Defining characteristics of soil moisture from meteorological satellites

Mathematical models of the atmosphere are actively used in the modern meteorology for the numerical weather forecasts. Satellite measurements become important in data assimilation systems. Modern systems for remote sensing, efficiency data transmission, a large area covering the Earth’s surface are advantages of MetOp satellites. Using meteorological satellites is an especially topical problem for the territory of Russia where the density of the observational network is insufficient. Soil moisture has a significant impact on the formation of sensible and latent surface heat fluxes. Therefore, the accuracy of soil moisture initialization significantly affects the accuracy of the atmosphere forecast state, especially in the boundary layer. However, regular instrumental measurements of soil moisture are not made. Therefore, a topical problem is the estimate of topsoil moisture involving other observations. This paper presents comparison of surface soil moisture content measured at the reference meteorological stations, and measurements obtained by the MetOp satellite. Stations in the United States belonging to the observation network Ameriflux are chosen as stationary observation points. Stations are located in different climatic zones and selected by the type of vegetation according to the classification of IGBP. The study period is the summer months of 2007-2012. The estimate was made for the following surfaces: evergreen needleleaf forests, deciduous broadleaf forests, mixed forests, open shrub-lands, closed shrub-lands, and woody savannas. Satellite observations are converted to direct measurements using a linear relationship. The estimate of the accuracy of measurements was made. The results of measurements of the surface soil moisture using ASCAT instrument showed a good agreement with the real data of direct measurements of humidity. For the most stations the correlation coefficient is greater than 0.5. The RMS error for all stations does not exceed 0.04 m3/m3, which allows using satellite measurements of soil moisture in data assimilation systems for numerical weather prediction models.

Keywords

Authors

References