Case study by SPH Engineering · Methane Detection · Oil and Gas
Drone methane detection at abandoned oil and gas wells
Researchers at the University of Alaska Anchorage built a drone methane survey method to show which abandoned wells are actively leaking. They tested it with controlled gas releases, then at three real wells in Houston, Alaska. The paper was peer-reviewed and published in Sensors on 2 April 2026.
The challenge
The United States has about 3.9 million documented abandoned and orphaned oil and gas wells, and fewer than 1,200 have ever had their emissions measured. Unplugged wells can leak methane to the surface. Agencies spending plugging funds need to know which wells are leaking to set priorities.
The usual way to find wells, aerial magnetometry, detects steel casings, not emissions. It cannot say whether a capped well leaks, and it fails where casings are missing or mixed with other metal debris.
The solution
The team flew a Laser Falcon TDLAS sensor on a DJI Matrice 300 RTK, with SkyHub and UgCS:
- The Laser Falcon measures methane along its laser path at 1.653 µm, with a range of 0 to 50,000 ppm·m and ±10% accuracy.
- SkyHub logged every reading and tagged it with the drone's GNSS position at the moment of capture. At 2 m/s, a timing error of half a second would put a reading about a metre out, which matters on a grid this fine.
- Before each visit, the team measured obstacle heights with a rangefinder and used them to set safe flight heights in UgCS.
- Line spacing and speed were set from the sensor's field of view and data rate, giving about one reading per metre.
- With the payload fitted, flight time fell from 55 to about 35 minutes. UgCS let the team pause for a battery swap and resume from the same waypoint, with no gaps in the grid.
- The payload blocked the downward obstacle sensors, so take-off and landing were flown manually.
Controlled release tests
At the Mat-Su campus, the team released a 5% methane calibration gas at 0.8 litres per minute and flew a grid of about 160 m x 160 m with 2 m line spacing at 2 m/s. One baseline flight at 17 m had no release; three leak flights flew at 32 m, 20 m and 25 m. The baseline skewness was 1.41, while the leak flights averaged 9.61. A threshold of 1.5, combined with a spatial clustering test (Moran's I), separated leak from no-leak with no false results.
Field results at three wells
- Ten flights across three visits: two before remediation and one during plugging work, at 25 m over all three wells and 20 m over Well 1.
- Skewness passed the threshold in 9 of 10 tests and clustering was significant in 8 of 10. The exceptions came from the first survey near a busy highway, where gusts from passing traffic disturbed the plume.
- Readings pointed to Wells 2 and 3, which matched the Alaska Oil and Gas Conservation Commission's knowledge that these were unplugged and leaking more.
- Tree cover changed the laser path and the readings. After the site was cleared during remediation, the results reached statistical significance, and a leak was still detected.
What the method does not do
The survey finds zones of unusually high methane. It does not identify which well is the source and does not give an emission rate, so ground checks are still needed. The researchers note that the same approach can apply to landfills, gas pipeline leak detection and repair (LDAR) and permafrost methane surveys.
Where this applies in India
This survey method suits sites where leaks must be ranked before ground crews are sent:
- old and suspended wells in onshore oil and gas fields;
- well pads and gathering lines in mature fields;
- landfills and city gas pipelines needing first-pass screening;
- projects that combine methane surveys with drone magnetometer surveys to locate casings.
SPH Engineering's methane detection training is also available through HobiTech.
The equipment, available from HobiTech
| Methane detector | Laser Falcon was used in this study. See our current methane detection systems. |
| Onboard computer | SkyHub by SPH Engineering |
| Flight planning | UgCS - Drone Flight Planning Software: see UgCS PRO, UgCS EXPERT and UgCS ENTERPRISE |
| Drone | DJI Matrice 300 RTK was used in this study. Ask us for current compatible drones. |
Case study by SPH Engineering, published in two parts in March 2026 and May 2026, based on Thomas and Wang (2026), Sensors, 26(7), 2205 (first posted as a preprint on Preprints.org). This page summarises both parts in our own words; the research was carried out by the University of Alaska Anchorage, not by HobiTech. Read the original case study (part 1) on SPH Engineering ↗ and the peer-reviewed follow-up (part 2) ↗
Need to find leaking wells or pipelines?
Tell us the sites, the terrain and the obstacles. We will recommend the methane detector, drone, SkyHub and UgCS setup, quoted in rupees with a GST invoice.