Each tephra clast leaves the eruption column at height H and falls under gravity while a crosswind advects it downwind. Its fall speed is its terminal settling velocity vt, found by balancing gravity against drag:
Weight = Drag
(π/6) d³ (ρp − ρa) g = Cd · (π/4) d² · ½ ρa v²
→ v_t = sqrt( 4 ρp g d / (3 Cd ρa) )
Cd(Re) = 24/Re · (1 + 0.15 Re^0.687) [intermediate drag, solved iteratively]
Re = ρa v d / μ
Small clasts (fine ash, tens of microns) sit in the low-Reynolds Stokes regime and fall slowly; large lapilli sit in the high-Reynolds Newtonian regime and fall fast — d appears squared in the fast-falling limit and to the ½ power in the slow-falling limit, so grain size changes fall speed far more than it changes drag shape. Air density ρa is modelled with an exponential atmosphere, ρa(y) = ρ0 e^(−y/8500 m), so the same clast falls faster once it drops into denser low-altitude air.
Downwind landing distance from the vent is approximately
x_land ≈ U · H / v_t
— fast-falling coarse clasts land near the vent, slow-falling fine ash rides the wind for tens of kilometres before settling. Accumulating landing points into distance bins reproduces the two hallmark signatures of a real tephra fall deposit: thinning (less mass per bin with distance) and fining (mean grain size dropping with distance) — exactly the isopach/isopleth pattern volcanologists map in the field to reconstruct plume height and wind conditions after an eruption. The simulation runs at accelerated time (real ash can take hours to settle) — because both fall and drift scale with the same time factor, the physically correct ratio x_land = U·H/v_t is preserved even though the clock is sped up.