Analysis of the recent evolution of dry spells in Burkina Faso

Abstract

Changes in the distribution of precipitation frequency and intensity can affect dry and wet spells, which will have an impact on climate-sensitive sectors, most notably agriculture. This study set out to analyze the variations of onset and cessation dates of the rainy season, to understand dry spells nature and their drivers in Burkina Faso. We used daily rainfall data of 10 synoptic stations from the National Meteorological Agency, the Madden-Julian Oscillation (MJO) and El Nino-Southern-Oscillation (ENSO) indices for the period 1991-2020 are then used to identify the onset and the cessation of the rainy season and their anomalies using the definitions of Sivakumar and (Kasei and Afuakwa) for season onset dates, and the definitions of Maikano and (Kasei and Afuakwa) for the season cessation dates, dry spells frequencies, longest dry spells, average dry spells duration and the modulation of the MJO and the ENSO in the occurrence of dry spells. All these parameters were computed between the onset and the cessation of the season and these dates were determined from May to October. Statistical analysis was made using Kendall's Nonparametric Test for Monotonic Trend, standard deviation, and simple linear model to assess trends in these variables. We used R programming language to extract and analyze the dry spells and the two modes, the onset and cessation dates are computed using second-order Markov chain with Instat software. We found that both changed with rainfall onsets having a more important change. The longest dry spells are identified at the cessation of the season compared with the onset of the season and there is a moderate frequency in the occurrence of DS. Across the studied period, most dry spells fluctuate between 5 and 15 days. Some phases of the MJO such as phases 5 and 7 increase the likelihood of DSs and phases 1 and 2 decrease it. At the interannual scale, the ENSO influences DS characteristics with El Nino that increase the likelihood of the DSs occurrence. To help different stakeholders with their decision-making, it would be helpful to identify the distributions and patterns of critical dry spells and explore the possibility of predicting the patterns of dry spells in the future.