Flood frequency and amplitude analysis under changing climate scenarios in the Mono River Basin, West Africa

Abstract This study examined the flood frequencies and amplitudes of annual maximum precipitation and discharge within the Mono River Basin (MRB) in Togo, West Africa. We scrutinized daily historical data from 1961 to 2020, alongside Shared Socio-economic Pathway SSP5-8.5 scenario derived from the Coupled Model Intercomparison Project phase 6 (CMIP6) data for 2035 to 2094. The observational data were initially evaluated in conjunction with innovative CMIP6 data collected between 1985 and 2014, and employing the Pearson correlation (r), Percent of Bias (PBIAS), Standardized Root Mean Square Error (RSR), Kling–Gupta efficiency (KGE) validation statistical metrics. Subsequently, Generalized Extreme Value (GEV) probability distributions were utilized to predict flood amplitudes and frequencies, leveraging annual maximum precipitation, discharge, and return period values. The findings revealed new insights regarding flood amplitudes for discharge and precipitation; categorized as moderate, limited, or critical. This classification depended on whether the amplitude fell below or exceeded the specified thresholds under various climate conditions. However, it’s essential to note the intricacies involved in this classification process, as it reflects the complex nature of climate impacts. Our investigation uncovered strong statistical metrics between observation and CMIP6 multi-model ensemble datasets, with acceptable RSR, KGE, and r values above 0.70, and PBIAS lower than − 20%. We determined that the Gumbel probability distribution is the most accurate model for annual maximums at a 95% confidence level. Flood amplitudes were initially categorized into three distinct classes: limited, moderate, or critical (based on their amplitude). This classification was applied to the northern, central, and southern regions. However, the frequency of return periods was observed to be below 25 years for more likely floods across all locations, but they exceeded 25 years for less probable floods in certain regions and scenarios. Although this analysis provides insightful data, it also reveals complexities that warrant further exploration. We projected an increase in less severe floods in the MRB over the next 25 years, with changing flood amplitudes and increased frequencies for most south regions from past to future periods. These results highlight the pressing need to enhance early warning systems and strengthen societal resilience through sustainable policies and governance.