Disentangling the predictability, drivers, and implications of observed changes in the Southern Hemisphere

Abstract

Despite rising global-mean temperatures, the Southern Ocean and tropical eastern Pacific Ocean have cooled during the satellite era and Antarctic sea ice has undergone rapid fluctuations. Earth System Models (ESMs) forced with historical emissions generally struggle to reproduce sea surface temperature (SST) and sea ice trends comparable to those seen in nature. In this dissertation, I examine several broad questions related to the drivers and predictability of observed changes in the Southern hemisphere. Chapter 1 frames and motivates the research questions explored in this dissertation. Chapter 2 employees a fully-coupled ESM nudged to observed winds to examine the drivers and predictability of record low Antarctic sea ice in austral winter 2023. Chapter 3 shows that ESMs tend to underestimate the strength of the subtropical Pacific cloud-SST feedback, leading to damped internal multidecadal Eastern Pacific SST variability. Consequently, the likelihood that internal variability has contributed to historical cooling in the Eastern Pacific is larger than previously appreciated. Chapter 4 links excessive precipitation in the tropical Pacific south of the equator to weakened wind-evaporation-SST and cloud-SST feedbacks. This model bias, known as the double intertropical convergence zone (ITCZ), is shown to be related to the strength of the Southern Ocean–eastern Pacific teleconnection. Chapter 5 presents DLESyM-Ocean, a probabilistic deep learning model for simulating present-day upper ocean and sea ice for use in subseasonal to seasonal forecasting and predictability studies. Conclusions and future works are discussed in Chapter 6.