Under construction
Micro Ultra-Cold Traps
Here on Earth, we have seasons because our planet’s rotational axis is tilted by about 23.5 degrees relative to our orbit around the Sun. Over the course of a year, the latitude of the “subsolar point” (the location where the Sun is directly overhead) shifts north and south. This also means that the path that the Sun takes across the sky each day changes with the seasons, tracing a higher path during the summer and appearing closer to the horizon in the winter.


On the Moon, the situation is quite different. Coincidentally, its axial tilt is almost exactly cancelled out by the tilt of its orbit around Earth, so its rotational axis is only about 1.5 degrees off from being perfectly perpendicular with the plane of our shared orbit around the Sun.

This arrangement has two important consequences. First, the Moon’s seasons are much smaller than Earth’s, as the subsolar point doesn’t wander far from the equator. Second, because the distance above the horizon that the Sun reaches at noon is equal to the difference in latitude between your position on the surface and the latitude of the subsolar point, the Sun is always close to the horizon in the lunar polar regions.


Because the Sun never appears more than a few degrees above the horizon for an observer near the lunar poles, it’s easy for the Sun to be blocked by local topographic features that rise above the ideal zero-elevation horizon. This most frequently occurs within large craters, whose high walls can obscure a significant portion of the sky. In these areas, the Sun never rises above the local topography, leaving them in eternal shadow. As such, they are appropriately given the name “permanently-shadowed regions,” or “PSRs.”


PSRs are of scientific interest because, in the absence of direct sunlight, they are much colder than the surrounding illuminated terrain. It’s possible that they may have been this cold for potentially billions of years, meaning that they could act as “cold traps” for molecules like water that would otherwise rapidly sublimate away, allowing potentially significant deposits of ice to accumulate within them.

The data that have been used to map PSRs come primarily from two instruments: the Diviner Lunar Radiometer Experiment (measures temperatures) and the Lunar Orbiter Laser Altimeter (LOLA; measures terrain elevation). The biggest problem with these datasets is their resolution, which averages around 240 metres per pixel. This is fine when you’re looking at a global scale, but 240 metres is still quite large, so we’re losing a lot of smaller details.
Hayne et al. (2020) examined the existence of “micro cold traps,” or cold traps that are smaller than the spatial scale of the orbital datasets. These micro cold traps could be anything from small craters to shadows cast by the small-scale roughness of the lunar surface. Their models found that 10–20% of the Moon’s total cold-trapping area is made up of cold traps at scales between 1 kilometre and 10 centimetres.
We wanted to extend this work to look within the larger PSRs themselves. At first glace, this might seem strange – it makes sense that small features could cast shadows on sunlit terrain, but the PSRs are, by definition, already shadowed. It’s not like you can have doubly-shadowed terrain, right?
As it turns out, you can! Although the PSRs aren’t exposed to direct sunlight, they aren’t completely dark. The biggest contributor to the illumination of PSRs is scattered sunlight. This is sunlight that has reflected off of directly-illuminated terrain into the PSRs. Although much fainter than direct sunlight, scattered sunlight is bright enough that, with a sensitive enough camera, it can be used to image the interior of PSRs from orbit. The ShadowCam instrument onboard the Korea Pathfinder Lunar Orbiter has been doing just that since 2023, assisting in the hunt for lunar water by trying to directly image any ice that might be exposed on the surface.

The darkest (and thus coldest) parts of a PSR are thus the areas that are protected from both direct and scattered sunlight. O’Brien & Byrne (2022) used higher-resolution LOLA topography maps (30 m/px at latitudes >75°, 5 m/px at latitudes >87.5°) to map these doubly-shadowed regions, finding that about 0.04% of the PSR area at both poles is doubly-shadowed.