Bridging the gap between water states and fluxes in land data assimilation systems

Authors: Crow, Wade

Submitted to: American Meteorological Society
Publication Type: Abstract Only
Publication Acceptance Date: 12/1/2020
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: The assimilation of L-band brightness temperature (Tb) from passive microwave satellite sensors has been demonstrated to improve the precision of soil moisture (SM) estimates acquired from land surface models (LSMs). However, much less success has been reported in subsequent efforts to improve analogous runoff or evapotranspiration (ET) flux forecasts. This lack of success has hampered the operational assimilation of L-band Tb into operational numerical weather and hydrologic forecasts that depend on water flux estimates. This challenge is illustrated by recent numerical weather prediction (NWP) experiments that demonstrate a degradation in 24-hour, 2-m air temperature (Ta2m) forecasts within the central United States following the assimilation of L-band Tb from the European Space Agency Soil Moisture and Ocean Salinity (SMOS) mission into the European Center for Medium-Range Weather Forecasting (ECMWF) NWP system. Results demonstrate that, while SMOS Tb assimilation clearly improves the precision of the ECMWF SM analysis, this improvement is not propagated into ET (and thus Ta2m) forecasts within the central United States. Using a combination of satellite- and ground-based estimates of SM and ET, and a synthetic twin data assimilation experiment, we demonstrate that the over coupling of SM and ET in the ECMWF LSM leads to a situation whereby late-summer improvements in the SM analysis degrade the quality of ET (and thus Ta2m forecasts). This counter intuitive result is linked to the underestimation of root-zone soil water holding capacity by the ECMWF LSM in the United Stated corn belt region. These results underscore the need to better parameterize the grid-scale relationship between land surface water states (e.g., SM) and fluxes (e.g., ET). Absent such improvement, it will provide difficult to reliably leverage data assimilation improvements in LSM states into the corresponding improvements in surface water fluxes. Prospects for using new and existing satellite-based SM and ET products for this purpose will be discussed