Remote Predictive Mapping of the Tunnunik Impact Structure in the Canadian Arctic using Multispectral and Polarimetric SAR Data Fusion

• Published in: Canadian Journal of Remote Sensing
• Main Authors: Byung-Hun Choe, Livio L. Tornabene, Gordon R. Osinski, Jennifer D. Newman
• Format: Article in Journal/Newspaper
• Language: English; French
• Published: Taylor & Francis Group 2018
• Subjects: Environmental sciences; GE1-350; Technology; T; Arctic; Canada; Victoria Island; victoria island
• Online Access: https://doi.org/10.1080/07038992.2018.1544846; https://doaj.org/article/4bb720746cef4821b19d4113331b58a6

The 28-km diameter Tunnunik impact structure in northern Victoria Island, Arctic Canada, was mapped using ASTER, Landsat 8, RADARSAT-2 polarimetric synthetic aperture radar (SAR), and Quickbird data. Multispectral analysis was accomplished through band ratios, MNF transform, and spectral matching algorithms, from which 4 distinct spectral units were defined. Polarimetric SAR decompositions also showed different scattering mechanisms for these 4 units indicating different surface roughness properties. These multispectral and polarimetric SAR observations were combined with detailed surface textures and morphological features as visible in very high-resolution Quickbird imagery (61 cm/pixel). Remote sensing parameters and their thresholds for characterizing each unit were implemented into a decision-tree algorithm and a remote predictive geological map was produced. Subsequent field and follow-up laboratory investigations enabled the ground-truthing of these predictions. The geological units were defined as follows: (i) (smooth) fluvioglacial deposits, (ii) (moderately rough) chert-bearing dolostone, (iii) (rough) dolostone, and (iv) (rough) dolostone covered by silicified surfaces. The rough surfaces characterized by multiple scattering in the polarimetric SAR decomposition correspond to the occurrences of weathered carbonate rocks, which are relatively resistant to weathering and form blocky surfaces. This shows that SAR-derived surface roughness properties can greatly contribute to defining geological units by combining with lithological mapping.