Get reliable rangeland science

Managers benefit from improved monitoring techniques that provide rapid, accurate, cost effective and robust measures of rangeland health and ecological trend. Remote sensing can be an effective tool for assessing rangelands efficiently and accurately. The purpose of this study is to evaluate and compare vegetation and bare-ground cover between high-resolution remotely sensed images with data collected from the field using common sampling practices. In 2011, high resolution (5cm) color and color-infrared images were taken of plots located throughout central and northern Utah using a UltraCam X (UCX) camera. Simultaneous to the aerial flights, field data (vegetation, rock and bare-ground cover) was collected by the Utah DWR Range Trend Program. Remotely sensed images were classified using a supervised classification in ERDAS Imagine software to generate vegetation classes and surface cover values. The reliability of these classifications were assessed using an Accuracy Assessment coupled with a generated Kappa Statistic value. Results indicate that total classification accuracy was 91% with a Kappa of 0.88. Total shrub cover measured using remote sensing was only 1% different than field based samples. Herbaceous vegetation differed by 5.6%. Differences in surface cover (rock, bare-grouse) between both methods was higher (26.8%). Remote sensing can be used by managers effectively quantify, monitor, and evaluate rangelands across temporal and spatial gradients.