In geophysics, crowdsourcing is an emerging nontraditional environmental monitoring approach that supports data acquisition from individual citizens. However, because of the involvement of undertrained citizens and imprecise low‐cost sensors, crowdsourced data applications suffer from different types of noises that can deteriorate the overall monitoring accuracy. In this study, we propose a machine learning approach for automatic crowdsourced data quality control (CSQC) that detects and removes noisy data inputs in spatially and temporally discrete crowdsourced observations coming from both fixed‐point sensors (e.g., surveillance cameras) and moving sensors (e.g., moving cars/pedestrians). We design a set of features from original and interpolated rainfall data and use them to train and test the CSQC models using both supervised and unsupervised machine learning algorithms. The performances of the CSQC models under various scenarios assuming no retraining are also tested (hereafter referred to as transferability). The results based on synthetic but realistic data show that the CSQC models can significantly reduce the overall rainfall estimate errors. Under the stationary assumption, the CSQC models based on both supervised and unsupervised algorithms perform well in noisy data identification and overall rainfall estimation error reduction; however, if the model is transferred to other cities with different rainfall patterns or noise compositions (without retraining), supervised multilayer perceptrons (MLPs) show the best performance.

Reservoirs, as critical nodes in regional water management, play an increasingly important role in drought mitigation. This study aims to optimize the drought-limited water level in the reservoir group and propose an evaluation method for selecting the optimal regulation scheme during drought periods. The reservoir water supply module within the Water Allocation and Simulation (WAS) model was enhanced to optimize the drought-limited water level of the reservoir group. A comprehensive adaptation index (CAI) was developed to quantitatively evaluate the effectiveness of water resource regulation under various drought scenarios. This methodology was applied to large and medium-sized reservoirs in the central Yunnan Province, China. The results show that the optimized drought-limited water level significantly improved the water supply performance of the reservoir group during drought years. Specifically, the optimized drought-limited water level notably reduced severe water shortage for water users in the long series and typical drought years, effectively mitigating the impacts of drought. Additionally, the most suitable water resource regulation strategies for different drought scenarios were identified. These research findings can provide technical references for reservoir management departments and drought operations authorities to formulate drought-limited level for the reservoir group and implement regional drought early warning and defense decision-making.

Accurate and reliable discharge estimation plays an important role in water resource management as well as downstream applications such as ecosystem conservation and flood control. Recently, data-driven machine learning (ML) techniques showed seemingly insurmountable performance in runoff forecasting and other geophysical domains, but they still need to be improved in terms of reliability and interpretability. In this study, focusing on discharge estimation and management, we developed an ML-based framework and applied it to the Huitanggou sluice hydrological station in Anhui Province, China. The framework contains two ML algorithms, the ensemble learning random forest (ELRF) and the ensemble learning gradient boosting decision tree (ELGBDT). The SHapley Additive exPlanation (SHAP) was introduced into our framework to interpret the impact of the model features. In our framework, the correlation analysis of the dataset can provide feature information for modeling, and the quartile method was utilized to solve the outlier problem of the dataset. The Bayesian optimization algorithm was adopted to optimize the hyperparameters of the ensemble ML models. The ensemble ML models are further compared with the traditional stage–discharge rating curve (SDRC) method and the single ML model. The results show that the estimation performance of the ensemble ML models is superior to that of the SDRC and the single ML model. In addition, an analysis of the discharge estimation without considering the flow state was performed. This analysis reveals that the ensemble ML models have strong adaptability. The ensemble ML models accurately estimate the discharge, with a coefficient of determination of 0.963, a root mean squared error of 31.268, and a coefficient of correlation of 0.984. Our framework can prove helpful to improve the efficiency of short-term hydrological estimation and simultaneously provide the interpretation of the impact of the hydrological features on estimation results.

The integration of multi-source data represents a defining trend in hydrological science, while the comprehensive quantification and characterization of inherent uncertainties in hydrological model prediction remains imperative. Data assimilation (DA) techniques offer a rigorous framework for integrating multi-source observational data with model simulations through systematic uncertainty characterization, thereby enhancing predictive accuracy while providing quantitative uncertainty estimates. This study systematically synthesizes and extracts the research hotspots and cutting-edge trends of DA within the hydrology domain. Specifically, from the perspectives of model structure, parameters, and states, it categorizes the development of data assimilation techniques in hydrology into system identification, parameter estimation, and state estimation. The research identifies several key challenges confronting the field of hydrological DA, including inherent nonlinear characteristics of hydrological processes, insufficient spatial coverage and limited availability of observational data, necessity for substantial modifications to existing hydrological models for DA compatibility, difficulties in quantifying errors within raw datasets, and computational complexity arising from high-dimensional state spaces during assimilation. Finally, using the Kalman filter as an illustrative example, the study demonstrates the concrete application of DA. It is proposed that the integration of deep learning with DA, coupled with the joint estimation of parameters and states, represents the promising and breakthrough directions for the future development of DA methodologies in hydrological research.

The uneven distribution of water resources in time and space restricts the development of society, which is a universal problem facing the world. To improve the ability to regulate water resources in water diversion projects and alleviate the contradiction between water resources and social development, China has decided to build water network projects (WNPs). A WNP is an integrated water supply system that includes multiple water sources and network water supply lines and integrates the functions of water supply, water delivery, and optimal allocation of water resources into a system that can supply water to many users. An appropriate water pricing mechanism is the key to ensuring the sustainable operation of water network projects. This paper presents a comprehensive water pricing model (WPM) for WNPs. The purpose of this model is to unify the water price measurement standard of WNPs and lay the foundation for water price marketization in the future. By applying the model to the Jiaodong WNP, it is found that the model is expected to improve the unbalanced use of water sources and the large difference in prices at water supply points (WSPs) in the project. This study not only provides a theoretical basis for water pricing reform but also has great potential to improve the efficiency of water resource use.

Winter water allocation in regulated plain river networks requires an operational link between hydrological low water benchmarks and replenishment decisions. This study focused on the Lixiahe plain river networks; utilizing winter water level records from nine stations during 2006–2021, five hydrological methods were applied to derive candidate water level benchmarks. A one-dimensional MIKE 11 model was applied to simulate total upstream replenishment scenarios of 40–400 m³/s and to establish station-specific stage–discharge relationships under fixed downstream boundaries and structure operation rules. Within the simulated range, the candidate benchmarks corresponded to scenario-specific replenishment requirements of 153.94–268.49 m³/s. The framework supports comparison of winter water allocation strategies and identification of controlling stations, but ecological sufficiency requires independent evidence from species, habitat, thermal, dissolved oxygen, and water quality responses.