Paleo Data Assimilation of Coral δ 18 O Part 2: 20th‐Century Trends and Variability of the Tropical Pacific

Abstract Limited instrumental observations have created uncertainties in the rates and patterns of historical variations in tropical sea‐surface temperature, precipitation, and salinity. Coral archives are well suited to monitor past variations in the tropical oceans; in particular, the oxygen isotope composition ( δ 18 O) has the unique potential to provide untapped insight into historic changes in the tropical hydrological cycle. We build upon Part 1 to emphasize the challenges and opportunities of using coral δ 18 O with the Last Millennium Reanalysis paleoclimate data assimilation (paleo‐DA) framework. We evaluate the sensitivity of our reconstructions to specific choices of paleo‐DA components, such as climate model priors, coral δ 18 O proxy system models, and salinity calibration products, and reflect on the combination of methodological choices that are required to best incorporate coral δ 18 O with a paleo‐DA approach. We then evaluate trends and variability across the global tropics, finding evidence for a strengthening of the equatorial Pacific sea‐surface‐temperature gradient and an intensification of the hydrological cycle. However, diagnosing the relative rates of warming across the equatorial Pacific is highly challenging, likely due to the high covariability of sea‐surface temperature and salinity anomalies across the global tropics. We also find that the hydrological cycle response to El Niño Southern Oscillation has intensified and marginally shifted eastward over the 19th and 20th centuries. This work emphasizes a need for greater monitoring and improved simulations of δ 18 O sw and salinity, and illustrates the potential for using coral δ 18 O to evaluate past hydroclimate variability with paleo‐DA.