Multi-Source Soil Moisture Estimations in Southern South America

OMAR V. MÜLLER; ERNESTO H. BERBERY; ALVARO SOLDANO; DANILO DADAMIA

Frascati

Congreso; Earth Observation for Water Cycle Science; 2015

Soil moisture content has a strong influence in different land surface processes. It controls heat and water exchangges between land and lower atmosphere. Then, it is a key variable for the forecasts of temperature, air moisture and precipitation, but also its satuarion is relevant to determine runoff and then, river flows. Soil moisture is recognized as an essential climate variable by the World Meteorological Organization. It has many areas of application, such as weather forecasts, agriculture, water resources administration or ecology. Yet, measurements of soil moisture are expensive and present technical difficulties. For these reasons there are no global networks (and very few at regional scales) recording the frequency, depth and spatial resolution required by different applications. Remote sensing is an additional source of soil moisture estimation at high resolution, but only for a shallow layer. Given the lack of observations, a frequently used method involves the use of land surface models forced by either observed or atmospheric model variables. This study focuses on the temporal and spatial representation of soil moisture from a combination of a unique network of in-situ observations, existing satellite products and regional models. The current evaluation is performed taking the in-situ observations as ground truth. The analysis centers on individual estimates to identify their specidfic attributes before a general intercomparison is performed. The inter-comparison was done with four datasets of soil moisture: (1) in-situ observations from the Argentine National Space Agency in support of their mission called Satélite Argentino de Observación Con Microondas (SAOCOM/CONAE); (2) from the Global Land Data Assimilation System (GLDAS); (3) from a Land Data Assimilation System that combines remote sensing data with a land surface model; and finally, (4) from a routhine weather forecasts produce with the Weather Research and Forecasting Model coupled to the Noah Land Surface Model. The evaluation takes into account the different properties of the datasets, showing that the soil moisture estimates are in the range of variation of in-situ observations, and presenting a notable resemblance in time and space with observations. The SAOCOM mission requires producing realistic soil moisture estimates to validate its remote sensing products when they become available. The evaluation discussed here is thus the first step before developing a data assimilation system that can combine multiple products to generate soil moisture estimates at high resolution, at different depths, calibrated by observations, and physically consistent in terms of mass, energy and momentum conservation. The results reported here are part of the collaborative initiative called Joint Assessment of Soil Moisture Indicators (JASMIN) for southern South America.