Thematic validation of high-resolution global land-cover data sets

This paper describes a procedure to validate the thematic accuracy of the International Geosphere-Biosphere Programme, Data and Information System (IGBP-DIS) DISCover (Version 1.0) 1-Kilometer Global Land-Cover Data Set. Issues of data set sampling design, image geometry and registration, and core sample interpretation procedures are addressed. Landsat Thematic Mapper and SPOT satellite image data were used to verify 379 primary core samples selected from Discover 1.0 using a stratified random sampling procedure. The goal was to verify a minimum of 25 samples per DISCover class this was accomplished for 13 of the 15 verified classes. Three regional Expert Image Interpreters independently verified each sample, and a majority decision rule was used to determine sample accuracy. For the 15 DISCover classes validated, the average class accuracy was 59.4 percent with accuracies for the 15 verified DISCover classes ranging between 40.0 percent and 100 percent. The overall area-weighted accuracy of the data set was determined to be 66.9 percent. When only samples which had a majority interpretation for errors as well as for correct classification were considered, the average class accuracy of the data set was calculated to be 73.5 percent.