Arctic polynyas: Retrieval of wintertime sea-ice production using satellite-derived thin-ice thickness distributions.

Knowledge of wintertime sea-ice production in polynyas is an important aspect ofatmosphere – sea-ice – ocean interactions in high latitudes, and a precise quantification canbe for instance utilized to verify frequently used climate and ocean models. This studyfeatures a high-resolution (2km) MODIS thermal infrared satellite data set with spatial andtemporal characteristics of 17 coastal polynya regions over the entire Arctic basin for2002/2003 to 2017/2018, as well as a similar data set based on lower resolution (6.25km)AMSR-E passive microwave satellite data for a nine-year overlapping comparison period (upto 2010/2011). The MODIS data set is relying on a 1D energy balance model, where quasi-dailythin-ice thickness composites (up to 20cm) are directly derived from ice-surfacetemperature swath-data and ERA-Interim atmospheric reanalysis data. Dedicated cloudscreening and spatial/temporal interpolation techniques are therein applied to effectivelyaccount for sensor-specific drawbacks. In case of the AMSR-E data set, thin-icethicknesses are derived using an empirical approach that utilizes a characteristicpolarization ratio (PR) – ice thickness relationship. In both data sets, the daily pan-Arcticmapping of thin-ice thicknesses allows for a long-term derivation of important polynyaproperties such as polynya area (POLA) and potential thermodynamic ice production(IP). It shows that the average POLA (average accumulated IP) for all Arctic polynyascombined is 1.99 x 105 km2 (1.34 x 103 km3) when derived from MODIS data and 2.23 x105 km2 (1.29 x 103 km3) in case of AMSR-E data. Although the two data sets areindependently derived, they show quite similar spatial and temporal variations of POLA andIP, which suggests a high reliability of both data sets. Hence, despite all methodicaldifferences, both data sets are to a large degree coherent in terms of capturing the generalspatial and temporal characteristics of Arctic polynyas for the overlapping 9-yr period.Emerging positive trends in POLA and IP over the long 16-yr period of the MODIS data setare mainly visible in the Eastern Arctic and are potentially related to large-scaleatmospheric modes and/or changing characteristics of the Transpolar drift system. [ABSTRACT FROM AUTHOR]