A review of sampling-based approaches for estimating landscape metrics

摘要

Landscape metrics are generally calculated from complete land cover and land use maps. However, a relatively new approach has been to use sample data for the same purpose. The objective of the study was to review the literature on sampling methods for estimating landscape metrics. The advantages of using sampling include the possibility to derive metrics at low cost, the possibility of using existing survey data, and the possibility in some cases to attain improved accuracy through careful assessments at a limited number of sampling units rather than by crude assessments for all patches of a land cover map. Among the disadvantages are the estimation of some metrics with bias or that cannot be estimated at all, and to some extent handling of sample data requires practitioners to acquire new skills. In conclusion, sampling-based approaches for the estimation of landscape metrics appear promising. Interesting areas for further studies involve (1) the optimization of sample-based inventories with respect to landscape metric estimation, (2) comparisons of selected sampling-based approaches and "wall-to-wall" based approaches, where all relevant error sources and costs are included, and (3) further investigations of whether metrics can be redefined to better suit sample-based assessments without compromising interpretability.