Geology Reference
In-Depth Information
implies an assumption that the stage of ice growth leaves
its “imprint” on the properties of the subsurface layer (i.e.,
the penetration layer where electromagnetic wave interacts
with the medium and captures information about it). This
is more pronounced in some cases such as the bubbly
subsurface layer that characterizes MY hummock ice
(section 4.4.3) or the very smooth surface that character-
izes Nilas. If this assumption is not satisfied, then the
radiometric data sets should better be linked to surface
conditions that actually engender the observations. In
other words data should be used to identify surface con-
ditions and parameters (e.g., bare versus snow covered
surface, rough versus smooth ice, flooded versus dry sur-
face, snow conditions, or skin temperature) rather than
bulk ice properties. Some data sets presented in this
chapter are oriented to support this criterion.
Radarsat‐1 SAR launch), the number of SAR ice back-
scatter studies peaked. A commonly adapted approach
involved linking the acquired backscatter data to ground
observations and measurements from snow‐covered
ice types. In addition to compiling backscatter signatures
of sea ice types, the aim was to develop empirical equa-
tions to relate the backscatter to ice types and surface
conditions. Ice classification was the main drive behind
these activities. However, the pressing question that occu-
pied the minds of the researchers at that time was about
the information contents of the observed backscatter sig-
nal: What were the dominant factors that trigger the
backscatter?
The rest of this section includes samples of backscat-
ter data from selected sources that span a long period
starting from 1980s until recently. The backscatter data
are grouped according to the traditional thickness‐based
ice types, but some data are grouped according to sur-
face conditions. In addition to the data sources used in
this chapter, other sources that are worth checking are
worth referring to. Kwok and Cunningham [1994] com-
piled a limited set of backscatter data from Arctic sea ice
types using ERS‐1 SAR. Onstott et al. , [1998] obtained
backscatter signature of artificial young ice types grown
in the outdoor tank at CRREL using a ground‐based
scatterometer. Drinkwater [1998] compiled a regional
and seasonal database of backscatter from Antarctic sea
ice using the C‐band scatterometer (20 km resolution)
onboard both ERS satellites. The data were compared to
ship‐borne radar measurements. Measurements showed
a large variability of backscatter from the same ice type
when surface conditions are different. For example, the
surface of Gray ice type may feature rough pancakes.
Similarly, Nila sheets may feature frost flower cover. In
each case the backscatter deviates from the “typical”
value of the original surface.
A first attempt to explore the possibility of mapping
Arctic sea ice using backscatter from a scatterometer is
presented in Ezraty and Cavanie [1999]. Going from FY
ice to MY ice, they observed a greater dynamic range of
radar backscattering in the Ku‐band (on NSCAT) than
the C band (on ERS‐1). They also observed a decrease
in backscatter caused by surface thawing during the
period from October 1996 to June 1997. Haas [2001]
compiled a seasonal backscatter database of Antarctic
perennial sea ice for the period between 1991 and 1999,
also from the scatterometers onboard ERS‐1 and ERS‐2
satellites. Nghiem and Bertoia [2001] measured polari-
metric radar backscatter signature of sea ice in the
Arctic during winter (March 1998) using an airborne
system on NASA's DC‐8 aircraft. They also investigated
the diurnal effects on sea ice backscatter and found it to
be significant.
8.1. RadaR BackscatteR
After the remarkable success of the first space‐borne
SAR onboard the short‐lived Seasat in 1979, it was rightly
felt at the time that radar was well suited as an opera-
tional tool for sea ice monitoring in polar oceans. As a
result, many research programs were dedicated to explore
the backscatter signature of different ice types with
different surface conditions in different seasons using a
variety of sensor's viewing geometry. That was a widely
accepted approach. A few programs were conducted
throughout the 1980s using surface‐based scatterometers
to measure radar backscatter from natural Arctic sea ice
[e.g., Onstott , 1979; Onstott and Gogineni , 1985]. A list of
surface‐based scatterometer investigations on natural
and laboratory‐grown sea ice conducted in the in the late
1970s and 1980s is presented in Onstott [1992]. A continu-
ation list that covers the period from 1989 to 2006 is pre-
sented in Geldsetzer et al. [2007]. Radar backscatter
measurements from thin ice types were also conducted on
laboratory‐grown ice in outdoor facilities. Of particular
importance were the two series of experiments conducted
in the Cold Regions Research and Engineering Laboratory
(CRREL), located in Hanover, New Hamsphere. The
first was conducted in the mid‐1980s under the title
CRREL Experiment (CRRELEX) [ Swift et  al. , 1992],
and the second was initiated by the U.S. Office of Naval
Research (ONR) in 1992 under the title Accelerated
Research Initiative (ARI) to study the electromagnetic
properties of sea ice [ Jezek et al. , 1998]. Airborne SAR
systems such as NASA's AirSAR and the CCRS's
Convair‐580 were used to acquire backscatter data from
ice types mainly on the western Arctic around Alaska
and the east coast of Canada respectively.
After the launch of SAR onboard ERS‐1 in1991 and
its successor on ERS‐2 in 1995 (almost coincident with
Search WWH ::




Custom Search