Validation of δ18O as a proxy for past monsoon rain by multi-GCM simulations

Stable oxygen isotope ratios (δ¹⁸O) of tree cellulose and speleothem carbonate are useful proxies for past monsoon rain in many tropical regions, as a decrease in rain δ¹⁸O is observed with increase in rainfall on a monthly time scale. This amount effect varies spatially; therefore a local calibration, with actual measurements of rain amount and its δ¹⁸O is required. Such observations, however, are quite limited in space and time. To circumvent this difficulty, many isotope enabled general circulation models (GCMs) are used to aid the interpretation of ¹⁸O proxies; nevertheless, all such simulations taken together are yet to be evaluated against observations over the Indian summer monsoon (ISM) region. Here we examine ten such GCM simulations archived by the stable water isotope INtercomparison Group, phase 2. The spatial patterns of simulated ISM rainfall and its δ¹⁸O are in good agreement with the limited observations available. Simulations nudged with observed wind fields show better skill in reproducing the observed spatio-temporal pattern of rainfall and its δ¹⁸O. A large discrepancy is observed in the magnitude of the simulated amount effect over the Indian subcontinent between the models and observation, probably because models simulate the spatial distribution of monsoon precipitation differently. Nudged simulations show that interannual variability of rainfall δ¹⁸O at proxy sites are controlled by either regional (rather than local) rainfall or upstream rain out. Interannual variability of rainfall δ¹⁸O over the East Asian region is well correlated with ENSO, while it is only weakly correlated over the Indian sub-continent.