Forest soil chemistry and terrain attributes in a Catskills watershed

Knowledge of soil chemistry is useful in assessing the sensitivity of fores ted areas to natural and anthropogenic disturbances, but characterizing lar ge areas is expensive because of the large sample numbers required and the cost of soil chemical analyses. We collected and chemically analyzed soil s amples from 72 sites within a 214-ha watershed in the Catskill Mountains of New York to evaluate factors that influence soil chemistry and whether ter rain features could be used to predict soil chemical properties. Using geog raphic information system (GIS) techniques, we determined five terrain attr ibutes at each sampling location: (i) slope, (ii) aspect, (iii) elevation, (iv) topographic index, and (v) flow accumulation. These attributes were in effective in predicting the chemical properties of organic and mineral soil samples; together they explained only 1 to 25% of the variance in pH(w), e ffective cation-exchange capacity (CECc), exchangeable bases, exchangeable acidity, total C, total N, and C/N ratio. Regressions among soil properties were much better; total C and pH, together explained 33 to 66% of the vari ation in exchangeable bases and CECc. Total C was positively correlated wit h N (r = 0.91 and 0.96 in Oa horizons and mineral soil, respectively), exch angeable bases (r = 0.65, 0.76), and CECc (r = 0.54, 0.44), indicating the importance of organic matter to the chemistry of these acidic soils. The fr action of CECc occupied by H explained 44% of the variation in pH(w). Soil chemical properties at this site vary on spatial scales finer than typical GIS analyses, resulting in relationships with poor predictive power. Thus, interrelationships among soil properties are more reliable for prediction.