BARK THICKNESS SHOULD YOU CARE?

1
BARK THICKNESSSHOULD YOU CARE?

• Jim Flewelling
• Jim.flew_at_yahoo.com
• GMUG, November 15, 2013

With a correction and addendum, Nov. 18, 2013.
2
OUTLINE

• YES You Should Care .
• Background 1994 NW Taper Coop
• Suspicious volume cutouts.
• Subsequent volume validations
• Bark thickness results
• Background Structure of Taper Predictions.
• Solutions.

3
Yes, You should care!

• For almost taper and volume equations were
  derived from data where the key variables
  included DBH (outside bark), and inside bark
  volume or stem profile.
• If the equations are applied to areas where bark
  thickness is different, results will be bad.

4
1994 West-sideTaper Coop

• 726 Douglas fir trees
• Across the diameter range within stands.
• 7 predefined regions.
• Eqns
• Bark thickness at breast height
• Inside-bark profiles (conditioned at BH)
• Upper stem bark thickness equations
• Excellent residuals.

5
1994 West-sideTaper Coop
6
Suspicious volume cutouts

• Client was noticing a pattern of low volume
  cutouts in one of the operating areas.
• Volume and taper equations are often the innocent
  usual suspects
• This client put them on trial instead of just
  complaining.

7
Subsequent volume validations

• 11 harvest age DF stands.
• 109 trees, selected across the DBH range.
• Standing measurements DBH, HT, and upper stem
  diameter (17, 41).
• Felled Outside bark profile, and bark thickness
  profile.

8
Subsequent volume validations

• 4 mean underestimate inside bark volume.
• 0.0 mean error in outside bark volume
• Mean predicted bark ratio .097
• Mean observed bark ratio .116
• Do the Math!
• No other problems of this magnitude.

9
Bark Thickness Results -1993
From same client
Dble Bark / DBH
10
Bark Thickness Results -2012
11
Bark Thickness Ratio ()

• Mean error 1.68 percentage points.
• Mean error by stand range of 7 points.
• No clear regional effects.

12
Bark Thickness Ratio - ANOVA

• Y Bark Thickness Ratio at breast height ().
• Variance between stands 3.95
• Standard error 1.99
• Variance between trees (within stands)5.43
• Standard error 2.33

13
Background - Structure

• Coop equations
• BR eqn at breast height f(DBH, HT, region).
• ib profile f(DBH, HT, region)
• Profiles then conditioned on DBHib
• Less common
• ib profile f(DBH, HT)
• Software. Ability to supply DBHib?

14
Solutions

• Cruising can include bark thickness on a small
  number of trees.
• For non-cruise applications, need survey models.
• Try to find correlated variables.

15
Solutions Few Trees / Stand
of Stands Trees/Stand of trees size of 95 CI on mean
20 10 200 1.92
20 20 400 1.93
20 40 800 1.89
40 10 400 1.30
40 20 800 1.30
40 40 1600 1.28
16
SOLUTIONS- CAUSES

• Methods?
• Climate / region / stand conditions ?
• Seed source ?
• Recent weather?
• ??????

17
ADDENDUM
The remaining slides were attached after the
presentation on November 15, 2013. These are
copies of several pages taken from the 1994
Northwest taper cooperative report. These pages
all refer to Douglas fir. English units (feet and
inches) are used.
18
1994 Coop Locations of plots in Washington with
Douglas fir sample trees.
19
1994 Coop Locations of plots in Oregon with
Douglas fir sample trees.
20
1994 Taper Coop. Douglas fir sample. Bark Ratio
double bark thickness / DBH.
21
SIZE GROUPS Trees are divided ibto twelve
categories by DBH and total height. Horizontal
lines represent cut lines. Not shown is a central
trend line, used to subdivide each height group
into to subsets representing a division between
large and small DBHs.
22
TAPER COOP. DF residuals by Region, and DCLASS
(DBH vs. Median for HT).
Double bark thickness (DBT), and inside-bark area
at BH
Area (ib), actual and predicted