negative coefficients, while shade-tolerant species, such as western redcedar 

 and western hemlock, had smaller negative coefficients. 



Site preparation increased heights for most of the shade-intolerant subse- 

 quent species, but had no effect or a negative effect on shade-tolerant subse- 

 quent species. 



Heights of subsequent regeneration increased as the number of trees on 

 the plot increased. This seems counter-intuitive, because increasing compe- 

 tition should decrease tree growth. However, increasing seedling density 

 could indicate better microsites. 



Heights of Excess Regeneration — Heights of excess trees were not meas- 

 ured in the field. However, a height must be assigned to all tree records being 

 passed to the Prognosis Model. Two things are known about excess trees. 

 First, an excess tree cannot be taller than the best tree of the same species 

 on the plot. Second, an excess tree is at least as tall as the minimum estab- 

 lishment height listed in table 1. 



Rather than assume a linear relationship between minimum establishment 

 height and height of the tallest tree of the same species on the plot, Weibull 

 distributions were developed for the heights of best trees. These distribu- 

 tions were used to define the shape of the distribution for heights of excess 

 trees. Coefficients are given in appendix B, table 19. Important variables 

 for the distributions are species and years since last disturbance. 



HOW THE PROGNOSIS MODEL AND REGENERATION 

 MODEL INTERACT 



The Prognosis Model is an individual-tree, distance-independent forest 

 growth and yield model. Originally developed for northern Idaho and adjacent 

 forests in Montana and Washington (Stage 1973; Wykoff and others 1982), the 

 Prognosis Model has been calibrated for other forest regions. Geographic 

 versions of the Prognosis Model are called variants. 



An inventory of trees from one or more plots represents a stand in the 

 Prognosis Model. The inventory is maintained in a tree list where attributes 

 of inventoried trees are kept — d.b.h., height, species, and so on. One of the 

 attributes in the tree list is the number of trees per acre each tree represents. 

 The trees per acre represented by a tree is initially a function of the inven- 

 tory plot size and the number of plots sampled in the stand. Each tree sampled 

 on a 1-acre plot represents 1.0 trees per acre, each tree sampled on two 1-acre 

 plots represents 0.5 trees per acre, each tree sampled on ten Vsoo-acre plots 

 represents 30 trees per acre, and so on. 



The Prognosis Model projects the increment of individual trees over time. 

 Time steps in the Prognosis Model are called cycles — cycles can vary in length, 

 the default being 10 years. Prognosis Model harvests and thinnings, if any, 

 are done at the beginning of the cycle; then growth and mortality are pre- 

 dicted for each tree. Each tree receives a predicted increment in d.b.h. and 

 height as well as a change in crown ratio. Mortality is simulated as a reduc- 

 tion in the trees per acre that the tree represents in the stand. 



Plot attributes that change are updated at the end of the cycle for reporting 

 purposes and for use by other submodels of the Prognosis Model; for example, 

 plot overstory density and plot overstory species composition. The regen- 

 eration model makes extensive use of plot information. 



Use of the regeneration model can be scheduled at the end of any Prognosis 

 Model cycle. The regeneration model uses the updated inventory to predict 



23 



