11 Snow and Ice
247
application of higher frequency data for ice thickness detennination has been investigated by Chang et ai. (1997) using airborne microwave data and a layer radiative
transfer model to calculate ice thickness, however, the small sample size limited
statistically significant results.
Active microwave remote sensing has an advantage over passive techniques for
lake ice information retrieval in that the higher spatial resolution means that lakes of
all sizes can be monitored. In addition to providing spatial and temporal information
on lake ice freeze-up and break-up and ice thickness, radar backscatter is related to
the internal structure of the lake ice. Hall (1993) provides an overview of the capability of active microwave remote sensing for retrieval of lake ice information. The
application of satellite SAR for lake ice remote sensing has been investigated and
demonstrated in recent years using data from the ERS-l satellite (e.g., Hall et aI.,
1994; Jeffries et aI., 1994; Leshkevich et aI., 1994).
Advantages. Passive microwave data provides several advantages not offered by
other satellite sensors. Studies have shown that passive microwave data offer the
potential to extract meaningful snowcover information, such as SWE, depth, extent
and snow state. SSM/I is a part of an operational satellite system, providing daily
coverage of most snow areas, with multiple passes at high latitudes, hence allowing
the study of diurnal variability. The technique has generally all-weather capability
(although affected by precipitation at 85GHz), and can provide data during darkness.
The data are available in near-real time, and hence can be used for hydrological
forecasting. SAR has an additional advantage of having resolution of about 25 m
making it very useful for mountain snowpacks.
Limitations. There are limitations and challenges in using microwave data for
deriving snow cover information for hydrology. The coarse resolution of passive
microwave satellite sensors such as SMMR and SSM/I (-25km) is more suited to
regional and large basin studies, although Rango et ai. (1989) did fmd that reasonable
SWE estimates could be made for basins ofless than 1O,OOOkm 2 . Heterogeneity of the
surface and the snowcover within the microwave footprint results in a mixed signature, which is ultimately represented by a single brightness temperature that is an
areally weighted mean of the microwave emission from each surface type within the
footprint. Hence, an understanding of the relationship between snow cover and
surface terrain and land cover (e.g., Goodison et aI., 1981) is as important for developing remote sensing applications in hydrology as for conventional hydrology analyses in snow covered areas.
Another challenge is to incorporate the effect of changing snowpack conditions
throughout the winter season. Seasonal aging, or metamorphism, results in a change
in the grain size and shape, and this will affect the microwave emission from the
snowpack. In very cold regions, depth hoar characterized by its large crystal structure
enhance the scattering effect on the microwave radiation, resulting in lower surface
emission producing an overestimate of SWE or snow depth (Hall, 1987 and Armstrong et aI., 1993). The increase in T B associated with wet snow conditions currently
prevents the quantitative determination of depth or water equivalent since algorithms
will tend to produce zero values under these conditions. The best way to view the
247
application of higher frequency data for ice thickness detennination has been investigated by Chang et ai. (1997) using airborne microwave data and a layer radiative
transfer model to calculate ice thickness, however, the small sample size limited
statistically significant results.
Active microwave remote sensing has an advantage over passive techniques for
lake ice information retrieval in that the higher spatial resolution means that lakes of
all sizes can be monitored. In addition to providing spatial and temporal information
on lake ice freeze-up and break-up and ice thickness, radar backscatter is related to
the internal structure of the lake ice. Hall (1993) provides an overview of the capability of active microwave remote sensing for retrieval of lake ice information. The
application of satellite SAR for lake ice remote sensing has been investigated and
demonstrated in recent years using data from the ERS-l satellite (e.g., Hall et aI.,
1994; Jeffries et aI., 1994; Leshkevich et aI., 1994).
Advantages. Passive microwave data provides several advantages not offered by
other satellite sensors. Studies have shown that passive microwave data offer the
potential to extract meaningful snowcover information, such as SWE, depth, extent
and snow state. SSM/I is a part of an operational satellite system, providing daily
coverage of most snow areas, with multiple passes at high latitudes, hence allowing
the study of diurnal variability. The technique has generally all-weather capability
(although affected by precipitation at 85GHz), and can provide data during darkness.
The data are available in near-real time, and hence can be used for hydrological
forecasting. SAR has an additional advantage of having resolution of about 25 m
making it very useful for mountain snowpacks.
Limitations. There are limitations and challenges in using microwave data for
deriving snow cover information for hydrology. The coarse resolution of passive
microwave satellite sensors such as SMMR and SSM/I (-25km) is more suited to
regional and large basin studies, although Rango et ai. (1989) did fmd that reasonable
SWE estimates could be made for basins ofless than 1O,OOOkm 2 . Heterogeneity of the
surface and the snowcover within the microwave footprint results in a mixed signature, which is ultimately represented by a single brightness temperature that is an
areally weighted mean of the microwave emission from each surface type within the
footprint. Hence, an understanding of the relationship between snow cover and
surface terrain and land cover (e.g., Goodison et aI., 1981) is as important for developing remote sensing applications in hydrology as for conventional hydrology analyses in snow covered areas.
Another challenge is to incorporate the effect of changing snowpack conditions
throughout the winter season. Seasonal aging, or metamorphism, results in a change
in the grain size and shape, and this will affect the microwave emission from the
snowpack. In very cold regions, depth hoar characterized by its large crystal structure
enhance the scattering effect on the microwave radiation, resulting in lower surface
emission producing an overestimate of SWE or snow depth (Hall, 1987 and Armstrong et aI., 1993). The increase in T B associated with wet snow conditions currently
prevents the quantitative determination of depth or water equivalent since algorithms
will tend to produce zero values under these conditions. The best way to view the