Predicting satellite-derived patterns of large-scale disturbances in forests of the Pacific Northwest Region in response to recent climatic variation

Across the Pacific Northwest, the climate between 1950 and 1975 was exceptionally cool and wet compared with more recent conditions (1995-2005). We reasoned that the changes in climate could result in expanded outbreaks of insects, diseases, and fire. To test this premise, we first modeled monthly variation in photosynthesis and growth of the most widely distributed species, Douglas-fir (Pseudotsuga menziesii), using a process-based model (3-PG) for the two periods. To compare with remotely sensed variables, we converted modeled growth potential into maximum leaf area index (LAImax), which was predicted to range from 1 to 9 across the region. On most sites, varying soil moisture storage capacity (θcap) from 200 to 300 mm while holding soil fertility constant, made slight but insignificant difference in simulated LAImax patterns. Both values of θcap correlated well with LAI estimates acquired from NASA's MODIS satellites in June, 2005 (r² = 0.7). To evaluate where 15 coniferous tree species might be prone to wide-scale disturbance, we used climatically-driven decision-tree models, calibrated in the 1950-1975 period, to identify vulnerable areas in 1995-2005. We stratified predictions within 34 recognized ecoregions and compared these results with large-scale disturbances recorded on MODIS imagery acquired between 2005 and 2009. The correlation between the percent of species judged as vulnerable within each ecoregion and the percent of forested areas recorded as disturbed with a MODIS-derived Global Disturbance Index was linear and accounted for 65 to 73% of the observed variation, depending on whether or not disturbance by fire was excluded from the analysis. Based on climate projections through the rest of the rest of the 21st century, we expect continued high levels of disturbance in ecoregions located beyond the climatically buffering influence of the Pacific Ocean.