10. EXPLOITATION AND EVALUATION OF RETRIEVAL
ALGORITHMS FOR GEOSTATIONARY SATELLITE DATA
PROCESSING
87
observed reflectance around 12 hours, i.e., when the Sun zenith angle is
minimum, decreases more rapidly than reflectances observed in the morning
or in the afternoon. At noon, the trajectory of photons that reach the soil
under the vegetation canopy and are reflected back to the sensor is shortest.
It translates into high reflectances, especially when the LAI is small. The
probability to be intercepted by the vegetation is indeed minimal. As the LAI
increases, the interception probability increases more rapidly for small Sun
zenith angles and the “reflectance peak” observed around 12 hours is less
pronounced. The high soil brightness is responsible for the same behavior in
the NIR band, despite of the importance of multiple scattering (Figure 6).
These frequent observations ease the discrimination between soil and
vegetation reflectance contributions for different LAI values. They should
also permit to verify that identical LAI values are retrieved over a short time
period, under varying illumination conditions. In turn, this constraint can be
exploited for the design of relevant quality control mechanism.
4.
DISCUSSION
In Section 2, we saw that validation and quality control are key activities
of operational remote sensing product extractions, as they ensure a
meaningful use of the disseminated information. Hence, the design of new
retrieval algorithms should include the associated quality control
mechanisms. Current space-borne polar observations of surfaces in the solar
spectral region suffer from an under-sampling of the surface anisotropy. This
poor angular sampling prevents the use of these observations for quantitative
applications with advanced and reliable algorithms. SEVIRI observations
will permit a better documentation of the radiation transfer processes in the
atmosphere-vegetation-soil system as a function of the Sun’s position, and
therefore provide improved constraints to solve the inverse problem. In turn,
this advantage can be used for the development of relevant quality control
mechanisms that generate information on the retrieved product reliability.
Through a simple but illustrative case study, we demonstrated that
geostationary observations, as will be provided by SEVIRI, will permit the
development of such advanced retrieval algorithms, thanks to a proper
sampling of the angular effects.
ALGORITHMS FOR GEOSTATIONARY SATELLITE DATA
PROCESSING
87
observed reflectance around 12 hours, i.e., when the Sun zenith angle is
minimum, decreases more rapidly than reflectances observed in the morning
or in the afternoon. At noon, the trajectory of photons that reach the soil
under the vegetation canopy and are reflected back to the sensor is shortest.
It translates into high reflectances, especially when the LAI is small. The
probability to be intercepted by the vegetation is indeed minimal. As the LAI
increases, the interception probability increases more rapidly for small Sun
zenith angles and the “reflectance peak” observed around 12 hours is less
pronounced. The high soil brightness is responsible for the same behavior in
the NIR band, despite of the importance of multiple scattering (Figure 6).
These frequent observations ease the discrimination between soil and
vegetation reflectance contributions for different LAI values. They should
also permit to verify that identical LAI values are retrieved over a short time
period, under varying illumination conditions. In turn, this constraint can be
exploited for the design of relevant quality control mechanism.
4.
DISCUSSION
In Section 2, we saw that validation and quality control are key activities
of operational remote sensing product extractions, as they ensure a
meaningful use of the disseminated information. Hence, the design of new
retrieval algorithms should include the associated quality control
mechanisms. Current space-borne polar observations of surfaces in the solar
spectral region suffer from an under-sampling of the surface anisotropy. This
poor angular sampling prevents the use of these observations for quantitative
applications with advanced and reliable algorithms. SEVIRI observations
will permit a better documentation of the radiation transfer processes in the
atmosphere-vegetation-soil system as a function of the Sun’s position, and
therefore provide improved constraints to solve the inverse problem. In turn,
this advantage can be used for the development of relevant quality control
mechanisms that generate information on the retrieved product reliability.
Through a simple but illustrative case study, we demonstrated that
geostationary observations, as will be provided by SEVIRI, will permit the
development of such advanced retrieval algorithms, thanks to a proper
sampling of the angular effects.