Hydrological evaluation of the Noah-MP land surface model for the Mississippi River Basin

[1] This study evaluates regional-scale hydrological simulations of the newly developed community Noah land surface model (LSM) with multiparameterization options (Noah-MP). The model is configured for the Mississippi River Basin and driven by the North American Land Data Assimilation System Phase 2 atmospheric forcing at 1/8° resolution. The simulations are compared with various observational data sets, including the U.S. Geological Survey streamflow and groundwater data, the AmeriFlux tower micrometeorological evapotranspiration (ET) measurements, the Soil Climate Analysis Network (SCAN)-observed soil moisture data, and the Gravity Recovery and Climate Experiment satellite-derived terrestrial water storage (TWS) anomaly data. Compared with these observations and to the baseline Noah LSM simulations, Noah-MP shows significant improvement in hydrological modeling for major hydrological variables (runoff, groundwater, ET, soil moisture, and TWS), which is very likely due to the incorporation of some major improvements into Noah-MP, particularly an unconfined aquifer storage layer for groundwater dynamics and an interactive vegetation canopy for dynamic leaf phenology. Noah-MP produces soil moisture values consistent with the SCAN observations for the top two soil layers (0–10 cm and 10–40 cm), indicating its great potential to be used in studying land-atmosphere coupling. In addition, the simulated groundwater spatial patterns are comparable to observations; however, the inclusion of groundwater in Noah-MP requires a longer spin-up time (34 years for the entire study domain). Runoff simulation is highly sensitive to three parameters: the surface dryness factor (α), the saturated hydraulic conductivity (k), and the saturated soil moisture (θmax).