Engaging Online Citizen Scientists and the Consensus Method to Monitor the Marine Biofouling Community

Abstract

Citizen science is an important and useful approach to research that broadens public science engagement and expands the scale at which science can be conducted. Monitoring for marine non-native species has been repeatedly identified as a venue for citizen scientists to make substantial contributions. In this study, we evaluated the accuracy of identifications made by volunteers of marine invertebrates on the project Invader ID, hosted on the online citizen science portal Zooniverse. We tested the efficiency and accuracy of invertebrate identifications made through consensus, where more than one volunteer must agree on a final identification before it is added to the database. Using the Matthew’s Correlation Coefficient, we found that four volunteers in consensus balanced efficiency and accuracy when compared to gold standard scientist identifications. Common, large taxa were identified most accurately; Branching Bryozoa identifications made by four volunteers were 85% accurate, Solitary Tunicates 91% accurate, and Colonial Tunicates 64%. In community-based comparisons, the identity of the taxonomist (volunteer or scientist) had a small impact on overall community composition, while site and salinity gradients had a strong impact on composition. We suggest that citizen science monitoring programs focused on highly recognizable taxonomic groups, or on a few species within those groups that can provide crucial information for managers aiming to quickly identify new invasions. Moreover, long-term assessments of composition could be successfully monitored by volunteers from photographs, helping to bolster understanding of ongoing impacts of invasive species or climate change.

Iguanas from above: Citizen scientists provide reliable counts of endangered Galápagos marine iguanas from drone imagery

“…Instead, in the task of detecting presence or absence of iguanas within an image, we found that the answer selected by five volunteers or more (from 20 or 30 total classifications) — which we refer to as the minimum threshold — was the most likely to be correct. This was somewhat unexpected, since previous work has shown that accuracy increased asymptotically with number of classifications per image (Swanson et al, 2016); however, other projects hosted by Zooniverse have found outcomes more similar to ours (Egna et al, 2020; Hennon et al, 2015; Lawson et al, 2022). We suspect that for some challenging images, highly skilled volunteers were required to identify the iguanas, and that these volunteers were relatively rare.…”

“…This was somewhat unexpected, since previous work has shown that accuracy increased asymptotically with number of classifications per image (Swanson et al, 2016); however, other projects hosted by Zooniverse have found outcomes more similar to ours (Egna et al, 2020;Hennon et al, 2015;Lawson et al, 2022). We suspect that for some challenging images, highly skilled volunteers were required to identify the iguanas, and that these volunteers were relatively rare.…”

“…Astronomy-related projects initially dominated the platform, however, recent years have seen a shift towards projects with a focus on Nature and Biology (Cox et al, 2015; Zooniverse.org website). The platform has hosted over 400 projects and has more than 2.5 million registered volunteers (Lawson et al, 2022) who — along with non-registered users — are free to participate in any project lodged there. In our cumulative curve of participation, we experienced a daily classification rate of 500-1500, which increased in response to promotional work; in keeping with results from other projects (Lintott et al, 2008; Spiers et al, 2019).…”

Tadpole fingerprinting: Using tail venation patterns to photo‐identify tadpole individuals of a threatened frog

“…For large datasets, confirmation by an expert reviewer may not be possible and in its absence, positive identification must be accepted once a certain level of consensus among non‐expert observers is achieved (Swanson et al, 2016). To reduce the chances of consensus being incorrect, the threshold set for the detection of a true recapture event using photo‐based CMR will need to be set based on the degree of variability in the target feature between individuals in the population, the size of the database, the number of participants who are used to compare all images, and verification tests on portions of the data (Lawson et al, 2022; Pennycuick, 1978; Swanson et al, 2016). We suggest that there are key aspects of the photographing process that are standardized, such as ensuring the camera lens is focused on the vein branches so that they are not blurry, and making sure tails are straight to prevent vein distortion.…”