Basin-Scale Evaluation of the Land Surface Water Budget in the NCEP Operational and Research NLDAS-2 Systems

The purpose of this study is to evaluate the components of the land surface water budget in the four land surface models (Noah, SAC, VIC, Mosaic) applied in the newly implemented NCEP operational and three upgraded research versions of the North American Land Data Assimilation System version 2 (NLDAS-2). This work focuses on monthly and annual components of the water budget over twelve National Weather Service (NWS) River Forecast Centers (RFCs). Monthly gridded FLUXNET evapotranspiration (ET) from the Max-Planck Institute (MPI) of Germany, USGS total runoff (Q), changes in total water storage (dS/dt, derived as a residual by utilizing MPI ET and USGS Q in the water balance equation), and GRACE (Gravity Recovery And Climate Experiment) total water storage anomalies are used as validating reference datasets. Compared to these ET and Q benchmarks, Mosaic and SAC (Noah and VIC) in the operational NLDAS-2 overestimate (underestimate) mean annual ET and underestimate (overestimate) mean annual reference Q. The multi-model ensemble-mean (MME) is closer to the mean annual reference ET and Q. All four land surface models are able to capture the annual cycles of mean monthly reference ET, Q and change in total water storage for all 12 RFCs except for the California-Nevada RFC (CNRFC) and Colorado Basin RFC (CBRFC), which are the two driest RFCs. An anomaly correlation (AC) analysis shows good AC values for simulated monthly mean Q and dS/dt, but significantly smaller AC values for simulated ET. Upgraded versions of the models (i.e., Noah-I, SAC-Clim, VIC4.0.5) utilized in the research side of NLDAS-2 yield largely improved performance with respect to the mean annual reference ET and Q as compared to that in the operational NLDAS-2. Also, the upgraded land models largely improve the annual cycle and anomalies of the simulated monthly mean water components. The Nash-Sutcliffe efficiency metric also reveals the improved performance of the upgraded models compared to their operational versions. These results demonstrate the value of continued research to operations upgrades in the operational NLDAS product suite by identifying and overcoming models weaknesses (e.g., lack of physical process/module – groundwater, irrigation, inappropriate model parameters' values, etc.).2015-->