Aims: Sand burial and precipitation play important roles in vegetation of inland deserts and desertified areas, and both are under strong influence of climate and land-use change. Sand deposition and precipitation both vary greatly in space and time, and different plant species, with diverse adaptations, occupy different niches along spatial gradients in the combination of sand and water availability dynamics. We hypothesized that species specificity in spatial and temporal niche occupation along such gradients is a mechanism for their co-existence and, thereby, a driver and stabilizer of biodiversity in dry, sandy areas. Location: Ordos Plateau, China. Methods: We conducted a 2-yr field experiment with factorial treatments of precipitation (control, slight enhancement, strong enhancement) and sand addition (control, medium addition, large addition). Plant cover of the six dominant species was followed over different seasons, as we expected different species to benefit from different treatments in different seasons and years with different weather regimes. Results: Sand burial alone significantly affected plant cover of all six dominant species, while precipitation enhancement had no significant effect. Effects of sand burial on plant cover changed overall during the two hydrologically contrasting years. Our key finding was that there were multiple significant two- and three-way interactions between species, sand burial and precipitation enhancement on plant cover, while there were also multiple two- and three-way interactions involving species, sand burial or precipitation treatment with year and/or season. Conclusion: Our results supported our hypothesis, i.e. the co-existence of species in a semi-arid sand dune can be explained from the different niches they occupy in terms of seasonal and year-to-year variation in precipitation in combination with sand deposition regime. The interactions of these drivers on plant cover with experimental enhancement of precipitation, mimicking a realistic scenario for this region, suggest that shifts in species composition are to be expected this century. More generally, our findings advance our understanding of what drives species co-existence and thereby biodiversity, now and in the future.