Evaluation of NLDAS-2 Simulated Daily Soil Moisture

Abstract:

Soil moisture observations from multiple networks located in different climate and biome conditions in the United States were used to evaluate three simulated soil moisture products. The three land surface models: Noah, Mosaic, and Variable Infiltration Capacity (VIC), were run within the North-American Land Data Assimilation System Project Phase 2 (NLDAS-2), with a 1/8^th degree spatial resolution and daily temporal resolution. Soil moisture observations from more than 50 stations, all with at least 12 continuous years of data, were used to evaluate model soil moisture at three different depth intervals in the soil column: 0 – 10 cm, 10 – 40 cm, and 40 – 100 cm. In addition, we evaluated the ability of the models to capture the temporal variations of the in situ observations in the context of model architecture and land surface parameterization. We find the similarity of in situ and model soil moisture varies both temporally and spatially, attributed to the representativeness of the point soil moisture observation compared to the model grid cell. In general, our results suggest the three components of the NLDAS-2 project can accurately simulate soil moisture seasonal variation, but do show systematic wet and dry bias at several stations across the contiguous United States.