Non-random sampling and its role in habitat conservation: a comparison of three wetland macrophyte sampling protocols

Aquatic macrophytes provide essential spawning and nursery habitat for fish, valuable food source for waterfowl, migratory birds and mammals, and contribute greatly to overall biodiversity of coastal marshes of the Laurentian Great Lakes. Two approaches have been used to survey the plant community in coastal wetlands, and these include the grid (GR) and transect (TR) methods. These methods have been used to identify the average species richness at different sites, but their suitability for determining total species richness of a site has not been tested. In this paper, we compare the performance of these two established methods with that of the Stratified method (ST), which uses the sampler's judgment to guide them to different habitat zones within the wetland. We used the three protocols to compare species richness of six coastal wetlands of the Great Lakes, three pristine marshes in eastern Georgian Bay (Lake Huron) and three degraded wetlands in Lake Ontario, Canada. The greatest species richness was associated with the ST method, irrespective of wetland quality. The ST method was also more efficient (fewer quadrats sampled), and revealed the most number of unique (those found with only one method) and uncommon species (those found in < 5% of the quadrats). Despite these statistical differences, we found that sampling method did not significantly affect the performance of a recently developed index of wetland quality, the Wetland Macrophyte Index. These results have important implications for designing macrophyte surveys to track changes in biodiversity and wetland quality.