|Hopkins, Karl - OREGON STATE UNIVERSITY|
Submitted to: Society of Range Management
Publication Type: Abstract Only
Publication Acceptance Date: January 1, 2005
Publication Date: February 1, 2005
Citation: Boyd, C.S., Hopkins, K., Svejcar, A.J. 2005. A photo-based monitoring technique for willow communties. [abstract] Society of Range Management. Paper No. 35. Technical Abstract: Willow (Salix spp.) and associated riparian shrubs are ecologically important in many riparian ecosystems. However, quantifying abundance and relative change in woody plants has proven difficult for reasons such as variability in measurement between observers and difficulties in developing standardized techniques. The objective of this study was to evaluate the variability associated with collection and analysis of field data for a photo-based monitoring technique for willow communities. To evaluate variation in data collection, we photographed 5 willow clumps, 10 times each and compared estimates within clump. This was assumed to mimic variability associated with repeat monitoring of a given clump over time. We used pin flags set at known distances apart to provide scale references in photographs and flag and camera locations were marked. We removed, and replaced, cameras and flags between successive photographs. We spatially rectified scanned images using digital imaging software and determined the 2-dimensional area of willow clumps by digitizing clump boundaries. To determine variability associated with image analysis, 6 observers were asked to determine willow clump area on a series of 5 images and results were compared across observers. Sampling error for each clump was calculated by dividing root mean square error by the mean value. Results indicate that field data collection produced minimal variability; sampling error averaged 1.82% across willow clumps. Sampling error between observers averaged 1.63% across clumps, and was <3% for all clumps. These results indicate our technique produces reliable estimates of willow abundance, while minimizing observer variability.