AGRICULTURAL DROUGHT ASSESSMENT IN LATIN AMERICA BASED ON A STANDARDIZED SOIL MOISTURE INDEX

摘要

We propose a relatively simple, spatially invariant and probabilistic year-round Standardized Soil Moisture Index (SSMI) that is designed to estimate drought conditions from satellite imagery data. The SSMI is based on soil moisture content alone and is defined as the number of standard deviations that the observed moisture at a given location and timescale deviates from the longterm normal conditions. Specifically, the SSMI is computed by fitting a non-parametric probability distribution function to historical soil moisture records and then transforming it into a normal distribution with a mean of zero and standard deviation of one. Negative standard normal values indicate dry conditions and positive values indicate wet conditions. To evaluate the applicability of the SSMI, we fitted empirical and normal cumulative distribution functions (ECDF and nCDF) to 32-years of averaged soil moisture amounts derived from the Essential Climate Variable (ECV) Soil Moisture (SM) dataset, and compared the root-mean-squared errors of residuals. SM climatology was calculated on a 0:25 grid over Latin America at timescales of 1, 3, 6, and 12 months for the long-term period of 1979-2010. Results show that the ECDF fits better the soil moisture data than the nCDF at all timescales and that the negative SSMI values computed with the non-parametric estimator accurately identified the temporal and geographic distribution of major drought events that occurred in the study area.