Ghost Murmur: A Long-Range Biomagnetic Sensor?

2 What Forbes Says Ghost Murmur Is

Forbes breaks the claim into three layers, each increasingly speculative.

The base layer: Ghost Murmur is a quantum magnetometer. Not a passive radio receiver, not an imaging system, not an acoustic sensor. It detects the magnetic field a living body emits through muscle contraction, primarily cardiac and respiratory signatures, using superconducting or atomic sensors descended from the same physics that makes medical magnetoencephalography work.

The middle layer: the system is deployed in a way that extends the range of those sensors beyond what academic literature shows. Forbes doesn't claim the physics is new. He claims the engineering is: some combination of airborne array coherence, signal-source separation, quantum-limited readout, and possibly active magnetic interrogation that turns a metre-scale medical device into a hundreds-of-metres field sensor.

The top layer: Ghost Murmur is part of the same classified technology cluster he argues underlies the MH370 orbs. Both depend on manipulating or sensing extremely weak electromagnetic structures. Both involve superconducting components operating at quantum-limited noise floors. Both have declassified precursors in the public patent and paper record. That's the thesis connection.

That top layer is the speculative one. The base layer is mainstream physics with commercial products on the market. The middle layer is where the argument lives, and it's the layer the public record cannot yet confirm or rule out.

3 SQUIDs, Josephson Junctions, and Femtotesla Sensing

A SQUID is a loop of superconductor interrupted by one or two Josephson junctions, thin insulating barriers that let Cooper pairs tunnel between the superconducting sections. Brian Josephson predicted the tunneling effect in 1962 as a 22-year-old PhD student at Cambridge. He received the 1973 Nobel Prize in Physics for it. Two years after the prediction, a team at Ford Research Labs (Jaklevic, Lambe, Silver, and Mercereau) turned the effect into a magnetometer. That device could measure changes in magnetic flux down to a single flux quantum, h/2e, about 2.07 femtowebers.

For reference: the Earth's magnetic field is around 50 microteslas. A neuron firing in your visual cortex produces a field of about 100 femtoteslas at the scalp. A SQUID can measure the second in the presence of the first, because its sensitivity is a linear readout of flux change, not absolute field.

Magnetoencephalography (MEG) exploits this. A clinical MEG helmet wraps the skull in 100 to 300 SQUID channels cooled to 4 kelvin by liquid helium, inside a magnetically shielded room that attenuates ambient fields by 10,000x or more. It resolves neural activity with millisecond timing and millimetre spatial precision. Magnetocardiography (MCG) does the same for the heart. Both are FDA-cleared. Both are in clinical use. Neither is remotely covert: the equipment is the size of a dental chair, the shielding is a room, and the measurement is point-blank.

The last decade has introduced a new sensor class. Optically-pumped magnetometers (OPMs), and their high-sensitivity variant the Spin-Exchange Relaxation-Free (SERF) magnetometer developed by Michael Romalis at Princeton in 2002, use a vapour of alkali atoms (usually rubidium or caesium) whose spins precess in a magnetic field. A laser polarises the spins, and a second laser reads the precession. No cryogenics. Sensitivity at the single-femtotesla floor. Wearable form factor.

That matters because it removes the main engineering constraint that kept SQUIDs inside hospital rooms. A clinical MEG system can now be a helmet the patient wears while walking around. DARPA funded this shift in 2010 through its Quantum-Assisted Sensing and Readout program (QuASAR) and extended it in 2013 through AMBIIENT, targeting atomic magnetometers for whole-body biomagnetic imaging. NIST's John Kitching group demonstrated chip-scale atomic magnetometers at femtotesla sensitivity in the same period.

If you want the physics in plain English, we have a companion 101 guide: Quantum Magnetometry. It covers SQUIDs, Josephson junctions, OPMs, and why the 1/r³ falloff makes the distance question so sharp.

4 What the Public Record Actually Shows

The Navy has been using Magnetic Anomaly Detection (MAD) since World War II. MAD is not quantum sensing in the strict sense; early systems used fluxgate magnetometers, and more recent systems use optically-pumped cells. But it's the closest existing public analogue to "find a thing by the magnetic field it emits or perturbs." A P-3 Orion maritime patrol aircraft flies a MAD boom at about 75 metres above the water and detects submerged steel hulls by the magnetic anomaly they produce in the Earth's ambient field. Detection ranges for quiet modern submarines are classified but discussed in the open literature as hundreds to low thousands of metres, depending on sensor noise floor and search geometry.

That's not human-heartbeat detection. It's several orders of magnitude easier: a submarine displaces tonnes of ferromagnetic material and distorts an ambient 50-microtesla field. A human heart produces a 100-picotesla dipole at skin contact, 1,000 times weaker than the Earth's field and, critically, a dipole source whose strength falls with the cube of distance.

What has been published for biomagnetic sources is stark. Cardiac signals at 1 metre are around 10 to 100 picotesla. At 5 metres, they're 10,000 times weaker, in the tens of femtoteslas, right at the single-sensor noise floor of the best commercial SQUID. At 10 metres, you've crossed the floor. At 100 metres, the signal is below the combined noise of every known sensor type operating in the Earth's field.

DARPA has published solicitations for programs that would extend this. BAA-14-27 (2014), BAA-19-27, and follow-ons describe quantum-enhanced magnetometers, array processing for source separation, and active interrogation concepts. None of the unclassified deliverables demonstrate long-range passive biomagnetic detection. That's the gap the Ghost Murmur claim tries to fill.

5 The 1/r³ Problem

Here's the hard constraint. A magnetic dipole's field strength at distance r is proportional to 1/r³. Double the distance, divide the signal by eight. Tenfold the distance, divide by a thousand. This is geometry, not engineering, and no amount of sensor improvement relaxes it.

What engineering can do is lower the noise floor. A SERF magnetometer at 1 femtotesla/√Hz, averaged for a second on a roughly periodic source, can detect a signal of a few hundred attotesla if the source has enough temporal structure. That gets you from 5 metres to maybe 15 or 20 for a cardiac source. It does not get you to a kilometre.

So the question becomes what else has to be true for long-range detection to work. Three possibilities sit inside the public literature, none of them demonstrated operationally.

Array coherence. A sensor array with N elements gives √N improvement against uncorrelated noise. A thousand-element coherent array buys you about 30x in SNR, which pushes the cardiac range from 5 metres to maybe 15. Not enough. But a million-element array, fabricated as a dense chip-scale OPM lattice of the kind NIST has prototyped, buys you 1,000x, which starts to matter at hundreds of metres. This is not currently deployed technology, but it's not physics-forbidden either.

Quantum correlation. Entanglement-assisted sensing beats the standard quantum limit. Theoretical ceiling: a √N further improvement on top of classical arrays (the Heisenberg limit). In a laboratory setting, this has been demonstrated for specific sensing tasks with small photon numbers. Scaling it to a fielded military system is where the engineering argument has to carry the weight.

Active interrogation. Passive detection at long range is hard. Active, where the sensor emits an interrogating field and measures the response, is easier because the source is under your control. A transmitter at a few microteslas can induce measurable eddy currents in moving tissue. Target no longer needs to emit a usable signal; it only needs to modulate an imposed one. This is the quietest of the three options in the unclassified record.

Any of these paths would produce a Ghost Murmur. All of them are currently at the boundary between physics and classified engineering. That boundary is exactly where Forbes' larger argument lives.

6 What Would Have to Be True

Suppose Ghost Murmur exists as Forbes describes. What else does the world have to look like?

First, a quantum-magnetometer research line inside the intelligence community that outran the academic state of the art by a decade or more. That's not unusual. GPS lived inside black programs for years before the Navstar constellation went operational. The Kennedy-era CORONA imagery program had pixel-resolution capabilities the public didn't know existed until the records were declassified in 1995. Capability outpacing the open literature is the normal classified pattern, not the exception.

Second, a delivery platform. A SERF magnetometer array doesn't function in the ambient Earth field without shielding or active cancellation; putting one on a helicopter or a drone means wrapping the sensor in a gradiometer configuration that rejects uniform fields and reads only local variation. The Navy's MAD systems already do this. Scaling to biomagnetic sensitivity is an engineering problem, not a physics problem.

Third, a fusion layer. The Iran rescue used more than one sensing modality: Gorgon Stare wide-area imagery from drones overhead, Palantir Maven AI to filter hundreds of thousands of pixels down to candidate targets, and then a terminal capability that closed the loop. Ghost Murmur only has to be the last link. Imagery brings you to the grid square; the biomagnetic sensor brings you to the man.

That architecture is consistent with how the US intelligence community actually fields new capabilities. Nobody builds a single magic box. They build layered sensor fusion that combines brittle systems into a robust one. A narrow-beam quantum magnetometer that only works inside a 500-metre cone, slewed by a Gorgon Stare cue, is a completely different engineering problem from "scan the mountain." The literature supports the first. The rescue fits the first.

7 Why It Matters for the MH370 Thesis

Forbes doesn't invoke Ghost Murmur to claim the CIA tracked MH370 with magnetometers. He invokes it to establish a capability baseline.

The standard rebuttal to the orb-and-wormhole framing of MH370 is that it requires technology the US doesn't have. Every time a new classified program gets named, declassified, or hinted at in operational reporting, that rebuttal gets harder to sustain. Ghost Murmur, if it's real, is one more data point in a pattern: quantum sensing, compact fusion magnets, superconducting arrays, exotic plasma containment. All connected at the physics level. All funded from the same part of the budget.

So when someone argues that a plasma-orb magnetic wormhole is too exotic for the US to have built, the counter is that the same physics cluster shows up in every declassified corner of the intelligence community. The orb-tracking capability Forbes talks about is not a separate miracle. It's the sensing side of a technology family whose weapons side is also deployed.

That's the connection. Ghost Murmur is relevant to MH370 not because it detected a 777 but because its existence makes the surrounding physics cluster harder to wave away.

8 Open Questions

Four things would move this analysis forward.

1. Confirm the WSJ attribution. Forbes cites the Wall Street Journal piece on the Iran rescue as the source for the Ghost Murmur code name. A direct citation (article title, date, author, paragraph) lets this be tested. At present, it's a secondary source chain.
2. Map the program lineage. If Ghost Murmur exists, it didn't appear from nothing. It's downstream of DARPA QuASAR, AMBIIENT, and unnamed successor solicitations. A FOIA request on the agency that holds the original contract, or a patent search for biomagnetic tracking assigned to the intelligence community, would corroborate or weaken the lineage.
3. Identify the delivery platform. Airborne, ground-based, or satellite-tier? Each implies a different array geometry and a different SNR budget. The rescue geometry (mountainous terrain, line-of-sight blocked from directly overhead) constrains this more than open sources admit.
4. Quantify the range claim. "Long range" is not a number. A credible Ghost Murmur operates at tens of metres, hundreds of metres, or kilometres, and the physics of each is a different conversation. Forbes has not committed to a number, and the public record cannot supply one.

None of these gaps invalidate the claim. They do set the bar for what would confirm it.

9 Timeline

10 Key Sources