IAPRS & SIS, Vol.34, Part 7, "Resource and Environmental Monitoring", Hyderabad, India,2002
Johannesen et al., 2000). These results and their consequences
have thrown open interesting and intriguing questions about
modeling climate change in the face of envisaged greenhouse
warming.
In view of the great importance attached to the role of
secularly (slowly) changing sea ice cover over the polar
regions in relation to future of global climate, it is extremely
important that each of the pace-based PMRs providing sea
ice information be calibrated absolutely and inter-calibrated
relative to each other. Due to well known difficulties
associated with collection of sea truth at the desired spatial
scales in the harsh polar environment, most of the efforts have
concentrated on relative calibration between sensors. Space-
based Landsat and NOAA Visible/IR band high resolution
images of polar sea ice region obtained during sunlit and
cloud free seasons/situations have been employed to obtain
and compare sea ice extent estimates derived from passive
microwave measurements. Short overlap between different
microwave sensors e.g. between SMMR and SSM/I during
1987 and between successive SSM/I sensors later on has
played a key role in the inter-calibration of sea ice estimates
from SMMR and SSM/. As a result we now have a highly
reliable 25 year long time series of sea ice measurements over
the Arctic and Antarctic regions (Comiso, 2000, Hanna and
Bamber, 2001, Gloersen et al., 1992). It may be mentioned
that this period has seen many important natural and human-
induced changes in the planetary atmosphere and the ocean
e.g. the occurrence of ‘ozone hole’ and several strong El-Nino
events which impact significantly on the global climate.
In this paper, ve present an inter-comparison of simultaneous
and coincident MSMR and 5SSM/ based brightness
temperature measurements as well as the estimated sea ice
characteristics from the two sensors.
2.0 MSMR AND SSM/I
MSMR onboard OCEANSAT-1 is a four frequency, eight
channel dual polarised Dicke switched PMR system operating
at 6.6, 10.65, 18 and 21 GHz. MSMR carries a black body
maintained at ambient system temperature and it has cold
space viewing horns for calibration of all the radiometers. The
extensive initial pre-launch calibration is carried out on
ground. The calibration performance and stability of the
radiometers have been analysed by Misra et al. (2002). All the
radiometers are shown to have a temperature sensitivity of
better than 1 K. Ali et al. (2000) have attempted the validation
of various geophysical parameters derived from MSMR over
the tropical region.
SSM/I is a similar dual polarisation total power radiometer
system operating at 19.35, 22.235, 37 and 85.5 GHz, with
22.235 GHz channel providing only vertical polarization
information. It also incorporates cold (sky horn) and hot
(reference absorber) calibration targets. After extensive
calibration and validation studies, SSM/I data products have
been made available on an operational basis.
For relative calibration of the MSMR observed brightness
temperatures as well as derived geophysical (sea ice)
parameters over the polar regions, we make use of the
simultaneous SSM/I measurements over the Antarctic region.
Characteristics of MSMR and SSM/I sensors along with the
orbital platforms from which they operate are summarized in
Table-1. Both MSMR and SSM/I are conically scanning
410
multi-frequency passive microwave radiometers. However,
there are several differences e.g. available frequency
channels and their center frequencies, the incidence angle,
the coverage swath, maximum north and south latitude
attainable and the local time of observations etc. While most
of these are minor differences as far observing Antarctic sea
ice is concerned, these must however be kept in mind during
the inter-comparison exercise. Since, 18/19 GHz channel is
the most suitable common ‘channel between MSMR and
SSM/L in providing surface characteristics with reasonably
good atmospheric transparency and with acceptably high
spatial resolution, we have restricted our analysis to data
from this channel. This channel is also common between the
three most used sensors for sea ice research viz. SMMR,
SSM/I and MSMR. Moreover, for developing the
algorithms for estimation of sea ice concentration, we have
used brightness temperatures of 10 and 18 GHz channels of
MSMR to derive the polarization and spectral gradient
ratios (Gloersen et al, 1992) to be used as in-dependent
variables.
Table —1: Characteristics of MSMR and SSM/I
Satellite OCEANSAT-1 DMSP -5D
Launch May 26, 1999 1987...
Orbit Polar Polar
Sun-Synchronous Sun-Synchronous
Orbital Height | 720 830-860
(Km)
Orbital 98.28 98.8
Inclination
(Deg)
Equatorial 12 noon 06:15
Crossing Time
Sensor MSMR_ SSM/I |
Frequency 6.6, 10.65,18, 21 19.35 22.235 1.31,
(GHz) 85.5
Polarisation H & V all H&V
channels (22.235 V-only)
Scanning Off-nadir Conical | Off-nadir Conical
Incidence 49.7 53.1
Angle (Deg.)
Swath (Km) 1360 1390
Repetivity Two days Two days
3.0 ANALYSIS OF BRIGHTNESS TEMPERATURES
FROM MSMR AND SSM/I
Fig. 1 provides a guide -map of the Antarctic region showing
various features of the continental ice as well as various
sectors of the Southern Ocean. It also shows small regions
over which MSMR — SSM/I comparison analysis is made.
In order to inter-calibrate MSMR observed brightness
temperatures with those from SSM/I, sample SSM/I data
sets covering the summer and winter seasons over the
Antarctic region were obtained from NASA/GSFC. The
SSM/I 19.35 GHz Tb values available at 25 km resolution
were first averaged using a 2x2 pixel window and then
resampled at 0.5 deg. x 0.5 deg. resolution to make these
observations compatible with MSMR. The SSM/I observed
