climateprediction.net is a new distributed Internet processing project which went live to the main Internet community in September 2003. As the name suggests, it runs climate prediction simulations. As such, it is relatively heavy on resources. "Block" sizes are much larger than with other projects such as distributed.net or even SETI@Home. One block of data takes a few weeks to process on a Pentium 4 2.4GHz. The good news is that individual stats and team stats are in terms of "model years", with a typical block of data being 45 model years long. Also the client reports in regularly, so your online stats are generally an accurate reflection of your client status.
There is a general consensus amongst the scientific community that global atmospheric temperatures will rise during the 21st Century, but there are wide variations in the predicted size of this temperature increase. The climateprediction.net project attempts to run a large Monte Carlo type simulation of many Global Climate Models (GCMs) testing the various different possible parameters. This will improve the models being used, and to quantify the error. One 'block' of data includes the parameters for a single GCM, and runs the model against historic data from the 19th Century, and then runs a prediction for the 21st Century.
The project originally started in 1999 with a letter in Nature by Myles Allen. David Stainforth from the University of Oxford, the Meteorological Office, the Rutherford Appleton Laboratory, and the University of Reading, all became actively involved. Following funding in 2002 from Natural Environment Research Council (NERC) and The Department of Trade and Industry (DTI), the project was able to advance. This also allowed the inclusion of expertise from the Knowledge Media Institute at the Open University, and the University of Oxford's Computing Laboratory.
A nice feature of the climateprediction.net client is that it can be set to display the temperature and cloud cover in the current climate model (see right). Although this display probably slows the simulation, it has been implemented well and has a slick zoom/pan interface.
To start, go to the climateprediction.net website, and download the client. Set this running, and register as a user (you can do this directly from the client). Once you are registered, you can join the team. Go to the team stats on the website, and find us under the letter "P" as "Prof. James Challis' Most Excellent UK Team". We are classified as a social club.
Climateprediction.net also have a BOINC client. New members are recommended to download the BOINC client instead of the classic client. BOINC users should see our BOINC climateprediction.net team page to join our BOINC team.
Team Graphs (produced by jatkins)
So Who is this Challis Bloke Anyway?
Prof. James Challis' Most Excellent UK Team also works on the SETI@Home Project, Grid.Org Projects, NFSNet Project, and the distributed.net RC5-72 and OGR Projects.
Return to the main Challis Team Page.
Rich's Home Page
Winwaed Software Technology LLC
Contact Page
© Copyright 1995-2008, Richard Marsden