Computational Sustainability is meant to encompass both efficient computing as Allesandro points out, but the primary focus of the Institute for Computational Sustainability at Cornell (who I believe coined the term) is to explore how to use computational methods to build a more sustainable world. This includes solving complex modeling, prediction and optimization problems in sustainability problems broadly understood. The domains I have seen addressed include:
- smart grid power systems
- species migration tracking
- occupancy prediction (ie. are there birds here even if we have no observations of them)
- migration corridor design for endangered species
- optimal placement of wind farms to maximize energy production
- control of invasive species
- optimal management of forests to reduce fires, increase diversity, sustainable harvests
For more on this check out http://blog.computational-sustainability.org/
If you are looking for ideas on a course on Computational Sustainability might look like, Alan Mackworth at the University of British Columbia is offering one next term: