THIN-ICE THICKNESSES AND ICE PRODUCTION IN ARCTIC POLYNYAS BETWEEN 2002 AND 2017 AS DERIVED FROM THERMAL INFRARED AND PASSIVE MICROWAVE REMOTE SENSING DATA.

A profound knowledge of the thin-ice thickness distribution and associated wintertime sea-ice production in polynyas is not only important to increase the understanding of atmosphere - seaice - ocean interactions, but also to evaluate frequently used climate and ocean models. In this study, we present a high-resolution (~2 km) MODIS thermal infrared satellite data set featuring spatial and temporal characteristics of 17 coastal polynya regions over the entire Arctic basin is presented for winter seasons 2002/2003 to 2016/2017. Besides the extended data record of 15 winter seasons, the data set features the most recent updates of the employed MOD/MYD29 sea ice product (Col. 6) which include improvements to the cloud masking and land/ocean separation. Thin-ice thickness distributions (≤ 20 cm) are calculated from MODIS ice-surface temperatures, combined with ECMWF ERA-Interim atmospheric reanalysis data in an energy balance model. Daily thin-ice thickness composites are subsequently computed. A gap-filling approach is applied to account for clouds and missing data in the thermal infrared imagery. The most recent winter season in the data set (2016/2017) features the so far highest annual ice production with a volume of 2,258 km³, presumably originating from highly active polynyas throughout the eastern Arctic (Laptev Sea, Kara Sea) and/or a particularly thin and less extensive wintertime sea-ice coverage that season. We additionally use thin-ice thicknesses from AMSR-E passive microwave data for the period 2002/2003 to 2010/2011 to evaluate and quantify sensor-specific differences. AMSR-E ice thickness data are based on an empirical approach that utilizes a distinct polarization ratio (PR) - ice thickness relationship for thin seaice. This direct comparison enables further insights into sensor-specific capabilities and error sources for a large-scale pan-Arctic polynya monitoring. Despite obvious differences, such as the acquired signal at the sensor, varying spatial resolutions or the individual sensitivity for cloud cover, both data sets are coherent in terms of capturing the general spatial and temporal characteristics of Arctic polynyas for the overlapping 9-yr period. [ABSTRACT FROM AUTHOR]