Geology Reference
In-Depth Information
influences are triggered by three parameters: integrated
water vapor along a vertical column of the atmosphere
(IWV), cloud liquid water contents (CLW), and the sur-
face wind speed over open ocean (
V
). Cloud thickness is
also a factor but it cannot be corrected for. Therefore,
pixels that include clouds thicker than a certain threshold
should be excluded from any surface parameter retrieval.
The water vapor is concentrated within the first 5 km of
the troposphere. The cloud liquid water is defined as the
integration of all forms of water including ice crystals in
a cloud volume.
Data of IWV and CLW and
V
, nearly coincident with
the satellite overpass, are needed to conduct the correc-
tion. There are two approaches to obtain these data. The
first is by using the data from operational weather mod-
els. The second is to calculate these parameters using
passive microwave observations from the water vapor
channel (around 22 GHz) in combination with channels
that are not strongly sensitive to these parameters. The
second approach has traditionally been used to obtain
those parameters over open ocean surface only. However
recent studies estimated the integrated water vapor from
the strong absorption lines spanning the highly opaque
183 GHz line. The lines are employed in spectral chan-
nels183 ± 1 GHz, 183 ± 3 GHz, and 183 ± 7GHz of the
Special Sensor Microwave/Temperature 2 (SSM/T2)
onboard the Defence Meteorological Satellite Program
(DMSP) series F‐11 to F‐15, as well as the Advanced
Microwave Sounding Unit‐B (AMSU‐B) onboard
NOAA‐15, 16, and 17. These channels measure radiation
originating from a number of different layers in the trop-
osphere. One of methods that have been proposed to
retrieve IWV from SSM/T‐2 over sea ice in the Arctic is
developed by
Miao
[1998] and in the upper troposphere
by
Sohn et al.,
[2003].
Qiao and Miao
[2003] present maps
of monthly averages IWV in the Arctic using AMSU‐B
channels.
One of the operational weather models that produce
IWV and CLW among a large suite of other atmospheric
parameters is the Global Environmental Multiscale (GEM).
This is the model used at the Canadian Meteorological
Centre (CMC) for short‐range regional forecasting and
medium‐range global forecasting [
Bélair et al.
, 2009]. The
model outputs the results at time steps of 7.5 min but
updates and archives the analysis (i.e., after assimilating
the model's results with remote sensing observations and
other ground measurements) at synoptic times 00, 06, 12,
and 18 GMT. GEM produces the meteorological param-
eters at a standard operational geographic grid of 33 km.
However, coincident and co‐located data with satellite
footprint from any overpass can be generated when the
model runs in a hindsight mode (to coincide with the
satellite overpass) and by interpolating of the model's
gridded results to be co‐located with the locations of the
observed footprints. This approach was used in
Shokr and
Markus
[2006] and
Shokr and Agnew
[2013].
There are two difficulties regarding the use of data
from weather models in a correction scheme that accounts
for the atmospheric influences on the passive microwave
observations. The first is the coarse spatial resolution of
the model's grid with respect to the spatial resolution
of the observations to be corrected (the current grid
spacing from GEM is 10 km or coarser while the foot-
print of the AMSR‐E 89 GHz is 4 × 6 km
2
). The second is
the unreliable estimate (or even the unavailability) of the
cloud liquid water contents parameter.
A review on using passive microwave observations to
determine the water vapor profile and the integrated
water vapor is presented in
Urban
[2013]. The IWV is
usually estimated over open ocean from the weak absorp-
tion line of water vapor near 23.0 GHz (e.g., the 23.8 GHz
channel of AMSR‐E). The retrieval equation usually
combines observations from this channel with other
low‐frequency channels. For example,
Boukabara
[1997]
proposed the following equation:
IWV 23 66 144
.
.
log(
280
T
)
247
.
log(
280
T
)
bV
,
19
bH
,
19
270
.
log(
280
T
)
0 811
.
log(
280
T
)
bV
,
22
bV
,
37
243
.
log(
280
T
)
bH
,
37
(7.101)
where IWV is in kg/m
2
,
T
b
is the brightness temperature
in K, and the subscripts denote the channel frequency
and polarization. Other equations are suggested by
Alishouse et al.
[1990] and
Petty
[1993].
A few methods to retrieve CLW over open water are also
available. Water in clouds emits microwave radiation, and
therefore CLW can be retrieved upon comparing the obser-
vation against a radiometrically cold background. This is
the case of clouds over ocean, where
T
b
above clouds is sig-
nificantly higher compared to the values from the radio-
metrically cold ocean. It contrasts the case of clouds over
land where cloud bottom reflects the surface emission
back, resulting in
T
b
above clouds being lower than
T
b
from
the surface.
O'Dell et al.
[2008] presents a review on some
of the methods of CLW retrieval from passive microwave
observations.
Bobylev et al.
[2010] present a neural network
approach to retrieve IWV and CLW over open water in the
Arctic using microwave radiometer data. They estimated
the error in the retrieval and compared it to other algo-
rithms.
Kern
[2001] found that the method of
Karstens
et al.
[1994] provided reliable results over open water. It is
based on the following equation where CLW is in kg/m
2
:
CLW 3232 16 228
.
.
(
T
)
492 98
.
ln(
280
T
)
bv
,
19
bv
,
22
)
1850 55
.
ln(
28
0
T
)
433 686
.
ln(
280
T
bv
,
37
bh
,
37
(7.102)
