along the roadside or in parks are in

danger of windthrow by a gale, potentially causing damage to properties or personal injury. Therefore,

reinforcement or felling of the trees is required. Recently, environmental protection concerns have resulted in a

tendency for large trees to be maintained or preserved rather than felled. Therefore, the establishment of a method

of quantitative evaluation for tree resistance to wind force is desirable.

The moment that causes windthrow is induced by the wind force applied on the tree crown and bole. Wind

tunnel studies have been undertaken to obtain drag coefficients for various tree species1-6. Mayhead4 analyzed the

test data from the full-scale experiments performed at a wind tunnel of the Royal Aircraft Establishment at

Farnborough, UK in the 1960’s. The results indicated a reduction in drag coefficient with increased wind velocity

because the projected frontal area of the tree crown decreased with wind velocity. He concluded that the drag

coefficients converged to nearly constant values at the range of wind velocity likely to cause windthrow. The drag

coefficients at 30 m/sec were determined as 0.41 for tree species having a dense crown (grand fir and Sitka

spruce), which receive a large wind pressure per projected area, and 0.31 for pines (Corsican pine, lodgepole pine,

and Scots pine).

Koizumi7 discussed the failure mode and critical wind velocity for plantation grown coniferous species of 19-

to 45-year-old trees. The possible failure modes were classified into bending failure of tree stem and uprooting.

The critical height at which bending failure occurs was estimated to be approximately 20% of the tree height for

plantation-grown conifers7. In contrast, aged trees having decayed or hollow trunks (examined in the present

study) were estimated to fail at the stem, near the ground level, because of the decrease in effective cross section

that is necessary to withstand the bending moment. Although Mattheck8 indicated that the thin wall of the hollow

tree trunk would fail by flattening of the cross section or by shell buckling, quantitative evaluation of the critical

moment that causes flattening or shell buckling is difficult because the vertical expansion of the hollow in a trunk,

as well as the horizontal distribution must be known.

The section modulus of an irregularly-shaped cross section, which determines the critical bending moment,

could not be solved analytically. The goal of the present study is to discuss the method by which to evaluate the

section modulus for trees having irregularly-shaped or hollow cross sections using bitmap images of photographs

of the cross section of the tree stems. The critical wind velocity that causes bending failure was also estimated for

1
the sample trees.

Materials and methods

Sample trees

The sample trees were twelve black poplars (Populus nigra var. italica) that were growing along a street running

north to south on the campus of Hokkaido University, in Sapporo (Fig. 1). The trees were planted in the 1930’s

and the ages of these trees were estimated to be approximately 60-70 years. Accurate ages could not be

determined because the heartwood near the pith was lost due to decay. The trees were felled due to the potential

for collapse based on a safety diagnosis by a tree doctor. The diagnosis report included measurements of the tree

height, crown length, crown width, diameter at breast height, and diameter and thickness of sound wood without

bark at the critical height near the ground level (Table 1). The thickness of sound wood listed in Table 1 was

measured by a tree doctor using a Resistograph9.

Evaluation of effective cross section

A butt log was collected from each tree after felling and the cross-cut surface at a height of one meter above

ground level was photographed with a digital camera for evaluation of effective cross section. A one-meter-long

bar lying across the crosscut surface was used as a scale pointing north and south directions (Fig. 2a). In order to

determine the thickness of sound wood without decay, the penetrating depth parallel to the grain was measured

using a Pilodyn tester every 20 mm from bark side in eight radial directions on the cross-cut surface, as shown in

Fig. 2a. In order to discuss the relationship between penetration depth and the extent of decay, 36 small specimens

(20 x 20 x 60 mm) were cut from the adjacent section of Pilodyn-tested points of sample logs and were tested for

compressive strength parallel to the grain in green condition. After the test, the specimens were dried at 20ºC and

60%RH and were measured with respect to wood density. In addition, six small clear specimens (20 x 20 x 300

mm) were cut from upper logs of the sample trees and were tested in green condition for bending strength in order

to estimate the critical wind velocity that causes breakage due to wind.

2
Evaluation of section modulus

Bitmap images of the photographs for the cross section were binarized into black and white images in which white

areas represent wood (with or without decay) (Figs. 2b and 2c). The section modulus for the wood area was

calculated based on the white areas of the binarized images as follows:

At first, the position of the neutral axis was determined by dividing the geometrical moment of the area by the

area of the wood, according to the following equation, by scanning the images:

λ1 =
∑ ydA
∑ dA (1)

where λ1 is the distance between a horizontal line passing through the lower edge of the wood area and the

neutral axis, y is the distance between the horizontal line and a finite small area (dA) which was taken as a pixel

among the white area in the binarized image (see Fig. 3).

To calculate for the cross section on which Young’s modulus of the wood is not uniform, ydA and dA in Eq. 1

should be weighted with Young’s modulus. In the present study, Young’s modulus of the wood was assumed to be

uniform over the wood area because it was difficult to weight with variable Young’s modulus using a binarized

image.

Then, the moment of inertia (I) and the section modulus (Z) were calculated by numerical analysis following Eqs.

2 and 3:

I = ∑ ( y −λ1 ) 2 dA (2)

I
Z=
max(λ1 ,λ2 ) (3)

Estimation of critical wind velocity

Critical wind velocity was estimated based on the simplification of assumptions as follows:

1. Uniform wind pressure acts on the projected crown area (A) calculated as a product of the crown width and

crown length.

3
2. The height of the wind pressure center (Hw) is the half of the crown length.

3. Trees are assumed to fail in bending at a height of 1 m above ground level without uprooting.

4. The effects of neighboring trees on wind velocity are neglected.

5. The wind pressure acting on the stem below the crown is neglected.

Trees were analyzed as vertical cantilever beams fixed at ground level. Drag coefficients (CD) for the poplar

crown were calculated by Eq. 4, as determined from the empirical equation for the data for the Scotch pine

reported by Mayhead4.

3.22
CD = + 0.186 (4)
v

where v is the wind velocity (m/s).

The critical bending moment at a height of 1 m above ground level (Mc) is expressed as follows:

1
Mc = C D ρvc A( H w − 1)
2

2 (5)
= Zσ b
　　

where ρ is the air density (1.20 kg/m3), vc is the critical wind velocity, and σb is the bending strength of the wood.

Finally, the critical wind velocity (vc) was calculated using the following equation, which was obtained by

rearranging Eqs. 4 and 5:

− 3.22 + 10.4 − 0.744C

where

2 Zσ b
C=−
1.2 A( H w − 1)

Results and discussions

Errors in numerical solution due to image resolution

The cross section of an elliptical cylinder (shown in Fig. 4) was considered in order to discuss the error that arose

in numerical solution compared to the analytical solution in relation to the image resolution. The analytical

4
solutions for the moment of inertia (I) and the section modulus (Z) of the elliptical cylinder were calculated by

following equations.

π
I= (ao bo − ai bi )
3 3
(7)
4

I
Z= (8)
ao

Numerical solutions for the ellipse image were calculated for 20, 40, 60, 80 and 100% resolution of the original

image, which spans 620 pixels in the direction of the major axis (Fig. 5). The ratios of the numerical solution to

the analytical solution were less than 1.01 under the condition that the long diameter (2bo) was divided into more

than approximately 400 pixels for the images. Although the cross-sectional shapes of the sample trees were

complicated with uneven circumferences, and thus more errors were included in the numerical solution compared

to the case of a simple elliptical cylinder, the errors due to the image resolution were considered to be negligible

because the resolution of the photograph for the sample trees exceeded 600 pixels in diameter.

Effective cross section

As shown in Fig. 6, the relationship between wood density (moisture content: 11%) and compressive strength

parallel to the grain of the small specimens in green condition changed at a wood density of approximately 240

kg/m3 and a compressive strength of 10 MPa. The compressive strength for the parts of specimens that had wood

densities of less than 240 kg/m3 showed lower values than predicted by linear regression, and these parts were

considered to contain decaying wood. This result was consistent with published data10 for the specific gravity and

compressive strength of balsam poplar (Populus balsamifera), which are 0.31 and 11.6 MPa, respectively. The

penetration depth obtained using the Pilodyn tester is a good indicator of wood density11. Negative correlation

between Pilodyn penetration and wood density for the small specimens was found, as shown in Fig. 7. The

boundary value between sound wood and decaying wood was taken to be 33 mm in Pilodyn penetration, which

corresponds to a wood density of 240 kg/m3, according to the regression line shown in Fig. 7.

The penetration depth obtained using the Pilodyn tester was greater on the inner side of the cross section than

on the outer side (Fig. 8). According to the above definition, the distances from the outside surface to the

boundary between the sound wood and the decaying wood were 8 cm on the west side and 10 cm on the southeast

5
side, respectively. As an example, approximate boundaries determined for the sample tree (No. 5) are shown in

Fig. 2, in which Image A shows the visually determined wood area and Image B shows the wood area determined

by the Pilodyn penetration method. The section modulus is considered to be in the range with upper bound of

Image A and lower bound of Image B.

Variation in section modulus concerning moment direction

The windbreak resistance varies with the load direction because the section modulus is not constant for

unevenly-shaped cross sections concerning the neutral axis direction. Six section moduli (Z) were calculated for

the directions designated by 30-degree intervals in the horizontal plane of the sample trees, as shown in Fig. 9,

where each section modulus was shown as a ratio to the maximum section modulus of the tree. Values related to

solid line (ZA) and broken line (ZB) were calculated based on Images A and B, respectively, in Fig. 2. Although the
B

sample trees were planted in a row running north to south, no correlation between the moment resistance and

orientation of the stem was found. The coefficients of variation for the ratio were 11.4% and 14.2% for ZA and ZB,B

respectively. Next, the average section modulus will be discussed for six directions.

Effects of the hollow cross-section on section modulus

The section modulus decreases to some extent as decay develops in the heartwood region. The calculated section

moduli for the sample trees are shown in Fig. 10 in relation to their location in the planted row. The section

moduli for the trees located at the end of the row were found to be large as a result of vigorous growth in open

space. The average decreases in section modulus caused by the hollow cross-section were 36% and 50% for

Images A and B, respectively.

The section modulus based on the cylindrical cross section (ZC) was calculated by Eq. 9 using the diameter

and sound wood thickness of sample trees measured by a tree doctor in advance of felling (see Table 1):

π (ro 4 − ri 4 )
ZC = (9)
4ro

where ro is the average radius of the trunk without bark and ri is average radius of the hollow space.

Adequate correlation was found for the section modulus between the calculation based on the cylindrical

6
assumption and that based on Image A (Fig. 11). Average section modulus ratios were 1.024 and 0.824 for ZA/ZC

and ZB/ZC, respectively. Therefore, rough estimation of section modulus can be performed using the
B

cylindrical-shape assumption. However, accurate estimation concerning individual trees and/or load directions

should be based on numerical solutions that consider irregular shape.

Critical wind velocity

Wind velocity that causes bending failure for the sample trees were estimated by Eq. 6. The bending strength (σb)

for a poplar tree was assumed as 27.7 MPa, which was derived as the lower limit (fifth percentile) of the bending

strength measured for the small clear specimens in the green condition. Effects of defects such as knots on

bending properties for trees or logs are considered to be small, because wood fibers at the sapwood surface are not

cut by lumber sawing12.

The tendency of critical wind velocity in relation to tree location (Fig. 12) was similar to that of section

modulus shown in Fig. 10. The effect of section modulus on critical wind velocity was found to be considerable.

The critical wind velocities calculated for five of 12 sample trees for Image B were less than 40 m/s, and these

trees were considered to be in danger of windthrow because the maximum wind velocity adopted for the building

design code for the Sapporo area is 32 m/s.

Conclusions

The method of evaluating section modulus and its effects on windthrow resistance for poplar trees having

irregularly-shaped cross sections were discussed and the following conclusions were obtained.

1. The section modulus of trees having an irregularly-shaped and hollow cross section could be calculated with

numerical analysis based on an image of the sound wood area. The error involved with image resolution was

simulated to less than 1% when the image was divided into more than 400 pixels.

2. The coefficient of variation in section modulus concerning six neutral axis directions was 11-14%.

3. The reduction in section modulus caused by decayed or hollow trunks was 36- 56% of the full cross section.

4. Rough estimation of the average section modulus is possible based on the assumption of a cylindrical section.

7
 However, accurate estimation for individual trees and/or load directions requires numerical solution considering

irregular shape.

5. The effect of the section modulus on the critical wind velocity was found to be considerable. The critical wind

velocity calculated for some of the sample trees was less than 40 m/s and these trees were considered to be in

danger of windthrow based on the assumptions used in this study.

The failure modes of uprooting or shell buckling for thin-walled cylinders were not considered in the present

study. The calculation for critical wind velocity was done based on average section modulus for six directions

instead of minimum section modulus. Therefore, the critical wind velocity evaluated herein may be too large and

overestimated.

Photographs of cross cut surfaces of sample trees were used for the analysis in the present study.

Non-destructive evaluation of effective cross section will be required for the purpose of safety diagnosis for trees

in parks or along the roadside.

8
Acknowledgment

The authors thank Ms. Akiko Miura for her kind assistance in the experimental work.
References

1. Hirata T (1953) Fundamental studies on the formation of cutting series (2) (in Japanese). Bull. Tokyo Univ.

Forests 45: 67-88

2. Fraser AI (1964) Wind tunnel and other related studies on coniferous trees and tree crops. Scottish forestry

18: 84-92

3. Leiser AT, Kemper JD (1973) Analysis of stress distribution in the sapling tree trunk. J. Amer. Soc. Hort. Sci.

98: 164-170

4. Mayhead GJ (1973) Some drag coefficients for British forest trees derived from wind tunnel studies.

Agricultural Meteorology 12: 123-130

5. Johnson RC, Ramey GE, O'Hagan DS (1982) Wind induced forces on trees. J. Fluids Eng. 104: 25-30

6. Murakami S, Deguchi K, Takahashi T (1984) Shelter effects of trees as wind-breaks (in Japanese). Proc.

Symposium Wind Eng. 129-136

7. Koizumi A (1987) Studies on the estimation of the mechanical properties of standing trees by non-destructive

bending test (in Japanese). Res Bull College Experiment Forests Hokkaido Univ 44: 1329-1415

8. Mattheck C (1998) Design in nature. Springer, Berlin, pp 1-276

9. Rinn F, Schweingruber F-H, Schär E (1996) RESISTOGRAPH and X-ray density charts of wood comparative

evaluation of drill resistance profiles and X-ray density charts of different wood species. Holzforschung 50:

303-311

10. Forest Products Laboratory (1987) Wood Handbook. USDA Forest Service, Madison, pp 4.1-4.45

11. Cown DJ (1978) Comparison of the Pilodyn and Torsiometer methods for the rapid assessment of wood

density in living trees. N.Z. J. For. Sci. 8: 384-391

12. Koizumi A, Ueda K (1987) Bending and torsional properties of logs (in Japanese). Res Bull College

Experiment Forests Hokkaido Univ 44: 355-380

1
Table 1. Dimension of sample poplar trees
No. H (m) CL(m) CW(m) DBH (cm) Hc (m) Dc (cm) Tc (cm)
1 27 23.9 7.7 123 1.5 97.0 16.5
2 21 18.6 5.4 95 0.6 80.5 8.0
3 28 23.0 9.1 97 0.9 85.0 19.0
4 30 26.3 7.8 82 1.0 76.5 11.0
5 30 25.0 8.5 91 1.0 78.0 12.5
6 30 25.2 8.9 100 1.5 90.0 10.0
7 30 25.2 8.3 97 1.1 83.0 9.5
8 30 24.5 9.4 107 1.1 98.5 9.0
9 29 24.5 8.8 106 1.1 86.5 12.0
10 28 24.5 7.4 90 1.5 75.0 8.5
11 28 24.0 8.0 85 1.4 82.5 8.0
12 29 24.6 8.9 105 1.1 94.5 20.0

H, Tree height; CL, Crown length; CW, Crown width; DBH, Breast height diameter; Hc,
Surveyed height considered to be critical concerning windbreak by a tree doctor; Dc,
Average diameter at surveyed height; Tc, Average sound wood thickness at surveyed height
Titles and legends for Figures

Fig. 1. Sample poplar trees planted in a row (before felling)

Fig. 2. Example of photograph and binarized images of a cross section (Tree No. 5)

(a) Photograph of a cross section with an one meter long bar. Black dots indicate

Pilodyn-tested points

(b) Binarized image of the photograph for wood area (Image A)

(c) Binarized image of the photograph for sound wood area estimated by Pilodyn tests

(Image B)

Fig. 3. Dimension of irregularly-shaped cross section

Fig. 4. Sample cross section of elliptical cylinder used for calculation for errors in numerical

solution.

bo=1.2 ao; t/bo=0.3

Fig. 5. Errors in numerical solution in relation to resolution of the elliptical cylinder image

shown in Fig. 4

Fig. 6. Relationship between wood density and compressive strength parallel to the grain for small

specimens cut from the sample logs including decaying wood

Fig. 7. Relationships between Pilodyn penetration and wood density for the small specimens

Fig. 8. An example of Pilodyn penetration trend observed in radial directions on the crosscut

surface

Fig. 9. Section modulus ratio concerning six moment directions for all sample trees. Twelve

trees were numbered sequentially from the south to north end in the planted row. Upward

direction indicates north. Solid line, calculation based on Image A; broken line, calculation

based on Image B. Both Images coincided with each other for Nos. 10 and 11

Fig. 10. Calculated section moduli of sample trees in relation to their location shown as the

distance from the south end tree (Tree No. 1). ZA,B, section moduli numerically calculated

for Images A and B

Fig. 11. Comparisons of section moduli. ZA,B, section moduli numerically calculated for Images

A and B; ZC, calculated for cylindrical cross section based on the average diameter and
 sound wood thickness measured by a tree doctor with a Resistograph; r, coefficients of

correlation between ZC and ZA,B

Fig. 12. Critical wind velocity estimated for the sample trees in relation to their location shown as

the distance from the south end tree (Tree No. 1). vA, calculated for Image A; vB, calculated

for Image B
 N
ｄA

λ1 λ2
N
Solution ratio (numerical/analytical)
1.1

Section modulus
Moment of inertia

0.9
0 100 200 300 400 500 600 700

Number of pixels in diameter of ellipse for major axis

10

0
100 200 300 400

Wood density (kg/m 3)

Wood density (kg/m 3)

200
r=-0.715
Y=-6.01X+440

100
10 20 30 40
Pilodyn penetration (mm)
 50
Pilodyn penetration (mm)

40

30

20
Tree No. 3
Southeast side
10 West side

0
0 5 10 15 20 25
Distance from bark (cm)
1 2 3 4

5 6 7 8

9 10 11 12
 0.1

Z (m3)

0.05

ZA
ZB
without hollow
0
0 70
Location of sample trees (m)
 0.08

0.06
ZA,B (m )
3

0.04

0.02
ZA r=0.906
ZB r=0.744
0
0 0.02 0.04 0.06 0.08
3
ZC (m )
 80
Critical wind velocity (m/s)

70
60
50
40
30
20 vA
10 vB

0
0 70
Location of sample trees (m)