Pre-earthquake geochemical anomalies in spring waters: site-dependent responses from western Türkiye

Understanding the processes occurring before earthquakes constitutes one of the fundamental challenges in earthquake sciences due to the complex and heterogeneous nature of crustal deformation in the earthquake preparation stage. Many geochemical worldwide observations have indicated that precursory anomalies prior to earthquakes may exist. Hydrogeochemical changes observed in spring waters are regarded as sensitive indicators of subsurface processes associated with stress accumulation and fluid migration in seismically active regions. This study aims to investigate hydrogeochemical anomalies in spring waters preceding earthquakes by distinguishing tectonically driven signals from seasonal and meteorological effects using rainfall data. Within this framework, hydrogeochemical parameters covering periods of 9 to 15 months were obtained from commercially bottled water samples from the Marmara Region and the Aegean Extensional Province (AEP) in western Türkiye. In this study, two distinct moderate (Mw ~5.0) earthquakes on the North Anatolian Fault Zone and one Mw 5.0 on the submerged section of the Menderes Fault Zone in the AEP were monitored using data from five distinct spring water sites. To this end, collected spring waters’ electrical conductivity (EC) and major ion concentrations (Cl⁻, SO₄²⁻, Na⁺, K⁺, Mg²⁺, Ca²⁺) have been measured.  The results suggest that anomalies exist at least 30 days in two spring waters before moderate earthquakes. However, no reliable anomaly was observed at some other spring water sites. This variability may be caused by the distance of these stations from fault zones and/or these spring waters are located on different tectonic blocks where pre-earthquake strain accumulation could not be transferred effectively to the adjacent blocks. Therefore, the effectiveness of hydrogeochemical monitoring appears to be strongly controlled by the structural connectivity between the spring system and the earthquake epicentre. These observations illuminate the importance of examining the relationship between hydrogeochemical variations observed in spring waters and pre-earthquake processes for their potential use as earthquake precursors. These findings suggest that future monitoring strategies incorporating denser spatial coverage, longer observation periods, and multi-parameter datasets are essential to better constrain the role of fault proximity and tectonic block configuration in controlling the detectability of pre-earthquake hydrogeochemical anomalies.