How GPS Jammers Actually Get Found
A question came up on an aerospace forum this week: how would US authorities actually find a terrestrial GPS jammer? It is a good question, because the public answer is less like a movie and more like an engineering funnel. Detection is now cheap and almost anyone can do it. Localization is still hard, and the last mile is still done by people driving around with antennas.
That split matters to anyone building or buying spectrum monitoring. Most of the effort and most of the vendor pitches sit at the detection end. Most of the operational value sits at the other end.
Detection is crowdsourced now
Aircraft already report how good their position is. ADS-B messages carry integrity and accuracy fields (NIC and NACp) alongside position, and when a GPS receiver is jammed, those numbers drop. Public networks such as ADS-B Exchange and the OpenSky Network collect these broadcasts worldwide.
GPSJam is the best-known example of turning that into a map. It aggregates 24 hours of aircraft accuracy reports into hexagons and colors a cell red when more than 10% of aircraft in it report low accuracy. The site’s own FAQ is careful to say that red does not prove jamming; the data shows degraded navigation, not its cause.
Researchers have refined the method. A 2024 paper in The Aeronautical Journal compares the error implied by NACp against the satellite geometry from the GPS almanac, so a bad constellation day is not mistaken for interference. That is the right instinct: before calling something an attack, rule out the boring explanation.
So the first stage works. You can know, within hours and for free, that something is wrong over a region.
Localization is an estimation problem
Knowing that aircraft lost GPS over an area is not the same as knowing where the transmitter is. Aircraft are moving sensors, unevenly spread, at different altitudes, and they only report a degraded flag, not signal strength or bearing. The problem becomes fitting a source location and power to a pattern of affected and unaffected aircraft.
The best public case study is Denver. In January 2022, pilots departing Denver International Airport reported losing GPS, and the disruption lasted more than a day, according to GPS World and a Stanford summary of the event. A Stanford group later used recorded ADS-B data from the event to test a non-linear least squares method, published in NAVIGATION in 2025. It located the source to within 0.1 degrees of latitude and longitude, and produced error bounds and probability heatmaps rather than a single point.
That is a strong result, and it is worth being precise about what it means. A tenth of a degree of latitude is about 11 kilometers. That is a search area, not an address. It tells a field team which part of a metro area to drive, which is valuable, but it does not hand anyone a jammer.
The method also depends on traffic. A related 2022 presentation on the Denver event at a UN GNSS workshop notes that missing data, poorly sampled regions, and aircraft that do not use GNSS positioning complicate the estimate. Where there are few flights, there is little to fit.
The last mile is still a person with an antenna
Every closed case in the public record ends the same way. Near Newark, the FAA reported interference with the airport’s Ground-Based Augmentation System, and FCC agents with direction-finding equipment traced it to a pickup truck. The driver was using a jammer to defeat his employer’s vehicle tracker, and the FCC proposed a $31,875 penalty, as Inside GNSS reported. In a later Newark case, agents traced a signal to a van on the New Jersey Turnpike, which ended in a 2019 consent decree with a $5,000 penalty.
The legal basis is not ambiguous. The FCC treats operating or selling a jammer as a violation of Sections 302(b) and 333 of the Communications Act, and its enforcement advisory warns of forfeitures, equipment seizure, and criminal sanctions.
The unsolved case is just as instructive. In October 2022, wide-area interference around Dallas-Fort Worth disrupted operations and closed a runway. Ten months later, Virginia Tech researchers writing in Inside GNSS reported that the source was still unknown, despite ADS-B and ground reference-station analysis. Good detection and a reasonable estimate did not turn into a found transmitter.
What this means for programs
If you are scoping a spectrum awareness or GNSS resilience effort, three points follow from the public record. These are our view, not settled doctrine.
Fund the whole funnel, not just the map. A dashboard that shows red hexagons is the cheapest part of the system. The expensive, decisive parts are the estimator that narrows the search and the field capability that closes it.
Ask for uncertainty, not a dot. The Stanford work is useful precisely because it reports error bounds. A localization tool that outputs one confident pin with no error ellipse will send a team to the wrong neighborhood with confidence.
Add your own sensors where traffic is thin. Aircraft are free sensors, but only where they fly. Fixed ground receivers that log signal strength, or a few SDR nodes with direction finding, fill exactly the gaps where ADS-B-based methods fail. This is also where modeling helps: an RF propagation model of the area tells you where a given jammer power would and would not show up, which constrains the search before anyone gets in a truck.
The forum question has a plain answer. Authorities find jammers by noticing degraded navigation, estimating a search area from whatever data they have, and then sending people with direction finders. The technology keeps improving the first two steps. The third is still where cases get closed.