Site 

 Index 



I ime 

 Si nee 

 Cutting 



Begi nni ng 

 Diameter 

 Distributions 



Diameter Distributions 

 at Start of Growth Period 



Basal Area 

 Growth Rates 



Mortal ity 

 Rates 



Ingrowth 

 Rates 



Conversion 

 Rates 



Diameter Distributions 

 at End of Growth Period 



Figure 8. --System interaction chart. 



People interested in validation have come to realize that none of the three philosopliies 

 fully meets validation needs individually or is applicable to all situations. As a result, a 

 utilitarian approach has evolved. Shannon defines the three steps of this approach: 



The first stage is to seek face validity of the internal structure of the model 

 based upon a priori knowledge, past research, and existing theory (page 215). 



The second stage is also concerned with the validation of the internal structure 

 of the model, and consists in empirically testing, whenever possible, the hypothesis 

 used (page 215) . 



The third stage attempts vigorously to verify the model's ability to predict the 

 behavior of the real world system (page 216) . 



Many applications to forestry require the simulator to be both a prediction (or "prognosis") 

 tool and a technique for examining the dynamics of the system. This requires the simulator to 

 be accurate not only in a behavioral sense but also quantitatively. Many of the system com- 

 ponents are slow to react and, coupled with the general long-range aspects of forestry, require 

 long simulation runs to fully monitor true system behavior. With these special characteristics 

 in mind, I slightly modified and expanded upon the preceding utilitarian philosophy to produce 

 the following tentative "rules of validation." 



1. The simulator components should have a sound theoretical basis that uses as much 

 experimental evidence and expert knowledge as possible. This would include identifying the 

 significant system components, the factors that can influence these components, and applicable 

 forms and techniques for modeling them. 



31 



