Geomagnetic observations from satellites have highlighted interannual variations in the rate of change of the magnetic field originating from Earth's core. Downward continued to the core surface, these variations primarily show up in the equatorial belt. First, we recall the main characteristics of these patterns, addressing their spatio-temporal resolution, as seen from field models. We then review the several dynamical frameworks proposed so far to understand and model these observations, which populate the frequency spectrum on time scales close to the Alfvén time A ≈ 2 yr, much shorter than the vortex turnover time U ≈ 150 yr in Earth's core. Magnetic-Archimedes-Coriolis (MAC) waves in a stratified layer below the core surface constitute a first possibility in the case of a sub-adiabatic heat flux at the top of the core. Their period may reach the interannual range for a layer thickness less than ≈ 30 km, for a buoyancy frequency of the order of the Earth's rotation rate. An alternative has been proposed in a context where the Coriolis force dominates the momentum balance, rendering transient motions almost invariant along the rotation axis (quasi-geostrophy, QG). Torsional Alfvén waves, consisting of axisymmetric QG motions, operate at periods similar to the Alfvén time, but are not sufficient to explain the interannual field changes, which require non-axisymmetric motions. QG Alfvén waves (involving the Coriolis and magnetic forces) constitute another possibility, with inertia playing an important role. They have been detected in the latest generation of geodynamo simulations, propagating in a ubiquitous manner at a speed slightly less than the Alfvén velocity. They are localized in longitude and as a result their description requires high azimuthal wavenumber. But the branch of QG waves with large extent in azimuth is also worth considering, as it reaches interannual periods as their radial wavenumber is increased. The excitation of such high-frequency dynamics is discussed with respect to the temporal spectrum of the core field, which presents a slope ∼ f −4 for periods approximately between A and U . We finally summarize the main geophysical implications of the existence of this interannual dynamics on core and lower mantle structure, properties, and dynamics.