414
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
demonstrated by Becker and Choudhury (1988) and Justice et al. (1989). NDVI
mainly responds to a thin layer of the canopy (leaves), while a microwave index
includes information about both the leafy and woody parts of the vegetation due
to greater penetration and sensitivity. In the current work, it is our hypothesis that
the microwave-derived vegetation indices can provide vegetation information that is
complementary to that provided by optical sensors.
In deriving vegetation indices from satellite measurements using either optical or
microwave observations, we face two problems: the effects of the atmosphere and
the background (surface underlying the vegetation) signals. Microwave observations
are less affected by atmospheric conditions than traditional optical methods. On the
other hand, the variability in the background emission signals resulting from the
soil state can have a greater effect on the microwave observations than when using
optical sensors that only sense the canopy. Not accounting for the variability in the
background contribution is a major reason why the microwave vegetation indices
derived in the previous studies may have not been widely used in monitoring global
vegetation information. The effect of the background emission signal must be incorporated to derive a reliable and useful vegetation index.
In this chapter, we explore and demonstrate a new technique for deriving microchapter, we explore and demonstrate a new technique for deriving microwe explore and demonstrate a new technique for deriving microwave vegetation indices (MVIs) using passive microwave radiometer AMSR-E data
(Kawanishi et al. 2003). Unlike microwave vegetation indexes derived in previ. Unlike microwave vegetation indexes derived in previous studies, the MVIs derived here are independent of soil surface emission signals
and depend only on vegetation properties such as vegetation fractional coverage,
biomass, water content, temperature, the characteristics of the scatterer size, and
the geometry of the vegetation canopy. This method provides a new opportunity
to establish a long-term global dataset for monitoring vegetation cover using allweather passive microwave instruments. In the next section, the physical principles
and the methodology used to develop the new MVIs based on the AMSR-E sensor configuration are introduced. In Section 18.3, how the MVIs can be derived
from AMSR-E measurements is described, and their general characteristics are discussed as well. Comparisons of the MVIs and NDVI (Moderate Resolution Imaging
Spectroradiometer [MODIS]) in assessing the global vegetation pattern during different seasons and for specific land cover types are presented in Section 18.4, which
is followed by our conclusions obtained from this study in Section 18.5.
18.2 THEORETICAL BASIS OF THE NEW
MICROWAVE VEGETATION INDICES
18.2.1 MicRowave eMiSSion Model
The microwave emission model that we used to derive the microwave vegetation indices
is the ω–τ model derived from a zeroth-order radiative transfer solution (Ulaby et al.
1982). This model is commonly used to describe microwave signals at low frequencies
and employed as an inversion model to retrieve soil moisture information for AMSR-E
(Njoku et al. 2003, 2006). For a satellite footprint with a fraction of vegetation cover F v
at a given viewing angle and frequency f, the measured brightness temperature without
considering atmospheric effects can be written as a four-component model:
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
demonstrated by Becker and Choudhury (1988) and Justice et al. (1989). NDVI
mainly responds to a thin layer of the canopy (leaves), while a microwave index
includes information about both the leafy and woody parts of the vegetation due
to greater penetration and sensitivity. In the current work, it is our hypothesis that
the microwave-derived vegetation indices can provide vegetation information that is
complementary to that provided by optical sensors.
In deriving vegetation indices from satellite measurements using either optical or
microwave observations, we face two problems: the effects of the atmosphere and
the background (surface underlying the vegetation) signals. Microwave observations
are less affected by atmospheric conditions than traditional optical methods. On the
other hand, the variability in the background emission signals resulting from the
soil state can have a greater effect on the microwave observations than when using
optical sensors that only sense the canopy. Not accounting for the variability in the
background contribution is a major reason why the microwave vegetation indices
derived in the previous studies may have not been widely used in monitoring global
vegetation information. The effect of the background emission signal must be incorporated to derive a reliable and useful vegetation index.
In this chapter, we explore and demonstrate a new technique for deriving microchapter, we explore and demonstrate a new technique for deriving microwe explore and demonstrate a new technique for deriving microwave vegetation indices (MVIs) using passive microwave radiometer AMSR-E data
(Kawanishi et al. 2003). Unlike microwave vegetation indexes derived in previ. Unlike microwave vegetation indexes derived in previous studies, the MVIs derived here are independent of soil surface emission signals
and depend only on vegetation properties such as vegetation fractional coverage,
biomass, water content, temperature, the characteristics of the scatterer size, and
the geometry of the vegetation canopy. This method provides a new opportunity
to establish a long-term global dataset for monitoring vegetation cover using allweather passive microwave instruments. In the next section, the physical principles
and the methodology used to develop the new MVIs based on the AMSR-E sensor configuration are introduced. In Section 18.3, how the MVIs can be derived
from AMSR-E measurements is described, and their general characteristics are discussed as well. Comparisons of the MVIs and NDVI (Moderate Resolution Imaging
Spectroradiometer [MODIS]) in assessing the global vegetation pattern during different seasons and for specific land cover types are presented in Section 18.4, which
is followed by our conclusions obtained from this study in Section 18.5.
18.2 THEORETICAL BASIS OF THE NEW
MICROWAVE VEGETATION INDICES
18.2.1 MicRowave eMiSSion Model
The microwave emission model that we used to derive the microwave vegetation indices
is the ω–τ model derived from a zeroth-order radiative transfer solution (Ulaby et al.
1982). This model is commonly used to describe microwave signals at low frequencies
and employed as an inversion model to retrieve soil moisture information for AMSR-E
(Njoku et al. 2003, 2006). For a satellite footprint with a fraction of vegetation cover F v
at a given viewing angle and frequency f, the measured brightness temperature without
considering atmospheric effects can be written as a four-component model: