The Impact of Ocean Dynamics on Atlantic Sea Surface Temperature Variability

Abstract

Internal variations of Atlantic sea surface temperature (SST) on decadal to multidecadal timescales have been linked to numerous climate phenomena with subsequent socioeconomic impacts, such as frequency and intensity of Atlantic hurricanes, droughts in Sahel and Arctic sea ice extent. Therefore, understanding the mechanisms of Atlantic SST variability is crucial to producing reliable predictions of these phenomena. Unfortunately, a consensus on the mechanisms of Atlantic SST variability has not been achieved, and as a result there has been intense debate about the relative roles of external forcing, stochastic atmospheric variability, and ocean dynamics in the Atlantic Multidecadal Variability (AMV). The goal of this dissertation is to improve our understanding of the relative roles of ocean dynamics and atmospheric forcing in internally generated Atlantic SST variability. To accomplish this goal, we compare a fully coupled atmosphere-ocean-ice model to a model in which the ocean is replaced by a motionless slab layer (henceforth slab ocean model). Differences in variability between the two models are diagnosed by an optimization technique that finds components whose variance differs as much as possible between the two models. A major result is that suppressing interactive ocean circulations produces significant differences in variability between the two models. In particular, suppressing ocean circulations enhances variability associated with the AMV and the tripole SST, suggesting that ocean dynamics tend to damp large-scale modes. On the other hand, variability associated with the Atlantic Nin Þo and subpolar gyre are suppressed, which is not surprising since ocean circulations are known to play a significant role in these modes. Contrary to initial expectations, the Atlantic Meridional Overturning Circulation has no significant influence on Atlantic SST variability in the fully coupled model. Suppressing ocean circulations impacts atmospheric sea level pressure (SLP) primarily through two modes. The two leading modes are associated with SST patterns similar to the leading SST differences, suggesting that internally forced component of the SLP differences arise primarily as the response to the SST differences. One of these modes is associated with the Atlantic Nin Þo, as one would expect from suppressing ocean dynamics. The differences found here defy simple explanation. For instance, the two models might have different memory time scales. However, significant differences were found even after regressing out SST at several lags. Alternatively, the fully coupled ocean might have more variability simply because it depends on more variables than the slab model. However, significant differences were found even after regressing out a host of variables that exist only in the fully coupled model (e.g., sea surface height). The above results are based on analysis of a model hierarchy. An alternative approach based on observations is utilized to diagnose relations between SST and other variables that indicate ocean dynamics. One such diagnostic is a positive correlation between temperature and salinity. Evaluating this diagnostic in multiple observational and reanalysis products confirms the notion that the subpolar North Atlantic is a region with strong ocean- dynamical influence. On the other hand, this diagnostic suggests that the tropical Atlantic is dominated by atmospheric forcing and potentially a thermodynamic wind-evaporation-SST feedback. Although an Atlantic Nin Þo exists in observations and requires ocean dynamics, it plays a relatively minor role in local regions.