Climatic and biophysical controls on conifer species distributions in mountain forests of Washington state, USA


McKenzie, D., Peterson, D.W., Peterson, D.L., Thornton, P.E. 2003. Climatic and biophysical controls on conifer species distributions in mountain forests of Washington state, USA. Journal of Biogeography 30:1093-1108.



The purpose of this study was to quantify relationships between conifer species distributions and climatic and biophysical variables, in order to provide better insight into the potential for redistribution of species on the landscape in response to climatic change.


Data are from 10,653 georeferenced sites in Washington State, USA, along a longitudinal gradient from west of the crest of the Cascade Range to the beginnings of the western slope of the Rocky Mountains, and across two physiographic provinces, the Northern Cascades, characterized by steep, rugged topography, and the Okanogan Highlands, presenting moderate slopes and broad rounded summits.


Tree data were drawn from the USDA Forest Service Area Ecology Program database, collected in mature, undisturbed stands. We compared simple climatic variables (annual temperature, growing-degree days, annual and seasonal precipitation) to biophysical variables (soil, hydrologic, and solar radiation) derived from climatic variables. Climatic and biophysical variables were taken from the output of climatological and hydrological simulation models (DAYMET and VIC) and estimated for each plot in the tree database. Generalized linear models were used, for each of 14 tree species, at multiple spatial extents, to estimate the probability of occurrence of that species as a function of climatic and biophysical predictors. Models were validated by a combination of bootstrapping and estimating receiver operating characteristic (ROC) curves.


For the majority of species, we were able to fit variables representing both moisture and temperature gradients, and in all but a few cases these models identified a unimodal response of species occurrence to these gradients. In some cases the ecological/environmental niche of a species had been clearly captured by the model, whereas in others a longer gradient in the predictor variable(s) would be needed. Responses of most species were consistent across three spatial scales.

Main Conclusions

By identifying the ecological niches of multiple species, we can forecast their redistribution on the landscape in response to climatic change, evaluate the predictions of simulation models, and alert managers to particularly sensitive or vulnerable ecosystems and landscapes.