No, not reliably: dogs may sometimes react a few seconds to a few minutes before an earthquake is felt by humans, but no scientific study has demonstrated a usable predictive ability, and most "alerts" reported are based on retrospective accounts rather than proof.
From P-waves to the electrical hypothesis: what dogs might perceive
An earthquake first releases P-waves (primary waves), which are faster but less destructive, arriving a few seconds before S-waves and surface waves responsible for damage. A dog's hearing, which can perceive frequencies up to about 45,000 Hz compared to 20,000 Hz in humans, along with vibrations transmitted through the ground via its paw pads, could theoretically allow it to detect these early micro-tremors before humans feel them.
Other hypotheses mention sensitivity to atmospheric pressure changes, radon gas emissions, or disturbances in Earth's electromagnetic field that might accompany rock fracturing at depth. These phenomena, if they exist, remain poorly characterized and their causal link to animal behavior has never been established reproducibly under controlled conditions.
Publicité
Publicité
The major methodological problem is confirmation bias: we remember dogs acting restless before an earthquake, but not the countless episodes of canine restlessness without an associated earthquake, nor earthquakes that occurred with no prior animal sign. No robust statistical correlation has been able to be established across large datasets.
Why no animal alert system has been validated
Several research programs, particularly in Japan and the United States, have attempted to quantify this phenomenon by equipping animals with motion sensors, without usable results to date. Any possible advance warning would also be too brief, often less than a minute, to constitute a useful alert on a human timescale. Current early warning systems rely exclusively on seismic networks, not on animal observation.
