Effects of the El Niño Southern Oscillation phenomenon and sowing dates on soybean yield and on the occurrence of extreme weather events in southern Brazil

Abstract

ABSTRACT Brazil accounts for 32% of world soybean production and potential yield losses related to climate variability in the country can have an adverse effect on global supply. Recent soybean production declines in Brazil (e.g. The 15.6% soybean yield drop occurred during the 2011/12 growing season) were a consequence of extreme weather events caused by the effects of the El Nino Southern Oscillation (ENSO) phenomenon. ENSO is one of the most important atmospheric-oceanic phenomena responsible for climate and crop yield variability in several agricultural producing regions of Brazil. The greatest challenge related to adapting to ENSO climate cycles is to understand how to minimize the negative impacts and take advantage of favorable conditions to improve yields. Thus, the aim of this study was to determine the spatial and temporal impacts of sowing dates and ENSO phases on soybean yield and the frequency of extreme weather events (high, low temperature and dry period) during the vegetative and reproductive soybean phenological phases in southern Brazil. For that, soybean yields for 16 different sowing dates (from September to February) were simulated using a well-calibrated crop growth and development model (CROPGRO-Soybean) for 12 locations across southern Brazil and considering historical weather series of 56 years (1961-2016). Homogeneous regions related to ENSO impacts were defined by cluster analysis to facilitate the understanding and use of the information produced. Alternative ENSO phases were found to affect the frequency of extreme weather events and soybean yields across the region studied. Potential negative ENSO effects could be minimized by adjusting the sowing windows according to the expected climate cycle. Our results revealed that ENSO impacts on soybean yields in southern Brazil vary according to the location and can be categorized into three main groups: Group I, where the impacts of the El Nino phase on soybean yield tend to be positive, and its positive impact did not vary with the different sowing dates; Group II, where the negative soybean yield anomalies were more frequent in Neutral years; and Group III, where the positive and negative impacts of ENSO phases varied with sowing dates.