PS 1-16
Uncertainties in impact studies of future climate change on natural and agricultural ecosystems by using modeled climate data with bias-correction

Monday, August 5, 2013
Exhibit Hall B, Minneapolis Convention Center
Mingliang Liu, Civil and Environmental Engineering, Washington State University, Pullman, WA
Jennifer C. Adam, Civil and Environmental Engineering, Washington State University, WA
Jennie Stephens, Clark University
Kirti Rajagopalan, Washington State University
Serena H. Chung, Laboratory for Atmospheric Research, Washington State University
Xiaoyan Jiang, NCAR, NCAR
Tsengel Nergui, LAR, WSU
John A. Harrison, School of Earth and Environmental Sciences, Washington State University Vancouver, Vancouver, WA
Alex Guenther, Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO
Christina L. Tague, Bren School of Environmental Science and Management, University of Calfornia, Santa Barbara, Santa Barbara, CA
Julian J. Reyes, Civil and Environmental Engineering, Washington State University
Background/Question/Methods

Bias-correction (BC) of modeled climate data as a post-process is widely used for climate change impacts (CCI) studies at global and regional levels. However, bias-correction could cause uncertainties in estimating changes in hydrological and biogeochemical processes in terrestrial ecosystems over future climate conditions due to the impaired spatial-temporal covariance of climate variations (such as temperature and precipitation) and physical conservation principles from BC process. Here we quantifies the differences of changes in simulated water variables (ET, runoff, snowpack water equivalent (SWE), and water demand for irrigation), crop yield, VOC and NO emissions, and dissolved inorganic nitrogen (DIN) export over the Pacific Northwest and gross primary production (GPP) over a watershed scale (HJ-Andrews). The model climate data were from WRF model runs with ECHAM -5-A1B scenario as boundary condition and a bunch of component models from a regional earth system model (BioEarth) were run as either offline or partially coupled approaches.  They include a macro-scale hydrological model (VIC), a coupled agricultural and hydrological model (VIC-CropSyst), an ecohydrological model (RHESSys), a gases and aerosols emission model (MEGAN),  and  nutrient leaching and export model (NEWS).  Series simulation experiments were conducted by using BC climate (on temperature and precipitation), and Non-BC climate.

Results/Conclusions

Simulation results demonstrate that BC and Non-BC treatment on climate data could provide consistent estimations on delta change during 1980s-2020s and 1980s-2050s in water fluxes (ET, runoff, and water demand), VOC (isoprene and monoprenes) and N emissions, mean crop yield (irrigated and non-irrigated), and DIN over regional scales such as the PNW domain as a whole. However, the BC and Non-BC treatment of climate data could lead to significant differences in SWE, crop yield from dry land, and ET from HG-Andrews. Even though most of the included variables have no significant differences between BC and Non-BC process at regional scale, there are large spatial variations across the study domain. In additions, certain months are presented with significant differences in delta change estimations from BC climate data. Sensitivity analysis indicated that BC treatment on precipitation and BC treatment on temperature almost contribute the same to these differences at region scale. We conclude that there are trade-offs by using BC climate data for offline CCI studies and coupled regional earth system model runs for regional studies.