|Taylor, Joshua - Bret|
Submitted to: Society for Range Management Meeting Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 8/28/2006
Publication Date: N/A
Technical Abstract: Half a century ago Aldo Leopold observed, “It is impossible fully to protect cheat country from fire...,” and chided, “We tilt windmills in behalf of conservation in convention halls... but on the back forty we disclaim even owning a lance.” “Lances” effective against cheatgrass (Bromus tectorum L.) or other invasive species are expensive. Even simple assessments of weed cover are hampered by the costs of labor and travel. Lack of statistically-adequate, detailed vegetative data often prevents large-area analysis of weed ecology. Recent advances in aerial imaging now allow routine acquisition of 1-mm/pixel (GSD), geocoded, aerial images. A co-incident development of a digital-point-frame (‘SamplePoint’ software) takes advantage of 1-mm GSD imagery in the same way that conventional point sampling takes advantage of sharp pins for ground-cover measurements. Aerial tools are a less-expensive way to obtain hundreds or thousands of intermittent images and, with SamplePoint and other image-analysis and GIS software programs, are a means of detecting, mapping, and examining the ecology of infestations and of evaluating control and management efforts across landscapes. These applications have been tested in projects evaluating the occurrence of cheatgrass on Nevada rangeland and of spotted knapweed (Centaurea maculosa) on Idaho rangeland. The Nevada study covered a 10,117-ha parcel using 649 aerial images collected during June 2005. The data were used with GIS databases to examine the relationship between cheatgrass cover and the study-site physiography. A significant association was established with aspect. The Idaho study was conducted north of Dubois at sites known to have spotted knapweed. Agreement on the presence of spotted knapweed between ground and aerial methods ranged from 25-83%. Density measurements between ground transects and image analysis were correlated (r=0.83) with means no different in 2 of the 3 study sites (P<0.05). We conclude that high-resolution surveys offer a promising “lance” against weeds.