The space-for-time substitution approach provides a valuable empirical assessment to infer temporal effects of disturbance from spatial gradients. Applied to predict the response of different ecosystems under current climate change scenarios, it remains poorly tested in microbial ecology studies, partly due to the trophic complexity of the ecosystems studied. The McMurdo Dry Valleys of Antarctica represent a trophically simple polar desert projected to experience drastic changes in water availability under current climate change predictions. We used this ideal model system to develop and validate a microbial space-for-time sampling approach, using the variation of geochemical profiles that follow alterations in water availability and reflect past and future changes in the system. Our framework measured soil electrical conductivity, pH, and water activity in situ to geochemically define 17 space-for-time transects from the shores of four dynamic and two static Dry Valley lakes. We identified microbial taxa that are consistently responsive to changes in wetness in the soils and reliably associated with long-term dry or wet edaphic conditions. Comparisons between transects defined at static and dynamic lakes highlighted the capacity for geochemically defined space-for-time gradients to identify lasting deterministic impacts of historical changes in water presence on the structure and diversity of extant microbial communities. We highlight the potential for geochemically defined space-for-time transects to resolve legacy impacts of environmental change when used in conjunction with static and dynamic scenarios, and to inform future environmental scenarios through changes in the microbial community structure, composition and diversity.