Multivariate assimilation of satellite-derived remote sensing datasets in the North American Land Data Assimilation System (NLDAS)

The North American Land Data Assimilation System (NLDAS) has produced over 34 years (Jan 1979 to present) of hourly land-surface meteorology and surface states, including soil moistures and temperatures, snow cover, runoff, and evapotranspiration. NLDAS uses the best-available observations and reanalyses to create near-surface forcing for land-surface models (LSMs) in “off-line” mode, but to-date has not included the assimilation of relevant hydrological remote sensing datasets. Several recent works have independently demonstrated the value of assimilating Advanced Microwave Scanning Radiometer – Earth observing system (AMSR-E) based soil moisture, AMSR-E based estimates of snow depth; Moderate Resolution Imaging Spectroradiometer (MODIS)-based estimates of Snow Covered Area (SCA); Gravity Recovery and Climate Experiment (GRACE)-based terrestrial water storage (TWS) and MODIS-based estimates of irrigation intensity. In this presentation, we will demonstrate results of assimilating these datasets in the NLDAS configuration using the NASA Land Information System (LIS), as part of the new phase of the NLDAS project. The results from individually assimilating of AMSR-E based soil moisture and snow depth into the Noah LSM indicate that systematic improvements are obtained not only in soil moisture and snow states, but also on evapotranspiration and streamflow estimates. We will also present results from the combined assimilation of the above-mentioned multi-sensor datasets in NLDAS and an evaluation of the resulting improvements and trends in soil moisture, snowpack, evapotranspiration and streamflow.