Timing and landowner access govern the effectiveness of spatially targeted culling for chronic wasting disease management: insights from a spatially explicit agent-based model

Chronic wasting disease (CWD) poses a persistent and expanding threat to free-ranging cervid populations across North America, yet the effectiveness of management interventions depends critically on their alignment with host behavior, disease dynamics, and operational constraints. Spatially targeted culling of entire deer social groups has been proposed as a strategy to disrupt localized transmission and environmental contamination, but its performance under realistic levels of landowner cooperation remains poorly understood. We developed a spatially explicit agent-based model parameterized with surveillance, harvest, and management data from southeastern Minnesota, USA to evaluate the potential and limitations of spatially targeted culling across different phases of disease emergence and spread. The model explicitly represents deer social organization, sex- and age-specific movement, environmental prion accumulation, harvest-based surveillance, and primarily private landowner-dependent access for targeted removal. Simulations reproduced observed disease dynamics during the calibration period (2016 to 2023) and successfully reproduced observed trends during an out-of-sample validation period (2024 to 2025), capturing both the initial emergence of CWD and its subsequent expansion under changing management conditions. Prospective projections indicated that, under current or substantially increased levels of landowner participation, spatially targeted culling alone was unlikely to prevent progression to endemic prevalence. In contrast, retrospective analyses showed that early increases in landowner access could have substantially delayed disease expansion and, under high-access scenarios, prevented endemic establishment altogether. The effectiveness of targeted culling declined sharply with delayed implementation, even when similar access levels were eventually achieved. Together, these results demonstrate that the utility of spatially targeted culling depends primarily on the timing of intervention relative to disease spread, with high removal intensity and sufficient access required to achieve meaningful impacts within a narrow window during which localized interventions can influence long-term disease trajectories. More broadly, this work underscores the importance of phase-aware, adaptive management strategies that prioritize early intervention in newly affected regions and provides a transferable modeling framework for evaluating wildlife disease control under realistic social and logistical constraints.