Evaluation of the e‐Surveyor Mobile Application for Undertaking Plant Surveys and Predicting Habitat Type

ABSTRACT

Mobile applications with automated species identification can assist citizen scientists in undertaking plant and habitat surveys. Whilst a high level of accuracy for these applications has been reported, very little testing has been done with citizen scientists in the field. If such applications are going to be used to support biodiversity research and conservation management, they need to be sufficiently accurate and functional for their intended use. We evaluated the accuracy of the e‐Surveyor mobile application, which includes automated identification for plant species and habitat prediction. We compared species lists and derived habitat associations collected by citizen scientists using the application in the field, with data recorded by expert botanists within the same survey plots. We also assessed the user experience via a questionnaire. Thirty‐seven citizen scientists attended the e‐Surveyor workshops and completed a questionnaire, with 51 individual plant surveys submitted across the three habitat types: calcareous grassland, neutral grassland and improved grassland. On average, experts recorded more plant species per plot compared with citizen scientists. Of the species recorded by citizen scientists on e‐Surveyor that were known to be present, 71% were correctly identified to species level, though typically only 45% of all observable species in the plot according to the expert botanists were captured correctly by the citizen scientists. Eighty per cent of surveys identified the correct first broad habitat and 25% identified the correct first phytosociological community suggested by the application. Citizen scientists provided valuable input through the questionnaire, including improvements to e‐Surveyor and future use cases. Most citizen scientists were able to accurately identify almost half of the observable plant species present and determine the correct broad habitat using e‐Surveyor, regardless of their botanical skill level. This suggests that the application is a useful tool for supporting biological recording, whilst improving confidence, knowledge and engagement with nature.