SAR and Passive Microwave Fusion Scheme: A Test Case on Sentinel‐1/AMSR‐2 for Sea Ice Classification

Abstract

In the new millennium, sea ice research has become an important topic due to an unprecedented sea ice decline in the Arctic (England et al., 2020). One reason is that sea ice plays an essential role in the polar ecosystem (Funder et al., 2010). Moreover, the knowledge about sea ice conditions is crucial for polar navigation, offshore operations, weather forecasting, and climate research (Sandven et al., 2006). The main sources of information about sea ice conditions and climatological studies are data from passive microwave radiometers (PMR), and synthetic aperture radars (SAR). The latter is preferably used for tactical navigational support and for local studies requiring data at high spatial resolution. Both sensors are commonly used due to their independence of cloud and light conditions and therefore their ability to provide imaging of the Earth's surface continuously during day and night, and for almost all weather situations (Wang et al., 2016).Spaceborne SAR provides high spatial resolution images and is one of the main sources from which detailed maps of sea ice conditions are produced for navigation in sea ice or at the ice margin (Karvonen, 2014). SAR is an active microwave sensor, which can achieve spatial resolutions ranging from about 1 m to 100 m (Johannessen et al., 2007). Image products of lower resolutions provide swath widths up to 500 km. An increased spatial resolution reduces the achievable swath width down to a few kilometers. The interpretation of SAR data is challenging due to the complex relationship between radar backscatter and sea ice surface and volume properties and strongly relies on the knowledge of sea ice experts (Zakhvatkina et al., 2019).Passive microwave radiometers are another type of sensor and are commonly used for large-scale sea ice observations. These sensors operate at multiple frequencies, each of which has different spatial resolutions and sensitivity to atmospheric parameters, in particular to cloud liquid water and atmospheric water vapor (Spreen et al., 2008). However, even the finest spatial resolution achievable with passive microwave sensors is about 3 km and hence significantly coarser than in the case of SAR. The finest spatial resolution is achieved at frequencies of around 90 GHz which, however, comes with a greater susceptibility to atmospheric noise. This effect is less severe at