Soil-Site Prediction of Stand Height 



The important relationship between soil characteristics and tree 

 growth has been apparent for many years. The complex relationship 

 between site quality for various tree species and soil factors has been 

 the point of concern of a number of excellent investigations. These 

 studies have deepened the understanding of tree growth and have pro- 

 vided a considerable amount of useful knowledge to forest managers 

 and landowners. 



It is important to note, however, that any soil-site study must 

 necessarily concentrate on only a limited portion of the total environ- 

 mental influence. Careful and complete evaluation of site quality must 

 take into consideration factors of climate, management, diseases and 

 injuries, competition, forest composition, and species as well as soil 

 characteristics. Nevertheless, an investigation of soil characteristics 

 within a limited climatic region on uncut, natural stands can provide 

 insight into site quality evaluation. 



A total of 109 sample plots were used in regression analyses to pre- 

 dict average dominant-codominant stand height in terms of age and 

 soil-site varialjles. These data included plot data from the 65 plots in 

 the growth study supplemented by plot data from 44 plots provided by 

 the Soil Conservation Service. 



The model used for predicting average height of dominant-codomi- 

 nant trees in terms of various independent variables was: 



Logio Height := bo + bi xi + b2 X2 + . . . + b, x -f e Model 1 



where: x = 1/age 



X2 to X = various soil-site characteristics including: 

 k 



Natural drainage class 

 Thickness of A horizon 

 Thickness of A + B horizons 

 Depths to A and B 



Soil pH 



Soil structure of B horizon 

 Texture class of A and B horizons 

 Percent clay in A and B horizons 

 Percent silt plus clay in A and B horizons 

 Percent coarse fragments greater than 2 mm. in B horizon 

 Percent slope 



Aspect and exposure of the site 

 Surface rockiness or stoniness 

 Stratification 



Family textural classification 

 e = the increment by which any individual dependent observation may fall off 

 the regression surface (called error) 



Although many variables approached significance (5-percent level), 

 only two proved significant. They were: 



1. Total age (at breast height) 



2. Natural drainage class 



