Most spatial epidemiological studies of nature-health relationships use generalized green space measures. For instance, coarse resolution spatial data containing normalized difference vegetative index (NDVI) values are prominent despite criticisms, such as the researcher’s inability to restrain exposure estimates to public (accessible) and private (largely inaccessible) land. Non-threatening natural landscapes can improve health through building capacities for health-promoting behaviors (e.g., physical activity). Such behaviors may be best activated by recreational and accessible parks. We curated the Parks and Protected Areas Database of the U.S. (PAD-US) to identify parks that are accessible for outdoor recreation. Our title adds “AR” to “PAD-US” where A=Accessible and R=Recreational. We validated the PAD-US-AR by comparisons with “greenspace” datasets and sociodemographics, which demonstrated its uniqueness from other commonly employed metrics of nature exposure. The PAD-US-AR presents a reliable estimate for exposure to parks accessible for outdoor recreation. It has strong associations with home prices, shares of female residents, and shares of older residents. The dataset can accompany nature exposure metrics in environmental epidemiology and allied fields.