Climate impacts on regional ecosystem services in the United States from CMIP3-based multimodel comparisons

Climate Research

Benjamin Felzer¹  Dork Sahagian¹ 

关键词: Inter-model comparison;    IPCC scenarios;    Regional downscaling;    Ecohydrology;   

DOI  :  10.3354/cr01249

来源: Inter-Research Science Publishing

【 摘 要 】

ABSTRACT: Projections of surface hydrology and local ecosystem responses to expected climate change in the 21st century can inform regional planners and land use managers in the broader context of climate change adaptation at regional scales. We use bias-corrected and downscaled projections for 3 IPCC scenarios (B1, A1B, and A2) to assess projected climate impacts on ecosystem function and services for different regions of the conterminous USA utilizing the Terrestrial Ecosystems Model version Hydro. Significance of model trends is analyzed for each of the 6 US National Climate Assessment megaregions for several climatological, hydrological, and ecological variables based on projections and consistency among the multimodel ensemble from phase 3 of the Coupled Model Intercomparison Project (CMIP3). Our regional analysis reveals that there are some robust and significant trends that can be useful to decision-makers, and that these trends are specific to each region, as each region responds to climate forcing differently in ways that reflect emergent behavior from the interaction of climate, ecosystem, and surface processes. Generally, runoff is simulated to increase in winter and decrease in summer throughout the northern USA, snowpack is reduced everywhere, and net primary productivity and maize yield increase except where limited by moisture. Model reconstructions of magnitudes and directions of some historical regional trends are incorrect, so predicted reversals may be spurious. Some model variables such as precipitation show no significant projected trends, yet in concert, they control the responses of other variables such as soil moisture, in which the trends are projected to be significant. As such, variables whose trends are less observable may be revealed by other variables controlled by them, and can thus be used as proxies to enhance predictive capacity.

【 授权许可】

Unknown