Anomalous fluid emissions at the Chung-lun mud pool, southwestern Taiwan: evidence for seismic–pore fluid interactions in a thrust fault system

Abstract The relationship between earthquakes and fluid pressure variations remains debated, particularly regarding whether surface fluid anomalies reflect pore pressure changes at seismogenic depth. Anomalous fluid emissions at the Chung-lun mud pool in southwestern Taiwan were continuously monitored between 2008 and 2010, with quantitative analyses focusing on periods of stable observations from mid-2009 to late 2010 using water level sensors and a digital camera system. Long-term anomalies were characterized by sustained variations in water levels and changes in degassing behavior that were independent of precipitation. In four out of five cases, these anomalies occurred in close temporal association with nearby earthquakes, indicating a systematic temporal relationship rather than a simple random coincidence. Short-term variations were linked primarily to rainfall events, while long-term discharge anomalies are interpreted as being consistent with pore pressure perturbations at depth, based on indirect surface observations and the exclusion of meteorological and shallow hydrological controls, rather than direct measurements within the seismogenic zone. The results hypothesize that pore fluid overpressure at depth can modulate both fault stability and surface fluid discharge, consistent with the concept of fault valve behavior. Some anomalies were recorded prior to seismic events, suggesting that gradual increases in pore pressure may contribute to stress evolution in critically stressed thrust faults, although the coupling is not deterministic. The integration of instrumental and photographic observations at Chung-lun demonstrates the potential of mud pools as natural observatories for geodynamically induced interactions between tectonic activity and fluid transport and emission. It highlights their important role as valuable components in regional earthquake monitoring networks. Graphical Abstract