Author(s): Giacomo De Carolis; Francesca De Santi
Linked Author(s):
Keywords:
Abstract: A new SAR inversion scheme to estimate the thickness of sea ice in the marginal ice zone through observation of wave decay from SAR imaging is proposed. This scheme applies to sea ice composed by mixtures of sea ice prevalently composed by grease and pancake ice (GPI) and thin ice floes. A close packing (CP) model (De Santi & Olla 2017) is adopted where GPI is treated as a three-layer model, with an infinitely thin top layer accounting for the effect of the pancakes, an infinite depth inviscid layer representing the ice-free ocean on the bottom and a viscous layer in the between accounting for the effect of grease ice. Predicted waves attenuation depends on GPI thickness and the grease ice viscosity. Comparison between waves decay observed from SAR imaging and waves decay predicted by the CP model is the bedrock of the proposed scheme (De Carolis 2001). Unfortunately, as it happens for any viscous layer model, infinite combinations of thickness and viscosity predict a same value for attenuation. The resulting SAR inversion problem results therefore undetermined (De Santi et al. 2018). Removal of such inconsistency is performed by calibrating the SAR inversion scheme with modeled GPI thicknesses in the Odden Sea (Pedersen & Coon 2004). A physically consistent GPI viscosity-thickness relationship is obtained. Two examples of application for the proposed scheme will be presented: a Sentinel1 C band, HH polarized SAR image taken in the Beaufort Sea on 1 November 2015 and a COSMOSkyMed X band, VV polarized SAR image taken in the Weddell Sea on 30 March 2019. The estimated GPI thicknesses are consistent with SMOS measurements with somewhat higher variability due to the higher spatial resolution of the estimates.
Year: 2020