Impact of Soil Moisture Active Passive Data Assimilation on Short-Term Numerical Weather Prediction during Warm Seasons

The Short-Term Prediction Research and Transition (SPoRT) Center at NASA Marshall Space Flight Center manages near-real-time runs of the Noah Land Surface Model (LSM) within the NASA Land Information System (LIS) over Continental U.S. (CONUS) and Eastern Africa domains. Soil moisture products from the CONUS model run are used by several NOAA/National Weather Service Weather Forecast Offices for experimental applications such as drought, flood, and wildfire situational awareness. The baseline “SPoRT-LIS” configuration is the Noah LSM driven by atmospheric and radar-gauge blended precipitation analyses, and input satellite-derived real-time green vegetation fraction on a ~3-km grid for the CONUS. This configuration is being enhanced by including data assimilation of Level 2 Soil Moisture Active/Passive (SMAP)-Enhanced soil moisture retrievals in a parallel SPoRT-LIS run beginning on 1 April 2015. This presentation will show results of daily short-term numerical weather prediction (NWP) sensitivity simulations using the NASA Unified-Weather Research and Forecasting (NU-WRF) modeling framework. A companion presentation at this annual meeting (Blankenship et al.) addresses the details of the Ensemble Kalman Filter data assimilation methodology within offline LIS simulations and soil moisture validation over the CONUS domain.

The NU-WRF sensitivity simulations are conducted by initializing all land surface variables with either the SPoRT-LIS simulation output, or the SMAP-Enhanced data assimilation output (hereafter SMAPENHDA). All other initial and boundary conditions are provided by the NCEP Global Forecast System (GFS) NWP model. The period of study for the NWP sensitivity simulations span May through August 2015 and 2016 for a domain at 9-km grid spacing spanning the full CONUS, along with a 3-km convection-allowing nest covering the central and eastern United States. A similar nested configuration is being designed for East Africa NWP simulations, with the period of study emphasizing transition periods from dry to wet seasons (e.g., February through April and October/November). Short-term forecasts of 2-3 days in length are conducted to examine the impacts of the SMAPENHDA on the NU-WRF model results compared to SPoRT-LIS-initialized NU-WRF runs. Results are examined for select high-impact case studies, and bulk verification statistics (emphasizing surface variables and precipitation) are generated for each season to quantify the influence of land surface fields from SMAP data assimilation on subsequent short-term NWP for seasonal scales.