Hydrologic Sensitivity to Climate Forcing and Spatial Resolution: A Study on the Mekong River Basin Using the Community Land Model (CLM5.0)

Abstract:

Hydrological modeling suffers from various sources of uncertainty. Meteorological forcing data, representation of the natural and anthropogenic processes, and spatial resolution are some of the key determinants of model performance. In this study, we analyze various hydrologic fluxes and state variables simulated by the Community Land Model (CLM5.0) at 0.05 and 0.5 degree spatial resolution over the Mekong river basin for various time scales (inter-annual and decadal) during 1979 to 2016 period. To account for forcing uncertainty, the model is driven by three different meteorological forcing datasets: (1) WATCH Forcing Data methodology applied to ERA-Interim reanalysis data (WFDEI), (2) The Global Soil Wetness Project Phase 3 (GSWP3), and (3) ERA5 the fifth generation ECMWF atmospheric reanalysis of the global climate available at a various temporal and spatial resolution. MOSART river routing scheme and the global scale hydrodynamic model CaMa-Flood are implemented using runoff derived from CLM5.0 to understand how the model structure influences the hydrograph peak and low flow. We use transient land use to consider land use and land cover changes over the years. Model simulated streamflow was validated against observed streamflow obtained from the Mekong River Commission (MRC), and Terrestrial water storage (TWS) was validated with the data from the GRACE satellites for the 2002 to 2016 period. Further, we use several precipitation data, including Tropical Rainfall Measuring Mission (TRMM) and Multi-Source Weighted-Ensemble Precipitation (MSWEP), to quantify the effects of precipitation in improving the simulation performance. We found significant deviation in streamflow, runoff, evapotranspiration, and TWS from the ensemble mean due to the uncertainty in the climate forcing data and model resolution. Precipitation uncertainty significantly affected soil moisture, groundwater recharge, and irrigation water use in the Mekong river basin. Although some uncertainty can be attributed to other human interventions in the recent decade when most of the dams were constructed, our results underscore the role of precipitation data for a realistic simulation of hydrologic fluxes and states.