Stable water isotopes in atmospheric general circulation models

The water isotopes (HDO)-O-16 and (H2O)-O-18 are widely used as tracers in the global hydrological cycle. The isotope composition of modern precipitat ion is strongly linked to local temperatures in high latitudes and, less cl early, to the amount of precipitation in low latitudes. As the isotopic com position of precipitation can be reconstructed from many palaeo-archives th ese isotopes serve as quantitative climate proxies for the past. In this re port we describe the different type of models that have been used to unders tand the global distribution of the isotopic composition of precipitation, today and in the past. The focus of this paper is on the results of atmosph eric general circulation models (AGCMs) that have been fitted with water is otope diagnostics. A good correspondence between simulated and observed wat er isotope signals is demonstrated on an annual time-scale. In particular t he regional and seasonal isotope-temperature relations estimated by the AGC Ms were in good agreement with global observations. Furthermore, we discuss the results of the models when they are forced with the climatic boundary conditions of 21 kyr BP (last glacial maximum) and of 6 kyr BP (mid-Holocen e). These numerical experiments allow us to compare directly the results of the AGCMs, that is the water isotopes, with the corresponding palaeo-obser vations. We confirmed that in general the water isotopes are valuable proxi es of temperature variations in high latitudes. The calibration of the 'iso tope- temperature thermometer', however, might be biased by local effects, such as the seasonal distribution of precipitation or the strength of the l ocal temperature inversion. Moreover, comparable 'small differences' in the forcing compared with today's climate, such as in the mid-Holocene experim ent, produce small responses of the water isotopes, which are often strongl y affected by atmospheric circulation changes. Finally, future applications of isotopic AGCMs, such as the coupling with the carbon cycle or the study of cloud processes, are discussed also. Copyright (C) 2000 John Wiley & So ns, Ltd.