Instabilidades pluviométricas induzem erros de previsão na agricultura do semiárido no Ceará

Abstract

This research aims to assess the instabilities associated with rainfall and the variables that define rice, beans, cassava, and corn production in the state of Ceará between 1945 and 2020. Therefore, we applied the Coefficients of Variation and the autoregressive, integrated and moving average (ARIMA) models, respectively. The results show high rainfall instability, which spilled over into the variables that define crop yields. It was shown that the errors generated in the models used to forecast the variables that define bean and corn yields, harvested areas and rice yields, as well as cassava yields, are influenced by annual rainfall. Based on this evidence, the probable errors that can be made in predicting rice, bean, cassava, and corn yields and the rainfall levels associated with them. The evidence found in the research confirms that the production of rainfed crops in the semi-arid region is subject to various risks, the greatest of which is associated with the difficulty of predicting prices. In these cases, prices tend to be lower, at the farmer level, when rainfall is more abundant because the supply of products tends to increase, and demand does not grow in the same proportion.