About NavLunar
The map comes before the ship. When Artemis astronauts report a position on the lunar surface, they won't read out latitude and longitude. They'll use the Lunar Grid Reference System (LGRS), an MGRS-style grid published by USGS and NASA in 2024 and 2025. NavLunar is a civilian reference instrument for that grid: the first public interactive viewer of the LGRS, with a permanent, linkable URL for every 25-km cell and every named feature, so anyone can point at a place on the Moon the way they share a map pin on Earth.
NavLunar is built entirely on open NASA, USGS, and IAU data. It is not affiliated with NASA or USGS. It sells nothing and claims nothing: this site does not sell lunar land, does not imply ownership of any part of the Moon, and never will. No one owns the Moon; under the Outer Space Treaty of 1967, no one can. This is a map, not a deed office.
Data sources
Full provenance (exact URLs, retrieval dates, verification steps) is recorded in the repository's DATA_SOURCES.md.
Methodology
Reference frame & projections. All coordinates are selenographic (planetocentric, east-positive) on the 1,737,400 m lunar sphere in the JPL DE421 mean-Earth frame, the same frame as LRO data products. Between 80°S and 80°N the grid uses the Lunar Transverse Mercator projection (45 zones of 8° per hemisphere, k₀ = 0.999, false easting 250 km); poleward of ±80° it uses Lunar Polar Stereographic (k₀ = 0.994, false origin 500 km), per USGS TM 11-E1.
Verification. Our TypeScript grid engine is validated by 1,300+ unit tests, including ~1,700 ground-truth conversions generated by executing the USGS reference implementation, all documented worked examples from the USGS/NASA publications, and round-trip property tests across zones, hemispheres, and the polar boundary. Where the specification is ambiguous at exact cell boundaries, we pin a deterministic convention (a boundary belongs to its north/east cell) and document every deviation in the repository.
Live terminator. The day/night line is computed from the Sun's direction in the Moon's body-fixed frame (IAU WGCCRE 2015 rotation model via astronomy-engine), refreshed every minute, and validated against JPL Horizons (subsolar point within 0.02 degrees at six instants from 1969 to 2099), the recorded Sun elevation at the Apollo 11 landing, known full and new moon dates and the ~12.2°/day terminator advance. Supported from 1950 to 2100.
Polar illumination. The pole views can shade every 400 m cell inside 85 degrees by the share of 2026 the Sun is up there. The method follows Mazarico and colleagues (Icarus, 2011): from each cell, the height of the horizon is measured in 64 directions across the NASA LOLA terrain model out to 200 km; then, for every hour of the year, the Sun's position from the same ephemeris that draws the day/night line is compared with that horizon, counting how much of the solar disc shows above it. A cell lit 80 percent of the year is a crater rim that sees the Sun most of the time; a cell at 0 percent is a floor the Sun never reaches. The panel also gives the longest run of hours with no Sun at all. Surface level, one year, so the figures are a few points under the published twenty-year averages.
Routing
How a route is found. Directions are computed on NASA's LRO LOLA elevation models: about 474 m per sample around the six Apollo landing sites, 160 m inside 80 degrees of either pole (400 m when a long polar route needs a lighter search), and about 1.9 km everywhere else, the global model of 16 samples per degree. The terrain route is a shortest-time search across those samples that avoids slopes steeper than the chosen vehicle's limit, and it is never slower than the direct line when the direct line is passable. Speeds and slope limits come from the documents cited in each route's panel. Every route page says which of these models it used.
What it cannot see. A slope is the average over one sample, so a steeper stretch, a boulder field, a small crater or a rille narrower than the sample is invisible to the search, and the steep flags are optimistic. Travel times are continuous driving or walking with no stops, rest, recharging or communication passes; stretches steeper than the limit are flagged and not timed. Daylight along a route assumes a smooth horizon with no relief, which is exactly wrong inside polar craters, where the illumination and shadowed region layers are the better guide. Routes longer than 1,500 km (300 km at the poles) are drawn as the great circle. The named routes run between the mapped centres of IAU features, not the exact stations the crews stopped at.
Horizon view
From any point, the horizon view traces a ray every half degree of azimuth across the same LOLA terrain the routes use, out to 200 km, and keeps the sample that stands highest above the local horizontal, with the curvature of the Moon included; the eye is 2 m above the ground. 200 km is enough because the surface falls away by the square of the distance over twice the radius: even a peak 7 km above the observer sinks below the horizontal plane beyond about 156 km. Each sample comes from the finest model covering it: 160 m within 30 km of a point near a pole, then 400 m, about 474 m around the Apollo sites, about 1.9 km elsewhere, so nearby foreground relief is smoothed wherever the coarser models answer. A peak is named only when an IAU feature's mapped footprint lies over the point that forms it; the Sun comes from the same ephemeris as the day and night line and is compared with the skyline in its direction, which is not the same as the Sun being above zero degrees.
Known limitations (v1). The map is equirectangular, so polar cells appear stretched; a dedicated polar stereographic view is planned. Grid math uses the reference sphere; terrain-induced distortion factors (USGS raster products) are not yet applied. ACC 6-character codes are 10-m references within a named 25-km cell, per the LGRS overview. Feature pages show gazetteer center points, not outlines.
Questions or corrections? Every number on this site traces to a source. If you find one that doesn't, it's a bug.