Author(s): Gogoua Habib Gogoua; Franziska Tugel; Reinhard Hinkelmann

Linked Author(s): Gogoua Habib Gogoua, Franziska Tuegel, Reinhard Hinkelmann

Keywords: Extreme rainfall events; Satellite rainfall data; Bias correction; Frequency analysis

Abstract: Extreme rainfall events (ERE) and subsequent floods are very destructive concerning their consequences on economics, infrastructure, and humans. During the last decades, the metropolitan city of Abidjan (Côte d’Ivoire) has been under severe rainstorms that impose the necessity to reevaluate the significance of ERE. This research aims to combine field and satellite rainfall estimates to define multiple scenarios for ERE through statistical analysis. For this purpose, ground rainfall data and four satellites real-time products (TRMM, CMORPH, PERSIANN, and SM2RAIN) were used. The approach consists of the statistical evaluation of the pattern of ERE temporal distribution and selecting five ERE periods for applying the linear scaling bias correction method to determine adequate satellite rainfall estimates for ERE. Next, the annual maximum daily ground rainfall from 1980 to 2019 has been adjusted with the log-Normal, log-Pearson type 3, and Gumbel distribution functions to assess the best-fit distribution functions for proceeding with the frequency analysis of ERE. Through comparison with measurements, statistical extremes, and bias-corrected satellite data, multiple ERE scenarios have been identified. The results show that from 2007 to 2019, 45% of ERE took place in May and June. CMORPH and TRMM were adjusted to the largest extent for ERE between 2014 and 2018. Gumbel distribution function performed best proven through the Chi-square and Kolmogorov-Smirnov tests. ERE scenarios with up to a 500-year return period were selected for further rainfall-runoff simulations.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221681

Year: 2022