Biogeochemical modeling of the response of forest watersheds in the northeastern U.S. to future climate change.

摘要

In this dissertation I assessed the potential hydrochemical responses of future climate change conditions on forested watersheds in the northeastern U.S. using climate projections from several atmosphere ocean general circulation models (AOGCMs) under different carbon dioxide (CO2) emissions scenarios. The impacts of changing climate on terrestrial ecosystems have been assessed by observational, gradient, laboratory and field studies; however, state-of-the-art biogeochemical models provide an excellent tool to investigate climatic perturbations to these complex ecosystems. The overarching goal of this dissertation was to apply a fully integrated coupled hydrological and biogeochemical model (PnET-BGC) to evaluate the effects of climate change and increasing concentrations of atmospheric CO2 at seven diverse, intensively studied, high-elevation watersheds and to evaluate aspects of these applications. I downscaled coarse scale results to local watersheds and applied these values as input to a biogeochemical model, PnET-BGC. I conducted my research in this dissertation in three phases. In phase one, I used PnET-BGC to evaluate the direct and indirect effects of global change drivers (i.e., temperature, precipitation, solar radiation, CO2) on biogeochemical processes in a northern hardwood forest ecosystem at the Hubbard Brook Experimental Forest (HBEF) New Hampshire, USA. A sensitivity analysis was conducted to better understand how the model responds to variation in climatic drivers, showing that model results are sensitive to temperature, precipitation and photosynthetically active radiation inputs. Model calculations suggested that future changes in climate that induce water stress (decreases in summer soil moisture due to shifts in hydrology and increases in evapotranspiration), uncouple plant-soil linkages allowing for increases in net mineralization/nitrification, elevated leaching losses of NO3- and soil and water acidification. Anticipated forest fertilization associated with increases in CO2 appears to mitigate this perturbation somewhat. In phase two, I compared the use of two different statistical downscaling approaches- Bias Correction-Spatial Disaggregation (BCSD) (Grid-based) and Asynchronous Regional Regression Model (ARRM) (station-based) - on potential hydrochemical projections of future climate at the HBEF. The choice of downscaling approach has important implications for streamflow simulations, which is directly related to the ability of the downscaling approach to mimic observed precipitation patterns. The climate and streamflow change signals indicate that the current flow regime with snowmelt-driven spring-flows in April will likely shift to conditions dominated by larger flows throughout winter. Model results from BCSD downscaling show that warmer future temperatures cause midsummer drought stress which uncouples plant-soil linkages, leading to an increase in net soil nitrogen mineralization and nitrification, and acidification of soil and streamwater. In contrast, the precipitation inputs depicted by ARRM downscaling overcame the risk of drought stress due to greater estimates of precipitation inputs. In phase three of this research, I conducted a cross-site analysis of seven intensive study sites in the northeastern U.S. with diverse characteristics of climate, soil and vegetation type, and historical land disturbances to assess the range of forest-watershed responses to changing climate. Model results show that evapotranspiration increases across all sites under potential future conditions of warmer temperature and longer growing season. Modeling results indicate that spruce-fir forests will likely experience temperature stress and decline in productivity, while some of the northern hardwood forests are likely to experience water stress due to early loss of snowpack, longer growing season and associated water deficit. This latter response is somewhat counter-intuitive as most sites are expected to have increases in precipitation. Following increases in temperature, ET and water stress associated with future climate change scenarios, a shifting pattern in carbon allocation in plants was evident causing significant changes in NPP. The soil humus C pool decreased significantly across all sites and showed strong negative relationship with increases in temperature. Cross-site analysis among different watersheds in the Northeast indicated that dominant type of vegetation, and historical land disturbances coupled with climate variability will influence future responses of watersheds to climate change. The variability in hydrochemical response across sites is due to vegetation type, soil and geological characteristics, and historical land disturbances.