Description of task
DAMOCLES will encompass the current state of the ice cover beyond its thickness, historical time series and operational sea ice types and properties from satellite sensors will be used for analysis of the current state and also feed into assimilation in oceanic and atmospheric models in WP4.
- Ice concentrations are determined at large scale from multi-frequency passive microwave sensors, which are independent of daylight and penetrate the atmosphere. Such data are available since 1978, however the many successive sensors of this period of nearly 30 years had slightly different characteristics so that the resulting ice concentrations are not directly comparable. Currently, a reanalysis of the complete period (1978 to date) is undertaken as a joint activity of DMI, met.no, DTU, the UK Met Office and NSIDC.
- Sea ice emissivity has turned out to be a crucial parameter for remote sensing of properties of both sea ice and the atmosphere above. The microwave signature, i.e. the sea ice emissivity and backscatter coefficient, is determined for the involved ice types with snow layers as a function of (1) the meteorological history (atmospheric temperature, precipitation, wind, ice drift), and (2) the different polarizations and frequencies between 6 and 190 Ghz in several studies:
- An analytical model based on the MEMLS emissivity model combined with a thermodynamic snow evolution model (see Figure) is developed and will be validated against field observations of snow and ice. A low-cost field experiment to provide adequate data for this purpose was carried out in April 2007 to be combined with field activities of Task 1.
- Comparing satellite observations to simulated ones using NWP atmospheric profiles in radiative transfer calculations.
Figure 1: AMSU first-year and multiyear sea ice emissivities, from Mathew et al. (2007).
Figure 2: AMSR first-year ice emissivities at 7 and 10 GHz. Emissivities of other frequencies and multiyear ice available from Mathew et al. (2007)
Figure 3: Correlations of all AMSR-E emissivities for first-year ice, Nov. 2005. From Mathew et al., 2007.
- Ice motion and deformation fields will be derived from several satellite sensors satellites (SSM/I, AMSR-E, QuikScat scatterometry, ENVISAT ASAR, RADARSAT), validated by drifting buoy and experimental in-situ data. In order to retain small scale ice motion features and the comparably lower noise level of the high resolution motion fields, separate developments are undertaken for high and low resolution sensors. A specific analytical effort will be undertaken to determine a relationship between small and large-scale ice motions.
Figure 4: Left: Displacement vectors at 2-days coarse resolution from QuikScat scatterometer (left) and from 1-day fine resolution from ENVISAT ASAR WSM data (right). Image: courtesy of Leif-Toudal Pedersen, ltp@oersted.dtu.dk.
Figure 5 : SeaWinds/QuikSCAT data over Arctic for 1st December 2004. Multi-year sea ice is in red, yellow and green. First year ice is in blue. Grey is open ocean. Image courtesy of Fanny.Ardhuin@ifremer.fr.
- Ice types discrimination into first-year ice and multiyear ice can be done with scatterometer (radar) data, as it has been shown by IFREMER. At present, the operating scatterometer is SeaWinds/QuikSCAT. The analysis of Arctic sea ice signatures highlights the difficulty to estimate the threshold value to be used for the discrimination. This is validated at the annual sea ice area minimum with radiometer data when sea ice becomes multiyear ice, by definition. Next step is to adapt this with MetOp/ASCAT data. The combination of ASCAT and QuikSCAT data at two different frequencies would greatly improved the ice types discrimination.
- N. Mathew, G. Heygster, C. Melsheimer, L. Kaleschke 2006: Surface emissivity of polar regions at AMSU window frequencies. Submited to IEEE Trans. Geosci. Rem. Sens.
