Species extinctions alter ecosystem services, and the magnitude of this impact is likely to change across environmental gradients. In Australia, soil-disturbing mammals that are now considered ecologically extinct are thought to be important ecosystem engineers. Previous studies have demonstrated microsite-level impacts of reintroduced soil-disturbing mammals on soil functions, but effects are yet to be tested across larger scales. Further, it is unclear how impacts vary across environmental gradients and if the restoration potential of reintroductions changes with climate. We examined the effects of soil-disturbing mammal reintroductions across a large rainfall gradient in Australia to test the hypothesis that ecosystem engineering effects on soil function depend on climate. We compared soil labile carbon, available nitrogen and the activity of four enzymes associated with nutrient cycling in three microsite types with and without soil-disturbing mammals in five sites along a large rainfall gradient (166-870 mm). Soil enzyme activity was greatest in the presence of soil-disturbing mammals and increased with rainfall, but soil available carbon and nitrogen varied across the gradient and among microsites. Microsite effects were often stronger than any effects of soil-disturbing mammals, with soil beneath vegetated patches (shrubs and trees) having greater enzyme activity, carbon and nitrogen than bare soils. However, soil-disturbing mammals homogenised nutrient distributions across microsites. The impacts of soil-disturbing mammals on soil function previously detected at micro-scales was detected at a landscape-scale. However, the overall effects of soil-disturbing mammals on soil functions varied with productivity (rainfall). The context of soil-disturbing mammal reintroductions is thus likely to be critical in determining their effectiveness in restoring soil function.