In recent decades, the west Antarctic Peninsula (WAP) has warmed more rapidly than anywhere else in the Southern Hemisphere. Associated with this, there has been a marked shortening of the sea ice season, a retreat of the majority of glaciers, and an increase in precipitation. Each of these changes in the freshwater system has the potential to exert significant influence on the ecosystem, via processes such as stabilisation of the upper water column, and supply of micronutrients to the mixed layer. Here we use a time series of hydrographic and stable oxygen isotope (δ18O) measurements collected at a near-coastal site in Marguerite Bay to quantify the prevalence of meteoric freshwater (glacial melt plus precipitation) separately from sea ice melt. During 2002–2009, meteoric water dominated, with summer water column inventories of order 4–6 m. Summer sea ice melt inventories were lower, ranging from −1 to 0.5 m (where a negative value indicates net sea ice formation from this water). In the near-surface layers, we find highest meteoric water prevalence in February 2006 (6%) and lowest in October 2007 (1%), whilst sea ice melt is highest in February 2005 (2%) and lowest in July 2002 (−2%). The ranges in both meteoric water and sea ice melt are significantly larger than derived previously using a subset of the data, reflecting the strong interannual variability present. The largest single determinant of the near-surface freshwater percentages is found to be changes in mixed layer depth. Notably deep layers occurred in the winters of 2003, 2007 and 2008, due to northerly winds associated with El Niño/Southern Oscillation and the Southern Annular Mode. These led to greatly reduced sea ice cover in northern Marguerite Bay, and allowed persistent air-sea heat fluxes and stronger rates of sea ice production, which is a key factor in controlling mixed layer depth. We also discuss the possible role of interannual changes in wind-induced mixing in this context. As climate change at the WAP continues, we expect further changes in each of the components of the freshwater budget, and also changes in the vertical redistribution of this freshwater by oceanographic processes. Our ongoing δ18O monitoring will help track these changes, and elucidate their consequences for the operation of the marine ecosystem.