International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
COMPARISON OF MICROWAVE BACKSCATTER ANISOTROPY
PARAMETERISATIONS OVER THE ANTARCTIC ICE SHEET
Alexander D. Fraser“, Neal W. Young“ ° and Neil Adams“
? Antarctic Climate & Ecosystems Cooperative Research Centre
University of Tasmania
Private Bag 80
Hobart, Tasmania, Australia, 7001
adfraser@utas.edu.au
b Australian Antarctic Division
Channel Highway
Kingston, Tasmania, Australia, 7050
^ Australian Bureau of Meteorology
GPO Box 727
Hobart, Tasmania, Australia, 7001
Commission VIII/10
KEY WORDS: Snow, Ice, Glaciology, Research, Active, Microwave, Satellite
ABSTRACT:
The Antarctic Ice Sheet exhibits a strong anisotropy in microwave backscatter, both as a function of azimuth angle and incidence
angle. This anisotropy arises as a result of i) the alignment of roughness elements and other wind-related surface and sub-surface
features, as well as ii) internal layers and snow grain size gradient within the snowpack. As a result of its antenna configuration, the
European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Advanced SCATterometer (ASCAT) satellite
instrument is able to observe much of the continent with a large azimuth and incidence angle diversity. A lack of azimuth and incidence
diversity has restricted previous backscatter parameterisations to relatively simple bi-sinusoidal (azimuth angle) and linear (incidence
angle) parameterisations. Using ASCAT, we show that a better fit can be obtained using a cubic incidence angle function and a Fourier
series of up to four terms for parameterisation of the azimuth angle anisotropy. Scatterometer instruments have previously been used
in Greenland to retrieve accumulation rate by observing the change in backscatter as a function of incidence angle. Here we present
preliminary results of an empirical study linking the isotropic component and incidence angle dependence to snow accumulation rate
in Antarctica, using snow stake measurements as ground truthing.
1 INTRODUCTION
Surface mass balance (SMB) is the net input term for ice mass
balance calculations and is thus a critical parameter for accurate
estimation of the Antarctic Ice Sheet (AIS) contribution to sea
level rise (Allison et al., 2009). Despite this, reliable in-situ mea-
surements of SMB are sparse throughout the AIS, particularly in
the interior of the continent where few staffed stations exist. The
mass input to the AIS can be determined from knowledge of the
spatial distribution of SMB (the spatial integral of which gives the
total mass input), and the temporal variations of this. Spatial ex-
tensions of sparse in-situ data can be provided by either modeling
or large-scale observation. Recent advances in atmospheric mod-
eling capability (e.g., Lenaerts et al. (2012) included the effects of
snow redistribution due to blowing snow) have given rise to high-
resolution, large-scale maps of SMB throughout the AIS. Satellite
remote sensing of SMB provides an indepentent large-scale spa-
tial extension of in-situ point and line measurements such as snow
stakes (e.g., Higham et al. (1997)) and ground penetrating radar-
derived isochronous horizons (e.g., Miiller et al. (2010)). Passive
microwave thermal emission signals are sensitive to SMB, and
several studies have explored these links in detail, e.g., Winebren-
ner et al. (2001); Arthern et al. (2006). Empirically-derived re-
lationships between SMB and scatterometer measurements have
been studied in Greenland, but no equivalent study has yet been
conducted in Antarctica.
573
Scatterometers are typically spaceborne, active instruments orig-
inally designed to obtain microwave backscatter measurements
from small scale roughness on the surface of the ocean for multi-
ple azimuth angles. From this information, surface wind strength
and direction can be inferred. For approximately two decades,
spaceborne scatterometer instruments have also been applied to
the great ice sheets in Greenland (e.g., Long et al. (1992)) and
Antarctica (e.g., Remy et al. (1992)), in order to retrieve surface
and near-surface snow parameters. Large parts of the surface
of both the Greenland Ice Sheet (GrIS) and Antarctic Ice Sheet
(AIS) are characterised by sastrugi (wind-aligned snow structural
features of magnitude depending on wind speed at the time of for-
mation, rarely greater than 50 cm in height (Kobayashi, 1980)).
Microwave backscatter is enhanced by several dB when look-
ing across these roughness features (Furukawa and Young, 1997).
For non-nadir microwave observations of the great ice sheets, a
large azimuthal diversity of observation is essential, in order to
accurately characterise this strong azimuthal anisotropy.
A more in-depth review of past ice sheet applications of scat-
terometer data to be provided in a forthcoming publication (Fraser
et al., in prep.) is summarised here. Much of the early scatterom-
eter work on ice sheets was focused around retrieval of sastrugi
parameters (magnitude and direction) using the bi-sinusoidal az-
imuthal modulation, e.g., Ledroit et al. (1993); Young et al. (1996);
Furukawa and Young (1997); Hyland and Young (1998); Long
