Impact of Sea Ice Thickness and Freeboard Products on Performance of Seasonal Forecasts

The ESA-funded A+5 study has constructed a flexible system for Arctic Mission Benefit Analysis (ArcMBA) that evaluates in a mathematically rigorous fashion the observational constraints imposed by individual and groups of EO data products in using the quantitative network design (QND) approach. The assessment of the observation impact (added value) is performed in terms of the uncertainty reduction in a four-week forecast of sea ice volume (SIV) and snow volume (SNV) for three regions along the Northern Sea Route by a coupled model of the sea ice-ocean system. The assessments covered seven EO products, three real products and four hypothetical products. The real products are monthly sea ice thickness, sea ice freeboard, and radar freeboard, all derived from CryoSat-2 by the Alfred Wegener Institute (AWI). These are complemented by two hypothetical monthly laser freeboard products (one with low accuracy and one with high accuracy), as well as two hypothetical daily snow depth products (again one with low accuracy and one with high accuracy). In total 15 (groups of) products were assessed, and their performance compared with each other and a reference case without any observations (prior). On the basis of the per-pixel uncertainty ranges that are provided with the CryoSat-2 sea ice thickness (SIT), sea ice freeboard (SIFB), and radar freeboard (RFB) products, the SIT achieves a much better performance for SIV than the SIFB product, while the performance of RFB is more similar to that of SIT. For SNV, the performance of SIT is low, the performance of SIFB higher and the performance of RFB yet higher. A hypothetical laser freeboard (LFB) product with low accuracy has a similar impact as RFB on both SIV and SNV. A reduction in the uncertainty of the LFB product yields a significant increase in performance. Combining either of the SIT/freeboard products with a hypothetical snow depth product achieves a significant performance increase. The uncertainty in the snow product matters: A higher accuracy product achieves an extra performance gain. The provision of spatial and temporal uncertainty correlations with the EO products would be beneficial not only for QND assessments, but also for assimilation of the products. The ArcMBA can be extended to cover further EO products and further target variables.