The effect of InSAR time series generation techniques on signals with small spatial scales

Paula Burgi, & Rowena B. Lohman

Published August 15, 2017, SCEC Contribution #7701, 2017 SCEC Annual Meeting Poster #211

As the rate of growth of InSAR catalogs continues to increase, analysis of these large datasets must be done thoughtfully and efficiently to manage the computational burden. One of the key decisions made during InSAR time series generation is the method for aligning the data to a common reference frame. This results in tradeoffs between coherence, resolution, and computational speed and depends on the spatial scale of the signal(s) of interest. Here, I explore these trade-offs by comparing times series generation techniques for data spanning a small, slow-moving landslide 15 km northeast of San José with Sentinel-1B data. In January 2017, the rate of downslope motion of this landslide increased from ~3 cm/yr to ~78 cm/yr, based on continuous GPS data from a monument located in the middle of the landslide. Although this rate is detectable with Sentinel data, the landslide has a spatial scale of only 1500 by 100 m. This small spatial scale and high deformation rate (for InSAR data) means that standard processing approaches fail in extracting the signal. Our initial results show that, during the period of most rapid deformation, we see approximately one full phase cycle, corresponding to 28 mm, of LOS deformation in a 10 pixel-wide swath for a 12-day pair. This results in a strain of ~2.8 mm/pixel, nearing the limit of observable strain before decorrelation. Additionally, this signal is only observed clearly at the crown of the landslide, which is slightly wider than the downslope dimensions and therefore defines the spatial resolution limit of this dataset.

Key Words
InSAR, landslide, processing techniques

Citation
Burgi, P., & Lohman, R. B. (2017, 08). The effect of InSAR time series generation techniques on signals with small spatial scales. Poster Presentation at 2017 SCEC Annual Meeting.

Related Projects & Working Groups
Stress and Deformation Over Time (SDOT)