Estimates of community attributes such as species richness, local extinction, and turnover are critical when evaluating ecological restoration efforts. Estimates of species richness based on counts can be biased by variation in the probability of detection among different species. We quantified the effects of livestock exclusion on riparian bird communities using mark-recapture models to account for variation in species detection rates. Specifically, we estimated species richness and other community parameters for fenced and grazed sites with robust design models where closed-captures were treated as mixtures, and then used transition rates to calculate derived vital rates for avian communities. Estimates of species richness based on unadjusted counts were correlated with estimates from robust design models, but counts failed to detect important temporal changes in species richness. Estimates of species richness from robust design models increased at fenced and grazed sites over an 8-year period, but community vital rates were unaffected by cattle exclusion. We examined qualitative changes in abundance of birds in four nesting guilds, and concluded that temporal changes may have been driven by regional dynamics in avian communities. Our mark-recapture analysis allowed us to compare standardized estimates of community parameters between habitats, observers, and time periods after accounting for variation in detection rates. Robust design models are a useful tool that will facilitate accurate assessments of community dynamics following future restoration efforts.