Buffer zones, calculated by flight‐initiation distance (FID), are often used to reduce anthropogenic disturbances to wildlife, but FID can vary significantly across life‐history stages. We examined the behavioral effect of potential natural (gulls and shorebirds) and anthropogenic (pedestrians) disturbance sources to staging roseate (Sterna dougallii) and common tern (S. hirundo) flocks from July to September in 2014 and 2015 at Cape Cod, Massachusetts, USA. We estimated the proportion of the flock exhibiting different responses to potential disturbance sources as a function of distance, flock size, percent roseate terns, and local disturbance rates, using Bayesian zero‐and‐one inflated beta regression. The proportion of tern flocks responding to the presence of shorebirds by flying was low (0.01 ± 0.001 [SE]) and did not vary by distance or other covariates, whereas the proportion flying in response to gulls increased as distance decreased, with smaller flocks, and with flocks with a larger proportion of roseate terns being more sensitive to gull presence. Prolonged flight response rapidly increased in probability from 0.0 to as much as 1.0 as distance from pedestrians to the flock decreased from 100 m and was much more likely with smaller flocks. Pedestrian activity levels also had an effect on flock responses; those engaged in active behaviors such as jogging were more likely to cause flushing than those engaged in passive behaviors. Terns seemed to view pedestrians as more of a threat than shorebirds and gulls, even though gulls are frequent kleptoparasites of terns. Pedestrians >120 m from a tern flock generally elicited the same probability of flight response as shorebirds and gulls. We recommend managers maintain anthropogenic disturbance levels at or below the intensity of those from natural sources at sites where recreation and wildlife values are both important. Because staging tern flocks may use a variety of areas within a site, we recommend instituting a 100‐m buffer around areas potentially used by staging flocks at Cape Cod, where we studied every location roseate terns are known to use in large numbers. For other sites used by mixed‐species tern flocks, we recommend the use of our field and analytical methods to develop appropriate buffer distances that will keep pedestrians far enough away to reduce the likelihood of flight and other non‐locomotive anti‐predator behaviors. These buffer zones will also benefit other species sensitive to human activity.