[Alex0420W]

Buffer-fits-inside-mosaic check against PGDA's LDEM_83S_10MPP_ADJ (±216.4 km from south pole, polar-stereographic Moon CRS) for the 9 USGS polygons + 100 km horizon-march buffer (the LOS-to-Earth criterion's physical horizon-search radius, not a tunable knob): only 3 of 9 polygons fit cleanly — HW, SP, G2. CB and N2 over-run the west/east edges by 50-75 km; N1, MM, MA, MP over-run the north edge by 4-60 km. A 3-region release was considered and rejected: G2 contributes 0/2 matched-within-5km against Wueller in v1.5, so the comparison collapses to HW+SP — too thin for a release, and truncating the buffer would bias LOS visibility asymmetrically on the truncated side.

PGDA's 5 m product (LDEM_87S_5MPP) covers only ~87°S to the pole and fits only 1 of 9 polygons (G2) with the buffer; even narrower coverage. NAC stereo DEMs (1-2 m, PGDA product 78) are per-ROI and would need a parallel integration path distinct from selene's polygon-tiled flow.

v1.6 deferred indefinitely pending wider public mosaic coverage or NAC stereo DEM mosaic generation through ASP. The 10 m work is genuinely blocked by data availability, not bandwidth or scheduling.

Roadmap entry added between v1.5.1 and v1.7. 'Where this goes next' updated with the corrected per-resolution coverage analysis (20 m clean, 10 m 3-of-9, 5 m 1-of-9, NAC per-ROI). No methodology, data, or validation changes; v1.5/v1.7/v1.8 headline numbers unchanged. Internal coverage notes at
docs/v1.6_internal_notes.md (uncommitted).

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).