Surface processes alter the water stable isotope signal of the surface snow after deposition. However, it remains an open question to which extent surface post-depositional processes should be considered when inferring past climate information from ice core records. Here, we present simulations for the Greenland Ice Sheet, combining outputs from two climate models with an isotope-enabled snowpack model. We show that surface vapor exchange and associated fractionation imprint a climate signal into the firn, resulting in an increase in the annual mean value of δ18O by +2.3‰ and a reduction in d-excess by −6.3‰. Further, implementing isotopic fractionation during surface vapor exchange improves the representation of the observed seasonal amplitude in δ18O from 65.0% to 100.2%. Our results stress that surface vapor exchange is important in the climate proxy signal formation and needs consideration when interpreting ice core climate records.