The hydrological degradation vulnerability (HDV) of a non-perennial river represents the progressive deterioration of its natural flow regime, characterized by shortened flow duration and reduced baseflow. When a river basin loses its capacity to store, transmit, and regulate water, its resilience to climatic variability declines and the system becomes increasingly vulnerable. This problem is especially pronounced in semi-arid and arid regions, where climatic variability, anthropogenic pressures, and geogenic hazards jointly intensify hydrological degradation. This study evaluates HDV in the Pennar River Basin, India, by integrating multiple controlling attributes: the Standardized Precipitation Evapotranspiration Index (SPEI), Standardized Groundwater Table Index (SGWTI), density of minor surface water bodies, groundwater–surface water head difference, and sinkhole density. The relative importance of these attributes is quantified using the Analytical Hierarchy Process. The results are validated using a newly developed depth–duration curve–based river–aquifer stress index and further cross-validated with the standardized total water storage anomaly derived from GRACE and GRACE-FO satellite observations (HI_STWSA).To improve spatial resolution, the total water storage anomaly is downscaled using machine learning techniques. Key independent predictors are identified, and the optimal model is selected using entropy-based performance metrics. The final model downscales STWSA using SPEI, SGWTI, canopy cover, and surface water pixel density, enabling a detailed assessment of spatial vulnerability patterns. The results reveal strong spatial variability in HDV across the basin. The Pennar River near Chennur, in its upstream reach, exhibits high vulnerability, primarily due to intensive groundwater extraction and extensive sinkhole development, which severely disrupts subsurface storage and flow continuity. In contrast, the downstream reach near Nellore shows comparatively lower vulnerability, reflecting reduced extraction pressure and better retention of hydrological connectivity. These findings highlight the urgent need for region-specific groundwater management strategies to mitigate hydrological degradation in non-perennial river systems under increasing climatic stress.

The Yangtze River Basin, as a crucial strategic water resource region in China, is vital to the development and security of nearly 400 million people. Under the dual influence of climate change and human activities, significant changes have occurred in the land use pattern of the Yangtze River Basin, accompanied by alterations in its hydrological regimes. In recent years, extreme hydrological events have occurred frequently in the basin. Accurately quantifying the impact of land use change on river runoff is of great significance for adapting to hydrological changes. However, most existing studies focus on the effects on runoff volume or depth, failing to deeply reveal how land use change influences the intrinsic processes of runoff generation, particularly its impacts on key hydrological process characteristics such as floods, baseflow, and intra-annual distribution uniformity. This study aims to explore in depth the effects of land use change on runoff characteristics including floods, baseflow, and intra-annual unevenness coefficient, thereby refining the mechanisms through which land use change affects hydrological processes. To this end, the research employs deep learning algorithms to simulate daily-scale runoff in the Yangtze River Basin and applies a baseflow separation algorithm to partition baseflow. Based on the simulation results, the impact of land use change on runoff characteristics is quantified. The findings indicate that land use change significantly influences runoff characteristics such as floods, baseflow, and intra-annual distribution in the Yangtze River Basin. This study quantifies the effects of land use change on runoff characteristics including floods, baseflow, and intra-annual unevenness coefficient in the Yangtze River Basin. It contributes to revealing the regulatory effects of land use change on basin hydrological processes, providing a scientific basis for water resource management and conservation practices in the Yangtze River Basin.

Under climate change, the frequency and intensity of algal blooms in rivers and lakes have increased. Chlorophyll is commonly used as an indicator of algal blooms; however, its relationship with algal density is not linear. Such nonlinearity may introduce uncertainty in predicting bloom intensity based on chlorophyll alone. This may affect subsequent algal bloom predictions; therefore, it is crucial to clarify the dynamic relationship between chlorophyll and algal density. This study analyzes water quality data (2021–2024) to explore the dynamic relationship between chlorophyll and algal density in six major river basins in China. This study first examined the spatiotemporal patterns of the correlation and found significant spatiotemporal differences. Subsequently, random forest models and SHAP were applied to identify key water quality indicators affecting the correlation. We further employed linear regression to categorize monitoring sites in China into four temporal patterns, providing a basis for regional LSTM modeling. Our results indicate that the correlation between chlorophyll and algal density changes with spatiotemporal variations. Nutrient concentrations and turbidity were identified as key drivers, although their effects varied among basins and years. This study shows that accounting for regional differences is essential when interpreting Chl-a-algal relationships.

Under global climate change, extreme precipitation events are occurring more frequently, and total nitrogen (TN) and total phosphorus (TP) in river basins exhibit asymmetric responses, altering the nutrient structure. The nitrogen-to-phosphorus ratio (N:P), an important indicator of nutrient limitation in freshwater ecosystems, may undergo abrupt shifts under extreme rainfall conditions. However, the precipitation thresholds triggering these shifts and their underlying mechanisms remain unclear. Existing studies have largely focused on the responses of individual indicators, such as TN or TP, under extreme precipitation, while systematic analyses of abrupt changes in N:P are still limited. This study focuses on the Yangtze River Basin and constructs a daily-scale precipitation–water quality dataset for 2020–2024. Nonparametric PELT was applied to detect change points in the N:P time series, and correlation analysis assessed their associations with precipitation characteristics. Threshold analysis identified critical precipitation levels associated with abrupt N:P changes, and machine learning combined with SHAP interpretability analysis revealed the driving factors under varying precipitation amounts and intensities. The results provide insights into nutrient limitation transitions under extreme rainfall and support threshold-based water quality management, as well as nonpoint source pollution and eutrophication control at the basin scale.

Wildfires can fundamentally alter hydrological and geomorphic processes in forested catchments, often triggering pronounced changes in sediment transport and downstream water quality. Although post-fire increases in sediment and other water quality constituents have been widely reported, understanding remains limited on how fire and hydrological conditions interact to drive these responses, and their variability across catchments.Here we investigate how wildfire disturbance and short-term hydrological conditions interact to control sediment mobilization at the event scale. We apply a Bayesian hierarchical modelling framework to multi-year, high-frequency turbidity and streamflow observations from a network of forested catchments in eastern Australia. These catchments experienced a wide range of burn severities during the 2019-2020 Black Summer fires and were subsequently impacted by multiple flood events. We further examine how post-fire sediment responses vary with the extent and spatial configuration of burning, as well as key catchment characteristics.Our results indicate that severe wildfire disturbance led to clear steepening of event-scale concentration-discharge (C-Q) relationships in a substantial subset of the most severely burned catchments. This implies that sediment mobilization has been enhanced consistently across catchments which experienced more severe wildfires. In contrast, short-term hydrological conditions had modest influences on C-Q behavior, suggesting that fire-induced landscape changes dominated sediment responses in the immediate post-fire period. Notably, the strength of post-fire sediment responses did not scale simply with the overall proportion of catchment burned but appeared to depend on the spatial distribution of severe burning and forest type. These findings provide event-scale, multi-catchment evidence of how wildfire alters sediment regimes in river basins, improving understanding of post-disturbance hydrological alteration and offering insights to support sediment modelling and management in fire-affected landscapes.Keywords: wildfire, bushfire, sediment, mobilization

Accurate precipitation nowcasting is crucial for mitigating the impacts of extreme weather events. However, existing radar-based nowcasting methods primarily rely on a single data source and the kinematic extrapolation of radar echo motion, exhibiting limited capability to predict chaotic dynamics of convective systems due to the lack of supplementary atmospheric information. Global Navigation Satellite System (GNSS)-derived Precipitable Water Vapor (PWV) provides an integrated measure of atmospheric moisture and serves as a key precursor for precipitation evolution. Therefore, we introduce PWV as an essential physical constraint to improve precipitation nowcasting. This study presents WAVE-NowcastNet, a novel multi-source fusion deep learning framework that effectively integrates high-resolution radar observations with GNSS-derived PWV. At the core of the network design is the Water Vapor Control (WVC) module, which utilizes a trainable copy of the NowcastNet encoder and zero convolution layers. This design preserves skilful nowcasting capabilities of pre-trained NowcastNet while enabling the model to learn the modulating effects of water vapor on precipitation evolution. Comprehensive evaluations demonstrate that WAVE-NowcastNet consistently outperforms established baselines, including NowcastNet, Optical Flow, ConvLSTM, and PredRNN. Notably, our model achieves a significant performance leap in long-term forecasts (2–3 hours) compared to the original NowcastNet, with Critical Success Index (CSI) and Fractions Skill Score (FSS) scores increasing by 64.28% and 35.55%, respectively. Furthermore, WAVE-NowcastNet achieves the highest accuracy in areal rainfall estimation, highlighting its potential for enhancing downstream hydrological applications. This study also provides a scalable, computationally efficient paradigm for incorporating auxiliary atmospheric variables into deep-learning-based nowcasting models.

Catchment transit time, the duration water takes to travel from precipitation to streamflow outlet, serves as a fundamental diagnostic of how river basins store, route, and transform water and solutes, and thus of how hydrological alterations propagate in river basins. In large river basins, transit times integrate the cumulative effects of climate variability, land-use change, topographic structrure, and groundwater-surface water interactions, yet existing knowledge remains fragmented across scales, regions, and methodologies. By synthesizing evidence from inter-comparison studies, we highlight two emergent insights relevant to hydrological alteration in large basins. First, dominant environmental controls on transit time exhibit scale-dependent shifts: transit times are most often associated with local soil and hillslope-riparian processes in small catchments, with topographic organization and hydrological connectivity at intermediate scales, and with groundwater-geologic mediation in large basins, while climatic condition modulates this transition by shaping catchment wetness. Second, transit time metrics show distinct sensitivities to environmental drivers; to our knowledge, this is the first synthesis to empirically reveal contrasting controls on different transit time metrics across studies.  Despite methodological uncertainties, these observations potentially indicate valuable insights into how catchments partition rapid near-surface routing versus longer-term storage and mixing. This knowledge can provide a powerful lens for diagnosing hydrological alterations driven by climate change and human activities, including damming, water extraction, and land-use change, and for anticipating their ecological consequences through altered flow timing, water quality, and habitat connectivity. Further, we highlight key needs for further cross-catchment studies, including broader empirical coverage beyond small temperate catchments, improved tracer sampling across flow conditions, and uncertainty-aware, method-consistent analysis. Meanwhile, more systematic explorations are needed into complex interactions among environmental drivers, transport processes, and hydrological responses. These efforts will advance a process-based understanding of transit times, improving predictions of water and solute transport under environmental changes, especially in ungauged regions.

Large tropical floodplain lakes play a critical role in regional water resources and ecosystem services, yet their ecological responses to hydrological variability remain insufficiently documented due to limited long-term and integrated observations. The Tonle Sap Lake system, characterized by pronounced seasonal water-level fluctuations driven by monsoonal forcing and river–lake connectivity, provides a unique natural setting to explore hydrology–ecosystem interactions under various hydrological conditions. In this study, we present an integrated analysis of multi-season fish sampling data collected over two consecutive years under distinct hydrological conditions, combined with concurrent hydrological observations and analysis. Fish were captured using trap-based methods at multiple locations across the lake, and individual-level morphological information was obtained for a wide range of taxa. The dataset spans multiple seasonal phases, allowing comparative examination of fish community characteristics under different water-level regimes. We analyze patterns in fish size structure and community-level responses in relation to inter-annual and seasonal hydrological variability. Our results indicate that hydrological conditions are associated with shifts in the distributional properties of fish assemblages across seasons and years, while some structural characteristics remain comparatively stable. These results suggest that floodplain lake ecosystems may exhibit both sensitivity and resilience to hydrological variability, depending on ecological traits and temporal context. In addition to fish sampling, water quality measurements and environmental DNA (eDNA) samples were collected concurrently during the surveys, providing complementary background information on the lake system, although the present analysis focuses primarily on hydrological variability and fish community characteristics. By integrating hydrological information with biological observations, this study contributes to interdisciplinary efforts to better understand the coupled dynamics of water resources and aquatic ecosystems in large tropical floodplain lake systems under ongoing environmental change.