Salt marsh birds in eastern North America are high conservation priorities due to their restricted ranges, threats from sea-level rise, and an overall lack of information related to their population status or trends. We calculated estimates of detection probabilities for clapper rail (Rallus crepitans), willet (Tringa semipalmata), seaside sparrow (Ammodramus maritimus), and saltmarsh sparrow (A. caudacutus) breeding in Delaware Bay, USA. Our objectives were to model the effects of survey covariates on detection probabilities for occupancy and abundance estimates, determine how survey frequency (no. visits to a site) and survey timing (when in the season sites are sampled) affected the estimates, and how varying the number and timing of visits affected the estimates compared to the recommendations based on the North American Marsh Bird Monitoring Protocol. To make these comparisons, we defined 4 survey frequency (2, 3, 4, and 8 visits) and 3 survey timing (early, early-middle-late, and late season) scenarios. Our results are based on 480 surveys (i.e., call broadcast point counts) to 30 sampling locations located at Bombay and Prime Hook National Wildlife Refuges during the 2008 and 2009 breeding seasons. The scenario of 3 visits early in the season provided the lowest error estimates of occupancy and abundance. Distributing the survey effort across the season, our early-middle-late scenario, provided high variance estimates and should be avoided when designing monitoring programs for these tidal marsh obligate species. Ó 2015 The Wildlife Society.