When the Signal Lies: GPS Interference and the Next Phase of Maritime Domain Awareness

Posted on | by Robbin Laird
10/15/2026

By Robbin Laird

Britain’s Maritime and Coastguard Agency has just told shipowners something that would have sounded alarmist a decade ago: satellite navigation can no longer be taken for granted in some of the world’s busiest shipping regions.

In a new Marine Guidance Note, MGN 719, the MCA warns that interference across the Baltic, the Black Sea, the Middle East Gulf and parts of East Asia means that satellite-derived positioning, navigation and timing information can no longer be relied upon everywhere. As Sam Chambers reported in Splash247 on October 13, the agency is urging operators to identify which of their systems depend on satellite data, to establish procedures for detecting interference, and to drill crews regularly in radar fixing, visual bearings, celestial navigation and dead reckoning.

The list of exposed equipment is striking. It is not just the electronic chart. The MCA identifies ECDIS, AIS, autopilots, dynamic positioning systems, radar, gyrocompasses and GMDSS communications equipment as potentially vulnerable. Ashore, vessel traffic services, pilotage systems, port management platforms and automated logistics systems are exposed as well. Citing earlier UK government research, the MCA estimates that a seven-day GNSS outage could cost the British maritime economy £1.5 billion, £1.3 billion of it in port operations alone.

The Splash article anchors the scale of the problem in data from Windward, the maritime AI company:

“Maritime intelligence firm Windward’s latest figures show an average of 5,131 GPS jamming incidents a day during the third quarter, with more than 124,000 individual vessels affected.”
— Splash247, citing Windward’s Q3 2026 data

That single sentence carries the argument of this piece. Five thousand incidents a day is not a battlefield anomaly. It is a condition of the operating environment. And 124,000 vessels is not a subset of suspicious ships; it is a large share of the world’s merchant fleet. According to Windward’s Q3 reporting as summarized by SAFETY4SEA, the daily average was broadly unchanged from Q2, while the number of distinct vessels affected rose 11.7 percent to 124,235. On September 1, during US strikes on Iran, Windward counted 1,699 jamming incidents in the Middle East Gulf and 6,838 globally, the highest daily figure of the quarter.

The MCA guidance is written as a navigation safety notice. But it is also, whether intended or not, a maritime domain awareness notice. And it confirms, in very practical terms, the argument I made in September about why the old model of MDA no longer works.

The Fakeable Signal, Now at Scale

In “The Transformation of Maritime Domain Awareness: Why the Old Model No Longer Works,” I argued that three forces broke the Cold War model of maritime surveillance: the scale of global shipping outgrew national sensor capacity; identity and location became both technically knowable and technically fakeable; and commercial capability displaced the state monopoly on maritime intelligence.

The second of those forces is where the GPS story sits. AIS gave the world its first global layer of maritime transparency.

But a cooperative signal is, by definition, a signal that can be switched off, falsified or spoofed. I described awareness and evasion as locked in a continuous technical race, a competitive dynamic that must be managed rather than solved.

What the MCA guidance and the Windward numbers show is that this race has entered a new phase.

The evasion problem I described in September was largely about individual actors: a shadow-fleet tanker switching off its transponder near a ship-to-ship transfer zone, or a vessel broadcasting a false identity to launder sanctioned cargo. That remains a central problem. But GPS jamming and spoofing operate at a different level. They do not corrupt one ship’s report. They corrupt the positioning layer that every ship in a region relies on, honest and dishonest alike.

AIS reports a position that the ship derives from GNSS. When GNSS is jammed or spoofed, AIS faithfully broadcasts the wrong answer. The Splash report notes that four tankers in the Black Sea apparently broadcast positions placing them in Lima, Peru, more than 12,000 kilometers away, and that separate monitoring identified 1,702 impossible position jumps involving 379 vessels in the Gulf of Oman in a single week in September.

That is the core shift. In the old framing, the question was whether a given vessel was telling the truth. In the new environment, the question is also whether the data environment itself is telling the truth.

When the Record Itself Is Corrupted

Windward’s own analysis of the Strait of Hormuz makes the point with real precision. Examining August 2026, Windward found that of 642 vessel transits logged through the strait, 393, or 61.2 percent, were false transits attributable to GPS jamming rather than real crossings.

Yet when it measured actual traffic through an independent pipeline, vessel activity averaged 1,635 vessels per hour during jammed periods and 1,655 per hour during quiet periods, a difference of just 1.2 percent.

Windward’s conclusion is worth quoting directly:

“The vessels aren’t fake, but the transits are.”
— Windward, “GPS Jamming and False Transits in the Strait of Hormuz”

And further: “The operational risk here isn’t that ships are being kept out of the strait, they aren’t, it’s that the record of what’s moving through it can no longer be taken at face value.”

This is the MDA problem in its sharpest form.

In one of the world’s most strategically important chokepoints, during a period of active conflict, a raw count of transits would have overstated traffic by a wide margin. Any analyst, insurer, sanctions authority or naval staff relying on that raw count would have been working from a picture in which most of the “events” never happened.

Windward reported a similar distortion elsewhere: since the start of Operation Epic Fury on February 28, it identified 3.35 million false ship-to-ship meetings worldwide, where injected positioning data made vessels appear to converge when no encounter took place. For an MDA system tuned to flag ship-to-ship transfers as a sanctions-evasion indicator, that is not background noise. It is a flood of false cues landing in the investigation queue.

In September I wrote that “going dark” should be treated as a cue for further investigation, not as proof of illicit conduct. The GPS problem extends that logic. A position jump, an apparent rendezvous or a logged transit is now also a cue to be verified, not a fact to be acted upon. The behavioral analysis that MDA depends on is only as good as the integrity of the positional data underneath it.

Two Sides of the Same Problem

The MCA guidance and the MDA challenge are two sides of one problem.

On the ship’s bridge, the answer is resilience: know which systems depend on GNSS, detect interference, and be able to navigate without it. The MCA’s prescription is, in effect, a return to seamanship. Radar fixing, visual bearings, celestial navigation and dead reckoning are skills that a generation of automation allowed fleets to let atrophy.

Ashore, in the operations center, the answer is verification. The analyst cannot go back to a sextant. The equivalent of seamanship for the watch floor is the ability to know where interference is active, to separate false signals from real activity, and to cross-check AIS against sources that do not depend on the ship’s own GNSS receiver: synthetic aperture radar, RF geolocation, electro-optical imagery, and historical patterns of life.

That is precisely the fusion discipline I argued has become the center of modern MDA.

What This Means for Governments

The implications I drew in September for workforce, acquisition and public-private integration all sharpen in light of this.

First, the workforce. In September I argued that the analyst of the future looks more like a financial-crimes investigator than a traditional watch officer. To that skill set must now be added a working understanding of electromagnetic interference: what jamming and spoofing look like in the data, where the persistent sources are, and how they distort the picture. Windward’s analysis notes that interference in the Black Sea is concentrated and persistent, consistent with fixed sources, while Hormuz is diffuse, short-lived and mobile. Those are different problems requiring different analytic responses, and analysts need to be trained to recognize the difference.

Second, acquisition. The UK is investing in a terrestrial backup through its enhanced Loran program, and that is the right call. But the timeline is instructive. As Splash notes, initial timing capability is not expected until 2028, with full positioning and navigation services targeted for 2030. That is a hardware answer on a hardware timeline. The interference problem is here now. In the interim, and probably well beyond, the practical defense of the maritime picture rests on data fusion and analytic tradecraft that can be fielded and updated continuously. This is the same lesson I drew in September: the marginal dollar spent on fusion and verification is likely to buy more actionable awareness, sooner, than the marginal dollar spent on additional platforms.

Third, public-private integration. The number that leads the Splash story did not come from a navy or a coast guard. It came from a commercial maritime intelligence company. A government regulator issuing guidance, and a trade publication explaining why it matters, both turned to commercial data to establish the scale of the threat. That is the structural inversion I described in September: governments are increasingly consumers of commercially generated maritime intelligence rather than its sole producers. The question is no longer whether to rely on such data, but how to integrate it in ways that are auditable, shareable with allies and partners, and trusted enough to support operational decisions.

Trust Is Now the Center of Gravity

This leads to the deepest point. If the positioning layer can be corrupted at scale, then the value of maritime intelligence rests on whether its users can trust it.

Trust in this context is not a marketing claim. It means being able to show how a conclusion was reached: which sources were used, which were discounted because they fell inside an active interference zone, what confidence attaches to a given position or behavior, and where human judgment entered the process. An allied navy, a coast guard or a sanctions authority being asked to act on a maritime cue needs to know not only what the system says, but why it says it, and whether that answer survives contact with a jammed environment.

That is why I see data integrity as the next frontier of maritime domain awareness. In the Cold War, the challenge was to find the platform. In the AIS era, the challenge became understanding the behavior. In the era of routine GNSS interference, the challenge is also verifying the record itself before anyone draws a conclusion from it.

Conclusion

The MCA’s MGN 719 is a navigation safety notice.

But read alongside Windward’s figures, it is also a strategic marker. When a G7 maritime regulator tells crews to practise celestial navigation because the satellite signal may lie to them, the assumption that the maritime picture is reliable by default has ended.

For seafarers, the response is resilience and seamanship. For those responsible for maritime security, the response is verification at scale: fusing independent sources, recognizing interference as part of the operational environment, and building intelligence that is auditable and shareable among allies.

Five thousand jamming incidents a day and more than 124,000 vessels affected in a single quarter is not a problem that more patrol hulls will solve. It is a problem of data integrity, analytic tradecraft and trusted public-private architecture.

That is where the MDA competition is now being fought, and it is where governments and alliances need to reorganize if they want the picture they act upon to match the sea as it actually is.

Sources