CryoSMOS is an European Space Agency (ESA) project issued in the Support to Science Element (STSE) framework. It is aimed at exploiting SMOS mission data (mainly brightness temperature) in order to investigate open issues about Antarctic cryosphere precesses.

Besides SMOS measurements, other data will be analysed, for instance data collected in previous ESA campaigns like DOMEX-1, DOMEX-2, DOMEX-3 and DomeCAir. Other relevant EO data like SSM/I, AMSRE-2 and AMSU data will be used in the project are free available by web portals (e.g. NSIDC). Also ground data available by web portals (e.g. Community Ice Sheet Model (CISM) data or NSIDC) or provided by data producer as AWI or BAS will be considered. Moreover relevant information will be obtained from BEDMAP-2 project datasets.

Up to now, four preliminary test cases have been individuated and will be considered in the first part of the study.

Quantification internal ice-sheet temperature

Ice sheet temperature is a primary factor in determining the ease at which ice deforms internally and also the rate at which the ice flows across the base. Consequently, ice temperature is fundamental for better understanding changes in ice sheet mass balance and dynamics. Temperature enters as a variable in glacier dynamics through a flow law typically taken as a power law relation between stress and strain rate (Nye, 1953; Glen, 1958). The flow rate is determined by the strain rate which in turn depends on a stiffness parameter that is highly temperature dependent (Hooke, 1981). Because ice sheet temperature varies with depth, the stiffness parameter and hence strain rates vary with depth. Temperature as a function of depth is therefore a key information for modeling ice sheet flow (see for example Van der Veen, 1999). Presently, the only information about ice sheet internal temperature arises from the small number of boreholes in which temperatures are directly measured and from modeling. Preliminary results presented demonstrate that SMOS Tb is sensitive to the temperature profile variation which is determined by the variation of bedrock topography and/or surface temperature. In order to better understand the relationship between TB of SMOS and temperature profile we plan to investigate in detail the area of Dome C where detailed ground information (e.g. temperature profile, snow layering, snow density ,etc.) are available and then to extend the analysis on other areas of Antarctica. The e.m. model previous suggested (DMRT or WALOMIS) are able to simulate the Tb of this area and will be useful for perform a sensitivity analysis and to develop a methodology able to infers temperature information from SMOS data. The use of other EO data, as for example Tb collected at higher frequencies by AMSRE-2, will be also investigated in order to derive temperature information for the first meters of the ice sheet and relate them to the deeper layers using geophysical models. Surface temperature information will be derived by Modis data - MOD11C3V41 while bedrock topography from BEDMAP-2 database. Boreholes data, available in other areas of Antarctica, will be used for algorithm validation.

Bedrock topography and/or geothermal heat flux

As pointed out in section 2 during DOMEC-Air experiment L-band microwave airborne data were collected by the EMIRAD radiometer system was flown over a 350 km by 350 km area near the Italian/French Concordia station close to Dome-C in Antarctica (Kristensen et al., 2013). The area is quite flat at an altitude of some 3000 m above sea level. The yearly mean temperature at Concordia is -55°C corresponding to 218 K. Figure 6 demonstrate that Tb data exhibits a variation of more than 10 K are noticed, and if a comparison with the bedrock topography beneath the ice is carried out, a convincing correlation is observed. In the area in question the bedrock range from some -1000 m below sea level to about +500 m above sea level, i.e. the ice thickness varies by 1500 m (from 4000 m to 2500 m). Moreover it is also demonstrated that the Tb is strictly related to gravity anomaly data. As pointed out previously, this results can be explained as a combination of effects as bedrock elevation, bedrock temperature which is due to geothermal heat flux and other parameters including snow year accumulation. Preliminary investigations demonstrate that a simple e.m. model is able to explain the obtained results at least for the area near to Concordia station. During the project execution it will be analyzed how the data acquired at high resolution scale from DOMEC-AIR are related to the low resolution scale of SMOS then how the local spatial properties of bedrock leads to an integrated results. A more advanced electromagnetic model will be also used to analyze and interpret the data and the up-scaling issue. Once that the DOMEC-Air data will be fully exploited and investigated, SMOS data collected in other region of Antarctica will be analyzed. For this purpose BEDMAP-2 data will be useful as auxiliary data.

Characterising Ice shelves

the other hand SMOS Tb data demonstrated his sensitivity to them. In particular we plan to investigate on three geophysical shelf ice parameters: the internal ice temperature (profile), the shelf ice thickness and the detection of basal marine ice. All of these parameters are poorly known and are fundamental to describe the dynamics of the ice shelves. In order to investigate on this issue it is important to observe a region wich is sufficiently large with resepct to the SMOS field of view. In this regards the Ross and Ronne-Filchner ice shelf are large enough to test the retrieval with real SMOS data. As previously pointed out we plan to use simulated temporal profile and an electromagnetic model for a theoretical sensitivity study to assess the vertical resolution of an ice temperature retrieval. In order to derive information about the vertical profile we plan to use Tb measurements at different incidence angles. A preliminary analysis using a simple radiative transfer model for a profile with ice surface temperature at T=-15°C and the basal temperature at T=-2°C suggests that the signal is saturated at a depth of about 500 m for a shallow incidence angle of 60°. Measurements at steeper near nadir angles carry information from layers deeper as 500 m. The set of polarised measurements at different incidence angles form an inverse problem for the retrieval. Similar as for atmospheric temperature sounding this problem can be solved with the optimal estimation method (Rodgers, 1976). The degrees of freedom (i.e. vertical resolution) will be analysed and what a-priori information (e.g. surface temperature) is required to constrain the inverse problem. It should be noticed that main retrieval problem are related to the relatively small scale of shelf ice areas and the coarse resolution of SMOS, the vicinity of coastal polynyas and their large variability and the sparseness or total absence of suitable validation measurements. In front of these potential problems, which will be investigated during the project execution , we have to admit the risk to obtain a large uncertainties in the derived products.

Characterizing surface processes (hoar, roughness, etc.)

The state of the surface, i.e. the geometrical shape of the surface and the snow physical properties at proximity of the surface (grain size, wetness, density), is important to understand and predict the surface energy and mass budgets which are two key nivo-meterological variables of interest for the study of the climate and contribution to sea level of the ice-sheets. Variations of near surface snow dielectric properties affect radiometer observations giving the possibility to retrieve information about the surface state. Despite the large penetration depth of L-band radiation, snow properties changes near the surface (up to one wavelength deep) do impact SMOS observations. In this context, the analysis of SMOS observations may provide information about the state of surface, such as melt event, the formation of hoar crystals on the surface, the variations of roughness, etc. First requirement will be the accurate characterization of the effect of surface state on SMOS observations at Dome C, to benefit from numerous dataset and studies performed in this area, in particular the two year-long time series of daily presence of hoar crystals, derived from in situ near-infrared photographs (Champollion et al., 2013). After that, electromagnetic modeling suggested in previous section (DMRT-ML, WALOMIS) will be used to extend understanding to the whole Antarctic ice-sheet. On the other hand, in snow wet zone, located mainly along coast and on ice-shelf, the state of the surface is affected by seasonal melting events, which influences strongly the snowpack. Indeed, the episodic presence of liquid water forms a refrozen snow layer and leads to sustained decrease of L-band brightness temperature observation after the melt season with respect to before the melt season. SMOS observations will be explored in synergy with other satellites, such as SSMI, AMSU and AMSRE and modeling studies will be performed in order to characterize snow properties variations during seasonal snowmelt. In a first time, analyses will be conducted on Ronne or Ross ice-shelves.

The SMOS mission

The SMOS mission, is one of the ESA's Earth Explorer missions which form the science and research element of ESA's Living Planet Programme. The single payload of the mission is the Microwave Imaging Radiometer by Aperture Synthesis (MIRAS) (Martin-Neira et al., 1997). This radiometer is the first spaceborne instrument to use interferometric aperture synthesis, a technique suggested back in the 80's as an alternative to real aperture radiometry for earth observation (Ruf et al., 1988).

The whole instrument includes 69 L-band state-of-the-art high-sensitivity receivers, three of them capable to operate as accurate, highly stable noise injection radiometers (NIR). Each receiver has a wide beamwidth antenna and can operate alternatively in two orthogonal polarizations, except the NIRs that are fully polarimetric. The antennas are evenly distributed along a "Y-shape" structure providing a spatial resolution similar to that of a real aperture antenna having the same overall dimension as the whole instrument. The instrument is tilted in an Earth-fixed attitude with a constant forward tilt angle of 32.5 deg between the instrument boresight and the local nadir, in the flight direction. This ensures an angular coverage of about -10 deg to +60 deg. The nominal swath is around 1000 Km. The average ground resolution is 43 km over land and the globe is fully imaged twice (ascending and descending orbits) every 3 days at 6 a.m. and 6 p.m. local solar time (equator crossing time).