Ice sheets are an active component of Earth’s climate system, influencing the atmospheric and ocean circulations, in the past and today. Because they are projected to lose a significant amount of their mass in the upcoming centuries, this input of freshwater to the global oceans will have serious consequences regarding sea-level rise, affecting a significant share of the World’s population. Ice sheet models can predict sea-level rise, but need to be evaluated against empirical data of past changes to be sure that they represent ice sheet processes accurately. Here we use large-scale, idealised, and regional-scale model setups to understand the main climate drivers behind changes in Dronning Maud Land (East Antarctica) for periods that were both warmer and colder than present. When evaluating high-resolution model results against empirical data, it becomes clear that just as numerical models use data to validate their results, the interpretation of in-situ datasets can be improved with the aid of numerical models. In short, the results presented here are a push towards closer collaboration between numerical modellers and field scientists, who should work in tandem at all stages, from experimental design and sampling planning to interpretation of results.