Distributed computing is a growing- and powerful- trend in research that puts your PC, and those of thousands of your friends and neighbors, to work in answering analytically difficult questions. I’ve written about how distributed computing is part of the wide range of opportunities out there for citizen scientists (See Action 88). By distributing analytical tasks over thousands of computers, researchers can avoid or delay costly investments in supercomputers that might be well outside their budgets. Distributed computing allows a small part of your computer to be used while its resting for a growing number of research efforts, that include searching for extraterrestrial life or letting it work on public health issues. And now- climate change!
The Union of Concerned Scientists is working with ClimatePrediction.net to run a distributed computing pilot study that tests climate model predictions within their bounds of uncertainty. By evaluating the effect of incremental changes in the values that go into the model, they can look at how sensitive the model is to over and under predicting responses to changes in the input variables, and use that to refine the selected models that are based on past climatological data, in order to select the best models for predicting climate data into the future. As you might imagine, running hundreds of thousands of models many thousands of times across a huge suite of variables is not a small computing task- and this is where the power of distributed computing is so valuable. Each person that signs up to allow their computer to be used during its downtime is helping researchers run more variations of the model and contributing to building model accuracy; which is a really good thing if we want to build better climate models. You can read more about the details of the study here, and go here to download the application you will need to run the program.