Leonid dust trails in the 2030s: stream simulations, parent element sets and encounter forecasts

Result files behind the Leonid forecasts and hindcasts of Abe (2026), Leonid dust trails in the 2030s. This is a data release: the output of the integrations — the grain states, the parent element sets they were anchored to, and the encounter forecasts evaluated from them. The model source code is not part of this release. Stream simulations Each file is self-describing: its params header records the grain count, ejection model, epoch, force model and the ephemeris-anchoring flag it was generated with. LEO-theory.json — epoch 2026, 18 trails (1433–1998), 180,000 grains; the 2030s forecasts LEO-epoch1998.json … LEO-epoch2002.json — 100,000 grains each; one per historical encounter Each historical year needs its own simulation because the grains are archived as frozen two-body elements, which carry no planetary perturbations: propagated far from their epoch the node distances shift by enough to misattribute the dominant trail. Encounter forecasts <dataset>-forecast.json holds the evaluated encounters — ZHR profile, per-trail contributions, nodal geometry, the Earth's path and the confidence class — for 2030–2035 and for 1998–2002 respectively. Parent element sets parents.json holds the 55P/Tempel-Tuttle elements the integrations start from, retrieved from the JPL Small-Body Database at full precision, including the full 8×8 covariance matrix used for the ensemble. Model REBOUND (IAS15) with REBOUNDx radiation forces; AU / day / GM (G=1), heliocentric ecliptic J2000 Ejection: Crifo & Rodionov (1997) gas-drag terminal velocity, sizes drawn from dN ∝ a−u, β = 5.74×10⁻⁴/(ρa), β ≤ 3.0×10⁻³ Ejection sites anchored to the JPL Horizons ephemeris of 55P for each apparition rather than to an internal backward integration Empirical density factor for old trails: 0.11 MANIFEST.json lists every file with its SHA-256 checksum and the provenance fields read back out of its header. Contact Shinsuke Abe — [e-mail] / [e-mail]