Sea ice and abrupt climate change

My PhD research focused on sea ice and abrupt climate change. Tools I used include construction and mathematical analysis of idealized physical models, numerical simulation with global climate models, and analysis of observations. Projects include work on (i) constructing an idealized Arctic sea ice and climate model to investigate the robustness of partial-year sea ice cover and the possibility of multiple equilibria and bifurcation thresholds; (ii) the impact of cloud simulation errors on simulated underlying sea ice in coupled global climate models (GCMs); (iii) an explanation for why the observed Arctic sea ice retreat during the past few decades has appeared to be most pronounced at the annual minimum; (iv) evidence for the mechanism that caused the Dansgaard-Oeschger abrupt warming events during the last glacial period, based on simulations with a coupled GCM; and (v) a proposed mechanism for the Younger Dryas abrupt cold interval 12,000 years ago, supported by the results of a coupled GCM.

Details of a low resolution last glacial maximum CCSM3 simulation we carried out are available here.

While being a graduate student, I also worked on El Niño modeling and observational analysis, past variability in sunlight reaching the earth (paleo insolation), and small-scale ocean mixing (salt fingers). These projects are described below.

El Niño

The term "El Niño" (or similarly "ENSO") describes the episodic warming and cooling of the eastern tropical Pacific on a timescale of several years. The figure above displays the past 20 years of average eastern tropical Pacific sea surface temperatures. A tendency to vary irregularly on a timescale of roughly 3-7 years (interannually) is evident.

A motivation to study the dynamics of El Niño is the possibility of improvement in prediction ability (the major events in 1982 and 1998 each had roughly 2,000 deaths attributed to them). Furthermore, in light of increased concern about climatic response to human-induced global change, improved understanding of El Niño dynamics may allow us to better assess potential feedbacks between tropical Pacific variability (which has global impacts) and global warming.

Rapidly varying weather, and especially westerly wind bursts (tropical Pacific weather events with a timescale of weeks), have been frequently suggested to drive ENSO. We are investigating the possibility that westerly wind bursts are in fact modulated by ENSO itself. In our first study of this hypothesized two-way feedback, we combined analysis of satellite scatterometer data with experiments carried out using an ENSO computer forecast model of intermediate complexity, and we found that the inclusion of observationally based westerly wind burst modulation by ENSO has a huge effect on simulated interannual variability. We extended this work using a more sophisticated ENSO model which combines an ocean general circulation model with a statistical atmospheric model (i.e., a hybrid coupled model). We added an explicit westerly wind burst component to the model atmosphere with guidance from a twenty-three year observational record, and we parameterized westerly wind burst occurrence such that the likelihood of an event depends on the western Pacific warm pool extent. The modulation of westerly wind bursts strongly affected simulated ENSO characteristics in this model.

Paleo insolation

This Matlab script computes daily average insolation (sunlight at the top of the atmosphere) as a function of day and latitude at any point during the past 5 million years. The self-contained script includes orbital parameter data from Berger and Loutre (1991). We posted a version of the script at the NOAA/NCDC Paleoclimatology Program archive (here).

daily_insolation.m

Ocean mixing (salt fingers)

When warm salty water lies above cold fresh water, vertical fingers of salty or fresh water can develop at the interface with a characteristic width of a few centimeters and height of a few meters. This mixing, which is readily observable in laboratory tank experiments as in the image above, occurs because of a scale-selective instability related to the different diffusion rates of heat and salt in water (heat diffuses about 100 times faster). The rapid diffusion of heat allows the energy locked up in the unstable temperature gradient to be released on small scales, even though the total density gradient is stable. Since sunlight heats the upper ocean and also causes enhanced salinity through evaporation, it is not suprising that much of the world ocean is fingering favorable. Observational evidence of salt fingers in the ocean is conflicting, but many oceanographers today believe that salt fingering plays a major role in water mass mixing and that hence this centimeter-scale phenomena could significantly influence global ocean circulation.

An idealized analytical model, consisting of a few equations, was proposed several decades ago to approximately explain the dynamic instability that leads to observed salt finger phenomena. The results of the model were shown to compare favorably with observed phenomena in salt finger experiments, and similar studies based on this model have followed since. I considered a mathematical approximation used in previous analyses of the model in which transient effects were neglected (a typical approximation in geophysical fluid dynamics stability analyses). I found that when the approximation was not made (i.e., transient effects were included), the results of the model changed significantly, leading to model predictions that agree less well with observations. This suggests that the idealized model may not be accurately capturing the key phenomena responsible for salt fingering.