A good Moon base site needs more than sunlight, a communications link and somewhere flat to land. It also needs ground that will remain usable. The Moon has no Earth-like system of moving tectonic plates, but it is not seismically silent: instruments left by Apollo astronauts recorded vibrations from its interior, its changing surface temperature and incoming impacts. For infrastructure planners, the useful question is not whether the Moon shakes. It is what kind of shaking can occur at a particular site, how often it might occur, and what that means for hardware expected to stay there for years. A rare event can be relatively unimportant for a brief visit yet deserve attention when locating a habitat, a power installation or a route into a crater. Hero: Lunar Reconnaissance Orbiter mosaic of Taurus-Littrow valley and the Lee-Lincoln thrust fault. The white cross marks the approximate Apollo 17 landing site. This is an equatorial-region case study, not a south-pole hazard map. Credit: NASA/ASU/Smithsonian. Four ways to make the Moon shake NASA's moonquake overview distinguishes several sources. Deep moonquakes occur hundreds of miles beneath the surface and are associated with tidal stresses: Earth's gravity deforms the Moon as the Earth-Moon system moves. These are different from the shallower events linked to faults in the lunar crust. Shallow tectonic moonquakes are associated with the Moon's cooling and contraction. NASA describes events at depths of roughly 20 to 30 kilometers. As the interior loses heat, the brittle crust cannot simply shrink smoothly. It breaks and moves along faults. The resulting surface features include lobate scarps, small cliff-like steps where one block of crust has been pushed over another. Two other sources originate close to the surface. Meteoroid impacts inject energy into the ground. Thermal moonquakes result from expansion and contraction as surface materials experience the transition between cold lunar night and hot daylight. A sensor recording ground motion therefore needs more context than a single spike in its data: not every signal is evidence of a nearby tectonic fault slipping. The distinction that matters “Moonquake” describes an observation, not one universal engineering load. Source depth, distance, local terrain and the character of the shaking all matter. A list of event magnitudes is not yet a design specification for a lunar building. What Apollo measured, and what it could not Between 1969 and 1977, the Apollo seismic network recorded thousands of lunar vibrations, according to NASA's 2025 account of moonquake research . Those measurements established that the Moon is active. They did not provide a dense, modern monitoring network surrounding every candidate settlement site. That geographic limitation is important. Knowing that an event occurred somewhere in a broad region is not the same as identifying the exact fault responsible or measuring the motion at a future habitat foundation. Orbital images can reveal scarps and displaced terrain, but a photograph alone cannot tell engineers the date of the next rupture. Duration also complicates comparisons with Earth. NASA's general overview describes some shallow events lasting more than ten minutes; its 2025 research article notes that lunar seismic signals can persist for hours. That should not be read as hours of constant peak shaking. A persistent signal and its strongest damaging motion are different quantities. Designers need the time history and frequency content of ground motion, not a sensational duration headline. Reported magnitudes require similar care. NASA summaries describe the strongest shallow events as roughly magnitude 5, with the overview citing values up to 5.5. Those numbers do not establish how hard a specific site would move. Distance from the source and local conditions separate the size of a quake from its consequences for equipment. Why the south pole needs a closer look A NASA-funded study described in January 2024 connected young thrust faults and possible shallow moonquakes with the south-polar region. Researchers modeled a fault scarp near de Gerlache and found that its formation could have been associated with a quake comparable to the strongest shallow event recorded by Apollo. The work is evidence of a plausible hazard, not a prediction that a particular landing zone will shake during a mission. The same study modeled slope stability and found that some south-polar terrain could experience regolith landslides even under light seismic shaking. That included areas within some permanently shadowed regions, which are attractive because they may contain accessible volatiles. The infrastructure issue is therefore broader than whether a habitat can resist vibration. A slope failure could obstruct a traverse, disturb equipment or change access to a resource site. This adds another layer to lunar water-ice prospecting . Finding a promising deposit does not establish a safe access route or a stable equipment location. Resource mapping and geotechnical assessment answer different questions, and a project needs both before treating a shadowed crater as a dependable supply source. The 2024 source uses the Artemis mission labels and candidate-region terminology current when it was written. This explainer uses its geological findings, not those historical mission assignments as a current flight schedule. Seismic planning remains relevant regardless of which mission first visits a particular region. How fallen boulders extend the record A short instrumental record is not the only evidence available. In the Taurus-Littrow valley, researchers studied boulder falls and a landslide associated with the Lee-Lincoln fault. Apollo 17 supplied an unusual advantage: astronauts brought back samples from boulders in the area. Exposure to cosmic radiation changes rock chemistry over time. A surface newly exposed after a boulder breaks away starts accumulating that exposure record. Combining sample measurements with the boulders' locations, sizes and tracks helps researchers reconstruct when movement occurred and estimate the shaking needed to cause it. Harrison Schmitt samples a boulder in Taurus-Littrow in 1972, photographed by Eugene Cernan. Samples and orbital images help reconstruct ancient movement. Credit: NASA/JSC/ASU. NASA's 2025 report summarizes an estimate of a magnitude 3 event along Lee-Lincoln about every 5.6 million years. That is a result for one fault and a particular reconstruction. It is not the interval between all moonquakes, not a south-pole forecast, and not a countdown to the next event. This is why “rare” and “potentially active” can both be true. A fault may produce very few events over human timescales while retaining the ability to move. The researchers' method broadens the evidence beyond Apollo's years of monitoring, but applying it elsewhere requires local observations and assumptions that must be tested rather than copied. Turn a hazard into an infrastructure decision Hazard is the possibility of damaging ground motion or slope failure. Risk also depends on what is exposed and how vulnerable it is. An unoccupied instrument on level terrain, a tall landed vehicle and a crew habitat below a steep slope do not have identical consequences if the ground moves. A practical site assessment would combine maps of faults, slopes, boulders and surface materials with local seismic measurements. It would also define the planned occupancy, equipment lifetime and acceptable loss of service. These are engineering implications of the research, not a published universal lunar building code or a claim that a particular base has passed certification. Before selecting the footprint: examine nearby scarps and unstable slopes, not just the flatness of the landing ellipse. Before freezing the design: specify credible ground-motion cases and assess foundations, equipment restraints, connections and tall structures against them. Before commi