is a required field.
Review summary
C-LEARN is an excellent tool for basic, rapid comparisons of climate change scenarios. Results from C-LEARN simulations can be compared in a relative sense, but output values should not be considered precise or absolute.Model function in the CoLab
Provides the "backbone" for climate impact simulation in each proposal generated by CoLab users. Users provide inputs to C-LEARN for three regions. Then C-LEARN models CO2 and GHG emissions and generates proposal outputs as well as data used by connected models (which in turn provide further information about climate impacts not simulated by C-LEARN).Model strengths
This very rapid simulation model reproduces the response properties of state-of- the-art three dimensional climate models very well--well within the uncertainties of the high resolution models--and with sufficient precision to provide useful information for its intended audience. The dynamic non-linear model is sufficiently sensitive that C-LEARN can be used as a decision support tool in understanding and, for example, discussing efforts aimed at limiting temperature change to 2 degrees C of warming - the currently espoused goal of the European Union that is supported by many (but not by all) in the scientific and NGO communities. The ability to rapidly test a range of policy proposals for future emissions is particularly useful for audiences wanting to understand the implications of different decisions in real time.The climate portion of the C-LEARN / C-ROADS model calculates globally averaged temperature by accounting for incoming solar energy to the Earth and outgoing infrared radiation. This method of calculation is intuitively reasonable and well established in climate science. It is thoroughly documented in the C-ROADS Simulator Reference Guide, available on the ClimateCoLab.org web site. Using this method, the interactive Web features of C-LEARN allow users to quickly compute the globally averaged temperature and sea level resulting from different assumptions about the world's use of fossil fuels and other resources.
Model weaknesses
In using C-LEARN, it is important to remember that like all other climate impact models, C-LEARN is a sensitivity tool, rather than a tool to provide precise quantitative estimates of projected emissions, CO2 concentrations, and temperature and sea level responses. The issues is that in C-LEARN, a single set of assumptions generates a single set of results. Actually no one knows exactly how much global warming, sea level rise, etc. will result from any particular scenario of fossil fuel use, deforestation, and reforestation. Predictions of future climate such as those illustrated in the IPCC Working Group 1 Summary for Policymakers span a wide range of possibilities even when the world's use of fossil fuels and other resources is "given."Other comments
C-LEARN is a modified, web-based version of C-ROADS. (See summary review of C-ROADS).Suggestions for model improvement
To ensure effective communication on the capabilities of the model simulations, we recommend that the simulations be more explicit on the uncertainties about the climate system, e.g., in the form of probability relationships within and outcomes of the model. At a minimum, the model developers should consider having no more that 2 significant figures for any quoted numerical result and reporting percentage changes in 5-point increments at best. In general, representation needs to be carefully considered. One possibility is a grey area surrounding the specific scenarios portrayed which includes dotted lines for 5th, 25th, 75th and 95th percentile runs if possible. This adds complication, to be sure, so perhaps showing this on a second page of graphs that can be accessed by a click would be best.
As the dramatic impacts associated with sea level rise are on a longer multi-century time scale, we recommend that sea level be represented for more than 100 years, or the risk be conveyed by alternate means. For instance, the new report from the Stabilization Committee for the National Academies provides an interesting graphic that links specific impacts drawn from the quantified literature to different temperature increases. Appending it to the right-hand side of the temperature change graph (yet another page of results) might be informative for people who are interested in looking at the implications of alternative futures. Click here to download the National Academies Report-in-Brief, which contains the graphic as its second figure.
We recommend including feedback from global temperature on ocean and land processes in the C-LEARN model, e.g., feedbacks on the carbon cycle, and consider a module on ocean acidification.
We also recommend that the model be further developed to explicitly include the six Kyoto gasses plus other human-induced drivers of climate change (e.g. soot, ozone, SO2).
A simple mapping extension could also be used show the land area impacted by rising sea levels.
Climate policy will be an iterative process, so it would be good if, at least, a second set of decisions (targets) could be allowed (a fundamental conclusion of IPCC-AR4 -- see page 22 of the SPM of the Synthesis Report). At the very least, consider something like allowing {Xi} for the the various country groupings for intermediate targets like 2050 (different for different country groupings) and {Yi} as a second set of targets for 2100. For example, in the current system, it is possible to do Stern's Global Deal for 2050 (80% reduction in emissions from the developing world, rapidly developing country emissions peaking in 2020, and developing country emissions peaking in 2030, but then it's not possible to change anything after that. The result is concentrations at 460 or so in 2050 (delta T at 1.7oC or so) rising to 538 (3.1oC) by 2100. Clearly what happens after 2050 is important. This would also allow investigation of what would have to be done if certain country groups did not meet those objectives, but these are decisions that would be made then; i.e., nobody is going to know in 2012 that this or that will or will not happen. Indeed, have two or three iteration points might be fun for a more advance exercise for people who want to explore something closer to reality. Perhaps intermediate numbers (say 2050) could be provided in any case.
