Project: Random Forest vs. Neural Networks for Estimating the Ocean Carbon Sink
The ocean has absorbed the equivalent of 41% of industrial-age fossil carbon emissions. In the future, this rate of this ocean carbon sink will determine how much of mankind’s emissions remain in the atmosphere and drive climate change. To quantify the ocean carbon sink, surface ocean pCO2 must be known, but cannot be measured from satellite; instead it requires direct sampling across the vast and dangerous oceans. Thus, there will never be enough observations to directly estimate the carbon sink as it evolves. Data science can fill this gap by offering robust approaches to extrapolate from sparse observations to full coverage fields given auxiliary data that can be measured remotely.
In this project, the intern will pick up on initial, well-documented work of MS students from the Fall 2018 Data Science Capstone. This intern’s work evaluate the potential of both neural network and random forest approaches in two contexts (1) 1% coverage pCO2 data to both train and test the algorithm, and (2) full-coverage pCO2 estimated from a climate model to train the algorithm and then all of the observed data to test. They will collaborate closely with Professor McKinley and senior graduate student Lucas Gloege.
One selected candidate will receive a stipend via the DSI Scholars program. Amount is subject to available funding.
- Professor Galen McKinley
- Department/School: Earth and Environmental Science Lamont Doherty Earth Observatory
- Location: at LDEO, Comer Building; at Morningside, Schermerhorn Extension 5th floor
- The McKinley Ocean Carbon Research Group studies how ocean physical and biogeochemical processes impact large-scale carbon cycling and primary productivity. These studies encompass fluid dynamics, climate processes, biogeochemistry and ecology. Our primary research tools are numerical models and large historical datasets.
- Earliest starting date: 03/01/2019
- End date: 05/30/2019
- Number of hours per week of research expected during Spring 2019: ~10
- Skill sets: Knowledge and experience of: Python and CUDA; – Architecture of GPUs and heterogeneous systems; – Data structures and algorithms for massively parallel systems- clusters and clouds; – Profiling and performance optimization of GPUs or other compute accelerators using different frameworks.
- Student eligibility (as of Spring 2019):
freshman, sophomore, junior, senior, master’s
- International students on F1 or J1 visa: eligible
- Additional comments: interest in earth science; great if had some earth science classes, but not required.