For the first time since the 1970s, concrete plans are in motion to bring humans back to the Moon through the Artemis programme. The dawning era of space exploration aims to set the groundwork for long-term off-Earth settlement, with sights on bringing the first humans to Mars. In order to provide a safe means of inhabiting extraterrestrial (ET) environments, a self-supporting habitat shield concept has been developed by the Specialist Modelling Group at Foster + Partners, which is based on the use of mechanically interlocking masonry manufactured from in-situ resources. Significant progress must still be made in the structural design and validation of such systems, however, before they can be safely implemented. Due to the novelty of the proposed structure, existing research does not provide numerical evidence of its stability under self-weight, or under seismic loading. Research in this direction is valuable, as it explores alternatives to popular design concepts based on monolithic additive manufacturing and addresses research gaps related to the analysis of mechanically interlocking structures. The aim of this project is to provide novel insight into the stability of self-supporting dry-stone vaults under the influence of microgravity and ground motions. This is achieved through a three-part analysis approach, which starts with an investigation into the fundamental structural behaviour of dry-stone, Nubian-type vaults based on a series of parametric studies. Key geometric parameters of the vaults are varied, and the resulting distinct element models are subjected to quasi-static pushover-type analysis in 3DEC. The parametric studies indicate what configuration of the vault geometry can maximise seismic capacity. This optimal vault geometry is modelled in Rhino using geometric constraint solving, which is used to produce a model which can be assembled from mechanically interlocking components. Through pushover-type testing and dynamic time history analysis, the performance of the resulting vault model is assessed with respect to possible moonquake loading. The final part of the design process accounts for uncertainties in material properties by conducting sensitivity studies, which indicate what degree of variation in performance is possible. After completing the three mains sets of analysis needed to design a moonquake-resistant vault, several comparative analyses are carried out to provide context to the structural performance of the Nubian-type vault. The first of these additional investigations models an equivalent monolithic vault geometry in DIANA and conducts finite element analysis to determine its capacity for lateral acceleration. The second comparative study determines how covering the vault in loose regolith may influence its structural performance. The third study investigates how the variation of gravity affects the stability and seismic performance of the vault. Several key findings are obtained in this project, leading to an understanding of the structural behaviour of Nubian-type vaults and how they may be a possible solution to shielding demands in ET contexts. The final vault model was shown to remain stable under a uniform lateral acceleration equivalent to the PGA for a possible moonquake with a 475 year return period. Dynamic analysis conducted using an artificial ground motion record obtained for the same return period produced less favourable results, resulting in partial collapse of the outermost six courses of the vault. Computational constraints necessitated the implementation of an undamped analysis, however, so these results are considered to be conservative. Further analysis is needed, but these results suggest that with minor adjustments, the vault may be able to safely resist moonquake loading. The comparative monolithic vault analysis demonstrates that an additively manufactured vault may provide slightly better resistance to static lateral acceleration than a component-based vault. When considering the structural redundancy and energy dissipation necessary for safe seismic design, however, the discrete vault is expected to perform better. Results show that covering the vault in loose regolith is expected to reduce its overall stiffness. Conclusions about changes in stability cannot be made with confidence due to modelling issues at the interface between the regolith and the structure. By studying the influence of gravitational variation on the vault, results indicate that it may perform favourably on the Earth or on Mars, though the possibility of material damage must be studied further.