Tree-quest: A citizen science app for collecting single-tree information

Quantifying single-tree structural attributes through crowdsourcing has strong potential to expand the availability and timeliness of ground-based data for terrestrial carbon assessment of trees both within and outside forests. Recently, a number of freely available augmented-reality (AR) mobile apps have enabled accurate measurement of carbon-relevant single-tree attributes such as tree diameter and height. However, crowdsourcing with these apps is constrained by limited data quality control and the need for a separate species-identification app. Here, we present a crowdsourcing-ready workflow with our novel and freely available Tree-Quest (TQ) app that, in addition to AR-based measurements of tree height and diameter, integrates the Pl@ntNet API for species identification and includes a crowdsourced data quality curation step based on its AR images. We have compiled a dataset comprising 700 measurements of single trees acquired from 30 volunteers across two Austrian urban environments. The trees had diameters at breast height (DBHs) ranging from 11.9 cm to 161.2 cm and tree heights (THs) ranging from 4.6 m to 29.0 m. The dataset was evaluated for the accuracy of the identified tree attributes, including TH and DBH. Compared with professional forest inventory measurements, volunteers using TQ achieved a mean absolute error (MAE) of 3 cm for DBH (R 2 = 0.97; rMAE = 6%) and 1.5 m for TH (R 2 = 0.91; rMAE = 11%). TQ results were also consistent with DBH and TH measurements acquired with other freely available mobile applications. Besides these encouraging results, TQ can work offline and is highly modular, allowing the design of customized quests, targeted questionnaires, and uploads of the collected information to a central database to share single tree data openly, and supports user interaction through gamification. These features provide a solid and unique framework for collecting citizen-science data on single trees .