Application of the Analog Ensemble Methodology to Land Surface Conditions

Soil state characteristics provide crucial input for land surface interactions, boundary layer meteorology, mobility models and drought prediction. Worldwide, there has been limited mapping of soil state characteristics. This changed with the launch of the soil moisture active passive (SMAP) satellite in January 2015. SMAP gathers observations on soil moisture and freeze/thaw state at varying resolutions, globally. The Cosmic-Ray Soil Moisture Observing System (COSMOS) sensor thermalizes and measures neutrons above the ground. The intensity of these neutrons is correlated with hydrogen content in soils, the largest of which is comprised of water. The Geospatial Research Laboratory and other partner organizations have collected stationary and roving COSMOS probe observations at specific locations and regions, respectively. This research investigates application of the Analog Ensemble (AnEn) methodology, developed by a team at the National Center for Atmospheric Research, to better understand uncertainty associated with soil state predictions.

The AnEn uses a set of past observations corresponding to the best analogs of a deterministic prediction model to generate a probability distribution of future environmental states: an ensemble of analogs. In the meteorological field, this methodology has proven effective at improving short-term prediction of wind speed and solar irradiance, downscaling and providing an understanding of the uncertainty associated with predictions but has never been applied to soil moisture. Land surface model output is used for historical deterministic predictions. Corresponding observational datasets come from the SMAP satellite and collected COSMOS data. This research attempts to understand the range of uncertainty associated with soil state characteristics by creating a probabilistic representation from deterministic output at minimal computational cost.