Sea Ice Detection With High Sampling Resolution Using Spaceborne GNSS-R Complex Waveform Data

Sea ice plays a crucial role in global climate patterns, making the acquisition of sea ice change information significant. The rapid development of GNSS and LEO satellites has facilitated the emergence of spaceborne GNSS-Reflectometry (GNSS-R) as a novel remote sensing approach. Previous studies predominantly used Delay-Doppler maps (DDM) to detect sea ice, which posed challenges in identifying small-scale sea ice information due to the sampling rate (1 Hz) of DDM. This paper proposes a method for high along-track spatial sampling resolution detection of sea ice and ice leads using complex waveform (CWF) products for the first time. CWF provides amplitude and phase information at a higher sampling rate (1000 Hz), which can potentially construct high-resolution sea ice detection observables. In this paper, the mean coherence coefficient (MCC) observable with 20 ms temporal resolution is extracted to quantify the difference in residual phase change between sea ice and open ocean. Then, a corresponding MCC observable threshold is established based on the training dataset to realize sea ice detection. The method is validated using TDS-1 CWF with OSISAF SIC datasets as references. The results show detection accuracies of 93.59% and 83.84% for the northern and southern hemispheres, respectively, coupled with a remarkable fifty-fold improvement in along-track spatial sampling resolution. In addition, the method was also used to identify ice leads in Davis Strait. The results revealed a high correlation coefficient of 78.42% between MCC observables and surface reflectance values extracted from MODIS products, preliminarily proving the feasibility and application potential of spaceborne GNSS-R in ice leads detection.