Proxy interpretation of coral‐recorded seawater δ¹⁸O using 1‐D model forced by isotope‐incorporated GCM in tropical oceanic regions

Abstract

The oxygen isotopic ratio in the seawater (δ 18 O sw ) recorded in δ 18 O of coral skeleton for several centuries in tropical regions is an important variable for reconstructing the past climate. However, the relationship between δ 18 O sw and hydrological balance has not been clearly uncovered yet. In this study, a one-dimensional ocean budget model forced by the global quasi-reanalysis isotope data, which is the output of an isotope-incorporated global spectral model, is employed to simulate δ 18 O sw . The results from the simulations are compared with 31 coral records in tropical oceanic regions. The 1-D model successfully reproduces δ 18 O sw in 21 coral records, indicating these corals have the potential to reconstruct the local hydrological budget. The results also confirm that the reproducibility of the 1-D model generally increases with the annual precipitation variability. However, in the regions where the precipitation variability is less, it is more difficult to reconstruct climatic information from the δ 18 O sw , since the nonlocal physical processes unaccounted for in the 1-D model may affect δ 18 O sw . The sensitivity analysis of the 1-D model shows that in tropical oceanic regions, the large-scale precipitation is the dominant factor for δ 18 O sw , confirming that δ 18 O sw recorded in the coral records is a proxy for the precipitation anomaly. This study implies that δ 18 O sw is more closely associated with precipitation and local water isotope budget in tropical oceanic regions than elsewhere, and that the 1-D model is a useful approach for the proxy interpretation in tropical oceanic regions.

Global Evaluation of Proxy System Models for Stable Water Isotopes With Realistic Atmospheric Forcing

“…The variability in δ 18 O recorded in coral skeleton aragonite (δ 18 O coral ) depends on the calcification temperature and local δ 18 O in seawater (δ 18 O sw ) at the time of growth (Epstein & Mayeda, ). Previous studies have modeled δ 18 O coral as a linear combination of sea surface temperature (SST) and the isotopic composition of seawater, as follows: where a and b are the constants (Brown et al, ; Juliet‐Laclerc & Schmidt, ; Liu, Kojima, et al, ; Liu, Kojima, et al, ; Thompson et al, ; Stevenson et al, ; Conroy et al, ). Generally, SST and δ 18 O coral are simply regressed using present‐day observations to obtain the gradient ( a ) by assuming that δ 18 O sw is constant.…”

“…Brown et al () selected a gradient of −0.20‰/°C to model δ 18 O coral . More recently, Liu et al (2013, Liu, Kojima, et al, ) calculated the gradient with consideration of the covariance between SST and δ 18 O sw . In the present study, a −0.20‰/°C gradient was selected following Brown et al () because the covariance between SST and δ 18 O sw was not available for all sites (Liu, Kojima, et al, ).…”

“…The isotopic composition of seawater is mainly determined by the net freshwater flux between the ocean and atmosphere (i.e., evaporation and precipitation; Cole & Fairbanks, ; Fairbanks et al, ). However, mixing with the deeper ocean also affects δ 18 O in sea‐surface water (Brown et al, ; Liu, Kojima, et al, ; Liu, Kojima, et al, ). Accordingly, variation in the surface water and its isotope content over time can be expressed as and where Q is the depth of the upper layer of the ocean (constant), P is the precipitation, E is the evaporation, and D is the mixing with the lower layer (damping factor).…”

“…The depth of the upper layer and the damping factor by mixing are the tunable parameters. We set these parameters as Q = 20 m and D / Q = 0.4 month following (Liu et al, ; Liu, Kojima, et al, ). The effects of horizontal advection, terrestrial runoff, and site‐by‐site differences on the strength of vertical mixing are neglected.…”

“…In particular, atmospheric simulation must be constrained by data assimilation; otherwise, atmospheric states can easily depart from the real states due to the chaotic nature of the atmosphere. However, only a few studies have evaluated PSMs with realistic inputs (Berkelhammer & Stott, 2012;Kanner et al, 2013;Ortega et al, 2014;Liu, Kojima, et al, 2014;Stevenson et al, 2015). Furthermore, most studies have evaluated PSMs at a single site or in a limited area (Brown et al, 2006;Keel et al, 2016;Vuille et al, 2003;Evans & Schrag, 2004;Anchukaitis et al, 2008).…”