Development of New Global Surface Type Maps from VIIRS Data

Accurate representation of terrestrial surface types or land cover at regional to global scales is an important element for a wide range of applications, such as land surface parameterization, modeling of biogeochemical cycles, and carbon cycle studies. Global surface type map is also a core input for National Weather Service’s (NWS) Numerical Weather Prediction (NWP) models, and could be used in other weather, climate, hydrology and ecology models. The National Oceanic and Atmospheric Administration (NOAA) Joint Polar Satellite System (JPSS) land surface type team has been generating global surface type classification maps using global observations acquired by the Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the Suomi National Polar-orbiting Partnership (S-NPP) satellite launched on 28 October 2011. The newest VIIRS surface type map with the 17-class International Geosphere-Biosphere Program (IGBP) classification scheme was created from 2014 VIIRS surface reflectance data using the support vector machines algorithm. Three additional tundra types were added to accommodate the requirement of the National Centers for Environmental Prediction’s (NCEP) model. Meanwhile, for some other studies, different classification types are also needed. For instance, the retrieval of global leaf area index (LAI) and fraction of photosynthetically active radiation absorbed by the vegetation (fPAR) relies on accurate discrimination of six major biome classes based on canopy structure, which include: Grass/Cereal Crops, Shrubs, Broadleaf Crops, Savannas, Broadleaf Forests, and Needleleaf Forests. Therefore, the original VIIRS surface type classification map in IGBP types is also converted to a 6-biome type classification. In this study, the development of the new VIIRS surface type map in 17 IGBP classes and the NCEP’s 20 types (IGBP + 3 tundra classes) are introduced. Then the conversion method from IGBP to biome types based on a look-up table and a secondary SVM classification is discussed, along with preliminary validations for both the IGBP and biome maps.