Flood Risk Research Articles

Enhancing transparency in data-driven urban pluvial flood prediction using an explainable CNN model

Mitigating severe losses caused by pluvial floods in urban areas with dense population and property requires effective flood prediction for emergency measures. Physics-based models face issues with low computational efficiency for real‐time flood prediction. An alternative approach for rapid prediction instead of physics-based models is to predict from a data-driven perspective. However, data-driven approaches for urban flood prediction are often perceived as “black box” and might raise concerns. In this study, we propose an explainable deep learning (DL) approach for rapid urban pluvial flood prediction with enhanced transparency using a convolutional neural network (CNN) and the explainable artificial intelligence (AI) framework Shapley additive explanation (SHAP). We process a systematic stepwise feature selection process and establish a CNN model considering topography, drainage networks and rainfall to predict maximum inundation depths. Then, SHAP is applied to provide trustworthy explanations for the decision making in model results. The results show that: 1) Forward selection can offer insights into selecting effective input variables for improved predictions and promote understanding of DL modelling. The spatial pattern of inundation depths predicted by the proposed CNN model shows good agreement with those predicted by the physics-based model, demonstrated by average correlation coefficient (CC) and mean absolute error (MAE) values of 0.982 and 0.021 m, respectively. 2) The CNN model substantially outperforms the physics-based model in computational speed when using the same hardware, achieving speedups of 210 times on GPU and 51 times on CPU in the case study (575167 grid cells, 14.38 km2). Particularly, it can still achieve good performance on a CPU-only standard laptop without high-performance hardware, with only a modest increase in computational time. 3) The SHAP explainable analysis confirms that the CNN model correctly captures the relationships between input variables and water depth, revealing a reasonable decision-making process, enhancing its transparency. The explainable DL approach incorporating SHAP for rapid urban pluvial flood prediction is promising to build trust among stakeholders and provide a trustworthy reference for prompt measures aiming at saving lives and protecting assets during flood emergencies. Additionally, the proposed DL approach can potentially be further expanded to analyze the causes of urban flooding events and serve as a foundation for exploring the transferability of data-driven urban flood prediction, providing benefits for better urban flood risk management.

Integrating Blue-Green Infrastructure strategies to enhance climate resilience in Colombia.

Colombia is a tropical developing country that faces frequent and unpredictable flood events. Previous efforts to mitigate the impacts of flood have revolved around traditional approaches like the construction of levees. However, as seen in other parts of the world, levees often make flooding worse. Considering the increasing frequency and severity of extreme weather associated with a changing climate, it is clear there is a pressing need for more effective strategies to mitigate the impacts of extreme weather events. However, poor data availability makes understanding flood risk and developing new approaches a difficult task. This research examines the feasibility of Blue-Green Infrastructure (BGI) as a nature-based methodological approach to deal with flooding focuses, focusing on the Guarapas River and Pitalito town, in regional Colombia. Utilizing PCSWMM as a modelling tool, the study demonstrates the applicability of BGI in reducing peak flow. Validating the model proved difficult given the substantial gaps in official data sources. So, it was necessary to use proxies such as social media posts of flood events, to determine the date and severity of flood events in the town. The analysis revealed that deploying a limited number of BGI elements can mitigate the adverse effects of floods in a cost-effective manner, and that the underpinning modelling could be carried out using data proxies such as local news reports and social media posts. This research contributes to the growing body of knowledge on BGI as a valuable solution for flood management in a changing climate but does so in a developing country context. By showcasing the feasibility of BGI in reducing peak flow and enhancing resilience in a region with limited data resources, this study shows that it is possible for developing countries to move away from traditional and sometimes increasingly ineffective solution. Barriers to adoption are still significant. But nature-based solutions to flooding represent a new, proactive measure in addressing the challenges posed by climate change, with implications for policymakers and stakeholders involved in sustainable planning. Keywords: Blue-Green Infrastructure, Climate change, Sustainable development, Developing countries, Water management.

Urban green streets and water management safety hubs in public schools lots

Abstract Green Stormwater Infrastructure (GSI) represents an effective approach to enhance urban resilience to climate change, contributing to runoff regulation and localized flood risk reduction. This study examines the potential for transforming open spaces in public schools into safety hubs for stormwater management in Mediterranean urban contexts, drawing on best practices from North America. Through a methodological analysis combining Research Through Design and an international literary scoping review, case studies of schools in the Metropolitan City of Naples are presented and compared with U.S. projects implementing GSI in school areas. The results highlight the importance of strategic coordination guidelines for local administrators, designers, and school principals, aimed at planning, designing, and managing highly performant green infrastructure networks. An integrated and shared metropolitan-level approach can maximize environmental benefits, transforming schools into community centers for urban resilience education and participatory management of risks related to extreme weather events.

Time-guided convolutional neural networks for spatiotemporal urban flood modelling

Urban flood modelling is key to understand flood risks and develop effective interventions in flood management. Deep learning (DL), known for its robust and automatic feature extraction capabilities, has been applied for urban flood predictions. However, the hybrid spatiotemporal structure of conventional DL-enabled urban flood models is limited in terms of accuracy and efficiency. To address this gap, this study develops a new DL model guided by time information. This model uses a classic CNN (Convolution Neural Network) architecture, Unet, as its backbone. Time information is integrated into inputs via an extra channel to specify the desired prediction time, facilitating the simulation of the spatiotemporal flood process. Additionally, a modified loss function is formulated to tackle the sample imbalance problem between flooded and non-flooded sites. The model performance is assessed in an urban area in Dalian, China with a total of 18 rainfall events of varying return periods. The model attains average precision and recall values of 0.90 and 0.81, respectively, across different time steps during various events. Furthermore, the model exhibits transferability in ungauged regions where a high influence of surrounding environments on local flood processes is identified by Grad-CAM (Gradient-weighted Class Activation Mapping) analysis. The results show that the new Unet model has great promise in efficiently providing accurate spatiotemporal flood simulations. The time-guided Unet model can serve as practical tools for rapid flood simulation in urban areas.

Integrated hydrological modelling and streamflow characterization of Gangotri Glacier meltwater

Runoff from glaciated catchments is an integrated process that includes glacier melt, snowmelt, rainfall and surface and subsurface runoff of meltwater from glacierized and non-glacierized areas. Monitoring and quantifying the contribution of the hydrologic components (snow, ice and rain) to river discharge in the Himalayan basins is essential for decision-making in the water sector, particularly in water resources management and flood risk reduction in the region. An attempt has been made to characterize and hydrologically model streamflow (Bhagirathi River) for the Gangotri Glacier (Central Himalaya, India). A semi-distributed conceptual hydrological model is used for the streamflow modelling and assessing the major streamflow components (snow melt, glacier melt and rainfall runoff). Initially, the model was calibrated using the available in situ hydro-meteorological records for the ablation seasons of 2013–14 to 2015–16 (3 years), and further validated for the ablation seasons of 2016–17 to 2018–19 (3 years). The model performed well for all the studied years except for some months, where abrupt changes in the contrasting weather parameters (precipitation and temperature) were recorded. In the Gangotri Glacier Valley (upper Bhagirathi River catchment), snowmelt contributed the largest portion (55.5%) to total streamflow followed by glacier melt (29.7%) and rainfall runoff components (14.7%).

Projecting urban flood risk through hydrodynamic modeling under shared socioeconomic pathways

Under future climate change, accurate risk assessment of urban flooding disasters is paramount for effective adaptation and mitigation strategies. However, conventional indicator-based assessment methods often fall short of accurately capturing the complexity of flooding dynamics. Current research predominantly focuses on predicting future hazard shifts while overlooking changes in other critical indicators. In this study, we establish a comprehensive index system for risk assessment, and quantified future changes in most indicators, utilizing the InfoWorks ICM model for hazard simulation and the CLUMondo model for land use predictions. Based on risk assessment results and regional characteristics, we further analyze the key factors driving future risk and discuss corresponding measures. The results indicate an exacerbation of future urban flood risk, with an 18% increase in high risk areas, primarily concentrated in the center of the study area. The dominant indicators are inundation depth and land use over the whole study area. However microtopography significantly affects risk in low-lying areas. Overall, under higher emission scenarios, the influence of GDP and population rises. These findings offer methodological insights for future urban flood risk assessment research and provide policymakers with valuable guidance to develop targeted adaptation measures in response to climate change.

Comparison and integration of physical and interpretable AI-driven models for rainfall-runoff simulation

Precise streamflow forecasting in river systems is crucial for water resources management and flood risk assessment. The Tagus Headwaters River Basin (THRB) in Spain is a key hydrological hub, providing regulated flow for agricultural, urban, and energy sectors, and facilitating water transfer to the Segura River Basin, a key semi-arid region. Given the basin's near-total allocation of water resources, accurate streamflow simulations are essential to optimize the socio-economic distribution and ensure sustainable management across both interconnected basins. This study conducts a comprehensive evaluation of the Soil and Water Assessment Tool (SWAT+), support vector regression (SVR), feed forward neural network (FFNN), and long short-term memory (LSTM) models in simulating the rainfall-runoff process at four gauging stations within the THRB. An ensemble machine learning technique is then applied to assess improvements in streamflow estimation. Results revealed that AI-based models significantly surpassed the SWAT+ model in performance. Furthermore, the application of an ensemble technique enhanced the precision of rainfall-runoff modeling by 18 to 26% during the calibration period and 4.1 to 9.2% during the validation period, compared to individual AI-based models. Additionally, the SWAT+ model's precision improved by 44 to 74% and 40 to 55% for the respective periods. The use of Shapley Additive Explanations (SHAP) methodology allowed the results of the ensemble with machine learning to be more interpretable by explaining how each model contributes to the prediction. This research offers significant contributions to hydrological modeling, highlighting the importance of ensemble techniques in elevating predictive accuracy for various river basins.

Urban flooding damage prediction in matrix scenarios of extreme rainfall using a convolutional neural network

The pivotal role of quantitative risk assessment in managing urban stormwater is underscored by the high spatial heterogeneity and complex non-stationarity complex of urban flooding. Conventional methods, reliant on measurable data, often fall short in accurately mapping spatial variations and gauging the full impacts of urban flooding. Addressing this gap, this study proposed a robust convolutional neural networks (CNN)-based tool, specifically designed for urban flooding risk and economic losses estimation under extreme design rainfall scenarios. Furthermore, the study validates the transferability of CNN models trained in data-abundant regions to similar but data-scarce regions, using Guangzhou as a case study for urban flooding damage prediction. The findings reveal that the most affected areas, particularly the old, densely built-up urban areas in the south-central Guangzhou, are susceptible to significant economic losses during extreme rainfall events. Notably, under the most severe scenario (Scenario 5), estimated economic losses amount to approximately $6359.91 million, with industrial and residential sectors bearing the brunt, accounting for 28.29 % and 39.94 % of the total losses, respectively. These insights are crucial for prioritizing mitigation efforts and formulating effective evacuation strategies in high-risk areas, ultimately aiding in the reduction of economic losses.

Spatiotemporal evolution and influencing factors of flood resilience in Beibu Gulf Urban Agglomeration

In the context of global climate change and rapid urbanization, enhancing flood resilience is essential for mitigating urban flood risk. However, few studies have conducted long-term, cross-scale dynamic evaluations of flood resilience and analyzed its influencing factors and mechanisms. Taking the Beibu Gulf urban agglomeration as the study area, a long-term, cross-scale dynamic evaluation index system based on the “Robustness-Resistance-Recovery” (3Rs) framework were adopted to assess the spatiotemporal evolution of flood resilience from 2000 to 2020. Furthermore, the optimal parameters-based geographical detector model is employed to identify the key influencing factors and mechanisms. The results reveal that pre-flood robustness is lower in coastal areas and higher in inland regions. In the during-flood stage, cities with greater comprehensive power exhibit stronger resistance. Post-flood recovery is higher in city centers and marginal mountainous areas, while coastal and inland low-lying regions show lower recovery. The flood resilience of urban agglomerations has improved in recent years, largely due to the enhancement of urban flood control infrastructure and healthcare capacity. However, disparities between cities persist. From 2000 to 2020, economic factors have been the primary drivers of improved flood resilience, while ecological factors have gained increasing importance over the past decade. These findings provide valuable insights for flood prevention, mitigation, and resilience management in urban agglomerations. The proposed dynamic evaluation index system offers a reference framework for evaluating flood resilience in other regions.

A novel framework for evidence-based assessment of flood resilience integrating multi-source evidence: A case study of the Yangtze River Economic Belt, China

Floods are extremely destructive to the society, economy and nature, and severely restrict sustainable urban development. There is an urgent need to assess the level of flood resilience, analyze its evolutionary characteristics and improve the flood resilience of the region. However, few previous studies have integrated multi-source evidence into a framework for assessing flood resilience over a longer time span. This study integrated multi-source evidence to analyze the Yangtze River Economic Belt, including literature-based evidence, policy texts and social media data. Subsequently, the Kruskal algorithm and Jaccard coefficient were used to construct collinear networks and identify the clustering dimensions of flood resilience. After evidence judgment, a flood resilience indicator system was constructed. The flood resilience index was calculated using the improved Critic method. The dynamic evolution characteristics and key influencing factors were illustrated by the Kernel density and Grey correlation methods, respectively. The principal contribution is to integrate multi-source evidence to assess flood resilience, with a newly constructed indicator system that systematically considers local public demand, policy priorities and common indicators from previous literature. The results show that: the flood resilience index of the Yangtze River Economic Belt is in the range of [0.16, 0.75] from 2000 to 2022, with higher values in the downstream compared to those in the midstream and upstream. The social dimension dominating the overall flood resilience. The kernel density curve of the YREB is characterized by a right trailing. The inter-regional differences tend to widen and have a polarization effect. The special funds for water conservancy finances and the length of flood prevention levees are key factors influencing the level of flood resilience. Finally, strategies for enhancing flood resilience were proposed. This study provides a feasible framework for constructing an indicator system and assessing flood resilience by integrating multi-source data. It is important for flood risk management and helps to the development of urban flood prevention and resilience strategies.

Investigation on flooding dynamic response and mitigation measure of river bridge

Flood-induced failure of river bridges is intense due to frequently extreme and undesigned precipitation for global climate change. To reveal the dynamic characteristics and failure mechanism of bridges, flooding dynamic analysis is fully investigated via fluid–structure interaction simulation and Arbitrary Lagrangian-Eulerian method. Parametric analysis of different fluids and bridge configurations and pre-ballast mitigation measure are further conducted based on a simply-supported box-girder bridge. This study reveals several significant findings: the narrow channel can accelerate downstream flood velocity of the bridge piers to over double that of the upstream flow; moreover, unsubmerged piers experience vertical downward wave force surpassing buoyancy; the box girders exhibit three flood-response states, i.e., intact, slip and collapse, dependent on the hydraulic loading and bridge resistance capacity. Maximum horizontal wave force and displacement increase exponentially with the depth of water storage. Post-slip damage, bearing stress on the wave-facing side tends towards zero, while the wave-backing side presents a twofold increase. Maximum horizontal displacement (Xmax) distinguishes the failure states, such as 0 m ≤ Xmax ≤ 0.03 m (intact), 0.03 m < Xmax ≤ 1.39 m (slip), Xmax > 1.39 m (collapse), whose relation to the maximum horizontal wave force has been proposed for design reference and the inversion analysis of the bridge force. Applying a pre-ballast equivalent to 75 % of girder weight reduces Xmax by up to 91.5 %, shifting failure from collapse to slip. These insights offer crucial guidance for enhancing bridge safety amidst escalating flood risks.

An operational discontinuous Galerkin shallow water model for coastal flood assessment

Hydrodynamic modeling for coastal flooding risk assessment is a highly relevant topic. Many operational tools available for this purpose use numerical techniques and implementation paradigms that reach their limits when confronted with modern requirements in terms of resolution and performances. In this work, we present a novel operational tool for coastal hazards predictions, currently employed by the BRGM agency (the French Geological Survey) to carry out its flooding hazard exposure studies and coastal risk prevention plans on International and French territories. The model, called UHAINA (wave in the Basque language), is based on an arbitrary high-order discontinuous Galerkin discretization of the nonlinear shallow water equations with SSP Runge–Kutta time stepping on unstructured triangular grids. It is built upon the finite element library AeroSol, which provides a modern C++ software architecture and high scalability, making it suitable for HPC applications. The paper provides a detailed development of the mathematical and numerical framework of the model, focusing on two key-ingredients : (i) a pragmatic P0 treatment of the solution in partially dry cells which garantees efficiently well-balancedness, positivity and mass conservation at any polynomial order; (ii) an artificial viscosity method based on the physical dissipation of the system of equations providing nonlinear stability for non-smooth solutions. A set of numerical validations on accademic benchmarks is performed to highlight the efficiency of these approaches. Finally, UHAINA is applied on a real operational case of study, demonstrating very satisfactory results.

A nonstationary stochastic simulator for clustered regional hydroclimatic extremes to characterize compound flood risk

Traditional approaches to flood risk management assume flood events follow an independent, identically distributed (i.i.d.) random process from which static risk measures are computed. Modern risk accounting strategies also consider nonstationarity or long-term trends in the mean and moments of the associated flood probability distributions. However, few approaches consider how extreme hydroclimatic events cluster in both space and time, compounding damage risks. Here we introduce a compound flood risk simulator that models and conditionally forecasts future variability in regional flooding events that cluster in time, given trends and oscillations in a variable climate signal. A modular, novel integration of wavelet signal processing, nonstationary time series forecasting, k-nearest neighbor (KNN) bootstrapping, multivariate copulas, and modified Neyman-Scott (NS) event clustering process provides users the ability to model interannual and sub-annual clustering of flood risk. Our semi-parametric flood generator specifically targets the clustered temporal dynamics of jointly modeled flood intensity, duration, and frequency over a finite future period of a decade or more, thereby providing a foundation for adaptation approaches that integrate temporally clustered flood risk into planning, response and recovery.

Rainwater Flood risk assessment in Zaatari Refugees Camp- Jordan: Towards Sustainable Solutions

Refugee settlements are frequently located in isolated and remote areas, characterized by poor-quality land and harsh climatic conditions. Zaatari camp in Jordan has suffered and continues to suffer from rainwater floods every winter. This study uses GIS analysis to assess flood risk in the camp, revealing that 61.1% of the area is moderately risky, while 0.84% and 22.42% are very high and high-risk zones respectively. The highest risk areas are near the camp boundary. This study recommends a structural intervention and first-class priority at the entrance area of the camp as it is located within a high-risk area and is of vital importance from a functional standpoint. These recommendations range to areas within high-risk zone in (1, 2, 7, 12) districts, to moderate risk zone in (4,8,11) districts with precautionary and preventive measures. In addition, some structural solutions can be implemented in several stages. This research is significant not only for managing flood risk in Zaatari but also for providing a replicable methodology for similar humanitarian settlements worldwide.

FLOOD RISK MAPPING IN THE IRTYSH RIVER BASIN USING SATELLITE DATA

Floods are among the most frequent and devastating natural disasters, causing significant economic damage and loss of life worldwide. Effective flood risk management relies on accurate modeling techniques that can predict vulnerable areas and assess potential impacts. In this study, flood dynamics are simulated in the Irtysh River Basin near Ust-Kamenogorsk, a city in East Kazakhstan prone to seasonal flooding using high-resolution satellite imagery and digital elevation data. The primary objective is to visually model flood risks based on terrain characteristics. The study utilizes imagery sourced from the Mapbox platform, which combines data from MODIS, Landsat 7, Maxar, and the Google Earth Engine, providing access to Sentinel-2 surface reflectance imagery at 10-meter resolution. Elevation data from the Copernicus global digital elevation model, with a 30-meter resolution, is used to simulate flood progression. The flood simulation involves calculating flood depth relative to the terrain’s elevation, allowing for a pixel-by-pixel determination of submerged areas. Each simulation incrementally increases water levels to generate a sequence of images, showcasing the progression of flooding over time. The study describes hydraulic soil characteristics usage, and focuses on visualizing flood risk based on terrain data and water level changes. The simulation results indicate that flooding initially impacts riverbanks as water flow starts from the northwest of the city with critical infrastructure becoming vulnerable once water levels exceed 2 meters from the lowest elevation point. These findings highlight the potential of high-resolution satellite imagery and terrain data for flood risk assessment and improving urban flood preparedness. The results provide valuable insights into flood progression enabling more informed decision-making for disaster mitigation.

HIGH-RESOLUTION SATELLITE ESTIMATION OF SNOW COVER FOR FLOOD ANALYSIS IN EAST KAZAKHSTAN REGION

The increasing frequency of extreme weather events linked to climate change has made flood forecasting an important issue, particularly in mountainous regions where snowmelt is a major driver of seasonal flooding. This study explores the application of snow cover estimation techniques to assess snowmelt dynamics and their potential impact on flood risks in the Ulba and Uba basins in East Kazakhstan. To achieve this, high-resolution multispectral satellite imagery from the Sentinel-2 Surface Reflectance dataset is used, focusing on images collected between March and October for the years 2021 to 2024. The images are processed in Google Earth engine platform with strict filtering based on spatial intersection with the basins and cloud cover pixels percentage, ensuring high-quality data for snow cover analysis. The study utilizes multiple remote sensing indices for snow cover estimation. The normalized difference snow index is calculated using the green and shortwave infrared bands to detect snow-covered pixels. Fractional snow-covered area is derived from the NDSI using the 'FRA6T' relationship, offering a more nuanced estimate of snow distribution across the basins. Additionally, a near-infrared to shortwave infrared ratio threshold is employed to minimize confusion between snow and water, improving the detection of snow cover, particularly in regions near water bodies or during melt periods. The resulting snow cover maps and fSCA estimates provide a detailed picture of snow distribution and melt dynamics, contributing to the assessment of snowmelt’s role in flood risk development. The obtained insights can assist in refining flood forecasting models, improving early warning systems, and supporting informed water resource management in vulnerable regions.

The Impact of Green Infrastructure on the Quality of Stormwater and Environmental Risk

Increasing urbanization and the associated sealing of areas and the use of storm sewer systems for drainage not only increase the risk of flooding but also reduce water quality in streams into which stormwater is discharged. Green infrastructure (GI) measures are applied with the aim of managing this stormwater sustainably and reducing the associated risks. To this end, a quantitative–qualitative approach was developed to simulate GI—namely, rain gardens, bioretention cells, and vegetative bioswales—at the urban catchment scale. The findings highlight the potential of applying GI measures to managing stormwater more effectively in urban environments and mitigating its negative pollution-related impacts. For the housing estate analyzed, a simulated implementation of GI resulted in a reduction in pollution, measured as total nitrogen (N; 9–52%), nitrate-N (5–30%), total phosphorus (11–59%), chemical oxygen demand (8–46%), total suspended solids (13–73%), copper (12–64%), zinc (Zn; 16–87%), polycyclic aromatic hydrocarbons (16–91%), and the hydrocarbon oil index (HOI; 15–85%). Reducing the concentrations of pollutants minimizes the risk to human health determined via the HOI from a low-risk level to zero risk and reduces the ecological risk in terms of Zn pollution from a significant risk to a low risk of adverse effects. The modeling conducted clearly shows that the GI solutions implemented facilitated a quantitative reduction and a qualitative improvement in stormwater, which is crucial from an environmental perspective and ensures a sustainable approach to stormwater management. Lowering the levels of stormwater pollution through the implementation of GI will consequently lower the environmental burden of pollutants in urban areas.

GIS Application for Creating Potential Flood Map Using AHP: A Case Study in Chaq-Chaq Valley, Sulaymaniyah City, Kurdistan Region, Iraq

This study aims to analyze and classify the hazards of flooding within the Chaq-Chaq sub-basin, the Sharazoor-Piramagroon basin, Sulaymaniyah City, Iraqi Kurdistan Region. The selected area is exposed to flooding from time to time. The study adopted many thematic maps, namely the topographic wetness index (TWI), elevation, slope, precipitation, land use–land cover (LULC), normalized difference vegetation index (NDVI), distance from the river, distance from the road, and drainage density. Modern technologies like remote sensing, GIS, ENVI, and AHP programs are used to obtain a digital map showing the classification of the flooding hazard within the sub-basin area. The results showed that large parts, about 50% of the basin area, are exposed to high and very high classes of flood risk. In comparison, areas of medium risk constituted about 27% of the total area of the sub-basin. Parts with low and very low risk recorded a percentage estimated at 22% of the total area. Large parts of the city of Sulaymaniyah face a high and very high risk of flooding, significantly impacting the tourist areas and economic facilities situated in the valley basin.

Downscaling, bias correction, and spatial adjustment of extreme tropical cyclone rainfall in ERA5 using deep learning

Hydrological models that are used to analyse flood risk induced by tropical cyclones often input ERA5 reanalysis data. However, ERA5 precipitation has large systematic biases, especially over heavy precipitation events like Tropical Cyclones, compromising its usefulness in such scenarios. Few studies to date have performed bias correction of ERA5 precipitation and none of them for extreme rainfall induced by tropical cyclones. Additionally, most existing works on bias adjustment focus on adjusting pixel-wise metrics of bias, such as the Mean Squared Error (MSE). However, it is equally important to ensure that the rainfall peaks are correctly located within the rainfall maps, especially if these maps are then used as input to hydrological models. In this paper, we describe a novel machine learning model that addresses both gaps, RA-Ucmpd, based on the popular U-Net model. The key novelty of RA-Ucmpd is its loss function, the compound loss, which optimizes both a pixel-wise bias metric (the MSE) and a spatial verification metric (a modified version of the Fractions Skill Score). Our results show how RA-Ucmpd improves ERA5 in almost all metrics by 3-28%—more than the other models we used for comparison which actually worsen the total rainfall bias of ERA5—at the cost of a slightly increased (3%) error on the magnitude of the peak. We analyse the behaviour of RA-Ucmpd by visualizing accumulated maps of four particularly wet tropical cyclones and by dividing our data according to the Saffir-Simpson scale and to whether they made landfall, and we perform an error analysis to understand under what conditions our model performs best.

Snow runoff modelling in the upper Indus River Basin and its implication to energy water food nexus

Pakistan's hydropower sector depends heavily on glacier and snowmelt water that originates from the Upper Indus Basin (UIB). It is expected that climate change may adversely affect future hydropower generation capacity as a result of fluctuations in the magnitude, seasonality and hydrological extremes of the Indus River flow. This study employed the Degree-Day Snowmelt Runoff Model alongside the Moderate Resolution Imaging Spectroradiometer MODIS and daily ground-based hydro-meteorological data to model the snowmelt runoff response in the UIB. The results indicated a significant increase in the annual and seasonal runoff under both RCP4.5 and RCP8.5 scenarios, suggesting more water availability for hydropower and irrigation. By the end of the century, annual river flow is projected to increase by 28 % to 69 % under the RCP4.5 and RCP8.5 climate scenarios. Consequently, rise in annual river flow is expected to increase the electricity generation capacity of future hydropower projects by 93 % to 167 % under the RCP4.5 and RCP8.5 scenarios, respectively. The construction of robust multipurpose dams may potentially reduce flood risks in downstream areas during peak flows, while also supplying water for hydropower generation and irrigation during low flows. This, in turn, may enhance the resilience of both the hydropower and agriculture sectors in the face of climate change.