Analyzing the Hydrological Disconnect Between Extreme Precipitation and River Discharge in Douala, Cameroon

Analyzing the Hydrological Disconnect Between Extreme Precipitation and River Discharge in Douala, Cameroon | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Analyzing the Hydrological Disconnect Between Extreme Precipitation and River Discharge in Douala, Cameroon Calvin Padji, Cyrille Meukaleuni, David Monkam This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8073038/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study investigates the critical temporal disconnect between extreme rainfall and subsequent flood discharge in Douala, Cameroon, a rapidly urbanizing coastal city increasingly vulnerable to flooding. Using a 13-year dataset (2010–2023) that integrates CHIRPS precipitation data Funk et al. ( 2015 ) and ERA5-Land discharge data (Muñoz-Sabater et al. 2021 ), we applied Extreme Value Theory (EVT) through a Peak-Over-Threshold (POT) framework to quantify and compare return levels and return periods for both meteorological and hydrological extremes. Our analysis reveals a marked hydrological transformation: while extreme rainfall exhibits heavy-tailed behavior (shape parameter ξ = 0.19), discharge extremes show even heavier tails (ξ = 0.44), suggesting that urban watershed processes amplify flood risks beyond what rainfall statistics alone would predict. Cross-correlation and event-based analyses indicate a systematic temporal lag of 1–2 days between precipitation peaks and discharge responses, with weak coupling (r ≈ 0.2–0.4) between daily rainfall and runoff intensity. Moreover, a 100-days return period rainfall event (98.08 mm) generates only a 37.93 mm·day⁻¹ discharge peak, reflecting over 60% attenuation between precipitation input and hydrological output. Diagnostic plots, threshold stability checks, and goodness-of-fit tests confirmed the robustness of the Generalized Pareto Distribution (GPD) models. The inclusion of return date analysis, following the methodology proposed by Padji et al. ( 2024 ), demonstrated the approach’s predictive capability for estimating the timing of flood events rather than their magnitude alone. These findings reveal that Douala’s flood hazard is not solely determined by rainfall extremes but is amplified by rapid urbanization, reduced infiltration, and altered drainage dynamics. The combination of weak rainfall–discharge coupling, temporal lag, and hydrological attenuation underscores the urgent need for integrated flood forecasting and water management strategies that explicitly incorporate urban hydrological dynamics rather than relying exclusively on rainfall forecasts. Extreme Rainfall Flood Discharge Return Period Temporal Lag Urban Hydrology EVT Generalized Pareto Distribution Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction Floods are among the most devastating natural disasters worldwide, causing significant human, economic, and environmental losses. Over the past decades, the frequency and severity of flood events have been amplified by the joint effects of climate variability, land-use change, and rapid urbanization (Legg 2021 ; Forster, Storelvmo, and Alterskjæ 2021). Extreme precipitation events, in particular, play a central role in triggering flash floods and riverine floods, especially in tropical and coastal regions where high-intensity rainfall events are recurrent (Alfieri et al. 2018; Jia et al. 2019 ). According to the Centre for Research on the Epidemiology of Disasters (CRED & UNDRR 2022), floods account for more than 40% of all natural disasters globally, with Asia and Africa being the most affected continents. The social consequences include the displacement of populations, the spread of waterborne diseases, and damages to infrastructure, while the economic costs often represent several percentage points of national GDP for highly vulnerable countries (UNDRR 2020 ). Despite significant progress in flood risk modeling, predicting the timing and magnitude of hydrological extremes remains a challenge, particularly in regions where observational data are scarce or incomplete (Blöschl et al. 2019 ). In Africa, floods are increasingly becoming one of the most pressing environmental challenges. Several studies report that the frequency of flood-related disasters has more than doubled in sub-Saharan Africa since the 1980 (Di Baldassarre et al. 2010; Tschakert et al. 2010). Major cities such as Lagos, Accra, Dakar, and Douala are particularly vulnerable due to their rapid demographic growth, poorly regulated urban expansion, and inadequate drainage infrastructure (A. C. Douglas et al. 2008 ; Adelekan 2010 ). In West and Central Africa, climate projections indicate that the intensity of extreme rainfall events is likely to increase under warming scenarios, thereby exacerbating flood hazards (Sylla et al. 2016 ; Dosio et al. 2021 ). However, most of the existing research on African floods has primarily focused on either precipitation extremes or hydrological modeling of river basins, with limited efforts to explicitly connect the two processes through joint statistical approaches. This constitutes an important scientific gap, as understanding the rainfall–runoff relationship at extreme scales is fundamental to designing effective flood management and early-warning systems in the region (Nka et al. 2015 ). The city of Douala, Cameroon’s economic capital and largest port, epitomizes this situation. Located in a low-lying coastal zone and traversed by the Wouri River, Douala has experienced recurrent and severe flooding over the past decades (Kometa and Ebot 2012; Tsalefac et al. 2015 ). The combined effect of extreme rainfall events, inadequate drainage, and unplanned settlement on floodplains has resulted in persistent human and material losses. For instance, the floods of August 2020 and July 2022 led to the displacement of thousands of households, the destruction of roads and houses, and increased vulnerability to cholera outbreaks (Akwa and Nguimbous 2021). Although several studies have analyzed rainfall variability and extremes in Douala (Adzandeh, Alaigba, and Nkemasong 2020 ; Nonki et al. 2021 ), very few have attempted to statistically link precipitation extremes to river discharge dynamics in the Wouri basin. Moreover, the majority of previous research has concentrated on annual maxima or seasonal variability, without considering the dynamics of return periods and return levels in terms of both precipitation and river discharge. This methodological gap prevents a comprehensive understanding of how extreme rainfall translates—or fails to translate—into extreme flows in Douala’s hydrological system. Recent literature in hydrology and climate extremes provides useful insights but also shows critical limitations. For example, Papalexiou and Koutsoyiannis, (2016) emphasized the need for robust statistical frameworks for extreme rainfall, highlighting the limitations of traditional approaches such as block maxima that often waste valuable data. Similarly, Coles et al. ( 2001 ); Katz ( 2013 ) demonstrated the relevance of the Generalized Pareto Distribution (GPD) and Peak Over Threshold (POT) methods for more accurate estimation of return levels. In the African context, studies like Nka et al. ( 2015 ) on Cameroon and Ogden et al. ( 2011 ) on West Africa have applied extreme value theory to precipitation, but without systematically incorporating discharge data. Other works, such as Di Baldassarre et al. (2010) in the Sahel or Kundzewicz et al. ( 2014 ) globally, have underlined the growing mismatch between rainfall extremes and observed flood damages, yet they stop short of quantifying this gap in terms of return periods and lag times. This indicates that while precipitation extremes are well documented, the hydrological response—especially the lag between rainfall and discharge peaks—remains underexplored, particularly in tropical urban basins. The originality of the present study lies precisely in addressing these gaps. By combining daily precipitation data from Douala with daily discharge data from the Wouri River, we aim to calculate and compare the return levels and return periods of both precipitation and discharge extremes. More importantly, we will investigate the potential temporal lag between extreme rainfall events and peak discharges, highlighting situations where extreme precipitation does not immediately translate into extreme flows. This lag effect, if confirmed, would provide critical insights into the dynamics of the Wouri River and the broader challenges of flood risk in Douala. To our knowledge, no previous study in Douala—or even more broadly in tropical Africa—has systematically quantified this rainfall–discharge lag in terms of extreme value theory. Internationally, although some studies have explored rainfall–runoff correlations under extreme conditions (e.g., (Blöschl et al. 2017 ; Berghuijs et al. 2019 ), they have largely focused on temperate regions with dense hydrometric networks, leaving tropical urban basins underrepresented. Our work therefore contributes an innovative methodological and geographical perspective, reinforcing the scientific and practical relevance of this research. The consequences of such an approach are twofold. First, by quantifying return levels and return periods for both precipitation and discharge extremes, we provide a more nuanced understanding of hydrological risks in Douala. This knowledge is vital for urban planning, flood defense design, and early warning systems. Second, by identifying potential lags or mismatches between rainfall and river response, our research sheds light on the complex dynamics of the Wouri basin, where human factors such as land use, drainage infrastructure, and river channel modifications may alter the natural rainfall–runoff relationship. Such findings could be transferable to other rapidly urbanizing tropical basins worldwide, where similar dynamics are at play but remain poorly documented in the scientific literature. This dissertation is organized as follows. Chapter One presents the global and African context of floods and extreme precipitation, with a focus on Douala’s vulnerability. Chapter Two introduces the data and methodology, particularly the use of extreme value theory (EVT) approaches such as the Generalized Pareto Distribution (GPD) and the Peak Over Threshold (POT) method for estimating return periods and levels. Chapter Three discusses the empirical results, comparing precipitation and discharge extremes, quantifying their return levels, and analyzing the temporal lag between them. The final section synthesizes the findings, highlights the originality and limitations of the study, and provides recommendations for flood risk management and future research in Douala and similar urban contexts. 2. Zone, Data and methods study 2.1. Study area. Douala, the economic capital of Cameroon, is located in the Littoral Region on the Atlantic coast of Central Africa. The city lies between latitudes 3°40′–4°10′ N and longitudes 9°40′–10°00′ E, covering an estimated surface area of about 210 km². It is the most populated city in Cameroon, with over 3.5 million inhabitants according to recent estimates, and it serves as the country’s principal seaport and commercial hub. Its geographical position at the mouth of the Wouri River and proximity to the Gulf of Guinea makes it highly strategic for trade, but also particularly exposed to hydrometeorological hazards (Fig. 1 ). The climate of Douala is classified as equatorial humid, characterized by very high annual rainfall exceeding 3,500 mm on average, with some years recording more than 4,000 mm. Precipitation is strongly seasonal, with a long rainy season extending from March to November and a short dry season between December and February. The peak of rainfall generally occurs between July and September, during which torrential rains often cause severe flooding. The city is also marked by persistently high humidity and temperatures ranging between 23°C and 31°C throughout the year, conditions that favor intense convective activity. Hydrologically, Douala is dominated by the Wouri River basin, which drains a large catchment before discharging into the Atlantic Ocean. Numerous small rivers, streams, and drainage channels cross the city, many of which are directly connected to the Wouri estuary. These watercourses, combined with the city’s flat and low-lying topography, increase its vulnerability to both riverine and pluvial flooding. In addition, the presence of extensive wetlands and mangrove ecosystems around the estuary plays a dual role: they act as natural buffers against flooding, but are increasingly threatened by rapid urbanization and land reclamation. From an urban perspective, Douala has experienced rapid and largely uncontrolled population growth and spatial expansion over the past decades. Informal settlements are widespread, often located in flood-prone zones such as riverbanks, marshes, and poorly drained lowlands. The lack of adequate drainage infrastructure, coupled with unplanned urban sprawl, significantly amplifies the impacts of extreme rainfall events. Consequently, floods in Douala regularly disrupt transportation, damage property, and pose severe public health risks, particularly through the spread of waterborne diseases such as cholera and typhoid fever. The geographical and environmental setting of Douala therefore makes it a critical hotspot for hydroclimatic risk studies in West and Central Africa. Its combination of high rainfall exposure, sensitive river systems, rapid urbanization, and limited adaptive infrastructure creates a unique context where the study of extreme precipitation and river discharge dynamics is particularly relevant. Understanding this geographical situation is essential for designing robust flood risk management and early warning strategies that can reduce vulnerability in this rapidly growing coastal megacity. 2.2. Data study. The Wouri basin, encompassing the densely populated and economically critical city of Douala, Cameroon, is increasingly vulnerable to flooding driven by extreme precipitation events. Understanding the dynamic relationship between these rainfall extremes and the subsequent hydrological response of the river system is paramount for effective water resource management and flood mitigation strategies. This study focuses on the Wouri basin from 2010 to 2023, aiming to quantify the link between extreme daily precipitation and the immediate runoff response, thereby elucidating the basin's behavior under severe meteorological conditions. To analyze precipitation patterns, this study utilized daily data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) dataset. CHIRPS incorporates satellite imagery with in-situ station data to create gridded precipitation time series, particularly suited for trend analysis and drought and flood monitoring in data-sparse regions Funk et al. ( 2015 ). The daily precipitation values were extracted for a point location representing the city of Douala (9.7°E, 4.05°N), providing a high-resolution, long-term record of rainfall inputs to the basin. The hydrological response of the basin was assessed using daily runoff data from the ERA5-Land reanalysis dataset, produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). ERA5-Land provides a globally consistent and comprehensive replay of land variables from 1950 to the present, offering a reliable representation of key hydrological fluxes like runoff and total precipitation at an enhanced resolution of ~ 9 km (Hersbach et al. 2019 ; Muñoz-Sabater et al. 2021 ; Hersbach et al. 2020 ). Daily runoff_sum and total_precipitation_sum data were averaged over the entire Wouri basin polygon to characterize the aggregate hydrological behavior of the catchment in response to meteorological forcings. The methodology involved extracting and collating daily time series for both precipitation and runoff. Extreme precipitation events were defined as days with rainfall exceeding 50 mm. A key metric, the runoff ratio (daily runoff / daily precipitation), was calculated to assess the efficiency of runoff generation. The analysis compared general statistics with conditions during extreme events to identify shifts in hydrological behavior. Furthermore, the temporal lag between precipitation peaks and runoff peaks was examined to infer the basin's response time. This integrated use of CHIRPS precipitation data (Climate Hazards Group InfraRed Precipitation with Station data; (Funk et al. 2015 ) and ERA5-Land reanalysis runoff data (Muñoz-Sabater et al. 2021 ) provides a robust framework for analyzing precipitation–runoff dynamics in a basin where traditional gauge data may be limited. 2.3. Method. The analysis of extremes was conducted using the Peaks Over Threshold (POT) framework of Extreme Value Theory (EVT) (Coles et al. 2001 ; R. W. Katz, Parlange, and Naveau 2002). Prior to applying the POT approach, the rainfall and discharge series were subjected to rigorous quality control. Outliers were identified and removed using the interquartile range (IQR) and z-scores, short missing gaps were filled by interpolation, and long gaps were excluded. Stationarity was tested with the Augmented Dickey–Fuller (ADF) test, while monotonic trends were assessed using the Mann–Kendall test (Kendall 1948 ; Helsel and Hirsch 1993). These pre-processing steps ensured the robustness of the datasets for extreme value analysis. Thresholds for extreme event extraction were determined using mean residual life (MRL) plots and threshold stability plots (Davison and Smith 1990). Exceedances above the selected thresholds were fitted to the Generalized Pareto Distribution (GPD), with shape (ξ) and scale (σ) parameters estimated via Maximum Likelihood Estimation (MLE). The adequacy of the fitted models was evaluated through diagnostic checks, including Quantile–Quantile (Q–Q), Probability–Probability (P–P), and Cumulative Distribution Function (CDF) plots, along with the Kolmogorov–Smirnov goodness-of-fit test (Choulakian and Stephens 2001). Return levels were then derived for recurrence intervals of 2, 5, 10, 20, and 50 years, with uncertainty quantified using parametric bootstrapping (Effron and Tibshirani 1993). In addition to conventional EVT analysis, the study employed the Padji Calvin Method (Padji et al. 2024 ), which innovatively introduces the concept of return dates. This method extends EVT by converting statistical return periods into calendar-based dates, achieved by mapping threshold exceedances to their empirical occurrence within the annual cycle and adjusting for leap years. Finally, rainfall–runoff coupling was investigated by analyzing discharge responses within a 0–5-day lag window after rainfall extremes, using cross-correlation functions (CCF) (Box and Jenkins 1976 ). This approach allowed the identification of potential temporal mismatches between meteorological and hydrological extremes, providing novel insights into flood dynamics in Douala. 3. Results and discussion 3.1. Rainfall–Discharge Variability and Hydrological Dynamics in Douala. Figure 2 presents a multi-panel analysis of rainfall and discharge in Douala over the period 2010–2023. Panel (a) illustrates the distributional characteristics of both variables through boxplots. Rainfall exhibits a positively skewed distribution with numerous outliers, reflecting the predominance of low-intensity events interspersed with episodic extremes, a pattern typical of tropical rainfall regimes (Koutsoyiannis 2004; R. W. Katz, Parlange, and Naveau 2002). In contrast, discharge displays a more symmetrical distribution with fewer extreme values, highlighting the dampening influence of watershed processes on the rainfall–runoff response (Beven 2012; Blöschl and Sivapalan 1995). The lower central tendency of discharge relative to rainfall underscores the role of hydrological losses such as infiltration, evapotranspiration, and storage (Dingman 2015). Panel (b) shows the temporal dynamics of rainfall and discharge. Both series reveal marked seasonality and inter-annual variability, with synchronized peaks corresponding to major hydrological events. However, discharge responses are attenuated relative to rainfall, particularly during extreme precipitation episodes, confirming the buffering capacity of the watershed through storage and transmission losses (Gupta and Waymire 1993; Sivapalan 2003). Intensified hydrological activity is visible during 2015–2017, likely linked to broader climatic anomalies such as ENSO-related variability (Nicholson 2013; Lyon and Vigaud 2017). Panels (c) and (d) present the frequency distributions of rainfall and discharge. The rainfall histogram confirms the predominance of low-intensity events, with an exponential decline in frequency as intensity increases (Coles et al. 2001; R. W. Katz, Parlange, and Naveau 2002). Conversely, the discharge histogram approximates a normal distribution, reflecting the watershed’s ability to regulate highly variable rainfall inputs into more stable streamflow outputs (Beven 2012). These frequency characteristics provide valuable measures of central tendency and variability for hydrological risk assessments (Legg 2021). Taken together, the comparative analysis demonstrates that rainfall acts as the primary driver of streamflow variability, but watershed characteristics substantially modify this relationship (Blöschl and Sivapalan 1995; Sivapalan 2003). The attenuation of discharge extremes and the temporal lag between rainfall peaks and discharge responses point to the critical role of antecedent soil moisture, vegetation cover, and geomorphological features in shaping hydrological behavior (Dingman 2015; Gupta and Waymire 1993). These findings provide valuable insights for flood forecasting, water resource management, and climate adaptation strategies (Legg 2021), while underscoring the persistence of rainfall–discharge dynamics revealed by daily observations over a 13-year period. 3.2. Extreme Value Modeling of Rainfall and Discharge in Douala: A Peak-Over-Threshold Approach for Flood Risk Assessment. The application of the Peak-Over-Threshold (POT) method hinges on a critical compromise: selecting a threshold that is sufficiently high to satisfy the asymptotic basis of the Generalized Pareto Distribution (GPD), yet low enough to retain an adequate number of excesses for precise parameter estimation. This selection is paramount, as an ill-chosen threshold can lead to biased estimates and invalid inferences regarding the tail behavior of the data. Therefore, prior to model fitting, the following section is devoted to the objective and diagnostic-driven selection of optimal thresholds for the rainfall and discharge series in Douala. 3.2.1. Modeling Extreme Hydrological Events in Douala: A Generalized Pareto Distribution (GPD) Approach. The validity of the Generalized Pareto Distribution (GPD) model is contingent upon the selection of an optimal threshold. An ill-chosen threshold violates the model's asymptotic assumptions, potentially leading to significant bias in the estimation of extreme quantiles and return levels. 3.2.1.1. Threshold Selection and Fitting of the Generalized Pareto Distribution (GPD) Model. The multi-panel figure 3 presents a comprehensive threshold selection analysis for both rainfall and discharge data using Mean Residual Life (MRL) plots and threshold choice plots, essential for implementing the Peak-Over-Threshold (POT) approach in extreme value analysis. Panels a), b), and c) focus on rainfall threshold determination. Panel a) (MRL plot) demonstrates the mean excess function for rainfall data, which exhibits initial instability followed by a linear trend above approximately 18 mm. The optimal threshold of 20 mm (green line) is identified where the mean excess becomes approximately linear, satisfying the key assumption of generalized Pareto distribution (GPD) applicability. The threshold range of 18-23 mm (red lines) indicates the region where the GPD model provides stable parameter estimates. The linearity above this threshold confirms that excess rainfall values follow the GPD, validating the threshold choice for extreme rainfall analysis. Panel b) shows the stability of the shape parameter (ξ) across different threshold values. The relative stability of ξ above 20 mm indicates that the distribution of rainfall excesses maintains consistent tail behavior, confirming the threshold appropriateness. The convergence of parameter estimates above this threshold suggests reliable extrapolation for return level estimation. Panel c) displays the scale parameter (σ) behavior, which should increase linearly with threshold if the GPD assumption holds. The linear trend observed above the selected threshold supports the validity of the chosen threshold for extreme rainfall modeling. Panels d), e), and f) present the corresponding analysis for discharge data. Panel d) (MRL plot) reveals an optimal discharge threshold of 16 mm/day, with a practical range of 13-20 mm/day. The linear mean excess above this threshold indicates that extreme discharge values follow the GPD, enabling reliable flood frequency analysis. Panel e) demonstrates the stability of the shape parameter for discharge above 16 mm/day, suggesting consistent tail behavior for extreme flow events. The stable ξ parameter above the threshold confirms that the discharge extremes belong to the same distribution family, crucial for accurate flood risk assessment. Panel f) shows the scale parameter behavior for discharge, with the expected linear relationship above the chosen threshold, further validating the threshold selection for extreme discharge analysis. The selected thresholds (20 mm for rainfall and 16 mm/day for discharge) represent the levels above which observations can be considered extreme events and appropriately modeled using the GPD framework. The stability of parameter estimates above these thresholds ensures reliable estimation of return levels and quantiles for rare hydrological events. These results provide statistically robust thresholds for extreme value analysis, enabling accurate estimation of design rainfall intensities and flood magnitudes for infrastructure planning and risk management. The concordance between different diagnostic methods (MRL plots and parameter stability plots) strengthens the confidence in the selected thresholds for subsequent extreme value modeling in hydrological applications. 3.2.1.2. Goodness-of-Fit Validation for the Generalized Pareto Distribution. The diagnostic plots for the fitted Generalized Pareto Distribution (GPD) models provide a comprehensive visual assessment of the model's adequacy in describing the tail behavior of extreme rainfall and discharge events in Douala. Figure 4. Diagnostic plots for the goodness-of-fit of the Generalized Pareto Distribution (GPD) model to extreme rainfall data (threshold: uR = 20 mm ) in Douala: (a) Probability-Probability (PP) plot, (b) Quantile-Quantile (QQ) plot, (c) Density plot, and (d) Cumulative Distribution Function (CDF) plot. For Extreme Rainfall ( uR = 20 mm ), shown in Figure 4: Probability-Probability (PP) Plot [4a]: The close alignment of the points along the 1:1 perfect fit line indicates an excellent agreement between the empirical probabilities of the observed excess rainfall and the probabilities predicted by the fitted GPD model. This suggests that the model accurately captures the overall distribution of the data. Quantile-Quantile (QQ) Plot [4b]: The strong linearity of the points, particularly in the upper tail, is a crucial result. It demonstrates that the model successfully replicates the magnitude and behavior of the most extreme rainfall events, which is the primary objective of Extreme Value Analysis. Density Plot [4c]: The close fit of the red GPD density curve to the histogram of observed excesses confirms that the model's shape (ξ) and scale (σ) parameters accurately describe the frequency of extreme rainfall intensities. Cumulative Distribution Function (CDF) Plot [4d]: The near-perfect overlap between the empirical CDF and the fitted model CDF provides strong evidence that the GPD is an appropriate model for the entire range of excess rainfall values. Conclusion for Rainfall: Collectively, these four diagnostic tools offer robust validation that the GPD model provides a statistically sound fit to the extreme rainfall data. Figure 5. Diagnostic plots for the goodness-of-fit of the Generalized Pareto Distribution (GPD) model to extreme discharge data (threshold: uD = 16 mm/day ) in Douala: (a) Probability-Probability (PP) plot, (b) Quantile-Quantile (QQ) plot, (c) Density plot, and (d) Cumulative Distribution Function (CDF) plot. For Extreme Discharge ( uD = 16 mm ), shown in Figure 5: PP and QQ Plots [5a and 5b]: The points in both plots adhere closely to the 1:1 line, indicating a good overall fit. The QQ plot shows that the model performs well across most quantiles, which is essential for predicting extreme flood discharges. Density Plot [5c]: The fitted density function follows the histogram of discharge excesses well. The slight underestimation in the very low-density region of the right tail suggests the model may slightly underestimate the frequency of the very largest events, but the overall fit is strong. CDF Plot [5d]: The strong agreement between the empirical and modeled CDFs confirms that the GPD accurately represents the probability structure of extreme discharge values. Conclusion for Discharge: The diagnostic plots in Figure 5 validate the application of the GPD model to extreme discharge events. The model provides a reliable fit, justifying its use for flood frequency analysis. Overall Synthesis: The successful application of the Peak-Over-Threshold method, confirmed by these diagnostic checks (Figures 4 and 5), demonstrates that the extreme values of both rainfall and discharge in Douala are well-characterized by the Generalized Pareto Distribution. This establishes a solid statistical foundation for estimating the probability and magnitude of future extreme events. The adequacy of the Generalized Pareto Distribution (GPD) models for both rainfall and discharge extremes was rigorously assessed using Kolmogorov-Smirnov (K-S) and Anderson-Darling (A-D) goodness-of-fit tests. The results presented in Table 1 demonstrate strong statistical evidence supporting the validity of the fitted models. For extreme rainfall events exceeding the 20 mm threshold, both statistical tests yielded non-significant results (K-S: D = 0.044, p = 0.711; A-D: A = 0.482, p = 0.765), indicating no evidence to reject the null hypothesis that the observed excess rainfall data follows the fitted GPD. Similarly, for extreme discharge values above the 16 mm/day threshold, both tests confirmed the model's adequacy (K-S: D = 0.077, p = 0.744; A-D: A = 0.535, p = 0.711). The consistency between both tests is particularly noteworthy given their different sensitivities to various aspects of distributional fit. The Kolmogorov-Smirnov test, more sensitive to deviations in the center of the distribution, and the Anderson-Darling test, more powerful for detecting discrepancies in the tails, both converge on the same conclusion. This concordance strengthens the validity of our threshold selection and parameter estimation. The presence of tied values in the dataset, common in environmental measurements due to instrumental precision, was addressed through appropriate statistical adjustments. The robustness of the Anderson-Darling test to such data characteristics provides additional confidence in our findings. These results collectively demonstrate that the GPD provides an excellent fit to the extreme values of both rainfall and discharge in Douala. The models adequately capture the tail behavior of the hydrological extremes, thereby validating their use for subsequent extreme value analysis, including the estimation of return levels and the quantification of probabilities for rare, high-magnitude events crucial for urban flood risk assessment and infrastructure planning. Table 1: Results of the Kolmogorov-Smirnov and Anderson-Darling goodness-of-fit tests for the fitted GPD models. Hydrological Variable Threshold (mm/day) Test Type Test Statistic P-Value Interpretation Rainfall 20 K-S A-D D = 0.0437 A = 0.4824 0.711 0.765 Adequate fit Adequate fit Discharge 16 K-S A-D D = 0.0775 A = 0.5351 0.744 0.711 Adequate fit Adequate fit 3.2.1.3 GPD parameter estimation results. The Generalized Pareto Distribution (GPD) was applied to exceedances above carefully selected thresholds for both rainfall and discharge series using the Maximum Likelihood Estimation (MLE) method. Threshold selection was guided by the Mean Residual Life (MRL) plot and quantile analysis, ensuring that a sufficient number of exceedances were retained for reliable parameter estimation while preserving the asymptotic properties of the GPD. As presented in Table 2, the scale parameter (σ) characterizes the dispersion of the exceedances, while the shape parameter (ξ) governs the behavior of the distribution tail. For rainfall extremes, the estimated shape parameter is positive (ξ = 0.19), indicating a heavy-tailed distribution of the Fréchet type. This implies that extreme rainfall events can potentially reach very large magnitudes. The 95% confidence interval for ξ (0.105–0.289) includes zero, which means that, although the point estimate suggests heavy tails, an exponential tail (Gumbel type) cannot be statistically rejected at the 5% significance level. For discharge extremes, the shape parameter is also positive (ξ = 0.44), suggesting a heavy-tailed distribution. This result indicates that even after catchment storage and attenuation processes, discharge extremes can attain very high values. However, the wider confidence interval (0.180–0.708) reflects greater uncertainty in parameter estimation relative to rainfall, likely arising from the nonlinear rainfall–runoff transformation and the hydrological complexity of the watershed. The two datasets remain directly comparable: rainfall and discharge analyses were based on the same observational period (13.99 years). Nevertheless, the number of exceedances differed (757 for rainfall and 173 for discharge), reflecting the relative rarity of hydrological extremes after the catchment transformation. Model diagnostics (deviance and AIC) were higher for rainfall than discharge, pointing to greater variability in rainfall extremes compared to discharge, consistent with the moderating effect of watershed processes on the hydrological response. Overall, these GPD parameter estimates provide the statistical foundation for the calculation of return levels and probabilities of rare hydrological extremes. Importantly, the differences observed between rainfall and discharge tail behaviors underscore the necessity of explicitly accounting for catchment processes in extreme value analysis, rather than assuming a direct one-to-one transfer of rainfall extremes into hydrological extremes. Table 2: GPD parameter estimates for rainfall and discharge extremes in the city of Douala from 2010 to 2023. In brackets are the 95% confidence intervals associated with these parameters. Parameter Rainfall Extremes Discharge Extremes Scale (σ) Shape (ξ) Number of Exceedances Data Span Deviance AIC 21.95 (19.41, 24.48) 0.19 (0.105, 0.289) 757 observations (1.48% of data) 13.99 years 6489.16 6493.16 19.26 (18.22, 20.3) 0.24 (0.2, 0.283) 173 observations (1.18% of data) 13.99 years 7993.72 7997.72 3.3. Comparison of Return Levels (RLs) and Return Periods (RPs) of Extreme Rainfall and Discharge Figure 6 provides an integrated assessment of Douala’s watercourse dynamics through six analytical panels. Panels a–c highlights the distributional properties of extremes. The Generalized Pareto Distribution (GPD) analysis for rainfall (Panel a) yields a shape parameter of ξ = 0.1973, indicating moderately heavy-tailed behavior, consistent with extreme value theory applications in hydrology (Coles et al. 2001; R. W. Katz, Parlange, and Naveau 2002). In contrast, discharge extremes (Panel b) exhibit a substantially heavier tail (ξ = 0.4443), underscoring the amplification of flood risks within the watershed (Koutsoyiannis 2004). The comparative analysis (Panel c) reveals a systematic temporal displacement between rainfall and discharge return periods: equivalent magnitude events occur more frequently in discharge than in rainfall, particularly at longer return periods. This suggests that urban watershed processes transform rare rainfall extremes into relatively frequent flood occurrences (Smith and Ward 1998; Villarini and Smith 2010). Panels d–f expands on these patterns by examining return-period dynamics. Rainfall returns levels (Panel d) display relatively stable growth, whereas discharge return levels (Panel e) show steeper gradients and higher variability, particularly beyond the 100-year event (Beirlant and Goegebeur 2004). The integrated view (Panel f) demonstrates that discharge confidence intervals are substantially wider than those for rainfall, reflecting greater uncertainty in flood prediction (Coles et al. 2001). This uncertainty stems from urban features such as impervious surfaces, drainage inefficiencies, and modified flow pathways (Arnell and Gosling 2016; I. Douglas et al. 2008). Taken together, these results highlight several key challenges: Hydrological amplification: Urban runoff processes compress rainfall returns periods into shorter discharge return periods, intensifying flood hazards (Miller and Hutchins 2017). Infrastructure limitations: The widening gap between rainfall and discharge uncertainty reflects systemic inefficiencies in drainage capacity (I. Douglas et al. 2008). Temporal compression of risks: Discharge extremes recur more frequently than their meteorological drivers, signaling reduced watershed storage and faster hydrological response (Seneviratne et al. 2012). Spatial heterogeneity: Variability in confidence intervals indicates uneven flood vulnerability across different areas of the city (Habitat 2018). Climate sensitivity: The heavy-tailed discharge distribution suggests disproportionate risks under projected increases in extreme rainfall (IPCC 2021). These findings emphasize the urgent need for integrated interventions, including drainage rehabilitation, green infrastructure, improved land-use planning, and enhanced early warning systems (Di Baldassarre et al. 2010; IPCC 2021). Figure 7 provides complementary evidence of systemic hydrological dysfunction through rainfall–discharge relationship analysis. Panel a show a weak linear correlation (r ≈ 0.2–0.3), indicating that daily rainfall explains only 4–9% of discharge variability. The wide scatter of data points highlights the decoupling between rainfall inputs and river response, while the shallow regression slope underscores limited predictive capacity (Gupta and Waymire 1990). Panel b examines temporal dynamics using cross-correlation. The maximum lagged correlation occurs after 1–2 days, suggesting a delayed discharge response. However, the peak correlation remains modest (r ≈ 0.3–0.4), reinforcing the weak coupling observed in Panel a. The flattened correlation profile indicates heterogeneous and prolonged hydrological responses across sub-catchments, shaped by urban modifications (Villarini and Smith 2010). Together, Panels a and b reveal: Systemic hydrological disconnection: Weak correlations confirm that urbanization has disrupted natural rainfall–runoff connectivity (I. Douglas et al. 2008). Urbanization effects: Impervious surfaces, drainage blockages, and fragmented flow pathways dominate discharge behavior, reducing rainfall predictability (Olivier and Prins 2009). Flood risk implications: The combination of weak correlations and limited lead time complicates flood forecasting and increases uncertainty (Di Baldassarre et al. 2010). Infrastructure deficiencies: Inadequate monitoring, poor drainage maintenance, and lack of integration between urban growth and water management exacerbate risks (Habitat 2018). Climate vulnerability: Non-linear and disorganized responses reduce resilience and magnify risks under intensified rainfall (IPCC 2021). Policy responses should prioritize distributed monitoring networks, systematic drainage rehabilitation, the expansion of green infrastructure, and better integration of hydrological considerations into urban planning (Seneviratne et al. 2012; A. C. Douglas et al. 2008). Figures 6 and 7 collectively demonstrate that Douala’s flood risks are not solely a function of extreme rainfall but are significantly shaped by watershed transformation and urbanization. The amplification of discharge extremes, the weak rainfall–discharge coupling, and the spatial–temporal compression of risks provide strong evidence for urgent reforms in both infrastructure and governance to reduce the city’s flood vulnerability (Di Baldassarre et al. 2010; IPCC 2022). 3.4. Estimation of Return Periods, Dates, and Levels for Extreme Precipitation and Discharge in Douala City. The quantitative analysis presented in Table 3 reveals significant hydrological disparities between the extreme regimes of precipitation and discharge within the urban watershed of Douala. This systematic comparison of return periods (RPs), return levels (RLs), and their 95% confidence intervals provides a mathematical demonstration of the profound disturbance of natural streamflow dynamics. Table 3 highlights a disproportionate hydrological attenuation, whereby for a 100-day return period, extreme precipitation intensity (98.08 mm) generates only an extreme discharge of 37.93 mm/day, representing a reduction of more than 60%. This nonlinear decline between precipitation intensity and hydrological response, as quantified in Table 3, corroborates the findings of Ngoran and Xue (2015) regarding the accelerated degradation of hydrological functions in urbanized catchments of Douala. Table 3 also shows a widening divergence in the widths of confidence intervals between precipitation and discharge. For the 1095-day return period by example, the confidence interval for discharge [88.24 – 186.51] spans a relative amplitude of 111%, compared to only 36% for precipitation [173.68 – 251.08]. This growing differential uncertainty, systematically documented in Table 3, reflects the complexity introduced by anthropogenic factors in hydrological response, confirming the conclusions of Uddin and Salah (2018) on the spatial heterogeneity of urban runoff processes in Douala. A closer examination of Table 3 demonstrates that the confidence intervals of discharge return periods are systematically wider and more asymmetric than those of precipitation. For instance, at a return level of 70.93 mm/day (discharge), the return period ranges from 267 to 332 days, while at a return level of 143.91 mm (precipitation), the return period ranges from 239 to 359 days. This structural asymmetry in uncertainties, as quantified in Table 3, illustrates the fundamental disruption of hydrological transfer processes, consistent with the findings of Adamu et al. (2025) on chronic obstructions within the Douala hydrographic network. The systematic comparison provided in Table 3 between extreme precipitation and discharge regimes offers a critical quantitative foundation for recalibrating flood management strategies. The significant discrepancies between confidence intervals highlight the need, as emphasized by Raouf et al. (2025), to develop differentiated approaches in the design of hydraulic infrastructure that account for the increased variability introduced by anthropogenic modifications of the watershed. Table 3 thus serves as quantitative evidence of the advanced degradation of streamflow dynamics in Douala, providing essential data to guide hydrological restoration policies and climate adaptation strategies in this rapidly expanding coastal metropolis. Table 3: Comparative Return Periods and Return Levels of Extreme Precipitation and Extreme Discharge in the Douala Watershed, with 95% Confidence Intervals RPs (day) RLs of the extreme Discharges ( mm/day) 95% CIs for RLs of the extreme Discharges (mm/day) 95% CIs for RPs (day) of the extreme Rainfall RLs of the extreme Rainfalls ( mm) 95% CIs for RLs of the extreme Rainfalls (mm) 95% CIs for RPs (day) of the extreme Rainfalls 50 24.03 [20.94, 27.11] [50,50] 73.87 [68.753,79.00] [46,53] 70 30.24 [26.06, 34.43] [69,71] 85.21 [78.81 91.60] [64,75] 100 37.93 [32.28,43.57] [97,103] 98.08 [89.90,106.25] [89,109] 140 46.38 [38.92, 53.85] [132,147] 111.07 [100.71,121.43] [121,157] 210 58.41 [47.76, 69.05] [193,227] 127.93 [114.19,141.67] [174,244] 280 68.35 [54.48, 82.21] [250,309] 140.73 [124.03,157.43] [225,334] 300 70.93 [56.14, 85.72] [267,332] 143.91 [126.42,161.40] [239,359] 365 78.71 [60.92, 96.50] [319,411] 153.19 [133.30,173.08] [285,444] 730.5 112.37 [78.27, 146.47] [592,868] 189.07 [158.45,219.68] [516,943] 1095 137.38 [88.24, 186.51] [844,1345] 212.38 [173.68,251.08] [726,1464] 1421 156.02 [94.34, 217.7] [1058,1784] 228.41 [183.7,273.12] [899,1943] 1825 2000 2500 3000 3500 4000 4500 176.08 183.99 204.69 223.18 240.04 255.60 270.11 [99.80, 252.36] [101.67,266.32] [105.83,303.54] [108.75,337.62] [110.80,369.27] [112.22,398.98] [113.14,427.08] [1311,2339] [1418,2583] [1712,3289] [1995,4007] [2268,4735] [2532,5471] [2789,6215] 244.59 250.72 266.12 279.21 290.66 300.86 310.08 [193.45,295.74] [197.05,304.39] [205.89,326.34] [213.18,345.24] [219.38,361.93] [224.79,376.92] [229.58,229.58] [1101,2550] [1183,2817] [1412,3590] [1626,4376] [1829,5175] [2023,5980] [2210,6791] The comparative analysis of discharge and rainfall extremes (Tables 4 and 5) provides critical insights into the hydrological functioning of the urban watershed of Douala. The results highlight not only the amplification of hydrological hazards but also the temporal and structural disconnections between precipitation inputs and river discharge responses. Discharge extremes display strong interannual variability, with the flood of May 28, 2013 (174 mm/day), representing the most exceptional event, characterized by a return period of 1781 days and a wide confidence interval [1284–2278 days]. This confirms the complexity of flood generation processes in the Wouri basin, as also noted by Besack et al. (2025). The breadth of these intervals indicates substantial predictive uncertainty for rare events, which complicates flood hazard assessment. A systematic lag was observed between extreme discharges and subsequent precipitation return dates, ranging from 2 to 37 months. For example, the flood of September 2, 2014 (145 mm/day), corresponds to a rainfall return date projected for January 6, 2018, reflecting a delay of 40 months. Such temporal displacement suggests the presence of hydrological memory within the catchment, a phenomenon consistent with the findings of (Boum-Nkot et al. 2024; Tume et al. 2025; Yonzoua et al. 2021). Rainfall extremes exhibit a clear upward trend between 2010 and 2023, with annual maxima increasing from 127 mm in 2010 to 248 mm in 2018—an intensification of 95% within less than a decade. This corroborates earlier observations by Molua and Lambi (2006); Grijsen (2014); Nonki et al. (2019), who reported an increase in the frequency and intensity of hydro-meteorological extremes across Cameroon under changing climatic conditions. The joint interpretation of Tables 4 and 5 reveals a marked asymmetry between rainfall and discharge extremes. Whereas rainfall extremes follow a relatively regular progression, discharge extremes display accentuated interannual fluctuations. This divergence reflects the growing influence of anthropogenic factors, such as soil sealing and altered drainage, on watershed hydrology (Cheng, Li, and Liu 2020; Anyangwa 2025). The efficiency of rainfall-to-discharge conversion is notably low. For instance, in 2015, a rainfall maximum of 227 mm produced only 57.7 mm of discharge, corresponding to a transfer efficiency of 25%. This attenuation can be explained by urban hydrological losses, including infiltration, evaporation, and depression storage, as described by Din et al. (2017). The wide confidence intervals associated with return dates delineate windows of heightened vulnerability. For example, the 2013 flood defines a critical vigilance window between December 2016 and August 2019, necessitating long-term preparedness. The variability of return periods further underscores the need for multi-scale early warning systems capable of managing both frequent (100–200 days) and exceptional (>1000 days) events, in line with the recommendations of MINEE (http://www.minee.cm). The findings demonstrate the necessity of adaptive hydrological models that account for land-use change, drainage infrastructure performance, and climatic variability. Moreover, the complementarity between rainfall and discharge extremes justifies the deployment of integrated monitoring systems that measure both variables simultaneously, thereby allowing continuous model calibration and improved risk forecasting. Together, Tables 4 and 5 provide quantitative evidence that Douala’s hydrological risks are not solely determined by extreme rainfall events but are amplified and reshaped by watershed transformations and rapid urbanization. The asymmetry in rainfall–discharge responses, the low transfer efficiency, and the growing uncertainties emphasize the urgent need for structural (drainage rehabilitation, green infrastructure) and non-structural (monitoring, early warning, planning) interventions to reduce flood vulnerability in the city. Table 4: Number of return days estimated for different maximum discharge (mm/day) over each year, as well as estimated return dates of maximum rainfall and 95% CIs using the adjustment of daily discharge totals from 2010-23 in Douala by the GPD model. Dates of extreme discharge in each year. Maximum discharge (mm/day) Estimated number of return days 95% CIs of number of return days 95% CIs of RLs (mm/day) Estimated dates of return of the next maximum rainfall. 95% CIs for RD. 15/07/2010 37.6 99 [96, 102] [32.1, 43.29] 22/10/2010 [15/07/2010, 25/10/2010] 31/08/2011 102 603 [500, 706] [73.45, 130.56] 25/04/2013 [ 12/01/2013 , 06/08/2013 ] 29/09/2012 44.7 131 [124, 138] [37.62, 51.78] 07/02/2013 [ 31/01/2013 , 14/02/2013 ] 28/05/2013 174 1781 [1284,2278] [99.29, 248.75] 13/04/2018 [ 02/12/2016 , 23/08/2019 ] 02/09/2014 145 1222 [930, 1516] [90.85,199.08] 06/01/2018 [ 20/03/2017 , 27/10/2018 ] 27/08/2015 57.7 205 [189,222] [47.26, 68.14] 19/03/2016 [ 03/03/2016 , 05/04/2016 ] 24/07/2016 94.5 519 [438, 601] [69.68, 119.30] 25/12/2017 [ 05/10/2017 , 17/03/2018 ] 16/06/2017 131 995 [777, 1212] [85.92, 176.17] 07/03/2020 [ 02/08/2019 , 10/10/2020 ] 26/07/2018 41.1 114 [109,119] [34.78,47.36] 17/11/2018 [ 12/11/2018 , 22/11/2018 ] 16/05/2019 122 861 [685, 1038] [82.37, 161.61] 23/09/2021 [ 31/03/2021 , 19/03/2022 ] 21/06/2021 13/08/2022 08/07/2023 118 80.6 40.8 806 382 113 [646, 965] [332,431] [108, 117] [80.73,155.35] [62.04, 99.19] [34.61, 47.1] 05/09/2023 30/08/2023 29/10/2023 [ 29/03/2023 , 11/02/2024 ][ 11/07/2023 , 18/10/2023 ][ 24/10/2023 , 02/11/2023 ] Table 5: Number of return days estimated for different maximum precipitation (mm) over each year, as well as estimated return dates of maximum rainfall and 95% CIs using the adjustment of daily rainfall totals from 2010-23 in Douala by the GPD model. Dates of extreme rainfall in each year. Maximum rainfall (mm) Estimated number of return days 95% CIs of number of return days 95% CIs of RLs (mm) Estimated dates of return of the next maximum rainfall. 95% CIs for RD. 23/08/2010 127 206 [171, 239] [113.54, 140.66] 17/03/2011 [ 10/02/2011 , 19/04/2011] 18/08/2011 170 511 [381, 641] [145.34, 194.61] 10/01/2013 [ 02/09/2012 , 20/05/2013 ] 07/07/2012 154 371 [288, 453] [133.88, 174.08] 13/07/2013 [ 21/04/2013 , 03/10/2013 ] 03/07/2013 146 290 [208, 362] [118.24, 146.30] 19/04/2014 [ 27/01/2014 , 30/06/2014 ] 25/08/2014 193 314 [249, 378] [128.01, 164.06] 05/07/2015 [ 01/05/2015 , 07/09/2015 ] 27/08/2015 227 1390 [883, 1897] [182.84, 271.20] 17/06/2019 [ 26/01/2018 , 05/11/2020 ] 21/08/2016 229 1435 [906, 1963] [184.07, 273.98] 26/07/2020 [ 13/02/2019 , 05/01/2022] 30/08/2017 215 1144 [752, 1535] [175.36, 254.68] 17/10/2020 [ 21/09/2019 , 12/11/2021 ] 26/07/2018 248 1921 [1146, 2697] [195.46,300.56] 29/10/2023 [ 14/09/2021 , 13/12/2025 ] 08/08/2019 198 856 [591, 1121] [164.37, 231.58] 11/12/2021 [ 21/03/2021 , 02/09/2022 ] 20/08/2020 11/08/2021 26/08/2022 08/07/2023 178 125 104 144 595 196 117 300 [434, 756] [164, 228] [103,131] [240, 361] [150.88, 205] [111.86,138.06] [94.90, 113.16] [126.42, 161.40] 07/04/2022 23/02/2022 21/12/2022 29/11/2023 [ 28/10/2021 , 15/09/2022 ] [22/01/2022, 27/03/2022] [07/12/2022, 04/01/2023] [04/03/2024, 03/07/2024] 4. Conclusion The objective of this study was to evaluate the hydrological dynamics of the Douala watershed by determining the Return Periods (RPs), Return Levels (RLs), and Return Dates (RDs) — together with their respective Confidence Intervals (CIs) — for extreme rainfall and discharge events. The analysis was based on daily rainfall and discharge data covering the period 2010–2023. The univariate and stationary Peak-Over-Threshold (POT) approach was applied using the Generalized Pareto Distribution (GPD), fitted through the Maximum Likelihood Estimation (MLE) method. Thresholds were selected through three complementary criteria: the Mean Residual Life (MRL) plot, the Model-Based Check (MBC), and the automatic threshold selection algorithm, ensuring the robustness of the GPD fitting process. This study also applied the method developed by Padji et al. ( 2024 ) for the simultaneous estimation of RPs, RLs, and RDs of extreme events, including the determination of their Confidence Intervals using the Delta Method. This innovative approach provides a dynamic framework capable of linking statistical extreme-value estimation with temporal prediction, thereby enhancing the practical utility of hydrological forecasting for flood early warning. The results confirm a marked interannual and seasonal variability in extreme events, with four main rainfall regimes identified: December–February (DJF), March–May (MAM), June–September (JJAS), and October–November (ON) (Padji et al. 2024 ). The JJAS period remains the most critical, characterized by the highest rainfall intensities and the strongest discharge peaks, leading to recurrent and destructive flooding (Padji et al. 2024 ). Historical flood events recorded in Douala are consistent with the RDs projected by the model, particularly during the major rainy season, validating both the GPD-based estimations and the methodological framework proposed by (Padji et al. 2024 ). The estimated GPD shape parameters indicate a heavy-tailed Fréchet behavior for rainfall, suggesting the potential occurrence of very intense precipitation events, while discharge extremes exhibit a relatively lighter tail, reflecting the influence of basin retention, infiltration, and drainage capacity. The comparison of rainfall and discharge RPs and RLs reveals a systematic asymmetry: rainfall extremes are transformed into disproportionately variable discharge responses. This nonlinearity highlights the hydrological amplification effect in an urbanized basin, consistent with earlier findings by Tchuikoua ( 2020 ) and Ndjama et al. ( 2014 ). Furthermore, the computation of Confidence Intervals for both RPs and RDs provides critical quantitative uncertainty bounds for risk analysis. For instance, the flood event of 28 May 2013, associated with a discharge of 174 mm/day, corresponds to a modeled return period of approximately 1781 days, with a 95% confidence interval ranging from 1284 to 2278 days. These results illustrate the temporal variability and uncertainty inherent in the prediction of extreme floods, offering key insights for adaptive flood management and infrastructure planning. The comparative interpretation of Tables 4 and 5 reinforces these findings. Table 4 shows pronounced interannual variability and uncertainty in discharge extremes, while Table 5 highlights an intensification of extreme rainfall magnitudes during the last decade. This asymmetry in recurrence behavior reveals that urbanization and land-use transformation have altered the rainfall–runoff coupling, reducing infiltration capacity and increasing hydrological response speed. Consequently, extreme rainfall events now produce more erratic and amplified discharge responses, underscoring the need for integrated watershed management. In summary, this study demonstrates that the integration of rainfall and discharge extremes within the POT–GPD framework, enhanced by the Padji et al. ( 2024 ) method, provides a coherent and operational approach for understanding and predicting hydrological risks in tropical urban basins. The methodology not only quantifies the magnitude and frequency of extreme events but also estimates their probable occurrence dates, offering valuable support for early warning systems and urban flood preparedness. This approach is transferable to other climatic parameters and catchments, provided high-resolution datasets are available. Future research should extend the analysis using sub-daily (hourly) data to refine the temporal prediction of extreme events and strengthen the city’s climate resilience framework. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declarations Author Contribution Padji Calvin – Conceptualization, Methodology, Data Analysis, Writing – Original Draft, Visualization.Calvin led the conception and design of the study, conducted the Extreme Value Theory analysis, interpreted the results, and prepared the manuscript draft.Mekaleuni Cyrille – Data Curation, Software, Validation, Writing – Review & Editing.Cyrille contributed to data acquisition, preprocessing of CHIRPS and ERA5-Land datasets, statistical validation, and assisted in revising and improving the manuscript.Monkam David – Supervision, Resources, Project Administration, Writing – Review & Editing.David supervised the research work, provided scientific guidance on hydrological interpretation, ensured methodological consistency, and contributed to the final review and editing of the manuscript. Clinical trial number: not applicable. References Adamu, Baba, Mohammadou Buba, Balgah Nguh, Mboza Yengeh, and Amos Zeh. 2025. “Urban Drainage Systems and Implications on the Inhabitants and Physical Environment of Buea Urban Area, South West Region of Cameroon.” Journal of Environmental and Geographical Studies 4 (April):61–88. https://doi.org/10.58425/jegs.v4i1.340. Adelekan, Ibidun O. 2010. “Vulnerability of Poor Urban Coastal Communities to Flooding in Lagos, Nigeria.” Environment and Urbanization 22 (2): 433–50. Adzandeh, E A, D Alaigba, and C N Nkemasong. 2020. “Application of Geospatial Techniques and Logistic Regression Model for Urban Growth Analysis in Limbe, Cameroon.” Niger. J. Environ. Sci. Technol.(NIJEST) 4:138–51. Akwa, Teh Exodus, and Simone Pierrette Nguimbous. 2021. “Investigation of Typhoid Fever and Their Associated Risk Factors in Children Attending ‘Deo Gratias’ Hospital in Douala, Littoral, Cameroon.” Electronic Journal of Medical and Educational Technologies 14 (2): em2107. Alfieri, Lorenzo, Sagy Cohen, John Galantowicz, Guy J P Schumann, Mark A Trigg, Ervin Zsoter, Christel Prudhomme, Andrew Kruczkiewicz, Erin Coughlan de Perez, and Zachary Flamig. 2018. “A Global Network for Operational Flood Risk Reduction.” Environmental Science & Policy 84:149–58. Anyangwa, Immanuel Ateh. 2025. “Effectiveness of Municipal Solid Waste Management Systems in Douala, Cameroon.” Arnell, Nigel W, and Simon N Gosling. 2016. “The Impacts of Climate Change on River Flood Risk at the Global Scale.” Climatic Change 134 (3): 387–401. Baldassarre, Giuliano Di, Alberto Montanari, Harry Lins, Demetris Koutsoyiannis, Luigia Brandimarte, and Günter Blöschl. 2010. “Flood Fatalities in Africa: From Diagnosis to Mitigation.” Geophysical Research Letters 37 (22). Beirlant, Jan, and Yuri Goegebeur. 2004. “Local Polynomial Maximum Likelihood Estimation for Pareto-Type Distributions.” Journal of Multivariate Analysis 89 (1): 97–118. Berghuijs, Wouter R, Shaun Harrigan, Peter Molnar, Louise J Slater, and James W Kirchner. 2019. “The Relative Importance of Different Flood‐generating Mechanisms across Europe.” Water Resources Research 55 (6): 4582–93. Besack, Felix, Gordon Ajonina, Seth Ebonji, Minette Tomedi, Landry Mbang Essome, Josué Tchogom, Raphaël Onguéné, and Olusegun Dada. 2025. “Change in Hydrological Signature of Tropical River-Estuaries During Extreme Flood Conditions: A Case Study of the Wouri-Nkam River Estuary (Cameroon) During the August 2020 Flood Event.” Estuaries and Coasts 48 (November):1–15. https://doi.org/10.1007/s12237-025-01599-w. Beven, Keith J. 2012. Rainfall-Runoff Modelling: The Primer . John Wiley & Sons. Blöschl, Günter, Julia Hall, Juraj Parajka, Rui A P Perdigão, Bruno Merz, Berit Arheimer, Giuseppe T Aronica, Ardian Bilibashi, Ognjen Bonacci, and Marco Borga. 2017. “Changing Climate Shifts Timing of European Floods.” Science 357 (6351): 588–90. Blöschl, Günter, Julia Hall, Alberto Viglione, Rui A P Perdigão, Juraj Parajka, Bruno Merz, David Lun, Berit Arheimer, Giuseppe T Aronica, and Ardian Bilibashi. 2019. “Changing Climate Both Increases and Decreases European River Floods.” Nature 573 (7772): 108–11. Blöschl, Günter, and Murugesu Sivapalan. 1995. “Scale Issues in Hydrological Modelling: A Review.” Hydrological Processes 9 (3‐4): 251–90. Boum-Nkot, Suzanne Ngo, Bertil Nlend, Frederic Huneau, Fabiola Kezebou, and Bernard Gwodog. 2024. “Groundwater Vulnerability Mapping of Hyper-Humid Urban Aquifers for an Improved Water Management in African Megacities, Opportunities and Limitations in the Case of Douala (Cameroon).” Arabian Journal of Geosciences 17 (3): 77. Box, George, and G M Jenkins. 1976. “Analysis: Forecasting and Control.” San Francisco . Cheng, Zhonghua, Lianshui Li, and Jun Liu. 2020. “The Impact of Foreign Direct Investment on Urban PM2. 5 Pollution in China.” Journal of Environmental Management 265:110532. Choulakian, Vartan, and Michael A Stephens. 2001. “Goodness-of-Fit Tests for the Generalized Pareto Distribution.” Technometrics 43 (4): 478–84. Coles, Stuart, Joanna Bawa, Lesley Trenner, and Pat Dorazio. 2001. An Introduction to Statistical Modeling of Extreme Values . Vol. 208. Springer. CRED & UNDRR. 2022. “2022 Disasters in Numbers. Centre for Research on the Epidemiology of Disasters & UN Office for Disaster Risk Reduction.” Davison, Anthony C, and Richard L Smith. 1990. “Models for Exceedances over High Thresholds.” Journal of the Royal Statistical Society Series B: Statistical Methodology 52 (3): 393–425. Din, Ndongo, Vanessa Maxemilie Ngo-Massou, Guillaume Léopold Essomè-Koum, Eugene Ndema-Nsombo, Ernest Kottè-Mapoko, and Laurant Nyamsi-Moussian. 2017. “Impact of Urbanization on the Evolution of Mangrove Ecosystems in the Wouri River Estuary (Douala Cameroon).” In Coastal Wetlands: Alteration and Remediation , 81–131. Springer. Dingman, S Lawrence. 2015. Physical Hydrology . Waveland press. Dosio, Alessandro, Martin W Jury, Mansour Almazroui, Moetasim Ashfaq, Ismaila Diallo, Francois A Engelbrecht, Nana A B Klutse, Christopher Lennard, Izidine Pinto, and Mouhamadou B Sylla. 2021. “Projected Future Daily Characteristics of African Precipitation Based on Global (CMIP5, CMIP6) and Regional (CORDEX, CORDEX-CORE) Climate Models.” Climate Dynamics 57 (11): 3135–58. Douglas, Alecia C, Juline E Mills, Mamadou Niang, Svetlana Stepchenkova, Sookeun Byun, Celestino Ruffini, Seul Ki Lee, Jihad Loutfi, Jung-Kook Lee, and Mikhail Atallah. 2008. “Internet Addiction: Meta-Synthesis of Qualitative Research for the Decade 1996–2006.” Computers in Human Behavior 24 (6): 3027–44. Douglas, Ian, Kurshid Alam, Maryanne Maghenda, Yasmin Mcdonnell, Louise McLean, and Jack Campbell. 2008. “Unjust Waters: Climate Change, Flooding and the Urban Poor in Africa.” Environment and Urbanization 20 (1): 187–205. Effron, Bradley, and Robert J Tibshirani. 1993. “An Introduction to the Bootstrap.” Chapman & Hall, New York. Forster, P, T Storelvmo, and K Alterskjæ. 2021. “IPCC Sixth Assessment Report (AR6) Working Group 1: The Physical Science Basis, Chap. 7.” University Press, UK, https://www. ipcc. ch/report/ar6/wg1/downloads (last …. Funk, Chris, Pete Peterson, Martin Landsfeld, Diego Pedreros, James Verdin, Shraddhanand Shukla, Gregory Husak, James Rowland, Laura Harrison, and Andrew Hoell. 2015. “The Climate Hazards Infrared Precipitation with Stations—a New Environmental Record for Monitoring Extremes.” Scientific Data 2 (1): 1–21. Grijsen, Johan. 2014. Understanding the Impact of Climate Change on Hydropower: The Case of Cameroon . World Bank, Washington, DC. Gupta, Vijay K, and Ed Waymire. 1990. “Multiscaling Properties of Spatial Rainfall and River Flow Distributions.” Journal of Geophysical Research: Atmospheres 95 (D3): 1999–2009. Gupta, Vijay K, and Edward C Waymire. 1993. “A Statistical Analysis of Mesoscale Rainfall as a Random Cascade.” Journal of Applied Meteorology and Climatology 32 (2): 251–67. Habitat, U N. 2018. “Tracking Progress Towards Inclusive, Safe, Resilient and Sustainable Cities and Human Settlements. SDG 11 Synthesis Report-High Level Political Forum 2018.” Helsel, Dennis R, and Robert M Hirsch. 1993. Statistical Methods in Water Resources . Vol. 49. Elsevier. Hersbach, Hans, Bill Bell, Paul Berrisford, Gionata Biavati, Dick Dee, András Horányi, Julien Nicolas, Carole Peubey, Raluca Radu, and Iryna Rozum. 2019. “The ERA5 Global Atmospheric Reanalysis at ECMWF as a Comprehensive Dataset for Climate Data Homogenization, Climate Variability, Trends and Extremes.” In Geophysical Research Abstracts . Vol. 21. Hersbach, Hans, Bill Bell, Paul Berrisford, Shoji Hirahara, András Horányi, Joaquín Muñoz‐Sabater, Julien Nicolas, Carole Peubey, Raluca Radu, and Dinand Schepers. 2020. “The ERA5 Global Reanalysis.” Quarterly Journal of the Royal Meteorological Society 146 (730): 1999–2049. IPCC. 2021. “Climate Change 2021: Impacts, Adaptation and Vulnerability. Contribution of Working Group II.” ———. 2022. “Sixth Assessment Report (AR6) – Impacts, Adaptation and Vulnerability.” https://www.ipcc.ch/report/ar6/wg2/. Jia, Weiping, Jianping Weng, Dalong Zhu, Linong Ji, Juming Lu, Zhiguang Zhou, Dajin Zou, Lixin Guo, Qiuhe Ji, and Li Chen. 2019. “Standards of Medical Care for Type 2 Diabetes in China 2019.” Diabetes/Metabolism Research and Reviews 35 (6): e3158. Katz, Michael B. 2013. The Undeserving Poor: America’s Enduring Confrontation with Poverty: Fully Updated and Revised . Oxford University Press. Katz, Richard W, Marc B Parlange, and Philippe Naveau. 2002. “Statistics of Extremes in Hydrology.” Advances in Water Resources 25 (8–12): 1287–1304. Kendall, Maurice George. 1948. “Rank Correlation Methods.” Kometa, Sunday Shende, and Mathias Ashu Tambe Ebot. 2012. “Watershed Degradation in the Bamendjin Area of the North West Region of Cameroon and Its Implication for Development.” Journal of Sustainable Development 5 (9): 75. Koutsoyiannis, Demetris. 2004. “Statistics of Extremes and Estimation of Extreme Rainfall: I. Theoretical Investigation/Statistiques de Valeurs Extrêmes et Estimation de Précipitations Extrêmes: I. Recherche Théorique.” Hydrological Sciences Journal 49 (4). Kundzewicz, Zbigniew W, Markus Stoffel, Ryszard J Kaczka, Bartłomiej Wyżga, Tadeusz Niedźwiedź, Iwona Pińskwar, Virginia Ruiz-Villanueva, Ewa Łupikasza, Barbara Czajka, and Juan Antonio Ballesteros-Canovas. 2014. “Floods at the Northern Foothills of the Tatra Mountains—a Polish-Swiss Research Project.” Acta Geophysica 62 (3): 620–41. Legg, Stephen. 2021. “IPCC, 2021: Climate Change 2021-the Physical Science Basis.” Interaction 49 (4): 44–45. Lyon, Bradfield, and Nicolas Vigaud. 2017. “Unraveling East Africa’s Climate Paradox.” Climate Extremes: Patterns and Mechanisms , 265–81. Miller, James D, and Michael Hutchins. 2017. “The Impacts of Urbanisation and Climate Change on Urban Flooding and Urban Water Quality: A Review of the Evidence Concerning the United Kingdom.” Journal of Hydrology: Regional Studies 12:345–62. Molua, Ernest L, and Cornelius M Lambi. 2006. “Climate, Hydrology and Water Resources in Cameroon.” CEEPA, Pretoria 37. Muñoz-Sabater, Joaquín, Emanuel Dutra, Anna Agustí-Panareda, Clément Albergel, Gabriele Arduini, Gianpaolo Balsamo, Souhail Boussetta, Margarita Choulga, Shaun Harrigan, and Hans Hersbach. 2021. “ERA5-Land: A State-of-the-Art Global Reanalysis Dataset for Land Applications.” Earth System Science Data 13 (9): 4349–83. Ndjama, J, A A Ako, D Kuitcha, T A Margaret, W Fantong, G Mafany, E G Takem, M C Y Fopouo, and K B V Kamgang. 2014. “Ecohydrology of Surface and Underground Water of the Mgoua Drainage Basin in Douala (Cameroon).” Ngoran, Suinyuy Derrick, and XiongZhi Xue. 2015. “Addressing Urban Sprawl in Douala, Cameroon: Lessons from Xiamen Integrated Coastal Management.” Journal of Urban Management 4 (1): 53–72. Nicholson, Sharon E. 2013. “The West African Sahel: A Review of Recent Studies on the Rainfall Regime and Its Interannual Variability.” International Scholarly Research Notices 2013 (1): 453521. Nka, Bernadette N, Ludovic Oudin, H Karambiri, Jean-Emmanuel Paturel, and P Ribstein. 2015. “Trends in Floods in West Africa: Analysis Based on 11 Catchments in the Region.” Hydrology and Earth System Sciences 19 (11): 4707–19. Nonki, Rodric M, André Lenouo, Christopher J Lennard, and Clément Tchawoua. 2019. “Assessing Climate Change Impacts on Water Resources in the Benue River Basin, Northern Cameroon.” Environmental Earth Sciences 78 (20): 606. Nonki, Rodric M, André Lenouo, Raphael M Tshimanga, Francine C Donfack, and Clément Tchawoua. 2021. “Performance Assessment and Uncertainty Prediction of a Daily Time-Step HBV-Light Rainfall-Runoff Model for the Upper Benue River Basin, Northern Cameroon.” Journal of Hydrology: Regional Studies 36:100849. Ogden, Fred L, Nawa Raj Pradhan, Charles W Downer, and Jon A Zahner. 2011. “Relative Importance of Impervious Area, Drainage Density, Width Function, and Subsurface Storm Drainage on Flood Runoff from an Urbanized Catchment.” Water Resources Research 47 (12). Olivier, Stephen L, and Jan F Prins. 2009. “Evaluating OpenMP 3.0 Run Time Systems on Unbalanced Task Graphs.” In International Workshop on OpenMP , 63–78. Springer. Padji, Calvin, Cyrille Meukaleuni, Cyrille Mezoue Adiang, Daniel Bongue, and David Monkam. 2024. “Estimation of Return Dates and Return Levels of Extreme Rainfall in the City of Douala, Cameroon.” Heliyon 10 (14): e34832. Papalexiou, Simon Michael, and Demetris Koutsoyiannis. 2016. “A Global Survey on the Seasonal Variation of the Marginal Distribution of Daily Precipitation.” Advances in Water Resources 94:131–45. Raouf, Abdou, Kasi Njeudjang, Jean Ghislain Tabue Youmbi, Xiaohui Lu, Jianhua Yue, Eucharia Chidinma Okoro, Elijah Chibueze Odii, and Ali Moussa Abdoul Wahab. 2025. “Factors Influencing Groundwater Development and Mitigation Strategies in Adamawa Region, Cameroon: A Critical Review.” Sustainable Water Resources Management 11 (1): 5. Seneviratne, Sonia, Neville Nicholls, David Easterling, Clare Goodess, Shinjiro Kanae, James Kossin, Yali Luo, Jose Marengo, Kathleen McInnes, and Mohammad Rahimi. 2012. “Changes in Climate Extremes and Their Impacts on the Natural Physical Environment.” Sivapalan, Murugesu. 2003. “Prediction in Ungauged Basins: A Grand Challenge for Theoretical Hydrology.” Smith, Keith, and Roy Ward. 1998. Floods: Physical Processes and Human Impacts. Sylla, Mouhamadou Bamba, Pinghouinde Michel Nikiema, Peter Gibba, Ibourahima Kebe, and Nana Ama Browne Klutse. 2016. “Climate Change over West Africa: Recent Trends and Future Projections.” Adaptation to Climate Change and Variability in Rural West Africa , 25–40. Tchuikoua, Louis. 2020. “Lecture de La Dynamique Territoriale à Douala et Son Impact Sur La Gestion Des Déchets Solides Ménagers Par La Théorie de La Formation Socio-Spatiale.” Environnement, Ingénierie & Développement . Tsalefac, Maurice, François Hiol Hiol, Gil Mahé, Alain Laraque, Denis Jean Sonwa, Paul Scholte, Wilfried Pokam, Andreas Haensler, Tazebe Beyene, and Fulco Ludwig. 2015. “Climate of Central Africa: Past, Present and Future.” Tschakert, Petra, Regina Sagoe, Gifty Ofori-Darko, and Samuel Nii Codjoe. 2010. “Floods in the Sahel: An Analysis of Anomalies, Memory, and Anticipatory Learning.” Climatic Change 103 (3): 471–502. Tume, Suiven John Paul, Yongka Bernard Nyuyse, Clarius Kuyar, Ngoran Larissa Tomla, Ngoran Hassan Afoni, Foka Sherifatu Kinyuy, and Ngala Elvis Ngum. 2025. “Vulnerability of Urban Housing to Climate Change in Cameroon.” In Designing Healthy Buildings and Communities: Shaping a Climate-Resilient Future , 49–84. Springer. Uddin, Mohammad Kashif, and Mukhtar M Salah. 2018. “Statistical Analysis of Litchi Chinensis’s Adsorption Behavior toward Cr (VI).” Applied Water Science 8 (5): 140. UNDRR. 2020. “Words into Action Guidelines: Developing Early Warning Systems. United Nations Office for Disaster Risk Reduction.” Villarini, Gabriele, and James A Smith. 2010. “Flood Peak Distributions for the Eastern United States.” Water Resources Research 46 (6). Yonzoua, C C Fandjio, U J M Pettang Nana, M B Manjiaa, and C Pettanga. 2021. “Mesure de La Vulnérabilité Des Milieux Urbains Au Cameroun Face Au Changement Climatique: Measuring the Vulnerability of Urban Areas in Cameroon to Climate Change.” Journal of the Cameroon Academy of Sciences 17 (2): 115–29. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8073038","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":554485448,"identity":"bbf02b58-c1ed-41b5-a04c-909cfd067b93","order_by":0,"name":"Calvin Padji","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7ElEQVRIiWNgGAWjYHACA4MEBgYZBmYGxgdAHg8fsVp4gFqYDUBa2IjRAiJ4gJhNAsQiqMWc/fCGggc1djwGx3mPVX7NsZNhY2B++OgGHi2WPWkFBgnHknkMDvOl3Zbdlgx0GJuxcQ4+Vx3IMTBIbGDmkWzmMbstuY0ZqIWHTRqvlvNvQFrqwVqKJbfVE6HlBtiWwzz8zDxmjB+3HSasxXLGM5BfjoO0GEszbjvOw8ZMwC/m/MnbDH/UVMux8Z8x/PhzW7U9P3vzw8d4HQaMCAMYh5kHTOJRDtXC/ADGYfxBQPUoGAWjYBSMTAAApwA9yBMpFbcAAAAASUVORK5CYII=","orcid":"","institution":"University of Douala","correspondingAuthor":true,"prefix":"","firstName":"Calvin","middleName":"","lastName":"Padji","suffix":""},{"id":554485449,"identity":"25dee0ba-dc87-4005-9ce3-712eb16db85e","order_by":1,"name":"Cyrille Meukaleuni","email":"","orcid":"","institution":"University Institute of the Coast","correspondingAuthor":false,"prefix":"","firstName":"Cyrille","middleName":"","lastName":"Meukaleuni","suffix":""},{"id":554485450,"identity":"97a19225-aa5d-44b0-a300-59b1ef27e0ea","order_by":2,"name":"David Monkam","email":"","orcid":"","institution":"University of Douala","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"Monkam","suffix":""}],"badges":[],"createdAt":"2025-11-10 05:38:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8073038/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8073038/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97528997,"identity":"82059d89-e5f3-4c38-bd10-6ec56bb76dc8","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2301254,"visible":true,"origin":"","legend":"","description":"","filename":"Articlespringer.docx","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/475db3df29a0d8347ca08fdf.docx"},{"id":97528993,"identity":"7c9c8f88-4724-4ade-bab9-87544ce6f791","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6192,"visible":true,"origin":"","legend":"","description":"","filename":"1bef670e0f9a43598e9ed978b3562135.json","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/3ba86990ba181a73d24d8b78.json"},{"id":97528999,"identity":"0fcfa0be-01f8-47d2-88e9-44437013581a","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192514,"visible":true,"origin":"","legend":"","description":"","filename":"1bef670e0f9a43598e9ed978b35621351enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/240b9f9deaef86b8158b9381.xml"},{"id":97670904,"identity":"a73cf536-b87d-448c-8670-6ab535c62295","added_by":"auto","created_at":"2025-12-08 09:31:30","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":75280,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/8e71c4e9e05dc61d1a973d23.jpeg"},{"id":97671351,"identity":"9eefdabc-4d9c-4b88-9c03-7f36ecbaec54","added_by":"auto","created_at":"2025-12-08 09:32:31","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":100275,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/330c04d538c366e35ac79595.jpeg"},{"id":97529004,"identity":"6e0de130-4f16-4559-9c8e-c0f0c2243541","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78156,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/4da24a3a4f97f9319c4622cb.jpeg"},{"id":97529002,"identity":"e35310bc-e27b-4de1-9b85-97919142b63d","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":115158,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/bb99faffca0f64e01c9992e1.jpeg"},{"id":97529009,"identity":"4555e030-9eba-4103-aa50-bf5c85a8590e","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82892,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/9074dac8b743d3579bd98a6a.jpeg"},{"id":97529007,"identity":"248ce01a-0942-4760-bfbc-a314107244d6","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":81710,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/1b770e0d4425834460cbe9b9.jpeg"},{"id":97671468,"identity":"8f77b881-f3fe-4d5f-8b9c-e9059833d767","added_by":"auto","created_at":"2025-12-08 09:32:38","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":79605,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/0fb4ab329cec75c831eb80e8.jpeg"},{"id":97529010,"identity":"a33ff251-6e9d-4a5d-ba89-c91dc6e96a72","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":90905,"visible":true,"origin":"","legend":"","description":"","filename":"groupimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/8ec9d3f821f1948746bc52fa.jpeg"},{"id":97671578,"identity":"6753a4ac-0763-4524-90d9-0515e4dd51e0","added_by":"auto","created_at":"2025-12-08 09:32:45","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14620,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/b23e7f53ae5ee86a2a6e418f.png"},{"id":97529011,"identity":"949f3fed-e8c9-4ebe-8f94-d02a1039411b","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46619,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/dd7262990514bac3d702452d.png"},{"id":97672479,"identity":"ada5a0d3-a62e-4607-af7d-3fcca2c9b97e","added_by":"auto","created_at":"2025-12-08 09:38:05","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23706,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/e80c91fb4e78dc05b6ebb554.png"},{"id":97529019,"identity":"b43471f6-3a35-485f-b34f-ccc608a07525","added_by":"auto","created_at":"2025-12-05 13:04:19","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31128,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/dd7da2abf08ffaedab131256.png"},{"id":97672085,"identity":"9785c64d-ffa5-4c1d-a712-778138731b35","added_by":"auto","created_at":"2025-12-08 09:34:07","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":32894,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/ce70053e79742d69a4cbb8d8.png"},{"id":97529014,"identity":"0def586a-5bfb-43bb-850c-6d8d41e18d96","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33300,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/5db793f9dc184f87e516f7f4.png"},{"id":97529015,"identity":"3b1d86c3-691f-4dff-a976-80e5dd299096","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20649,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/fbdf6c7157abd072c68adbba.png"},{"id":97529016,"identity":"36b30827-7453-4234-8719-8b1ce2434f2a","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28146,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinegroupimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/5f86b60525f8897f9a36ca4f.png"},{"id":97670922,"identity":"9df1569f-acb2-4d67-913e-32d8f1389082","added_by":"auto","created_at":"2025-12-08 09:31:31","extension":"xml","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":188776,"visible":true,"origin":"","legend":"","description":"","filename":"1bef670e0f9a43598e9ed978b35621351structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/16103b15acd90df4eb3611a7.xml"},{"id":97529018,"identity":"9de59b0f-e9fd-496f-859e-22f8de24b684","added_by":"auto","created_at":"2025-12-05 13:04:19","extension":"html","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":200464,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/50bb67d28a7944e2737405d6.html"},{"id":97671953,"identity":"17049fae-ac1c-4d8b-aa2b-28cb73665134","added_by":"auto","created_at":"2025-12-08 09:33:29","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1072062,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eGeographical location of the city of Douala, the Urban Zone and the main watersheds in the map of Africa and Cameroon.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"ArticlespringerCopy1.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/df9ee7658d5bb692088cc54e.png"},{"id":97528995,"identity":"bcce13d9-766c-4419-a8ea-237f7bc86d29","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":58900,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMulti-panel analysis of rainfall and discharge in Douala (2010–2023).\u003c/strong\u003e (a) Boxplots show positively skewed rainfall distributions with frequent outliers, contrasted with more moderated discharge responses. (b) Time series highlight seasonal and inter-annual variability, with synchronized peaks but dampened discharge amplitudes. (c) Rainfall histogram indicates the dominance of low-intensity events, while (d) discharge histogram approximates a normal distribution, reflecting watershed regulation of rainfall inputs.\u003c/p\u003e","description":"","filename":"ArticlespringerCopy2.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/7e7978bc3ecb7d668176d137.png"},{"id":97528992,"identity":"46c9ae7f-e343-46be-b56b-c4acc654b552","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":76950,"visible":true,"origin":"","legend":"\u003cp\u003eThreshold selection for rainfall (Panels a–c) and discharge (Panels d–f) using Mean Residual Life (MRL) plots and parameter stability diagnostics. The selected thresholds (20 mm for rainfall and 16 mm/day for discharge) are indicated by green lines, with practical stability ranges shown in red. Linearity in MRL plots and stability of shape (ξ) and scale (σ) parameters confirm the suitability of these thresholds for Generalized Pareto Distribution (GPD) modeling under the Peak-Over-Threshold (POT) framework.\u003c/p\u003e","description":"","filename":"ArticlespringerCopy3.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/9bc1d7e722489bee4c58c34a.png"},{"id":97528996,"identity":"8552675e-a0d3-44d3-a56d-247fc28c2b31","added_by":"auto","created_at":"2025-12-05 13:04:18","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":64008,"visible":true,"origin":"","legend":"\u003cp\u003eDiagnostic plots for the fitted Generalized Pareto Distribution (GPD) model of extreme rainfall in Douala (threshold = 20 mm). (a) Probability–Probability (PP) plot, (b) Quantile–Quantile (QQ) plot, (c) density plot, and (d) cumulative distribution function (CDF). All diagnostics confirm a robust GPD fit to rainfall exceedances.\u003c/p\u003e","description":"","filename":"ArticlespringerCopy4.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/ef33f41199321e964bbbe1e2.png"},{"id":97671342,"identity":"c176b090-06fc-44fd-808c-74c874f80b8d","added_by":"auto","created_at":"2025-12-08 09:32:30","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":67316,"visible":true,"origin":"","legend":"\u003cp\u003eDiagnostic plots for the fitted Generalized Pareto Distribution (GPD) model of extreme discharge in Douala (threshold = 16 mm/day). (a) Probability–Probability (PP) plot, (b) Quantile–Quantile (QQ) plot, (c) density plot, and (d) cumulative distribution function (CDF). The model provides a reliable fit to discharge exceedances, with slight underestimation of the most extreme flows.\u003c/p\u003e","description":"","filename":"ArticlespringerCopy5.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/380c5f240419e2cae7899772.png"},{"id":97670772,"identity":"9f727b8f-5d81-41d6-af82-5844d92be778","added_by":"auto","created_at":"2025-12-08 09:31:18","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":75841,"visible":true,"origin":"","legend":"\u003cp\u003eReturn level–return period relationships for rainfall and discharge extremes in Douala: (a) rainfall, (b) discharge, (c) combined rainfall–discharge comparison, (d) rainfall (analytical), (e) discharge (analytical), (f) combined rainfall–discharge (analytical).\u003c/p\u003e","description":"","filename":"ArticlespringerCopy6.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/15e0d3b64bfe259f2811c788.png"},{"id":97671190,"identity":"cdb21568-0565-46de-893c-b419a35fce36","added_by":"auto","created_at":"2025-12-08 09:32:07","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":60641,"visible":true,"origin":"","legend":"\u003cp\u003eRainfall–discharge connectivity in Douala: (a) scatterplot of daily rainfall versus discharge (Pearson r = 0.318), (b) cross-correlation analysis showing lagged hydrological response.\u003c/p\u003e","description":"","filename":"ArticlespringerCopy7.png","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/d086d7ff306ec9de5dd5bf45.png"},{"id":100042620,"identity":"babc5a19-2e5b-48cd-bb74-1639df217943","added_by":"auto","created_at":"2026-01-12 11:25:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3756117,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8073038/v1/d418e749-1bf0-492e-afbc-58bfb810a53e.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Analyzing the Hydrological Disconnect Between Extreme Precipitation and River Discharge in Douala, Cameroon","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eFloods are among the most devastating natural disasters worldwide, causing significant human, economic, and environmental losses. Over the past decades, the frequency and severity of flood events have been amplified by the joint effects of climate variability, land-use change, and rapid urbanization (Legg \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Forster, Storelvmo, and Alterskj\u0026aelig; 2021). Extreme precipitation events, in particular, play a central role in triggering flash floods and riverine floods, especially in tropical and coastal regions where high-intensity rainfall events are recurrent (Alfieri et al. 2018; Jia et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). According to the Centre for Research on the Epidemiology of Disasters (CRED \u0026amp; UNDRR 2022), floods account for more than 40% of all natural disasters globally, with Asia and Africa being the most affected continents. The social consequences include the displacement of populations, the spread of waterborne diseases, and damages to infrastructure, while the economic costs often represent several percentage points of national GDP for highly vulnerable countries (UNDRR \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Despite significant progress in flood risk modeling, predicting the timing and magnitude of hydrological extremes remains a challenge, particularly in regions where observational data are scarce or incomplete (Bl\u0026ouml;schl et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn Africa, floods are increasingly becoming one of the most pressing environmental challenges. Several studies report that the frequency of flood-related disasters has more than doubled in sub-Saharan Africa since the 1980 (Di Baldassarre et al. 2010; Tschakert et al. 2010). Major cities such as Lagos, Accra, Dakar, and Douala are particularly vulnerable due to their rapid demographic growth, poorly regulated urban expansion, and inadequate drainage infrastructure (A. C. Douglas et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Adelekan \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). In West and Central Africa, climate projections indicate that the intensity of extreme rainfall events is likely to increase under warming scenarios, thereby exacerbating flood hazards (Sylla et al. \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Dosio et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, most of the existing research on African floods has primarily focused on either precipitation extremes or hydrological modeling of river basins, with limited efforts to explicitly connect the two processes through joint statistical approaches. This constitutes an important scientific gap, as understanding the rainfall\u0026ndash;runoff relationship at extreme scales is fundamental to designing effective flood management and early-warning systems in the region (Nka et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe city of Douala, Cameroon\u0026rsquo;s economic capital and largest port, epitomizes this situation. Located in a low-lying coastal zone and traversed by the Wouri River, Douala has experienced recurrent and severe flooding over the past decades (Kometa and Ebot 2012; Tsalefac et al. \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The combined effect of extreme rainfall events, inadequate drainage, and unplanned settlement on floodplains has resulted in persistent human and material losses. For instance, the floods of August 2020 and July 2022 led to the displacement of thousands of households, the destruction of roads and houses, and increased vulnerability to cholera outbreaks (Akwa and Nguimbous 2021). Although several studies have analyzed rainfall variability and extremes in Douala (Adzandeh, Alaigba, and Nkemasong \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Nonki et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), very few have attempted to statistically link precipitation extremes to river discharge dynamics in the Wouri basin. Moreover, the majority of previous research has concentrated on annual maxima or seasonal variability, without considering the dynamics of return periods and return levels in terms of both precipitation and river discharge. This methodological gap prevents a comprehensive understanding of how extreme rainfall translates\u0026mdash;or fails to translate\u0026mdash;into extreme flows in Douala\u0026rsquo;s hydrological system.\u003c/p\u003e\u003cp\u003eRecent literature in hydrology and climate extremes provides useful insights but also shows critical limitations. For example, Papalexiou and Koutsoyiannis, (2016) emphasized the need for robust statistical frameworks for extreme rainfall, highlighting the limitations of traditional approaches such as block maxima that often waste valuable data. Similarly, Coles et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2001\u003c/span\u003e); Katz (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) demonstrated the relevance of the Generalized Pareto Distribution (GPD) and Peak Over Threshold (POT) methods for more accurate estimation of return levels. In the African context, studies like Nka et al. (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) on Cameroon and Ogden et al. (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) on West Africa have applied extreme value theory to precipitation, but without systematically incorporating discharge data. Other works, such as Di Baldassarre et al. (2010) in the Sahel or Kundzewicz et al. (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) globally, have underlined the growing mismatch between rainfall extremes and observed flood damages, yet they stop short of quantifying this gap in terms of return periods and lag times. This indicates that while precipitation extremes are well documented, the hydrological response\u0026mdash;especially the lag between rainfall and discharge peaks\u0026mdash;remains underexplored, particularly in tropical urban basins.\u003c/p\u003e\u003cp\u003eThe originality of the present study lies precisely in addressing these gaps. By combining daily precipitation data from Douala with daily discharge data from the Wouri River, we aim to calculate and compare the return levels and return periods of both precipitation and discharge extremes. More importantly, we will investigate the potential temporal lag between extreme rainfall events and peak discharges, highlighting situations where extreme precipitation does not immediately translate into extreme flows. This lag effect, if confirmed, would provide critical insights into the dynamics of the Wouri River and the broader challenges of flood risk in Douala. To our knowledge, no previous study in Douala\u0026mdash;or even more broadly in tropical Africa\u0026mdash;has systematically quantified this rainfall\u0026ndash;discharge lag in terms of extreme value theory. Internationally, although some studies have explored rainfall\u0026ndash;runoff correlations under extreme conditions (e.g., (Bl\u0026ouml;schl et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Berghuijs et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), they have largely focused on temperate regions with dense hydrometric networks, leaving tropical urban basins underrepresented. Our work therefore contributes an innovative methodological and geographical perspective, reinforcing the scientific and practical relevance of this research.\u003c/p\u003e\u003cp\u003eThe consequences of such an approach are twofold. First, by quantifying return levels and return periods for both precipitation and discharge extremes, we provide a more nuanced understanding of hydrological risks in Douala. This knowledge is vital for urban planning, flood defense design, and early warning systems. Second, by identifying potential lags or mismatches between rainfall and river response, our research sheds light on the complex dynamics of the Wouri basin, where human factors such as land use, drainage infrastructure, and river channel modifications may alter the natural rainfall\u0026ndash;runoff relationship. Such findings could be transferable to other rapidly urbanizing tropical basins worldwide, where similar dynamics are at play but remain poorly documented in the scientific literature.\u003c/p\u003e\u003cp\u003eThis dissertation is organized as follows. Chapter One presents the global and African context of floods and extreme precipitation, with a focus on Douala\u0026rsquo;s vulnerability. Chapter Two introduces the data and methodology, particularly the use of extreme value theory (EVT) approaches such as the Generalized Pareto Distribution (GPD) and the Peak Over Threshold (POT) method for estimating return periods and levels. Chapter Three discusses the empirical results, comparing precipitation and discharge extremes, quantifying their return levels, and analyzing the temporal lag between them. The final section synthesizes the findings, highlights the originality and limitations of the study, and provides recommendations for flood risk management and future research in Douala and similar urban contexts.\u003c/p\u003e"},{"header":"2. Zone, Data and methods study","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Study area.\u003c/h2\u003e\u003cp\u003eDouala, the economic capital of Cameroon, is located in the Littoral Region on the Atlantic coast of Central Africa. The city lies between latitudes 3\u0026deg;40\u0026prime;\u0026ndash;4\u0026deg;10\u0026prime; N and longitudes 9\u0026deg;40\u0026prime;\u0026ndash;10\u0026deg;00\u0026prime; E, covering an estimated surface area of about 210 km\u0026sup2;. It is the most populated city in Cameroon, with over 3.5\u0026nbsp;million inhabitants according to recent estimates, and it serves as the country\u0026rsquo;s principal seaport and commercial hub. Its geographical position at the mouth of the Wouri River and proximity to the Gulf of Guinea makes it highly strategic for trade, but also particularly exposed to hydrometeorological hazards (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe climate of Douala is classified as equatorial humid, characterized by very high annual rainfall exceeding 3,500 mm on average, with some years recording more than 4,000 mm. Precipitation is strongly seasonal, with a long rainy season extending from March to November and a short dry season between December and February. The peak of rainfall generally occurs between July and September, during which torrential rains often cause severe flooding. The city is also marked by persistently high humidity and temperatures ranging between 23\u0026deg;C and 31\u0026deg;C throughout the year, conditions that favor intense convective activity.\u003c/p\u003e\u003cp\u003eHydrologically, Douala is dominated by the Wouri River basin, which drains a large catchment before discharging into the Atlantic Ocean. Numerous small rivers, streams, and drainage channels cross the city, many of which are directly connected to the Wouri estuary. These watercourses, combined with the city\u0026rsquo;s flat and low-lying topography, increase its vulnerability to both riverine and pluvial flooding. In addition, the presence of extensive wetlands and mangrove ecosystems around the estuary plays a dual role: they act as natural buffers against flooding, but are increasingly threatened by rapid urbanization and land reclamation.\u003c/p\u003e\u003cp\u003eFrom an urban perspective, Douala has experienced rapid and largely uncontrolled population growth and spatial expansion over the past decades. Informal settlements are widespread, often located in flood-prone zones such as riverbanks, marshes, and poorly drained lowlands. The lack of adequate drainage infrastructure, coupled with unplanned urban sprawl, significantly amplifies the impacts of extreme rainfall events. Consequently, floods in Douala regularly disrupt transportation, damage property, and pose severe public health risks, particularly through the spread of waterborne diseases such as cholera and typhoid fever.\u003c/p\u003e\u003cp\u003eThe geographical and environmental setting of Douala therefore makes it a critical hotspot for hydroclimatic risk studies in West and Central Africa. Its combination of high rainfall exposure, sensitive river systems, rapid urbanization, and limited adaptive infrastructure creates a unique context where the study of extreme precipitation and river discharge dynamics is particularly relevant. Understanding this geographical situation is essential for designing robust flood risk management and early warning strategies that can reduce vulnerability in this rapidly growing coastal megacity.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Data study.\u003c/h2\u003e\u003cp\u003eThe Wouri basin, encompassing the densely populated and economically critical city of Douala, Cameroon, is increasingly vulnerable to flooding driven by extreme precipitation events. Understanding the dynamic relationship between these rainfall extremes and the subsequent hydrological response of the river system is paramount for effective water resource management and flood mitigation strategies. This study focuses on the Wouri basin from 2010 to 2023, aiming to quantify the link between extreme daily precipitation and the immediate runoff response, thereby elucidating the basin's behavior under severe meteorological conditions.\u003c/p\u003e\u003cp\u003eTo analyze precipitation patterns, this study utilized daily data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) dataset. CHIRPS incorporates satellite imagery with in-situ station data to create gridded precipitation time series, particularly suited for trend analysis and drought and flood monitoring in data-sparse regions Funk et al. (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The daily precipitation values were extracted for a point location representing the city of Douala (9.7\u0026deg;E, 4.05\u0026deg;N), providing a high-resolution, long-term record of rainfall inputs to the basin.\u003c/p\u003e\u003cp\u003eThe hydrological response of the basin was assessed using daily runoff data from the ERA5-Land reanalysis dataset, produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). ERA5-Land provides a globally consistent and comprehensive replay of land variables from 1950 to the present, offering a reliable representation of key hydrological fluxes like runoff and total precipitation at an enhanced resolution of ~\u0026thinsp;9 km (Hersbach et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Mu\u0026ntilde;oz-Sabater et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Hersbach et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Daily runoff_sum and total_precipitation_sum data were averaged over the entire Wouri basin polygon to characterize the aggregate hydrological behavior of the catchment in response to meteorological forcings.\u003c/p\u003e\u003cp\u003eThe methodology involved extracting and collating daily time series for both precipitation and runoff. Extreme precipitation events were defined as days with rainfall exceeding 50 mm. A key metric, the runoff ratio (daily runoff / daily precipitation), was calculated to assess the efficiency of runoff generation. The analysis compared general statistics with conditions during extreme events to identify shifts in hydrological behavior. Furthermore, the temporal lag between precipitation peaks and runoff peaks was examined to infer the basin's response time. This integrated use of CHIRPS precipitation data (Climate Hazards Group InfraRed Precipitation with Station data; (Funk et al. \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) and ERA5-Land reanalysis runoff data (Mu\u0026ntilde;oz-Sabater et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) provides a robust framework for analyzing precipitation\u0026ndash;runoff dynamics in a basin where traditional gauge data may be limited.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Method.\u003c/h2\u003e\u003cp\u003eThe analysis of extremes was conducted using the Peaks Over Threshold (POT) framework of Extreme Value Theory (EVT) (Coles et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; R. W. Katz, Parlange, and Naveau 2002). Prior to applying the POT approach, the rainfall and discharge series were subjected to rigorous quality control. Outliers were identified and removed using the interquartile range (IQR) and z-scores, short missing gaps were filled by interpolation, and long gaps were excluded. Stationarity was tested with the Augmented Dickey\u0026ndash;Fuller (ADF) test, while monotonic trends were assessed using the Mann\u0026ndash;Kendall test (Kendall \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1948\u003c/span\u003e; Helsel and Hirsch 1993). These pre-processing steps ensured the robustness of the datasets for extreme value analysis.\u003c/p\u003e\u003cp\u003eThresholds for extreme event extraction were determined using mean residual life (MRL) plots and threshold stability plots (Davison and Smith 1990). Exceedances above the selected thresholds were fitted to the Generalized Pareto Distribution (GPD), with shape (ξ) and scale (σ) parameters estimated via Maximum Likelihood Estimation (MLE). The adequacy of the fitted models was evaluated through diagnostic checks, including Quantile\u0026ndash;Quantile (Q\u0026ndash;Q), Probability\u0026ndash;Probability (P\u0026ndash;P), and Cumulative Distribution Function (CDF) plots, along with the Kolmogorov\u0026ndash;Smirnov goodness-of-fit test (Choulakian and Stephens 2001). Return levels were then derived for recurrence intervals of 2, 5, 10, 20, and 50 years, with uncertainty quantified using parametric bootstrapping (Effron and Tibshirani 1993).\u003c/p\u003e\u003cp\u003eIn addition to conventional EVT analysis, the study employed the Padji Calvin Method (Padji et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), which innovatively introduces the concept of return dates. This method extends EVT by converting statistical return periods into calendar-based dates, achieved by mapping threshold exceedances to their empirical occurrence within the annual cycle and adjusting for leap years. Finally, rainfall\u0026ndash;runoff coupling was investigated by analyzing discharge responses within a 0\u0026ndash;5-day lag window after rainfall extremes, using cross-correlation functions (CCF) (Box and Jenkins \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1976\u003c/span\u003e). This approach allowed the identification of potential temporal mismatches between meteorological and hydrological extremes, providing novel insights into flood dynamics in Douala.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results and discussion","content":"\u003cp\u003e\u003cstrong\u003e3.1. Rainfall\u0026ndash;Discharge Variability and Hydrological Dynamics in Douala.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 2 presents a multi-panel analysis of rainfall and discharge in Douala over the period 2010\u0026ndash;2023.\u003c/strong\u003e Panel (a) illustrates the distributional characteristics of both variables through boxplots. Rainfall exhibits a positively skewed distribution with numerous outliers, reflecting the predominance of low-intensity events interspersed with episodic extremes, a pattern typical of tropical rainfall regimes (Koutsoyiannis 2004; R. W. Katz, Parlange, and Naveau 2002). In contrast, discharge displays a more symmetrical distribution with fewer extreme values, highlighting the dampening influence of watershed processes on the rainfall\u0026ndash;runoff response (Beven 2012; Bl\u0026ouml;schl and Sivapalan 1995). The lower central tendency of discharge relative to rainfall underscores the role of hydrological losses such as infiltration, evapotranspiration, and storage (Dingman 2015).\u003c/p\u003e\n\u003cp\u003ePanel (b) shows the temporal dynamics of rainfall and discharge. Both series reveal marked seasonality and inter-annual variability, with synchronized peaks corresponding to major hydrological events. However, discharge responses are attenuated relative to rainfall, particularly during extreme precipitation episodes, confirming the buffering capacity of the watershed through storage and transmission losses (Gupta and Waymire 1993; Sivapalan 2003). Intensified hydrological activity is visible during 2015\u0026ndash;2017, likely linked to broader climatic anomalies such as ENSO-related variability (Nicholson 2013; Lyon and Vigaud 2017).\u003c/p\u003e\n\u003cp\u003ePanels (c) and (d) present the frequency distributions of rainfall and discharge. The rainfall histogram confirms the predominance of low-intensity events, with an exponential decline in frequency as intensity increases (Coles et al. 2001; R. W. Katz, Parlange, and Naveau 2002). Conversely, the discharge histogram approximates a normal distribution, reflecting the watershed\u0026rsquo;s ability to regulate highly variable rainfall inputs into more stable streamflow outputs (Beven 2012). These frequency characteristics provide valuable measures of central tendency and variability for hydrological risk assessments (Legg 2021).\u003c/p\u003e\n\u003cp\u003eTaken together, the comparative analysis demonstrates that rainfall acts as the primary driver of streamflow variability, but watershed characteristics substantially modify this relationship (Bl\u0026ouml;schl and Sivapalan 1995; Sivapalan 2003). The attenuation of discharge extremes and the temporal lag between rainfall peaks and discharge responses point to the critical role of antecedent soil moisture, vegetation cover, and geomorphological features in shaping hydrological behavior (Dingman 2015; Gupta and Waymire 1993). These findings provide valuable insights for flood forecasting, water resource management, and climate adaptation strategies (Legg 2021), while underscoring the persistence of rainfall\u0026ndash;discharge dynamics revealed by daily observations over a 13-year period.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2.\u003c/strong\u003e \u003cstrong\u003eExtreme Value Modeling of Rainfall and Discharge in Douala: A Peak-Over-Threshold Approach for Flood Risk Assessment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe application of the Peak-Over-Threshold (POT) method hinges on a critical compromise: selecting a threshold that is sufficiently high to satisfy the asymptotic basis of the Generalized Pareto Distribution (GPD), yet low enough to retain an adequate number of excesses for precise parameter estimation. This selection is paramount, as an ill-chosen threshold can lead to biased estimates and invalid inferences regarding the tail behavior of the data. Therefore, prior to model fitting, the following section is devoted to the objective and diagnostic-driven selection of optimal thresholds for the rainfall and discharge series in Douala.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2.1. \u003cstrong\u003eModeling Extreme Hydrological Events in Douala: A Generalized Pareto Distribution (GPD) Approach.\u003c/strong\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe validity of the Generalized Pareto Distribution (GPD) model is contingent upon the selection of an optimal threshold. An ill-chosen threshold violates the model\u0026apos;s asymptotic assumptions, potentially leading to significant bias in the estimation of extreme quantiles and return levels.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2.1.1. Threshold Selection and Fitting of the Generalized Pareto Distribution (GPD) Model.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe multi-panel figure 3 presents a comprehensive threshold selection analysis for both rainfall and discharge data using Mean Residual Life (MRL) plots and threshold choice plots, essential for implementing the Peak-Over-Threshold (POT) approach in extreme value analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePanels a), b), and c)\u003c/strong\u003e focus on rainfall threshold determination.\u0026nbsp;\u003cstrong\u003ePanel a)\u003c/strong\u003e (MRL plot) demonstrates the mean excess function for rainfall data, which exhibits initial instability followed by a linear trend above approximately 18 mm. The optimal threshold of 20 mm (green line) is identified where the mean excess becomes approximately linear, satisfying the key assumption of generalized Pareto distribution (GPD) applicability. The threshold range of 18-23 mm (red lines) indicates the region where the GPD model provides stable parameter estimates. The linearity above this threshold confirms that excess rainfall values follow the GPD, validating the threshold choice for extreme rainfall analysis.\u0026nbsp;\u003cstrong\u003ePanel b)\u003c/strong\u003e shows the stability of the shape parameter (\u0026xi;) across different threshold values. The relative stability of \u0026xi; above 20 mm indicates that the distribution of rainfall excesses maintains consistent tail behavior, confirming the threshold appropriateness. The convergence of parameter estimates above this threshold suggests reliable extrapolation for return level estimation.\u0026nbsp;\u003cstrong\u003ePanel c)\u003c/strong\u003e displays the scale parameter (\u0026sigma;) behavior, which should increase linearly with threshold if the GPD assumption holds. The linear trend observed above the selected threshold supports the validity of the chosen threshold for extreme rainfall modeling.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePanels d), e), and f)\u003c/strong\u003e present the corresponding analysis for discharge data.\u0026nbsp;\u003cstrong\u003ePanel d)\u003c/strong\u003e (MRL plot) reveals an optimal discharge threshold of 16 mm/day, with a practical range of 13-20 mm/day. The linear mean excess above this threshold indicates that extreme discharge values follow the GPD, enabling reliable flood frequency analysis.\u0026nbsp;\u003cstrong\u003ePanel e)\u003c/strong\u003e demonstrates the stability of the shape parameter for discharge above 16 mm/day, suggesting consistent tail behavior for extreme flow events. The stable \u0026xi; parameter above the threshold confirms that the discharge extremes belong to the same distribution family, crucial for accurate flood risk assessment.\u0026nbsp;\u003cstrong\u003ePanel f)\u003c/strong\u003e shows the scale parameter behavior for discharge, with the expected linear relationship above the chosen threshold, further validating the threshold selection for extreme discharge analysis.\u003c/p\u003e\n\u003cp\u003eThe selected thresholds (20 mm for rainfall and 16 mm/day for discharge) represent the levels above which observations can be considered extreme events and appropriately modeled using the GPD framework. The stability of parameter estimates above these thresholds ensures reliable estimation of return levels and quantiles for rare hydrological events.\u003c/p\u003e\n\u003cp\u003eThese results provide statistically robust thresholds for extreme value analysis, enabling accurate estimation of design rainfall intensities and flood magnitudes for infrastructure planning and risk management. The concordance between different diagnostic methods (MRL plots and parameter stability plots) strengthens the confidence in the selected thresholds for subsequent extreme value modeling in hydrological applications.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2.1.2. Goodness-of-Fit Validation for the Generalized Pareto Distribution.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe diagnostic plots for the fitted Generalized Pareto Distribution (GPD) models provide a comprehensive visual assessment of the model\u0026apos;s adequacy in describing the tail behavior of extreme rainfall and discharge events in Douala.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 4. Diagnostic plots for the goodness-of-fit of the Generalized Pareto Distribution (GPD) model to extreme rainfall data (threshold:\u0026nbsp;\u003c/strong\u003euR = 20 mm\u003cstrong\u003e) in Douala: (a) Probability-Probability (PP) plot, (b) Quantile-Quantile (QQ) plot, (c) Density plot, and (d) Cumulative Distribution Function (CDF) plot.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFor Extreme Rainfall (\u003c/strong\u003euR = 20 mm\u003cstrong\u003e), shown in Figure 4:\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003eProbability-Probability (PP) Plot [4a]:\u003c/strong\u003e The close alignment of the points along the 1:1 perfect fit line indicates an excellent agreement between the empirical probabilities of the observed excess rainfall and the probabilities predicted by the fitted GPD model. This suggests that the model accurately captures the overall distribution of the data.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eQuantile-Quantile (QQ) Plot [4b]:\u003c/strong\u003e The strong linearity of the points, particularly in the upper tail, is a crucial result. It demonstrates that the model successfully replicates the magnitude and behavior of the most extreme rainfall events, which is the primary objective of Extreme Value Analysis.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eDensity Plot [4c]:\u003c/strong\u003e The close fit of the red GPD density curve to the histogram of observed excesses confirms that the model\u0026apos;s shape (\u0026xi;) and scale (\u0026sigma;) parameters accurately describe the frequency of extreme rainfall intensities.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eCumulative Distribution Function (CDF) Plot [4d]:\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eThe near-perfect overlap between the empirical CDF and the fitted model CDF provides strong evidence that the GPD is an appropriate model for the entire range of excess rainfall values.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion for Rainfall:\u003c/strong\u003e Collectively, these four diagnostic tools offer robust validation that the GPD model provides a statistically sound fit to the extreme rainfall data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 5. Diagnostic plots for the goodness-of-fit of the Generalized Pareto Distribution (GPD) model to extreme discharge data (threshold:\u0026nbsp;\u003c/strong\u003euD = 16 mm/day\u003cstrong\u003e) in Douala: (a) Probability-Probability (PP) plot, (b) Quantile-Quantile (QQ) plot, (c) Density plot, and (d) Cumulative Distribution Function (CDF) plot.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFor Extreme Discharge (\u003c/strong\u003euD = 16 mm\u003cstrong\u003e), shown in Figure 5:\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e\u003cstrong\u003ePP and QQ Plots [5a and 5b]:\u003c/strong\u003e The points in both plots adhere closely to the 1:1 line, indicating a good overall fit. The QQ plot shows that the model performs well across most quantiles, which is essential for predicting extreme flood discharges.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eDensity Plot [5c]:\u003c/strong\u003e The fitted density function follows the histogram of discharge excesses well. The slight underestimation in the very low-density region of the right tail suggests the model may slightly underestimate the frequency of the very largest events, but the overall fit is strong.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eCDF Plot [5d]:\u003c/strong\u003e The strong agreement between the empirical and modeled CDFs confirms that the GPD accurately represents the probability structure of extreme discharge values.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion for Discharge:\u003c/strong\u003e The diagnostic plots in Figure 5 validate the application of the GPD model to extreme discharge events. The model provides a reliable fit, justifying its use for flood frequency analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverall Synthesis:\u003c/strong\u003e The successful application of the Peak-Over-Threshold method, confirmed by these diagnostic checks (Figures 4 and 5), demonstrates that the extreme values of both rainfall and discharge in Douala are well-characterized by the Generalized Pareto Distribution. This establishes a solid statistical foundation for estimating the probability and magnitude of future extreme events.\u003c/p\u003e\n\u003cp\u003eThe adequacy of the Generalized Pareto Distribution (GPD) models for both rainfall and discharge extremes was rigorously assessed using Kolmogorov-Smirnov (K-S) and Anderson-Darling (A-D) goodness-of-fit tests. The results presented in Table 1 demonstrate strong statistical evidence supporting the validity of the fitted models.\u003c/p\u003e\n\u003cp\u003eFor extreme rainfall events exceeding the 20 mm threshold, both statistical tests yielded non-significant results (K-S: D = 0.044, p = 0.711; A-D: A = 0.482, p = 0.765), indicating no evidence to reject the null hypothesis that the observed excess rainfall data follows the fitted GPD. Similarly, for extreme discharge values above the 16 mm/day threshold, both tests confirmed the model\u0026apos;s adequacy (K-S: D = 0.077, p = 0.744; A-D: A = 0.535, p = 0.711).\u003c/p\u003e\n\u003cp\u003eThe consistency between both tests is particularly noteworthy given their different sensitivities to various aspects of distributional fit. The Kolmogorov-Smirnov test, more sensitive to deviations in the center of the distribution, and the Anderson-Darling test, more powerful for detecting discrepancies in the tails, both converge on the same conclusion. This concordance strengthens the validity of our threshold selection and parameter estimation.\u003c/p\u003e\n\u003cp\u003eThe presence of tied values in the dataset, common in environmental measurements due to instrumental precision, was addressed through appropriate statistical adjustments. The robustness of the Anderson-Darling test to such data characteristics provides additional confidence in our findings.\u003c/p\u003e\n\u003cp\u003eThese results collectively demonstrate that the GPD provides an excellent fit to the extreme values of both rainfall and discharge in Douala. The models adequately capture the tail behavior of the hydrological extremes, thereby validating their use for subsequent extreme value analysis, including the estimation of return levels and the quantification of probabilities for rare, high-magnitude events crucial for urban flood risk assessment and infrastructure planning.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1: Results of the Kolmogorov-Smirnov and Anderson-Darling goodness-of-fit tests for the fitted GPD models.\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"648\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\u003cstrong\u003eHydrological Variable\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\u003cstrong\u003eThreshold (mm/day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\u003cstrong\u003eTest Type\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003e\u003cstrong\u003eTest Statistic\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 99px;\"\u003e\u003cstrong\u003eP-Value\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 127px;\"\u003e\u003cstrong\u003eInterpretation\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 118px;\"\u003e\u003cstrong\u003eRainfall\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e20\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\u0026nbsp; \u0026nbsp; K-S\u003cbr\u003e\u0026nbsp; \u0026nbsp; A-D\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003eD = 0.0437\u003cbr\u003eA = 0.4824\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 99px;\"\u003e\u0026nbsp;0.711\u003cbr\u003e\u0026nbsp;0.765\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 127px;\"\u003eAdequate fit\u003cbr\u003eAdequate fit\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 118px;\"\u003e\u003cstrong\u003eDischarge\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e16\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\u0026nbsp; \u0026nbsp; K-S\u003cbr\u003e\u0026nbsp; \u0026nbsp; A-D\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 102px;\"\u003eD = 0.0775\u003cbr\u003eA = 0.5351\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 99px;\"\u003e\u0026nbsp;0.744\u003cbr\u003e\u0026nbsp;0.711\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 127px;\"\u003eAdequate fit\u003cbr\u003eAdequate fit\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e3.2.1.3 GPD parameter estimation results.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Generalized Pareto Distribution (GPD) was applied to exceedances above carefully selected thresholds for both rainfall and discharge series using the Maximum Likelihood Estimation (MLE) method. Threshold selection was guided by the Mean Residual Life (MRL) plot and quantile analysis, ensuring that a sufficient number of exceedances were retained for reliable parameter estimation while preserving the asymptotic properties of the GPD.\u003c/p\u003e\n\u003cp\u003eAs presented in Table 2, the scale parameter (\u0026sigma;) characterizes the dispersion of the exceedances, while the shape parameter (\u0026xi;) governs the behavior of the distribution tail. For rainfall extremes, the estimated shape parameter is positive (\u0026xi; = 0.19), indicating a heavy-tailed distribution of the Fr\u0026eacute;chet type. This implies that extreme rainfall events can potentially reach very large magnitudes. The 95% confidence interval for \u0026xi; (0.105\u0026ndash;0.289) includes zero, which means that, although the point estimate suggests heavy tails, an exponential tail (Gumbel type) cannot be statistically rejected at the 5% significance level.\u003c/p\u003e\n\u003cp\u003eFor discharge extremes, the shape parameter is also positive (\u0026xi; = 0.44), suggesting a heavy-tailed distribution. This result indicates that even after catchment storage and attenuation processes, discharge extremes can attain very high values. However, the wider confidence interval (0.180\u0026ndash;0.708) reflects greater uncertainty in parameter estimation relative to rainfall, likely arising from the nonlinear rainfall\u0026ndash;runoff transformation and the hydrological complexity of the watershed.\u003c/p\u003e\n\u003cp\u003eThe two datasets remain directly comparable: rainfall and discharge analyses were based on the same observational period (13.99 years). Nevertheless, the number of exceedances differed (757 for rainfall and 173 for discharge), reflecting the relative rarity of hydrological extremes after the catchment transformation. Model diagnostics (deviance and AIC) were higher for rainfall than discharge, pointing to greater variability in rainfall extremes compared to discharge, consistent with the moderating effect of watershed processes on the hydrological response.\u003c/p\u003e\n\u003cp\u003eOverall, these GPD parameter estimates provide the statistical foundation for the calculation of return levels and probabilities of rare hydrological extremes. Importantly, the differences observed between rainfall and discharge tail behaviors underscore the necessity of explicitly accounting for catchment processes in extreme value analysis, rather than assuming a direct one-to-one transfer of rainfall extremes into hydrological extremes.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Table 2: GPD parameter estimates for rainfall and discharge extremes in the city of Douala from 2010 to 2023. In brackets are the 95% confidence intervals associated with these parameters.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"663\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 194px;\"\u003e\u003cstrong\u003eParameter\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 241px;\"\u003e\u003cstrong\u003eRainfall Extremes\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 228px;\"\u003e\u003cstrong\u003eDischarge Extremes\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 194px;\"\u003e\u003cstrong\u003eScale (\u0026sigma;)\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eShape (\u0026xi;)\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eNumber of Exceedances\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eData Span\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eDeviance\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eAIC\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 241px;\"\u003e21.95 (19.41, 24.48)\u003cbr\u003e0.19 (0.105, 0.289)\u003cbr\u003e757 observations (1.48% of data)\u003cbr\u003e13.99 years\u003cbr\u003e6489.16\u003cbr\u003e6493.16\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 228px;\"\u003e19.26 (18.22, 20.3)\u003cbr\u003e0.24 (0.2, 0.283)\u003cbr\u003e173 observations (1.18% of data)\u003cbr\u003e13.99 years\u003cbr\u003e7993.72\u003cbr\u003e7997.72\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Comparison of Return Levels (RLs) and Return Periods (RPs) of Extreme Rainfall and Discharge\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFigure 6 provides an integrated assessment of Douala\u0026rsquo;s watercourse dynamics through six analytical panels. Panels a\u0026ndash;c highlights the distributional properties of extremes. The Generalized Pareto Distribution (GPD) analysis for rainfall (Panel a) yields a shape parameter of \u0026xi; = 0.1973, indicating moderately heavy-tailed behavior, consistent with extreme value theory applications in hydrology (Coles et al. 2001; R. W. Katz, Parlange, and Naveau 2002). In contrast, discharge extremes (Panel b) exhibit a substantially heavier tail (\u0026xi; = 0.4443), underscoring the amplification of flood risks within the watershed (Koutsoyiannis 2004). The comparative analysis (Panel c) reveals a systematic temporal displacement between rainfall and discharge return periods: equivalent magnitude events occur more frequently in discharge than in rainfall, particularly at longer return periods. This suggests that urban watershed processes transform rare rainfall extremes into relatively frequent flood occurrences (Smith and Ward 1998; Villarini and Smith 2010).\u003c/p\u003e\n\u003cp\u003ePanels d\u0026ndash;f expands on these patterns by examining return-period dynamics. Rainfall returns levels (Panel d) display relatively stable growth, whereas discharge return levels (Panel e) show steeper gradients and higher variability, particularly beyond the 100-year event (Beirlant and Goegebeur 2004). The integrated view (Panel f) demonstrates that discharge confidence intervals are substantially wider than those for rainfall, reflecting greater uncertainty in flood prediction (Coles et al. 2001). This uncertainty stems from urban features such as impervious surfaces, drainage inefficiencies, and modified flow pathways (Arnell and Gosling 2016; I. Douglas et al. 2008).\u003c/p\u003e\n\u003cp\u003eTaken together, these results highlight several key challenges:\u003c/p\u003e\n\u003col start=\"1\" type=\"1\"\u003e\n \u003cli\u003eHydrological amplification: Urban runoff processes compress rainfall returns periods into shorter discharge return periods, intensifying flood hazards (Miller and Hutchins 2017).\u003c/li\u003e\n \u003cli\u003eInfrastructure limitations: The widening gap between rainfall and discharge uncertainty reflects systemic inefficiencies in drainage capacity (I. Douglas et al. 2008).\u003c/li\u003e\n \u003cli\u003eTemporal compression of risks: Discharge extremes recur more frequently than their meteorological drivers, signaling reduced watershed storage and faster hydrological response (Seneviratne et al. 2012).\u003c/li\u003e\n \u003cli\u003eSpatial heterogeneity: Variability in confidence intervals indicates uneven flood vulnerability across different areas of the city (Habitat 2018).\u003c/li\u003e\n \u003cli\u003eClimate sensitivity: The heavy-tailed discharge distribution suggests disproportionate risks under projected increases in extreme rainfall (IPCC 2021).\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThese findings emphasize the urgent need for integrated interventions, including drainage rehabilitation, green infrastructure, improved land-use planning, and enhanced early warning systems (Di Baldassarre et al. 2010; IPCC 2021).\u003c/p\u003e\n\u003cp\u003eFigure 7 provides complementary evidence of systemic hydrological dysfunction through rainfall\u0026ndash;discharge relationship analysis. Panel a show a weak linear correlation (r \u0026asymp; 0.2\u0026ndash;0.3), indicating that daily rainfall explains only 4\u0026ndash;9% of discharge variability. The wide scatter of data points highlights the decoupling between rainfall inputs and river response, while the shallow regression slope underscores limited predictive capacity (Gupta and Waymire 1990).\u003c/p\u003e\n\u003cp\u003ePanel b examines temporal dynamics using cross-correlation. The maximum lagged correlation occurs after 1\u0026ndash;2 days, suggesting a delayed discharge response. However, the peak correlation remains modest (r \u0026asymp; 0.3\u0026ndash;0.4), reinforcing the weak coupling observed in Panel a. The flattened correlation profile indicates heterogeneous and prolonged hydrological responses across sub-catchments, shaped by urban modifications (Villarini and Smith 2010).\u003c/p\u003e\n\u003cp\u003eTogether, Panels a and b reveal:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eSystemic hydrological disconnection: Weak correlations confirm that urbanization has disrupted natural rainfall\u0026ndash;runoff connectivity (I. Douglas et al. 2008).\u003c/li\u003e\n \u003cli\u003eUrbanization effects: Impervious surfaces, drainage blockages, and fragmented flow pathways dominate discharge behavior, reducing rainfall predictability (Olivier and Prins 2009).\u003c/li\u003e\n \u003cli\u003eFlood risk implications: The combination of weak correlations and limited lead time complicates flood forecasting and increases uncertainty (Di Baldassarre et al. 2010).\u003c/li\u003e\n \u003cli\u003eInfrastructure deficiencies: Inadequate monitoring, poor drainage maintenance, and lack of integration between urban growth and water management exacerbate risks (Habitat 2018).\u003c/li\u003e\n \u003cli\u003eClimate vulnerability: Non-linear and disorganized responses reduce resilience and magnify risks under intensified rainfall (IPCC 2021).\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003ePolicy responses should prioritize distributed monitoring networks, systematic drainage rehabilitation, the expansion of green infrastructure, and better integration of hydrological considerations into urban planning (Seneviratne et al. 2012; A. C. Douglas et al. 2008).\u003c/p\u003e\n\u003cp\u003eFigures 6 and 7 collectively demonstrate that Douala\u0026rsquo;s flood risks are not solely a function of extreme rainfall but are significantly shaped by watershed transformation and urbanization. The amplification of discharge extremes, the weak rainfall\u0026ndash;discharge coupling, and the spatial\u0026ndash;temporal compression of risks provide strong evidence for urgent reforms in both infrastructure and governance to reduce the city\u0026rsquo;s flood vulnerability (Di Baldassarre et al. 2010; IPCC 2022).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4. Estimation of Return Periods, Dates, and Levels for Extreme Precipitation and Discharge in Douala City.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe quantitative analysis presented in Table 3 reveals significant hydrological disparities between the extreme regimes of precipitation and discharge within the urban watershed of Douala. This systematic comparison of return periods (RPs), return levels (RLs), and their 95% confidence intervals provides a mathematical demonstration of the profound disturbance of natural streamflow dynamics.\u003c/p\u003e\n\u003cp\u003eTable 3 highlights a disproportionate hydrological attenuation, whereby for a 100-day return period, extreme precipitation intensity (98.08 mm) generates only an extreme discharge of 37.93 mm/day, representing a reduction of more than 60%. This nonlinear decline between precipitation intensity and hydrological response, as quantified in Table 3, corroborates the findings of Ngoran and Xue (2015) regarding the accelerated degradation of hydrological functions in urbanized catchments of Douala.\u003c/p\u003e\n\u003cp\u003eTable 3 also shows a widening divergence in the widths of confidence intervals between precipitation and discharge. For the 1095-day return period by example, the confidence interval for discharge [88.24 \u0026ndash; 186.51] spans a relative amplitude of 111%, compared to only 36% for precipitation [173.68 \u0026ndash; 251.08]. This growing differential uncertainty, systematically documented in Table 3, reflects the complexity introduced by anthropogenic factors in hydrological response, confirming the conclusions of Uddin and Salah (2018) on the spatial heterogeneity of urban runoff processes in Douala.\u003c/p\u003e\n\u003cp\u003eA closer examination of Table 3 demonstrates that the confidence intervals of discharge return periods are systematically wider and more asymmetric than those of precipitation. For instance, at a return level of 70.93 mm/day (discharge), the return period ranges from 267 to 332 days, while at a return level of 143.91 mm (precipitation), the return period ranges from 239 to 359 days. This structural asymmetry in uncertainties, as quantified in Table 3, illustrates the fundamental disruption of hydrological transfer processes, consistent with the findings of Adamu et al. (2025) on chronic obstructions within the Douala hydrographic network.\u003c/p\u003e\n\u003cp\u003eThe systematic comparison provided in Table 3 between extreme precipitation and discharge regimes offers a critical quantitative foundation for recalibrating flood management strategies. The significant discrepancies between confidence intervals highlight the need, as emphasized by Raouf et al. (2025), to develop differentiated approaches in the design of hydraulic infrastructure that account for the increased variability introduced by anthropogenic modifications of the watershed.\u003c/p\u003e\n\u003cp\u003eTable 3 thus serves as quantitative evidence of the advanced degradation of streamflow dynamics in Douala, providing essential data to guide hydrological restoration policies and climate adaptation strategies in this rapidly expanding coastal metropolis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;Table 3: Comparative Return Periods and Return Levels of Extreme Precipitation and Extreme Discharge in the Douala Watershed, with 95% Confidence Intervals\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"718\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003eRPs (day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\u003cstrong\u003eRLs of the\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eextreme Discharges\u003c/strong\u003e\u003cbr\u003e(\u003cstrong\u003emm/day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e\u003cstrong\u003e95% CIs for RLs\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;of the extreme\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eDischarges (mm/day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\u003cstrong\u003e95% CIs for \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eRPs (day)\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eof the extreme\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eRainfall\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\u003cstrong\u003eRLs of the\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eextreme Rainfalls\u0026nbsp;\u003c/strong\u003e(\u003cstrong\u003emm)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\u003cstrong\u003e95% CIs for RLs\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;of the extreme\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;Rainfalls (mm)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e\u003cstrong\u003e95% CIs for \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eRPs (day)\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eof the extreme Rainfalls\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e50\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e24.03\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[20.94, 27.11]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[50,50]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e73.87\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[68.753,79.00]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[46,53]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp; 70\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e30.24\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[26.06, 34.43]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[69,71]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e85.21\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[78.81 91.60]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[64,75]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp;100\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e37.93\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[32.28,43.57]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[97,103]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e98.08\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[89.90,106.25]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[89,109]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp;140\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e46.38\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[38.92, 53.85]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[132,147]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e111.07\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[100.71,121.43]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[121,157]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp;210\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e58.41\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[47.76, 69.05]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[193,227]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e127.93\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[114.19,141.67]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[174,244]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp;280\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e68.35\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[54.48, 82.21]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[250,309]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e140.73\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[124.03,157.43]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[225,334]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp;300\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e70.93\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[56.14, 85.72]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[267,332]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e143.91\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[126.42,161.40]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[239,359]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp; 365\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e78.71\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[60.92, 96.50]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[319,411]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e153.19\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[133.30,173.08]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[285,444]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e730.5\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e112.37\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[78.27, 146.47]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[592,868]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e189.07\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[158.45,219.68]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[516,943]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp;1095\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e137.38\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[88.24, 186.51]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[844,1345]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e212.38\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[173.68,251.08]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[726,1464]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp;1421\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e156.02\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[94.34, 217.7]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[1058,1784]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e228.41\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[183.7,273.12]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[899,1943]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\u003cstrong\u003e\u0026nbsp;1825\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e2000\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e2500\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e3000\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e3500\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e4000\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e4500\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e176.08\u003cbr\u003e183.99\u003cbr\u003e204.69\u003cbr\u003e223.18\u003cbr\u003e240.04\u003cbr\u003e255.60\u003cbr\u003e270.11\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 132px;\"\u003e[99.80, 252.36]\u003cbr\u003e[101.67,266.32]\u003cbr\u003e[105.83,303.54]\u003cbr\u003e[108.75,337.62]\u003cbr\u003e[110.80,369.27]\u003cbr\u003e[112.22,398.98]\u003cbr\u003e[113.14,427.08]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e[1311,2339]\u003cbr\u003e[1418,2583]\u003cbr\u003e[1712,3289]\u003cbr\u003e[1995,4007]\u003cbr\u003e[2268,4735]\u003cbr\u003e[2532,5471]\u003cbr\u003e[2789,6215]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e244.59\u003cbr\u003e250.72\u003cbr\u003e266.12\u003cbr\u003e279.21\u003cbr\u003e290.66\u003cbr\u003e300.86\u003cbr\u003e310.08\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e[193.45,295.74]\u003cbr\u003e[197.05,304.39]\u003cbr\u003e[205.89,326.34]\u003cbr\u003e[213.18,345.24]\u003cbr\u003e[219.38,361.93]\u003cbr\u003e[224.79,376.92]\u003cbr\u003e[229.58,229.58]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 16px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 98px;\"\u003e[1101,2550]\u003cbr\u003e[1183,2817]\u003cbr\u003e[1412,3590]\u003cbr\u003e[1626,4376]\u003cbr\u003e[1829,5175]\u003cbr\u003e[2023,5980]\u003cbr\u003e[2210,6791]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThe comparative analysis of discharge and rainfall extremes (Tables 4 and 5) provides critical insights into the hydrological functioning of the urban watershed of Douala. The results highlight not only the amplification of hydrological hazards but also the temporal and structural disconnections between precipitation inputs and river discharge responses.\u003c/p\u003e\n\u003cp\u003eDischarge extremes display strong interannual variability, with the flood of May 28, 2013 (174 mm/day), representing the most exceptional event, characterized by a return period of 1781 days and a wide confidence interval [1284\u0026ndash;2278 days]. This confirms the complexity of flood generation processes in the Wouri basin, as also noted by Besack et al. (2025). The breadth of these intervals indicates substantial predictive uncertainty for rare events, which complicates flood hazard assessment.\u003c/p\u003e\n\u003cp\u003eA systematic lag was observed between extreme discharges and subsequent precipitation return dates, ranging from 2 to 37 months. For example, the flood of September 2, 2014 (145 mm/day), corresponds to a rainfall return date projected for January 6, 2018, reflecting a delay of 40 months. Such temporal displacement suggests the presence of hydrological memory within the catchment, a phenomenon consistent with the findings of (Boum-Nkot et al. 2024; Tume et al. 2025; Yonzoua et al. 2021).\u003c/p\u003e\n\u003cp\u003eRainfall extremes exhibit a clear upward trend between 2010 and 2023, with annual maxima increasing from 127 mm in 2010 to 248 mm in 2018\u0026mdash;an intensification of 95% within less than a decade. This corroborates earlier observations by Molua and Lambi (2006); Grijsen (2014); Nonki et al. (2019), who reported an increase in the frequency and intensity of hydro-meteorological extremes across Cameroon under changing climatic conditions.\u003c/p\u003e\n\u003cp\u003eThe joint interpretation of Tables 4 and 5 reveals a marked asymmetry between rainfall and discharge extremes. Whereas rainfall extremes follow a relatively regular progression, discharge extremes display accentuated interannual fluctuations. This divergence reflects the growing influence of anthropogenic factors, such as soil sealing and altered drainage, on watershed hydrology (Cheng, Li, and Liu 2020; Anyangwa 2025).\u003c/p\u003e\n\u003cp\u003eThe efficiency of rainfall-to-discharge conversion is notably low. For instance, in 2015, a rainfall maximum of 227 mm produced only 57.7 mm of discharge, corresponding to a transfer efficiency of 25%. This attenuation can be explained by urban hydrological losses, including infiltration, evaporation, and depression storage, as described by Din et al. (2017).\u003c/p\u003e\n\u003cp\u003eThe wide confidence intervals associated with return dates delineate windows of heightened vulnerability. For example, the 2013 flood defines a critical vigilance window between December 2016 and August 2019, necessitating long-term preparedness. The variability of return periods further underscores the need for multi-scale early warning systems capable of managing both frequent (100\u0026ndash;200 days) and exceptional (\u0026gt;1000 days) events, in line with the recommendations of MINEE (http://www.minee.cm).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;The findings demonstrate the necessity of adaptive hydrological models that account for land-use change, drainage infrastructure performance, and climatic variability. Moreover, the complementarity between rainfall and discharge extremes justifies the deployment of integrated monitoring systems that measure both variables simultaneously, thereby allowing continuous model calibration and improved risk forecasting.\u003c/p\u003e\n\u003cp\u003eTogether, Tables 4 and 5 provide quantitative evidence that Douala\u0026rsquo;s hydrological risks are not solely determined by extreme rainfall events but are amplified and reshaped by watershed transformations and rapid urbanization. The asymmetry in rainfall\u0026ndash;discharge responses, the low transfer efficiency, and the growing uncertainties emphasize the urgent need for structural (drainage rehabilitation, green infrastructure) and non-structural (monitoring, early warning, planning) interventions to reduce flood vulnerability in the city.\u003c/p\u003e\n\u003cp\u003eTable 4: \u0026nbsp;Number of return days estimated for different maximum discharge (mm/day) over each year, as well as estimated return dates of maximum rainfall and 95% CIs using the adjustment of daily discharge totals from 2010-23 in Douala by the \u003cem\u003eGPD model.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"100%\" style=\"margin-right: calc(75%); width: 25%;\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eDates of extreme discharge in each year.\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eMaximum discharge (mm/day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003eEstimated\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003enumber\u0026nbsp;of return days\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e95% CIs of number of return days\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e95% CIs of RLs (mm/day)\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003eEstimated dates of return of the next maximum rainfall.\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e95% CIs for RD.\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e15/07/2010\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e37.6\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e99\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[96, 102]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[32.1, 43.29]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e22/10/2010\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[15/07/2010, 25/10/2010]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e31/08/2011\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e102\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e603\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[500, 706]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[73.45, 130.56]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e25/04/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e12/01/2013\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e06/08/2013\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e29/09/2012\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e44.7\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e131\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[124, 138]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[37.62, 51.78]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e07/02/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e31/01/2013\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e14/02/2013\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e28/05/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e174\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e1781\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[1284,2278]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[99.29, 248.75]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e13/04/2018\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e02/12/2016\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e23/08/2019\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e02/09/2014\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e145\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e1222\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[930, 1516]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[90.85,199.08]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e06/01/2018\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e20/03/2017\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e27/10/2018\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e27/08/2015\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e57.7\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e205\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[189,222]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[47.26, 68.14]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e19/03/2016\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e03/03/2016\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e05/04/2016\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e24/07/2016\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e94.5\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e519\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[438, 601]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[69.68, 119.30]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e25/12/2017\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e05/10/2017\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e17/03/2018\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e16/06/2017\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e131\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e995\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[777, 1212]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[85.92, 176.17]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e07/03/2020\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e02/08/2019\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e10/10/2020\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e26/07/2018\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e41.1\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e114\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[109,119]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[34.78,47.36]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e17/11/2018\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e12/11/2018\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e22/11/2018\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e16/05/2019\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e122\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e861\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[685, 1038]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[82.37, 161.61]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e23/09/2021\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e31/03/2021\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e19/03/2022\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e21/06/2021\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e13/08/2022\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e08/07/2023\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e118\u003cbr\u003e80.6\u003cbr\u003e40.8\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e806\u003cbr\u003e382\u003cbr\u003e113\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[646, 965]\u003cbr\u003e[332,431]\u003cbr\u003e[108, 117]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[80.73,155.35]\u0026nbsp;\u003cbr\u003e[62.04, 99.19]\u003cbr\u003e[34.61, 47.1]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4839%;\"\u003e\u003cstrong\u003e05/09/2023\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e30/08/2023\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e29/10/2023\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 19.4838%;\"\u003e[\u003cstrong\u003e29/03/2023\u003c/strong\u003e, \u003cstrong\u003e11/02/2024\u003c/strong\u003e][\u003cstrong\u003e11/07/2023\u003c/strong\u003e, \u003cstrong\u003e18/10/2023\u003c/strong\u003e][\u003cstrong\u003e24/10/2023\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e02/11/2023\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 5: Number of return days estimated for different maximum precipitation (mm) over each year, as well as estimated return dates of maximum rainfall and 95% CIs \u0026nbsp; using the adjustment of daily rainfall totals from 2010-23 in Douala by the \u003cem\u003eGPD model.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003eDates of extreme rainfall in each year.\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003eMaximum rainfall (mm)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003eEstimated\u0026nbsp;number\u0026nbsp;of return days\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e95% CIs of number of return days\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e95% CIs of RLs (mm)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003eEstimated dates of return of the next maximum rainfall.\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e95% CIs for RD.\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e23/08/2010\u0026nbsp;\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e127\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e206\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[171, 239]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[113.54, 140.66]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e17/03/2011\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e10/02/2011\u003c/strong\u003e, 19/04/2011]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e18/08/2011\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e170\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e511\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[381, 641]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[145.34, 194.61]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e10/01/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e02/09/2012\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e20/05/2013\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e07/07/2012\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e154\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e371\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[288, 453]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[133.88, 174.08]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e13/07/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e21/04/2013\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e03/10/2013\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e03/07/2013\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e146\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e290\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[208, 362]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[118.24, 146.30]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e19/04/2014\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e27/01/2014\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e30/06/2014\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e25/08/2014\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e193\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e314\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[249, 378]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[128.01, 164.06]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e05/07/2015\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e01/05/2015\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e07/09/2015\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e27/08/2015\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e227\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e1390\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[883, 1897]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[182.84, 271.20]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e17/06/2019\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e26/01/2018\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e05/11/2020\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e21/08/2016\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e229\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e1435\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[906, 1963]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[184.07, 273.98]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e26/07/2020\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e13/02/2019\u003c/strong\u003e, 05/01/2022]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e30/08/2017\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e215\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e1144\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[752, 1535]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[175.36, 254.68]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e17/10/2020\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e21/09/2019\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e12/11/2021\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e26/07/2018\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e248\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e1921\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[1146, 2697]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[195.46,300.56]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e29/10/2023\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e14/09/2021\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e13/12/2025\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e08/08/2019\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e198\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e856\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[591, 1121]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[164.37, 231.58]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e11/12/2021\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e21/03/2021\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e02/09/2022\u003c/strong\u003e]\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e20/08/2020\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e11/08/2021\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e26/08/2022\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;\u003cstrong\u003e08/07/2023\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e178\u003cbr\u003e125\u003cbr\u003e104\u003cbr\u003e\u0026nbsp;144\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e595\u003cbr\u003e196\u003cbr\u003e\u0026nbsp;117\u003cbr\u003e\u0026nbsp;300\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e[434, 756]\u003cbr\u003e[164, 228]\u003cbr\u003e\u0026nbsp;[103,131]\u003cbr\u003e\u0026nbsp;[240, 361]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e[150.88, 205]\u003cbr\u003e[111.86,138.06]\u003cbr\u003e\u0026nbsp;[94.90, 113.16]\u003cbr\u003e\u0026nbsp;[126.42, 161.40]\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\u003cstrong\u003e07/04/2022\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e23/02/2022\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e21/12/2022\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e29/11/2023\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 198px;\"\u003e[\u003cstrong\u003e28/10/2021\u003c/strong\u003e,\u0026nbsp;\u003cstrong\u003e15/09/2022\u003c/strong\u003e]\u003cbr\u003e\u003cstrong\u003e[22/01/2022,\u003c/strong\u003e \u003cstrong\u003e27/03/2022]\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e[07/12/2022, 04/01/2023]\u003c/strong\u003e\u003cbr\u003e\u003cstrong\u003e[04/03/2024, 03/07/2024]\u003c/strong\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"4. Conclusion","content":"\u003cp\u003eThe objective of this study was to evaluate the hydrological dynamics of the Douala watershed by determining the Return Periods (RPs), Return Levels (RLs), and Return Dates (RDs) \u0026mdash; together with their respective Confidence Intervals (CIs) \u0026mdash; for extreme rainfall and discharge events. The analysis was based on daily rainfall and discharge data covering the period 2010\u0026ndash;2023. The univariate and stationary Peak-Over-Threshold (POT) approach was applied using the Generalized Pareto Distribution (GPD), fitted through the Maximum Likelihood Estimation (MLE) method. Thresholds were selected through three complementary criteria: the Mean Residual Life (MRL) plot, the Model-Based Check (MBC), and the automatic threshold selection algorithm, ensuring the robustness of the GPD fitting process.\u003c/p\u003e\u003cp\u003eThis study also applied the method developed by Padji et al. (\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) for the simultaneous estimation of RPs, RLs, and RDs of extreme events, including the determination of their Confidence Intervals using the Delta Method. This innovative approach provides a dynamic framework capable of linking statistical extreme-value estimation with temporal prediction, thereby enhancing the practical utility of hydrological forecasting for flood early warning.\u003c/p\u003e\u003cp\u003eThe results confirm a marked interannual and seasonal variability in extreme events, with four main rainfall regimes identified: December\u0026ndash;February (DJF), March\u0026ndash;May (MAM), June\u0026ndash;September (JJAS), and October\u0026ndash;November (ON) (Padji et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The JJAS period remains the most critical, characterized by the highest rainfall intensities and the strongest discharge peaks, leading to recurrent and destructive flooding (Padji et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Historical flood events recorded in Douala are consistent with the RDs projected by the model, particularly during the major rainy season, validating both the GPD-based estimations and the methodological framework proposed by (Padji et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe estimated GPD shape parameters indicate a heavy-tailed Fr\u0026eacute;chet behavior for rainfall, suggesting the potential occurrence of very intense precipitation events, while discharge extremes exhibit a relatively lighter tail, reflecting the influence of basin retention, infiltration, and drainage capacity. The comparison of rainfall and discharge RPs and RLs reveals a systematic asymmetry: rainfall extremes are transformed into disproportionately variable discharge responses. This nonlinearity highlights the hydrological amplification effect in an urbanized basin, consistent with earlier findings by Tchuikoua (\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Ndjama et al. (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFurthermore, the computation of Confidence Intervals for both RPs and RDs provides critical quantitative uncertainty bounds for risk analysis. For instance, the flood event of 28 May 2013, associated with a discharge of 174 mm/day, corresponds to a modeled return period of approximately 1781 days, with a 95% confidence interval ranging from 1284 to 2278 days. These results illustrate the temporal variability and uncertainty inherent in the prediction of extreme floods, offering key insights for adaptive flood management and infrastructure planning.\u003c/p\u003e\u003cp\u003eThe comparative interpretation of Tables \u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e and \u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e reinforces these findings. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e shows pronounced interannual variability and uncertainty in discharge extremes, while Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e highlights an intensification of extreme rainfall magnitudes during the last decade. This asymmetry in recurrence behavior reveals that urbanization and land-use transformation have altered the rainfall\u0026ndash;runoff coupling, reducing infiltration capacity and increasing hydrological response speed. Consequently, extreme rainfall events now produce more erratic and amplified discharge responses, underscoring the need for integrated watershed management.\u003c/p\u003e\u003cp\u003eIn summary, this study demonstrates that the integration of rainfall and discharge extremes within the POT\u0026ndash;GPD framework, enhanced by the Padji et al. (\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) method, provides a coherent and operational approach for understanding and predicting hydrological risks in tropical urban basins. The methodology not only quantifies the magnitude and frequency of extreme events but also estimates their probable occurrence dates, offering valuable support for early warning systems and urban flood preparedness. This approach is transferable to other climatic parameters and catchments, provided high-resolution datasets are available. Future research should extend the analysis using sub-daily (hourly) data to refine the temporal prediction of extreme events and strengthen the city\u0026rsquo;s climate resilience framework.\u003c/p\u003e\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003ePadji Calvin \u0026ndash; Conceptualization, Methodology, Data Analysis, Writing \u0026ndash; Original Draft, Visualization.Calvin led the conception and design of the study, conducted the Extreme Value Theory analysis, interpreted the results, and prepared the manuscript draft.Mekaleuni Cyrille \u0026ndash; Data Curation, Software, Validation, Writing \u0026ndash; Review \u0026amp; Editing.Cyrille contributed to data acquisition, preprocessing of CHIRPS and ERA5-Land datasets, statistical validation, and assisted in revising and improving the manuscript.Monkam David \u0026ndash; Supervision, Resources, Project Administration, Writing \u0026ndash; Review \u0026amp; Editing.David supervised the research work, provided scientific guidance on hydrological interpretation, ensured methodological consistency, and contributed to the final review and editing of the manuscript.\u003c/p\u003e\u003cp\u003eClinical trial number: not applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAdamu, Baba, Mohammadou Buba, Balgah Nguh, Mboza Yengeh, and Amos Zeh. 2025. \u0026ldquo;Urban Drainage Systems and Implications on the Inhabitants and Physical Environment of Buea Urban Area, South West Region of Cameroon.\u0026rdquo; \u003cem\u003eJournal of Environmental and Geographical Studies\u003c/em\u003e 4 (April):61\u0026ndash;88. https://doi.org/10.58425/jegs.v4i1.340.\u003c/li\u003e\n\u003cli\u003eAdelekan, Ibidun O. 2010. \u0026ldquo;Vulnerability of Poor Urban Coastal Communities to Flooding in Lagos, Nigeria.\u0026rdquo; \u003cem\u003eEnvironment and Urbanization\u003c/em\u003e 22 (2): 433\u0026ndash;50.\u003c/li\u003e\n\u003cli\u003eAdzandeh, E A, D Alaigba, and C N Nkemasong. 2020. \u0026ldquo;Application of Geospatial Techniques and Logistic Regression Model for Urban Growth Analysis in Limbe, Cameroon.\u0026rdquo; \u003cem\u003eNiger. J. Environ. Sci. Technol.(NIJEST)\u003c/em\u003e 4:138\u0026ndash;51.\u003c/li\u003e\n\u003cli\u003eAkwa, Teh Exodus, and Simone Pierrette Nguimbous. 2021. \u0026ldquo;Investigation of Typhoid Fever and Their Associated Risk Factors in Children Attending \u0026lsquo;Deo Gratias\u0026rsquo; Hospital in Douala, Littoral, Cameroon.\u0026rdquo; \u003cem\u003eElectronic Journal of Medical and Educational Technologies\u003c/em\u003e 14 (2): em2107.\u003c/li\u003e\n\u003cli\u003eAlfieri, Lorenzo, Sagy Cohen, John Galantowicz, Guy J P Schumann, Mark A Trigg, Ervin Zsoter, Christel Prudhomme, Andrew Kruczkiewicz, Erin Coughlan de Perez, and Zachary Flamig. 2018. \u0026ldquo;A Global Network for Operational Flood Risk Reduction.\u0026rdquo; \u003cem\u003eEnvironmental Science \u0026amp; Policy\u003c/em\u003e 84:149\u0026ndash;58.\u003c/li\u003e\n\u003cli\u003eAnyangwa, Immanuel Ateh. 2025. \u0026ldquo;Effectiveness of Municipal Solid Waste Management Systems in Douala, Cameroon.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eArnell, Nigel W, and Simon N Gosling. 2016. \u0026ldquo;The Impacts of Climate Change on River Flood Risk at the Global Scale.\u0026rdquo; \u003cem\u003eClimatic Change\u003c/em\u003e 134 (3): 387\u0026ndash;401.\u003c/li\u003e\n\u003cli\u003eBaldassarre, Giuliano Di, Alberto Montanari, Harry Lins, Demetris Koutsoyiannis, Luigia Brandimarte, and G\u0026uuml;nter Bl\u0026ouml;schl. 2010. \u0026ldquo;Flood Fatalities in Africa: From Diagnosis to Mitigation.\u0026rdquo; \u003cem\u003eGeophysical Research Letters\u003c/em\u003e 37 (22).\u003c/li\u003e\n\u003cli\u003eBeirlant, Jan, and Yuri Goegebeur. 2004. \u0026ldquo;Local Polynomial Maximum Likelihood Estimation for Pareto-Type Distributions.\u0026rdquo; \u003cem\u003eJournal of Multivariate Analysis\u003c/em\u003e 89 (1): 97\u0026ndash;118.\u003c/li\u003e\n\u003cli\u003eBerghuijs, Wouter R, Shaun Harrigan, Peter Molnar, Louise J Slater, and James W Kirchner. 2019. \u0026ldquo;The Relative Importance of Different Flood‐generating Mechanisms across Europe.\u0026rdquo; \u003cem\u003eWater Resources Research\u003c/em\u003e 55 (6): 4582\u0026ndash;93.\u003c/li\u003e\n\u003cli\u003eBesack, Felix, Gordon Ajonina, Seth Ebonji, Minette Tomedi, Landry Mbang Essome, Josu\u0026eacute; Tchogom, Rapha\u0026euml;l Ongu\u0026eacute;n\u0026eacute;, and Olusegun Dada. 2025. \u0026ldquo;Change in Hydrological Signature of Tropical River-Estuaries During Extreme Flood Conditions: A Case Study of the Wouri-Nkam River Estuary (Cameroon) During the August 2020 Flood Event.\u0026rdquo; \u003cem\u003eEstuaries and Coasts\u003c/em\u003e 48 (November):1\u0026ndash;15. https://doi.org/10.1007/s12237-025-01599-w.\u003c/li\u003e\n\u003cli\u003eBeven, Keith J. 2012. \u003cem\u003eRainfall-Runoff Modelling: The Primer\u003c/em\u003e. John Wiley \u0026amp; Sons.\u003c/li\u003e\n\u003cli\u003eBl\u0026ouml;schl, G\u0026uuml;nter, Julia Hall, Juraj Parajka, Rui A P Perdig\u0026atilde;o, Bruno Merz, Berit Arheimer, Giuseppe T Aronica, Ardian Bilibashi, Ognjen Bonacci, and Marco Borga. 2017. \u0026ldquo;Changing Climate Shifts Timing of European Floods.\u0026rdquo; \u003cem\u003eScience\u003c/em\u003e 357 (6351): 588\u0026ndash;90.\u003c/li\u003e\n\u003cli\u003eBl\u0026ouml;schl, G\u0026uuml;nter, Julia Hall, Alberto Viglione, Rui A P Perdig\u0026atilde;o, Juraj Parajka, Bruno Merz, David Lun, Berit Arheimer, Giuseppe T Aronica, and Ardian Bilibashi. 2019. \u0026ldquo;Changing Climate Both Increases and Decreases European River Floods.\u0026rdquo; \u003cem\u003eNature\u003c/em\u003e 573 (7772): 108\u0026ndash;11.\u003c/li\u003e\n\u003cli\u003eBl\u0026ouml;schl, G\u0026uuml;nter, and Murugesu Sivapalan. 1995. \u0026ldquo;Scale Issues in Hydrological Modelling: A Review.\u0026rdquo; \u003cem\u003eHydrological Processes\u003c/em\u003e 9 (3‐4): 251\u0026ndash;90.\u003c/li\u003e\n\u003cli\u003eBoum-Nkot, Suzanne Ngo, Bertil Nlend, Frederic Huneau, Fabiola Kezebou, and Bernard Gwodog. 2024. \u0026ldquo;Groundwater Vulnerability Mapping of Hyper-Humid Urban Aquifers for an Improved Water Management in African Megacities, Opportunities and Limitations in the Case of Douala (Cameroon).\u0026rdquo; \u003cem\u003eArabian Journal of Geosciences\u003c/em\u003e 17 (3): 77.\u003c/li\u003e\n\u003cli\u003eBox, George, and G M Jenkins. 1976. \u0026ldquo;Analysis: Forecasting and Control.\u0026rdquo; \u003cem\u003eSan Francisco\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eCheng, Zhonghua, Lianshui Li, and Jun Liu. 2020. \u0026ldquo;The Impact of Foreign Direct Investment on Urban PM2. 5 Pollution in China.\u0026rdquo; \u003cem\u003eJournal of Environmental Management\u003c/em\u003e 265:110532.\u003c/li\u003e\n\u003cli\u003eChoulakian, Vartan, and Michael A Stephens. 2001. \u0026ldquo;Goodness-of-Fit Tests for the Generalized Pareto Distribution.\u0026rdquo; \u003cem\u003eTechnometrics\u003c/em\u003e 43 (4): 478\u0026ndash;84.\u003c/li\u003e\n\u003cli\u003eColes, Stuart, Joanna Bawa, Lesley Trenner, and Pat Dorazio. 2001. \u003cem\u003eAn Introduction to Statistical Modeling of Extreme Values\u003c/em\u003e. Vol. 208. Springer.\u003c/li\u003e\n\u003cli\u003eCRED \u0026amp; UNDRR. 2022. \u0026ldquo;2022 Disasters in Numbers. Centre for Research on the Epidemiology of Disasters \u0026amp; UN Office for Disaster Risk Reduction.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eDavison, Anthony C, and Richard L Smith. 1990. \u0026ldquo;Models for Exceedances over High Thresholds.\u0026rdquo; \u003cem\u003eJournal of the Royal Statistical Society Series B: Statistical Methodology\u003c/em\u003e 52 (3): 393\u0026ndash;425.\u003c/li\u003e\n\u003cli\u003eDin, Ndongo, Vanessa Maxemilie Ngo-Massou, Guillaume L\u0026eacute;opold Essom\u0026egrave;-Koum, Eugene Ndema-Nsombo, Ernest Kott\u0026egrave;-Mapoko, and Laurant Nyamsi-Moussian. 2017. \u0026ldquo;Impact of Urbanization on the Evolution of Mangrove Ecosystems in the Wouri River Estuary (Douala Cameroon).\u0026rdquo; In \u003cem\u003eCoastal Wetlands: Alteration and Remediation\u003c/em\u003e, 81\u0026ndash;131. Springer.\u003c/li\u003e\n\u003cli\u003eDingman, S Lawrence. 2015. \u003cem\u003ePhysical Hydrology\u003c/em\u003e. Waveland press.\u003c/li\u003e\n\u003cli\u003eDosio, Alessandro, Martin W Jury, Mansour Almazroui, Moetasim Ashfaq, Ismaila Diallo, Francois A Engelbrecht, Nana A B Klutse, Christopher Lennard, Izidine Pinto, and Mouhamadou B Sylla. 2021. \u0026ldquo;Projected Future Daily Characteristics of African Precipitation Based on Global (CMIP5, CMIP6) and Regional (CORDEX, CORDEX-CORE) Climate Models.\u0026rdquo; \u003cem\u003eClimate Dynamics\u003c/em\u003e 57 (11): 3135\u0026ndash;58.\u003c/li\u003e\n\u003cli\u003eDouglas, Alecia C, Juline E Mills, Mamadou Niang, Svetlana Stepchenkova, Sookeun Byun, Celestino Ruffini, Seul Ki Lee, Jihad Loutfi, Jung-Kook Lee, and Mikhail Atallah. 2008. \u0026ldquo;Internet Addiction: Meta-Synthesis of Qualitative Research for the Decade 1996\u0026ndash;2006.\u0026rdquo; \u003cem\u003eComputers in Human Behavior\u003c/em\u003e 24 (6): 3027\u0026ndash;44.\u003c/li\u003e\n\u003cli\u003eDouglas, Ian, Kurshid Alam, Maryanne Maghenda, Yasmin Mcdonnell, Louise McLean, and Jack Campbell. 2008. \u0026ldquo;Unjust Waters: Climate Change, Flooding and the Urban Poor in Africa.\u0026rdquo; \u003cem\u003eEnvironment and Urbanization\u003c/em\u003e 20 (1): 187\u0026ndash;205.\u003c/li\u003e\n\u003cli\u003eEffron, Bradley, and Robert J Tibshirani. 1993. \u0026ldquo;An Introduction to the Bootstrap.\u0026rdquo; Chapman \u0026amp; Hall, New York.\u003c/li\u003e\n\u003cli\u003eForster, P, T Storelvmo, and K Alterskj\u0026aelig;. 2021. \u0026ldquo;IPCC Sixth Assessment Report (AR6) Working Group 1: The Physical Science Basis, Chap. 7.\u0026rdquo; University Press, UK, https://www. ipcc. ch/report/ar6/wg1/downloads (last \u0026hellip;.\u003c/li\u003e\n\u003cli\u003eFunk, Chris, Pete Peterson, Martin Landsfeld, Diego Pedreros, James Verdin, Shraddhanand Shukla, Gregory Husak, James Rowland, Laura Harrison, and Andrew Hoell. 2015. \u0026ldquo;The Climate Hazards Infrared Precipitation with Stations\u0026mdash;a New Environmental Record for Monitoring Extremes.\u0026rdquo; \u003cem\u003eScientific Data\u003c/em\u003e 2 (1): 1\u0026ndash;21.\u003c/li\u003e\n\u003cli\u003eGrijsen, Johan. 2014. \u003cem\u003eUnderstanding the Impact of Climate Change on Hydropower: The Case of Cameroon\u003c/em\u003e. World Bank, Washington, DC.\u003c/li\u003e\n\u003cli\u003eGupta, Vijay K, and Ed Waymire. 1990. \u0026ldquo;Multiscaling Properties of Spatial Rainfall and River Flow Distributions.\u0026rdquo; \u003cem\u003eJournal of Geophysical Research: Atmospheres\u003c/em\u003e 95 (D3): 1999\u0026ndash;2009.\u003c/li\u003e\n\u003cli\u003eGupta, Vijay K, and Edward C Waymire. 1993. \u0026ldquo;A Statistical Analysis of Mesoscale Rainfall as a Random Cascade.\u0026rdquo; \u003cem\u003eJournal of Applied Meteorology and Climatology\u003c/em\u003e 32 (2): 251\u0026ndash;67.\u003c/li\u003e\n\u003cli\u003eHabitat, U N. 2018. \u0026ldquo;Tracking Progress Towards Inclusive, Safe, Resilient and Sustainable Cities and Human Settlements. SDG 11 Synthesis Report-High Level Political Forum 2018.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eHelsel, Dennis R, and Robert M Hirsch. 1993. \u003cem\u003eStatistical Methods in Water Resources\u003c/em\u003e. Vol. 49. Elsevier.\u003c/li\u003e\n\u003cli\u003eHersbach, Hans, Bill Bell, Paul Berrisford, Gionata Biavati, Dick Dee, Andr\u0026aacute;s Hor\u0026aacute;nyi, Julien Nicolas, Carole Peubey, Raluca Radu, and Iryna Rozum. 2019. \u0026ldquo;The ERA5 Global Atmospheric Reanalysis at ECMWF as a Comprehensive Dataset for Climate Data Homogenization, Climate Variability, Trends and Extremes.\u0026rdquo; In \u003cem\u003eGeophysical Research Abstracts\u003c/em\u003e. Vol. 21.\u003c/li\u003e\n\u003cli\u003eHersbach, Hans, Bill Bell, Paul Berrisford, Shoji Hirahara, Andr\u0026aacute;s Hor\u0026aacute;nyi, Joaqu\u0026iacute;n Mu\u0026ntilde;oz‐Sabater, Julien Nicolas, Carole Peubey, Raluca Radu, and Dinand Schepers. 2020. \u0026ldquo;The ERA5 Global Reanalysis.\u0026rdquo; \u003cem\u003eQuarterly Journal of the Royal Meteorological Society\u003c/em\u003e 146 (730): 1999\u0026ndash;2049.\u003c/li\u003e\n\u003cli\u003eIPCC. 2021. \u0026ldquo;Climate Change 2021: Impacts, Adaptation and Vulnerability. Contribution of Working Group II.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003e\u0026mdash;\u0026mdash;\u0026mdash;. 2022. \u0026ldquo;Sixth Assessment Report (AR6) \u0026ndash; Impacts, Adaptation and Vulnerability.\u0026rdquo; https://www.ipcc.ch/report/ar6/wg2/.\u003c/li\u003e\n\u003cli\u003eJia, Weiping, Jianping Weng, Dalong Zhu, Linong Ji, Juming Lu, Zhiguang Zhou, Dajin Zou, Lixin Guo, Qiuhe Ji, and Li Chen. 2019. \u0026ldquo;Standards of Medical Care for Type 2 Diabetes in China 2019.\u0026rdquo; \u003cem\u003eDiabetes/Metabolism Research and Reviews\u003c/em\u003e 35 (6): e3158.\u003c/li\u003e\n\u003cli\u003eKatz, Michael B. 2013. \u003cem\u003eThe Undeserving Poor: America\u0026rsquo;s Enduring Confrontation with Poverty: Fully Updated and Revised\u003c/em\u003e. Oxford University Press.\u003c/li\u003e\n\u003cli\u003eKatz, Richard W, Marc B Parlange, and Philippe Naveau. 2002. \u0026ldquo;Statistics of Extremes in Hydrology.\u0026rdquo; \u003cem\u003eAdvances in Water Resources\u003c/em\u003e 25 (8\u0026ndash;12): 1287\u0026ndash;1304.\u003c/li\u003e\n\u003cli\u003eKendall, Maurice George. 1948. \u0026ldquo;Rank Correlation Methods.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eKometa, Sunday Shende, and Mathias Ashu Tambe Ebot. 2012. \u0026ldquo;Watershed Degradation in the Bamendjin Area of the North West Region of Cameroon and Its Implication for Development.\u0026rdquo; \u003cem\u003eJournal of Sustainable Development\u003c/em\u003e 5 (9): 75.\u003c/li\u003e\n\u003cli\u003eKoutsoyiannis, Demetris. 2004. \u0026ldquo;Statistics of Extremes and Estimation of Extreme Rainfall: I. Theoretical Investigation/Statistiques de Valeurs Extr\u0026ecirc;mes et Estimation de Pr\u0026eacute;cipitations Extr\u0026ecirc;mes: I. Recherche Th\u0026eacute;orique.\u0026rdquo; \u003cem\u003eHydrological Sciences Journal\u003c/em\u003e 49 (4).\u003c/li\u003e\n\u003cli\u003eKundzewicz, Zbigniew W, Markus Stoffel, Ryszard J Kaczka, Bartłomiej Wyżga, Tadeusz Niedźwiedź, Iwona Pińskwar, Virginia Ruiz-Villanueva, Ewa Łupikasza, Barbara Czajka, and Juan Antonio Ballesteros-Canovas. 2014. \u0026ldquo;Floods at the Northern Foothills of the Tatra Mountains\u0026mdash;a Polish-Swiss Research Project.\u0026rdquo; \u003cem\u003eActa Geophysica\u003c/em\u003e 62 (3): 620\u0026ndash;41.\u003c/li\u003e\n\u003cli\u003eLegg, Stephen. 2021. \u0026ldquo;IPCC, 2021: Climate Change 2021-the Physical Science Basis.\u0026rdquo; \u003cem\u003eInteraction\u003c/em\u003e 49 (4): 44\u0026ndash;45.\u003c/li\u003e\n\u003cli\u003eLyon, Bradfield, and Nicolas Vigaud. 2017. \u0026ldquo;Unraveling East Africa\u0026rsquo;s Climate Paradox.\u0026rdquo; \u003cem\u003eClimate Extremes: Patterns and Mechanisms\u003c/em\u003e, 265\u0026ndash;81.\u003c/li\u003e\n\u003cli\u003eMiller, James D, and Michael Hutchins. 2017. \u0026ldquo;The Impacts of Urbanisation and Climate Change on Urban Flooding and Urban Water Quality: A Review of the Evidence Concerning the United Kingdom.\u0026rdquo; \u003cem\u003eJournal of Hydrology: Regional Studies\u003c/em\u003e 12:345\u0026ndash;62.\u003c/li\u003e\n\u003cli\u003eMolua, Ernest L, and Cornelius M Lambi. 2006. \u0026ldquo;Climate, Hydrology and Water Resources in Cameroon.\u0026rdquo; \u003cem\u003eCEEPA, Pretoria\u003c/em\u003e 37.\u003c/li\u003e\n\u003cli\u003eMu\u0026ntilde;oz-Sabater, Joaqu\u0026iacute;n, Emanuel Dutra, Anna Agust\u0026iacute;-Panareda, Cl\u0026eacute;ment Albergel, Gabriele Arduini, Gianpaolo Balsamo, Souhail Boussetta, Margarita Choulga, Shaun Harrigan, and Hans Hersbach. 2021. \u0026ldquo;ERA5-Land: A State-of-the-Art Global Reanalysis Dataset for Land Applications.\u0026rdquo; \u003cem\u003eEarth System Science Data\u003c/em\u003e 13 (9): 4349\u0026ndash;83.\u003c/li\u003e\n\u003cli\u003eNdjama, J, A A Ako, D Kuitcha, T A Margaret, W Fantong, G Mafany, E G Takem, M C Y Fopouo, and K B V Kamgang. 2014. \u0026ldquo;Ecohydrology of Surface and Underground Water of the Mgoua Drainage Basin in Douala (Cameroon).\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eNgoran, Suinyuy Derrick, and XiongZhi Xue. 2015. \u0026ldquo;Addressing Urban Sprawl in Douala, Cameroon: Lessons from Xiamen Integrated Coastal Management.\u0026rdquo; \u003cem\u003eJournal of Urban Management\u003c/em\u003e 4 (1): 53\u0026ndash;72.\u003c/li\u003e\n\u003cli\u003eNicholson, Sharon E. 2013. \u0026ldquo;The West African Sahel: A Review of Recent Studies on the Rainfall Regime and Its Interannual Variability.\u0026rdquo; \u003cem\u003eInternational Scholarly Research Notices\u003c/em\u003e 2013 (1): 453521.\u003c/li\u003e\n\u003cli\u003eNka, Bernadette N, Ludovic Oudin, H Karambiri, Jean-Emmanuel Paturel, and P Ribstein. 2015. \u0026ldquo;Trends in Floods in West Africa: Analysis Based on 11 Catchments in the Region.\u0026rdquo; \u003cem\u003eHydrology and Earth System Sciences\u003c/em\u003e 19 (11): 4707\u0026ndash;19.\u003c/li\u003e\n\u003cli\u003eNonki, Rodric M, Andr\u0026eacute; Lenouo, Christopher J Lennard, and Cl\u0026eacute;ment Tchawoua. 2019. \u0026ldquo;Assessing Climate Change Impacts on Water Resources in the Benue River Basin, Northern Cameroon.\u0026rdquo; \u003cem\u003eEnvironmental Earth Sciences\u003c/em\u003e 78 (20): 606.\u003c/li\u003e\n\u003cli\u003eNonki, Rodric M, Andr\u0026eacute; Lenouo, Raphael M Tshimanga, Francine C Donfack, and Cl\u0026eacute;ment Tchawoua. 2021. \u0026ldquo;Performance Assessment and Uncertainty Prediction of a Daily Time-Step HBV-Light Rainfall-Runoff Model for the Upper Benue River Basin, Northern Cameroon.\u0026rdquo; \u003cem\u003eJournal of Hydrology: Regional Studies\u003c/em\u003e 36:100849.\u003c/li\u003e\n\u003cli\u003eOgden, Fred L, Nawa Raj Pradhan, Charles W Downer, and Jon A Zahner. 2011. \u0026ldquo;Relative Importance of Impervious Area, Drainage Density, Width Function, and Subsurface Storm Drainage on Flood Runoff from an Urbanized Catchment.\u0026rdquo; \u003cem\u003eWater Resources Research\u003c/em\u003e 47 (12).\u003c/li\u003e\n\u003cli\u003eOlivier, Stephen L, and Jan F Prins. 2009. \u0026ldquo;Evaluating OpenMP 3.0 Run Time Systems on Unbalanced Task Graphs.\u0026rdquo; In \u003cem\u003eInternational Workshop on OpenMP\u003c/em\u003e, 63\u0026ndash;78. Springer.\u003c/li\u003e\n\u003cli\u003ePadji, Calvin, Cyrille Meukaleuni, Cyrille Mezoue Adiang, Daniel Bongue, and David Monkam. 2024. \u0026ldquo;Estimation of Return Dates and Return Levels of Extreme Rainfall in the City of Douala, Cameroon.\u0026rdquo; \u003cem\u003eHeliyon\u003c/em\u003e 10 (14): e34832.\u003c/li\u003e\n\u003cli\u003ePapalexiou, Simon Michael, and Demetris Koutsoyiannis. 2016. \u0026ldquo;A Global Survey on the Seasonal Variation of the Marginal Distribution of Daily Precipitation.\u0026rdquo; \u003cem\u003eAdvances in Water Resources\u003c/em\u003e 94:131\u0026ndash;45.\u003c/li\u003e\n\u003cli\u003eRaouf, Abdou, Kasi Njeudjang, Jean Ghislain Tabue Youmbi, Xiaohui Lu, Jianhua Yue, Eucharia Chidinma Okoro, Elijah Chibueze Odii, and Ali Moussa Abdoul Wahab. 2025. \u0026ldquo;Factors Influencing Groundwater Development and Mitigation Strategies in Adamawa Region, Cameroon: A Critical Review.\u0026rdquo; \u003cem\u003eSustainable Water Resources Management\u003c/em\u003e 11 (1): 5.\u003c/li\u003e\n\u003cli\u003eSeneviratne, Sonia, Neville Nicholls, David Easterling, Clare Goodess, Shinjiro Kanae, James Kossin, Yali Luo, Jose Marengo, Kathleen McInnes, and Mohammad Rahimi. 2012. \u0026ldquo;Changes in Climate Extremes and Their Impacts on the Natural Physical Environment.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eSivapalan, Murugesu. 2003. \u0026ldquo;Prediction in Ungauged Basins: A Grand Challenge for Theoretical Hydrology.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eSmith, Keith, and Roy Ward. 1998. \u003cem\u003eFloods: Physical Processes and Human Impacts.\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eSylla, Mouhamadou Bamba, Pinghouinde Michel Nikiema, Peter Gibba, Ibourahima Kebe, and Nana Ama Browne Klutse. 2016. \u0026ldquo;Climate Change over West Africa: Recent Trends and Future Projections.\u0026rdquo; \u003cem\u003eAdaptation to Climate Change and Variability in Rural West Africa\u003c/em\u003e, 25\u0026ndash;40.\u003c/li\u003e\n\u003cli\u003eTchuikoua, Louis. 2020. \u0026ldquo;Lecture de La Dynamique Territoriale \u0026agrave; Douala et Son Impact Sur La Gestion Des D\u0026eacute;chets Solides M\u0026eacute;nagers Par La Th\u0026eacute;orie de La Formation Socio-Spatiale.\u0026rdquo; \u003cem\u003eEnvironnement, Ing\u0026eacute;nierie \u0026amp; D\u0026eacute;veloppement\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eTsalefac, Maurice, Fran\u0026ccedil;ois Hiol Hiol, Gil Mah\u0026eacute;, Alain Laraque, Denis Jean Sonwa, Paul Scholte, Wilfried Pokam, Andreas Haensler, Tazebe Beyene, and Fulco Ludwig. 2015. \u0026ldquo;Climate of Central Africa: Past, Present and Future.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eTschakert, Petra, Regina Sagoe, Gifty Ofori-Darko, and Samuel Nii Codjoe. 2010. \u0026ldquo;Floods in the Sahel: An Analysis of Anomalies, Memory, and Anticipatory Learning.\u0026rdquo; \u003cem\u003eClimatic Change\u003c/em\u003e 103 (3): 471\u0026ndash;502.\u003c/li\u003e\n\u003cli\u003eTume, Suiven John Paul, Yongka Bernard Nyuyse, Clarius Kuyar, Ngoran Larissa Tomla, Ngoran Hassan Afoni, Foka Sherifatu Kinyuy, and Ngala Elvis Ngum. 2025. \u0026ldquo;Vulnerability of Urban Housing to Climate Change in Cameroon.\u0026rdquo; In \u003cem\u003eDesigning Healthy Buildings and Communities: Shaping a Climate-Resilient Future\u003c/em\u003e, 49\u0026ndash;84. Springer.\u003c/li\u003e\n\u003cli\u003eUddin, Mohammad Kashif, and Mukhtar M Salah. 2018. \u0026ldquo;Statistical Analysis of Litchi Chinensis\u0026rsquo;s Adsorption Behavior toward Cr (VI).\u0026rdquo; \u003cem\u003eApplied Water Science\u003c/em\u003e 8 (5): 140.\u003c/li\u003e\n\u003cli\u003eUNDRR. 2020. \u0026ldquo;Words into Action Guidelines: Developing Early Warning Systems. United Nations Office for Disaster Risk Reduction.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003eVillarini, Gabriele, and James A Smith. 2010. \u0026ldquo;Flood Peak Distributions for the Eastern United States.\u0026rdquo; \u003cem\u003eWater Resources Research\u003c/em\u003e 46 (6).\u003c/li\u003e\n\u003cli\u003eYonzoua, C C Fandjio, U J M Pettang Nana, M B Manjiaa, and C Pettanga. 2021. \u0026ldquo;Mesure de La Vuln\u0026eacute;rabilit\u0026eacute; Des Milieux Urbains Au Cameroun Face Au Changement Climatique: Measuring the Vulnerability of Urban Areas in Cameroon to Climate Change.\u0026rdquo; \u003cem\u003eJournal of the Cameroon Academy of Sciences\u003c/em\u003e 17 (2): 115\u0026ndash;29.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Extreme Rainfall, Flood Discharge, Return Period, Temporal Lag, Urban Hydrology, EVT, Generalized Pareto Distribution","lastPublishedDoi":"10.21203/rs.3.rs-8073038/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8073038/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study investigates the critical temporal disconnect between extreme rainfall and subsequent flood discharge in Douala, Cameroon, a rapidly urbanizing coastal city increasingly vulnerable to flooding. Using a 13-year dataset (2010\u0026ndash;2023) that integrates CHIRPS precipitation data Funk et al. (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) and ERA5-Land discharge data (Mu\u0026ntilde;oz-Sabater et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), we applied Extreme Value Theory (EVT) through a Peak-Over-Threshold (POT) framework to quantify and compare return levels and return periods for both meteorological and hydrological extremes. Our analysis reveals a marked hydrological transformation: while extreme rainfall exhibits heavy-tailed behavior (shape parameter ξ\u0026thinsp;=\u0026thinsp;0.19), discharge extremes show even heavier tails (ξ\u0026thinsp;=\u0026thinsp;0.44), suggesting that urban watershed processes amplify flood risks beyond what rainfall statistics alone would predict. Cross-correlation and event-based analyses indicate a systematic temporal lag of 1\u0026ndash;2 days between precipitation peaks and discharge responses, with weak coupling (r\u0026thinsp;\u0026asymp;\u0026thinsp;0.2\u0026ndash;0.4) between daily rainfall and runoff intensity. Moreover, a 100-days return period rainfall event (98.08 mm) generates only a 37.93 mm\u0026middot;day⁻\u0026sup1; discharge peak, reflecting over 60% attenuation between precipitation input and hydrological output. Diagnostic plots, threshold stability checks, and goodness-of-fit tests confirmed the robustness of the Generalized Pareto Distribution (GPD) models. The inclusion of return date analysis, following the methodology proposed by Padji et al. (\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), demonstrated the approach\u0026rsquo;s predictive capability for estimating the timing of flood events rather than their magnitude alone. These findings reveal that Douala\u0026rsquo;s flood hazard is not solely determined by rainfall extremes but is amplified by rapid urbanization, reduced infiltration, and altered drainage dynamics. The combination of weak rainfall\u0026ndash;discharge coupling, temporal lag, and hydrological attenuation underscores the urgent need for integrated flood forecasting and water management strategies that explicitly incorporate urban hydrological dynamics rather than relying exclusively on rainfall forecasts.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e","manuscriptTitle":"Analyzing the Hydrological Disconnect Between Extreme Precipitation and River Discharge in Douala, Cameroon","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-05 13:04:13","doi":"10.21203/rs.3.rs-8073038/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8d8b6df1-0611-4ffc-ac42-5152da51fef5","owner":[],"postedDate":"December 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-27T00:23:26+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-05 13:04:13","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8073038","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8073038","identity":"rs-8073038","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}