Coccidioidomycosis or Valley Fever (VF) is an emerging soil-borne fungal zoonosis affecting humans and animals. Most non-human cases of VF are found in dogs, which we hypothesize may serve as sentinels for estimating the human exposure risk. The objective of this study is to use the spatial and temporal distribution and clusters of dogs seropositive for VF to define the geographic area in Texas where VF is endemic, and thus presents a higher risk of exposure to humans. The included specimens were seropositive dogs tested at a major diagnostic laboratory between 1999 and 2009. Data were aggregated by zip code and smoothed by empirical Bayesian estimation to develop an isopleth map of VF seropositive rates using kriging. Clusters of seropositive dogs were identified using the spatial scan test. Both the isopleth map and the scan test identified an area with a high rate of VF-seropositive dogs in the western and southwestern parts of Texas (relative risk = 31). This location overlapped an area that was previously identified as a potential endemic region based on human surveys. Together, these data suggest that dogs may serve as sentinels for estimating the risk of human exposure to VF.