Large-Scale Assessment of Rapid Monitoring Methods for Estimating Moist-Soil Seed Production

Federal, state, and private entities manage seasonal flooded, shallow wetlands to provide food and other habitat resources for wetland-dependent migratory birds, including migrating and wintering waterfowl. Individual National Wildlife Refuges managed by the U.S. Fish and Wildlife Service annually monitor seed production in moist-soil wetlands to track performance relative to regional foraging habitat objectives and to evaluate local habitat management activities. The National Wildlife Refuge System does not currently have a standard sampling protocol, and thus seeks a reliable rapid assessment method for estimating seed production to achieve standardized estimates and to avoid inconsistencies in data collection, metrics used, and usefulness of the monitoring efforts. We compared seed yield estimates derived from a suite of commonly used seed production assessment methods with those from soil core samples across six National Wildlife Refuges in the southeastern United States. The most parsimonious model included only common plant species and a single visual assessment of overall coverage (1-5) and seed quality (1-4) for each moist-soil unit (r(adj)(2) = 0.71). Generally, models that included only common plant species and a visual estimate of seed yield for moist-soil wetlands overall had greater support than models that included all plant species and those that included data from subplots (n = 10) nested within moist-soil wetlands. Experience level of observer had a moderate effect on accuracy (r(mar)(2) = 0.20) and geographic range increased variation in overall seed yield estimates within moist-soil wetlands. Notably, we found that similar indices developed in different geographic regions performed well across the Southeast, but a widely used index based on estimates of seed yield for individual plant species performed poorly in this study. Standardizing the use of a single, efficient, and reliable method to estimate seed abundance in moist-soil wetlands will provide wetland mangers the ability to consistently estimate performance relative to objectives, evaluate management actions, and track trends on National Wildlife Refuges in the southeastern United States.

Automated summary

The study compared the impact of different sampling methods on seed yield estimates in moist-soil wetlands in National Wildlife Refuges in the southeastern United States, finding that models including only common plant species and a single visual assessment were more reliable. Observer experience level and geographic range had a moderate effect on seed yield estimates.