D.G. BARBER, A. THOMAS, AND T.N. PAPAKYRIAKOU
K* gains and large L * losses experienced near solar noon on day u8 indicate clear-sky
conditions, whereas the observations on day 126 are attributed to a low-cloud, precipitating atmosphere, with strong surface winds. More detailed analysis of this cloud-cover effect on scattering has shown that changes in the surface dielectrics induced by cloud
effects on the longwave flux are detectable within the winter scattering season (Barber
and Thomas 1997). Concomitant enlargement of snow grains under temperature gradient morphosis during this period (Barber et al. 1994) served to enhance the radiative
gains from K*, further enhancing Q*. Note that Q* was measured directly, thus deviations in the summation of K* and L * stem from sensor precision and alignment differences. Dielectrically, the sea ice surface and snow volume followed a pattern which
appeared to covary with the radiative and conductive fluxes of the snow and sea ice volumes. These forcings can be considered as a causal relationship since the available energy will directly affect the fractional volumes of ice, air and brine and the phase proportions of water within the snow and sea ice, thereby causing a change in the dielectrics.
We used a combined surface/volume scattering model (Barber and LeDrew 1994) to
diagnose the scattering characteristics of this condition. In time series 1 the gradual
rise in both E' and E" corresponded to an observed decrease in 0'0. The downturn in
both E' and E" after day 123 corresponded to an increase in 0'0 back to a magnitude in
scattering similar to the beginning of the time series. The inverse relationship between
the complex dielectric constant and 0'0 suggests that volume scattering 0'0 v is occurring
(Ulaby et al.1986). Our diagnostic model predicted a decrease in the volume scattering
0'0 v term and an increase in the surface scattering 0'0, term for the range of parameters
in time series 1. The decrease in 0'0 v was smaller than the increase in 0'0, because of the
low magnitude of the dielectric constant in the snow volume, so that the total scattering term 0'0 tended to increase over the period where the ERS-1 0'0 was observed to be
decreasing. In time series 2 the same inverse relationship existed between E', E", and 0'0,
again suggesting that volume scattering was responsible for at least part of the total 0'0.
It is also important to note that the ice surface dielectrics illustrated both a large magnitude and similarity to 0'0 in time series 1 and that the snow volume did not. In time
series 2 there was an overall increase in the magnitude of E' and E", and the time series
patterns also varied in phase with 0'0.
Determining the exact location and magnitude of the surface and volume scattering
contributions to 0'0 cannot be determined using our diagnostic model since it is not
designed to separate ice from snow volume scattering. We can however use the dielectric properties to evaluate the microwave penetration depths, thereby illuminating the
important depths over which scattering may occur. It would appear that for dielectric
properties and snow densities typical of time series 1, the minimum penetration depth
is approximately equivalent to the observed snow depth (approx. 30 cm). Under conditions typical of time series 2 the minimum penetration depth was reduced well under
this 30-cm level to values of between 5 and 10 cm. This observation lends support to
the supposition that the surface scattering term 0'0, dominated the total scattering
cross-section 0'0 in time series 1 (at least before about day 120), and that the volume
term a o v contributed a significant amount to the total0'° in time series 2. Given the
decrease in 0'0 with an increase in E' and E" prior to day 120, it is likely that the brinesoaked, large, elongated snow grains immediately next to the ice surface were acting as
volume scattering centers, thereby contributing a 0'0 v term to the ic~ surface 0'0, in specifying the total first-year ice 0'0.
K* gains and large L * losses experienced near solar noon on day u8 indicate clear-sky
conditions, whereas the observations on day 126 are attributed to a low-cloud, precipitating atmosphere, with strong surface winds. More detailed analysis of this cloud-cover effect on scattering has shown that changes in the surface dielectrics induced by cloud
effects on the longwave flux are detectable within the winter scattering season (Barber
and Thomas 1997). Concomitant enlargement of snow grains under temperature gradient morphosis during this period (Barber et al. 1994) served to enhance the radiative
gains from K*, further enhancing Q*. Note that Q* was measured directly, thus deviations in the summation of K* and L * stem from sensor precision and alignment differences. Dielectrically, the sea ice surface and snow volume followed a pattern which
appeared to covary with the radiative and conductive fluxes of the snow and sea ice volumes. These forcings can be considered as a causal relationship since the available energy will directly affect the fractional volumes of ice, air and brine and the phase proportions of water within the snow and sea ice, thereby causing a change in the dielectrics.
We used a combined surface/volume scattering model (Barber and LeDrew 1994) to
diagnose the scattering characteristics of this condition. In time series 1 the gradual
rise in both E' and E" corresponded to an observed decrease in 0'0. The downturn in
both E' and E" after day 123 corresponded to an increase in 0'0 back to a magnitude in
scattering similar to the beginning of the time series. The inverse relationship between
the complex dielectric constant and 0'0 suggests that volume scattering 0'0 v is occurring
(Ulaby et al.1986). Our diagnostic model predicted a decrease in the volume scattering
0'0 v term and an increase in the surface scattering 0'0, term for the range of parameters
in time series 1. The decrease in 0'0 v was smaller than the increase in 0'0, because of the
low magnitude of the dielectric constant in the snow volume, so that the total scattering term 0'0 tended to increase over the period where the ERS-1 0'0 was observed to be
decreasing. In time series 2 the same inverse relationship existed between E', E", and 0'0,
again suggesting that volume scattering was responsible for at least part of the total 0'0.
It is also important to note that the ice surface dielectrics illustrated both a large magnitude and similarity to 0'0 in time series 1 and that the snow volume did not. In time
series 2 there was an overall increase in the magnitude of E' and E", and the time series
patterns also varied in phase with 0'0.
Determining the exact location and magnitude of the surface and volume scattering
contributions to 0'0 cannot be determined using our diagnostic model since it is not
designed to separate ice from snow volume scattering. We can however use the dielectric properties to evaluate the microwave penetration depths, thereby illuminating the
important depths over which scattering may occur. It would appear that for dielectric
properties and snow densities typical of time series 1, the minimum penetration depth
is approximately equivalent to the observed snow depth (approx. 30 cm). Under conditions typical of time series 2 the minimum penetration depth was reduced well under
this 30-cm level to values of between 5 and 10 cm. This observation lends support to
the supposition that the surface scattering term 0'0, dominated the total scattering
cross-section 0'0 in time series 1 (at least before about day 120), and that the volume
term a o v contributed a significant amount to the total0'° in time series 2. Given the
decrease in 0'0 with an increase in E' and E" prior to day 120, it is likely that the brinesoaked, large, elongated snow grains immediately next to the ice surface were acting as
volume scattering centers, thereby contributing a 0'0 v term to the ic~ surface 0'0, in specifying the total first-year ice 0'0.