GROUNDPULSE

Pipeline Corridor Displacement Screening — Methodology Exhibit

Permian Basin crude corridor (Midland–Monahans–Wink), West Texas · Sentinel-1 C-band InSAR displacement baseline, 2025-08-02 → 2025-12-30

This exhibit demonstrates GroundPulse’s per-segment displacement screening on a real Sentinel-1 interferometric time series over the corridor. Every figure is measured.

Permian corridor per-segment displacement screening map — Per-segment line-of-sight displacement rate (mm/yr) along the ≈71-mi corridor, Midland (NE) → Monahans/Wink (SW), over the 15-acquisition baseline (2025-08-02 → 2025-12-30). Centerline colored by ordinary-least-squares LOS rate — the same per-segment values tabulated below. Basemap © OpenStreetMap, © CARTO. Positive = motion away from the sensor (LOS, relative baseline) — consistent with subsidence documented in this basin (Staniewicz et al. 2020). Geometry is an approximate public corridor centerline, not an operator alignment sheet; LOS ≠ axial pipe strain.

Corridor screened

71 mi

Segments

Acquisitions

LOS rate range

−7.3–+46.3 mm/yr

Per-Segment Line-of-Sight Displacement Rate

Measured LOS velocity per 2-mile segment over the 15-acquisition baseline (2025-08-02 → 2025-12-30), ordered NE → SW along the corridor. Positive = motion away from the sensor (LOS, relative baseline; consistent with basin subsidence per Staniewicz et al. 2020). Each segment carries the full 15-epoch series. No risk tier or hazard class is asserted here — see “What this shows” below. Geographic anchors are approximate: segments 1–6 are the NE Midland lobe; segments 8–19 cross the West Odessa near-stable section; segments 20–31 cover the central Monahans–Wink lobe; segments 32–36 taper toward the SW end. Exact attribution to operator stationing requires the operator alignment sheet.

Segment	Approx. milepost	LOS rate (mm/yr)	Obs.
1	0.0	+37.2	15
2	2.0	+44.9	15
3	4.0	+46.3	15
4	6.0	+29.7	15
5	8.0	+21.0	15
6	10.0	+15.3	15
7	12.0	+7.2	15
8	14.0	−2.1	15
9	16.0	−1.8	15
10	18.0	−1.0	15
11	20.0	−6.9	15
12	22.0	−7.3	15
13	24.0	−5.8	15
14	26.0	−2.2	15
15	28.0	+1.6	15
16	30.0	+4.9	15
17	32.0	+2.5	15
18	34.0	+4.2	15
19	36.0	+6.4	15
20	38.0	+14.3	15
21	40.0	+20.7	15
22	42.0	+17.1	15
23	44.0	+20.6	15
24	46.0	+22.4	15
25	48.0	+23.3	15
26	50.0	+25.4	15
27	52.0	+20.5	15
28	54.0	+20.6	15
29	56.0	+29.1	15
30	58.0	+25.9	15
31	60.0	+18.4	15
32	62.0	+18.6	15
33	64.0	+15.2	15
34	66.0	+15.4	15
35	68.0	+1.9	15
36	70.0	−7.2	15

Spatial pattern: two coherent subsidence lobes — a sharp peak (≈46 mm/yr) at the NE (Midland) end and a broader secondary lobe (≈25–29 mm/yr) along the central corridor — separated by a near-stable to slightly-upward mid-section (West Odessa) and tapering to near-stable at the SW. The structure is spatially organized, not scattered noise.

What This Shows

Shows: where, and by how much, the ground along the corridor moved over the 2025-08-02–2025-12-30 window, per segment, from independent satellite radar — a prioritization of segments by measured movement.

Hazard Mechanism & Independent Corroboration

The measured subsidence gradient is consistent with the documented driver of deformation in this part of the Permian: hydrocarbon and produced-water extraction and subsurface fluid movement, which have produced widespread, spatially-variable surface deformation across the Pecos–Reeves–Ward–Winkler oil patch — including localized subsidence of several cm/yr near legacy wells and injection-related uplift. The independent peer-reviewed record:

Staniewicz, S., et al. (2020), “InSAR reveals complex surface deformation patterns over an 80,000 km² oil-producing region in the Permian Basin,” Geophysical Research Letters, 47, e2020GL090151. doi:10.1029/2020GL090151

Method Notes

Sensor: Sentinel-1 A/B, C-band. Source: GroundPulse Dolphin CSLC surface-displacement product (GroundPulse in-house), over the corridor. Per-segment aggregates computed over pixels within each segment’s right-of-way buffer; up to 4,666,813 coherent samples per acquisition across the corridor AOI. Velocity is an ordinary-least-squares rate over the 15-epoch series.

Model & Parameters

Sensor/product: Sentinel-1 A/B C-band; GroundPulse in-house dolphinRust CSLC

Dolphin product lineage: GroundPulse dolphinRust CSLC run; frame F20697

Window: 2025-08-02 → 2025-12-30; 15 acquisitions

Rate model: ordinary least squares over the 15-epoch segment series

Coherence gate: temporal coherence ≥ 0.5 before AOI/ROW containment

ROW buffer: 500 m centerline buffer in the EIA import path

Outlier policy: MAD reject σ=3.0 for along-pipe kinematic indicators

Samples: up to 4,666,813 coherent pixels per acquisition across the corridor AOI

Reference handling: relative Dolphin LOS baseline; no GNSS datum tie asserted here

Storage: per-pixel evidence remains raster/COG; report uses aggregate serving values

Line-of-sight (LOS) geometry: Sentinel-1’s near-polar orbit heading (~13° off north) makes the radar largely blind to N–S motion, so N–S-trending segments are under-measured; at ~29–46° incidence, reported rates mix vertical and E–W components and are LOS, not vertical.
LOS velocity is not axial pipe strain. True strain requires pipe geometry, boundary conditions, and soil–pipe interaction modeling (ALA/ASCE guidelines). These rates screen where ground is moving, not pipe stress.
InSAR measures surface line-of-sight motion. It does not model trench settling, pipe embedment, or depth-of-cover change; those are separate geotechnical / integrity analyses.
Reference-point stability is assumed; a drifting reference is an InSAR-internal bias source distinct from trench settlement or depth-of-cover change.
Temporal decorrelation in vegetated, disturbed, or actively-worked terrain reduces coherence and rate reliability. Dig-ins and open ROW work can both decorrelate the radar signal and carry settling/depth-of-cover changes this screening layer may not capture.
Processing-method sensitivity: different InSAR time-series methods can yield materially different results on the same stack.
Corridor geometry is an approximate centerline reconstructed from public route references — NOT an operator alignment sheet. Per-segment values are sampled along that approximate line; exact-alignment attribution requires the operator’s stationing.

Data Currency & Forward Availability

Baseline window: 2025-08-02 → 2025-12-30 (15 Sentinel-1 acquisitions, 12-day nominal revisit). Sentinel-1 acquisition over this corridor continues, so the record extends forward with every pass, and each pass re-checks alert thresholds. Operator strain-gauge or other in-situ strain data would be a natural cross-validation input for any pilot that needs to connect surface motion to pipe response.

Corridor map (this build): the figure above renders the screened centerline over a labeled basemap with scale bar and north orientation, each ≈2-mi segment colored by its measured OLS LOS rate — the same values tabulated above. Geometry remains an approximate public corridor centerline reconstructed from public route references, not an operator alignment sheet; exact-alignment attribution requires operator stationing. The source geometry comes from public-domain EIA Energy Atlas crude-oil pipeline data fetched by scripts/fetch-eia-pipelines.py, filtered to the corridor and imported as GeoJSON centerline segments with 500 m buffers.

Generated 2026-06-25 from GroundPulse serving data; per-segment rates are ordinary-least-squares LOS velocities over the 15-epoch series. Not a PHMSA filing. Not engineering certification. Figures are satellite-derived screening estimates over an approximate corridor; field verification and operator stationing required before any integrity action. © Morton Analytics LLC.