Use of Climatic Water Balance Metrics as Site Productivity Predictors

Estimates of long-term forest site productivity are required to inform multiple forest management objectives including growth and yield assessments, silvicultural planning, and biomass/carbon projections. Estimates traditionally have been quantified in the form of site index by measuring the average height-age relationships of dominant and codominant trees or using regional site index equations. Site index implementation requires that trees are free from suppression and that height growth results from the integration of the biological determinants of growth. While useful in even-aged stands, early age height growth suppression is common in uneven-aged forest structures making existing site indices difficult to assess. Additionally, the individual biological determinants of growth are not identified and do not provide a basis for site index to be mapped across the landscape or predicted under alternative climate scenarios. This research aims to characterize the major physiographic and climatic determinants of growth. We obtained site index estimates for 203 ponderosa pine, 343 Douglas-fir, 232 lodgepole pine and 99 western larch trees throughout the state of Montana using regional equations (Milner 1992). Terrain descriptors (slope, aspect and elevation), climate normals (min/max temperatures, vapor pressure deficit), and climatic water balance (actual evapotranspiration and deficit) were derived for each site index tree at various resolutions (list range of resolutions ). Regression analysis was performed using a hierarchy of terrain, climate and mixed models. Slope, aspect, and elevation were able to explain approximately half the variation in site index for ponderosa pine, lodgepole pine and western larch. Geographically localizing the model increased the variance explained by the terrain models for all species except western larch. A simple climatic water balance interaction model (AET x DEF) was unable to explain much of the variation in site index. However, when climatic water balance was added to the terrain model the variance explained increased for all species. A biophysical model utilizing only water balance and climate variables explained more of the variation in site index than terrain based models for all species. Implications of spatial accuracy of the climatic data products and fine scale variation in tree data are discussed and recommendations for future research are provided