In early 2007 something happened deep within Earth that satellites detected but no one noticed at the time. Orbiting hundreds of kilometers above the surface, the twin spacecraft of the Gravity Recovery and Climate Experiment recorded a gravity signal unlike anything else in its fifteen years of operation. The signal was vast, powerful, and appeared almost overnight. It reached its maximum strength in January of that year, stretched across seven thousand kilometers of the Atlantic boundary off the coast of Africa, and then faded away within two years. Only a later reanalysis revealed it, and only then did geophysicists begin to understand what it meant. What they found pointed not to floods, storms, or tides, but to a mineral transformation nearly three thousand kilometers below the surface, at the contact point between the solid mantle and the molten outer core.
The GRACE satellites were built for an entirely different purpose. They were designed to track the movement of mass on Earth’s surface, especially water and ice. By flying in tandem, one following the other, they constantly measured the tiny changes in distance between them as they encountered stronger or weaker gravity fields below. A mountain range would pull slightly harder, groundwater depletion would reduce the pull, ice sheets growing or shrinking would tug at the satellites with subtle but measurable force. The system was sensitive enough to reveal the draining of aquifers in India, the retreat of Greenland’s ice, and the rise of oceans from melting glaciers. But in January 2007, over the eastern Atlantic, it recorded something far more unusual.
The anomaly was not subtle. It was the most intense signal recorded in that region during the entire GRACE mission, well above the statistical noise that usually obscures small variations. At its peak it measured more than one micro-Eötvös in gravity gradient intensity, a value that placed it in the upper one percent of all anomalies in the dataset. Its sheer scale was unlike normal hydrological or atmospheric effects. Instead of spanning a few hundred or a thousand kilometers, it extended across an area thousands of kilometers wide. It was aligned north to south, cutting across the continent–ocean boundary between Africa and the Atlantic. And it lasted not decades but just a few years, rising in 2006, peaking in 2007, and dissipating by 2008.
When researchers first encountered the anomaly in the reprocessed data, they tested the obvious explanations. Could this be water? Floods, rainfall, or seasonal hydrology often create mass shifts large enough to be detected by satellites. Five independent hydrological models were consulted. None produced anything remotely resembling the 2007 anomaly. Even when hydrology was reconstructed directly from GRACE data under the assumption that the signal had to come from water, the scale remained wrong. The characteristic length of the modeled hydrological signals was about two to four thousand kilometers, while the detected anomaly extended across seven thousand. The shape was wrong as well. Water models placed the strongest effects inland, while the actual anomaly straddled land and ocean. Even an unrealistic scenario where entire continents were covered in water could not reproduce the signal.
Next the team looked to the oceans. Could a change in sea level, driven by currents or thermal expansion, have created the anomaly? Again, models and observations failed to match. Ocean simulations showed nothing unusual in the eastern Atlantic during 2006 to 2008. Altimetric data from satellites measuring sea surface height confirmed the region was quiet. In situ salinity and temperature profiles were also normal. There was no evidence of an El Niño or La Niña during the period, no major oscillations that could have shifted mass on the necessary scale. Oceanic eddies, even at their largest, extend no more than a few thousand kilometers and last weeks rather than years. Nothing at the surface could explain what GRACE had recorded.
The anomaly also stood out because of its timing. It coincided with one of the strongest geomagnetic jerks on record, an abrupt change in Earth’s magnetic field detected in 2007. A geomagnetic jerk occurs when the secular variation of the field suddenly accelerates or decelerates. These events have puzzled scientists for decades. They happen irregularly, often about a decade apart, and suggest rapid changes in the flow of the outer core. The 2007 jerk was one of the most intense observed in modern times. The fact that the gravity anomaly and the geomagnetic disturbance happened simultaneously, in the same geographic sector, pointed to a common cause.
Attention shifted to the boundary between the mantle and the core, a region known as the D” layer. This zone lies at a depth of about 2,900 kilometers, where pressures exceed 130 gigapascals and temperatures rise above 2,500 degrees Celsius. It is one of the least understood parts of the planet. Seismic studies have revealed that it is heterogeneous, with regions of unusually slow seismic velocities, patches of ultra-low velocity zones, and hints of layered complexity. One of the most important processes suspected in this region is the phase transition of bridgmanite, the dominant mineral of the lower mantle. Under extreme pressure and temperature, bridgmanite can transform into a denser form called post-perovskite.
This transformation is abrupt. Laboratory experiments show that once conditions are met, the mineral structure rearranges rapidly. The density difference between the two phases is significant, on the order of a hundred kilograms per cubic meter. If a region hundreds or thousands of kilometers wide undergoes such a transformation, the change in mass distribution is large enough to affect the gravity field detectable at the surface. The transition is also highly sensitive to temperature. Hotter material transforms deeper, cooler material shallower. In rising plumes or sinking heterogeneities, the depth of the boundary can shift quickly. This means that within the convective flow of the mantle, the phase transition can occur on short timescales as material moves vertically.
Modeling by the research team suggested that two cold regions, each about a thousand kilometers wide, undergoing perovskite to post-perovskite transition at the base of the African Large Low Shear Velocity Province, could account for the 2007 anomaly. The transformation would depress the mineral boundary by roughly 180 meters, generate a surface gravity change of about 0.4 millimeters in the geoid, and deform the core–mantle boundary by twelve centimeters. These values matched the observed signal. The rapid onset and decay of the anomaly could be explained by the growth of the transformed regions over about eighteen months followed by relaxation in the surrounding D” layer.
The African Large Low Shear Velocity Province, or LLSVP, is a massive thermochemical anomaly deep beneath Africa and the Atlantic. It is one of two such structures on Earth, the other lying beneath the Pacific. These regions are thought to be composed of chemically distinct material that is hotter and less dense than surrounding mantle. They extend thousands of kilometers laterally and may reach several hundred kilometers in height. They play a role in directing mantle plumes to the surface, influencing hotspot volcanism and continental dynamics. Within such an environment, steep thermal gradients and vertical motions are common. The conditions are ripe for the perovskite to post-perovskite transition to occur repeatedly.
The 2007 anomaly revealed that these processes are not only real but fast. Instead of millions of years, the transformation unfolded in less than two years. That is nearly instantaneous on geological timescales. It means Earth’s deep interior is capable of sudden, powerful shifts that can be measured at the surface. The coincidence with the geomagnetic jerk suggests that these shifts may be responsible for triggering disturbances in the outer core, where the magnetic field is generated. A deformation of the boundary by only a few centimeters can alter the flow of molten iron, changing the pattern of convection and producing observable magnetic effects.
Such a link between mantle processes and geomagnetic variations has long been suspected but rarely observed. The 2007 anomaly provides the first strong evidence. A mass redistribution at the base of the mantle corresponded in time and space with a magnetic shock at the surface. The implication is that Earth’s magnetic field is sensitive not just to the core itself but to the dynamics of the overlying mantle. Events hidden deep below our feet may leave signatures in the very shield that protects the planet from solar radiation.
The magnitude of the event also raises questions about frequency. Was 2007 a rare occurrence, or do similar transformations happen more often than we realize? The GRACE mission spanned fifteen years, and its successor GRACE-FO continues to operate. Researchers plan to comb through the full record in search of other anomalies. Already, a smaller blip in 2004 has been identified, though it lacked the strength of the 2007 event. If more are found, it would mean that Earth undergoes periodic pulses of deep mantle transformation, each capable of altering the core–mantle boundary and disturbing the magnetic field.
The existence of such pulses would redefine how scientists view the stability of Earth’s interior. Instead of a steady, slow convective system, the mantle would be punctuated by sudden transitions that send ripples through the planet. These would be invisible to us at the surface yet powerful enough to affect gravity and magnetism. They could even play a role in long-term magnetic field reversals or excursions, events where the field weakens or flips polarity.
For now, the 2007 anomaly remains the strongest evidence that the mantle can shift in this way. It was detected only because satellites built for an entirely different purpose happened to be watching. GRACE was intended to monitor water resources, not the deep interior of the planet. Yet by stripping away hydrological and atmospheric signals, scientists uncovered a hidden record of Earth’s most inaccessible layer. It shows that satellites can act as eyes not just on the surface but into the very core of the planet, revealing changes no other method can reach.
The gravity anomaly itself is gone. It rose, peaked, and faded, leaving only traces in orbital data. But those traces tell a story that is both unsettling and profound. In 2007, the base of the mantle shifted. Minerals changed their structure. Mass was redistributed. The boundary with the core flexed. The outer core responded. The magnetic field jerked. The Earth convulsed silently, invisibly, in a way we could not feel but satellites could sense.
What else lies hidden in the records of past missions? How many more signals from the deep mantle have passed unnoticed? The data remain, waiting for careful analysis. If events like the 2007 anomaly occur more often, then Earth is a far more restless planet than we imagine. Its stability is an illusion created by our short lifespans. Beneath us, vast mineral structures can transform in the span of months, altering gravity and magnetism alike.
The revelation that satellites captured such a moment changes how we look at Earth. Not because it adds a new narrative, but because it exposes a reality that was always there. The deep mantle can shift on human timescales. It can move with enough force to deform the boundary with the core. It can stir the flows that generate our magnetic shield. And it can do so silently, leaving no sign on the surface except for a ghostly signal in the gravity field.
In 2007, our planet shifted from within. The record of that shift lay hidden until researchers brought it to light. What they found was a glimpse into a world almost three thousand kilometers down, where minerals change their form, where the mantle touches the core, and where unseen processes shape the fate of the magnetic field above. The anomaly lasted only a short time, but its legacy is a permanent reminder that the deep Earth is not inert. It is alive, restless, and capable of sudden change.
Source:
Gaugne Gouranton, C., Panet, I., et al. (2025). GRACE Detection of Transient Mass Redistributions During a Mineral Phase Transition in the Deep Mantle. Geophysical Research Letters, 52(17). https://doi.org/10.1029/2025GL108689
