Oxygen and carbon isotope (δ 18 O and δ 13 C) data from bulk carbonates are widely applied proxies for temperature, the precipitation/evaporation ratio and productivity in lacustrine palaeohydrology and palaeoclimatology. In case of the terminal and alkaline Lake Van, however, previous studies have shown that bulk oxygen isotope compositions are in disagreement with other proxies when interpreted in a conventional manner. Similarly, the reports on the nature and the timing and site of carbonate precipitation in Lake Van are inconsistent. This study provides evidence on the mineralogy (X-ray powder diffraction analysis, scanning electron microscope imaging, confocal Raman microscopy, electron microprobe analysis) and isotope composition (δ 18 O and δ 13 C) of non-skeletal carbonate minerals in a Lake Van sedimentary profile spanning the last ca. 150 kyr. Carbonate phases present in the sediment include aragonite, low-Mg calcite, and calcian dolomite. Dolomite forms as an early diagenetic phase and occurs episodically in high concentrations driving the bulk isotope record towards the higher dolomite δ 18 O and varying δ 13 C values. Aragonite and low-Mg calcite precipitate in the surface water and are present in the sediments in varying amounts (relative aragonite to calcite content for dolomite-poor samples Ar/(Ar + Cc) of 93 to 41 wt%). In an attempt to explain this variation, we revised a precipitation model based on annually laminated sediments containing both aragonite and calcite spatially separated in light and dark coloured laminae, respectively. According to our model, spring calcite precipitation, under close-to-freshwater conditions, is followed by evapoconcentration-driven aragonite precipitation in late summer. The precipitation of these carbonate polymorphs from chemically differing surface waters (i.e. freshwater-influenced and evapoconcentrated) leads to distinctly different oxygen and carbon isotope signatures between sedimentary penecontemporaneous aragonite and calcite. The δ 18 O and δ 13 C values of aragonite relative to calcite are significantly higher by several per mille than inferred from aragonite-calcite fractionation factors alone, suggesting that the generalised assumption of sedimentary coeval calcite and aragonite precipitating from water with the same isotopic composition is flawed. The here proposed revised hydrologically-separated carbonate precipitation model is not only taking (i) differences in the isotopic fractionation between carbonate minerals into account, but also (ii) considering the hydrological conditions and the processes favouring the precipitation of a given mineral and ultimately controlling its isotopic composition. If mixed mineralogies are present, this mineralogy-specific approach has the potential of refining environmental reconstructions and reconciling apparently equivocal interpretations of different proxy records.