value BRDF applications may be employed to further reduce angular variations.
This has been accomplished in the Nadir BRDF-Adjusted Reflectance (NBAR)
MODIS product, which generates nadir reflectances through BRDF model inversions applied to seven or more good quality, cloud-free acquisitions within a
composite cycle. The MODIS VI and NBAR products are produced at local solar
zenith and local solar noon angle, respectively, and thus still contain latitudinal
and seasonal sun angle influences. Although a sun angle seasonal bias may be
present, it is less of a problem in inter-annual time series data and trend analyses,
providing there is no sensor orbital drift. Los (1993) documented AVHRR orbital
drift issues in the NDVI time series (Pathfinder AVHRR Land) and adjusted the
data to a standard illumination and viewing geometry by applying MODIS derived
kernels, resulting in a 50–85 % reduction of BRDF effects.
1.4 Validation
As one of the most widely used satellite data products by the research and
applications user community, the validation of VIs is essential to assess their
quality, accuracy, and reliability. Previous validation efforts have involved finer
resolution airborne and satellite imagery, field radiometer measurements, biophysical field sampling, and automated ground observation networks (Privette
et al. 2000), with the objective of validating VIs through independent radiometric
and biophysical measures.
Radiometrically, VIs are considered validated by independent, top-of-canopy
reflectance measurements as uncertainties in VI measures of greenness are solely
attributed to atmosphere correction accuracies and BRDF influences. VIs are
readily computed with in situ spectroradiometers, tower-mounted sensors, and
airborne instruments. Calibrated and traceable ‘‘transfer radiometers’’ mounted on
light aircraft and unmanned airborne vehicles (UAVs) can be flown at altitudes of
150–300 m above ground level (AGL) and acquire top-of-canopy reflectances with
nadir looks and prescribed sun angles for independent characterization of VIs. This
extends locally-constrained, sub-canopy scale field radiometric measurements to
kilometer length scales enabling sampling of site heterogeneity. Airborne sensors
can also be flown at higher altitudes (1–20 km AGL) for scaling and larger area
analyses encompassing a range of terrestrial biome types. The MODLAND Quick
Airborne Looks (MQUALS) is an example of light aircraft sensor and spectraldigital camera deployment flown at the Jornada Experimental Range for ground
truth characterization of surface conditions and sampling of landscape variability
(Huete et al. 1999).
The quality of satellite-based VI retrievals vary in space and time due to
geographic and seasonal variations in cloud persistence, unresolved clouds,
aerosols, quality of atmosphere correction, view-sun angle geometries, topography, and sensor performance (Wolfe et al. 1998; Miura et al. 2000; Samanta et al.
2010; Kobayashi and Dye 2005). Cloud and aerosol residual effects and artifacts
12
A. Huete et al.
This has been accomplished in the Nadir BRDF-Adjusted Reflectance (NBAR)
MODIS product, which generates nadir reflectances through BRDF model inversions applied to seven or more good quality, cloud-free acquisitions within a
composite cycle. The MODIS VI and NBAR products are produced at local solar
zenith and local solar noon angle, respectively, and thus still contain latitudinal
and seasonal sun angle influences. Although a sun angle seasonal bias may be
present, it is less of a problem in inter-annual time series data and trend analyses,
providing there is no sensor orbital drift. Los (1993) documented AVHRR orbital
drift issues in the NDVI time series (Pathfinder AVHRR Land) and adjusted the
data to a standard illumination and viewing geometry by applying MODIS derived
kernels, resulting in a 50–85 % reduction of BRDF effects.
1.4 Validation
As one of the most widely used satellite data products by the research and
applications user community, the validation of VIs is essential to assess their
quality, accuracy, and reliability. Previous validation efforts have involved finer
resolution airborne and satellite imagery, field radiometer measurements, biophysical field sampling, and automated ground observation networks (Privette
et al. 2000), with the objective of validating VIs through independent radiometric
and biophysical measures.
Radiometrically, VIs are considered validated by independent, top-of-canopy
reflectance measurements as uncertainties in VI measures of greenness are solely
attributed to atmosphere correction accuracies and BRDF influences. VIs are
readily computed with in situ spectroradiometers, tower-mounted sensors, and
airborne instruments. Calibrated and traceable ‘‘transfer radiometers’’ mounted on
light aircraft and unmanned airborne vehicles (UAVs) can be flown at altitudes of
150–300 m above ground level (AGL) and acquire top-of-canopy reflectances with
nadir looks and prescribed sun angles for independent characterization of VIs. This
extends locally-constrained, sub-canopy scale field radiometric measurements to
kilometer length scales enabling sampling of site heterogeneity. Airborne sensors
can also be flown at higher altitudes (1–20 km AGL) for scaling and larger area
analyses encompassing a range of terrestrial biome types. The MODLAND Quick
Airborne Looks (MQUALS) is an example of light aircraft sensor and spectraldigital camera deployment flown at the Jornada Experimental Range for ground
truth characterization of surface conditions and sampling of landscape variability
(Huete et al. 1999).
The quality of satellite-based VI retrievals vary in space and time due to
geographic and seasonal variations in cloud persistence, unresolved clouds,
aerosols, quality of atmosphere correction, view-sun angle geometries, topography, and sensor performance (Wolfe et al. 1998; Miura et al. 2000; Samanta et al.
2010; Kobayashi and Dye 2005). Cloud and aerosol residual effects and artifacts
12
A. Huete et al.