Extensive observations of sea ice cover are essential for understanding the development of these ice masses and of their interactions with other components of the climate system. Among the least known parameters of sea ice cover are snow and ice thicknesses. Both parameters are crucial for the calculation of heat and salinity fluxes between the ocean and the atmosphere over ice covered areas. Snow and ice are good thermal conductors and have relatively high albedo, but their impact can vary considerably depending on thickness. Sensitivity studies using several global circulation models have also indicated that accurate values for the snow and ice thicknesses are required for consistent predictions.

Comparison to other instruments

• In contrast with freeboard retrievals based on lidar or radar topography, THOR is expected to retrieve separate thickness values for snow and sea ice. This is important because uncertainties in snow thickness can cause several times larger uncertainties in freeboard ice thickness estimates.
• In contrast with freeboard estimates (based on the altitude difference of the observed top surface and the expected sea level), THOR retrievals can work even when the sea level is not known precisely. This is important because an uncertainty in sea level causes a 9 times larger uncertainty in sea ice thickness, as 90% of the ice is underwater.
• Unlike radar retrievals, THOR retrievals are not contaminated significalntly by sub-resolution size leads and melt ponds.
• In contrast with coarse-resolution radar measurements, THOR can yield information on surface roughness.
• Unlike radars, THOR cannot yield sea ice thicknesses if clouds obscure its view.
• Unlike topographical radars and lidars, THOR can estimate only snow but not sea ice thickness during daytime, or if the ice is covered by a thick layer of fresh snow that dampens the ice signal substantially. In such cases we plan to revert to freeboard estimates that will be aided by THOR's data on snow thickness.

Feasibiltiy of concept

The results of ground-based experiments suggest to us that observations of diffuse halos can indeed be used to estimate sea ice thickness (see Table 1 in Haines et al. 1997).

Experimental setup of in-situ ice thickness measurements in Haines et al. (1997). Ice thickness was retrieved using only reflectance measurements; transmission measurements were used only for estimating the anisotropy of brine pockets and absorption by algae.

Ground-based observations suggest that halo observations can also reveal snow thickness. (See Várnai, T., and R. F. Cahalan, 2009: Modeling and analysis of offbeam lidar returns from thick clouds, snow, and sea ice. International Conference on Advances in Mathematics, Computational Methods, and Reactor Physics, American Nuclear Society, ISBN: 978-0-89448-069-0)

Retrieval methodology

Snow and ice thickness is retrieved by comparing THOR observations to simulated data generated by a three-dimensional Monte Carlo radiative transfer model for a wide range of snow and ice conditions.

Time-integrated return signals for various snow and ice conditions, as simulated by a Monte Carlo radiative transfer model