Can Plot-Level Photographs Accurately Estimate Tundra Vegetation Cover in Northern Alaska?

Plot-level photography is an attractive time-saving alternative to field measurements for vegetation monitoring. However, widespread adoption of this technique relies on efficient workflows post-processing images and the accuracy resulting products. Here, we estimated relative cover using both traditional sampling methods (point frame) semi-automated classification photographs (plot-level photography) across thirty 1 m2 plots near Utqiaġvik, Alaska, from 2012 2021. Geographic object-based image analysis (GEOBIA) was applied generate objects based three spectral bands (red, green, blue) images. Five machine learning algorithms were then classify into groups, random forest performed best (60.5% overall accuracy). Objects reliably classified following classes: bryophytes, forbs, graminoids, litter, shadows, standing dead. Deciduous shrubs lichens not classified. Multinomial regression models used gauge if estimates plot-level could accurately predict point frame space or time. yielded useful graminoids. predictive performance varied by class whether it being in new locations change over time previously sampled plots. These results suggest that may maximize use time, funding, available technology monitor Arctic, but current sufficient detect small changes cover.