An investigation of GSWP‐2 Mississippi River basin surface water and energy budgets

Abstract

This paper demonstrates the usefulness of watershed‐scale surface water and energy budget analysis for evaluating model skill and deficiencies. Mississippi river basin land surface model water and energy budget simulations from the Second Global Soil Wetness Project (GSWP‐2) are compared with available observations from 1986 to 1995. Forced by the same meteorological parameters, the evaluated GSWP‐2 land surface models (SSiB, Noah and CLM2‐TOP) exhibit similarities and differences in depicting basin averaged water and energy balance components. The CLM2‐TOP evaporation approaches the reanalysis and runoff is closer to the gauge measurement, while the Noah model's evaporation is underestimated and the corresponding runoff is overestimated, which could be related to the limitation in the parameterization of turbulent flux and infiltration process. The energy components of the three land surface models have a similar seasonal cycle but with different magnitudes. Despite the seasonal differences, the interannual variations of water and energy budget components are comparable and correlated. Overall, the more complex CLM2‐TOP model outperforms the SSiB and Noah models in reproducing watershed‐scale seasonal and interannual water and energy variations. All three models consistently reveal that the most extreme hydrological events are associated with El Niño‐Southern Oscillation (ENSO) phases. However, the relationship of surface water components to ENSO varies geographically. For future studies, land‐atmosphere coupled processes need to be included to determine the ability of land surface models to represent surface water and energy processes.