Report 178 · Defense Tech
Every few weeks a screenshot of a public GPS-interference map goes around, a region painted red, captioned as proof of jamming. The maps are genuinely useful. But they are not maps of jammers. They are maps of what airplanes said about their own navigation accuracy, aggregated over a day, at altitude. Read that way, a red hex tells you something real and leaves out three things people assume it shows.
I spent my service as a Counter-IED and Electronic Warfare Officer, and the first habit that job builds is asking what a sensor actually measured before you believe what a display says it measured. The public GPS-interference maps are a good test of that habit, because they are honest about their method and almost nobody reads it.
The best known is GPSJAM, built by John Wiseman. Its FAQ is short, candid, and worth reading in full. What follows is that FAQ, read alongside a peer-reviewed Stanford paper on the same raw data and the federal rule that defines what "good" accuracy means.
There is no network of jamming detectors behind these maps. The data is ADS-B, the position broadcast most aircraft transmit for air traffic surveillance. In the FAQ's words, "Many aircraft broadcast digital radio messages (ADS-B) that contain information about their GPS accuracy." GPSJAM pulls those reports from two volunteer receiver networks, airplanes.live and ADS-B Exchange.
Each ADS-B position comes with quality fields the aircraft computes about itself. The Stanford GPS Lab paper by Zixi Liu, Sherman Lo, Juan Blanch, Yu-Hsuan Chen and Todd Walter, published in NAVIGATION in September 2025, describes the two that matter. NACp is the estimated position uncertainty. NIC is "an integrity containment radius within which the current horizontal position is guaranteed to lie with 99.999% probability." The federal ADS-B rule, 14 CFR 91.227, sets the bar: the aircraft's NIC "must be less than 0.2 nautical miles," which the paper translates as NIC of 7 or better under normal conditions. Its Denver example shows the contrast plainly: on a normal day "all aircraft consistently report NIC values greater than or equal to seven," and during interference affected aircraft drop as low as NIC 0.
So the input to the map is a self-report. An aircraft's navigation system decided its own fix had gotten worse and said so.
GPSJAM aggregates those reports over 24 hours into hexagons. Green means "more than 98% of all aircraft who flew through that area reported good navigation accuracy." Yellow means "between 2% and 10% of aircraft reported low navigation accuracy." Red means "more than 10%."
The FAQ also publishes the formula, which I respect a great deal:
percent_bad_aircraft = 100 * (num_bad_aircraft - 1) / (num_good_aircraft + num_bad_aircraft)
The minus one is a deliberate noise filter. Wiseman explains that it limits false positives where few aircraft flew, and says it "does bias the map to not show potential interference in hexes where there's very little data." Worked through (my arithmetic): a hex with 20 aircraft and one bad report scores 0%. Two bad reports score 5%, three score 10%, and it takes four, 15%, to turn the hex red. That is a sensible design. It also means a quiet hex can hide a real event that touched one or two aircraft.
This is the point that matters most to anyone who has hunted emitters. A jammer on the ground and an aircraft at cruise altitude can see each other over enormous distances. The Stanford group's earlier conference paper on localizing jammers from ADS-B gives the radio line-of-sight rule of thumb: horizontal range in kilometers is about 4.12 times the sum of the square roots of the two heights in meters.
Plug in numbers (again my arithmetic, assuming a jammer at ground level and no terrain in the way): an aircraft at 10,000 meters, roughly airliner cruise, has line of sight to that jammer out to about 412 kilometers. At 3,000 meters it is still about 226 kilometers. Two people standing on flat ground, by the same rule, see each other's antennas out to only about 12 kilometers.
Power and antenna patterns decide whether line of sight turns into actual degradation, and the NAVIGATION paper models exactly that. But the geometry explains two things the FAQ says outright. First, the map can light up a region hundreds of kilometers from the emitter. Second, people inside a red zone often notice nothing: aircraft "have line-of-sight to many more potential sources of GPS interference than you do on the ground." A red hex over a city is a statement about the sky above it, not about the phone in your pocket.
Finding the actual transmitter is a different and harder problem. The NAVIGATION paper takes the same NIC values and solves for a jammer's location and power, and on a real 2022 interference event at Denver International Airport it reports localization "within a 4 km radius of the actual jammer location, with a 95% confidence interval equivalent to a 10 km-radius circle." That is real work, with error bounds. A colored hex map is not that work, and it does not claim to be. I wrote about how emitter hunting is done by time-difference-of-arrival in Report 025.
The FAQ is blunt about this. Asked whether red and yellow always mean jamming, it answers "No," and says the data "doesn't tell me what's causing the low accuracy." It even notes the ADS-B fields don't say which navigation system is in use. Most aircraft use GPS, but the source could be another satellite constellation or an inertial system.
Its best example is close to home for American readers: interference that shows up in the southwestern United States, especially Texas, which the FAQ says "isn't conflict-related and it's not jamming." It is military trainer aircraft "performing aggressive aerial maneuvers that temporarily block their own GPS antennas." Same red, entirely different cause. The FAQ also lists deliberate jammer testing outside conflict zones among the causes.
And there is a failure mode the map can't see by construction. A successfully spoofed receiver may report a confident, wrong position. The FAA's own pilot guide warns that a corrupted receiver "may appear to be functioning normally while providing false information," which I covered in Report 073. A map built on self-reported accuracy can only show the degradation a receiver knows about.
This one is subtle. The NAVIGATION paper cites the ADS-B standard, RTCA DO-260B: if no new GNSS position arrives within two seconds, the transmitter "will clear all but the altitude and status subfields of the airborne position message." In other words, an aircraft that loses its fix completely stops broadcasting a position at all. The paper notes that interference produces "a loss of airborne position messages and/or an increase in the claimed integrity bounds," and that earlier methods, such as one EUROCONTROL developed, worked from trajectory gaps for this reason.
An aircraft with no reported position is hard to place into a hexagon. GPSJAM's FAQ does not say how, or whether, it counts such aircraft, and I could not confirm it. The safe reading is that the map is strongest at showing degraded aircraft and weakest exactly where degradation is total. Add the FAQ's other limits: blank areas mean no aircraft or no receivers, which covers active war zones, and the archive has documented gaps.
I did not inspect GPSJAM's code or raw data. Everything about its method comes from its public FAQ, including the thresholds and formula. Its definition of "low accuracy" is not stated in the FAQ; the background image description mentions NACp, but I have not confirmed which field and cutoff the hexes use.
The line-of-sight numbers are idealized. They come from a rule of thumb in the Stanford conference paper and assume a ground-level jammer and no terrain. Real coverage depends on power, antennas, terrain and the receiver.
I read DO-260B only as quoted in the NAVIGATION paper, not the standard itself.
Public GPS-jamming maps are built from aircraft grading their own navigation, binned into hexes over a day, with a filter that leans toward showing less. That makes them a fair indicator that something degraded navigation in a region of sky. It does not make red a jammer's location, a diagnosis of jamming, or a statement about the ground. When you see the screenshot, ask the operator's question: what did this sensor actually measure?
Scope note: this report explains how public interference maps are built and how to read them. It contains no jamming or evasion instructions and is not operational guidance for pilots or mariners.
U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.