Water

This study proposes a novel integrated framework that combines a Hippopotamus-Optimized Support Vector Regression (HO-SVR) prediction model with a Retrieval-Augmented Generation-enhanced Large Language Model (RAG-LLM)-based intelligent decision module, addressing the core challenge of bridging prediction and prevention in coal mine water inrush disasters. It represents the first application of the combined HO-SVR and RAG-LLM approach in this field. Methodologically, a hybrid data augmentation technique (SMOTE–GN–Bootstrap) alleviates data scarcity and imbalance, while feature selection and dimensionality reduction optimize the input features. The developed HO-SVR model demonstrates superior prediction accuracy over benchmark models. The key innovation lies in the RAG-LLM module which automatically generates interpretable reports and actionable prevention strategies based on the prediction results and key influencing factors, thereby establishing a closed-loop intelligent system from accurate prediction to informed prevention. Practically, this framework enables proactive risk management through data-driven predictions, significantly reduces water inrush incidents, and provides intelligent decision support for field operations, substantially enhancing mine safety. Furthermore, the study discusses the model’s potential and challenges across different geological settings, charting a course for developing more generalized models

The risk of urban flooding has escalated with increasing rainfall intensity and the expansion of impervious surfaces. While commercial models such as XP-SWMM provide reliable hydraulic analyses, their closed-source structure limits transparency and integration with external tools. In contrast, the Grid-Based Urban Drainage System Analysis Model (GUDS), developed on the Weighted Cellular Automata 2D (WCA2D) framework, offers greater flexibility for process verification and coupling with platforms such as GIS and spreadsheets. This study presents a comparative assessment of numerical stability and velocity estimation schemes between XP-SWMM and GUDS. Moving beyond previous validation-focused studies, it quantitatively examines how algorithmic formulations—particularly in flow velocity computation and numerical treatment—affect inundation propagation and model stability under varying topographic conditions. Results demonstrate that XP-SWMM yields higher analytical precision but is prone to numerical instability on steep slopes, whereas GUDS maintains stable simulations due to its simplified water-level-difference approach, albeit with reduced responsiveness to rapidly changing flows. The differences in maximum inundation depth, inundation area, and propagation speed were relatively minor—approximately 11.6%, 10.7%, and 9.2% on average, respectively. This work provides a novel quantitative perspective on the trade-offs between precision and stability in urban flood modeling, highlighting GUDS’s robustness and practical applicability as an open and extensible alternative to conventional equation-based models.

Satellite and reanalysis rainfall estimates (SREs) are valuable alternatives to gauge data in data-scarce regions; however, their reliability in areas with complex terrain and variable precipitation remains uncertain. This study evaluated six SREs (CHIRPS v2, ERA5, ERA5-Land, IMERG v07, MSWEP v2.8, and TRMM 3B42) against gauge observations over the period 2005 to 2019. The evaluation was conducted using multiple statistical, categorical, and distributional metrics at daily to seasonal timescales. Terrain-based classification and rainfall intensity categories were used to explore the influence of topography and event magnitude on product performance. The accuracy of SREs improves with temporal aggregation, the monthly scale offering the highest reliability for water resource management. However, their tendency to overestimate light and underestimate heavy daily rainfall requires careful bias adjustment in flood and extreme event analysis. MSWEP, CHIRPS, and IMERG provided balanced and consistent performance across all metrics, rainfall intensities, and terrain zones. Notably, ERA5 and ERA5-Land consistently overestimated average rainfall. All SREs identified dry days well, and their performance declined with increasing intensity. No significant performance variation was observed across different altitudes. This study provides valuable insights into the selection of rainfall products, supporting climate and hydrological studies in data-scarce areas of the Ethiopian highlands.