What Controls the Skill of General Circulation Models to Simulate the Seasonal Cycle in Water Isotopic Composition in the Tibetan Plateau Region?

This study evaluates the simulation of the seasonal cycle of water isotopic composition over Tibetan Plateau regions (TP) from six isotope-enabled general circulation models (GCMs) participating in the second Phase of Stable Water Isotope Intercomparison Group. For both meteorological factors (precipitation rate and wind field) and isotopic composition, GCMs generally agree with reanalysis data and in-situ observations, but there is a significant spread across models and the isotopic seasonality is systematically underestimated. In the southern TP, the precipitation isotopic composition is more depleted in summer than in winter, and the amplitude of the simulated isotopic seasonal variations is primarily driven by the amplitude of the simulated upstream precipitation. In contrast, in the northern TP, the precipitation isotopic composition is more depleted in winter than in summer, and the amplitude of the simulated seasonal variability of isotopes is mainly driven by the simulated strength of the zonal wind. We conclude that the skill of a GCM to simulate the seasonal cycle in the isotopic composition depends mainly on the skill of the GCM to simulate the Indian summer monsoon precipitation and the westerlies. The same causes contributing to the underestimated seasonality at present-day may also contribute to the underestimated delta O-18 change at the mid-Holocene.

Automated summary

This study evaluates the simulation of the seasonal cycle of water isotopic composition over Tibetan Plateau regions from six isotope-enabled general circulation models. The results show that there are differences among the models and the simulated isotopic seasonality is generally underestimated. The simulations demonstrate that in the southern Tibetan Plateau, the precipitation isotopic composition is more depleted in summer than in winter, while in the northern plateau it is the opposite. The simulations are primarily influenced by the simulated upstream precipitation and the simulated strength of the zonal wind.