Using soil properties to predict forest productivity in southern Illinois

The purpose of this study was to help explain and extrapolate expert site index value estimates for important tree species of southern Illinois soils. Statistical models were used to quantify the relationship between soil properties and expert-derived values for tree growth. Sixteen physical and chemical soil properties of 68 soils found in southern Illinois were used, along with published site index values, in a multivariate stepwise multiple regression analysis. The tree species selected for estimated site index regression were white oak (Quercus alba L.), yellow poplar (Liriodendron tulipifera L.), and northern red oak (Quercus rubra L.). These tree species were chosen based on availability of site index data, the site conditions required by the species, and ability to allow the prediction of tree growth for all soil series in the area. Stepwise regression procedures were used to select the most important soil parameters for each species productivity estimate from 16 original physical and chemical soil properties. The most important soil parameters in models to extrapolate site index predictions were total rooting depth, thickness of the A horizon, bulk density of the A and E horizon, bulk density of the subsoil, and percentage clay found in the B horizon. The parameter estimates were used to construct site index prediction equations. The soil property equations explained 61% of the variation in white oak site index estimates, 70% of northern red oak site index estimates, and 80% of the variation in site index estimates of yellow poplar. The projected productivity estimates will be useful to land managers who wish to allocate time and other resources to land based on the potential productivity of the site. Use of this method resulted in tree height growth projections for many soils published in regional soil surveys, including soils that currently lack forest cover and have high seasonal water tables or are subject to flooding.