Remote Sensing of Surface Signal and Infrastructure Exposure – Salt Lake City Airport Corridor

Term: Fall 2025

Courses: GEOG 6130 – Advanced Active Remote Sensing & GEOG 6150 – Geospatial Big Data

Tools: Sentinel-1 SAR (ASF HyP3), Google Earth Engine, QGIS, Python,
OpenStreetMap data, zonal statistics, grid-based spatial modeling

 What I built

I combined short-term InSAR line-of-sight (LOS) analysis with large vector datasets to evaluate surface signal behavior and infrastructure exposure around the Salt Lake City International Airport.

The project began as a radar signal and displacement assessment (6130) and evolved into a multi-layer geospatial big-data workflow (6150) integrating,

• Sentinel-1 InSAR products

• Building footprints

• Land-use polygons

• Grid-based aggregation

• A composite exposure index

The problem

Can short-term InSAR signal patterns, even when dominated by atmospheric effects, still provide meaningful spatial insight when combined with infrastructure data?

A 12-day interferogram alone is not enough to confirm ground deformation. The challenge was determining whether radar-derived surfaces could still be useful when integrated into a structured spatial workflow.

My approach

I treated the LOS raster as a measurable environmental surface rather than confirmed deformation.

InSAR Processing (6130):

• Masked low-coherence pixels (< 0.3)

• Converted LOS from meters to centimeters

• Built a 500 m UTM grid and averaged LOS per cell

Verified that displacement patterns were atmospheric, not subsidence

Big Data Integration (6150):

• Performed zonal statistics on thousands of building footprints

• Summarized LOS behavior by land-use class

• Calculated building density per grid cell

• Developed a composite vulnerability index:

V = (Normalized LOS + Building Density + Land-Use Weight) / 3

This created a scalable exposure model rather than a simple raster map.

Key Findings

• No measurable ground deformation in the 12-day interferogram

• LOS signal shows a clear south-to-north gradient

• Strongest LOS values align with the southeastern industrial corridor

• Industrial land use + high building density + elevated LOS produce the highest composite index values

• Farm and open land consistently fall in the lowest exposure class
  
  

While the LOS field was atmospheric, the integrated workflow revealed consistent spatial structure tied to development patterns.

Technical Contributions

• End-to-end InSAR preprocessing and coherence masking

• Raster-to-vector integration at building and grid scale

• Multi-criteria index development and normalization

• Entity-relationship data model design

• Structured geospatial big-data pipeline combining heterogeneous datasets

Outcome

This combined project demonstrates how SAR products can be integrated with vector infrastructure data to produce meaningful exposure models, even when short-term displacement signals are dominated by atmospheric noise.

It reinforced two key lessons:

1. Not all radar-derived displacement is deformation

2. A strong data architecture can extract value from imperfect signals

The workflow is scalable and transferable to other infrastructure corridors where monitoring surface stability and exposure overlap is critical.