A comparative analysis of three different MODIS NDVI datasets for Alaska and adjacent Canada

Abstract

Moderate Resolution Imaging Spectroradiometer (MODIS) data offer great potential for monitoring vegetation dynamics in Alaska. However, certain MODIS image quality issues, such as geometric distortion, have been analyzed and documented in high latitudes and regions distant from the Greenwich Meridian. To improve MODIS data usability, the Canada Centre for Remote Sensing (CCRS) developed a seven-band reflectance dataset (10 day composite) at 250 m resolution for Canada and North America. More recently, the US Geological Survey Earth Resources Observation and Science Center produced an eMODIS dataset that includes a 7 day composite of normalized difference vegetation index (NDVI) and surface reflectance data at 250 m, 500 m, and 1 km resolutions for the conterminous United States and Alaska. Although these two datasets are based on the same MODIS level 1B data as those of the standard MODIS products, they are processed to improve their use in high-latitude regions. In this study, we conducted a comparative analysis of the standard MODIS, CCRS MODIS, and eMODIS Alaska 250 m NDVI products using the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) images as a reference source. CCRS MODIS and eMODIS Alaska NDVI images have significantly improved geometric features over those of the standard MODIS product. Pixel-by-pixel comparisons of the MODIS datasets indicated that all retained the original MODIS radiometric characteristics, but considerable mismatches at the pixel level were found due to geometric distortions caused by resampling. All three MODIS datasets agreed well as images were degraded to 5 or 10 km resolution.