Monitoring spotted knapweed with very-large-scale-aerial imagery in sagebrush-dominated rangelands.

Authors: Taylor, Joshua - Bret; Moffet, Corey; Booth, D

Submitted to: Society for Range Management Meeting Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 12/1/2009
Publication Date: 2/8/2010
Citation: Taylor, J.B., Moffet, C.A., Booth, D.T. 2010. Monitoring spotted knapweed with very-large-scale-aerial imagery in sagebrush-dominated rangelands. Society for Range Management Meeting Abstracts. P B-95.

Interpretive Summary: Spotted knapweed (Centaurea stoebe L.) invades and destroys productive rangelands. Because of remoteness, rough terrain, and extensiveness of many rangelands, locating and monitoring spotted knapweed infestations can be difficult and costly. We have established that spotted knapweed plants can be identified and measured from very-large-scale-aerial (VLSA) images that were acquired from low-flying (100 m) aircraft. Furthermore, approximately 30% less time was required to collect and evaluate weed infestation data using VLSA-based methods rather than ground-based methods. Based on our research, spotted knapweed infestations in sagebrush-dominated rangelands can be efficiently identified and quantified using VSLA-based methods

Technical Abstract: Spotted knapweed (Centaurea stoebe L.) invades and destroys productive rangelands. Monitoring weed infestations across extensive and remote landscapes can be difficult and costly. We evaluated the efficacy of very-large-scale-aerial (VLSA) imagery for detection and quantification of spotted knapweed in sagebrush (Artemisia tripartita Rydb., Artemisia tridentata Nutt. ssp. vaseyana Rydb.)-dominated rangelands. Three sites, having unique disturbances and varying infestations of spotted knapweed, were selected for the experiment. VLSA images of predetermined photographic targets within sites were acquired from 100 m above ground level. For each site, 30 or 40 photo points were randomly selected; photo points were randomly divided into groups of 10 for independent measures of image- and ground-based spotted knapweed cover and density; cover and density means were calculated for each group; and image- and ground-based means were randomly paired for analysis. Limit-of-agreement (LoA) analysis was used to determine method repeatability and method agreement (image vs. ground). Repeatability coefficients for density were ± 1.7 and 1.5 plants/m2 for image- and ground-based measurements, respectively. The ½LoA width, generated from comparing the methods, was 1.3 plant/m2, which indicated agreement. Repeatability coefficients for cover were ± 3.7 and 1.7% for image- and ground-based measurements, respectively. The ½LoA width was intermediate to the repeatability coefficients at 2.95%, and as cover increased, measurement differences between methods increased. Approximately 30% less time was required to collect and evaluate data using VLSA-based methods than ground-based methods. We conclude that spotted knapweed infestations in sagebrush-dominated rangelands were efficiently identified and quantified using VSLA-based methods.