A transform (strike-slip) fault sits between two plates sliding past each other at a steady far-field rate, but friction locks the fault itself down to a "locking depth" D. Below D the crust creeps freely; above it, the two sides are stuck together and simply bend elastically while the plates keep moving underneath — this is Reid's elastic rebound theory (1906, formulated after the San Francisco earthquake).
Interseismic surface velocity (Savage & Burford, 1973):
v(x) = (v_plate / π) · atan(x / D)
Slip deficit (locked, unreleased offset):
δ(t) = v_plate · t since the last rupture
Coulomb failure: rupture when δ(t) reaches the fault's
strength threshold δ_crit → instantaneous coseismic slip,
δ resets to 0, permanent offset steps forward by δ_crit.
The bending fence line crossing the fault is exactly the classic geodetic signature of a locked fault — real GPS stations near the San Andreas or North Anatolian faults show this same smooth atan velocity gradient during the "interseismic" wait, and a sudden step when the fault finally ruptures. The white arrows are that velocity field, scaled for visibility.
- Plate rate — the steady far-field relative motion driving the whole cycle.
- Locking depth D — how deep the fault stays stuck; a shallower D concentrates bending closer to the trace.
- Fault strength — how much slip deficit (≈ future coseismic slip) the fault can store before it fails — bigger stores mean rarer but larger earthquakes.
- Trigger rupture now — force an early, smaller earthquake, releasing whatever deficit has built up so far.
Offsets in the scene are visually exaggerated (~500×) — real interseismic bending is centimetres, and real coseismic slip is typically 1–8 m, not the metres-wide kink you see rendered here.