*2.4. Validation Scheme*

To validate the surface area classification using the Otsu thresholding approach, we downloaded five orthorectified and multispectral SkySat images [17] (Blue: 0.450–0.515 μm, Green: 0.515–0.595 μm, Red: 0.605–0.695 μm, and NIR: 0.740–0.900 μm) at sub-meter (0.66–0.73 m) spatial resolution (Table 3). For each image, we overlaid the OFR geometry and manually delineated the OFR surface area, which resulted in 144 validation surface areas from 71 different OFRs for multiple observations in time. Then, we conducted a pairwise comparison of the validation surface area with the surface area obtained from PlanetScope, Basemap, and Planet Fusion. When the PlanetScope surface area date did not correspond exactly to the SkySat dates, we used the closest observation in time, which had a maximum difference of three days before or after the SkySat date. In addition, we assessed the uncertainties of PlanetScope, Basemap, and Planet Fusion for different surface area size classes: 0.1–5 ha (*n* = 50), 5–10 ha (*n* = 46), and 10–50 ha (*n* = 48).


**Table 3.** SkySat image identification, acquisition date, number of OFRs surface area observations per image, percent clear (indicates the presence or absence of cloud cover; higher values indicate fewer clouds), ground-control ratio (defines the image positional accuracy; values closer to 1 mean higher accuracy), and ground-sampling distance in meters.
