Physiological responses to fertilization recorded in tree rings: isotopic lessons from a long-term fertilization...

Nitrogen fertilizer applications are common land use management tools, but details on physiological responses to these applications are often lacking, particularly for long-term responses over decades of forest management. We used tree ring growth patterns and stable isotopes to understand long-term physiological responses to fertilization using a controlled fertilization experiment begun in 1964 in Washington State (USA), in which three levels of nitrogen fertilizer were applied: 157, 314; and 471 kg/ha. Basal area increment (BAI) increased more than fourfold in the highest treatment to twofold in the lowest, and a significant increase in BAI was observed for 20 years. Latewood delta 13C sharply decreased by 1.4 per thousand after fertilization and was significantly lower than controls for four years, but no differences existed between fertilization levels, and the effect disappeared after four years, indicating that intrinsic water use efficiency (A/gs) increased in response to fertilization. Earlywood delta 13C showed similar trends but was more variable. Latewood delta 18O increased significantly above controls by approximately 2 per thousand in all treatments, but the duration differed with treatment level, with the effect being longer for higher levels of fertilization and lasting as long as nine years after fertilization. Because source water and relative humidity were the same between experimental plots, we interpreted the delta 18O increase with treatment as a decrease in leaf-level transpiration. Earlywood delta 18O did not show any treatment effects. Because the Pacific Northwest has a mediterranean climate with dry summers, we speculated that fertilization caused a substantial increase in leaf area, causing the trees to transpire themselves into drought stress during the late summer. We estimate from the delta 18O data that stomatal conductance (gs) was reduced by approximately 30%. Using the delta 13C data to estimate assimilation rates (A), A during the late season was also reduced by 20-30%. If leaf-level A decreased, but BAI increased, we estimated that leaf area on those trees must have increased by fourfold with the highest level of treatment within this stand. This increase in leaf area resulting from fertilization caused a hydraulic imbalance within the trees that lasted as long as nine years after treatment at the highest levels of fertilization.