Derivation of high-quality three-dimensional surface deformation velocities through multi-source point cloud fusion: Application to Kīlauea volcano

Abstract

Abstract Kīlauea volcano, located in the southeast of Hawai’i, is one of the most active volcanoes in the world. Multi-source InSAR measurements acquired from ascending and descending high-resolution COSMO-SkyMed SAR imagery are used to monitor Kīlauea volcano’s deformations. In order to obtain identical observations from multi-source SAR data, the InSAR measurements acquired by different platforms are generally geocoded and resampled into a unified coordinate system. Such method is however inapplicable to the point clouds provided by high-resolution SAR data. In addition, the geocoding accuracy of high-resolution InSAR measurements would be greatly affected by external DEM error, orbit error, propagation delay, reference point parameter error. In this study, a method for estimating 3-D deformations on the basis of SAR point cloud fusion is proposed. Firstly, iterative closest point (ICP) method is introduced to fuse InSAR point clouds from ascending and descending COSMO-SkyMed datasets to reduce the influence of geocoding offsets. Then, the precise 3-D surface deformations of Kīlauea volcano are obtained by employing the homogeneous weighted observations in the local space based on the strain model. The maximum deformation rates between April 2012 and April 2013 of Kīlauea volcano are estimated as 32\xa0mm/year and 42\xa0mm/year in the east and up directions, respectively. The accuracies of the InSAR measurements are validated by the deformation rates obtained from the precise GNSS observations.