| Summary: | Higher spatial and temporal resolutions of remote sensing data are likely to be useful for ecological monitoring efforts. There are many different treatment approaches for the introduced European genotype of <i>Phragmites australis</i>, and adaptive management principles are being integrated in at least some long-term monitoring efforts. In this paper, we investigated how natural color and a smaller set of near-infrared (NIR) images collected with low-cost uncrewed aerial vehicles (UAVs) could help quantify the aboveground effects of management efforts at 20 sites enrolled in the <i>Phragmites</i> Adaptive Management Framework (PAMF) spanning the coastal Laurentian Great Lakes region. We used object-based image analysis and field ground truth data to classify the <i>Phragmites</i> and other cover types present at each of the sites and calculate the percent cover of <i>Phragmites</i>, including whether it was alive or dead, in the UAV images. The mean overall accuracy for our analysis with natural color data was 91.7% using four standardized classes (Live <i>Phragmites</i>, Dead <i>Phragmites</i>, Other Vegetation, Other Non-vegetation). The Live <i>Phragmites</i> class had a mean user’s accuracy of 90.3% and a mean producer’s accuracy of 90.1%, and the Dead <i>Phragmites</i> class had a mean user’s accuracy of 76.5% and a mean producer’s accuracy of 85.2% (not all classes existed at all sites). These results show that UAV-based imaging and object-based classification can be a useful tool to measure the extent of dead and live <i>Phragmites</i> at a series of sites undergoing management. Overall, these results indicate that UAV sensing appears to be a useful tool for identifying the extent of <i>Phragmites</i> at management sites.
|
|---|