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.

Hydrogeochemical characterization studies are regarded as an important method for determining the origin of hot springs. The major elements, trace elements, and stable isotopes of four groups of hot spring water samples and two groups of gas samples collected from the intersection of the Altyn Tagh fault zone and the East Kunlun fault belt were investigated in this study. The hot spring water temperature ranged between 6 °C and 14 °C. The water chemistry types of the hot springs were Na·Mg-Cl, Mg·Na-Cl·HCO₃, Na-Cl·SO₄, and Na-Cl·HCO₃. The δD values ranged from −50.00% to −68.60%, while the δ¹⁸O values ranged from −6.90% to −8.60%. The hot spring water was recharged mainly by infiltrating precipitation, with a recharge elevation of 3390~3676 m. The heat storage temperature ranged from 66.7 to 164.9 °C. The circulation depth was estimated to range between 1043 and 2679 m. The strontium isotopic composition of the water samples in response to the main weathering sources comprised carbonate and sulfate. CO₂ was the main component in the hot spring gas in the study region, and its content was over 95%. The ³He/⁴He-R/Ra relationship diagram revealed that the mantle-sourced helium from the Yitunbulake spring was 3.06%. In comparison, that from the Aiken spring was 7.38%, which indicated an intrusion of mantle-source material mixed into the hot springs in the study region. The crustal marine limestone contributed significantly to the carbon inventory of the hot spring gas samples (>75%). The dissolution of the marl aquifer resulted in the release of CO₂. Carbon was primarily obtained through metamorphism and hydrothermal reactions in the basement lithologies. The Yitunbulake and Aiken hot springs are found near the intersection of the Arjin and East Kunlun fractures, where the water–rock response is relatively strong and the depth of circulation and thermal storage temperature are both high. This causes relatively high ambient pressure to be released from the deep fluid, resulting in microseismic activity in this region. The continuous observation of Aiken spring water chemistry allows for the monitoring of fracture activity in the region. The results of the study could serve as a foundation for further exploration of the relationship between geothermal water and deep faults, shallow geological formations, hydrogeological conditions, and geothermal resource development in the region.

Studying the hydrogeochemical characteristics of hot springs provides essential geochemical information for monitoring earthquake precursors and understanding the relationship between fluids, fractures, and earthquakes. This paper investigates the hydrogeochemical characteristics of hot springs along the Tingri–Nyima Rift (TNR) in southern Tibet, a seismically active zone at the collision front of the Indian and Asian-European plates. The major elements, hydrogen, and oxygen isotopes of seven thermal springs were analyzed from July 2019 to September 2021. The findings indicate that Mount Everest’s meteoric water, which has a recharge elevation of roughly 7.5–8.4 km, is the main source of recharge for the hot springs. The water samples have two main hydrochemical types: HCO₃-Na and Cl-Na. The temperature of the geothermal reservoir is between 46.5 and 225.4 °C, while the circulation depth is between 1.2 and 5.0 km based on silica-enthalpy mixing models and traditional geothermometers. Furthermore, continuous measurements of major anions and cations at the Yundong Spring (T06) near Mount Everest reveal short-term (8 days) seismic precursor anomalies of hydrochemical compositions before an M_L4.7 earthquake 64.36 km away from T06. Our study suggests that seismicity in the northern section of the TNR is controlled by both hydrothermal activity and tectonic activity, while seismicity in the southern section is mainly influenced by tectonic activity. In addition to magnitude and distance from the epicenter, geological forces from deep, large fissures also affect how hot springs react to seismic occurrences. A fluid circulation model is established in order to explain the process of groundwater circulation migration. The continuous hydrochemical monitoring of hot springs near Everest is critical for studying the coupling between hot springs, fractures, and earthquakes, as well as monitoring information on earthquake precursory anomalies near Everest.

The 2020 Jiashi M6.4 earthquake occurred in the Kaping fold-and-thrust belt, a major south-verging active thin-skin system in the southwestern Chinese Tien Shan Mountain, north of the Tarim Basin. Within 50 km from the epicentral area, seismic hazard is high, as suggested by the occurrence of the 1902 Mw 7.7 Artux (Kashgar) earthquake and 1997 Jiashi strong earthquake swarm. The seismogenic structure responsible for the 2020 event is not well constrained and is a subject of debate. We relocated the 2020 Jiashi earthquake sequence and assessed the relocation uncertainties, using eight seismic velocity models and based on detailed local and regional subcrustal structures from seismic profiles. Then we compared the temporal variation in the Gutenberg–Richter b-values of the 2020 sequence with those of the 1997, 1998, and 2003 earthquake sequences. Our results show that most events cluster at depths greater than 10 km, suggesting that the events most likely occurred beneath the décollement and inside the Tarim Craton. The spatiotemporal evolution of the sequence suggests that two groups of structures at depth were involved in the 2020 sequences: NW–SE-trending lateral strike-slip faults and E–W-trending reverse faults. The b-values of the 2020 sequence exhibits relatively stable temporal evolution, unlike those of the multi-shock sequence that occurred inside the Tarim Craton. It indicates that the 2020 sequence perhaps was influenced by the stress interaction with the 10 km thick overlying strata. Our study provides a new perspective on the seismogenic structure of the earthquakes that occurred because of reactivation of ancient structures developed in a stable craton.

The exploration of multi‐layer coupling mechanisms between earthquakes and the ionosphere is crucial for utilizing ionospheric precursors in earthquake prediction. A significant research task involves continuously tracking the spatio‐temporal changes in ionospheric parameters, acquiring comprehensive seismic anomaly information, and capturing “deterministic” precursor anomalies. Based on data from the China Seismo‐Electromagnetic Satellite (CSES), we enhance the Pattern Informatics (PI) Method and propose an Improved Pattern Informatics (IPI) Method. The IPI method enables the calculation of the spatio‐temporal dynamics of electron density anomalies detected by the CSES satellite. The seismic signals in the electron density during earthquake on 2021 at Maduo are investigated in this work. The results show that: (a) Compared to original electron density images, the IPI method‐derived models extract distinct electron density anomaly signals, regardless of the data whether are collected during descending (daytime) or ascending (nighttime) orbits, or across different time scales of change window. (b) The electron density anomalies appear about 40 days prior to the Maduo Mw7.3 earthquake. The evolution of these anomalies follows a pattern of appearance, persistence, disappearance, re‐emergence, and final disappearance. Moreover, the evolution trends of the IPI hotspot images at daytime and nighttime are similar. These results suggest that the IPI method can capture the spatio‐temporal trends of ionospheric parameters and effectively extract electronic precursors related to strong earthquakes.