Evaluation of Advanced Land Surface Models in the NLDAS Testbed

Land surface models (LSMs) have been developed dramatically over the past years by extending single-layer to multi-layer snow model, adding groundwater component, carbon and nitrogen cycles, dynamic vegetation processes, urbanization and water quality sub-models. These additions have advanced the intermediate LSMs currently used in NLDAS-2 to more complex LSMs. This study focused on evaluating the performance of both intermediate and advanced LSMs in simulating water fluxes and soil moisture by using the datasets and tools developed from the NLDAS-2 (North American Land Data Assimilation System Phase 2) testbed. Community Land Model version 4 (CLM4) and Community Noah LSM with Multi-Parameterization Options (Noah-MP) are two advanced land surface models as they have incorporated most of these developments; while Noah and VIC are two intermediate complex models. GRACE (Gravity Recovery and Climate Experiment) terrestrial water storage data is utilized to evaluate the performance of the four LSMs in modeling the total terrestrial water fluctuation. USGS streamflow from the 961 small basins, the monthly mean MODIS and Fluxnet Evapotranspiration (ET), and the SCAN (Soil Climate Analysis Network) soil moisture are also used to evaluate the performance of four models. Preliminary results show that all four land surface models can capture the terrestrial water cycle and capture the broad spatial and temporal features of water fluxes such as total runoff and ET. Mean relative biases analysis shows that Noah overestimates (underestimates) mean annual observed streamflow (ET), Noah-MP underestimates (overestimates) mean annual streamflow (ET), and CLM4 and VIC are close to the observations. Overall, both advanced models perform more robustly than current NLDAS-2 Noah LSM. This study will help identify the direction for the development of next-generation land surface models in the future and help identify the candidates of LSMs for next generation NLDAS system.