216
C.J. Donlon
channels are used), although instability of calibration, cloud-detection failures,
and episodes of atmospheric aerosol can each degrade this potential significantly.
With only the 11 and 12 μm channels available (as in day-time), coefficient-based
retrievals have been limited to accuracies of about 0.4 K. Recent work on METOPA AVHRR data shows that optimal estimation techniques can drive accuracy down
to ~0.3 K for SST estimates where the retrieval cost is low (Merchant et al., 2008a).
All these quoted errors have a random element, but are in large part correlated on
the synoptic scales of the atmosphere. Further research is required to develop and
refine SST optimal estimation techniques for TIR sensors that maximize the error
reduction from having multiple complementary observing systems in space.
13.3.6 Improving SST Provision in the High Latitude Regions
Accurate retrieval of SST at high latitudes using TIR satellite sensors requires that
(a) the discrimination between ice-free and ice-covered water at the resolution,
temporal and spatial, of the SST retrieval schemes is well known; and (b) the atmospheric attenuation on the infrared radiation as it propagates from the sea surface to
the satellite radiometer is determined. For infrared SST retrievals, during the day,
reflected sunlight provides a powerful mechanism for identifying open, cloud-free
water.
During the polar night the problem of identifying ice becomes more difficult. A
simple temperature threshold test might be adequate to identify pack ice but this
would not be sufficient in the more complex marginal ice zone. Surface temperature retrievals below –1.8 ◦ C, the freezing point of seawater, can be classified as ice
cover. However, this is prone to error as (a) there is noise in the satellite-derived
surface temperature, so that ice-free retrievals could fall below the threshold, and
ice-covered pixels fall above the threshold; and (b) when melting, sea ice, especially if covered by snow, may remain frozen at temperatures above the threshold.
More effort should be given to define and implement ice masking procedures and
techniques in Polar Regions for TIR satellite observations.
Considering the impact of atmospheric attenuation on the water leaving signal
it is clear that the polar atmosphere is generally very dry and cold, and is thus
an extreme in terms of the climatological distribution of atmospheric properties.
It represents an anomalous set of conditions for routine SST atmospheric correction
algorithms optimized for the global range of atmospheric variability (e.g. Walton
et al., 1998; May et al., 1998). It is to be expected that systemic retrieval errors
in the derived SSTs will result: bias errors, usually result in warm SST errors that
can be greater than 1 K (Vincent et al., 2008b). Loss of the correlation between
the brightness temperatures measured at 10.5 and 11.5 μm with the atmospheric
water vapour that occurs in very dry atmospheres and Vincent et al. (2008a, b)
show using AVHRR brightness temperature data collocated with ship-based radiometric skin SST measurements that a simple, single channel retrieval algorithm
can produce improved accuracy in the measurement of skin SST and Ice Surface
Temperature. Single-channel algorithms appear to be better suited to the problem
C.J. Donlon
channels are used), although instability of calibration, cloud-detection failures,
and episodes of atmospheric aerosol can each degrade this potential significantly.
With only the 11 and 12 μm channels available (as in day-time), coefficient-based
retrievals have been limited to accuracies of about 0.4 K. Recent work on METOPA AVHRR data shows that optimal estimation techniques can drive accuracy down
to ~0.3 K for SST estimates where the retrieval cost is low (Merchant et al., 2008a).
All these quoted errors have a random element, but are in large part correlated on
the synoptic scales of the atmosphere. Further research is required to develop and
refine SST optimal estimation techniques for TIR sensors that maximize the error
reduction from having multiple complementary observing systems in space.
13.3.6 Improving SST Provision in the High Latitude Regions
Accurate retrieval of SST at high latitudes using TIR satellite sensors requires that
(a) the discrimination between ice-free and ice-covered water at the resolution,
temporal and spatial, of the SST retrieval schemes is well known; and (b) the atmospheric attenuation on the infrared radiation as it propagates from the sea surface to
the satellite radiometer is determined. For infrared SST retrievals, during the day,
reflected sunlight provides a powerful mechanism for identifying open, cloud-free
water.
During the polar night the problem of identifying ice becomes more difficult. A
simple temperature threshold test might be adequate to identify pack ice but this
would not be sufficient in the more complex marginal ice zone. Surface temperature retrievals below –1.8 ◦ C, the freezing point of seawater, can be classified as ice
cover. However, this is prone to error as (a) there is noise in the satellite-derived
surface temperature, so that ice-free retrievals could fall below the threshold, and
ice-covered pixels fall above the threshold; and (b) when melting, sea ice, especially if covered by snow, may remain frozen at temperatures above the threshold.
More effort should be given to define and implement ice masking procedures and
techniques in Polar Regions for TIR satellite observations.
Considering the impact of atmospheric attenuation on the water leaving signal
it is clear that the polar atmosphere is generally very dry and cold, and is thus
an extreme in terms of the climatological distribution of atmospheric properties.
It represents an anomalous set of conditions for routine SST atmospheric correction
algorithms optimized for the global range of atmospheric variability (e.g. Walton
et al., 1998; May et al., 1998). It is to be expected that systemic retrieval errors
in the derived SSTs will result: bias errors, usually result in warm SST errors that
can be greater than 1 K (Vincent et al., 2008b). Loss of the correlation between
the brightness temperatures measured at 10.5 and 11.5 μm with the atmospheric
water vapour that occurs in very dry atmospheres and Vincent et al. (2008a, b)
show using AVHRR brightness temperature data collocated with ship-based radiometric skin SST measurements that a simple, single channel retrieval algorithm
can produce improved accuracy in the measurement of skin SST and Ice Surface
Temperature. Single-channel algorithms appear to be better suited to the problem