Interferência da instabilidade pluviométrica na previsão da produção de grãos no semiárido do Ceará, Brasil

Abstract

There were created models of rainfall distribution and prediction of endogenous variables associated with grain production in the semi-arid region of Ceará State between 1947 and 2017. The data used are from IBGE and FUNCEME. ARIMA models are used in order to predict values of areas harvested, yield and average price of grains. The effects of rainfall on the generated noise of each of these forecast models are tested. The results showed that it is possible to classify the climate of the semi - arid of Ceará State in periods of drought, normal and rainy. The prediction models presented good adjustments under statistical point of view and are influenced by the rainfall observed in the period. Contrary to the variables harvested areas and yields that showed positive coefficients, the noises associated with the grain price forecasting model fluctuate negatively with rainfall. The main conclusion of the study is that the rainfall instability interferes in the prediction of the decision variables of grain producers in Ceará State, inducing technical and economic instabilities in this production.