Impact of SMAP Soil Moisture Data Assimilation on Soil Moisture and on Warm Season Convection Forecasts

On July 13-14, 2016, a severe squall line moved through the Ohio Valley and Great Lakes region, causing widespread high wind reports, some hail, and at least one tornado. We examine the impact of assimilating retrievals of soil moisture from the Soil Moisture Active Passive (SMAP) satellite on numerical weather forecasts of this storm. SMAP retrievals are assimilated into the Noah land surface model via an Ensemble Kalman Filter within the NASA Land Information System (LIS) framework. An additional control LIS run is conducted with no data assimilation. Profiles of soil temperature and moisture from both LIS runs are used to initialize a Weather Research and Forecasting (WRF) 48-hr forecast beginning at 00Z 13 July 2016. Soil moisture differences between the control and the SMAP assimilation simulations are significant, in part because of anomalously dry conditions in southern Ontario stemming from an unrealistic gradient near the US-Canada border in the forcing precipitation dataset. Consequently, differences in the diurnal evolution of boundary layer temperature, moisture, and Convective Available Potential Energy (CAPE) are noticeably different between the two model runs. The SMAP assimilation simulation does a better job of representing both the shape and timing of the squall line, based on comparisons to radar reflectivity observations. It also correctly suppresses a convective cell that arises over Southern Ontario, likely due to changes in the CAPE. Further validation is performed for a series of 48-hour forecasts for the May to August warm seasons of 2015 and 2016 for the entire CONUS. Validation for these forecasts is conducted by comparing precipitation forecasts to hourly Multi-Radar Multi-Sensor (MRMS) quantitative precipitation estimates (QPE), using neighborhood verification techniques over a range of accumulation intervals (3-24 hours). Near-surface temperature, dewpoint, and wind are validated against conventional point observations from the Meteorological Assimilation Data Ingest System (MADIS). An additional experiment was performed over East Africa, focusing on the time frames leading up to the two rainy seasons during Boreal Spring and Autumn months, with daily forecasts occurring from mid-February to mid-April, and mid-September to mid-November. Precipitation forecasts for East Africa are validated against the Integrated Multi-satellitE Retrievals for GPM (IMERG) satellite QPE, with temperature, dewpoint and wind validated against Global Data Assimilation System observations.