Use of SMOS L3 soil moisture data: validation and drought assessment for Pernambuco State, Northeast Brazil

Abstract

The goal of this study was to validate soil moisture data from Soil Moisture Ocean Salinity (SMOS) using two in situ databases for Pernambuco State, located in Northeast Brazil. The validation process involved two approaches, pixel-station comparison and areal average, for three regions in Pernambuco with different climatic characteristics. After validation, the SMOS data were used for drought assessment by calculating soil moisture anomalies for the available period of data. Four statistical criteria were used to verify the quality of the satellite data: Pearson correlation coefficient, Willmott index of agreement, BIAS, and root mean squared difference (RMSD). The average RMSD calculated from the daily time series in the pixel and the areal assessment were 0.071 m3m-3 and 0.04 m3m-3, respectively. Those values are near to the expected 0.04 m3m-3 accuracy of the SMOS mission. The analysis of soil moisture anomalies enabled the assessment of the dry period between 2012 and 2017 and the identification of regions most impacted by the drought. The driest year for all regions was 2012, when the anomaly values achieved -50% in some regions. The use of SMOS data provided additional information that was used in conjunction with the precipitation data to assess drought periods. This may be particularly relevant for planning in agriculture and supporting decision makers and farmers.