Parameter Sensitivity and Transferability for Simulating ET and GEP of Sagebrush Ecosystems across a Climate Gradient

Environmental gradients exert controls on water, carbon, and energy fluxes across montane landscapes, impacting the magnitude and timing of evapotranspiration and carbon uptake. Ecohydrology models are essential tools to better understand and quantify water and carbon fluxes for different ecosystems and across climate gradients. Parameter uncertainty limits the ability of these models to simulate ecosystem processes. The focus of this study was to: evaluate the sensitivity of input parameters an ecohydrology model for simulating evapotranspiration (ET) and gross ecosystem productivity (GEP) for four sagebrush ecosystems across an elevation/climate gradient; and asses the transferability of parameter values across the climate gradient. The Simultaneous Heat and Water (SHAW) model was applied to four sites within the Reynolds Creek Critical Zone Observatory located at elevations of 1425, 1680, 1808 and 2111 m with mean annual precipitation of 290, 337, 425, and 795 mm. Annual average GEP for the sites is 420, 615, 665, and 849 gC m-2. Root mean square deviation (RMSD) of simulated daily ET using a priori parameters were 0.49, 0.49, 0.43, and 0.56 mm and for GEP was 0.72, 0.88, 2.08, and 1.02 gC m-2 d-1. A Monte-Carlo approach of 10,000 to 20,000 model runs was used to assess sensitivity and distribution of parameter values for each site. Parameters found to be sensitive were optimized using a coupling of the Parameter Estimation & Uncertainty (PEST) software with the SHAW model. Post-optimization RMSD values for ET were 0.37, 0.47, 0.48, and 0.53 mm, and for GEP were 0.53, 0.57, 0.68, and 0.76 gC m-2 d-1. Parameter value distributions were found to site-specific for some parameters while others were found to be transferrable from site to site.