To minimize the damage from contaminant accidents in rivers, early identification of the contaminant source is crucial. Thus, in this study, a framework combining Machine Learning (ML) and the Transient Storage zone Model (TSM) was developed to predict the spill location and mass of a contaminant source. The TSM model was employed to simulate non-Fickian Breakthrough Curves (BTCs), which entails relevant information of the contaminant source. Then, the ML models were used to identify the BTC features, characterized by 21 variables, to predict the spill location and mass. The proposed framework was applied to the Gam Creek, South Korea, in which two tracer tests were conducted. In this study, six ML methods were applied for the prediction of spill location and mass, while the most relevant BTC features were selected by Recursive Feature Elimination Cross-Validation (RFECV). Model applications to field data showed that the ensemble Decision tree models, Random Forest (RF) and Xgboost (XGB), were the most efficient and feasible in predicting the contaminant source.

A Transient Storage Model (TSM), which considers the storage exchange process that induces an abnormal mixing phenomenon, has been widely used to analyze solute transport in natural rivers. The primary step in applying TSM is a calibration of four key parameters: flow zone dispersion coefficient ($K_f$), main flow zone area ($A_f$), storage zone area ($A_s$), and storage exchange rate ($\alpha$); by fitting the measured Breakthrough Curves (BTCs). In this study, to overcome the costly tracer tests necessary for parameter calibration, two dimensionless empirical models were derived to estimate TSM parameters, using multi-gene genetic programming (MGGP) and principal components regression (PCR). A total of 128 datasets with complete variables from 14 published papers were chosen from an extensive meta-analysis and were applied to derivations. The performance comparison revealed that the MGGP-based equations yielded superior prediction results. According to TSM analysis of field experiment data from Cheongmi Creek, South Korea, although all assessed empirical equations produced acceptable BTCs, the MGGP model was superior to the other models in parameter values. The predicted BTCs obtained by the empirical models in some highly complicated reaches were biased due to misprediction of $A_f$. Sensitivity analyses of MGGP models showed that the sinuosity is the most influential factor in $K_f$, while $A_f$, $A_s$, and $\alpha$, are more sensitive to $U/U_{*}$. This study proves that the MGGP-based model can be used for economic TSM analysis, thus providing an alternative option to direct calibration and the inverse modeling initial parameters.

Passive remote sensing is a practical and widely used method for bathymetry mapping in shallow rivers. However, the accuracy of this approach is limited because of different riverbed types; therefore, it is important to classify the riverbed types for improving bathymetry mapping accuracy and providing useful information for fluvial systems. In this study, we proposed a Gaussian mixture model (GMM)-based clustering method that utilizes hyperspectral imagery to classify riverbed types without sampling the bed material. We evaluated the proposed method in two shallow streams with different bed mixture conditions: (i) sand and vegetation and (ii) sand and moss-covered sand. The results showed that the GMM method accurately identified the spectral variability caused by diverse riverbed materials, enabling the precise classification of riverbed types. Moreover, by combining the GMM method with optimal band ratio analysis, we observed a reduction in error for the bathymetry mapping results by approximately 0.05 to 0.07 m. While our proposed method exhibits potential applications in various river environments, further research is needed to validate its effectiveness in classifying more complex riverbed types and conditions. Overall, our study findings suggest that the GMM-based clustering method using hyperspectral imagery is a promising tool for improving bathymetry mapping accuracy and classifying riverbed types in shallow rivers.

Digital twin (DT) technology is gaining attention for effective water quality management by integrating diverse data sources and enabling real-time insights. The practical implementation of DT technology for intelligent river water quality management requires extensive spatiotemporal big data, underscoring the critical need to integrate physical sensors, soft sensors, and remote sensing technologies. Here, we synthesized recent advancements in hybrid physical–soft sensing systems and highlighted their potential to address the inherent limitations of conventional water quality monitoring methods, such as limited spatiotemporal resolution and high operational costs. Soft sensors, driven by machine learning (ML), estimated difficult-to-measure water quality parameters by leveraging easily measurable variables from physical sensors. Therefore, soft sensors significantly expanded the range of measurable parameters and improved data collection frequency. In addition, remote sensing offers broad spatial coverage, enabling large-scale monitoring of optically active constituents, algal blooms, and sediment dynamics. We critically review methodologies and applications that integrate these sensing technologies into DT frameworks, and identify critical knowledge gaps, particularly the lack of a fully unified integration framework combining these technologies for next-generation DT systems. By assessing the strengths and limitations of each approach and proposing integration strategies, this study offers practical guidance and integration recommendations for DT-based river management.