Re-investigation on the potential of paleoclimate proxies using Reanalysis of oxygen isotopic composition of seawater with an isotope incorporated AGCM

Oxygen isotopic composition of seawater (δ18Osw), which primarily reflects hydrologic balance at the sea surface and/or terrestrial runoff, is an important parameter for paleoclimate reconstruction. However, a complete understanding of the nature of δ18Osw has not been achieved yet because mutual approach between the observational and theoretical researches on δ18Osw is currently limited. An approach using reanalysis data of δ18Osw is, therefore, important for better understanding of the meaning of δ18Osw signal. In this study, we calculated δ18Osw from the reanalyzed climate data and compared with observed δ18Osw data and coral-derived δ18Osw time-series. We employed a simple one-dimensional budget model, which is forced by hydrologic and isotopic mass-exchange at the ocean-atmosphere interface (i.e. precipitation, evaporation and their isotopic compositions) derived from an isotope-incorporated AGCM IsoGSM [Yoshimura et al. 2008]. For calibration of the model we firstly examined δ18Osw time-series derived from a Porites coral taken from inside a gulf of the east coast of the Philippines, where horizontal advection and terrestrial runoff is relatively negligible. The model successfully reproduced the coral-inferred local hydrologic budget of the site. We also simulated δ18Osw considering terrestrial runoff and three-dimensional seawater advection. As the second step, we extended our examination to 32 coral-derived δ18Osw time-series. It is revealed that the method again successfully reproduced the coral-derived δ18Osw time-series at many sites. The good consistency between the model and the coral records infers the potential of δ18Osw as an indicator of local hydrologic balance, but the relationship between model-based δ18Osw estimates with coral-derived δ18Osw reconstruction varies in regions. We found a tendency that both δ18Osw is more consistent in the region with higher precipitation. On the other hand, neither seasonal nor inter-annual variability of δ18Osw reflects local hydrologic balance, where the annual precipitation is less than 1000 mm/year, except for isolated areas. In addition, the dominating factor on δ18Osw is possibly different with time scales owing to the fact that the reproducibility of the model varies with time scales.