Key Findings

The review compiled published reports of proposed earthquake precursors, including changes in seismicity, ground deformation, groundwater, gas emissions, temperature, and electric or magnetic fields.

The authors found that reported signals varied considerably in timing, duration, location, magnitude, and reliability. Some observations appeared repeatedly and had plausible physical explanations, but the available evidence did not establish any single precursor as a reliable, generally applicable earthquake-prediction method.

This paper catalogues reported observations and possible mechanisms; it does not validate dependable earthquake prediction.

Limitations

The review draws on heterogeneous reports that used different instruments, thresholds, locations, time windows, and definitions of a precursor. Many observations were retrospective, lacked consistent controls, or had not been independently replicated.

Publication bias and incomplete reporting of null results may also affect the literature. Apparent anomalies can have non-seismic explanations, and a signal observed before one earthquake may not generalize to other events or regions.

The review therefore supports continued investigation of proposed precursors, but not operational earthquake prediction without prospective, independently replicated testing that measures false alarms and missed events.

Methods Summary

This paper used a structured survey of published scientific literature rather than collecting new experimental data. The authors examined six broad classes of proposed earthquake precursors: electric and magnetic fields, gas emissions, groundwater-level changes, temperature changes, surface deformation, and seismicity.

For each class, they catalogued reported characteristics including spatial extent, timing relative to the earthquake, duration, amplitude, and signal-to-noise ratio. They compared the reported anomalies with earthquake magnitude, distance from the eventual epicentre, and time before the main shock, and reviewed proposed physical explanations.

The analysis was comparative and descriptive. It looked for broad patterns across varied published reports; it did not conduct a prospective prediction trial or independently replicate the underlying observations.