Sampling efforts for estimating fish species richness in western USA river sites

Fish species richness is an important indicator of river ecological condition but it is particularly difficult to estimate in large unwadeable rapidly flowing rivers. Intensive multi-gear sampling is time consuming, logistically complex and expensive. However, insufficient sampling effort underestimates species richness and yields inaccurate data about the ecological condition of river sites. We raft-electrofished 10 large river sites in 10 different ecoregions and six western USA states for distances equal to 300 times their mean wetted channel widths (MCWs) to estimate the effort needed to approach asymptotes in fish species richness. To collect 90% of the observed fish species at the sites, we found that an average of 150 MCWs (ranging 80–210 MCWs) were needed, with the number of MCWs increasing in rivers with a higher proportion of spatially rare species. Frequently, the second or third additional 100 MCWs produced only one or two additional singletons or doubletons (species occurring only once or twice at a site). Before initiating sampling programs for estimating species richness, we recommend assessing sampling effort, particularly if rare or uncommon species are expected or desired.