By Serena Cheung

– Executive Summary –

The escalating unpredictability and severity of natural disasters, compounded by the effects of climate change, underscore the critical need for innovative forecasting methods. This urgency is highlighted by the record-breaking 28 weather and climate disasters in the United States in 2023—each causing well over $1 billion in damages—marking a sharp rise from the historical yearly average of 8.5 billion-dollar events from 1980 to 2023.[1] Earthquakes and tropical cyclones are the world’s most perilous types of natural hazards, posing the greatest risks to life and property.[2] However, forecasts often fall short in predicting these natural disasters, despite advancements in traditional forecasting techniques, resulting in significant fatalities and extensive damage.

Animals seem to exhibit conspicuous behavioral changes that enhance their survival in anticipation of natural disasters, serving as potential “bioindicators” of environmental changes, an ecosystem’s health, and impacts to humans.[3] Guided by the research question—How can scientific research on changes in animal behavior before natural disasters be leveraged to improve the accuracy of forecasts?—this thesis employs a multidisciplinary approach, weaving together insights from zoology, behavioral ecology, meteorology, and seismology. To evaluate the potential benefits of including adaptive animal behavior into disaster forecasting, a foundational understanding of the complexities of tropical cyclones and earthquakes as well as existing forecasting techniques is essential. This analysis serves to bridge the gap between traditional forecasting methods and the possibility of animal-based forecasting approaches, offering a new perspective on disaster forecasting. The thesis systematically reviews existing studies on anomalous animal behaviors before natural disasters, illustrating animals’ forecasting potential for tropical cyclones and earthquakes.

This thesis uncovered three main findings that contribute to the understanding of the potential impact of adaptive animal behavior on disaster forecasting. First, the research unearthed compelling evidence that various avian, terrestrial, and aquatic animals demonstrate observable behavioral adaptations in response to pre-disaster environmental cues. These changes occur at varying times before disasters, in multiple locations, and involve a range of anomalous behaviors. For tropical cyclones, studies have highlighted changes in the behavior of migratory birds, coastal sharks, and oceanic fish observed hours to months before impending storms.[4] For earthquakes, anomalous behaviors of snakes, farm animals (e.g., cows, sheep, and dogs), ants, and rodents have been observed hours to several weeks in advance of seismic events.[5] These behavioral responses not only ensure the survival of these species but could also provide sufficient advance notice to enhance the accuracy of forecast lead times.

The second finding is that adaptive animal behavior before the onset of natural disasters has the potential to be a “diagnostic precursor”—defined by Thomas H. Jordan et al. as “some kind of signal observable before [disasters] that indicates with high probability the location, time, and magnitude of an impending event”—of tropical cyclones and earthquakes.[6] Moreover, studying multiple species simultaneously may prove most effective, as the research indicates that the signals from single species may not consistently show a significant statistical correlation with environmental conditions before disasters. Therefore, aggregating bioindicators from several species into a collective analysis could reduce environmental and biological variability and noise, potentially creating a more reliable and predictive model for forecasting natural disasters.

The third finding is that the growing body of evidence of adaptive animal behavior before natural disasters warrants further investigation. Future research should focus on confirming these preliminary observations in a methodical manner and creating a comprehensive and varied data set of bioindicators. Such efforts may reveal behavioral patterns that can enhance current probabilistic forecasts and integrate into an ensemble model, which is a composite of multiple individual forecasts.[7] By supplementing current forecast models with adaptive animal behavior data, the accuracy and reliability of disaster forecasts could improve. Through this field of research, scientists could also discover new meteorological and seismic variables that offer deeper insights into pre-disaster conditions and might even lead to biomimetic or bioinspired solutions.[8]

The integration of adaptive animal behavior as a reliable component in disaster forecasting faces several challenges, including skepticism from scientific authorities and the public, technical hurdles in advancing research, financial constraints, and ethical concerns. To navigate these obstacles, the thesis evaluates the feasibility of this area of study using Rogers’s five factors of adoption/diffusion—relative advantage, compatibility, complexity, trialability, and observability.[9] This framework critically assesses how adaptive animal behavior indicators can be effectively incorporated into disaster forecast models. By employing rigorous scientific methods and advanced computational technologies, integrating adaptive animal behavior could represent a viable approach to enhancing natural disaster forecasting.

Adaptive animal behavior may serve as the contemporary equivalent of a canary in the coal mine, conspicuously sounding the alarm for the onset of natural disasters through instinctual biological responses. This novel approach could transform disaster forecasting by strengthening the United States’ preparedness against these natural threats. Based on the findings, this thesis proposes the following five recommendations:

1. Expand and strengthen research on pre-disaster adaptive animal behavior through a systematic methodology that incorporates multi-year studies of various species across several natural disaster events.
2. Employ advanced computational technologies, such as artificial intelligence and machine-learning algorithms, to refine the data analysis of bioindicators, filter out noisy data and false positives, and identify animal behaviors that signal upcoming natural disasters.
3. Secure reliable research funding through high-quality proposals that emphasize the unique advantages of studying adaptive animal behavior—not only enhancing disaster forecasting and preparedness but also raising awareness of ecological and animal conservation issues and the effects of climate change on ecosystems.
4. Foster international and interdisciplinary collaboration with diverse partners to study adaptive animal behavior as a precursor to natural disasters and potentially establish dedicated monitoring centers.
5. Prioritize ethical and environmental integrity in research to ensure the humane treatment of animals and address the wider ecological impacts of natural disasters.

---

[1] “Billion-Dollar Weather and Climate Disasters,” National Centers for Environmental Information, accessed January 16, 2024, https://www.ncei.noaa.gov/‌access/billions/; Department of Homeland Security, National Preparedness Report (Washington, DC: Department of Homeland Security, 2021), 68, https://www.hsdl.org/c/abstract/?docid=862169. This average reflects damages adjusted to 2023 dollars using the consumer price index.

[2] “Tropical Cyclone,” World Meteorological Organization, December 16, 2022, https://wmo.int/topics/‌tropical-cyclone.

[3] Trishala K. Parmar, Deepak Rawtani, and Y. K. Agrawal, “Bioindicators: The Natural Indicator of Environmental Pollution,” Frontiers in Life Science 9, no. 2 (2016): 110, https://doi.org/10.1080/21553769.‌2016.1162753.

[4] Christopher M. Heckscher, “A Nearctic-Neotropical Migratory Songbird’s Nesting Phenology and Clutch Size Are Predictors of Accumulated Cyclone Energy,” Scientific Reports 8 (July 2018), https://doi.‌org/10.1038/s41598-018-28302-3; Bradley A. Strickland et al., “Movements of Juvenile Bull Sharks in Response to a Major Hurricane within a Tropical Estuarine Nursery Area,” Estuaries and Coasts 43, no. 5 (July 2019): 1144–57, https://doi.org/10.1007/s12237-019-00600-7; Nathan M. Bacheler et al., “Tropical Storms Influence the Movement Behavior of a Demersal Oceanic Fish Species,” Scientific Reports 9 (2019): 1481, https://doi.org/10.1038/‌s41598-018-37527-1.

[5] Kelin Wang et al., “Predicting the 1975 Haicheng Earthquake,” Bulletin of the Seismological Society of America 96, no. 3 (June 2006): 757–64, https://doi.org/10.1785/0120050191; Martin Wikelski et al., “Potential Short-Term Earthquake Forecasting by Farm Animal Monitoring,” Ethology 126, no. 9 (2020): 931–41, https://doi.org/10.1111/‌eth.‌13078; Gabriele Berberich et al., “Early Results of Three-Year Monitoring of Red Wood Ants’ Behavioral Changes and Their Possible Correlation with Earthquake Events,” Animals 3, no. 1 (February 2013): 63–84, https://doi.org/10.3390/‌ani3010063; Yonghong Li et al., “Behavioral Change Related to Wenchuan Devastating Earthquake in Mice,” Bioelectromagnetics 30, no. 8 (December 2009): 613–20, https://doi.org/10.1002/bem.20520.

[6] Thomas H. Jordan et al., “Operational Earthquake Forecasting: State of Knowledge and Guidelines for Utilization,” Annals of Geophysics 54, no. 4 (2011): 319, https://doi.org/10.4401/ag-5350.

[7] “NHC Track and Intensity Models,” National Hurricane Center and Central Pacific Hurricane Center, June 11, 2019, https://www.nhc.noaa.gov/modelsummary.shtml.

[8] Biomimetics is the science of mimicking nature for synthetic design while bioinspiration applies ideas from nature to innovations. See Francois Barthelat, “Biomimetics for Next Generation Materials,” Philosophical Transactions of the Royal Society 365, no. 1861 (September 2007): 2907, https://www.‌jstor.‌org/stable/25190632; Fermanian Business & Economic Institute, Bioinpiration: An Economic Progress Report (San Diego: San Diego Zoo Global, 2013), 5, https://cnnespanol.cnn.com/wp-content/uploads/‌2014/‌05/bioreport13.final.sm.pdf.

[9] Mithun Sridharan, Puneet Maheshwari, and Pallavi Mundhada, “Rogers’ Five Factors—How to Appraise Innovations?,” Think Insights, December 7, 2021, https://thinkinsights.net/strategy/rogers-five-factors/.