Evaluation of the water balance of Southeast Asia with a land surface model and era interim reanalysis
Monday, 15 December 2014
We report estimates of the surface water balance of seven Southeast Asia (SEA) river basins simulated using the Variable Infiltration Capacity (VIC) land surface hydrology model forced with climate data from the NCEP-NCAR reanalysis and the WRF regional climate model (VIC-NCEP and VIC-WRF, respectively), in addition to equivalent water balance terms from ERA Interim reanalysis (ERA-INT). We compare modeled streamflow with observations from several rivers in the region (Mekong, Sittang and Irrawaddy). The VIC streamflow simulations matched observations quite well particularly for the Mekong basin with both forcings from NCEP as well as WRF; however, the simulation using the NCEP data has slightly higher correlation as well as Nash Sutcliffe efficiency than simulation using WRF data. All of the flux terms, precipitation (P), evapotranspiration (ET), and runoff (Q) from all the three sources (VIC-NCEP, VIC-WRF and ERA-INT) had similar seasonal patterns. However, ERA-INT ET had considerably lower seasonal variability than the other two models. While the VIC model balances the surface water budget by construct, the non-closure for ERA-INT was 7.1% of the annual mean P for the SEA region overall. We compared precipitation elasticity and temperature sensitivity to assess the impact of changes in P and T on runoff. The entire SEA region taken together has an elasticity of 1.78, implying that a 1% change in P would bring on average a 1.78% change in runoff. The SEA region has a temperature sensitivity of -1.79% indicating that a 1°C increase in T brings about a 1.79% decrease in mean annual SEA runoff. Of the seven basins, Mekong runoff has the highest elasticity (2.38) and sensitivity (large negative: -2.61%); while Irrawaddy and Sittang runoffs have the lowest elasticity (1.52) and sensitivity (-0.85), respectively.