Any assessment of future climate change requires knowledge of the full range of natural variability in the El Niño/Southern Oscillation (ENSO) phenomenon. Here we splice together fossil-coral oxygen isotopic records from Palmyra Island in the tropical Pacific Ocean to provide 30-150-year windows of tropical Pacific climate variability within the last 1,100 years. The records indicate mean climate conditions in the central tropical Pacific ranging from relatively cool and dry during the tenth century to increasingly warmer and wetter climate in the twentieth century. But the corals also document a broad range of ENSO behaviour that correlates poorly with these estimates of mean climate. The most intense ENSO activity within the reconstruction occurred during the mid-seventeenth century. Taken together, the coral data imply that the majority of ENSO variability over the last millennium may have arisen from dynamics internal to the ENSO system itself.

The El Niño-Southern Oscillation (ENSO) drives large changes in global climate patterns from year to year, yet its sensitivity to continued anthropogenic greenhouse forcing is uncertain. We analyzed fossil coral reconstructions of ENSO spanning the past 7000 years from the Northern Line Islands, located in the center of action for ENSO. The corals document highly variable ENSO activity, with no evidence for a systematic trend in ENSO variance, which is contrary to some models that exhibit a response to insolation forcing over this same period. Twentieth-century ENSO variance is significantly higher than average fossil coral ENSO variance but is not unprecedented. Our results suggest that forced changes in ENSO, whether natural or anthropogenic, may be difficult to detect against a background of large internal variability.

We demonstrate that the carbon isotopic signal of mid‐depth waters evolved differently from deep waters in the South Atlantic sector of the Southern Ocean during the Pleistocene. Deep sites (>3700 m) exhibit large glacial‐to‐interglacial variations in benthic δ13C, whereas the amplitude of the δ13C signal at Site 1088 (∼2100 m water depth) is small. Unlike the deep sites, at no time during the Pleistocene were benthic δ13C values at Site 1088 lower than those of the deep Pacific. Reconstruction of intermediate‐to‐deep δ13C gradients (Δ13CI‐D) supports the existence of a sharp chemocline between 2100 and 2700 m during most glacial stages of the last 1.1 myr. This chemical divide in the glacial Southern Ocean separated well‐ventilated water above ∼2500 m from poorly ventilated water below. The Δ13CI‐D signal parallels the Vostok atmospheric pCO2 record for the last 400 kyr, lending support to physical models that invoke changes in Southern Ocean deep water ventilation as a mechanism for changing atmospheric pCO2. The emergence of a strong 100‐kyr cycle in Δ13CI‐D during the mid‐Pleistocene supports a change in vertical fractionation and deep‐water ventilation rates in the Southern Ocean, and is consistent with possible CO2‐forcing of this climate transition.