Get reliable rangeland science

Rangelands are spatially heterogeneous and characteristically so; however, the sampling protocols used to estimate rangeland attributes assume spatial homogeneity, which is paradoxical. This study tests the hypothesis that stratified sampling techniques that assume homogeneity across the sampled space, like line-point intercept, generate statistically meaningful estimates of rangeland cover attributes.� By repeatedly generating line-point estimates over simulated data with known spatial attributes, a distribution of possible percent cover estimates is created. By analyzing this distribution, the precision of line-point intercept estimates can be parametrized in terms of a truth standardized confidence interval. The results indicate a tremendous lack of precision (i.e. wide variation) in estimates generated by line-point intercept, especially for low percent cover attributes. Therefore, homogeneity of the strata cannot be assumed and stratified sampling cannot be validly applied to estimate rangeland attributes; science, policy and debate of rangelands must rest upon alternative estimating procedures.