Search

Soil moisture flash droughts often pose significant challenges to humans and ecosystems, with wide‐ranging socioeconomic consequences. However, the underlying mechanisms of flash droughts and their changes remain unquantified. Taking China as a case study, we present a novel framework that combines machine learning with interpretable and cluster techniques to investigate flash drought mechanisms from 1980 to 2018. We first quantified the temporal contribution of drivers and further identified different mechanisms during drought onsets. We subsequently investigated the temporal changes in different mechanisms and classified drought event types. We identified four driving mechanism types triggering drought: Concurrent precipitation, Antecedent‐concurrent precipitation, Antecedent temperature‐concurrent precipitation, and Antecedent transpiration‐concurrent precipitation. The total effects from vegetation transpiration contributed to around 50% of the impacts for mechanisms involving antecedent transpiration and concurrent precipitation, highlighting the non‐neglectable role of vegetation water consumption in drought occurrences. Remarkably, about 60% of flash drought onsets exhibited close association with the antecedent anomalies, which contribute approximately 50% of overall effects, emphasizing the importance of the cumulative effects of drivers. Moreover, driving mechanisms associated with temperature and transpiration increased significantly over time, implying an elevated influence of these factors on droughts. Our classification of drought events reveals that nearly 70% of events were driven by at least two mechanisms, underscoring a complex time‐varying pattern of driving factors during drought events. The proposed holistic framework not only sheds insight into the multifaceted mechanisms driving flash droughts within China but also extends its potential applicability to broader geographical contexts. Key Points: A novel framework combining interpretable machine learning and cluster techniques was proposed to investigate flash drought mechanismFour distinct types of driving mechanisms of flash drought onset in eastern China were discoveredNearly 70% of flash drought events involve multiple mechanisms, emphasizing the time‐varying feature of drought drivers [ABSTRACT FROM AUTHOR]

The rise of extreme precipitation events has attracted wide attention throughout the world. However, the analysis of precipitation extremes is not sufficient in the short period of instrumental observation and it is complicated by their non‐stationary behaviours. As a period dominated by natural forcing, the last millennium is helpful for understanding long‐term changes in precipitation extremes. Therefore, this study aimed to (1) investigate changes in daily and subdaily precipitation extremes in China from the last millennium to the end of the twenty‐first century, (2) detect the non‐stationarity of precipitation extremes and (3) compare design storms by using stationary and non‐stationary distributions. The annual maximum 1‐day precipitation (RX1day) and annual maximum 3‐hour precipitation (RX3hour) are used to quantify variations of precipitation extremes. The results show that the trend of RX1day was insignificant for the Medieval Climate Anomaly (MCA, 950–1250) and the Little Ice Age (LIA, 1500–1800). For the historical period (1850–2014), the trend of RX1day and RX3hour was insignificant in most subregions. In addition, the differences for RX1day between the historical period and the MCA/LIA were mostly in the range of −5% to 5%. For the future period (2015–2100), RX1day and RX3hour are projected to increase in almost all land grids compared with the historical period, and they show greater increases in western China than eastern China. The non‐stationarity of RX1day and RX3hour is mainly found under the high‐emission scenario. Moreover, stationary generalized extreme value (GEV) distribution underestimates 50‐year return levels for RX1day and RX3hour in most areas compared with non‐stationary GEV distribution under the high‐emission scenario. Overall, this study suggests that daily and subdaily extreme precipitation will intensify over China in the future. For design storms, non‐stationary methods need to be used for risk management and engineering design under the high‐emission scenario. [ABSTRACT FROM AUTHOR]

This study proposes a new hybrid model for monthly streamflow predictions by coupling a physically based distributed hydrological model with a deep learning (DL) model. Specifically, a simplified hydrological model is first developed by optimally selecting grid cells from a distributed hydrological model according to their soil moisture characteristics. It is then driven by bias corrected general circulation model (GCM) predictions to generate soil moistures for the forecasting months. Finally, model‐simulated soil moisture along with other predictors from multiple sources are used as inputs of the DL model to predict future monthly streamflows. The proposed hybrid model, using the simplified Variable Infiltration Capacity (VIC) as the hydrological model and the combination of Convolutional Neural Network and Gated Recurrent Unit (CNN‐GRU) as the DL model, is applied to predict 1‐, 3‐, and 6‐month ahead reservoir inflows for the Danjiangkou Reservoir in China. The results show that the hybrid model consistently performs better than VIC and CNN‐GRU models with great improvement in Kling‐Gupta efficiency (KGE) values for lead times up to 6 months. Additional tests indicate that hybrid models based on CNN‐GRU outperform those based on LASSO, XGBoost, CNN, and GRU models. Moreover, compared with the distributed hydrological model, the hybrid model greatly reduces the computation burden of rolling prediction. It also saves decision‐makers the time and effort of trying different combinations of predictors, which is indispensable when building DL models. Overall, the new hybrid model demonstrates great potential for monthly streamflow prediction where training data are limited. Key Points: Deep learning and physically based distributed hydrological models are coupled for monthly streamflow predictionsA simplified hydrological model is developed on the basis of the distributed model to reduce computational cost in streamflow predictionsThe hybrid model is an efficient and accurate surrogate tool for real‐time monthly streamflow predictions [ABSTRACT FROM AUTHOR]

Drought is a natural disaster and its occurrence and development can be attributed to a range of driving factors, such as deficient precipitation, wind, temperature, land use/land cover changes, and so on. Therefore, attribution and monitoring of droughts are complicated and challenging. Standardized Precipitation Evapotranspiration Index (SPEI) has been widely used in meteorological drought monitoring. Nonstationarity has not been considered in SPEI-based drought monitoring. Here we propose a modified SPEI considering the nonstationarity of hydrometeorological processes, i.e. NSPEI hereafter. We observe nonstationarity SPEI in the Northeast, the Huang-Huai-Hai Plain, the Yangtze River Delta, and the Qinghai-Tibet Plateau. The goodness-of-fit can be determined for 76% of the total stations. Under the assumption of nonstationarity, the drought monitoring performance of the NSPEI is more robust than the traditional SPEI. Besides, we evaluate drought conditions across China under four shared socioeconomic pathways (SSPs), i.e. SSP126, SSP245, SSP370 and SSP585. We identify intensifying droughts in southwest China but a wetting tendency in northern China in winter under the SSP126 scenario; but a significant wetting tendency in southwest China and a drying tendency in northeast China under SSP245, SSP370 and SSP585 scenarios. This study provides a novel meteorological drought index and highlights a fresh picture of drought conditions across China under different SSPs. • A non-stationary SPEI index is developed. • The drought monitoring performance of the NSPEI is robust than the traditional SPEI. • We reevaluated drought conditions across China in future. [ABSTRACT FROM AUTHOR]

Understanding changes in human water utilization under the policy environment of the national top-down water unified allocation and regulation (TWAR) is crucial for promoting river health, sustainable resource utilization, and high-quality socioeconomic development along the whole Yellow River, a congenitally water-deficient river in China. This study assessed the TWAR policy effects on the total/sectoral/per capita water uses using the canonical difference-in-difference (DID) model, synthetic control method (SCM), and generalized synthetic control method (GSCM). The net effects of TWAR were assumed to be spatially heterogeneous and associated with variations in water-conserving efficiency, described by two indicators: water-conserving irrigation (WCI) and industrial recycled water (IRW). The suitable threshold intervals of WCI and IRW for benefiting decelerating water utilization were analyzed using a linear mediating effect (ME) model and panel smoothing transformation regression (PSTR) model, to explain the rebound effect in water use under the impact of the prevailing TWAR. The results show that TWAR initially had sustained positive effects on curbing water use growth, with the highest contribution to per capita agricultural water use in the lower section and high-quota region of the Yellow River Basin, but these effects decreased after a decade. Only when the WCI shifted within the threshold interval [0.283, 0.771] and the IRW did not exceed 0.962 could TWAR restrain water use. Emphasizing the conservation-oriented improvement of water use efficiency may necessitate extra attention to regions where water-conserving practices already reach high levels beyond the maximum WCI or IRW thresholds. These regions might be inclined to overutilize saved water to boost economic benefits and thereby potentially exacerbate the undesirable rebound effect from new uncontrolled water extraction. • TWAR policy effects on total/sectoral/per capita water uses were assessed. • Three quasi-natural experimental methods were employed: DID, SCM, and GSCM. • Water use initially slowed by TWAR may rebound after decadal periods. • TWAR's causality showed heterogeneity due to sectoral water-conserving efficiency. • Inhibiting role on water use was functional within suitable efficiency thresholds. [ABSTRACT FROM AUTHOR]

Optimizing reservoir drawdown operations holds significant implications for hydropower generation, water supply, and drought mitigation strategies. However, achieving multi-objective optimization in reservoir drawdown operations poses fundamental challenges, particularly considering emergency storage capacity and seasonal drought patterns. This study introduces a novel multi-objective optimization framework tailored for a mega reservoir, focusing on drawdown operations to enhance hydropower generation and water supply reliability. A drawdown operation model leveraging a multi-objective ant lion optimizer is developed to simultaneously maximize reservoir hydropower output and minimize water shortage rates. China's Three Gorges Reservoir (TGR), situated over the upper reaches of the Yangtze River, constitutes the case study, with the standard operation policy (SOP) serving as a benchmark. Results showcase the efficacy of the proposed method, with substantial improvements observed: a 10.6% increase in hydropower output, a 6.0% reduction in water shortage days, and a 9.5% decrease in minimal reservoir water release compared to SOP. This study provides robust technical and scientific bolster to optimize reservoir ESC and enhance the synergy between hydropower generation, water supply, and drought resilience. Additionally, it offers decision-makers actionable strategies that account for emergency water supply capacities. These strategies aim to support mega reservoir's resilience against extreme drought events facilitating the collaboration between modelers and policy-makers, by means of intelligent optimization and decision-making technologies., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

Monitoring suspended sediment concentration (SSC) in rivers is pivotal for water quality management and sustainable river ecosystem development. However, achieving continuous and precise SSC monitoring is fraught with challenges, including low automation, lengthy measurement processes, and high cost. This study proposes an innovative approach for SSC identification in rivers using multimodal data fusion. We developed a robust model by harnessing colour features from video images, motion characteristics from the Lucas-Kanade (LK) optical flow method, and temperature data. By integrating ResNet with a mixed density network (MDN), our method fused the image and optical flow fields, and temperature data to enhance accuracy and reliability. Validated at a hydropower station in the Xinjiang Uygur Autonomous Region, China, the results demonstrated that while the image field alone offers a baseline level of SSC identification, it experiences local errors under specific conditions. The incorporation of optical flow and water temperature information enhanced model robustness, particularly when coupling the image and optical flow fields, yielding a Nash-Sutcliffe efficiency (NSE) of 0.91. Further enhancement was observed with the combined use of all three data types, attaining an NSE of 0.93. This integrated approach offers a more accurate SSC identification solution, enabling non-contact, low-cost measurements, facilitating remote online monitoring, and supporting water resource management and river water-sediment element monitoring., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)