The University of Maryland's 1km Global Land Cover Classification was used for vegetation classes for NCA-LDAS. The data was observed between 1981 and 1994 using the AVHRR satellites. For more info on the procedure used, please see Hansen et al., 2000, Int. J. Remote Sensing, Vol. 21, No. 6 & 7, 1331-1364.
There are 13 vegetation classes in the dataset, plus one class for water:
INDEX CLASS ----- ----- 0 Water 1 Evergreen Needleleaf Forest 2 Evergreen Broadleaf Forest 3 Deciduous Needleleaf Forest 4 Deciduous Broadleaf Forest 5 Mixed Cover 6 Woodland 7 Wooded Grassland 8 Closed Shrubland 9 Open Shrubland 10 Grassland 11 Cropland 12 Bare Ground 13 Urban and Built-Up 14 Water
The NCA-LDAS group re-gridded the land cover from the UMD 1km grid to the 1/8th-degree NCA-LDAS grid by counting the number of 1km points of each vegetation type within the larger 1/8th-degree grid size. The LIS software determined the dominant vegetation type in each grid point, as the initial NCA-LDAS simulations only use one tile per grid point.