Estimating hydrological regimes from observational soil moisture, evapotranspiration, and air temperature data

Authors: KOSTER, R - Nasa Goddard Institute For Space Studies; FELDMAN, A - Goddard Space Flight Center; HOLMES, T - Goddard Space Flight Center; Anderson, Martha; Crow, Wade; HAIN, C - Nasa Marshall Space Flight Center

Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/29/2024
Publication Date: 3/1/2024
Citation: Koster, R., Feldman, A., Holmes, T., Anderson, M.C., Crow, W.T., Hain, C. 2024. Estimating hydrological regimes from observational soil moisture, evapotranspiration, and air temperature data. Journal of Hydrometeorology. 25(3):495-513. https://doi.org/10.1175/JHM-D-23-0140.1.
DOI: https://doi.org/10.1175/JHM-D-23-0140.1

Interpretive Summary: Land-surface models used in weather and climate forecasting systems characterize the exchange of water between the soil, plants and atmosphere, converting soil moisture into evaporative fluxes. The accurate modeling of these exchanges is critical to forecasts – evaporation both cools the near-surface air temperature and humidifies the atmospheric, promoting rainfall recycling. Many models use simple functional relationships between soil moisture and evaporation to govern the exchange; however, these functions – important as they are – are difficult to test for robustness over large areas using ground-based measurements. In this study, we use remotely sensed measures of soil moisture from the Soil Moisture Active/Passive (SMAP) satellite and evapotranspiration from geostationary satellites to test functional relationships over the continental United States. The results demarcate evaporative regions across the continent, from waterlimited conditions in the west to energy-limited conditions in the east. The relationships between soil moisture and 2m air temperature (a key output from forecast models) are qualitatively similar, giving confidence that by improving (or verifying) model relationships using remote sensing, we may improve our ability to forecast temperature.

Technical Abstract: Evapotranspiration has long been understood to vary with soil moisture in drier regions and to be relatively insensitive to soil moisture in wetter regions. A number of recent studies have quantified this behavior with various model and observational datasets, but given the disparate approaches and datasets used, uncertainty persists in how the underlying relationships vary in space and time. Here we complement the existing studies by analyzing two datasets as yet untapped for this purpose: a satellite-based evapotranspiration (ET) product and a meteorological station-based dataset of daily 2m air temperature (T2M) diurnal amplitudes. Both datasets are analyzed synchronously with soil moisture from the Soil Moisture Active/Passive (SMAP) satellite. We derive maps of evaporative regimes that vary in space and time within expectations: the water-limited regime grows as expected across the conterminous United States (CONUS) as spring moves into summer, only to shrink again going into winter. The relationship between the ET and soil moisture data appears particularly tight, which is encouraging given that the ET data (like the T2M data) were not constructed using any soil moisture information. The general agreement between the two independent sets of results gives us confidence that the generated maps correctly represent, to first order, evaporative regime behavior in Nature. The T2M results have the added benefit of highlighting the significant connection between soil moisture and overlying air temperature, a connection relevant to T2M predictability.