Independent Variables 



Several variables of forest 

 stand condition thought to be good 

 candidates for predicting changes 

 in timber value associated with 

 gypsy moth outbreaks v^^ere se- 

 lected for analysis as independent 

 variables. These variables are meas- 

 ures of: 



Timber stand size composition 



Average tree diameter 



Timber stocking 



Timber stand age 



Species composition 



Crown position 



Crown condition 



Site index 



Land capability 



Elevation 



Aspect 



Slope 



Position on slope 



The choice of these variables was 

 based primarily on findings of Kegg 

 (1971), Campbell and Sloan (1977), 

 Houston and Valentine (1977), 

 Gansner and co-workers (1978), and 

 Herrick and co-workers (1979). 



These variables were then 

 screened with the help of corre- 

 lation and stagewise regression 

 analysis to select the best ones for 

 the model. We were looking for pre- 

 dictors that (1) are easy to measure, 

 (2) explain large amounts of 

 variance in the dependent variable, 

 and (3) are not highly correlated 

 with one another. Three were 

 selected: 



• Basal area per acre in tree species 

 that gypsy moth tends to avoid.' 



• Percent of basal area in trees 3.0 

 to 4.9 inches dbh. 



' Such as yellow-poplar, ash, black lo- 

 cust, and sycamore. 



• Percent of basal area in trees with 

 poor crowns.^ 



Results 



Robust regression analysis 

 gave us a simple equation for 

 predicting the rate of change in 

 timber value for infested stands: 



R = 1.143 -I- 0.065(BSA) + 

 0.082(PBS) - 0.107(PBP) 



where 



R = 



BSA = 



PBS = 



PBP = 



Compound rate of change 

 in timber value 



Basal area per acre in tree 

 species that gypsy moth 

 tends to avoid 



Percent of stand basal 

 area in trees 3.0 to 4.9 

 inches dbh 



Percent of stand basal 

 area in trees with poor 

 crowns. 



Only three of many elements of 

 stand condition analyzed as in- 

 dependent variables are included in 

 this equation. Including them makes 

 especially good sense because (1) 

 stands with greater stocking in 

 species avoided by gypsy moth are 

 usually attacked less severely and 

 thus tend to have lower mortality 

 rates and higher value growth rates; 

 (2) trees in the 3.0- to 4.9-inch dbh 

 class have little or no value for 

 timber products, but those that sur- 

 vive gypsy moth attacks soon grow 

 into merchantable size and thus 

 have high value growth rates; (3) 

 trees with poor crowns have lower 

 vigor, are more likely to die after 

 defoliation, and have lower (often 

 negative) value growth rates. 



^ Crowns were classed as poor when 

 50 percent or nnore of the branches were 

 dead (allowances pernnitted for non-self- 

 pruning species); when foliage density, 

 size, or coloration was of subnormal 

 quality, or when epicormic sprouting was 

 heavy. 



Implications for Management 



The model has limitations. It 

 has not been field tested, and we do 

 not know how well it will work on a 

 new frontier of infestation. Plans 

 have been made to test it. Also, it 

 takes no account of important non- 

 timber impacts on esthetic quality 

 or the nuisance of caterpillars in 

 recreation areas and back yards. 



Despite its problems, the model 

 and others like it seem to offer use- 

 ful tools for decisionmaking. Es- 

 pecially where timber value is 

 important. Consider these two 

 typical stands: 



Stand #1 Stand #2 



Basal area per 

 acre in species 

 avoided (ft.^) 



Percent of basal 

 area in trees 3.0 

 to 4.9 inches 

 dbh 



Percent of basal 

 area in trees 

 with poor 

 crowns 



10 



65 



35 



25 



According to our model, if Stand #1 

 suffered a gypsy moth outbreak un- 

 treated, we could expect a com- 

 pound rate of change in timber 

 value of about - 5 percent over the 

 next 8 years or so. Whereas, the 

 rate of change for Stand #2 would 

 be -I- 5 percent. With this kind of in- 

 formation, choices of how to man- 

 age for the gypsy moth can be 

 made more objectively. 



Acknowledgments 



Our sincere thanks to per- 

 sonnel of Forest Pest Management 

 staff. Northeastern Area State and 

 Private Forestry, USDA Forest Serv- 

 ice, for support in collecting and 

 analyzing data. 



2 



