Sea ice forward modelling

Forward models relate the physical properties of the snow and ice with the microwave thermal emission or backscatter. The models are used:
1) in sensitivity studies, where the microwave signature sensitivity to different physical properties can be simulated,
2) model inversion to derive snow or ice parameters, where the model with its physical input is fitted to a measured signature, or
3) forward modelling to estimate a surface signature using simulated physical input.

Both the backscatter and emission models are reseach level model with a rather large number of input parameters. Some of these parameters are rarely measured in the field and/or difficult to measure properly e.g. surface roughness.  Anyway the models are applied for both sensitivity studies and model inversion in research projects.

Sea ice microwave emissivity

The microwave brightness temperature of a lossy half-space such as snow and sea ice is the product of the effective temperature and the emissivity. The effective temperature is the integrated emitting layer thermometric temperature. During winter the surface skin temperature is usually lower than the effective temperature. The emissivity at a certain polarisation, frequency and angle is a function of subsurface extinction and reflections between layers with different permittivity. These parameters and processes are described in the emission model. We use aand develop a sea ice version of Microwave  Emission Model for Layered Snowpacks (MEMLS). MEMLS treating the processes for fresh snow on any given surface is described in Wiesmann & Mätzler, RSE 70: 307-316 (1999). The sea ice version is described in Mätzler et al. (eds) Thermal microwave radiation - Applications for Remote Sensing, IEE Electromagnetic Wave Series, 2006.

Brightness temperature and emissivity simulations
Input to the forward model is a stack of layers with information about: temperature, layer thickness, grain or inclusion sizes, salinity, liquid water content for each layer.  These parameters can be measured in snow pits and by ice coring or they can be simulated using thermodynamic models. Earlier studies with snowpacks on land indicate that the thermodynamic models tend to underestimate the sharp desity contrasts between snow layers. This may also be a problem in snow pits since snow layers on sea ice are often thin and the snow cover often only 10-20cm thick. Smal scale spatial variability is difficult to characterise with both methods. Regional climatic differences and interchannel correlations may be characterised in thermodynamic models using seasonal simulations.
Seasonal emissivity simulation

Radar altimeter backscatter modelling

The backscatter coefficient for a space-borne altimeter system of snow free, level sea ice is primarily a function of ice roughness and coherent surface scattering surface scattering mechanisms witin the snow/ice system. Attenuation together with volume scattering detirmines the exinction of the propagating signal within the snow and ice layers. Nadir backscatter of snow covered sea ice is dominated by ice surface surface scattering. Yet the link between the prominent scattering horisons and the effective scattering surface, when the altimeter is used for surface elevation measurements, is unclear. These issues are important for the success of the upcoming CryoSat. A theoretical prediction is given using a backscatter model. The model is described in Tonboe et al. IEEE GRSL 3(2): 237-240. A gnu-octave (similar to matlab) version of the model can be downloaded here (tar-file).

Read more (pdf):
Tonboe, R. T., S. Andersen, R. S. Gill, L. Toudal Pedersen. The simulated seasonal variability of the Ku-band radar altimeter effective scattering surface depth in sea ice. In: Wadhams & Amanatidis (Eds.) Arctic Sea Ice Thickness: past, present and future. Climate change and natural hazards series 10, EUR 22416, 2007.

The figure show the simulated leading edge of the multiyear ice pulse form for different snow cover depths. The curve to the left is for the 5-cm snow cover, the middle curve the 25-cm snow cover and the right curve the 50-cm snow cover.
return pulse form

Collaboration

The models were developed in the former EU projects IOMASA lead by Georg Heygster and GreenICE lead by Peter Wadhams developments are still ongoing in the EU contribution to IPY DAMOCLES project lead by J.-C. Gascard.