For both business and wildlife conservation, success can mean doing more with less.
A novel decision model derived from business operations detects emerging wildlife disease months earlier, or with lower costs, than the current traditional strategies, according to a collaborative study published Sept. 15 in the Proceedings of the National Academy of Sciences.
Researchers at the College of Veterinary Medicine (CVM) and Smith School of Business at Queen’s University, Canada, developed the model, which could detect a chronic wasting disease incursion in wild New York state white-tailed deer 8.4 months earlier than current methods in a case study.
“This is a crucial benefit, as wildlife populations are seeing increased risk from emerging diseases just as the agencies that manage them are experiencing ever-tightening budgets,” said co-lead author Krysten Schuler, wildlife disease ecologist and associate research professor at CVM.
Typically, wildlife-management agencies focus on reactive management strategies, with resources allocated only after the threat is detected. However, “by the time the first case is detected, the disease or species may have been spreading unnoticed for an extended period,” Schuler said.
But proactivity is often expensive, presenting a dilemma for underfunded management groups. Wildlife managers must make careful calculations to enact costly prevention activities, with balancing surveillance activities that are expensive initially but become cheaper over time.
Schuler’s team collaborated with co-lead author Jue Wang, associate professor of management science at Queen’s University, who was a visiting scholar at the Cornell Johnson Graduate School of Management at the time of the study. Wang devised a new AI-driven optimization model that allocates a given surveillance budget across many geographical sites to minimize unnoticed disease spread before the first case is detected.
Using chronic wasting disease (CWD) in New York state as a case study, the researchers tested the model with a theoretical budget of $500,000 across all 62 counties in the state for 10 years. When they compared the outcomes of this optimized AI model with the one currently being used by the New York State Department of Environmental Conservation (co-authors on the paper), in New York state, they found that the model was able to detect CWD in deer on average 8.4 months earlier than the traditional models, or reduce the current spending by 22% without compromising the performance.
“The biggest challenge in tackling emerging diseases is that they can appear almost anywhere across vast landscapes, making it difficult to decide where to focus attention,” Wang said. “With limited budgets, this new model provides clear guidance on when and where to invest in prevention or surveillance.”
Any large-scale system threatened by the introduction and spread of a rare foreign agent can benefit from the AI tool, with applications including fighting invasive species, infectious disease in agriculture or zoonotic disease in public health.
“The applications of this model are endless,” Schuler said.
Lauren Cahoon Roberts is director of communications at the College of Veterinary Medicine.