Brian Huang, 2013, Computer Science
This summer, I conducted research in the Program in Atmospheric and Oceanic Sciences (AOS) at Princeton’s Forrestal Campus. The main goal of my research was to contribute to the search for reliable early warning metrics for abrupt climate change. A few studies have analyzed early warning signals by testing them on paleoclimate records which record abrupt climate shifts in the distant past, and although the metrics seem promising, their development is still at an exploratory level. I initially planned to use Monte Carlo simulations to quantitatively measure the power of these early warning indicators using a previously created methodology, but partway through my internship, I reexamined this approach. In the end, after much reading and with my mentor’s help, I designed a novel and more rigorous approach to analyzing the power of our early warning metrics. Among other things, this entailed finding and coding a model for the underlying mechanics of a generic abrupt climate shift, creating an algorithmic method for determining success in the metrics, and a means of comparing metrics despite uncertainty about the optimal combination of their parameters. My internship also gave me the opportunity to attend several interesting seminars and talks within AOS as well as at the NOAA-run Geophysical Fluid Dynamics Laboratory. This experience not only taught me about the current questions in the field of climate science, but also gave me the opportunity to participate in the effort to answer them.