Understanding spatial and temporal Douglas-fir fertilizer response in the Pacific Northwest using boosted regression trees and linear discriminant analysis

Fertilizer response of Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) plantations can vary considerably throughout the Pacific Northwest due to differences in soil and site productivity. In this study, we calculated significant two-, four-, and six-year cumulative volume growth response per tree due to urea fertilization on 71 Douglas-fir installations using a paired t-test. To understand the biogeoclimatic factors affecting fertilizer response, climate, location, soil, and productivity predictor variables were used in boosted regression tree (BRT) and linear discriminant (LDA) analyses to produce models that could predict significant regional response to fertilization. The ability to predict significant volume response decreased with time since fertilization (71-37% from 2 to 6 years), yet installations that were predicted to respond in all years of BRT models had similar response to significantly responding installations (>5000 cm3/yr). The most common predictors of volume response in both model types and all measurement periods were high elevation (>400 m), low-moderate site index (<42 m at 50 years), and cold winter temperatures (<4.4 C). The Oregon (OR) Klamath Mountain region contained the greatest coverage of predicted volume response. The Middle Cascade Mountains also had large areas predicted to respond, but in this study OR installations responded better than Washington (WA) installations. The Coast Range, Olympic Mountains, Puget Trough, and Northern Cascade Mountains were predicted to have very little response. This study demonstrates the spatial relationships between climate, soil, and productivity variables that indicate fertilization response across the coastal Pacific Northwest.