Defending Against Agroterrorism: Modeling Pathogen Dispersion Pathway

Home / Articles / Journal Article /

It is widely recognized that American farmers and ranchers make a major contribution to the U.S. economy by providing a safe and reliable food supply, which promotes energy security, job growth, and economic development. The U.S. Department of Agriculture (USDA) reports that in 2015, agriculture, food, and allied industries contributed more than $992 billion to the gross domestic product (GDP)—representing a 5.5 percent share of the total [1]. America’s farms alone contributed $136.7 billion in GDP [1]. Furthermore, the USDA indicates that the U.S. food and beverage manufacturing sectors together employ more than 1 percent (1.5 million people) of all non-farm employment [1]. An attack on American agricultural assets would threaten the security, safety, and availability of U.S. food supplies and cause significant economic damage to the U.S.

Response Planning

Agroterrorism is the deliberate introduction of one or more animal or plant diseases in order to generate fear, cause economic loss, and undermine social stability [2]. It is a subset of bioterrorism, a broader designation that generally refers to attacks aimed at human populations. Agroterrorism is attractive to some nation-states or groups because successfully disrupting a food supply can cause social as well as economic harm [3,4]. These concerns were exacerbated by the 9/11 terrorist attacks and subsequent anthrax attacks (which killed five and infected 17 others). Current federal strategies for preventing, detecting, and responding to agroterrorism and bioterrorism alike are informed by what is known about the Soviet Union’s biological warfare agency, Biopreparat (Биопрепарат), and the 1979 anthrax leak at Sverdlovsk (now Yekaterinburg), which infected at least 94 people and led to 64 fatalities [5]. The resulting U.S. agro-defense strategy, valid at the time, has undergone significant refinement and must continue to do so as scientific breakthroughs continue to occur in biotechnology, cyber, and the fast approaching quantum computing.

Fortunately, there has not yet been an actual agroterrorism event in the U.S., but other countries have experienced examples where state actors used biological weapons against agriculture [4]. Serious questions exist regarding how we can most efficiently and accurately detect and—most importantly—discriminate between naturally occurring and terror-inspired disease events. Timely detection of an agroterrorism event is critical if we are to protect food supply and consumer health. Moreover, understanding the scope and nature of an agroterrorism attempt soon after its incidence may provide an early warning of potential ramifications for human populations.

The book Agroterrorism: A Guide for First Responders by R. Withers [6] explains why livestock are a more likely target for terrorism than crops—“Developing biological weapons to destroy livestock is much easier than developing chemical weapons for the same purpose. There are far more highly contagious and lethal biological agents that can be used. … [and] the high-mobility breeding and rearing practices contribute to the ease of dissemination of a disease.” In line with the greater threat posed by animal-focused agroterrorism, we limit the discussion in this article to potential attacks on animal agriculture.

Potential Threats

Understanding the scope and nature of an agroterrorism attempt is important for protecting against the possibility of harm to human populations. “Because animals may be sentinels of disease in humans and many of the high-threat bio-agents…are zoonoses,” Withers writes, “it is possible that veterinarians might recognize an event in animals before it is recognized in humans [7].” Only moderate levels of expertise and fiscal investment are needed to weaponize certain biological materials, and that task is made even easier if a weaponeer has access to pathogens through veterinary and/or medical diagnostic laboratory systems. It is safe to surmise that any advanced bioweapons program developed by an adversary would pursue the use of biological weapons in multipronged attacks that target both civilian and military targets.

As RAND Corporation expert Bruce W. Bennett testified before the House Armed Services Committee in 2013, “While there is evidence of North Korean biological weapons, little is known with certainty about the biological weapon agents the North has developed, which of these agents it has weaponized, and how it would use them [8].”

A recent unclassified report by scholars at Harvard contends that North Korea may already possess as many as 13 potential biological warfare agents, several of which could be pathogenic to both animals and humans [9]. As history has shown, biological weapons programs can be exceedingly difficult to detect. As Bennett continued, “Biological weapon programs are easier to hide than most military programs because they can be developed in a university setting or hidden within efforts to develop related vaccines [8].”

Given the ease with which biological weapons programs can be obscured from surveillance, it is imperative that the U.S. bolster its capabilities for detection, containment, and remediation of the agriculture domain, given that the confirmation that a hostile bioweapons program exists may only occur at the point of weapon delivery. In any agroterrorism event, delay of any type exponentially increases the scope and ultimate effects of the event.

Identifying Intentional Versus Natural Infections

The Department of Homeland Security has declared the Food and Agriculture industrial sector, which accounts for approximately one-fifth of all economic activity in the U.S., to be critical infrastructure and thus a necessary element of national security. The animal production industry is a network—both dispersed and concentrated—composed of complex production, processing, and delivery systems [4].

The mobility of animals and their by-products across jurisdictional boundaries (e.g., states, counties, nations, etc.) creates opportunities for uncontrolled natural and terrorist-planned outbreaks [10]. Such incidents could bring significant physical and economic harm. For example, in 2011, an outbreak of foot-and-mouth disease in the U.K. caused an estimated $25–30 billion in losses [11]. The 2014–15 outbreak of avian influenza (originating from wild birds) resulted in an estimated $3.3 billion in losses to the U.S. economy—the single largest animal health disaster in the United States [12].

Compounding the difficulties of detecting an incident, agroterrorism defense programs must consider the potential for contaminants to persist on farm grounds, facilities, and processing and transport equipment. It took several years to fully decontaminate the congressional office buildings and U.S. Postal Service sorting facilities where anthrax spores were spread in 2001, and those facilities were out of service until decontamination was completed.

An agroterror agent could similarly spread throughout the animal production chain for an indeterminate length of time, requiring ample resources for recommissioning. Differentiating an intentional disease outbreak from a naturally-occurring one remains challenging [13], and scholarly works have covered this topic extensively [14,15]. Time plays a critical role in making this determination. Many of the downstream impacts of mortality events, whether natural or terrorism, remain the same. The presentation of animal mortality can and has been confusing when deaths could be linked to a natural event just as easily as to an infection.

A classic example of unclear mortality presentation is anthrax, whose spores can lie dormant in the soil for decades and then produce sudden high-mortality outbreaks in unvaccinated, healthy animals. A July 2016 anthrax outbreak in the Yamalo-Nenets region of Siberia was Russia’s first in 75 years, causing the deaths of thousands of reindeer and displacing dozens of traditional reindeer-herding families [16]. Initial reports on the event postulated that lightning strikes from a single storm event were the cause. Only after reindeer continued to die was it correctly identified as a natural anthrax outbreak. Ultimately, as many as 250,000 reindeer were destroyed by the end of the year as a regional animal depopulation measure [17].

It is reasonable to have concluded that the reindeer deaths were the result of lightning strikes. Indeed, shortly after the Yamalo-Nenets event, a separate lightning event killed 300 reindeer in Norway. The obvious question that follows is, what would this delay have meant in an actual agroterrorism event?

The Regulatory Context of Detection

The USDA has delegated its authority to prevent, detect, and respond to agroterrorism to the Office of Homeland Security and Emergency Coordination (OHSEC). Responsibility for complying with Homeland Security Presidential Directive/HSPD–9— Defense of United States Agriculture and Food also falls to OHSEC. This directive lays out national priorities for the defense of agriculture and food systems against terrorist attacks, major disasters, and other food supply emergencies. However, a recent audit report completed in March 2017 by the department’s Office of the Inspector General (OIG) concluded that OHSEC’s planning efforts for agroterrorism prevention, detection, and response were deficient. At present, the OIG concluded, OHSEC does not present an integrated and actionable statement of critical needs [18].

The functions necessary to do this correctly include intelligence analysis, law enforcement, animal health, plant health, public health, environmental remediation, and outbreak response and recovery. The 2008 Food and Agriculture Incident Annex (FAIA) to the National Response Framework, which addresses only the response and recovery element of agrodefense, lists USDA and HHS as Coordinating Agencies and a suite of strategic supporting agency partners. The update to the FAIA (expected in 2018) will provide additional specifics to the span of responsibility, along with blueprints for coordination.

Outbreaks first present themselves at the farm or facility level, so that USDA’s National Animal Health Emergency Response Plan (NAHERP) assumes that detection of an animal disease takes place at the most local level [6]. Initial detection is therefore highly dependent on farming staff recognizing an unusual clinical presentation or a spike in cases, and thereafter reacting properly and in a timely manner. Ideally, these observations would then be immediately reported to a qualified veterinarian. In the United States, veterinarians who work with livestock herds must be accredited by the USDA and must be trained to recognize the clinical syndromes of Foreign Animal Diseases of concern, thereby increasing the likelihood of rapid recognition that an animal disease event has occurred. If the index of suspicion of an outbreak is high, the State Veterinarian’s Office is then alerted.

Response and Containment: Modeling the Host Environment

Per NAHERP, should the presence of a foreign animal disease be confirmed, the premises will be placed under a quarantine order and a “movement hold” of all susceptible and affected animals will be established for a minimum of 72 hours [6]. The USDA Animal and Plant Health Inspection Service (APHIS) directs the response (containment and remediation), while local law enforcement establishes control zones and controls access. The FBI and USDA OIG oversees documentation, evidence collection, and chain-of-custody procedures with support from local/state law enforcement agencies [2,6,19].

When an outbreak is determined to be verified and a coordinated response is mounted, one challenge is the lack of environmental modeling tools, which we describe below. Necessary planning components include the key area of characterization, risk determination, potential course of action, and a means of assessing the value of these measures to health. Outside the realm of biosurveillance, ecologists and epidemiologists have long been using environmental models to predict the transport, fate, and effects of contaminants on ecological receptors and on humans [20]. The transport and fate of materials released from natural disasters or anthropogenic events vary—as do their deleterious impacts to all life—by the biogeochemical processes of the host ecosystem. Therefore, regardless of why or how pathogens are released, both preparedness and response should be informed by those biogeochemical processes [21].

Rapid diagnostics, including patient-side diagnostics, may arguably be the most important element of an animal disease stockpile. The National Veterinary Stockpile, administered by USDA’s APHIS, lacks both therapeutics and rapid diagnostics. Current biosensors for the detection of pathogens in food, livestock, and agricultural products include antibody- or immunologically- based tests and polymerase chain reaction (PCR) assays. Antibody tests have an advantage over PCR testing when it comes to speed and cost—little to no sample preparation is required to conduct an antibody test. However, compared to a PCR assay, an antibody test has a relatively low sensitivity and specificity, and it typically requires laboratory-based PCR testing for confirmation. PCR testing, considered the gold standard, provides high-confidence results with excellent sensitivity and specificity, although a PCR test requires a rather rigorous sample preparation. The ability to quickly deploy a user-friendly diagnostic capability is essential to successfully contain an agroterrorism event.

One purpose of the Food Systems Institute at Auburn University is to provide meaningful preparedness and response tools based on risks to livestock, starting with poultry, beef, and pork, which are consumed by large segments of the human population. Epidemiological risk assessments have been conducted on historical animal disease events, but they lack comprehensive risk analysis and do not incorporate a broader picture to include ecological complexity along with human health.

Since 1900, there have been at least 15 documented cases involving the use of pathogenic agents/substrates to contaminate foods [2,22]. We analyzed example events from the last 100 years to inform both models and algorithms. We have further calibrated the resulting tools based on our event analysis. These models need appropriate parameters that illustrate the physiological landscape over which the events occur, including the routes that could lead to exposure of livestock.

With this approach, we can begin to predict a temporal likelihood for the incidence of infections and public health disasters, and optimize our surveillance and identification methodologies depending on data inputs such as the food source species and transport/processing routes leading to consumers, thereby producing a range of potential outcomes to inform public health and disaster response officials.

Data for Agroterrorism Event Modeling

Predictive models used in the forecasting of pathogen pathways are commonly based on a conceptual framework that considers the interactions among pathogens, hosts, and environmental conditions. Knowledge of the survivability of a given pathogen and an accurate characterization of the environment (ecosystem features, current weather conditions, etc.) are fundamental elements in forecasting dispersion patterns [11]. A variety of means exist for characterizing the natural landscape. For our purposes, ecoregions, used for characterizing the environment over large areas with similar physical and biological components, provide an appropriate and targeted representation of the landscape [23]. They are classified into hierarchical spatial groups that nest over large to small scales by analyzing the patterns and composition of a given area’s geology, landforms, soils, vegetation, climate, land use, wildlife, and hydrology [24].

In our analysis, the highest resolution ecoregion level (IV) of the United States will be used to characterize each of the 15 documented case environments as an initial step in determining the environmental parameters that are favorable to pathogen survivability, which in turn can be used to assess the temporal and spatial risk of pathogen dispersal.

This modeling exercise requires a geographic tie, and spatial analytical tools within geographic information systems (GIS) need to (and should be) used as a platform for epidemic spread models [25,26]. Along with ecoregions, there are a variety of geospatial datasets available that support emergency management and ecological modeling which are applicable to our analyses.

Such GIS data includes land ownership, transportation, hydrology, elevation, the distribution of natural hazard zones, environmentally sensitive areas, land cover, animal species, climate variables, and soils, among others. Global food distribution models such as the Gridded Livestock of the World, v2 might be mined to inform the resource flows within our model.

Linking these datasets will ultimately aid in forecasting potential pathways, and also provide a model that can be used to simulate future scenarios under different pressures, such as increasing population, regulatory changes, climate change, natural disasters, etc. The utility of a successful comprehensive predictive model is to pinpoint a release origin, determine factors that influence origin type (natural or deliberate), and perhaps most importantly, to identify where response actions may be taken in the appropriate time frame. Armed with such a model, the ultimate product is risk reduction and preservation of human lives.


1. USDA (2017 October). Retrieved from

2. Olson, D. (2012). Agroterrorism – Threats to America’s economy and food supply. FBI Law Enforcement Bulletin. Retrieved from

3. Yeh, J.Y., Lee, J.H., Park, J.Y, Cho, Y.S., & Cho, I.S.. (2013). Countering the livestock-targeted bioterrorism threat and responding with an animal health safeguarding system. Transboundary and Emerging Diseases, 60(4), 289-297.

4. Monke, J. (2007). CRS report for Congress: Agroterrorism: Threats and preparedness. Order Code RL32521. Retrieved from

5. Meselson, M., Guillemin, J., Hugh-Jones, M., Langmuir, A., Popova, I., Shelokov, A., & Yampolskaya, O.. (1994, Nov. 18). Science, 18, 266(5188), 1202-8.

6. Moats, J. (2007). Agroterrorism: A guide for first responders. College Station, TX: Texas A&M University Press.

7. Withers, R. (Ed.). (2014 September). USAMRIID’S medical management of biological casualties handbook (Eighth Edition), 11-12. United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, Maryland: USAMRIID.

8. The challenge of North Korean biological weapons, Committee on Armed Services Subcommittee on Intelligence, Emerging Threats and Capabilities, United States House of Representatives (2013, Oct. 11) (testimony of Bruce W. Bennett, The RAND Corporation). Retrieved from https://www. CT400/CT401/RAND_CT401.pdf

9. Kim, H-K., Philipp, E., & Chung, H. (2017). North Korea’s biological weapons program: The known and unknown (Rep.). Retrieved

10. Yeh, J., Park, J., Cho, Y. S., & Cho, I. (2012). Animal Biowarfare Research: Historical Perspective and Potential Future Attacks. Zoonoses and Public Health, 59(8), 536-544. doi:10.1111/j.1863-2378.2012.01505.x

11. Manuel, F. Z. (2017). Agro terrorism: A global perspective. Journal of Political Sciences & Public Affairs, 5, 262. doi:10.4172/2332-0761.1000262

12. Blue Ribbon Study Panel on Biodefense. (2017 October). Special focus: Defense of animal agriculture. Blue Ribbon Study Panel on Biodefense: Washington, DC. Retrieved from

13. Federal Bureau of Investigation, Counterterrorism Division, Food and Drug Administration, Office of Criminal Investigations, Food and Drug Administration, Center for Food Safety and Applied Nutrition, Department of Homeland Security, & U.S. Department of Agriculture, Office of the Inspector General. (2008). Criminal Investigation Handbook for Agroterrorism (p. 34). Retrieved from

14. Hugh-Jones, M. L., & Brown, C. C. (2006, April). Accidental and intentional animal disease outbreaks: Assessing the risk and preparing an effective response. Revue Scientifique Et Technique, 25(1), 21-33.

15. Gilpen, J. L., Carabin, H., Regens, J. L., & Burden, R. W. (2009). Agriculture emergencies: A primer for first responders. Biosecurity and bioterrorism: Biodefense Strategy, Practice, and Science, 7(2), 187–198. doi:10.1089/bsp.2008.0037

16. Samuelson, K. (2016, Aug. 5). Russia hada fatal anthrax outbreak. Could it happen in the U.S.? Time. Retrieved from

17. Guarino, B. (2016, Nov. 16). Starvation killed 80,000 reindeer after unusual arctic rains cut off the animals’ food supply. The Washington Post. Retrieved from

18. USDA. (2017 March). Agroterrorism prevention, detection, and response. Retrieved from

19. Federal Bureau of Investigation, Counterterrorism Division, Food and Drug Administration, Office of Criminal Investigations, Food and Drug Administration, Center for Food Safety and Applied Nutrition, Department of Homeland Security, & U.S. Department of Agriculture, Office of the Inspector General. (2008). Criminal Investigation Handbook for Agroterrorism. Retrieved from

20. Suter, G.W., Vermeire, T., Munnsjr, W., & Sekizawa, J. (2005). An integrated framework for health and ecological risk assessment. Toxicology and Applied Pharmacology, 207(2), 611-616. doi:10.1016/j.taap.2005.01.051

21. Till, J. E., & Grogan, H. A. (2008). Radiological risk assessment and environmental analysis. Oxford: Oxford University Press. doi:10.1093/acprof:oso/9780195127270.001.0001

22. Centers for Disease Control and Prevention. (2017, Oct. 6). List of selected multistate foodborne outbreak investigations. Retrieved from

23. Bailey, R. G., Avers, P. E., King, T., & McNab, W. H. (1994). Ecoregions and subregions of the United States (map). Washington, DC: USDA Forest Service. 1:7,500,000. With supplementary table of map unit descriptions, compiled and edited by W.H. McNab and R.G. Bailey.

24. Omernik, J. M., & Griffith, G. E.. (2014). Ecoregions of the conterminous United States: Evolution of a hierarchical spatial framework. Environmental Management, 54(6), 1249-1266. doi:10.1007/s00267-014-0364-1

25. Cromley, E. K. (2003). GIS and disease. Annual Review of Public Health 24(1), 7-24. doi:10.1146/annurev. publhealth.24.012902.141019

26. Perez, L., & Dragicevic, S. (2009). An agentbased approach for modeling dynamics of contagious disease spread. International Journal of Health Geographics, 8(1), 50. doi:10.1186/1476-072x-8-50