>
| By Ivo Grigorov http://www.the-eggs.org/articles.php?id=141 |
A series of recent events has fuelled a hot debate over the transparency and credibility of climate research. While the debate between sceptics and believers may continue, the circumstances have provided good context for “citizen science” to spill over into climate research.
The concept is not a new one and already applied in astronomy & planetary science, archaeology and biodiversity studies. The idea is that volunteers participate in tasks where human perception and common sense are needed, without the time-consuming scientific training. So could the concept work in something as technical, multi-disciplinary and complex as Global Climate Change modelling? Earlier this year, the Clear Climate Code Project (CCC; http://clearclimatecode.org) set up by the staff of the Cambridge-based Ravenbrook Limited software engineering consultancy (http://www.ravenbrook.com), published their own version of NASA’s Goddard Institute for Space Studies GISTEMP Model. CCC is a volunteer-based project founded on the premise that “The results of some climate-related software are used as the basis for important public policy decisions. If the software is not clearly correct, decision-making will be obscured by debates about it”. The goals of the small group of software engineers are to: 1) produce clear climate science software; 2) encourage the production of clear climate science software; 3) increase public confidence in climate science results, without judgement or arbitration of climate science. Why start with GISTEMP1? GISTEMP is just one of the instrumentation record analyses openly available2 but one has to start somewhere. The CCC team took the original version and re-wrote it in a single software, Python, in order to restructure the code for clarity for competent users who are not necessarily scientists, while attempting to independently reproduce Hansen’s originally published results. The results were not only reproduced3 (Figure 1), but the Python version of the model is significantly lighter (40% of original code), clearer (with half the codelines carrying explanation and comments) and significantly faster. Moreover, Hansen’s collaborator at NASA GISS, Dr Reto Ruedy, has openly praised the re-coding of the model by saying “I hope to switch to your version of that program …Ideally, we would like to replace our whole code”4. What next? Clear Climate Code are currently working on an integrated graphic visualisation tools for GISTEMP. Beyond that, CCC are looking to repeat the demonstration with other global models focussed on Arctic Sea Ice Extent and past temperature reconstructions. The goals of the computer engineers are also very complimentary to those of Marine Ecosystem Evolution in a Changing Environment (MEECE)5 Project. Funded by Framework Program 7, the MEECE project, coordinated by Plymouth Marine Laboratory (UK), aims to 1) improve the knowledge base on marine ecosystems and their response to climate and anthropogenic pressure, as well as 2) develop innovative predictive management tools based on the current generation of marine ecoystem models. A central step in that ambition is making the current generation of marine ecosystem models more transparent and usable by any competent user outside the original development team. Making source code accessible and readily usable is a skill in itself and a task that often does not make the list of priorities when there are pressing scientific questions to be answered. The CCC demonstration shows that the benefits can be beyond simple transparency and public confidence in research. Accessible and readily usable model code can invite constructive contribution from outside the research domain, and poses the question whether the GISTEMP code clarification can spill over into other of Global Climate Change modelling fields, if “citizen scientist” are given the minimum of technical documentation and access to the source code? Clear Climate Code (http://clearclimatecode.org) is set up by the staff of the Cambridge-based Ravenbrook Limited software engineering consultancy (http://www.ravenbrook.com). Contact: Nick Barnes, nb@ravenbrook.com. MEECE Integrated Project (http://www.meece.eu) is a research project funded by Framework Programme 7. Through its Model Library (http://www.meece.eu/library.html) the projects aims to bring transparency to marine ecosystem models by providing access to the minimum technical information necessary for a competent non-expert to apply the models. References 1. Hansen, J.E., and S. Lebedeff, 1987: Global trends of measured surface air temperature. J. Geophys. Res., 92, 13345-13372 2. Code source: http://data.giss.nasa.gov/gistemp/sources/ 3. http://ccc-gistemp.googlecode.com/files/ccc-gistemp-0.2.0-comparison-2010-01-11.html 4. Reto Ruedy-CCC communication on Google Groups – http://groups.google.com/group/ccc-gistemp-discuss/msg/bdba6c032080f05b 5. MEECE Integrated Project is funded by Framework Program 7, www.meece.eu Ivo Grigorov (ivo_grigorov@hotmail.com) is a European Programs Officer at CNRS, France (IUEM, Place Copernic, Technopole, Plouzane, France 29200) and DTU-Aqua, Denmark.
|
Uma (in)certa antropologia 
Figure 1. Global annual temperature anomaly. Without an offset, the CCC-version (red) replicates GISS original output (black) so well that it is barely visible. For full GISTEMP-CCC comparison, visit http://clearclimatecode.org/category/status/