Use of Cosmic-ray Soil Moisture Observing System (COSMOS) to Verify and Improve Land Surface Model Output

Soil state characteristics provide crucial input for land surface interactions, hydrologic processes, boundary layer meteorology, mobility models, and climate. The Cosmic-Ray Soil Moisture Observing System (COSMOS) is a stationary sensor that uses data from thermalized neutrons to generate soil moisture observations. A number of stationary COSMOS sensors were distributed across the continental U.S. as a part of a NSF research effort in the past and many are still reporting.    Efforts are underway to develop a mobile COSMOS capable of collecting data along a route to obtain high resolution soil moisture observations with large spatial coverage in regions of interest.

The Geospatial Research Laboratory and partner organizations are collecting stationary and roving COSMOS observations. This research investigates the verification of land surface model soil moisture products using contiguous and non-contiguous COSMOS observations. Machine learning is then used to determine if Noah land surface model output can be improved through post processing when COSMOS observations are available and the length of record necessary is available. Significant impacts of this long-term project include a) an understanding of the trade-off between spot size and signal intensity for the COSMOS sensors and its impact in the verification process, b) verification of land surface model conditions using COSMOS sensors, and c) develop an understanding of the confidence intervals associated with the land surface model and downscaled data for future use in analytic tools or further modeling.