Intra-sensor Spectral Compatibility Analysis of the Enhanced Vegetation Index for Moderate Resolution Imagers Using Hyperion
Abstract
Spectral vegetation indices (VIs) are one of the more important satellite products for monitoring and characterizing seasonal dynamics and interannual variability of terrestrial vegetation in regional to global scales. The enhanced vegetation index (EVI), designed for the National Aeronautics and Space Administration (NASA) Earth Observing System (EOS) Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, has been shown to effectively characterize global vegetation states and ecosystem processes, and encompass the range of biophysical/biochemical information in manners to complement the conventional, normalized difference vegetation index (NDVI). Because the EVI is limited to sensor systems designed with a blue band, a 2-band EVI, without a blue band, was recently developed (referred to as the EVI2), which has the best similarity with the 3-band EVI for the MODIS sensors, particularly when residual atmospheric effects are insignificant and data quality is good. In this study, we evaluated intra-sensor spectral compatibility between the EVI and EVI2 and their sensitivities to aerosol contaminations for select moderate resolution imaging sensors, including Visible/Infrared Imager/Radiometer Suite (VIIRS), MODIS, Second Generation Global Imager (S-GLI), and SPOT-4 VEGETATION, and for Landsat-7 Enhanced Thematic Mapper Plus (ETM+), using a global set of Earth Observing One (EO-1) Hyperion hyperspectral data. Hyperion scenes were spectrally aggregated into red, near-infrared (NIR), and blue bandpasses of the five sensors and spatially aggregated into 1 km resolution pixels. Two atmospheric correction scenarios were also applied to examine the impact of the atmosphere on intra-sensor EVI/EVI2 compatibility: (1) Rayleigh/ozone/water vapor (ROW)-corrected and (2) total-atmosphere-corrected 'top-of-canopy (TOC)' reflectances. The EVI of all the five sensor bandpasses exhibited great resistance to atmospheric aerosol contaminations. The EVI2 of every sensor was, on the other hand, impacted by the aerosol effects as it no longer possesses the aerosol resistance term. The EVI and EVI2 showed the best similarities for MODIS and ETM+ bands. Their similarities degraded for VIIRS bands and were the worst for VEGETATION and S-GLI bands for which EVI and EVI2 differences reached 0.04 EVI units. These results indicate that the EVI2 may require an adjustment or optimization for some sensor bandpasses in order to retain intra-sensor compatibility with the EVI.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2013
- Bibcode:
- 2013AGUFM.B41D0444M
- Keywords:
-
- 0480 BIOGEOSCIENCES Remote sensing;
- 1640 GLOBAL CHANGE Remote sensing