Effects of land use, climate variation, and N deposition on N cycling and C storage in northern hardwood forests

Abstract

We hypothesized that much of the variability in dissolved inorganic nitrogen (DIN) loss from forested catchments can be explained by land use history and interannual climatic variation, and that these factors determine the degree to which N deposition results in increased storage of C in forests. We used an existing model of C, N, and water balances in forest ecosystems in conjunction with long‐term climate and N leaching loss data from several northern hardwood forest ecosystems to predict the effects of land use, climate variability and N deposition on C storage and N cycling and loss. Six sites from the White Mountains of New Hampshire with very different land use histories and annual stream DIN losses were used. The only model parameter that varied between sites was land use or disturbance history. Each site was simulated using both mean climate data for each year and actual time series climate data. Vegetation removal resulted in a period of increased DIN leaching, followed by losses below those in control stands for both measured and simulated data. One site with an extreme fire event over 170 years ago still showed reduced N losses in both modeled and measured data. Significant interannual variation in DIN loss is evident in the field data. Model predictions using actual climate time series data captured much of this variation. This high interannual variability along with the slow rate of change in DIN loss predicted by PnET‐CN using mean climate throughout the simulations suggests that statistically significant increases in DIN leaching losses due to long‐term increases in N deposition will not be detectable for several decades, given current rates of N deposition. N deposition increased C storage in all simulations, but the quantity stored was about 50% that predicted by another published model. This difference results from differences in the efficiency with which added N is retained in the ecosystem. The previous model used an 80% retention value, while retention was closer to 50% over most of the time period examined here.