Retrospective studies of unusual animal behavior as an earthquake predictor.
Summary
WIVVI summary (paraphrased): Researchers used a standardized retrospective interview method to examine reports of unusual animal behavior before seven earthquakes in North, Central, and South America. For two earthquakes, interviews conducted at greater distances supplied an estimate of the background rate at which unusual behavior would be reported without a presumed local precursor. Relative to this baseline, one earthquake was preceded by a statistically significant increase in reported unusual animal behavior, whereas four were not. The mixed results suggest that unusual behavior may have occurred before at least one event but was neither consistent across earthquakes nor sufficient to establish a reliable forecasting method.
Source
Geophysical Research Letters
Keywords
unusual animal behavior, earthquake prediction, earthquake precursors, retrospective interviews, baseline comparison, reporting frequency, behavioral observations, seismic events, Americas, observational study
Source Type
research article
Key Findings
Across seven studied earthquakes, only one showed a statistically significant increase in reports of unusual animal behavior relative to the estimated baseline, while four did not. This provides mixed rather than uniformly positive evidence. It shows that elevated reporting can occur before a particular earthquake but also that most examined events did not display the same pattern. The findings therefore do not support unusual animal behavior as a consistent or reliable standalone earthquake predictor.
Limitations
The evidence was collected after the earthquakes through interviews and was therefore vulnerable to recall bias, selective reporting, altered memories, media influence, and awareness of the event. Baseline rates were available from more distant comparison areas for only two earthquakes, limiting consistent control across all seven events. “Unusual behavior” encompassed varied species and behaviors rather than one predefined biological signal. The sample contained few earthquakes, and negative observations or instances of unusual behavior not followed by earthquakes could not be captured prospectively. The study could identify differences in reporting frequency but could not determine a mechanism, establish causality, or validate operational forecasting performance.
Methods Summary
Researchers applied a standardized retrospective interview schedule after seven earthquakes in North, Central, and South America to collect reports of animal behavior before each event. The interview approach was designed to improve consistency and reduce some reliability problems common to anecdotal accounts. For two earthquakes, reports collected at locations farther from the affected area were used to estimate a baseline frequency of unusual-animal-behavior reports. Frequencies near the earthquakes were then compared with this baseline to determine whether reported unusual behavior increased significantly before individual events.