Optimizing free-roaming dog control programs using agent-based models

Urban free-roaming dog populations in the developing world are managed by a patchwork of local veterinary practitioners, government programs, and non-governmental organizations with varied effectiveness. While lethal removal is still commonly practiced, vaccination and fertility control methods are increasingly being adopted. Identifying which method(s) provides the most cost effective management is needed to inform dog population managers who seek to limit conflicts like dog bites, the spread of disease, and predation on wildlife. Here we describe an agent-based model that simulates the population of free-roaming dogs in Jaipur, a northwestern Indian city. We then apply various lethal and fertility control methodologies to identify which most effectively lowered the dog population size. This spatially explicit model includes temporal and demographic details of street dog populations modeled after data from the study city. We tested each pairing of control type (lethal or fertility) with search method (how to target efforts) to see their efficacy at altering the city’s dog population size, age structure, sterilization coverage, as well as the number of dogs handled. Models were run for 15 years to assess the long term effects of intervention. We found that the fertility control method that targets areas of the city with the highest percentage of intact bitches outperforms all other fertility control and lethal removal programs at reducing the population size while sterilizing a significantly higher proportion of the population. All lethal program methods skewed population demographics towards significantly younger dogs, thus likely increasing the frequency of conflict with humans. This work demonstrates the benefits of modeling differing management policies in free-roaming dogs.