[Alex0420W]

Adds distance-to-nearest-scarp criterion using Mishra & Kumar 2022's 704 south-polar lunar scarp catalog (Geophysical Research Letters, CC-BY 4.0) as the primary data source, with the LROC SOC POLAR_SCARP_LOCATIONS shapefile (NASA-hosted, citing Watters 2015 verbatim) bundled alongside as an attribution anchor.

Eight of eight selene-base criteria now contribute to the aggregate score: slope, illumination, thermal, ice, hazard, coupling, LOS-to-Earth, and seismic. Default seismic weight 0.10 (slightly below slope, reflecting that scarp distance is one of multiple safety factors most relevant on long habitat-occupation timescales); the other seven weights scaled by 0.928 so the simplex still sums to 1.00.

Result on the v1.5 20 m catalog (same per-region tiled HLS filters, same NMS, only the aggregator's input set + weights change):

metric	v1.5 (7)	v1.8 (8)	Δ
selene sites	69	69	0
selene matched within 5 km	56 (81%)	56 (81%)	identical
Wueller matched within 5 km	44 (60%)	46 (63%)	+2, +2.7 pp
median matched-pair distance	1.76 km	1.69 km	-0.07 km

Same 69 sites, same top-3 in every region; 13 sites in HW/MMP/N1/N2 reshuffle within ranks 4-10 because seismic at 0.10 weight lifts the aggregate of slightly-further-from-scarp cells just enough to flip a few interior orderings. Two of those reshuffled cells now pair within 5 km of a Wueller target where their v1.5 predecessor didn't, lifting bidirectional from 44/73 to 46/73.

Scoring shape: logistic 1/(1+exp(-(d-25)/8)) — d=5 km → 0.08, d=25 km → 0.50, d=50 km → 0.96. The transition is centred on Civilini et al. 2023's documented Apollo-era shallow-moonquake-to-scarp clustering distance; logistic shape preferred over hard linear ramp because Civilini's distance estimates are themselves noisy with epicentral uncertainty of order 10 km.

Bundled data:

• src/selene_base/criteria/data/scarps_mishra_kumar_2022/ main_segments.{shp,shx,dbf,prj} (~1.4 MB) — 704 LineString / MultiLineString features, polar-stereo Moon CRS matching selene's analysis CRS exactly. License: CC-BY 4.0 (Mishra & Kumar 2022, doi:10.1029/2022GL098505, Zenodo doi:10.5281/zenodo.6624114).
• src/selene_base/criteria/data/scarps_watters_2015_polar/ polar_scarp_locations.{shp,shx,dbf,prj} (~25 KB) — 48 named-scarp Points in IAU Moon 2000 geographic. NASA public, attribution required. Cites Watters et al. 2015 verbatim as data citation; bundled as attribution anchor + sanity reference, not consumed by the criterion (sparse for selene's polar-grid scope).

Both bundles ship full-text README.md with citation, license, source URL, and rename provenance.

Code:

• src/selene_base/criteria/seismic.py: replaces the v1.4-era linear- ramp compute() with a logistic; adds load_bundled_catalog() pointing at the in-repo Mishra & Kumar primary file plus BUNDLED_MISHRA_KUMAR_2022 / BUNDLED_WATTERS_2015_POLAR path constants for downstream consumers.
• src/selene_base/pipeline/score.py: extends the seismic-score candidate list to fall back to the bundled Mishra & Kumar shapefile when no user-supplied override at data/raw/scarps/ is present, so the criterion fires out of the box without any download step.
• config/weights_default.yaml: rebalanced 8-criterion vector; documents the 0.928 scaling and the 0.03 → 0.10 seismic bump.
• tests/test_seismic.py: rewritten for the logistic + new load_bundled_catalog API; 19 tests including hand-verified logistic values, monotonicity, range, NaN/inf handling, parameter validation, and an end-to-end smoke test that runs the criterion chain on the bundled catalog over a coarse polar-stereo grid.

README + docs:

• New "Seismic criterion (v1.8)" Methodology subsection with the Civilini-derived logistic parameters and the v1.5-vs-v1.8 result comparison.
• v1.8 Roadmap entry with the headline numbers.
• Default-weights paragraph rewritten with the new 8-criterion vector.
• Mapping-to-NASA-FOM table: seismic row updated from "catalog TODO (planned)" to "Mishra & Kumar 2022 + Watters 2015 (consistent; bundled in-repo since v1.8)".
• References: Civilini 2023 entry expanded with the v1.8 logistic- parameter justification; Mishra & Kumar 2022 added as a citation; the LROC SOC bundling location noted.
• docs/v1.5_catalog_report.md §8 honest-limitations: the seismic- criterion limitation is now resolved; v1.8 update note added in place. The v1.5 report's headline numbers (81.2 % / 1.76 km) remain valid as the v1.5 7-criterion baseline; v1.8's 56/69 selene-matched count is identical and the median tightens to 1.69 km.

Tests: 372 passed (+10 over v1.7), 9 skipped (LEND + Robbins, expected). Lint + format clean.

Summary

What this PR does, in 1–3 sentences.

Motivation

Why this change is worth making (link the issue if there is one).

Checklist

• ruff check . passes
• ruff format --check . passes
• pytest passes locally
• New behaviour has a test (or an explicit # noqa reason if not)
• Public-facing changes (CLI flags, criterion math, file layout) are
  reflected in the README or the relevant module docstring
• No real-data fixtures were committed; new tests run in CI without
  downloaded data (or are guarded with pytest.mark.skipif)

Scope notes

• Things this PR explicitly does not address (so reviewers know what
  to expect in follow-ups).