Response to Zöller et al.'s critique on “Potential short-term earthquake forecasting by farm-animal monitoring”
Summary
WIVVI summary (paraphrased): The authors respond to the statistical critique of their farm-animal monitoring study. They acknowledge that only one animal group, one earthquake sequence, and one region were observed, and state clearly that earthquake predictions are not possible—and should not be attempted—from this dataset. They nevertheless argue that the collective animal-activity data may contain information relevant to future local earthquake forecasting when combined with conventional geophysical observations. They propose that future research integrate rigorous forecasting-evaluation methods with longer-term, multisite animal-monitoring data.
Source
Ethology, Volume 127, Issue 3, pages 307–308
Keywords
animal behaviour, earthquake forecasting, earthquake precursors, farm animals, biologging, collective animal activity, forecast evaluation, multisite monitoring, geophysical observations, statistical critique
Source Type
response, peer reviewed, commentary
Key Findings
The authors stated that:
• The original study observed one group of farm animals during one earthquake sequence in one region.
• Earthquake prediction is not possible—and should not be attempted—from this dataset.
• They nevertheless maintain that collective animal-activity data may contain information relevant to future local forecasting when combined with conventional geophysical observations.
• They argue that future studies should combine rigorous forecasting-performance evaluation with appropriately designed animal-monitoring programmes.
These statements describe a research hypothesis and proposed future direction, not an established forecasting capability.
Limitations
• This response presents no new experimental data or independent validation.
• It relies on the same single-site, single-animal-group dataset examined in the original paper and critique.
• The proposed value of combining animal activity with conventional geophysical observations remains hypothetical.
• The response does not establish prospective forecasting thresholds, false-alarm rates, missed-event rates, or performance on unseen data.
• Independent, preregistered, multisite studies with predefined alerts and out-of-sample evaluation are required before forecasting utility can be assessed.
Methods Summary
This paper is a methodological response rather than a new observational study. The authors address the statistical critique, clarify the intended scope of their original conclusions, acknowledge the dataset’s limits, and distinguish prediction from exploratory evidence that might inform future forecasting research. They propose integrating animal-monitoring data with conventional geophysical measurements and rigorous forecast-verification methods in future studies.