Regional locale and its influence on the prediction of loblolly pine diameter distributions

Data gathered from intensively and nonintensively managed loblolly pine (Pinus taeda L.) plantations were used to model the diameter distributions of stands across the southeastern United States. Weibull scale and shape parameters were predicted using stand density, site index, and stand age as covariates. Including geographic locale (latitude and longitude) of the stand improved the diameter distribution prediction from 1.9 to 16.9% when two common goodnessof-fit-statistics were applied to the models. Cumulative distribution function regression methods performed up to 13% better than a moment-based parameter recovery approach for estimating the parameters of the diameter distribution. The resultant models indicate that for a given set of stand conditions, plantations at northern latitudes exhibit a distribution shifted toward larger diameter classes. Equations predicted that plantations subject to intensive management at eastern latitudes similarly exhibit a distribution shifted toward larger diameter classes; however, nonintensively managed plantations at eastern locales exhibit a reverse trend: diameter distributions were predicted to shift toward smaller diameter classes, with a larger mean diameter predicted to occur at western locales. These results highlight the importance of quantifying differences in management practices and the gain from incorporating regional locale information in predicting loblolly pine growth and yield throughout its natural range.