Table 8 . --Comparison of actual log of basal area growth data from managed, even-aged plots to 

 log of basal area growth predicted by various equations developed using data from 

 those plots indicated 



Vigor class 



Plots used 

 in equation 

 development 



Number of 

 observations 



Mean squared 

 error 



Average 

 residual 



Percent 

 average 

 residual 



Blackjack pine 



All 



5,886 



0.224909 



+0.023643 



-0.814989 



Blackjack pine 



Managed, 

 even-aged 



5,886 



.187551 











Table 9 . --Comparison of actual log of basal area 

 aged plots to log of basal area growth 

 using data from those plots indicated 



growth data from managed , even- 

 predicted by various equations 



cii in uiicvai 



developed 



Vigor class 



Plots used 

 in equation 

 development 



Number of 

 observations 



Mean squared 

 error 



Average 

 resiilual 



Percent 

 average 

 residual 



Blackjack pine 



All 



24,956 



0.771353 



-0.000004 



+0.000140 



Blackjack pine 



All 



managed 



24,956 



.767860 











Also in those tables are the MSE's, average residuals, and percent average residuals for 

 selected equations developed in phases 2 and 3 of the analysis using just the data from the 

 given set. The equations chosen for comparison were those minimizing RMSQR for each data set. 

 In choosing these equations, no concern was given to whether they were reasonable in behavior, 

 and most of them were not. These equations do, however, represent the best that could be done 

 for the given data set and, therefore, provide a benchmark for how good the final equations 

 are at predicting growth. 



For blackjack pine, the results for the virgin uneven-aged data (table 6) indicate that 

 the two equations are quite similar at predicting growth. For the managed, uneven-aged data 

 (table 7), the equation developed with the even-aged data shows a larger (but still small) 

 average residual, but the MSE's are almost the same. This positive percent residual is offset 

 by a negative percent residual for the even-aged data (table 8) . When the managed even- and 

 uneven-aged data are put together (table 9), the average residual becomes insignificant. The 

 conclusion, therefore, is that the two blackjack equations give about the same results as far 

 as MSE's and average residuals are concerned. 



For yellow pine, the results indicate not much predictive capability was lost when the 

 two data sets (virgin and managed uneven-aged data sets) were combined. 



Mortality 



Mortality is often separated into a catastrophic (irregular) or a noncatastrophic (regular) 

 form (Lee 1971; Stage 1973; Monserud 1976). As its name implies, catastrophic mortality is 

 the result of usually unpredictable, massive disturbances such as fire, hurricanes, tornadoes, 

 and insect or disease epidemics. Noncatastrophic mortality accompanies normal stand develop- 

 ment. The models developed in this simulator predict only noncatastrophic mortality. 



12 



