Stable Isotope Mapping to Predict Forest Ecosystem Nitrogen Retention and Responses to Fertilization.

The degree to which forest ecosystems retain added nitrogen (N), whether through fertilization or atmospheric deposition, regulates terrestrial productivity and N loading to aquatic ecosystems.  The challenge lies in identifying forests that retain added N, and in the case of managed forests, those that exhibit a positive growth response.  Patterns in the stable isotopic composition of N (δ¹⁵N) have the potential to reveal trends in, and controls over, terrestrial N cycles.  Further, recent work in imaging spectroscopy demonstrates the ability to predict foliar δ¹⁵N at leaf and canopy scales.  Applying this general framework to the two largest actively managed forest agroecosystems in the US [i.e., loblolly pine (Pinus taeda L.) in the southeastern and Douglas-fir (Pseudotsuga menziesii) in the Pacific Northwest US], the objectives of this study are to quantify relationships between leaf-level hyperspectral reflectance, natural abundance δ¹⁵N signatures, and forest ecosystem N dynamics.  We have developed species-specific multiple regression models using specific reflectance features to predict foliar δ¹⁵N with R² = 0.54 and 0.81 for loblolly pine and Douglas-fir, respectively. Nitrogen stable isotope signatures correlate significantly with fertilizer growth response across the two species, with stronger relationships in Douglas-fir than loblolly pine.  These results suggest that δ¹⁵N can be used to predict forest ecosystem N dynamics and responses to N additions, and that opportunities exist to use hyperspectral remote sensing to provide a framework for the rapid prediction of ecosystems susceptible to N leaching losses across larger spatial scales and potentially guide precision silviculture/fertilization efforts.