Animal movement and habitat selection are in part a response to landscape heterogeneity. Many studies of movement and habitat selection necessarily use environmental covariates that are readily available over large‐scales, which are assumed representative of functional habitat features such as resource availability. For widely distributed species, response to such covariates may not be consistent across ecosystems, as response to any specific covariate is driven by its biological relevance within the context of each ecosystem. Thus, the study of any widely distributed species within a limited geographic region may provide inferences that are not widely generalizable. Our goal was to evaluate the response of a marine predator to a suite of environmental covariates across a wide ecological gradient. We identified two behavioral states (resident and transient) in the movements of shortfin mako sharks (Isurus oxyrinchus) tracked via satellite telemetry in two regions of the western North Atlantic Ocean: the tropical Caribbean/Gulf of Mexico marginal sea (CGM), and the temperate waters off the east coast of North America (OWA). We compared patterns of resident behavior between regions, and modeled relationships between oceanographic variables and resident behavior. We tracked 39 sharks during 2013 – 2015. Resident behavior was associated with shallow, continental shelf and slope waters in both regions. In the OWA resident behavior was associated with low sea surface temperature and high primary productivity, however, sharks exhibited no response to either variable in the CGM. There was a negative relationship between sea‐surface height gradient (a proxy for oceanic fronts) and resident behavior in the OWA, and a positive relationship in the CGM. Our observations likely reflect shark responses to regional variability in factors responsible for the distribution and availability of prey. Our study illustrates the importance of studying widely distributed species in a consistent manner over large spatial scales.