| Ground-based rangeland monitoring methods conducted by rangeland professionals have been the standard for over a century. However, costs to conduct ground-based monitoring are rising. Additionally, the use of ground-based monitoring methods can incur variation and error among observers and species and between observers, years, methods, plot sizes, and species. For these reasons, remote sensing has become a method of interest for monitoring rangeland ecosystems. The first experiment was conducted on very large scale aerial (VLSA) imagery acquired from the Grand River National Grassland (GRNG) near Hettinger, ND, USA. Kendall's coefficient of concordance (W), a non-parametric statistic, was used to analyze the agreement among observers of the same set of VLSA imagery for 19 ground cover classes. Two groups of three female observers were used. The first group consisted of observers under the age of 30 years who had at least two years of experience in the GRNG. The second group consisted of observers older than 30 years of age with little to no experience in the GRNG. Kendall's coefficient of concordance showed that the group of older observers with limited experience was consistently different than the younger observers with more experience. The second experiment compared the use of 1 mm per pixel VLSA imagery acquired from the GRNG to results for ground cover obtained by traditional, ground-based monitoring methods. Simple linear regressions were applied to indicator species and life forms using VLSA imagery analysis as the independent variable. VLSA imagery worked well to predict ground cover variables that had easily identifiable features such as distinct color or characteristic inflorescence and were abundant in the study area. VLSA imagery predicted results similar to basal cover acquired by ground-based monitoring methods but with considerable variability. The use of VLSA imagery to replace ground-based monitoring methods in the northern mixed prairie will depend on how much variability a manager is willing to accept. |
