Climate Science Challenges and Opportunities for Machine Learning
Abstract:
In this lecture we will focus on three areas: 1) monitoring environmental change; 2) modelling our climate; and 3) quantifying climate risks. The first section will highlight some of the key observational datasets available for assessing how our atmosphere and oceans have changed over the recent past (from 1900 to the present day), and some of the challenges surrounding the non-uniform distribution of in-situ measurements. The second section will briefly explain what a climate model simulator is, what they are good at doing and the challenges associated with comparing their output with real-world climate change. Then in the third section we will discuss how the climate research community increases spatial granularity of our climate models and zoom-in on specific regions of interest, such as densely populated regions or vulnerable environments. At the end we will go through a Google Colab notebook using some gridded climate model simulation output to setup a Multifidelity Climate Modelling data challenge!
repo with Google Colab Notebook: https://github.com/scotthosking/mf_modelling
Relevant links and publications:
- D. Bannister, A. Orr, S. K. Jain, I. P. Holman, A. Momblanch, T. Phillips, A. J. Adeloye, B. Snapir, T. W. Waine, J. S. Hosking, C. Allen-Sader (2019) “Bias correction of high-resolution regional climate model precipitation output gives the best estimates of precipitation in Himalayan catchments”, Journal of Geophysical Research: Atmospheres
- The Synthesis Report (SYR) of the IPCC Fifth Assessment Report (AR5) provides an overview of the state of knowledge concerning the science of climate change, emphasizing new results since the publication of the IPCC Fourth Assessment Report (AR4) in 2007. https://www.ipcc.ch/report/ar5/syr/
- UK and US scientists are collaborating to investigate one of the most unstable glaciers in Antarctica, the Thwaites Glacier, roughly the same size as Florida or Britain. https://thwaitesglacier.org/
- Pangeo is a community of people working collaboratively to develop software and infrastructure to enable Big Data geoscience research. http://pangeo.io
- GitHub repo to help download and work with weather station data from across the globe. https://github.com/scotthosking/get-station-data
- GitHub repo with python library to work with Argo float oceanic data, as mentioned during the lecture. https://github.com/euroargodev/argopy