Spatial Modeling of Caprine Brucellosis in Rural Goat Herds: A Cellular Automaton-Based Framework and Global Sensitivity Analysis

Brucellosis remains a major constraint for goat production and public health in rural regions, where limited sanitary infrastructure and heterogeneous environmental conditions complicate disease control. These characteristics motivate spatially explicit representations of individual animals and their environment. In this study, we develop a cellular automaton model for caprine brucellosis in rural goat production systems, representing individual goats, daily movement patterns, reproductive status, and spatially distributed environmental contamination. Disease transmission emerges from interactions between susceptible animals and contaminated land and water sources, while infection-induced abortion generates new environmental contamination. We performed a global sensitivity analysis using Sobol quasi-random sampling and partial rank correlation coefficients over 1498 model simulations, evaluating epidemic magnitude, epidemic timing, and reproductive losses. Epidemic magnitude was most strongly associated with the spatial scale of the production area, the probability of infection through contaminated land, and the proportion of pregnant goats, whereas reproductive losses were particularly sensitive to the proportion of pregnant animals. In contrast, the timing of epidemic and reproductive events showed comparatively weak partial associations with the parameters considered. These results illustrate how spatial structure, environmental exposure, and reproductive composition can differentially shape simulated disease outcomes in resource-limited production systems.