Contribution of frequency analysis to the predetermination of floods in the nazinon watershed at dakaye in Burkina Faso

Flood risk remains a constant threat to populations, prompting growing interest in water sciences. Within this context, the present study, “Contribution of Frequency Analysis to the Predetermination of Floods in the Nazinon Watershed at Dakaye”, aims to implement a suitable frequency model for flood prediction. The methodology involved analyzing precipitation variability, characterizing the hydrological regime, and understanding watershed behavior. A frequency analysis was then applied to maximum flow rates. The sample was constructed using the annual maximum method. Basic assumptions of frequency analysis were tested, followed by calibration and validation of various models to identify the best-fitting distribution. Rainfall analysis in the Nazinon watershed revealed an upward trend. However, the annual maxima method failed to meet the stationarity assumption. Among the tested distributions, Gumbel’s law provided the best fit. This model enabled the estimation of return periods of 2, 5, 10, 20, 50 and 100years using the method of moments. In terms of probability, a flood with a 100-year return period has a 1% chance of occurring in any given year. These quantiles are essential for designing hydraulic structures and can be used with a safety coefficient to ensure resilience. This study highlights the importance of frequency analysis in flood prediction and infrastructure planning, especially in regions experiencing increasing rainfall variability.