Temporal dynamics of rainfall using the gamma distribution for the municipality of Chimoio, Mozambique

Abstract

Knowledge of rainfall patterns improves agricultural planning and assists in the development of public policies. Thus, the objective was to estimate the monthly and annual rainfall patterns for the municipality of Chimoio, Mozambique. For this, we apply to a series of 32 years (1989-2020) of monthly and annual scale data, the Gamma probability distribution function (GPD). The estimate of rain occurrence was performed at levels of 25, 30, 40, 50, 70, 75, 80, and 90% probability. We performed the adequacy of the GPD using the Kolmogorov-Smirnov (KS) adherence tests; Chi-square and Shapiro-Wilk, at 1% significance. Our study showed that the GPD met the KS test, indicating that it can be used to estimate rainfall patterns. Historically, the rainy season in Chimoio comprises December to March, and the months from April to November correspond to the dry season. We also show that there is a 75% probability of rainfall between 30 and 100 mm in the rainy season, with higher values in December (104.3 mm) and January (133.33 mm). Annually, the average rainfall is 1027.7 mm, with a probability of 40 and 50% of rainfall equal to 1065.3 and 990.1 mm, respectively. In the period studied, we noticed that 1997 (1724 mm) and 2006 (546.6 mm) were the wettest and driest years, respectively, showing the great variability of local rainfall patterns. The GPD adequately expressed the probability of rainfall in Chimoio, and can be an important tool for agricultural planning and public policy development in the municipality.