Dome C as a Candidate Cal/Val
Site for SMOS
Inputs by Frederique Remy and Mark Drinkwater
Dome Concordia (Dome C) is a broad topographic dome roughly centred
at 75° 06'06"S, 123° 23'42"E on the polar plateau of East Antarctica
(at 3233 m elevation a.s.l.), and is situated more than 700 km from the
coast. This location is about 65 km south of the old U.S. Dome C camp,
and has been selected as the optimal site for a new collaborative European
(EPICA) deep ice core [Tabacco and others, J. Glac, 2000]. The chosen core
site will allow recovery of a core to a depth of 3250m with a climate history
of some 400,000 years. In conjunction, the French and Italians have agreed
to cooperate in the establishment of a research programme, including construction
and operation of a scientific base "Concordia". The use of this station
for scientific research is open to the world-wide scientific community
and is considered here as a candidate calibration site for SMOS.
The objectives of a calibration/validation
project at this site is twofold:
This site is well suited for cal/val. activity for the following reasons:
To take advantage of this well documented and well suited site for calibration
and validation or long-term stability of satellite instruments.
To provide the scientific community a unique satellite and in-situ data
base with which to better interpret the ice core.
Cold, dry snow is relatively transparent to microwaves, ensuring relative
stability of the emissivity. The penetration of microwaves in layered Antarctic
firn has been estimated using Ku-band radar altimeter data [Legresy and
Remy, 1998], and using a combination of C-band scatterometer data and empirical
scattering models [Bingham and Drinkwater, 2000]. At Ku-band, the estimated
penetration depth is of the order of 6-7 m, and at P-band is of the order
of 4000 m. L-band extinction is dominated primarily by temperature-regulated
absorption in the upper 10 m (dependent on imaginary part of complex permittivity
of snow/air mixture). The resulting penetration depth is estimated to be
of the order of 100-150 m. The near-surface layer (i.e. top 10m), although
experiencing seasonal temperature variability is largely transparent at
L-band, and so the emissivity variability should be reasonably small. The
primary thermally-forced brightness temperature variability (caused by
surface air temperature generated waves in the vertical temperature profile),
and is known, provided in-situ data is available. The seasonal cycle in
air temperature at Dome C is of the order of 30° K. Short time-scale
diurnal variability caused during polar daylight hours is low-pass filtered
by the thermal conductivity of the snow, and lagged response of bulk firn
temperature. Below 10m the firn temperature does not exhibit any seasonal
homogeneity of snow surface at the 100 km scale
topography is known from satellite altimetry to high precision at 100km
surface roughness is minimal relative to other ice sheet locations
the sky is clear, and the atmosphere is extremely dry and stable
ionospheric activity is minimal (and daily and long term variation in Total
Electron Content is minimized)
snow accumulation is low (around 3.7cm/yr).
Recent data analyses by Bingham and Drinkwater  have identified
long-term climate/calibration drifts of SSM/I over the ice sheet (when
modeled seasonal variations are removed) at high frequencies. These imply
a long-term drifts of the order of XX K on time scales of XX years, whilst
exhibiting short-term anomalies in Tb which appear to result from precipitation
events or changes in the snow interface. The present revisit of SMOS at
high latitudes suggest N?? overlapping measurements with an accuracy of
?? Kelvins, OR alternatively XX-day mean Tbs with an accuracy of ±??
The Dome C Site is well characterized,
Further in-situ measurement activities are planned in the frame of the
10 years of consistent automatic weather station data (air temperature,
pressure, wind speed and direction)
Topography is known at the kilometric scale (from ERS geodetic phase altimetry
and GPS measurements), and ICESAT will provide metre accuracy laser
profiles over Dome C in 2001-2002.
Bedrock is mapped at the same kilometer scale using P-band ground-penetrating
radar, by subtracting ice thickness (along profiles) from the detailed
altimeter topographic profiles.
Ice surface velocity field is characterized by ERS SAR interferometry,
GPS, and DORIS tracking.
Ice core data provides the mean snow accumulation rate, and density profile
Embedded thermistor strings provide firn temperature profile temporal variability.
Ancillary data sets exist from microwave radar scatterometers (5 and 13.6
GHz), radiometers (5, 19, 22, 37, 85 GHz), and radar altimeters (13.6 GHz).
Priority Objectives (for Science Definition
Back to main page
Establish temporal and spatial variability in SMMR C-band Tb over Dome
C, and establish time-space correlation statistics/length scales in this
region and rms variability information (*this should be done in view of
the SMOS temporal revisit which is possible at high latitudes).
Establish seasonal variation in Tb due to seasonal temperature cycle
Correct for temperature and investigate variation in emissivity (independent
of physical temperature).
Establish whether radiative transfer models for ice sheet can be generalized
to L-band (is dense-medium theory still appropriate, or are simple approximations
Establish whether the Dome C ice sheet target is stable enough to check
for long term stability and drift in total power (i.e. calibration).
Is Dome C appropriate as a distributed target for checking consistency
of image reconstruction, by using known trajectories of target through
the FOV (in incidence and azimuth - through simulated swaths).
Can the stability of the ice sheet emission be used to test the consistency
in reconstructed brightness temperatures (for a reasonably uniform radiator)
Does radiometric anisotropy in the surface matter at L-band for this pixel
scale? If so, extrapolate existing definitions of anisotropy from NSCAT/SSMI/QuikScat
data to define modulation pattern (as a function of local incidence and
azimuth angle), and define Tb trajectories and range of variability according
to relative orbit swath direction relative to surface anisotropy.
Can Dome C indicate what the averaging requirement is to achieve a known
level of accuracy such as a 0.1 K criterion (i.e. defined reproducibility
in Tb over time).