by slower growth. Yet, notice that the average height of best trees for Douglas- 

 fir in the Pseudotsuga menziesii/Symphoricarpos alhus habitat type is 4.9 feet 

 for the budworm projection and 4.2 feet without budworm (table 7). This 

 occurs because budworm changes the proportion of advance and subsequent 

 regeneration within a species. Budworm reduces the probability of subse- 

 quent Douglas-fir (appendix B, table 13) but does not change the probability 

 of advance Douglas-fir (appendix B, table 12). Thus, there is a greater pro- 

 portion of taller advance Douglas-fir, increasing the average height for that 

 species. 



Nonhost species can be taller as well. This effect coiild be due to earlier estab- 

 lishment of nonhost species or better growth rates because of less competition. 



In smnmary, there are four effects of western spruce budworm represented 

 in the regeneration model. First is a reduction in the probability of stocking. 

 Second is reduced nimiber of trees per stocked plot. The significance of reduc- 

 tions in the probability of stocking (PS) and the number of trees per stocked 

 plot (TPSP) becomes clear when densities are assigned using equation 4: 



TPA = (PS * TPSP * 300)/iV 



PS and TPSP interact multiplicatively, so small reductions in either (or both) 

 can have large impacts. 



The third effect of budworm is to change species composition of the new 

 stand. Fewer host species regenerate in the new stand. The fourth effect 

 is to change heights of regeneration. Nonhost species may become taller 

 relative to host species that are under attack by budworm. Subsequent host 

 species in defoliated stands may be shorter, primarily due to delays in be- 

 coming established. However, average heights of host species may be taller 

 with budworm defoliation than without defoliation. This happens when the 

 proportion of advance regeneration for a species increases relative to subse- 

 quent regeneration. The taller advance regeneration increases the average 

 height for the species. 



OTHER MODEL FEATURES 



Most of the details about the regeneration model have been discussed. 

 A few additional topics may be of interest to others developing a similar model. 



Calibration Model calibration is a process of adjusting the probability of stocking curve 



to coincide with an inventory. This adjustment takes place whenever the 

 year of inventory is within the years the regeneration model is making pre- 

 dictions. A stand more densely stocked than expected has less room for ad- 

 ditional regeneration, while a stand less densely stocked than expected has 

 more room for regeneration. 



We considered using the ratio of stocked plots to calibrate the probability 

 of stocking. However, this calibration method would not work for inventories 

 with plot sizes other than Vaoo-acre. 



This drawback was eliminated by developing an equation predicting the 

 proportion of plots that are stocked from the number of trees per acre. The 

 equation is 



RATIO = [1 + exp - (-5.17397+0.85131*ln(TPA))]-i (5) 

 where 



RATIO = proportion of plots that are stocked 



TPA = trees per acre in the stand <3 inches d.b.h. 



33 



