Bottom-fast ice delineation with PolSAR and InSAR techniques in the Mackenzie Delta region, Northwest Territories, Canada

Abstract

Bottom-fast ice (BFI) refers to sea or lake ice that freezes to the sea or lake bed. The formation of BFI plays an important role in the Mackenzie Delta Region's seasonal environmental changes. This study involves examining the use of Synthetic Aperture Radar (SAR) images for delineating BFI from floating ice on a study site within the coastal region of the Mackenzie Delta. The ultimate aim is to establish robust BFI monitoring tools using SAR data processing techniques. BFI detection methodologies based on polarimetric SAR (PolSAR) image classification and interferometric SAR (InSAR) coherence segmentation have been examined using RADARSAT-2 data. Polarimetric analysis can use individual SAR images and is not susceptible to decorrelation factors such as weather events. By using these analysis techniques it is thus possible to delineate BFI extent throughout the winter. A limitation of these techniques is the sensitivity to the water and consequently the ice saline content. The polarimetric decompositions and classifications for delineation of BFI apply only to areas where the water has a low saline concentration. Conversely, InSAR methods for BFI delineation require interferometric datasets with high temporal coherence. The results of the interferometric datasets can eliminate false alarms in BFI maps caused by similar backscattered SAR signal from floating ice with varying degrees of salinity. The delineation results using both methods were analyzed and compared. It was found that though each method is able to detect and delineate BFI on its own, reliable and more accurate results occur when the two methods are combined, as the reliance on different phenomena allow for false alarm identification.