188
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
as an input: the MODIS (Level 1B) radiance data (Guenther et al. 2002), the MODIS
cloud mask (Ackerman et al. 1998; Platnick et al. 2003), and the MODIS geolocation
product for latitude and longitude, viewing geometry data and the land/water mask
(Wolfe et al. 2002). Only the general methodology is presented in this chapter. Full
details of the mapping algorithm are available in the “Algorithm Theoretical Basis
Document” (Hall et al. 2001) and can be seen at the NSIDC and MODIS Snow and
Sea Ice Global Mapping Project Web pages (http://www.nsidc.org and http://modissnow-ice.gsfc.nasa.govweb).
The mapping approach exploits the high reflectance in the visible and the low
reflectance in the shortwave infrared part of the spectrum by the normalized difference snow index (NDSI; Hall et al. 2001). The NDSI allows us to distinguish snow
from many other surface features such as clouds that have high reflectance in both
the visible and the shortwave infrared parts of the spectrum (Hall et al. 1998). The
NDSI can usually separate cumulus clouds from snow, but it cannot always separate
optically thin cirrus clouds (Hall and Riggs 2007). For Terra data, the NDSI calculation is based on MODIS bands 4 (0.55 µm) and 6 (1.6 µm):
NDSI TERRA = (Band 4 – Band 6)/(Band 4 + Band 6).
(9.1)
MODIS band 6 detectors failed on Aqua shortly after launch, so band 7 (2.1 µm) is
used instead to calculate the NDSI for Aqua (Hall et al. 2000, 2003):
NDSI AQUA = (Band 4 – Band 7)/(Band 4 + Band 7),
(9.2)
where “Band” stands for the reflectance of the channel. The fractional snow cover
map is estimated based on the regression technique (Salomonson and Appel 2004).
The fractional area (in percent) of each pixel covered by snow is calculated for both
land and inland water bodies not covered by clouds and over the range of NDSI
values from 1 to 100. Fractional snow may be mapped over the whole NDSI range
indicative of snow (Salomonson and Appel 2006).
Snow Fraction = –0.01 + 1.45 × NDSI.
(9.3)
The MODIS snow mapping algorithm is automated, which means that a consistent data set may be generated for long-term climate studies that require snow cover
information (Hall et al. 2002a). Its main advantages rest on the fact that there is very
efficient tradeoff between spatial and temporal resolution and mapping accuracy
and that it is adaptable to a range of illumination conditions. The main limitation
is that there is always inevitable missing information during cloud coverage and
beneath dense forest canopies (Hall et al. 2001). The reduction of clouds is possible,
and the potential methods are summarized in detail in Section 9.4. Mapping snow in
forested locations is based upon a combination of the normalized difference vegetation index (NDVI) and the NDSI (Hall et al. 1998). Applications of the NDVI allow
for the use of different NDSI thresholds for forested and nonforested pixels without
compromising the algorithm performance for other land cover types. However, such
a mapping approach can only be applied to the Terra data. The NDSI/NDVI test for
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
as an input: the MODIS (Level 1B) radiance data (Guenther et al. 2002), the MODIS
cloud mask (Ackerman et al. 1998; Platnick et al. 2003), and the MODIS geolocation
product for latitude and longitude, viewing geometry data and the land/water mask
(Wolfe et al. 2002). Only the general methodology is presented in this chapter. Full
details of the mapping algorithm are available in the “Algorithm Theoretical Basis
Document” (Hall et al. 2001) and can be seen at the NSIDC and MODIS Snow and
Sea Ice Global Mapping Project Web pages (http://www.nsidc.org and http://modissnow-ice.gsfc.nasa.govweb).
The mapping approach exploits the high reflectance in the visible and the low
reflectance in the shortwave infrared part of the spectrum by the normalized difference snow index (NDSI; Hall et al. 2001). The NDSI allows us to distinguish snow
from many other surface features such as clouds that have high reflectance in both
the visible and the shortwave infrared parts of the spectrum (Hall et al. 1998). The
NDSI can usually separate cumulus clouds from snow, but it cannot always separate
optically thin cirrus clouds (Hall and Riggs 2007). For Terra data, the NDSI calculation is based on MODIS bands 4 (0.55 µm) and 6 (1.6 µm):
NDSI TERRA = (Band 4 – Band 6)/(Band 4 + Band 6).
(9.1)
MODIS band 6 detectors failed on Aqua shortly after launch, so band 7 (2.1 µm) is
used instead to calculate the NDSI for Aqua (Hall et al. 2000, 2003):
NDSI AQUA = (Band 4 – Band 7)/(Band 4 + Band 7),
(9.2)
where “Band” stands for the reflectance of the channel. The fractional snow cover
map is estimated based on the regression technique (Salomonson and Appel 2004).
The fractional area (in percent) of each pixel covered by snow is calculated for both
land and inland water bodies not covered by clouds and over the range of NDSI
values from 1 to 100. Fractional snow may be mapped over the whole NDSI range
indicative of snow (Salomonson and Appel 2006).
Snow Fraction = –0.01 + 1.45 × NDSI.
(9.3)
The MODIS snow mapping algorithm is automated, which means that a consistent data set may be generated for long-term climate studies that require snow cover
information (Hall et al. 2002a). Its main advantages rest on the fact that there is very
efficient tradeoff between spatial and temporal resolution and mapping accuracy
and that it is adaptable to a range of illumination conditions. The main limitation
is that there is always inevitable missing information during cloud coverage and
beneath dense forest canopies (Hall et al. 2001). The reduction of clouds is possible,
and the potential methods are summarized in detail in Section 9.4. Mapping snow in
forested locations is based upon a combination of the normalized difference vegetation index (NDVI) and the NDSI (Hall et al. 1998). Applications of the NDVI allow
for the use of different NDSI thresholds for forested and nonforested pixels without
compromising the algorithm performance for other land cover types. However, such
a mapping approach can only be applied to the Terra data. The NDSI/NDVI test for