Successive occurrences of compound heatwaves and extreme precipitation events (CHEPs) can amplify their adverse socioeconomic consequences. However, the role of anthropogenic forcings in driving their changes remains incompletely understood. Here, by leveraging multiple Single Model Initial-condition Large Ensembles, we systematically partition the effects of anthropogenic greenhouse gas (GHG) and aerosol (AER) emissions on the evolution of CHEPs. We find that CHEPs occur more frequently in high latitudes of the Northern hemisphere, East and Southeast Asia, North Central South America and Central Africa. CHEP frequency has significantly intensified by 1.3 times over global land areas during 1984-2014 relative to the pre-industrial period. This increase is predominantly driven by the rapid escalation of GHG emissions despite an approximately 50% offset by AER emissions and other external forcings. However, this dampening effect has weakened in recent years due to declining aerosol emissions, resulting in an accelerated global increase in CHEP occurrence. Our bivariate attribution and projection framework demonstrates that anthropogenic climate change exerts a substantially stronger influence on CHEPs in tropical regions than in other areas. Furthermore, most rare and extreme CHEPs are predominantly driven by anthropogenic GHG and are projected to increase more dramatically than normal events as global warming continues. Under a 3°C level of anthropogenic warming, CHEP frequency is projected to amplify more than fourfold across most land areas. Our findings underscore the critical need to cut fossil fuel emissions to safeguard against exacerbated compound extremes.

Effective water scarcity assessment is frequently challenged by spatial mismatches between fine-resolution hydrologic water supply and administratively aggregated water-use statistics, which can obscure spatial heterogeneity and reduce the interpretability of results for planning. To address this spatial mismatch, a scale-matched framework is proposed for water scarcity risk screening by coupling multi-scale water demand simulation with runoff-derived renewable potential supply. Water demand is simulated by spatially explicit variables via a cellular automata (CA) model to produce gridded water demand at three scales: 1 km × 1 km scale, the appropriate spatial scale, and prefecture scale. Runoff-derived renewable potential supply is derived as runoff depth multiplied by a regionally specified water resources exploitation ratio, and this potential is also expressed as gridded fields at the three scales. Water scarcity risk screening indicators are then computed on these scale-matched grids and subsequently aggregated into basin and administrative units. The resulting scale-matched water scarcity risk signals are quantified by three complementary indicators—deficit ratio, temporal reliability, and non-deficit fraction, which characterize the mismatch magnitude, the long-horizon stability, and the spatial prevalence of non-water scarcity conditions, respectively. Results indicate that screening signals are systematically altered across the three different scales: the 1 km × 1 km scale reveals fragmented high-pressure clusters, the appropriate spatial scale retains coherent regional patterns while preserving meaningful sub-unit contrasts, and prefecture-scale aggregation yields smoother signals but can mask localized risk pockets. Persistently screened hotspots are concentrated in demand-intensive corridors such as the North China Plain and in arid inland systems in the Northwest, with strengthened screening signals toward end-century conditions in parts of the Northwest and the Southwest. Overall, a reproducible, scale-consistent screening tool designed to support spatial prioritization, basin–administrative coordination, and targeted planning diagnostics.

To mitigate flood disasters and reduce their risks, a lot of reservoirs, as a reservoir group, and some flood detention basins (FDBs) have been constructed. As the flood control targets of the reservoir group and the FDB are often the same or are interrelated, it is necessary to integrate the reservoir group and FDB operation instead of operating them independently. Hydropower benefit is one of the functions of the reservoir group besides its flood control. There should be a tradeoff between the flood risk and the hydropower benefit from the integrated flood control operation of a reservoir group and FDB. A framework is proposed to determine and understand the tradeoff for enhancing decision-making in flood control. It has been applied in a case study of the Jingjiang FDB and a mixed reservoir group (Three Gorges, Shuibuya, and Geheyan) in the Changjiang River Basin. It has been found that the integrated operation of Jingjiang FDB can enhance the exchange between extreme flood risk and hydropower benefit, while this exchange is weakened for the non-extreme flood conditions due to the operational inflexibility of Jingjiang FDB. The safety flow at the downstream protection target has a significant impact on the tradeoff under non-extreme flood conditions but only a slight impact under extreme flood scenarios. The damage coefficient exerts a significant and consistent influence on the tradeoff across both extreme and non-extreme flood conditions. It is critical to determine the damage coefficient and the safety flow for improving the performance of the flood control operation. The proposed framework not only determines the tradeoff between flood risk and hydropower benefit but also helps facilitate the integrated flood control operation of a reservoir group and FDB.

Global warming and intensifying human activities are altering the global hydrologic cycle, raising concerns about future water availability. Although many studies have examined changes in runoff, the global-scale association between socioeconomic development and changes in rainfall-runoff relationships (RRRs) remains insufficiently quantified. Here, changes in RRRs were diagnosed across 1492 basins during 1990-2015 using annual runoff coefficients, and dominant covariates were identified using a conceptual rainfall-runoff model with a covariate-driven time-varying parameter. Candidate climatic, vegetation, and socioeconomic covariates were tested individually in comparable schemes, and dominance was defined diagnostically as the covariate that produced the largest improvement in runoff simulation skill relative to a constant-parameter baseline. A changepoint in the mean runoff coefficient was detected in 1201 basins (80.5%). Among diagnosed basins, natural environmental covariates (climate and vegetation) were most frequently diagnosed as dominant (67.7%), whereas socioeconomic covariates accounted for the remaining 32.3%. Within the socioeconomic category, macro indicators (GDP, population, and per capita GDP) were more frequently diagnosed as dominant than intervention-specific proxies. Within socioeconomically dominated basins, runoff coefficients decreased more often than increased (65.9% versus 34.1%). For the macro-socioeconomic subset, explainable machine learning indicated that both climatic covariates and development-related proxies (e.g., impervious expansion and water use) were strongly associated with runoff variability. Overall, the results map where rainfall-runoff relationships have shifted globally and quantify the relative importance of socioeconomic covariates, supporting the incorporation of socioeconomic pathways in large-scale hydrologic assessment and water-resource planning.

Accurate flood simulation and forecasting are crucial for reservoir operation and disaster mitigation. Traditional process-based hydrological models are often limited in capturing nonlinear runoff responses under complex hydro-meteorological conditions, whereas deep learning models perform well yet lack physical consistency. To address these challenges, a novel hybrid dynamic parameter network (HyDPNet) deep learning model was proposed, which integrates a differential form of the Xinanjiang (XAJ) model into recurrent neural network units with parameters generated by the auxiliary neural networks. An LSTM post-processor further captures long-term dependencies and enhances the simulation accuracy. The application to the Lushui River basin demonstrates that the HyDPNet model achieves superior performance, with Nash-Sutcliffe Efficiency (NSE) values of 0.98 and 0.97 in the training and test periods, respectively. Compared to the benchmark models, HyDPNet yields the lowest simulation errors and flood peak errors across the flood events. By combining physical process representation and data-driven learning, the HyDPNet model provides a physically interpretable, process-informed framework for hydrological simulation, supporting disaster prevention and effective water management.

Compound hot and dry events (CHD) severely threaten terrestrial ecosystems, yet their non-linear impacts remain insufficiently quantified. Synthesizing 198 unique flux sites (213 temporal records), multi-source observations, and climate simulations, we reveal that historical CHD severity and coincidence rates intensified by 105–154

Risk assessment of hydrological extremes as well as process-based simulation and projection under climate change rely on high-quality watershed datasets. However, existing large-sample hydrology datasets are usually constrained by limited spatial coverage and lack future climate change scenarios at a catchment scale. Here, we introduce a Global Large-Sample Watershed Synthesis dataset (GloWS), which covers 23,029 basins from 24 international, national and sub-national sources. GloWS is generated after rigorous station screening, careful visual inspection by Google Map satellite imagery, and systematical data quality control. This dataset includes metadata, hydrological indices, meteorological variables, land-use and land-cover characteristics, water storage terms, and other static attributes. To empower future projections and assess climate change impacts, GloWS also includes nine daily catchment-mean bias-corrected variables from 22 Global Climate Models (GCMs) under the Coupled Model Intercomparison Project Phase 6 (CMIP6), and provides the multi-model ensemble mean scenarios throughout the 21st century. We envision that GloWS, with its comprehensive attributes and extensive spatiotemporal coverage, could provide support for a broad range of large-sample hydrology applications.

Microplastics are pervasive in inland waters, yet large-scale association patterns of abundance and traits remain unclear. Here we compiled a global lake database integrating surface-water and sediment observations with 56 hydroclimatic and anthropogenic indicators. Most hydroclimatic variables show positive associations with abundance in surface waters but act oppositely in sediments. Leading contribution patterns differ: leaf area index and 2-m dew-point temperature rank highest in surface waters, whereas solar radiation and wind speed lead in sediments. To enable finer-scale evaluation, we conducted dense sampling in China, which hosts one of the world's largest and most intensively studied lake systems. Hydroclimatic and anthropogenic fingerprints show clear cross-compartment contrasts: surface-water traits associate strongly with aquaculture production and rural income, while sediment traits more link to population aged 0-14 and chemical oxygen demand. Our findings imply a hydroclimate-human game across compartments, highlighting fingerprints structuring microplastic abundance and traits.

The uncertainties of inflow forecasts pose significant challenges to real-time reservoir scheduling decisions and risk assessment. This study proposed a novel method for risk analysis of real-time reservoir scheduling decisions that integrates inflow process forecasting uncertainties. A three-member ensemble inflow forecasting scheme was constructed for the Three Gorges Reservoir, and the uncertainty of inflow process forecasts was quantified based on the Copula-based Transition Density Bayesian Model Averaging (CTD-BMA) and the Gaussian Copula Approach (GCA)-based probabilistic inflow process forecasting. A coupled feature-temporal dual-attention mechanism Long Short-Term Memory (DA-LSTM) neural network scheduling model was constructed to make the real-time reservoir scheduling decisions. Combining the forecasted inflow process ensembles with the DA-LSTM scheduling decisions, the forecasted water level processes were calculated based on the water balance principle, thereby evaluating the rationality of scheduling decisions through the probability of exceeding target water levels. The results demonstrated that the CTD-BMA and GCA methods can achieve reliable probabilistic inflow process forecasts and generate inflow process ensembles with temporal correlation structures. The proposed risk analysis method successfully reflects the water level exceeding risks associated with different scheduling decisions in a probabilistic form, providing an intuitive and reliable reference for decision-makers. This study offers a new perspective on integrating probabilistic forecasting with reservoir scheduling decisions.

Abstract Bankfull discharge, the maximum flow a river can convey before spilling over its banks, is central to modelling flood risk and understanding river-channel evolution. Global flood inundation models assume a 2-year return period for bankfull conditions, but this assumption remains untested globally. Here we use observations and machine learning to estimate bankfull discharge at ~1-km resolution along a recently developed global river network. We show that the widely used 2-year discharge reasonably approximates bankfull-flow magnitude, but bankfull return periods vary substantially within and across climate zones. Bankfull conditions occur more frequently in tropical and temperate regions (median return periods of 1.5 and 1.8 years; interquartile ranges of 2.5 and 3.2 years, respectively) and less frequently in cold and arid regions (2.8 and 4.3 years; interquartile ranges of 4.8 and 6.0 years). This variation across climate zones provides a basis for improving the representation of channel capacity in global flood models.

Heatwaves and extreme precipitations are the two prevalent types of weather-related extreme events globally. Compared with univariate extremes, impacts of compound extreme precipitations preconditioned by heatwaves (CHEPs) on the society and economy can be amplified. Previous studies demonstrated that heatwaves can trigger extreme precipitations by enhancing atmospheric instability and moisture-holding capacity. Other studies projected future changes in CHEPs under various greenhouse gas emission scenarios. However, there is a lack of studies assessing the time of emergence (ToE) of CHEP change signals, especially for record-shattering events. Since current water resource management strategies and infrastructures are based on historical data, it is crucial to understand when hydro-meteorological conditions will surpass unprecedented levels to develop effective adaptation and mitigation strategies for climate change.Here, we present a global analysis of ToE for record-shattering CHEPs as well as their exposed GDP and population (POP). Both the frequency and magnitude of observed CHEPs have substantially increased during the past 65 years at the global scale. Using climate models from Detection and Attribution Model Intercomparison Project, we find that rarer CHEPs are increasingly attributable to anthropogenic greenhouse gas emissions, while aerosol emissions have a mitigating effect on their occurrences. To detect when historical record-shattering events will become normal, we develop a novel framework based on advanced Single Model Initial-condition Large Ensemble simulations. Our results indicate that CHEP hotspots, including East and Southeast Asia, North-central South America, and Central Africa, are likely to experience earlier ToE compared to other regions. In contrast, arid regions, such as North Africa, West Asia, and southwestern Australia, show no signs of ToE until at least 2100. GDP and POP exposure to such events reveal an alarming upward trend throughout the 21st century. By the late 21st century, 41% (29%) of sub-regions defined by the Sixth Assessment Report of the Intergovernmental Panel on Climate Change are projected to experience GDP exposure exceeding 4,000 billion USD (at 2010 purchasing power parity) to record-shattering frequency (magnitude), while 34% (27%) are expected to have POP exposure exceeding 100 million under the SSP2-4.5 scenario. Record-shattering CHEPs pose a distinct threat to the economy between 21.75°N and 53.25°N, with the most significant impact between 35.25°N and 39.75°N. Compared to the GDP exposure, the POP exposure hotspots shift toward lower latitudes, with a broader range extending from 0.75°S to 53.25°N. Additionally, we classify areas based on the Human Development Index and income levels defined by the World Bank. The unequal distribution of GDP and POP exposure reveals the poorest and least developed countries will experience more extended impacts compared to wealthier nations. This study highlights the urgent need for region-specific mitigation and adaptation strategies to combat climate change, especially for the vast high-risk and low-income regions.

Reliable flood projection is crucial for designing suitable flood protection structures and for enhancing resilience in vulnerable regions. However, projections of future flooding suffer from cascading uncertainties arising from the climate model outputs, emission scenarios, hydrological models, and the shortage of observations in data-sparse regions. To overcome these limitations, we design a new hybrid model, blending machine learning and climate model simulations, for global-scale projection of river flooding. This is achieved by training a random forest model directly on climate simulations from 20 CMIP6 models over the historical period (1985−2014), with extreme discharges observed at approximately 15,000 hydrologic stations as the target variable. The random forest model also includes static geographic predictors including land cover, climate, geomorphology, soil, human impacts, and hydrologic signatures. We make the explicit assumption that the random forest model can ‘learn’ systematic biases in the relationship between the climate simulations and flood regimes in different regions of the globe. We then apply the well-calibrated random forest model to a new vector-based, global river network in approximately 18.51 million reaches with drainage areas greater than 100 km2. Global changes in flood hazard are projected for the 21st century (2015−2100) under SSP2-4.5 and SSP5-8.5. We show that the data-driven method reproduces historical annual maximum discharges better than the physically-based hydrological models driven by bias-corrected climate simulations in the ISIMIP3b experiment. We then use the machine learning model with explainable AI to diagnose spatial biases in the climate simulations and future flood projections in different regions of the globe.

Streamflow seasonality, characterized by its unequal intra-annual distribution, is predominantly shaped by precipitation dynamics in non-snowmelt-dominated regions. However, climate change and human activities can alter the hydrological cycle, weaken the consistency in precipitation-streamflow seasonality, and further modify streamflow seasonality. Despite its importance, the mechanisms modulating precipitation-streamflow seasonality consistency remain underexplored. This study employs time-varying copulas and partial correlation analysis to identify and quantify the evolving seasonality inconsistency between precipitation and streamflow under the influence of multiple anthropogenic and climatic factors. In the Yangtze River Basin (YRB), recent environmental changes, including climate warming, vegetation greening, and reservoir regulation, have increasingly influenced the hydrological cycle. In this study, a slight intensification of precipitation seasonality is observed in half of the YRB watersheds, whereas streamflow seasonality is moderated in 90 % of the watersheds. The climatic and human-induced changes consistently weaken the dependency between precipitation and streamflow across seasons. Comparing the dependency strength between the periods 2006-2015 and 1982-1991, a decline of 17 % was observed in spring, 1 % in summer, 42 % in autumn, and 75 % in winter. Partial correlation analysis highlights three key factors driving the observed decrease in summer streamflow and increase in winter streamflow, with reservoir regulation exerting the most pronounced influence in winter. This study sheds new light on investigating streamflow seasonality in a changing environment.

Water use simulation plays a pivotal role in water resource management globally. Simulating water use at regional raster scale enables better alignment with available water resources, facilitating efficient allocation. However, there remains a deficiency in methods of spatiotemporal scale selection for ensuring the simulation accuracy while also guaranteeing the information density at each raster scale. A novel framework has been proposed to select the appropriate spatiotemporal scales for water use simulation. The framework utilizes an iterative input variables selection (IIS) algorithm to identify optimal input variables for water use simulation and an end-to-end deep learning-based spatiotemporal scale adaptive selection (SSAS) model to determine the appropriate spatiotemporal scales. Due to China's substantial population, water demand, and the growing challenges of global warming, the country is particularly susceptible to water scarcity. The proposed framework was applied to select the appropriate spatiotemporal scales for simulating irrigation, domestic, and industrial water use across 341 prefectures in China. The results indicate that the appropriate spatial scales for irrigation water use simulation range from 1 km to 5 km in most places, while they vary from 1 km to 4 km for domestic and industrial water use simulation. Furthermore, the appropriate temporal scale generally spans from 10 days to 45 days for all three types of water use simulation. It is interesting to find that the simulation accuracy is significantly impacted by the selection of appropriate temporal scales through the parameter sensitivity analysis. Our proposed framework supports water resource management and facilitates efficient water resource allocation to mitigate water scarcity.

As warmer temperatures enhance atmospheric moisture, hydrological droughts tend to intensify in most regions of the globe. Consequently, younger generations are expected to face a more severe risk of hydrological drought during their lifetimes, emphasizing the critical issue of intergenerational inequity due to climate change. To quantify exposure to hydrological drought across generations, we constructed a cascade model chain for drought simulation using hybrid terrestrial models, based on 5 GCM outputs under SSP5-85, five hydrological models and a deep learning model. We then projected future univariate and bivariate hydrological drought evolution in 4091 river basins, and quantified lifetime exposure to drought for the age groups born in 2020 and 1960. Drought severity and duration are projected to increase substantially in the Eastern America, Southern Brazil and Western Europe, over 79 % of basins. Extreme droughts far beyond historical records are expected to become more frequent and impact Western Europe in particular. Of note, the exposure of the different age groups to hydrological drought shows a notable disequilibrium. Exposure of people born in 2020 to hydrological drought hazards is projected to increase by 12 % over the late 21st century compared to those born in 1960, indicating that the acceleration of climate change is expected to increase the lifetime risk of future generations. The exposure factor of the newborns is 1.4 times higher than that of 80 years of age under warming condition. Our findings underscore that future drought conditions under extreme warming pose a significant threat to the living conditions of younger generations.

Sea ice albedo is a critical parameter for quantifying the energy budget in the Antarctic region. High spatiotemporal resolution sea ice albedo product is essential for Antarctic climate and environmental research. In this study, based on Visible Infrared Imaging Radiometer Suite (VIIRS) reflectance data, we use the Multiband Reflectance Iteration (MBRI) algorithm to calculate sea ice albedo. This algorithm fully utilizes multi-band observations from single-angle/date observed information to correct the anisotropy of the sea ice surface. Additionally, spatiotemporal information is utilized to reconstruct albedo under cloudy-sky conditions, while correcting for cloud radiative forcing effects. A new daily seamless Antarctic sea ice albedo product with a 1 km resolution is then generated for the period 2012 to 2021. Monte Carlo simulations show that the average retrieval uncertainty of this product is 0.022, with higher uncertainty in backward observations. MBRI albedo product was validated using in situ measurements from Automatic Weather Stations (AWS). The results show that the bias is 0.02, with a root mean square error (RMSE) of 0.071. After upscaling to 25 km resolution and applying a 5 d temporal aggregation, the RMSE decreased to below 0.055. Compared to existing albedo products, the MBRI product exhibits improved spatial continuity due to the reconstruction of cloudy-sky pixels. Statistical analysis shows that albedo under cloudy-sky conditions is higher than under clear-sky conditions (mean difference: 0.035-0.064). The MBRI albedo product can be used to estimate sea ice albedo feedback, energy balance analysis and sea ice monitoring. The latest version of our albedo (version 2) and uncertainty datasets are available at 10.5281/zenodo.11216156 (Ma et al., 2024) and 10.5281/zenodo.15067607 (Ma et al., 2025), respectively.

The frequency and severity of meteorological drought across Eastern Asia (EA) are observed and projected to increase with ongoing global climate change. However, the temporal and spatial responses of land surface elements including soil moisture and runoff to meteorological drought remain insufficiently understood and characterised, especially in terms of quantifying the differentiated lag time of soil moisture and hydrological drought responses. In this study, we investigate the spatiotemporal response of soil moisture and hydrological droughts to meteorological drought across EA, using the Standardised Precipitation Index (SPI; a measure of precipitation deficit), Standardised Soil Moisture Drought Index (SSMDI; quantifying soil moisture anomalies) and Standardised Streamflow Index (SSI; assessing streamflow deviations) at short- (1-3), medium- (6-9) and long-scales (12-24 months). The results show that in northern EA, soil moisture droughts indicate weak negative correlations with meteorological droughts at 1-6 month time scales, while in southern regions, stronger positive correlations are observed at 9-24 month time scales. Hydrological droughts exhibit consistently strong positive correlations with meteorological droughts across all time scales, especially in the southern regions of EA. Northern regions exhibit longer lag times of up to 8 months of soil moisture droughts in response to meteorological drought, whereas southern regions exhibit rapid responses (1 month). In all time scales, hydrological droughts in northern regions of EA have a delayed lag times extending up to 10 months. As climate change and land use changes may amplify drought impacts, long-term forecasting and adaptive management are essential to mitigate water shortages and ecological stress across the EA region.