In dynamic environments, resource availability may change by several orders of magnitude, over hours to months, but the duration of resource availability is not often included as a characteristic attribute of resources even though temporal resource dynamics might limit patterns of use. In our study of wading birds foraging in intertidal areas, tides cause large changes in the areal extent of shallow‐water foraging habitat (i.e., the resource), but tides also constrain the duration of availability, which is often overlooked. We hypothesized that temporal constraints on habitat availability from tides would be reflected in patterns of habitat use by foraging birds. We estimated the time‐integrated habitat availability and compared it to traditional habitat attributes (seagrass cover, substrate type, instantaneous water depth, and proximity to mangrove islands or deep water) that have strong evidential support for influencing patterns of use. To evaluate our hypotheses, we quantified habitat attributes at intertidal areas in the Florida Keys, USA, where wading birds were observed foraging (Little Blue Heron, Egretta caerulea: N = 183; Great White Heron, Ardea herodias occidentalis: N = 162). We tested for nonrandom use by sampling habitat attributes at two spatial scales around the observed feeding locations and we analyzed the data using a conditional logistic regression model. There was no evidence that seagrass cover or substrate explained patterns of use. The proximity of foraging locations relative to mangroves and to deep water were important at both spatial scales but had lower effect sizes (odds ratios) than time‐integrated habitat availability and water depth, and the latter may only serve as a physical constraint on access. We found support that time‐integrated habitat availability was a distinct resource attribute, had the greatest effect size (four‐ to eightfold change in relative probability of use), and best explained patterns of habitat use at the largest spatial scale. In studies of resource use where changes in resource availability are nonlinear or when strong constraints on access are imposed by behavior, incorporating time‐integrated estimates of resource availability into analyses can improve insights into spatiotemporal patterns of resource use.