Questions Raised by the Global Warming Hiatus: The Predictability of Tropical Pacific Decadal Variability and Subsurface Warming of the Tropical Atlantic Ocean

This dissertation explores two aspects of interdecadal climate variability: the predictability of variability in the tropical Pacific region on longer-than-interannual timescales, and changes to subsurface heat content in the tropical Atlantic Ocean over a decade. Chapter 1 assesses the ability of a general circulation model (GCM) and an intermediate- complexity model (ICM) to simulate persistent cool states of the Tropical Pacific. These states, which are La Niña-like on average, last from seven to ten years and induce prolonged droughts in many parts of the world. We find that these models produce analogs to real-world examples of these states and that they are modestly predictable in the ICM. We also find some predictability of the interdecadal shifts in the mean state in this model. In Chapter 2, an attractor reconstruction technique is used to investigate the predictability of interdecadal variability in the ICM further. We find that the interdecadal regimes of the ICM emerge as a pair of distinct orbits in a three-dimensional state space, and that the observed system possesses some comparable characteristics. The ICM is found to spend over a third of the time in states from which the regime of the following fifteen years is predictable with high confidence. The implications for developing an interdecadal prediction scheme are discussed. Chapter 3 examines changes in the heat content of the tropical Atlantic below the thermocline during the recent global warming hiatus. We use simulated Lagrangian particles in an ocean reanalysis dataset to analyze the changes to the circulation of the upper tropical Atlantic Ocean that occurred between the hiatus decade and the decade prior to it; and develop a testable hypothesis for how variability in the tropical Pacific region may have influenced the subsurface heat content of the Atlantic