Development of a well-behaved site index equation: jack pine in north central Ontario

Abstract

A base-age invariant site index equation for jack pine based on the Chapman–Richards function was produced that satisfied nine criteria of preferred behavior for site index equations. A difference form of the Chapman–Richards equation produced the best behavior; height equalled site index at base age, and the shape of the curves reflected the data. The data structure used to fit the difference equation was all possible differences rather than the conventional nonoverlapping sequential intervals because this improved the behavior of the model. Height-prediction equations typically use height at base age (site index) as a predictor variable. As site index is measured with error, the equation will be biased. This bias will be evident in the predicted height at base age and in the shape of the curves. Base-age invariant equations predict height and site index with the same equation and thus diminish the effect of stochastic predictor variables. The equation performed comparably to a previously published equation with a specific base age of 50 years.