Confronting CHIRPS dataset and in situ stations in the detection of wet and drought conditions in the Brazilian Midwest

Abstract

The states of Mato Grosso (MT) and Mato Grosso do Sul (MS) are located in Midwest Brazil and are dependent on agribusiness, which makes their water regimes of fundamental importance. However, the existing weather stations in these states are limited, and rainfall products are therefore a valuable alternative. The objectives of this study are: (a) to validate the CHIRPS datasets for the states of MT and MS in the Brazilian Midwest region; (b) to evaluate the intraseasonal variability of regional rainfall via CHIRPS; and (c) to define the drought and wet periods on the decennial scale, based on the annual SPI. Three seasons were defined (rainy, drought, and transition), and the differences showed that the MS rainy season (56.65%) was reduced due to the veranicos that occur in February, with an increase in the percentages of the drought (13.83%) and transition (29.52%) seasons. The highest monthly rainfall occurs in the north of MS (328–390 mm), while intermediate rainfall (109–273 mm) occurs in the south of MS. This satisfactorily represents the formation of a rainfall gradient in a N–S direction during the rainy season, similar to other characterizations around the world. Based on the annual SPI assessed per decade, there were moderate droughts annually in the 1990s, 2000s, and 2010s, but not in the 1980s. There were significant differences between the decades due to the intensity and categorization of the ENSO phases (strong and moderate), and drier decades with moderate rainfall reduction in both states can be defined in the last 40 years. The data obtained via CHIRPS are satisfactory for the spatio-temporal evaluation of regional rainfall, and for the use of SPI to detect drought and wet periods in the states of MT and MS.