Excerpt from a paper by a CS prof at University of Toronto (from 2010) discussing some obvious, direct ways that people with coding skills etc. can contribute to climate science:
Software quality is a particular concern. Climate scientists build a variety of software tools to support their work.
At the heart of the field are the Global Circulation Models (GCMs) that simulate the atmosphere, oceans, cryosphere and biosphere, to study the processes of climate change on a global scale, and generate future projections used in the IPCC assessments [14].
Less glamourous, but equally important, a large number of data handling and analysis tools are used for processing the raw observational data and the results of simulation runs, and for sharing climate data with the broader scientific community.
Most of this software is built by the climate scientists themselves, who have little or no training in software engineering.
As a result the quality of this software varies tremendously: The GCMs tend to be exceptionally well engineered [5], while some data processing tools are barely even tested.