Citizen science data accurately predicts expert-derived species richness at a continental scale when sampling thresholds are met

Abstract

Understanding species richness patterns in time and space is critical for conservation management and ecological analyses. But estimates of species richness for a given place are often imprecise and incomplete, even when derived from expert-validated range maps. The current uptake of citizen science in natural resource management, conservation, and ecology offers great potential for extensive data to define species occurrence and richness patterns in the future. Yet, studies are needed to validate these richness patterns and ensure these data are fit-for-purpose. We compared data from a continental-scale citizen science project—FrogID—with expert-derived range maps to assess how well the former predicts species richness patterns in space. We then investigated how many citizen science submissions are necessary to fully sample the underlying frog community. There was a strong positive association between citizen science species richness estimates and estimates derived from an expert-derived map of frog distributions. An average of 153 citizen science submissions were necessary to fully-sample frog richness based on the expert-derived frog richness. Sampling effort in the citizen science project was negatively correlated with the remoteness of an area: less remote areas were more likely to have a greater number of citizen science submissions and be fully sampled. This suggests that scientists will likely need to rely on professionals for data collection in remote regions. We conclude that a citizen science project that has been running for ~ 18 months, can accurately predict frog species richness at a continental scale compared with an expert-derived map based on ~ 240 years of data accumulation. At large-scales, biodiversity data derived from citizen science projects will likely play a prominent role in the future of biodiversity and conservation.