Retrieval of Ice and Snow Geophysical Parameters - Sea Ice: Physics and Remote Sensing - page 467

laser beam is reflected off the snow surface. Therefore,
the freeboard includes the snow depth. On the other
hand, the transmitted radar pulse is reflected off the
snow‐ice interface. Therefore, the freeboard in the case of
radar altimeter does not include snow depth. This was
verified in a few laboratory experiments. Beaven et al .
[1995] have shown that, under dry cold snow, an incident
radar signal of 13.4 GHz at nadir is reflected off the
snow‐ice interface. Both laser and radar altimeters meas-
ure the freeboard from a reference of available water sur-
face. This can be found usually in leads or OW area within
the ice cover. The following discussions address laser and
radar altimetery separately with reference to the param-
eters presented in Figure 10.34.
For laser altimeter the equation that has to be solved to
retrieve the ice thickness H , given the measured freeboard
f laser , is written as

Antarctic ice sheets [ Zwalley et al ., 2002b], but it was used
during its lifetime to determine polar sea ice thickness
profiles along the ground track of the satellite. GLAS is a
profile sensor that acquires data from narrow stripes
of 70 m width (cross track) with footprints 3 km apart.
Images can be constructed from sequential satellite
overpasses. A general overview of the ice thickness esti-
mation using ICESat‐1 data, including a description of
the freeboard retrieval approach, is given in Kwok et al .
[2004]. A review of the achievements of ICESat‐1 is
presented in Kwok [2010]. According to Kwok et al . [2007],
ice thickness was calculated over the Arctic region from
ICESat‐1 during a number of data acquisitions cam-
paigns (a total of 10 were achieved by the end of February
2007). Each campaign spanned approximately one 33 day
subcycle of the 91 day repeat cycle of the satellite's orbit.
An example of ice thickness profile along a 160 km
ICESat‐1 ground track north of Ellesmere Island in the
Arctic is presented in Kwok et al . [2004] and reproduced
in Figure 10.35. The coincident Radarsat image is
included in the figure to provide a closer inspection of the
sea ice feature along the ICESat‐1 track. The image is
resampled into 150 m pixel spacing (i.e., at approximately
the same sample spacing of ICESat‐1). The established
freeboard thickness was achieved using a reference of
water surface obtained from open leads or leads covered
with very thin ice. The primary ice types that appear in
the Radarsat image are thick FY ice and MY ice. The
dark objects are either open water or refrozen smooth
thin ice in leads. The figure highlights the use of the sub-
jective analysis of Radarsat data as a tool to verify the ice
thickness retrieval. The precision of the ice thickness
retrieval over a flat ice surface is reported by the authors
to be about 2 cm. This accuracy allows for the derivation
of surface roughness. However, unknown snow depth
introduces a large uncertainty in the thickness estimates.
In an extended study on the accuracy of ICESat‐1, Kwok
et al . [2007] concluded that the accuracy depends on the
approach of retrieving the reference level of the seawater.
They studied three approaches in detail and concluded
that using new open leads in the ice sheet, identified either
in ICESat or SAR data, provide the best sea surface level
reference .
Although thickness estimates from the laser altimeter
are produced at a resolution acceptable for ship naviga-
tion, the satellite overpasses provide measurements only
along the ground track of the satellite. This means that
filling gaps to produce a regional map of ice thickness
requires data from numerous passes that can only be
acquired over a long period. This is operationally unac-
ceptable. The real value of the laser altimeter data resides
in providing climate‐related information on ice thickness
(and therefore volume) that reveals trends of interannual
variability at a synoptic scale (e.g., covering the entire

1

Hf Hh
Hh

i sn

(10.89)

1

laser

ow

where the subscripts i , sn and ow denote ice, snow, and
open water, respectively, and ρ is the density of the
medium. The snow depth is required to calculate the ice
thickness.
For radar altimeter the equation, which is based on the
buoyancy law, takes the form

1

Hf 1

i

hB (10.90)

radar i

ow

where ρ is the density and B i is the buoyancy of sea ice;

B i / (10.91)

snow ow i

The accuracy of the measured freeboard in equation
(10.89) and (10.90) is crucial because freeboard represents
only a small fraction of the ice thickness (10%-15%,
depending on the snow cover). Therefore, even a small
error in its measurement will be magnified when estimating
ice thickness. This is the most crucial step in converting
the radar altimeter measurements to ice thickness.
NASA's laser altimeter Geoscience Laser Altimeter
System (GLAS) onboard the Ice, Cloud and land
Elevation Satellite (ICESat‐1) [ Zwalley et al ., 2002b] was
the first system that mapped the ice thickness across the
polar regions. It was launched on 13 January, 2003 and
ended its mission on 14 August, 2010. The system oper-
ated at wavelengths of 1064 and 532 nm. The longer
wavelength was used for surface altimetry. The next gen-
eration of this satellite (ICESat‐2) is scheduled for launch
in 2016. The GLAS system was originally designed to
measure changes in the elevation of Greenland and