Applicability of Predictive Models of Drought-Induced Tree Mortality between the Midwest and Northeast United States

Abstract

Regression models developed in the upper Midwest (United States) to predict drought-induced tree mortality from measures of drought (Palmer Drought Severity Index) were tested in the northeastern United States and found inadequate. The most likely cause of this result is that long drought events were rare in the Northeast during the period when inventory data were available. Therefore, new predictive models of drought mortality for the Northeast were built using USDA Forest Service inventory data and national climate data from 1969 to 2007. The Standardized Precipitation Index was better correlated with tree mortality in the Northeast than the Palmer Drought Severity Index, and new models were estimated. The reliability of the northeast models varied considerably by drought-sensitivity class, with the model for drought-intolerant species being particularly suspect. I argue that the Midwest models may nevertheless have some value in the Northeast because my tests were unable to cover the range of drought conditions under which the models were developed, there is no obvious reason why the same species should respond differently in a very similar ecological province, and some northeast models are very weak.