Pacific magnetic records point to the whole solid Earth tilting just before the Hawaiian bend formed

The evidence comes from rock that records where magnetic north was. Magnetic stripes in slowly cooled magmas and lavas point toward the magnetic pole as they form. Once solid, they keep that direction while the plate carries them away ^[4]. In recent geologic history the magnetic pole has usually sat over the spin axis. A dated stripe's change in direction therefore tells researchers how the Pacific plate has moved relative to the spin axis since the stripe formed ^[4]. Gordon said the pair combined "a new paleomagnetic pole" with "a data set that had taken us decades to collect and create." ^[2] The results appear in JGR Solid Earth ^[3].

"Which means we had two data sets showing different movements," Woodworth said, comparing the magnetic record with the models of what drives plates ^[6]. A plate cannot travel in two directions at once ^[10]. So one of the reference frames had to be wrong. Woodworth said that "ultimately we had to reconsider the assumption that the spin axis was stationary relative to the solid Earth and its hot spots." ^[7]

Their answer is that between about 56 and 48 million years ago "the entire solid Earth reoriented relative to the spin axis, which gave us the jump found in the paleomagnetic record," Woodworth said ^[8]. On that reading, the southward jump was the globe tilting under a fixed axis, and the plate's own course need not have reversed ^[8][10]. Gordon said he thinks the shift was large enough to count as true polar wander ^[14]. "These data indicate that contrary to some current thought, substantial true polar wander could have happened," he said, placing it 48 to 56 million years ago ^[14].

The current thought he means goes back decades. "When I was a student, many geophysicists assumed that the solid Earth does not change its location relative to the spin axis," Gordon said. He added that "some geologists think that Earth's orientation relative to the spin axis hasn't significantly changed in the past 100 million years." ^[13]

The timing fits. The proposed tilt spans about 8 million years and ends roughly a million years before the bend's estimated age ^[12][11]. The phys.org account does not give the size of the reorientation in degrees, or say how much of the bend, if any, the tilt explains. Gordon's own phrasing is "could have happened." ^[14] The result also rests on one team's reconstruction. Untangling it took, Woodworth said, "a fair amount of cross-referencing with other geological records." ^[9]