Sensitivity of Rainfall to Cumulus Parameterization Schemes from a WRF Model over the City of Douala in Cameroon

Abstract

With the recurrence of extreme weather events in Central Africa, it becomes imperative to provide high-resolution forecasts for better decision-making by the Early warning systems. This study assesses the performance of the Weather Research and Forecasting (WRF) model to simulate heavy rainfall that affected the city of Douala in Cameroon during 19–21 August 2020. The WRF model is configured with two domains with horizontal resolutions of 15 and 5[Formula: see text]km, 33 vertical levels using eight cumulus parameterization schemes (CPSs). The WRF model performance is assessed by investigating the agreement between simulations and observations. Categorical and deterministic statistics are used, which include the probability of detection (POD), the success ratio (SR), the equitable threat score (ETS), the pattern correlation coefficient (PCC), the root mean square error (RMSE), the mean absolute error (MAE), and the BIAS. K-index is finally used to assess the capacity of the WRF model to predict the instability of the atmosphere in Douala during the above-mentioned period. It is found that (1) The POD, SR and ETS decrease when the threshold increases, showing the difficulty of the WRF model to predict and locate heavy rainfall events; (2) There are important differences in the rainfall area simulated by the eight CPSs; (3) The BIAS is negative for the eight CPSs, implying that all of the CPSs tested underestimate the rainfall over the study area; (4) Some of the CPSs have good agreement with observations, especially the new modifed Tiedtke and the Betts–Miller–Janjic schemes; (5) The K-index, an atmospheric instability index, is well predicted by the eight CPSs tested in this work. Overall, the WRF model exhibits a strong ability for rainfall simulation in the study area. The results point out that heavy rainfall events in tropical areas are very sensitive to CPSs and study domain. Therefore, sensitivity tests studies should be multiplied in order to identify most suitable CPSs for a given area.

Projected changes in extreme rainfall and temperature events and possible implications for Cameroon's socio‐economic sectors

“…For instance, in Cameroon, the Far North and Littoral regions are currently experiencing frequent flooding caused by extreme rainfall events during the rainy season. The last floods in these two regions were in August 2020, with more than 38,000 people affected, homes destroyed, roads eroded by erosion, bridges collapsed, agricultural plantations destroyed and the resurgence of waterborne diseases such as cholera (Final‐Report, 2021; Iroume et al, 2022; Tanessong et al, 2021). We found that the total annual rainfall amount of wet days (PRCPTOT) is projected to increase over the whole country through enhancement in the simple daily rainfall intensity (SDII) and in the rainfall amount of very wet days (R95PTOT).…”

“…The socio‐economic sectors of Cameroon are highly dependent on the climatic regime as agriculture is mainly rainfed and electricity is sourced majorly from hydroelectric dams (Nonki et al, 2019). The country's main cities, notably Douala and Yaounde as well as the Far North region, experience flooding during the rainy season (August–September) with losses estimated at several billion CFA francs (Final‐Report, 2021; Iroume et al, 2022; Tanessong et al, 2021). The western Cameroon, especially the city of Bafoussam, has not yet forgotten the disaster of 29 October 2019, due to a landslide responding to extreme rainfall, causing the death of ∼50,000 people and hundreds of animals, destroying 155 houses, numerous shops, factories and hundreds of hectares of landscape and farms (Aretouyap et al, 2021).…”