Growth Model

The Pack Forest LMS portfolio uses the FVS growth model which is tailored to Northwest forest conditions by using the PN variant. It is recognized that FVS running in LMS tends to significantly over-predict forest growth at Pack Forest. Growth model outputs can be modified by key files, which provide multipliers that adjust algorithms used to predict various stand characteristics. As Pack Forest Managers collect more growth data and compare it to predicted values, the key file can be adapted to bring modeling closer to reality.

The latest key file was developed using data from Pack Forest's Continuous Forest Inventory (CFI). This inventory provides 20-years of growth data. Confidence in the model has been increased by comparing actual growth with LMS predicted growth using a variety of key files. The current key file is based on the CFI data and leaves a sufficient margin for uncertainty by under-predicting the growth rates determined by the CFI data. More information is available by reading Pack Forest growth Model Calibration using CFI.

Following is a comparison of existing growth models adapted for use in the Pack Forest region. The comparison is a useful reminder that growth models are not prefect representatives of forest growth. This comparison also serves as a broad sensitivity analysis of the models. A sensitivity analysis is a technique to evaluate the sensitivity of a system to a given variable. For example, a stand of trees could be projected two or more times by the same model, each time with different adjustments to particular algorithms, such as those related to site index or mortality. By controlling only one variable at a time, the sensitivity of the model to the given changes can be established. This is useful when it is necessary to make changes to growth predictions.

The following analysis is much broader in that the variable is the model itself. One modeling run is with FVS, one with FVS using the original Pack Forest key file and the third uses Organon. Though this analysis is not particularly useful in terms of making adjustments to the modeling ability at Pack Forest, it does serve as a comparison of models available for the geographic area.

Graphical Comparisons:

Summary Values

SUMMARY VALUES
	FVS (No Keyfile)	FVS (Keyfile)	Organon (DLL, SMC Variant)
BIODIVERSITY
COARSE FILTER (Oliver5C structures)	51%	49%	42%
COARSE FILTER (HCSSPT Structures)	47%	43%	44%
LATE SUCCESSIONAL STRUCTURES	100%	100%	100%
OLD GROWTH STRUCTURES	45%	36%	0%
PRODUCTIVE CAPACITY
TOTAL GROWTH (million board feet)	322	105	240
FINAL/INITIAL STANDING VOLUME	611%	267%	481%
SUSTAINED STANDING QUALITY	23%	35%	27%

Normalized Values

NORMALIZED SUMMARY VALUES
	FVS (No Keyfile)	FVS (Keyfile)	Organon (DLL, SMC Variant)
BIODIVERSITY
COARSE FILTER (Oliver5C structures)	5	5	4
COARSE FILTER (HCSSPT Structures)	5	4	4
LATE SUCCESSIONAL STRUCTURES	10	10	10
OLD GROWTH STRUCTURES	5	4	0
PRODUCTIVE CAPACITY
TOTAL GROWTH (million board feet)	10	9	10
FINAL/INITIAL STANDING VOLUME	0	0	0
SUSTAINED STANDING QUALITY	2	3	3
HEALTH, SOIL & WATER
WIND SAFETY	4	3	2

Coarse Filter Biodiversity

FVS (No keyfile)	FVS (Keyfile)
Organon (SMC)

Old Growth Structures

Standing Volume

Total Growth

Trees > 24" DBH, Volume

Trees 12 to 24" DBH, Volume

Trees < 12" DBH, Volume

Low wind risk, number of acres

Visualization

Another approach to sensitivity analysis is to compare photographs with computer generated visualizations. The visualizations can be generated at both the stand and landscape level. An example of landscape visualization compared to a photograph is show below. These images are from Wilson and McGaughey (2000).

Photograph taken by Jeremy Wilson using 35mm camera. Pictures were then scanned and rescalled to 640x480 pixel resolution.	Visualization generated at 640x480 pixel resolution using Envision.