Meridional atmospheric transport is an important process in the climate system and has implications for the availability of heat and moisture at high latitudes. Near-surface cold and warm temperature advection over the ocean in the context of extratropical cyclones additionally leads to important air–sea exchange. In this paper, we investigate the impact of these air–sea fluxes on the stable water isotope (SWI) composition of water vapour in the Southern Ocean’s atmospheric boundary layer. SWIs serve as a tool to trace phase change processes involved in the atmospheric water cycle and, thus, provide important insight into moist atmospheric processes associated with extratropical cyclones. Here we combine a 3-month ship-based SWI measurement data set around Antarctica with a series of regional high-resolution numerical model simulations from the isotope-enabled numerical weather prediction model COSMOiso. We objectively identify atmospheric cold and warm temperature advection associated with the cold and warm sector of extratropical cyclones, respectively, based on the air–sea temperature difference applied to the measurement and the simulation data sets. A Lagrangian composite analysis of temperature advection based on the COSMOiso simulation data is compiled to identify the main processes affecting the observed variability of the isotopic signal in marine boundary layer water vapour in the region from 35 to 70◦ S. This analysis shows that the cold and warm sectors of extratropical cyclones are associated with contrasting SWI signals. Specifically, the measurements show that the median values of δ18O and δ2H in the atmospheric water vapour are 3.8 ‰ and 27.9 ‰ higher during warm than during cold advection. The median value of the second-order isotope variable deuterium excess d, which can be used as a measure of non-equilibrium processes during phase changes, is 6.4 ‰ lower during warm than during cold advection. These characteristic isotope signals during cold and warm advection reflect the opposite air–sea fluxes associated with these large-scale transport events. The trajectory-based analysis reveals that the SWI signals in the cold sector are mainly shaped by ocean evaporation. In the warm sector, the air masses experience a net loss of moisture due to dew deposition as they are advected over the relatively colder ocean, which leads to the observed low d. We show that additionally the formation of clouds and precipitation in moist adiabatically ascending warm air parcels can decrease d in boundary layer water vapour. These findings illustrate the highly variable isotopic composition in water vapour due to contrasting air–sea interactions during cold and warm advection, respectively, induced by the circulation associated with extratropical cyclones. SWIs can thus potentially be useful as tracers for meridional air advection and other characteristics associated with the dynamics of the storm tracks over interannual timescales.