vegetated and non-vegetated areas, and capture the overall condition, magnitude,
and phenology of vegetation growth. Additional vegetation indices are formulated
on essentially the similar principle as NDVI with additional improvement such as
accounting for atmospheric noise through adding the blue band information to
specific indices, etc. Examples of these alternative vegetation indices include the
soil adjusted vegetation index (SAVI, Huete 1988), the atmospherically resistant
vegetation index (ARVI, Kaufman and Tanre 1992), and the soil and atmospherically resistant vegetation index (SARVI, Huete and Liu 1994), which integrates the former two algorithms. These efforts culminated in the development of
the enhanced vegetation index (EVI, Huete and Justice 1999), which was specifically targeted at improved vegetation monitoring through the MODIS platforms. The EVI corrects both soil and atmospheric effects akin to SARVI and has
improved sensitivity to high biomass, which leads to NDVI saturation (Huete et al.
2002). Both NDVI and EVI are widely employed vegetation indices and are used
for LSP metrics derivation and studies. For boreal regions where snow cover
affects vegetation recognition, the normalized difference water index (NDWI; Gao
1996) is useful for more accurate LSP analysis, as NDWI decreases with snowmelt
and increases with canopy growth (from leaf water content) versus NDVI, which
increases with snowmelt and therefore introduces biases if used alone (Delbart
et al. 2005). Recent development also includes an attempt to combine the use of
NDVI and the normalized difference infrared index (NDII) to produce a remote
sensing phenology index (PI) to better overcome the background soil and snow
contaminations (Gonsamo et al. 2012),
Biophysical variables corresponding to detailed vegetation processes can also
be estimated from satellite data and in turn used for LSP metric development. Such
variables include notably the leaf area index (LAI) and fraction of absorbed
photosynthetically active radiation (FAPAR or FPAR) as described in another
chapter of this book. The field observation-originated LAI can be remotely sensed
and is more directly related to vegetation properties and functions, such as
structure, evapotranspiration, and primary production. The FPAR relates to similar
vegetation activities and is a basis for remote sensing of the gross and net primary
productions (GPP and NPP; Running et al. 2004). Chapters 2 and 5 of this book
outline the details of these aspects of vegetation parameterization. Therefore, in
addition to the use of vegetation indices, there were studies deriving LSP metrics
from LAI and/or FPAR (Ahl et al. 2006; Kang et al. 2003; Wang et al. 2005). The
LAI and/or FPAR are strongly correlated with NDVI. In particular, LAI has an
approximately linear correspondence with NDVI when LAI is low, yet a highly
non-linear relationship when LAI is higher (http://earthobservatory.nasa.gov/
Features/LAI/LAI3.php). The NDVI, along with EVI, is still most commonly used
for generating phenology information from data acquired through earth resource
satellites.
The remote sensing data used for LSP monitoring and studies are primarily
derived from multispectral sensors onboard sun-synchronous polar orbiting satellites. The AVHRR onboard NOAA series polar orbiting environmental satellites
(POES) has provided long-term global NDVI products since the early 1980s
104
J. M. Hanes et al.
and phenology of vegetation growth. Additional vegetation indices are formulated
on essentially the similar principle as NDVI with additional improvement such as
accounting for atmospheric noise through adding the blue band information to
specific indices, etc. Examples of these alternative vegetation indices include the
soil adjusted vegetation index (SAVI, Huete 1988), the atmospherically resistant
vegetation index (ARVI, Kaufman and Tanre 1992), and the soil and atmospherically resistant vegetation index (SARVI, Huete and Liu 1994), which integrates the former two algorithms. These efforts culminated in the development of
the enhanced vegetation index (EVI, Huete and Justice 1999), which was specifically targeted at improved vegetation monitoring through the MODIS platforms. The EVI corrects both soil and atmospheric effects akin to SARVI and has
improved sensitivity to high biomass, which leads to NDVI saturation (Huete et al.
2002). Both NDVI and EVI are widely employed vegetation indices and are used
for LSP metrics derivation and studies. For boreal regions where snow cover
affects vegetation recognition, the normalized difference water index (NDWI; Gao
1996) is useful for more accurate LSP analysis, as NDWI decreases with snowmelt
and increases with canopy growth (from leaf water content) versus NDVI, which
increases with snowmelt and therefore introduces biases if used alone (Delbart
et al. 2005). Recent development also includes an attempt to combine the use of
NDVI and the normalized difference infrared index (NDII) to produce a remote
sensing phenology index (PI) to better overcome the background soil and snow
contaminations (Gonsamo et al. 2012),
Biophysical variables corresponding to detailed vegetation processes can also
be estimated from satellite data and in turn used for LSP metric development. Such
variables include notably the leaf area index (LAI) and fraction of absorbed
photosynthetically active radiation (FAPAR or FPAR) as described in another
chapter of this book. The field observation-originated LAI can be remotely sensed
and is more directly related to vegetation properties and functions, such as
structure, evapotranspiration, and primary production. The FPAR relates to similar
vegetation activities and is a basis for remote sensing of the gross and net primary
productions (GPP and NPP; Running et al. 2004). Chapters 2 and 5 of this book
outline the details of these aspects of vegetation parameterization. Therefore, in
addition to the use of vegetation indices, there were studies deriving LSP metrics
from LAI and/or FPAR (Ahl et al. 2006; Kang et al. 2003; Wang et al. 2005). The
LAI and/or FPAR are strongly correlated with NDVI. In particular, LAI has an
approximately linear correspondence with NDVI when LAI is low, yet a highly
non-linear relationship when LAI is higher (http://earthobservatory.nasa.gov/
Features/LAI/LAI3.php). The NDVI, along with EVI, is still most commonly used
for generating phenology information from data acquired through earth resource
satellites.
The remote sensing data used for LSP monitoring and studies are primarily
derived from multispectral sensors onboard sun-synchronous polar orbiting satellites. The AVHRR onboard NOAA series polar orbiting environmental satellites
(POES) has provided long-term global NDVI products since the early 1980s
104
J. M. Hanes et al.