Trend analysis of the climate database used in agricultural climate risk zoning

Abstract

The objective of this work was to identify trends in the series of historical rainfall, temperature, and evapotranspiration, considering the meteorological database used in studies on agricultural climate risk zoning (ZARC). The daily meteorological databases used in ZARC were compiled from about 4,000 station points over a continuous period from 1984 to 2022. The trend analysis was performed using Mann-Kendall’s test and Sen’s slope estimator for the entire Brazilian territory. The results highlighted the emergence of marked spatial and temporal climate trends in Brazil, with a potential impact on agricultural risks. Although more than 79% sites have shown no significant rainfall trends, a decreasing pattern was observed mainly at the beginning of the rainy season in August-September over the central region of Brazil, as well as a decrease in rainfall in April and an increase in October, in the South of the coutry. This decline in the central region coincides with a significant rise of minimum/maximum temperatures and evapotranspiration, leading to an extension of the dry season. The spatial-temporal change in rainfall associated with an annual increase in temperature (5%) and evapotranspiration (10%), could significantly affect agricultural production across many grain-producing regions in Brazil, particularly in areas where more than one crop is grown per season.