Earth's rotational variations by electromagnetic coupling due to core surface flow on a timescale of ∼1 yr for geomagnetic jerk

Abstract

SUMMARY Sudden changes of the core surface flow associated with geomagnetic jerks may cause changes in Earth’s length of day (LOD) and polar motion on a timescale of ∼1 yr. Here, I examined the LOD and polar motion on this timescale due to simple single-harmonic toroidal core surface flows by assuming electromagnetic (EM) core–mantle coupling and rigid rotation of the outer core. The predicted time derivative of LOD is sensitive to both the flows and electrical conductivity structure of the lowermost mantle. Most zonal and some non-zonal flows, with a discontinuity of its time derivative, produce a similar discontinuity for LOD derivative. In the conductivity structure with a conductance, for example, 10 8 S, the predicted LOD derivative is very sensitive to the thickness of conducting layer. The LOD derivative for a thin (∼200 m) conducting layer faithfully reflects the time-dependent behaviour of velocity perturbation at the core surface. With increasing thickness of conducting layer, however, the magnitude decreases, and phase lag between the LOD derivative and core surface flow increases. The velocity perturbations with a short duration ( 1 yr). That is, the EM coupling due to sudden changes of the flows, with the velocity of ∼10 −5 ms −1 and duration longer than ∼1 yr, would explain the observed LOD derivative with ∼0.1 ms yr −1 for a conductivity model with a conductance of 10 8 S or greater, particularly for a thin (∼200 m) conducting layer. The sensitivities of polar motion to conductivity structure are almost the same as those for the LOD. The magnitude of polar motion is dominantly contributed by flow components with degree one and order one, whereas some other components are significant. Even if we adopt 10 −4 ms −1 for core surface flows and a conductivity model with ∼10 8 S or greater, the magnitude for polar motion is ∼10 −3 arcsec at most. That is, it would be difficult to quantitatively explain the observed changes in phase of the Chandler wobble. However, the polar motion is sensitive to the time-dependent behaviour of velocity perturbation as well as the duration. The time-dependent behaviour, characterized by a step function type perturbation with duration of ∼1 yr followed by a relaxation with duration of ∼1 yr, may affect changes in phase of the Chandler wobble.