Citizen science data are a reliable complement to quantitative ecological assessments in urban rivers

Abstract

Abstract Conservation and management agencies may benefit from volunteer programs, or citizen science, to monitor a variety of ecosystem types. Citizen science can simultaneously engage individuals living in or near an ecosystem and provide environmental data that may otherwise be lacking. However, data produced by untrained volunteers is not always trusted among agency and academic researchers due to uncertainty surrounding accuracy and lingering stigma associated with ‘amateur’ assessments. Here we provide site-specific comparisons between quantitative ecological monitoring data and qualitative data produced by citizen scientists. Two citizen science organizations in the metro Detroit (Michigan, USA) area conduct regular monitoring of aquatic macroinvertebrates to assess urban stream quality. We conducted analogous sampling at seven locations over three years (12 total samples) using quantitative stream ecology methods to provide a site-specific comparison of stream assessments. Invertebrate assemblage composition varied between methods, overlapping by an average of 30% per site. Assemblage differences largely corresponded to under-sampling of small sessile taxa and large mobile taxa by volunteer and quantitative surveys, respectively. Most of the invertebrate taxa which differed between data types were naturally rare taxa. This resulted in predictable differences in quality assessment by quantitative and qualitative methods. Citizen volunteers regularly rated sites with low invertebrate richness as higher in quality than did quantitative assessments, resulting in a conservative indicator of stream degradation. Though qualitative citizen data and quantitative data varied, both methods had short-comings and strengths. Citizen science organizations can thus provide a useful complement to traditional monitoring practices, particularly given the long-term, spatially broad, and repeated nature of their sampling.