Authors: Adam Berland*, Ball State University, Lara Roman, USDA Forest Service, Jess Vogt, DePaul University
Topics: Environmental Science, Natural Resources
Keywords: citizen science, emerging technologies, urban forestry
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Congressional A, Omni, West
Presentation File: No File Uploaded
Municipalities need street tree inventory data to effectively manage street trees, but field inventories are often cost prohibitive. It is possible that emerging technologies like Google Street View can be used to complement or replace field inventories, but there are fundamental questions about the level of data quality for tree inventories generated remotely using Google Street View. We addressed this issue by assessing the quality of street tree data produced by self-identified novice, intermediate, and expert analysts for the following tree attributes: number of trees, diameter class, genus, and species. We measured agreement among analysts within each expertise group, and we also measured agreement between each analyst and field data from the same locations in suburban Chicago. Based on data from 16 analysts who inventoried trees using Google Street View, we found that data collection was faster for experts (1.5 minutes/tree) compared to field data collection (3.1 minutes/tree), while novice and intermediate groups were comparable to the field crew on average. Analysts across all expertise groups excelled at quantifying the number of trees on street segments, but results were less consistent for diameter class and genus/species identification. In general, experts and some intermediate analysts exhibited the best performance. Our findings suggest that communities can rely on analysts of any expertise level to reliably quantify the locations and number of street trees, but a higher level of expertise is needed to produce quality data on more detailed tree information. We offer recommendations for using this technique in practice.