We studied the occupancy dynamics of breeding crawfish frogs (Lithobates areolatus) at 45 randomly selected grasslands across 208 km2 at Big Oaks National Wildlife Refuge, in southeastern Indiana, USA, during March 2010 and 2011. We developed a suite of hypotheses explaining the relationship between occupancy, detection, and environmental covariates. We fit our hypotheses using multiseason occupancy models, and compared them in an information-theoretic, model-selection framework. Our top model suggested that the detection probability had a positive, linear relationship with time, temperature, and the amount of rain 24 hours before the survey, and had a quadratic relationship with date, which peaked on 19 March. Our top model supported our hypothesis that occupancy probability was positively correlated with grassland size; larger grasslands were more likely to be occupied by crawfish frogs. Based on our results, we recommend that managers conserve large tracts of grasslands near breeding sites. We recommend that in southeastern Indiana crawfish frog-breeding surveys be conducted in mid- to late-March, and that each call point be surveyed for 15 minutes. We provide a model to increase the precision of detection probability estimates for call surveys that target calling crawfish frogs. Published 2012. This article is a U.S. Government work and is in the public domain in the USA.