DEM differencing vs InSAR for pipeline slope monitoring
Pipeline integrity teams tracking slope movement over buried pipe usually land on one of two remote methods: differencing a digital elevation model against itself over time, or comparing radar phase between InSAR passes. Both are trying to answer the same question, where is the ground moving enough to threaten the pipe, but they get there differently, and they fail differently too.
What each method measures
DEM differencing compares two elevation surfaces built from stereo imagery, one from this year and one from last, and looks at the vertical change between them. A slope that's crept a few tens of centimeters, or a drainage that's cut a new channel across the ROW, shows up as a bump or a dip in the difference raster. It's a direct measurement of surface height rather than a proxy for it.
InSAR works from the phase difference between two radar passes over the same ground. It's extremely sensitive to small displacement, sometimes down to a few millimeters, but what it measures is displacement along the satellite's line of sight, so slope angle and look direction change what comes out of the processing. It also needs coherent targets between passes. Bare rock and gravel hold up fine. A ROW with seasonal grass, cultivation, or a fresh mow can decorrelate enough to lose the signal in exactly the stretch where a pipeline crosses farmland.
Where each one breaks down on a corridor
InSAR's trouble spots for pipeline corridors line up with where pipelines run: vegetated right-of-way, agricultural crossings, forested slopes, and anywhere the look angle sits close to parallel with the slope face. The method is built for a different kind of target, mostly open or urban ground with stable scatterers, and a long rural corridor doesn't always offer that.
DEM differencing doesn't have the same vegetation-decorrelation problem, since it only needs the stereo imagery to resolve the ground surface and give a clean elevation delta instead of a line-of-sight component. Dense canopy still hides bare earth from an overhead sensor, that part doesn't go away, but mowed or sparse ROW cover reads fine. What DEM differencing doesn't give you is InSAR's millimeter-scale sensitivity. At ≤0.5 m ground sample distance and an annual cadence, it's built to catch the kind of movement that matters for buried pipe, several centimeters to a few tens of centimeters of slope creep, or a washout scar, not a slope that's crept 2 mm since spring.
Picking a method for a corridor screening pass
An integrity engineer deciding where to send a geotech crew first doesn't need the method that's more precise in a lab sense. They need the one that produces a usable signal across the ground cover the corridor actually has. A line running through cultivated bottomland and brushy slopes is rough InSAR territory for a good chunk of its length. A line through open range or sparse scrub is friendlier to both.
Plenty of integrity programs end up running both: InSAR where coherence holds and the finer displacement sensitivity earns its keep, elevation differencing where ground cover or sensor geometry would kill the radar signal, or as a first annual pass to rank segments before committing to anything more sensitive. The two methods aren't competing for the same job. They measure different things and go dark in different places.
That's the gap Geohazard Screening sits in: a once-a-year elevation-differencing pass over the whole corridor that flags slope creep and washout signatures without needing coherent radar targets or a crew walking the right-of-way, so you know which segments earn a closer look before spending time on anything more detailed.
If your corridor has the kind of ground cover that leaves InSAR patchy, it's worth seeing what an annual elevation-differencing pass turns up instead.