This study explores the potential of machine learning algorithms for earthquake prediction, utilizing fluid chemical anomaly data from hot springs. Six hot springs, located within an active fault zone along the southeastern coast of China, were carefully chosen as hydrochemical monitoring sites for an extended period of two and a half years. Using this data, a prediction model integrating six algorithms was developed to forecast M ≥ 5 earthquakes in Taiwan. The model's performance was validated against recorded earthquake events, and the factors influencing its predictive capability were analyzed. Our comprehensive analysis conclusively demonstrates the superiority of machine learning algorithms over traditional statistical methods for earthquake prediction. Additionally, including sampling time in the data sets significantly improves the model's predictive performance. However, it is important to note that the model's predictive performance varies across different hot spring and indicators type, highlighting the importance of identifying optimal indicators for specific scenarios. The model parameters, including the anomaly detection rate (P) and earthquake response time threshold (M), significantly impact the model's predictive capabilities. Therefore, adjustments are needed to optimize the model's performance for practical use. Despite limitations such as the inability to differentiate pre‐earthquake anomalies from post‐earthquake anomalies and pinpoint the precise location of earthquakes, this study successfully showcases the potential of machine learning algorithms in earthquake prediction, paving the way for further research and improved prediction methods.

The change of climate and environmental conditions has obviously affected the evolution and propagation of drought in river basins. The Hun River basin (HRB) is a region seriously troubled by drought in China, so it is particularly urgent to evaluate the evolution of hydrological drought and investigate the threshold of triggering hydrological drought in the HRB. In this study, the standardized runoff index (SRI) was applied to reveal the evolution characteristics of hydrological drought. Meanwhile, based on drought duration and severity identified by the run theory, the copula function with the highest goodness of fit was selected to calculate the return period of hydrological drought. Furthermore, the propagation time from meteorological to hydrological drought was determined by calculating the Pearson correlation coefficients between 1-month SRI and multi-timescale standardized precipitation index (SPI). Finally, based on the improvement of the drought propagation model, the drought propagation thresholds for triggering different scenarios of hydrological drought and its potential influence factors were investigated. The results show that (1) the hydrological drought showed a gradually strengthened trend from downstream to upstream of the HRB from 1967 to 2019; (2) downstream of the HRB were districts vulnerable to hydrological drought with longer drought duration and higher severity; (3) the most severe drought with drought duration of 23 months and severity of 28.7 had corresponding return periods that exceed the thresholds of both duration and severity of 371 and 89 years, respectively; (4) the propagation time from meteorological to hydrological drought downstream of reservoir has been significantly prolonged; and (5) the drought propagation threshold downstream of the HRB was remarkably higher than that upstream in all drought scenarios. Additionally, midstream showed the highest drought propagation threshold at moderate and severe drought scenarios, while downstream showed the highest drought propagation threshold in the extreme drought scenario.

Groundwater is the main irrigation water source in the Upper Peacock River. As fast enlargement of irrigation areas continues in recent years, the groundwater level declines continuously and has posed a threat to the sustainability of local agriculture and ecology. A numerical model was established with the code MODFLOW–2000 in order to predict the declining trend of groundwater level and formulate measures to counter the overexploitation, in which the river–aquifer interaction was elaborated and characterized by field survey. The results show that under current intensity of groundwater withdrawal, the levels of both unconfined and confined waters would decline continuously in 7 years from 2015. To stop the groundwater level from declining on the regional scale, the withdrawal rate should be compressed by 45% with respect to that in 2015. Moreover, taking consideration of the constraint of maintaining the ecological water level in the vicinity of the Euphrates Poplar forest in the study area, the withdrawal rate should be compressed 70% for seven towns around the forest.