Comparisons between climate proxies and instrumental records through the last two centuries are often used to understand better the controls on palaeoarchives and to find relationships that can be used to quantify changes in pre-instrumental climate. Here we compare an 80-year-long annually resolved oxygen isotope record from Nar Golu, Turkey, a varved lake sequence, with instrumental records of temperature, precipitation, wind speed, relative humidity and calculated values of evaporation, all of which are known to be possible controls on lake oxygen isotope systems. Significant relationships are found between the isotope record and summer temperatures and evaporation suggesting these are dominant controls on the isotope hydrology of this non-outlet lake. Modelling the stable isotope hydrology of the lake system allows these relationships to be tested independently. We show that the isotope record follows the same trends in the temperature and evaporation records but that, even when combined, these two climatic factors cannot fully explain the magnitude of change observed in the isotope record. The models show the lake system is much less sensitive to changes in evaporation and temperature than the climate calibration suggests. Additional factors, including changes in the amount of precipitation, are required to amplify the isotope change. It is concluded that proxy-climate calibrations may incorrectly estimate the amplitude of past changes in individual climate parameters, unless validated independently.