MIT
Earth systems are simulated using numerical models of increasing resolution that play a key role in helping us understand and predict climate. The MIT general circulation model and MIT Darwin Project developers strive to provide open-source and user-friendly solutions to a wide user community of researchers and educators. In this talk, I will present an initial effort to interface these powerful and versatile tools with Julia. Emphasis, for now, is on porting the gcmfaces toolbox to Julia. This Matlab / Octave toolbox allows users to analyze ocean model output using code that is readily applicable to all supported grid types. Porting it to Julia notably aims to (1) improve scalability to increasingly large data sets, (2) alleviate costs associated with proprietary software, (3) increase integration with cloud services, and (4) facilitate access for educators and researchers via jupyter notebooks. Examples will be taken from a recent simulation of marine ecosystems by the MIT Darwin Project. In the longer term, this effort aims to allow users to leverage MITgcm capabilities (parallel solvers, automatic differentiation, virtual particle tracking, etc.) via Julia.
I work as a research scientist at the Massachusetts Institute of Technology (MIT) where I investigate oceanography and climate. As part of the Department of Earth, Atmospheric and Planetary Sciences, my work focuses on ocean modeling and the analysis of global ocean data sets such as Argo profile collections and satellite altimetry. Amongst other approaches, I carry out ocean state estimation using the MIT general circulation model to interpolate and interpret ocean observations. I also participate in the development of the MITgcm and its adjoint. My scientific interests include: Ocean circulation and Climate variability; tracer transport and turbulent transformation processes; interaction of bio-geochemistry and physical processes; global cycles of heat, water, and carbon; observational statistics; forward and inverse modeling.
