Variabilité pluviométrique en zone soudanienne au Niger de 1960 à 2022 : cas de la ville de Gaya

Abstract

Rainfall variability in the Sudanian zone of Niger from 1960 to 2022: The case of the city of Gaya

This study is based on the configuration of an algorithm for which we sought a maximum of four break points in the level of the annual rainfall totals (average) of the rainfall observed in Gaya located in the Sudanian climatic zone. The algorithm is based on the single change point search method extended to several change points by iteratively repeating the single point method on the different subsets obtained from the first iteration. The trends obtained are only representative when they are statistically significant at the 5% threshold according to the non-parametric Mann-Kendall test. This test was applied to the time series of anomalies of the Standardized Precipitation Index (SPI) over the study period (1960-2022). Thus, the SPI was used to determine dry (SPI < -0.5), normal (-0.5 < SPI < 0.5) or wet (SPI > 0.5) years. In addition, exceptionally wet (SPI > 1) or dry (SPI < -1) years were also sought, suggesting the existence of more than one break point as in the previous method. From precipitation anomalies, we looked for the existence of trends and/or periods of stationarity by identifying break points. The evolution of precipitation anomalies does not indicate a period of abrupt change with a non-significant increasing trend over the study period (1960-2022). Two data processing methods were applied. These are principal component analysis (PCA) and coefficient of variation (CV). The PCA result shows that the rainfall is homogeneous over the study period in Gaya. The coefficient of variation of rainfall is low (CV = 13.62% < 15%), which indicates a low variation in rainfall. Although the station is well watered, the evolution of annual precipitation highlighted the dry period of the 1970s and 1980s and the resumption of rainfall in the 1990s.

Keywords: Rainfall; Trend; Variability; Gaya.