Abstract

Global land cover is the characterizations of Earth's surface, and it has been recognized as key drivers of biodiversity changes. Most biological conservation studies are focused on using land cover data to study the pattern of biological environment. However, few studies considered the accuracy of land cover products, which may induce further errors. Ecologists should be aware of those problems before usage. Former comparative analysis only measures the disagreements between different land cover products. But it can not assess the accuracy for individual dataset, due to the differences in classification scheme and cartographic standards. A better way to analyze the uncertainties in land cover products is to use dataset with same input data source, legend, and classifiers. MODIS Collection 5 global land cover dataset is the latest version of land cover type product, including eight annual land cover layer from 2001 to 2008. In this paper, we identified the uncertain areas and figured out the problematic land cover types in this dataset without ground validation samples. Results show that unreasonable changes are concentrated in six main regions, and mostly are in mountain areas and transition zones. The most confusing land cover types are herbaceous and shrub, woody, cultivated land, since they are highly mixable in feature space. Future ecological applications should be aware of these areas and types, and more efforts should be devoted into developing new methods and tools in promoting accuracy of those particular areas and types.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call