IHE Delft Institute for Water Education

Urban trees provide a wide range of ecosystem services, including urban cooling, flood control, biodiversity, and air quality improvement. However, their impact depends on their physical attributes, seasonality, placement, and distribution in urban settings. This study focuses on the role of trees in stormwater management. Using the modified Gash and Rutter formulation, we developed a spatially distributed model to simulate individual tree’s interception capacity and throughfall. The model was applied to individual trees with time varying leaf area index (LAI) under different rainfall intensity, and canopy saturation conditions. Model validation against measurements showed strong agreement for total throughfall (R² ≃ 0.83) and interception (R² ≃ 0.61). At the urban scale, applying the model to a pilot area in Copenhagen, Denmark, resulted in a total runoff reduction of 15.2%, with local contributions varying between 2% and 43% based on tree characteristics. However, tree effectiveness was primarily driven by rainfall intensity and tree seasonality. Further research is needed to refine species-specific interception capacity and improve LAI-based parameterization in tree interception. Integrating this approach with hydrodynamic simulations could enhance understanding of trees’ local flood control contributions, supporting urban planning and tree-planting strategies to optimize city resilience.

Effectively characterizing uncertainty and error in flood prediction is essential for informed decision-making. This study combines advanced deep neural network architectures, i.e., Neural Hierarchical Interpolation for Time Series Forecasting (N-HiTS), and Long Short-Term Memory (LSTM), with multiple uncertainty quantification frameworks to evaluate flood forecasts across several watersheds in the southeastern United States. Bayesian inference, Monte Carlo–based methods, and quantile regression are applied to estimate predictive uncertainty. The comparative analysis examines how different uncertainty approaches perform across a range of flood magnitudes, highlighting their respective advantages and limitations at multiple scales. Results indicate that N-HiTS generally yields narrower and more reliable uncertainty bounds than LSTM. The findings further demonstrate that prior specification in MCMC sampling strongly influences uncertainty estimates and requires careful calibration. While Monte Carlo dropout, which is an approximate Bayesian technique, primarily captures uncertainty near flood peaks, MCMC offers a more complete characterization across the full hydrograph. In addition, this study investigates multi-site training to evaluate model adaptability under diverse hydrological regimes. Collectively, these results advance the integration of deep neural networks and uncertainty quantification to enhance flood modeling capabilities and risk management.

Extreme sea level (ESL) events, resulting from the combined effects of tides, storm surge, waves, and mean sea level rise, are a key driver of coastal flooding risk worldwide. While future increases in ESL magnitude and frequency due to anthropogenic climate change are well established, it remains unknown whether such changes, unlikely to be caused by internal variability alone, are already detectable in historical records, and whether their amplitude has emerged beyond the range of natural/internal variability at the global scale. Although a few studies have reported long-term increases in ESLs using historical tide-gauge records, the extent to which such changes are detectable and have emerged remains largely unknown, particularly at the global scale. This knowledge gap persists due to several challenges: (1) the infrequent, short-lived, and highly localized nature of ESL events, requiring high temporal and spatial resolution data; (2) the need for long, consistent time series to robustly characterize internal variability; and (3) the methodological complexity of detecting climate change signals and their emergence in extremes, compared to mean climate variables. Here, we assess the detection and emergence of climate change signals in ESLs along the global ice-free coastline using a 74-year (1950–2023) hydrodynamically modelled dataset with global coverage. Applying a detection and emergence framework tailored for extremes, we identify statistically significant increasing trends in ESL magnitude (global median 2.6 mm yr⁻¹) along approximately half of the global ice-free coastline. At nearly all computational points where a significant trend is detected, the associated ESL signal has already emerged beyond internal variability during the 1970–2023 period. The earliest times of emergence (regional-median ToE = between 1979 and 1982) occur in several IPCC reference regions, including the Equatorial Atlantic Ocean, Central Africa, the Equatorial Indian Ocean, Western Africa, Northeastern South America, Arabian Sea, and Northern South America. Linking times of emergence with national-level socio-economic indicators reveals that socioeconomically vulnerable countries with minimal historical CO₂ emissions have experienced increasing ESL magnitudes for the longest period.

Flood intensification due to climate variability and urbanization necessitates advanced forecasting tools, particularly in regions undergoing rapid transformation where drainage infrastructure data is often scarce. This study presents a methodological framework for flood forecasting in the Ribeirão Anhumas watershed (Campinas, Brazil), specifically applied to the International Hub for Sustainable Development (HIDS–Unicamp). As an innovation district currently under implementation, HIDS represents a unique opportunity to integrate predictive modeling into early-stage urban planning.The methodology addresses data scarcity by integrating physical modeling with machine learning. We have established a simulation environment using PCSWMM to replicate hydrological behavior under distinct infrastructure scenarios. These simulations, driven by high-resolution precipitation (10-min) and geospatial data (1 m DTM), generate the necessary synthetic training data for regions where sensor networks are yet to be deployed. The proposed architecture is designed to perform a binary classification of flood occurrence (Flood/No Flood), utilizing a multi-model approach: Recurrent Neural Networks (RNNs) for temporal dynamics, Convolutional Neural Networks (CNNs) for spatial patterns, and Graph Neural Networks (GNNs) to explicitly model the hydrological connectivity of the watershed.In this contribution, we present the complete data processing pipeline and the defined model architecture. The study focuses on evaluating the comparative performance of these architectures using classification metrics (accuracy, precision, recall, F1-score, and ROC curve). Furthermore, to ensure the model is transparent for decision-makers, we outline the application of Explainable AI (XAI) techniques, specifically SHAP and LIME. These are intended to identify the contribution of input variables to flood predictions, bridging the gap between "black-box" deep learning and interpretable hydrological processes. The final results aim to demonstrate how hybrid modeling can support the strengthening of early warning systems and resilience strategies in developing urban territories.

Anticipatory Action is increasingly put forward as a key approach to managing the emerging risks of climate change, by using forecasts to deliver vital resources to communities before disaster strikes. However, with climate change driving unprecedented weather extremes, how are anticipatory triggers, actions and implementation plans being designed to effectively prepare for and manage changing and emerging risks?In this research we identify examples of existing good practice, potential obstacles to progress, and ways in which weather and climate science can be better harnessed to strengthen anticipatory action as a climate adaptation tool. We use a mixed-methods approach, combining literature reviews, key informant interviews and stakeholder workshops. We find that while anticipatory action programming is usually informed by analysis of past events, there are emerging examples of good practice. These include addressing changing patterns of risk, undertaking scenario planning and simulation exercises, adapting triggers to account for upward trends in event frequency, and working to address the dangers of emerging risks such as heat waves and glacial lake outburst floods. However, in complex settings, for example in 'temporary' displacement camps, there is a need for longer-term thinking supported by integrated anticipatory action and resilience programming.

ABSTRACT This article examines whether economic growth reduces inequalities in access to water and sanitation across 64 countries over an average period of 13.5 years. Drawing on disaggregated data by income quintiles and rural–urban location, and employing ordinary least squares (OLS), two‐stage least squares (2SLS), and Seemingly Unrelated Regression Estimation, the study finds that GDP per capita growth is associated with overall improvements but limited distributive gains. Disparities persist among lower‐income quintiles and rural populations, and sanitation inequality shows no significant correlation with growth. The findings underscore the bidirectional relationship, as inequality in water access negatively impacts GDP per capita. Governance effectiveness, water use efficiency, and educational attainment emerge as stronger predictors of equitable access than income inequality. These results challenge growth‐centric development paradigms and call for equity‐oriented policy that integrates social and spatial justice into water and sanitation governance.

Wastewater characteristics vary significantly across regions due to a combination of climatic, socioeconomic, and infrastructural factors. Understanding these variations is important for accurate influent characterization and the development of region-specific guidelines. This study investigates geographical and regional differences in wastewater characteristics by analyzing influent data from multiple locations, including a detailed comparative analysis between two distinct regions with contrasting environmental and socioeconomic conditions. A comprehensive dataset spanning several years was utilized to examine key wastewater parameters, including BOD, COD, TSS, total Kjeldahl nitrogen, and total phosphorus. To ensure data reliability, preprocessing techniques incorporating machine learning models such as principal component analysis were applied. Statistical methods, including descriptive analysis, ANOVA, and correlation analysis, were applied to assess the variability and relationships between parameters. The results highlight significant differences in wastewater characteristics, not only between geographically distant locations but also within the same country. Whereas climatic conditions influence wastewater characteristics indirectly, socioeconomic factors such as water consumption patterns, industrial activities, and waste disposal practices play a significant role in shaping influent characteristics. These findings highlight the limitations of applying standardized design criteria and regulatory frameworks to wastewater treatment systems without accounting for local conditions. The study emphasizes the necessity of adaptive approaches that integrate regional variability, ensuring more precise influent characterization and enhancing overall system efficiency.

Abstract Indonesia’s southern Java coast has long been recognized for its high wave-energy potential, yet localized assessments—especially along the Malang coastline—remain sparse. This lack of site-specific data poses a barrier to effective Wave Energy Converter (WEC) deployment, which depends on accurate resource mapping and siting strategies. This study addresses the data gap by conducting a high-resolution wave energy resource assessment and identifying optimal WEC deployment sites along the south coast of Malang Regency, East Java. We utilized the Delft3D-WAVE module incorporating the SWAN spectral model, forced with ERA5 for the full year of 2023. The model domain was downscaled to 200–500 m resolution to capture nearshore variability, and outputs included significant wave height, peak wave period, direction, and wave power density (WPD). A multi-criteria approach was applied to evaluate five candidate sites based on resource level, water depth, and distance from shore. The results reveal a strong seasonal signal, with southeast monsoon waves producing peak offshore WPD values exceeding 40 kW/m, while calmer northwest monsoon conditions generate WPD values of 5–10 kW/m. Spatially, nearshore WPD ranged from 12–18 kW/m, with Site 1 identified as the most technically and economically balanced option. Site 3 offers comparable resources but is very shallow, favoring shoreline or nearshore device types. This study provides a replicable framework for wave energy site selection in data-sparse coastal regions of Indonesia. It highlights the importance of integrating seasonal dynamics, local bathymetry, and engineering criteria to support the sustainable development of marine renewable energy.