Differential Interferometry and Multiple-Aperture Interferometry for Retrieving Three-Dimensional Measurements of Glacial Surface Velocity

Sammanfattning: The measurement and monitoring of glacial surface velocity is important for many aspects of glaciology, such as determining the mass balance, for characterising the stability or instability of glaciers, or the identification of potential hazards from surging glaciers or Jökulhlaups, a type of glacial outburst flood. Predominately measurements of glacial surface velocity have been produced using either differential interferometry (DInSAR) applied to radar data, or offset-tracking applied to either optical or radar data. Both of these methods have their own set of limitations, notably the one-dimensional nature of DInSAR measurements, and the relatively low accuracy of offset-tracking. Instead using DInSAR and multiple-aperture interferometry (MAI) applied to ERS-1/2 Tandem SAR data, measurements of glacial surface displacements were obtained in the line-of-sight (LOS) and along-track directions respectively. Then using a weighted-least squares adjustment, the method for producing the full three-dimensional surface velocity field is presented and applied to the Svartisen glacial system, Norway and the Petermann Glacier, Greenland. The advantages and disadvantages of applying such a method were explored, of which interferometric coherence is found to be the largest factor in retrieving accurate measurements using MAI. Low interferometric coherence due to temporal decorrelation resulted in the inability to extract the full three-dimensional surface velocity field over the Bagley Icefield, Alaska, and the Mýrdalsjökull & Eýjafjallajökull ice caps, Iceland. A feasibility analysis into the use of Sentinel-1 data, revealed that the current revisit period is too large to maintain interferometric coherence between acquisitions, preventing the application of either DInSAR or offset-tracking in order to measure the surface velocity of the Blåmannsisen Glacier, Norway. Despite the limitations encountered, in part due to the selection of source data, MAI in tandem with DInSAR has been shown to be capable of measuring the three-dimensional surface velocity to a higher accuracy than offset-tracking when coherence is high. The methods used within have been developed to work with pre-processed single look complex (SLC) SAR data rather than raw unfocused SAR data, in an effort to improve their adoption and enable more accurate estimates of glacial surface velocity.