IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
3081
Microwave Signatures of Snow
on Sea Ice: Observations
Thorsten Markus, Member, IEEE, Donald J. Cavalieri, Member, IEEE, Albin J. Gasiewski, Fellow, IEEE,
Marian Klein, Member, IEEE, James A. Maslanik, Dylan C. Powell, B. Boba Stankov,
Julienne C. Stroeve, and Matthew Sturm
Abstract—Part of the Earth Observing Sysytem Aqua Advanced Microwave Scanning Radiometer (AMSR-E) Arctic sea
ice validation campaign in March 2003 was dedicated to the
validation of snow depth on sea ice and ice temperature products.
The difficulty with validating these two variables is that neither
can currently be measured other than in situ. For this reason,
two aircraft flights on March 13 and 19, 2003, were dedicated
to these products, and flight lines were coordinated with in situ
measurements of snow and sea ice physical properties. One flight
was in the vicinity of Barrow, AK, covering Elson Lagoon and the
adjacent Chukchi and Beaufort Seas. The other flight was farther
north in the Beaufort Sea (about 73◦ N, 147.5◦ W) and was coordinated with a Navy ice camp. The results confirm the AMSR-E
snow depth algorithm and its coefficients for first-year ice when
it is relatively smooth. For rough first-year ice and for multiyear
ice, there is still a relationship between the spectral gradient ratio
of 19 and 37 GHz, but a different set of algorithm coefficients is
necessary. Comparisons using other AMSR-E channels did not
provide a clear signature of sea ice characteristics and, hence,
could not provide guidance for the choice of algorithm coefficients.
The limited comparison of in situ snow–ice interface and surface
temperatures with 6-GHz brightness temperatures, which are
used for the retrieval of ice temperature, shows that the 6-GHz
temperature is correlated with the snow–ice interface temperature
to only a limited extent. For strong temperature gradients within
the snow layer, it is clear that the 6-GHz temperature is a weighted
average of the entire snow layer.
Index Terms—Advanced Microwave Scanning Radiometer
(AMSR), passive microwave, sea ice, snow on sea ice, validation.
Manuscript received November 14, 2005; revised March 24, 2006.
T. Markus is with the Hydrospheric and Biospheric Sciences Laboratory,
Goddard Space Flight Center, National Aeronautics and Space Administration,
Greenbelt, MD 20771 USA (e-mail: [email protected]).
D. J. Cavalieri is with the Goddard Space Flight Center, National Aeronautics
and Space Administration, Greenbelt, MD 20771 USA.
A. J. Gasiewski is with the Department of Electrical and Computer Engineering, University of Colorado, Boulder, CO 80309 USA.
M. Klein is with the Environmental Technology Laboratory, National
Oceanic and Atmospheric Administration, Boulder, CO 80305 USA.
J. A. Maslanik is with the Center for Astrodynamics Research, University of
Colorado, Boulder, CO 80309 USA.
D. C. Powell was with the Department of Physics and the Joint
Center for Earth Systems Technology, University of Maryland Baltimore
County, Baltimore, MD 20715 USA. He is now with Earth observing Systems,
Lockheed Martin, Greenbelt, MD 20770 USA.
B. B. Stankov is with the Earth Systems Research Laboratory, National
Oceanic and Atmospheric Administration, Boulder, CO 80305 USA.
J. C. Stroeve is with the National Show and Ice Data Center, Cooperative
Institute for Research in Environmental Sciences, University of Colorado,
Boulder, CO 80309 USA.
M. Sturm is with the U.S. Army Cold Regions Research and Engineering
Laboratory-Alaska, Fort Wainwright, AK 99703 USA.
Digital Object Identifier 10.1109/TGRS.2006.883134
I. I NTRODUCTION
S
NOW depth on sea ice is a standard product of the Earth
Observing System Aqua Advanced Microwave Scanning
Radiometer (AMSR-E) instrument [3]. The algorithm was developed through the comparison of in situ snow depth measurements with Special Sensor Microwave/Imager brightness
temperatures of Southern Ocean sea ice [8]. Currently, the
same algorithm coefficients are applied to the Arctic, although
there are significant physical and corresponding radiometric
differences between the ice and snow in the Arctic and the
Antarctic. The most important difference is the presence of
multiyear sea ice in the Arctic. Multiyear ice is flagged in the
Arctic, and no snow depth is retrieved over this sea ice type
because of the ambiguity between the radiometric signals of
multiyear ice and deep snow. Other complications such as salinity variability and the presence of a slush layer at the snow–ice
interface also influence the retrieval of snow depth, but the
lack of validation data, particularly over Arctic first-year sea
ice, has prohibited a refinement of algorithm coefficients. The
data collected during this aircraft campaign are a first effort to
evaluate the algorithm coefficients for the Arctic.
Similar to the retrieval of snow water equivalent on land (e.g.,
[2] and [7]) the algorithm for snow depth on sea ice makes use
of the difference in scattering by snow between the 19- and
37-GHz frequencies. Explicitly, snow depth hs is calculated
using the spectral gradient ratio of 37 and 19 GHz at vertical
polarization, i.e.,
hs = a+b × GR = a+b ×
T B37V (ice) − T B19V (ice)
T B37V (ice)+T B19V (ice)
(1)
where, for AMSR-E, a = 2.9 and b = −782.4 are regression
coefficients (in centimeters) and T B37V (ice) and T B19V (ice)
are brightness temperatures that have been corrected for the
open water fraction within each pixel using passive microwave
ice concentration estimates [9]. First-year sea ice has a high
emissivity (about 0.95) for both the 19V and 37V channels.
Thus, the difference between T B37V and T B19V , or GR,
is close to zero for no snow. With increasing snow depth,
the radiation emitted by the sea ice is increasingly scattered.
This scattering is greater at 37 GHz than at 19 GHz so that
increasing snow depth results in relatively greater brightness
temperatures at 19 GHz compared to 37 GHz and, thus, more
negative GRs.
In March 2003, seven aircraft flights over the Alaskan Arctic
were carried out as part of the AMSR-E sea ice validation
0196-2892/$20.00 © 2006 IEEE
3082
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
Fig. 1. Overview of PSR flights during the 2003 AMSR-E sea ice validation campaign over a Moderate Resolution Imaging Spectroradiometer (MODIS) image
from March 20, 2003. The “true” color image was created using MODIS bands 2, 4, and 3.
TABLE I
CHARACTERISTICS OF THE PSR SYSTEM
Fig. 2. Locations of the PSR onboard the NASA P-3 aircraft.
effort [1] (Fig. 1). The main instrument on the National Aeronautics and Space Administration’s (NASA) P-3 airplane was
the Polarimetic Scanning Radiometer (PSR) from the National
Oceanographic and Atmospheric Administration’s Environmental Technology Laboratory. Two PSR scanheads, namely
1) PSR/A and 2) PSR/CX, were integrated onto the P-3 in the
plane’s bomb bay (Fig. 2). These scanheads provide the same
(and more) frequency coverage as the AMSR-E instrument
(Table I).
Two of these flights, i.e., on March 13 and 19, 2003,
were coordinated with extensive surface measurements, which
MARKUS et al.: MICROWAVE SIGNATURES OF SNOW ON SEA ICE: OBSERVATIONS
3083
Fig. 3. (a) High-altitude and low-altitude PSR 37-GHz (V-pol) data for the March 13, 2003, flights over the Barrow area. Locations of in situ measurements are
indicated in gray. (b) Subset with only the low-altitude flights.
provided a ground-truth data set for the validation and evaluation of the AMSR-E snow depth on sea ice and sea ice temperature retrievals. The in situ observations consisted of snow depth
measurements as well as detailed measurements of snow and
ice physical properties such as grain size, density, and stratigraphy at selected locations (see [13] for further details).
In this paper, we compare the PSR microwave measurements
with the surface measurements to evaluate the performance of
the AMSR-E snow depth algorithm and to identify shortcomings or limitations. We also use the in situ snow depth and
temperature measurements near Barrow to analyze the performance of the AMSR-E ice temperature product.
II. M EASUREMENTS
In situ measurements were carried out in two locations,
namely: 1) along two transects in the vicinity of Barrow,
specifically over Elson Lagoon and the adjacent Beaufort and
Chukchi Seas coincident with the aircraft flights on March 13
and 2) along a transect close to a Navy ice camp in the
outer Beaufort Sea coincident with the aircraft flights on
March 19, 2003.
A. Barrow Area
Fig. 3(a) shows the coastline of the Barrow area together
with PSR 37-GHz brightness temperatures and the location of
the in situ measurements (in gray). The region was overflown
at two different altitudes, i.e., 1) at a low altitude of 152 m
(500 ft) directly over the transects [dense symbols in Fig. 3(a)]
and 2) at a higher altitude of 1307 m (4300 ft) to map the
entire region [colored circles in Fig. 3(a)] so that findings could
be extrapolated to a larger area [10]. During the low-altitude
flights, the PSR was operated in staring mode, i.e., the PSR
was not scanning. The spatial resolution of the PSR at this
altitude is about 30 m. Because of geolocation inaccuracies
and corrections for aircraft drift, these transects were overflown
several times to ensure good spatial agreement between the
aircraft observations and the ground measurements [Fig. 3(b)].
For the high-altitude flights, the PSR was operating in scanning
mode with a spatial resolution of about 500 m. One can see
the good spatial coherency and agreement between the two
data sets sampled at different altitudes. Elson Lagoon shows a
relatively uniform brightness temperature distribution, whereas
brightness temperatures are lower outside of Elson Lagoon
where the sea ice is rougher and the snow is deeper.
Snow depth along the lines was measured about every 2.5 m.
At certain locations along the snow depth line (about every kilometer), detailed measurements of the snow physical properties
were also taken. These consisted of snow grain size and snow
density for the individual snow layers as well as temperature
profiles (see also [13] for details). All of these parameters can
potentially affect the brightness temperature.
B. Navy Ice Camp
During the time of the AMSR-E validation campaign, the
U.S. Navy operated an ice camp about 175 km northeast of
Barrow (at roughly 73◦ N, 147.5◦ E) in the main pack of the
Beaufort Sea. Snow depth and ice thickness were measured
at about 5-m intervals along a 4.5-km surface transect. The
transect was a mix of smooth and rough first-year and multiyear
ice. This area was mapped using the PSR by high-altitude
transects (1307 m or 4300 ft) only. It required about 8 h to
complete the in situ measurements, whereas the PSR mapping
of the area required less than an hour. During the duration
of in situ measurements, the ice drifted significantly so that
3084
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
Fig. 4. Comparison of coincident brightness temperatures from the PSR and AMSR-E instruments from flights on March 18 and 22, 2003. These two flights
covered the marginal sea ice zone and provide the full range of expected brightness temperatures from open water to consolidated sea ice. The dashed line is the
diagonal line, and the solid line is the regression line.
an adjustment for ice drift was needed to colocate PSR measurements with the in situ measurements. The sea ice at the
Navy camp was drifting in the SSW direction at a speed of
roughly 190 m an hour (M. Sturm, personal communication).
Because the transects were flown in approximately the northnortheast–south-southwest direction, the ice drift was easily
corrected during the 8-h period when the measurements were
taken. All surface measurements began at about 10 A . M . local
time, which is approximately the same time when the aircraft
was in the area.
III. A NALYSIS
We regressed the PSR and the AMSR-E brightness
temperature data for corresponding channels to use the
AMSR-E algorithm coefficients with the PSR data. PSR brightness temperatures are compared with coincident AMSR-E
brightness temperatures for the flights on March 18 and 22
(see Fig. 1). These two flights were over the marginal sea
ice zone and, therefore, covered the whole range of expected
brightness temperatures from open water to consolidated sea ice
(Fig. 4). The results show that, overall, there is a good
MARKUS et al.: MICROWAVE SIGNATURES OF SNOW ON SEA ICE: OBSERVATIONS
3085
TABLE II
REGRESSION COEFFICIENTS THAT MATCH PSR BRIGHTNESS
TEMPERATURES WITH AMSR-E BRIGHTNESS TEMPERATURES
AS W ELL AS C ORRELATION C OEFFICIENTS
linear correlation between the two sensors (the correlation
coefficients and the regression coefficients are summarized in
Table II), although the 6- and 89-GHz data show some significant differences and significant scatter. The problem with the
AMSR-E 6-GHz data is that the spatial resolution is so coarse
(43 km × 75 km [6]) that the PSR data did not cover the entire
width of the 6-GHz AMSR-E footprint. For the regression, we,
therefore, use only the end points of the distribution where
both the PSR and the AMSR data received radiances from
pure open water and pure consolidated sea ice. The scatter and
offset at 89 GHz cannot be explained by the footprint size.
One reason for these differences could be the differences in
atmospheric burden because of the low altitude of the P-3. For
all flights, we had clear sky conditions. Atmospheric radiative
transfer calculations for a clear winter atmosphere show expected differences between PSR and AMSR-E data from about
0.5 K for 6 GHz to 3 K for 89 GHz. These differences are
too small to explain the observed differences, except for the
18- and 37-GHz vertical polarization channels. There also
seems to be a consistent shift with frequency in the calibration
differences. At 6 GHz, PSR T Bs are significantly lower than
the AMSR-E T Bs, but this difference becomes smaller with
increasing frequency. At 89 GHz, PSR T Bs are greater than
the AMSR-E T Bs. Because the correlation between the PSR
and the AMSR-E data is very good (excellent for the 10-, 18-,
and 37-GHz channels), differences in relative calibration pose
no problem in applying the PSR data to the AMSR-E algorithms. To evaluate and validate the snow depth and temperature
algorithms, PSR brightness temperatures are regressed onto
AMSR-E brightness temperatures.
A. Barrow Area
As mentioned earlier in this paper, the ground transects were
overflown several times to ensure optimum spatial coincidence.
Fig. 5. In situ snow depth measurements along the Elson line and corresponding aircraft altitude as well as PSR-derived GR(37V19V).
TABLE III
CORRELATION AND DIFFERENCES BETWEEN IN SITU AND PSR-DERIVED
SNOW DEPTH FOR DIFFERENT AVERAGES
As seen in Fig. 3(b), there is some swerving along the transect.
Variable winds made it difficult for the plane to stay exactly
over the transect, and the plane’s autopilot was correcting for
any drift away from the desired transects. Autopilot corrections were moderate to keep aircraft attitude variations and,
thus, PSR incidence angle variations at a minimum (Fig. 5).
Although the PSR was set to a fixed beam position pointing
forward at 55◦ from nadir, a yaw motion results in a slightly
offset direction so that aircraft subtrack and PSR footprints are
3086
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
Fig. 6. Comparison of in situ and PSR-derived snow depth along the Elson
and Beaufort lines. The thin line is the 50-m average, and the thick line is the
200-m average.
not necessarily coincident. However, most measurements were
taken within 100 m of the transect [Fig. 5(e)], and the PSR data
show good spatial agreement [Fig. 5(f)].
1) Snow Depth: To investigate the effect of spatial variability and geolocation uncertainties in more detail, we averaged
the in situ snow depth and the PSR-derived snow depth over
50, 100, 150, 200, and 250 m (Table III). Overall, there is
a relatively good correlation between the data sets, and with
increasing spatial averaging, the correlation coefficient also
increases, and the average absolute difference between in situ
and PSR snow depth decreases. With more than 200-m spatial
averaging, the improvements are not as great, and we will,
therefore, use 200-m averages in Section IV.
Fig. 6(a) shows the surface snow depth measurements averaged every 50 m along the transect (thin line) as well as the
snow depth when 200-m averages are taken. Fig. 6(b) shows the
corresponding PSR snow depth retrievals. Although reduced
compared to the full-resolution snow depth, the plot still shows
significant small-scale variability in the 50-m in situ data, which
is not resolved in the PSR data. The main reason is probably
the much lower PSR resolution, although the in situ data are
averaged. The variability of the snow cover perpendicular to the
in situ transect can be the reason because averages of the in situ
data are still just along the transect, whereas the PSR data are
area averages and are not just linear. Overall, one can see that
peaks in snow depth correspond in the two data sets and that
there is no significant offset in the snow depth. Furthermore,
the deeper snow along the Beaufort line is well captured in the
PSR data.
2) Snow Depth and Ice Temperature: As mentioned earlier,
at certain locations along these transects (approximately every
kilometer), detailed snow property measurements as well as
measurements of the snow–ice interface and air temperatures
were taken [13]. To obtain adequate statistics, in situ measurements at these along-transect locations were taken from
11 individual snow pits at distances from 0 to 100 m from
the transect. The temperature measurements from these pits
are used to validate the AMSR-E ice temperature product as
well as to investigate the consistency between snow depth and
ice temperature retrievals. Sturm et al. [13] found a strong
positive correlation between these two variables, as would be
expected. During winter when the air temperature is well below
freezing, the thermal insulation of the snow causes increasing
ice temperatures with increasing snow depth.
The AMSR-E ice temperature is derived from the 6-GHz
vertical polarization brightness temperatures and a constant
value of 0.98 for the ice emissivity using Ti = T B6V /ε. At
6 GHz, there is no emissivity difference between first-year and
multiyear ice (see, e.g., [5]).
The comparison of PSR 6-GHz T Bs averaged over 100-m
areas (corresponding to the snow pit data) with in situ ice
temperatures showed that to get the passive-microwave-derived
ice temperature close to the measured ice temperature, a nonphysical value for the emissivity of 1.189 was necessary; when
using an emissivity of 0.98, the resulting ice temperature has a
significant negative bias. While an emissivity of 0.98 is close to
what has been measured with ground-based radiometers (see,
e.g., [4]), the simple product of emissivity and ice temperature
ignores the atmospheric contribution. However, radiative transfer calculations show a potential bias of less than 1 K for a clear
winter atmosphere at 6 GHz.
A comparison of the in situ and PSR-derived ice temperatures (Fig. 7, black and blue lines, respectively) shows that
the two generally agree (particularly for the Elson line), but
peaks in in situ ice temperature, which are mainly due to deeper
snow (Fig. 7, bottom), are barely reflected in the microwave
ice temperatures. This discrepancy is most clearly seen at the
end of the Beaufort line, where the in situ ice temperature
sharply increases because of a jump in snow depth. In contrast,
the PSR-derived ice temperature decreases slightly, following
the sharp decrease in surface temperature. On the other hand,
variations in surface temperature are not immediately reflected
in corresponding variations of the ice temperature because the
snow temperature profile adjusts slowly to changing surface
temperatures so that thermal equilibrium cannot necessarily be
assumed.
The limited amount of data, though, prohibits definitive conclusions. For example, even within the 100-m averaging bins,
there is a significant range in PSR-derived ice temperatures.
The light blue lines in Fig. 7 (top) indicate the minimum and
maximum PSR ice temperatures within each bin. We note that
the in situ ice temperatures are within this range at most pit
locations.
This bias and the lack of strong correlation demonstrates that
even at 6 GHz, the emission originates not entirely from the
snow–ice interface, and the snow pack temperature gradient
as well as the conditions of the pack affect ice temperature
retrieval. A synergistic analysis of AMSR-E brightness temperatures and snow depths along with MODIS surface temperatures together with a simple one-dimensional dynamic model
of heat storage in the snow layer (e.g., [12]) may help to
better estimate the temperature profile within the snow layer,
including the temperature at the snow–ice interface. This way,
not only AMSR-E-derived ice temperatures may improve, but
we may also be able to estimate model-based sea ice thickness.
MARKUS et al.: MICROWAVE SIGNATURES OF SNOW ON SEA ICE: OBSERVATIONS
Fig. 7. Comparison of in situ snow–ice interface and air temperature with
(top) PSR-derived ice temperature and (bottom) corresponding in situ snow
depth.
3087
Fig. 9. (a) PSR-derived snow depth, (b) P R10, (c) P R19, and
(d) GR(19V10V) versus in situ snow depth using 200-m averages. The
different colors indicate the different locations as labeled in (a). Labels e1–e7
and b1–b4 refer to the snow physical properties listed in Table IV.
PSR-derived snow depth. The correlation coefficient between
PSR and in situ snow depth is approximately 0.5 for both firstyear and multiyear ice types. The large scatter is believed to
be the result of the PSR’s relatively broad spatial resolution
of about 500 m at 1307 m (4300 ft) altitude, which averages
snow signatures both along the ice camp transect as well as
orthogonal to the transect.
IV. D ISCUSSION
A. Validation of the AMSR-E Snow Depth Algorithm
Fig. 8. (a) PSR-derived snow depth and (b) corresponding in situ snow depth
for the transect near the Navy ice camp. Multiyear (MY) ice areas are shaded
in gray. The thick line in the snow depth data is a 200-m running mean.
B. Navy Ice Camp
Fig. 8(a) shows snow depth derived from the PSR brightness
temperatures, and Fig. 8(b) shows the corresponding in situ
snow depths for the Navy Ice Camp line. Multiyear ice areas are
shaded in gray. For each in situ observation, we used the closest
PSR footprint for the comparison. Thus, values of constant PSR
snow depth indicate that the same PSR measurement was used
for consecutive in situ measurements. Over first-year ice, one
can see a reasonable correlation between in situ snow depth and
Fig. 9(a) shows a scatterplot of all PSR-derived snow depths
versus the corresponding in situ snow depths using 200-m bin
averages. For the Navy Ice Camp, to have some confidence
in the statistics and to avoid mixed-pixel issues, only those
measurements were plotted where the first-year or multiyear
fractions were at least 30%, with at least ten in situ measurements of snow depth for a given PSR measurement. One can
see that the data from the Elson and Chukchi lines agree very
well, with most of the measurements falling very close to the diagonal. Furthermore, most of the first-year Ice Camp data agree
fairly well. There are some outliers, but as mentioned earlier,
the PSR resolution for the Ice Camp flights was more coarse
than for the Elson, Beaufort, and Chukchi lines. Although the
slope is in good agreement with the in situ snow depth, the PSR
tends to slightly underestimate snow depth by ∼3.4 cm with
the current coefficients. We note that a similarly small negative
bias of 3.5 cm had previously been observed by Markus and
Cavalieri [8] when comparing regional in situ and passive
microwave-derived snow depth distributions.
We note that data from the Beaufort line and from multiyear
ice from the Ice Camp transect (also few first-year and Elson
line points) tend to form a separate cluster of larger snow
3088
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
TABLE IV
SNOW PHYSICAL PROPERTIES FROM SNOW PIT LOCATIONS
depths. While the multiyear pixels show little correlation, the
Beaufort line PSR measurements seem to overestimate snow
depth initially (which corresponds to more negative GRs) but
agree well with the in situ measurements for deeper snow depth.
While there is good correlation between the two snow depth
data sets for the Beaufort line (Table III), the data suggest
a need for different (perhaps dynamically varying) algorithm
coefficients.
To investigate whether the observed differences between the
Elson and the Beaufort lines can be explained by differences in
the physical properties of the snow, pixels where detailed snow
measurements were taken are indicated as e1–e7 for the Elson
line and b1–b4 for the Beaufort line (Table IV). An analysis
of the data does not show a clear dependence of PSR scatter
on snow physical properties, i.e., deviations from the algorithm
line cannot be explained by variations in grain size or density. A
companion modeling paper [11] shows that the PSR signatures
for both the Elson and Beaufort lines can be reproduced with
mainly different emissivities for sea ice. As mentioned earlier,
sea ice was significantly rougher for the Beaufort line, and this
roughness could be the reason for the differences. The open
question is whether an optimum set of coefficients for specific
snow–ice classes can be identified from the passive microwave
data themselves.
B. Potential Improvements
Fig. 9(b)–(d) shows in situ snow depth versus (b) P R10,
(c) P R19, and also (d) GR(19V10V), respectively. These
plots show that P R10 and P R19 are largely independent of
snow depth. Furthermore, both polarization ratios seem to
distinguish between the Elson line and Beaufort line data, with
P R10 showing a stronger separation. It could be suggested
that P R10 might also be used to distinguish between first-year
and multiyear ice, but there is only a small separation between
the clusters.
V. S UMMARY
Brightness temperatures from the PSR obtained during extensive P-3 aircraft flights regressed onto AMSR-E brightness
temperatures were used to estimate snow depth on Arctic sea
ice. A comparison with in situ snow depth measurements shows
good agreement and confirms algorithm coefficients for at
least relatively smooth first-year ice as observed in the Elson
Lagoon. The results also suggest that different sets of algorithm
coefficients will be needed for rough (multiyear) sea ice. The
current snow depth algorithm using GR(37V19V) seems to
work less well for multiyear than first-year ice. Analysis of
the other channels suggests that P R10 may be able to be
used to distinguish between rough and smooth sea ice and
may, therefore, be useful in determining algorithm coefficients.
An analysis on AMSR-E pixel scales is needed to investigate
whether these findings are also detectable at those scales before adjustments are made and implemented in the AMSR-E
processing. Future work will explore the use of combined
AMSR-E data with scatterometer data. Scatterometer data can
give an estimate of surface roughness, which could be used to
refine AMSR-E snow algorithm coefficients.
ACKNOWLEDGMENT
The authors would like to thank the reviewers for a thorough
reading of the manuscript and for their constructive comments.
R EFERENCES
[1] D. J. Cavalieri, T. Markus, J. A. Maslanik, M. Sturm, and E. Lobl, “March
2003 EOS Aqua AMSR-E sea ice field campaign,” IEEE Trans. Geosci.
Remote Sens., vol. 44, no. 11, pp. 3003–3008, Nov. 2006.
[2] A. T. C. Chang, J. L. Foster, and D. K. Hall, “Nimbus-7 SMMR derived
global snow cover parameters,” J. Glaciol., vol. 9, pp. 39–44, 1987.
[3] J. C. Comiso, D. J. Cavalieri, and T. Markus, “Sea ice concentration, ice
temperature, and snow depth using AMSR-E data,” IEEE Trans. Geosci.
Remote Sens., vol. 41, no. 2, pp. 243–252, Feb. 2003.
[4] D. T. Eppler et al., “Passive microwave signatures of sea ice,” in
Microwave Remote Sensing of Ice, Geophysical Monograph Series,
vol. 68, F. Carsey, Ed. Washington, DC: AGU, 1992, pp. 47–71.
[5] P. Gloersen, W. J. Campbell, D. J. Cavalieri, J. C. Comiso,
C. L. Parkinson, and H. J. Zwally, Arctic and Antarctic Sea Ice
1978–1987: Satellite Passive Microwave Observations and Analysis.
Washington, DC: NASA, 1992. NASA Spec. Publ., SP-511.
[6] T. Kawanishi, T. Sezai, S. Ito, K. Imaota, T. Takeshima, Y. Ishido,
A. Shibata, M. Miura, H. Inahata, and R. W. Spencer, “The Advanced
Microwave Scanning Radiometer for the Earth Observing System
(AMSR-E), NASDA’s contribution to the EOS global energy and
water cycle studies,” IEEE Trans. Geosci. Remote Sens., vol. 41, no. 2,
pp. 184–194, Feb. 2003.
[7] R. E. Kelly, A. T. C. Chang, L. Tsang, and J. L. Foster, “A prototype
AMSR-E global snow area and snow depth algorithm,” IEEE Trans.
Geosci. Remote Sens., vol. 41, no. 2, pp. 230–242, Feb. 2003.
[8] T. Markus and D. J. Cavalieri, “Snow depth distribution over sea ice in
the Southern Ocean from satellite passive microwave data,” in Antarctic Sea Ice Physical Processes, Interactions and Variability, Antarctic
Research Series, vol. 74, M. O. Jeffries, Ed. Washington, DC: AGU,
1998, pp. 19–40.
[9] ——, “An enhancement of the NASA Team sea ice algorithm,” IEEE
Trans. Geosci. Remote Sens., vol. 38, no. 3, pp. 1387–1398, May 2000.
[10] J. A. Maslanik, M. Sturm, M. Belmonte Rivas, A. J. Gasiewski, J. F.
Heinrichs, U. C. Herzfeld, J. Holmgren, M. Klein, T. Markus,
D. Perovich, J. G. Sonntag, J. C. Stroeve, and K. Tape, “Spatial variability
MARKUS et al.: MICROWAVE SIGNATURES OF SNOW ON SEA ICE: OBSERVATIONS
of Barrow-area shore-fast ice and its relationships to passive microwave
emissivity,” IEEE Trans. Geosci. Remote Sens., vol. 44, no. 11, pp. 3021–
3031, Nov. 2006.
[11] D. C. Powell, T. Markus, D. J. Cavalieri, A. J. Gasiewski, M. Klein,
J. A. Maslanik, J. C. Stroeve, and M. Sturm, “Microwave signatures of
snow on sea ice: Modeling,” IEEE Trans. Geosci. Remote Sens., vol. 44,
no. 11, pp. 3091–3102, Nov. 2006.
[12] A. J. Semtner, Jr., “A model for the thermodynamic growth of sea ice
in numerical investigations of climate,” J. Phys. Oceanography, vol. 6,
pp. 379–389, 1976.
[13] M. Sturm, J. A. Maslanik, D. K. Perovich, J. C. Stroeve,
J. Richter-Menge, T. Markus, J. Holmgren, J. F. Heinrichs, and
K. Tape, “Snow depth and ice thickness measurements from the Beaufort
and Chukchi Seas collected during the AMSR-Ice03 campaign,”
IEEE Trans. Geosci. Remote Sens., vol. 44, no. 11, pp. 3009–3020,
Nov. 2006.
Thorsten Markus (M’05) received the M.S. and
Ph.D. degrees in physics from the University of
Bremen, Bremen, Germany, in 1992 and 1995,
respectively.
He is currently a Research Scientist with the
NASA Goddard Space Flight Center (GSFC), Greenbelt, MD. From 1995 to 1996, he was a National
Research Council Resident Research Associate with
GSFC before joining NASA-UMBC Joint Center
for Earth Systems Technology, where he worked
until 2002. His research interests include satellite
microwave remote sensing of primarily ice and the utilization of satellite data
to study oceanic and atmospheric processes.
Dr. Markus is a member of the American Geophysical Union.
Donald J. Cavalieri (M’05) received the B.S. degree
in physics from the City College of New York, New
York, in 1960, the M.A. degree in physics from
Queens College, New York, in 1967, and the Ph.D.
degree in meteorology and oceanography from New
York University, New York, in 1974.
From 1974 to 1976, he was a National Research
Council Postdoctoral Resident Research Associate
with the National Oceanic and Atmospheric Administration’s (NOAA) Environmental Data Service,
Boulder, CO, where he continued his doctoral research on stratospheric–ionospheric coupling. From 1976 to 1977, he was a
Visiting Assistant Professor with the Department of Physics and Atmospheric
Science, Drexel University, Philadelphia, PA, where he worked on stratospheric
temperature retrievals from satellite infrared radiometers. In the fall of 1977,
he was a Staff Scientist with Systems and Applied Sciences Corporation,
Riverdale, MD, working on sea ice retrieval algorithms, in preparation for
the launch of Nimbus-7 SMMR. In 1979, he joined the Laboratory for Atmospheres, NASA Goddard Space Flight Center, Greenbelt, MD, where he
is currently a Senior Research Scientist in the Cryospheric Sciences Branch
of the Hydrospheric and Biospheric Sciences Laboratory. His current research
interests include polar ocean processes and microwave remote sensing of the
cryosphere.
Dr. Cavalieri is a member of the American Geophysical Union and the
American Meteorological Society.
3089
Albin J. Gasiewski (S’81–M’88–SM’95–F’02) received the B.S. and M.S. degrees in electrical engineering and the B.S. degree in mathematics from
Case Western Reserve University, Cleveland, OH, in
1983, and the Ph.D. degree in electrical engineering
and computer science from the Massachusetts Institute of Technology, Cambridge, in 1989.
He is a Professor of electrical and computer engineering with the University of Colorado at Boulder
and the Director of the CU Center for Environmental
Technology. From 1989 to 1997, he was a Faculty
Member with the School of Electrical and Computer Engineering at the Georgia
Institute of Technology, where he became an Associate Professor. From 1997
to 2005, he was with the U.S. National Oceanic and Atmospheric Administration’s (NOAA) Environmental Technology Laboratory in Boulder, CO, where
he was the Chief of the ETL’s Microwave Systems Development Division.
He has developed and taught courses on electromagnetics, remote sensing,
instrumentation, and wave propagation theory. His technical interests include
passive and active remote sensing, radiative transfer, antennas and microwave
circuits, electronic instrumentation, meteorology, and oceanography.
Prof. Gasiewski is the Past President (2004–2005) of the IEEE Geoscience
and Remote Sensing Society. He is a member of the American Meteorological
Society, the American Geophysical Union, the International Union of Radio
Scientists (URSI), Tau Beta Pi, and Sigma Xi. He currently serves as Vice
Chair of USNC/URSI Commission F. He served on the U.S. National Research
Council’s Committee on Radio Frequencies (CORF) from 1989–1995. He is
the General Cochair of IGARSS 2006, to be held in Denver, CO.
Marian Klein (M’95) received the M.S. and Ph.D.
degrees in electrical engineering from Technical
University of Košice (TU Košice), Košice, Slovak
Republic, in 1986 and 1996, respectively.
From 1987 to 1996, he was a member of the Faculty of Electrical Engineering and Informatics, TU
Košice. From September 1996 to June 1997, he was a
Fulbright Scholar with Georgia Institute of Technology, where he worked on the Laboratory for Radio
Science and Remote Sensing. Since August 1998, he
has been a Research Associate with the Cooperative
Institute for Research in Environmental Sciences, University of Colorado,
Boulder. He is also the Laboratory Manager for the CU Center for Environmental Technology. He has extensive knowledge and experience in radiometer
systems design for harsh environments, whether airborne or ground based.
He successfully led many field deployments of radiometric systems, which
were used in many experiments of the National Aeronautics and Space
Administration, the National Oceanic and Atmospheric Administration
(NOAA), the Department of Energy, and the Department of Defense including
several Soil Moisture Experiments (1999, 2002, 2003, and 2004), the Wakasa
Bay experiment (2003), the Cold Land Processes Experiments (2002 and
2003), and AMSR-E Arctic and Antarctic Sea Ice Experiments (2003 and
2004). He was a Project Leader for the Ground-based Scanning Radiometer
deployed in Barrow, Alaska, in 2004. He is a Lead Field Engineer and Design
and Fabrication Leader for several instruments of the NOAA/Environmental
Technology Laboratory and the Center for Environmental Technology, such as
the Polarimetric Scanning Radiometer system including its PSR-A, PSR-CX,
and PSR-S scanheads. He is also currently the Chief Executive Officer of
Boulder Environmental Sciences and Technology, LLC. His research interests
include passive microwave remote sensing, radiative transfer theory, and
development of millimeter- and submillimeter-wave instrument systems for
environmental studies.
James A. Maslanik received the B.S. degree in
forest science and the M.S. degree in environmental pollution control from the Pennsylvania State
University, University Park, PA, in 1980 and 1978,
respectively, and the Ph.D. degree in geography from
the University of Colorado, Boulder, in 1984.
He is a Research Professor with the Department
of Aerospace Engineering Sciences, University of
Colorado. His research interests include the interactions of sea ice with atmosphere and ocean, remote
sensing and field investigations of sea-ice properties,
effects of climate change on Arctic coastal communities, and development and
deployment of unpiloted aerial vehicles for polar research.
3090
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 11, NOVEMBER 2006
Dylan C. Powell received the B.S. degree in physics
from the Shippensburg University of Pennsylvania,
Shippensburg, in 2000, and the M.S. and Ph.D. degrees in atmospheric physics from the University of
Maryland Baltimore County (UMBC), Baltimore, in
2003 and 2005, respectively.
He is currently a Research Scientist with Lockheed Martin’s Earth Observing Systems, Greenbelt,
MD. He supported policy research and development
while serving on the NOAA Transition of Research
to Application Board in 2005 and was a Research
Assistant with the Joint Center for Earth Systems Technology at UMBC during
his graduate studies. His research interests include satellite remote sensing
techniques using visible, infrared, and microwave instruments for retrieving
geophysical parameters such as sea ice, snow cover, clouds, and vegetation.
Dr. Powell is a member of the American Geophysical Union.
B. Boba Stankov received the B.S. degree in meteorology from Belgrade University, Belgrade, Yugoslavia, in 1964, the M.S. degree in atmospheric
sciences from Colorado State University, Fort
Collins, in 1970, and Ph.D. degrees from Macquarie
University, Sydney, Australia, and the University of
Colorado, Boulder, in 1978 and 1998, respectively.
From 1964 to 1968, she was a Forecaster with the
Yugoslav Weather Bureau, and from 1968 to 1978,
she was a Graduate Student at Colorado State University and the University of Colorado. From 1978 to
1985, she was a Scientist with the Boundary Layer Research Division, National
Center for Atmospheric Research, Boulder, CO, and from 1985 to present, she
has been a Scientist with the Earth Systems Research Laboratory, National
Oceanic and Atmospheric Administration, Boulder. Until 1987, her research
interests have concentrated on boundary layer (BL) turbulence studies using in
situ airborne high-frequency sampling instrumentation. She studied BL length
scales in convective and stable conditions and ozone concentration in the BL.
Since 1987, her interests have focused on the ground-based, space-based, and
combined ground- and space-based microwave sounding of the atmosphere.
She developed a method for multisensor retrieval of atmospheric properties
using both active and passive remote sensor measurements to improve atmospheric sounding resolution within the BL. Since 2001, she has been a member
of the team conducting airborne polarimetric scanning radiometer measurements of the Earth’s surface properties.
Dr. Stankov is a member of the Yugoslav Meteorological Society, the
American Meteorological Society, and the American Institute of Aeronautics
and Astronautics (AIAA).
Julienne C. Stroeve received the B.S. and M.S.
degrees in aerospace engineering and the Ph.D. degree in geography from the University of Colorado,
Boulder, in 1989, 1991, and 1996, respectively,
where she focused on surface energy balance studies
of the Greenland ice sheet using satellite imagery.
Since 1996, she has been with the National
Snow and Ice Data Center, Cooperative Institute for
Research in Environmental Sciences, University of
Colorado, Boulder, as a Research Scientist, specializing in remote sensing of snow and ice. She has
extensive experience in remote sensing of the polar regions using satellite
imagery that spans the optical to the microwave spectral region. She has
participated in several field campaigns in Greenland and the Arctic for the
purpose of validation of various geophysical parameters retrieved from spacecraft such as sea ice concentration, surface temperature, and surface reflectivity.
Additional research projects include monitoring the rapid decline in ice cover
in the Arctic and increased melt of the Greenland ice sheet. At NSIDC, she is
responsible for the sea ice products derived from satellite passive microwave
data, which includes aiding in the design of Web pages providing general sea
ice information and data sets regarding the state of sea ice that may be useful to
a broad audience.
Dr. Stroeve is a member of the IEEE Geoscience and Remote Sensing
Society, the American Geophysical Union, and the Association of American
Geographers.
Matthew Sturm received the M.S. and Ph.D. degrees in geophysics from the University of Alaska,
Fairbanks, AK, in 1984 and 1989, respectively.
He is currently a Research Scientist with the
U.S. Army Cold Regions Research and Engineering
Laboratory-Alaska, Fort Wainwright, AK. He has led
over 15 Arctic expeditions. His research interests
include Arctic climate change, snow on land, and
snow on sea ice.
Dr. Sturm is a member of the American Geophysical Union, the International Glaciological Society,
and the Arctic Institute of North America.