A soil-site study for inland Douglas-fir
A soil-site study was conducted for inland Douglas-fir growing in northern Idaho and north western Montana. The hypothesis was that standard soil survey procedures would provide edaphic data that could predict site index in the absence of site trees. Soil profile descriptions and physical analyses were obtained on 133 plots, along with several physiographic site descriptors. Chemical analyses were performed on soil samples from a third of these plots, and moisture availability was determined on 60% of the plots. Site index was based on felled-tree stem analyses. Elevation was the strongest predictor, accounting for a third of the variation in site index. The addition of habitat type information resulted in a significant improvement (as did longitude and precipitation), but still left over half the variation unexplained. After examining numerous soil properties the standard error could only be reduced 0.3 m, a disappointing result in light of the considerable time and expense necessary for soil sampling. The causes of these low soil–site correlations could not be conclusively determined, but the most likely explanations are (i) that the number of important site factor interactions occurring in this large and complex study area far exceeded the sample size, and (ii) failure to measure the true causes of site productivity.