We propose to produce an operational data assimilation (DA) system for optimal integration of thermal infrared (TIR) and microwave (MV) soil moisture (SM) information and near real-time green vegetation fraction (GVF) into the Noah land-surface model component of the National Land Data Assimilation System (NLDAS). NLDAS produces hydrologic products (e.g. soil moisture, evapotranspiration, and runoff) used by NCEP for operational drought monitoring, but these products are sensitive to model input errors in soil texture (affecting infiltration rates) and prescribed precipitation rates. These types of model errors can be compensated for by periodically updating SM state variables in LSMs through assimilation of remotely sensed SM information. The work proposed here will build on a project currently funded under the Climate Test Bed Program entitled “A GOES Thermal-Based Drought Early Warning Index for NIDIS”, which is developing an operational TIR SM index (Evaporative Stress Index; ESI) based on maps of the ratio of actual-to-potential ET (fPET) generated with the Atmosphere-Land Exchange Inverse (ALEXI) surface energy balance algorithm.
The assembled research team has demonstrated that diagnostic information about SM and evapotranspiration (ET) from MW and TIR remote sensing can significantly reduce SM drifts in LSMs such as Noah. The two retrievals have been shown to be quite complementary: TIR provides relatively high spatial resolution (down to 100 m) and low temporal resolution (due to cloud cover) retrievals over a wide range of GVF, while MW provides relatively low spatial (25 to 60 km) and high temporal resolution (can retrieve through cloud cover), but only over areas with low GVF. Furthermore, MW retrievals are sensitive to SM only in the first few centimeters of the soil profile, while in vegetated areas TIR provides information about SM conditions integrated over the full rootzone, reflected in the observed canopy temperature. The added value of TIR over MW alone is most significant in areas of moderate to dense vegetation cover where MW retrievals have very little sensitivity to SM at any depth.
Building on this work, the proposed study will develop an optimal strategy for assimilating TIR and MW SM signals into the Noah model over the NLDAS domain using the Land Information System (LIS) developed by NASA. Additionally, near real-time green vegetation fraction (GVF) data products generated in NESDIS will be ingested, replacing climatological fields currently used in NLDAS, which are not always representative of actual conditions on the ground, especially in areas suffering from drought. We propose to use relative TIR / MW skill maps developed by Co-I Hain to spatiotemporally modify error characteristics needed by the EnKF as a function of GVF.
Assimilation results will be validated in comparison with in-situ SM observations and using a data denial validation methodology. Outputs from the operational DA system will include near real-time (updated each night) maps of surface and root-zone SM, ET and runoff. Anomalies computed from these improved hydrologic products will be compared to ALEXI ESI and standard drought metrics, including the operational NLDAS output. Output will be distributed in real-time to NCEP-CPC for use in the North America Drought Briefing and to the National Drought Mitigation Center in support of the U.S. Drought Monitor.