Parallelized DInSAR Processing Using High-Performance Computing Techniques

Abstract:

The recent growth in access to Synthetic Aperture Radar (SAR) data through the European Space Agency (ESA) Sentinel-1A/B satellites and upcoming NASA-ISRO Synthetic Aperture Radar (NISAR) mission provide increased opportunities for differential interferometric synthetic aperture radar (DInSAR) monitoring. Although a wealth of data is now available, processing software licenses are often cost prohibitive. Existing open source options have a steep learning curve, after which months of processing time is often necessary to produce research. We combined open source processing software InSAR Scientific Computing Environment (ISCE), a Docker container developed by the Alaskan Satellite Facility (ASF), and a software wrapper to allow large scale parallel processing of SAR images through the Rocky Mountain Advanced Computing Consortium (RMACC) Summit supercomputer at the University of Colorado Boulder. We modified the original ASF Docker for compatibility with Summit and modified it to allow for incorporation of a higher resolution DEM than those available from the Shuttle Topography Radar Mission (SRTM) at 10 or 30 meter pixel resolution. We employ freely available tools to develop a software wrapper that allows for updates to DInSAR time series in near real time, bypassing common issues in installation, updates and processing. We provide computer benchmarks for comparison of processing time on various architectures. Our workflow simplifies DInSAR analysis in both setup and processing time, allowing for a greater ease of access and a focus on the dynamics of the underlying geophysical processes at high spatial and temporal resolution. Here we provide preliminary results from Lagos, Nigeria and Barranquilla, Colombia, coastal cities undergoing local and regional subsidence as a result of natural and anthropogenic changes.