Environmental Engineering Reference
In-Depth Information
good enough for heterogeneous areas. A feasible and less expensive approach is to
integrate Landsat and MODIS data for applications.
There are varieties of approaches to integrate Landsat and MODIS data. In this
chapter, we limit our discussions on several selected examples and focus on three
categories of integration. The first category focuses on the algorithm integration and
tries to create consistent MODIS data product for Landsat using similar algorithm
such as the Landsat Ecosystem Disturbance Adaptive Processing System
(LEDAPS). The second category focuses on the data fusion approach and attempts
to integrate high temporal information from MODIS with high spatial information
from Landsat such as the Spatial and Temporal Adaptive Reflectance Fusion Model
(STARFM). The third category focuses on producing consistent data products using
MODIS data products as references.
16.2 Algorithm Integration
In order to integrate Landsat and MODIS data product, an intuitive idea is to use
MODIS algorithm for Landsat and produces similar data products at Landsat spatial
resolution. A first step of this effort is to produce surface reflectance using MODIS
algorithm. The surface reflectance product strives to remove atmospheric effects
(scattering and absorption) and is the basis for many high-level products and
quantitative applications. The LEDAPS is a NASA project to map disturbance,
regrowth, and permanent forest conversion across the continent (Masek et al.
2006
).
It processes Landsat imagery to surface reflectance, using atmospheric correction
routines developed for the Terra MODIS instrument (Vermote et al.
2002
).
The LEDAPS first calibrates Landsat data in digital number to the top-of-
atmosphere (TOA) reflectance using calibration coefficients provided in the
metadata file. TOA reflectance is then atmospherically corrected using the 6S
radiative transfer code (Vermote et al.
1997
) similar to the MODIS surface reflec-
tance product. Atmosphere correction procedure needs ancillary information on
ozone and water vapor, etc. Ozone concentrations are derived from Total Ozone
Mapping Spectrometer (TOMS) data aboard the Nimbus-7, Meteor-3, and Earth
Probe platforms. Column water vapor uses data from the NCEP reanalysis data. The
ozone and water vapor data are downloaded and organized into daily ancillary data
for LEDAPS processing. Digital topography and NCEP surface pressure data are
used to adjust Rayleigh scattering to local conditions (Masek et al.
2006
).
Similar to the atmospheric correction scheme in the MODIS surface reflectance
product, LEDAPS retrieves aerosol optical thickness (AOT) from Landsat imagery
using the dark and dense vegetation concept (Kaufman et al.
1997
). Aerosol optical
depth is retrieved at 1-km coarse spatial resolution first and then interpolated
spatially between the dark targets. The 6S radiative transfer algorithm uses the
interpolated AOT, ozone, atmospheric pressure, and water vapor to retrieve surface
reflectance.
Search WWH ::
Custom Search