Detection of agricultural drought impacts on soybeans production in Brazil (1983-2016) using precipitation anomalies, NDVI and ESPI

Abstract

Droughts are one of the most common natural hazards in the world that can affect social and economic aspects. The objectives of this dissertation is: (i) to characterize agricultural drought on Brazilian soybeans producing areas in terms of frequency and severity, during 1983-2016 period, through precipitation anomalies and ENSO Precipitation Index (ESPI) event; (ii) establish the degree of significance between Normalized Difference Vegetation Index (NDVI) and precipitation and (iii) assess the impacts on production drought months with critical soybean yield periods. The data used are NDVI, using the AVHRR and MODIS sensors, monthly precipitation data from the soybean producing states, soybean production data and ESPI events. The methods consist of using the percentiles in conjunction with the time series analysis, the lower threshold percentile of 25% was calculated and 75% represented the highest threshold. The thresholds values are plotted in the time series, identifying extreme dryness and wetness as the two categories. The results of the anomalies pointed out the pluviometric variability throughout Brazil, registering events considered extremely dryness, during consecutive years (1985, 1989, 1990, 1991, 1994, 1995, 1999, 2001, 2002, 2005, 2007, 2011, 2012 and 2015). It is possible to observe that precipitation anomalies followed moderately to soybean production. Then, the precipitation anomalies are correlated with NDVI to demonstrate the efficiency of the index for each soybean production state. The year 2012 recorded the lowest NDVI value, reaching 0.53 and the precipitation anomaly was -0.91, in which it can be observed that the NDVI values are correlated with precipitation. It is compared the years of negative precipitation anomalies under the ESPI phenomenon, it is possible to observe that whenever ESPI obtains negative values it generates significant changes in precipitation patterns. Thus, this research is important for Brazil, as it provides alternatives to monitor and evaluate the impacts of agricultural drought on soybean production.