Background and justification of selected algorithm

Sea Ice Concentration (SIC3H)

The proposed algorithm for SIC3H stems from the long history of developments by the EUMETSAT OSI SAF and the ESA CCI Sea Ice teams.

Since the early 2000s, the OSI SAF have been developing state-of-the-art algorithms including swath-based retrievals, dynamic generation of tie-points, atmospheric correction of the brightness temperatures using RTMs, and per-FoV uncertainties. Traditionally, these algorithms have used the KU and KA microwave bands from the SSMIS and later the AMSR2 missions.

Since the mid 2010s, the CCI Sea Ice projects have contribute critical R&D input to the original OSI SAF approach. Of particular relevance to the proposed CIMR Level-2 SIC algorithm are the demonstration of the benefits of using C-band (in combination with KA-band) radiometry, and the demonstration of using pan-sharpening techniques to take advantage of high(er) resolution microwave imagery. Lavergne et al. [2019] gives an overview of the algorithm baseline including the use of C-band radiometry, and Lavergne et al. [2021] describes the use of pan-sharpening (using near-90 GHz imagery).

The proposed CIMR Level-2 SIC algorithm take full advantage of these approaches by exploiting the C-, KU-, and KA-band imagery of CIMR with dynamic tuning of the tie-points, and pan-sharpening techniques to obtain <5 km, <5% SICs meeting the CIMR mission requirements ([AD-1]).

Although many developments have been materialized through OSI SAF and CCI SIC R&D, the proposed algorithm also relates to earlier and recent developments by other investigators, e.g. the Bootstrap and Bristol SIC algorithms [Comiso, 1986, Smith, 1996] and the use of pan-sharpening techniques [Kilic et al., 2020, Kloster, 1996].

Sea Ice Concentration 1H (SIC1H)

The CIMR SIC1H Level-2 product has two driving requirements: the Near Real Time 1H latency, and the <5 km resolution to support operational sea-ice services with safe navigation. Compared to the nominal SIC3H Level-2 product, the requirement for accuracy of the SIC1H Level-2 product is relaxed.

We propose to implement the CIMR SIC1H Level-2 processing chain with building blocks from the SIC3H algorithm but removing some steps that require most computing time. For example, the SIC1H product might be built only on KU and KA imagery, and do not use RTM correction of the brightness temperature. Another time-saving step can be to not estimate uncertainties, or at least with a simplified algorithm.

Sea Ice Edge (SIED)

We derive Sea Ice Edge from the SIC, by finding the FoVs whose SIC is higher than a threshold (TBD, but generally 15%). This allows a low-cost derivation of this Level-2 product, that ensures consistency with the SIC product.