Landing Site Analysis (Touchdown)

Touchdown is our custom toolkit that directly reduces the risk of landing on the Moon and operating hardware in the lunar environment.

It involves terrain reconstruction and enhancement: we extend existing DEMs (LOLA, NAC, and/or SfS) down to lander- and rover-relevant scales (as low as cm-resolution) using band-limited spectral synthesis calibrated to published roughness statistics. We preserve the measured terrain and generate ensembles of plausible smaller-scale roughness, craters, and boulders consistent with the available observations and terrain statistics.

One 5 m LOLA terrain view branches into three 10 cm terrain-and-rock renders labeled lowest, medium and highest hazard, with scenario hazard fractions. — Native 5 m LOLA over 100 × 100 m generates nine 10 cm conditional terrain members. Shown: the lowest, median, and highest hazard fractions among those nine, using a 9.4 m footprint slope / 0.35 m rock-height screen. All four views share a camera and 6× height exaggeration. The added fine terrain is simulated; hazard fractions are scenario-screening values.

We produce tens to hundreds of ensemble scenarios that are used to calculate probabilistic landing-safety products, and photorealistic renders for the complete landing sequence.

What problem does this solve?

Landing on the Moon is difficult, as proven by recent missions that have tipped over. It's especially difficult at the poles with significantly rougher terrain than mare, and exceptionally difficult lighting conditions.

Two labeled oblique terrain images at identical scale and height exaggeration: equatorial mare from merged LOLA and Kaguya data under high sunlight, and south-polar LOLA terrain under grazing sunlight. — Matching 40 × 40 km measured terrain. Left: Apollo 11-area mare, SLDEM2015 merged LOLA–Kaguya (~59 m posting), Sun 60°. Right: polar Haworth region, LOLA LDEM_83S_10MPP_ADJ, Sun 2°. Shared 62.5 m display sampling, orthographic camera, and 2× height exaggeration. Representative angles, not an epoch-matched comparison.

For polar sites >87.5°, the best topography products are generally from LOLA at 5 m/px resolution, but upwards of 90% of these maps are interpolated:

Two square 2 km maps of the same terrain: hillshade, and sparse native 5 m populated count cells in cyan, each with a 500 m scalebar. — Identical 2 × 2 km native 5 m LOLA windows: elevation left, positive count cells right. Each cyan mark is one native 5 × 5 m map cell, with no enlarged markers or dilation — these are populated raster cells, not individually located laser spots. In this window 18,184 of 160,000 cells (11.4%) have returns.

Some rare sites have NAC DTMs at ~1–2 m/px, or SfS, which uses image illumination plus topographic control rather than a stereo pair. But SfS products commonly have significant smears and other artifacts:

An oblique gray shape-from-shading render with labels and arrows marking a smooth patch and a texture discontinuity. — A 2 km USGS Haworth SfS render, true height scale and 25° presentation light. Arrows indicate suspect smoothing and texture changes; they are visual interpretation, not a verified pixel-level fill or error mask.

At 5 m resolution, boulders and craters up to ~15 m in diameter cannot be spatially resolved. The sub-pixel reconstructions from Touchdown produce physically based conditional scenarios for the expected boulder and small crater populations.

Adapting it to your work

Send us your candidate landing sites and lander or rover requirements. We'll tailor the terrain scenarios, hazard assessments, and renders to your vehicle and mission.

Our published research

Menges, JD, and K. M. Cannon (2026). Natural Landing Pads on the Moon and Mars. Acta Astronautica, 242, 219–235.

Cannon, K. M., and D. T. Britt (2020). Accessibility Dataset for Large Permanent Cold Traps at the Lunar Poles. Earth and Space Science, 7, e2020EA001291.