Spatial assessment of two widely used land-cover datasets over the continental U.S

Abstract

Satellite-based land-cover datasets are required for various environmental studies. Two of the most widely used land-cover datasets for the U.S. are the National Land-Cover Data (NLCD) at 30-m resolution and the Global Land-Cover Characteristics (GLCC) at 1-km nominal resolution. Both datasets were produced around 1992 and were expected to contribute similar land-cover information. This study investigated the NLCD distribution within each of 11 GLCC classes at 1-km unit in ten U.S. states. Our analyses showed that the NLCD had similar spatial distribution as the GLCC for the classes of grassland, shrubland, as well as deciduous and evergreen forests. Meanwhile, the GLCC class of cropland and pasture was highly correlated to the NLCD classes of row crops and pasture/hay. The GLCC savanna was appropriately related to the NLCD grassland, pasture/hay, and deciduous forest. The NLCD classes of row crops, pasture, and deciduous forest mainly dominated the GLCC class of cropland/woodland mosaic. Spatial similarity was lower for the GLCC classes of mixed forest, wooded wetland, and cropland/grassland mosaic. In addition to the NLCD urban areas, the GLCC urban and built-up lands were consistently related to the NLCD vegetated areas due to the common mixture of urban and vegetated lands. A set of subclass land-cover information provided through this study is valuable to understand the degrees of spatial similarity for the global vegetated and urban-related classes in selected study areas. The results of this study provide great reference for interchanging less-detailed global land-cover datasets for detailed NLCD to support environmental studies.