Online Help

From MAGICC6 Wiki
Revision as of 18:12, 17 June 2013 by Mmalte (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

With the live.magicc.org interface for running MAGICC6 on our servers, you have a simple 3-step process to generate your own climate scenario output. Below, we provide assistance with the choices you can make during this three step process;

Step 1: Choosing Emissions

Select one or multiple existing emission scenarios =

On the left of the first tab, you can select an emission scenario. Hold down the control key to select multiple emission scenarios.

Browse through the selected emission scenarios =

You can browse through the global emissions from your selected scenarios, for example Fossil CO2, Methane etc. This helps you selecting the scenarios (with admittedly cryptic names sometimes). Note: any selection you make here only affects the graph that is immediately right to it. The climate model runs will always take into account all of the emissions of your selected scenarios


Upload your own emission scenario

You have the option to upload your own MAGICC emission scenario as an ASCII file. You can either provide only globally aggregate emissions or regional emissions. If you want to put your own emission data into such a scenario input file, see this page of how to create your own .SCEN file .

Step 2: Select Model Settings

You can run MAGICC in two distinct modes, a "standard" one and a "probabilistic" one.

Selecting "Probabilistic" run mode

If you choose the "probabilistic" setting, your selected emission scenarios will be run multiple times by MAGICC, each time with a slightly different parameter setting. The results will not be single temperature or CO2 concentration outcomes for each scenario, but actually uncertainty distributions.

Note: Since probabilistic runs will require the climate model to be run 171 or even 600 times for each scenario, some patience is required. Normally, finishing a probabilistic run will merely take a couple of minutes, perhaps up to 10min. You will be provided with distributions of key climate outputs for every decade, but not for every year, as under the "standard" runs.

Choose the "multi-model ensemble" probabilistic run mode

If you select this option, we will run your emission scenarios 171 times, with all combinations of 19 AOGCM calibrations and 9 carbon cycle model calibrations. These AOGCMs and carbon cycle models are from the IPCC Fourth Assessment Report and belong to the so-called "CMIP3" and "C4MIP" intercomparisons. If you assume that all those 19 AOGCMs and 9 carbon cycle models are equally likely and sampling the full uncertainty space (and there are good reasons, by the way, not to make that assumption), then you can interpret your outcome as a probabilistic distribution of expected future climate change. Otherwise, simply call it what it is: a "multi-model ensemble" without assigning this statistical property to it.

Step 3: View Climate output

This last panel provides you with the outputs for your simulations.

Select the scenarios to plot

On the last panel, you find on the left hand side a list with all the simulations that you ran. For viewing these, choose one or multiple scenario runs that you did, whether they were using default climate settings or a probabilistic setup.

Note: If you cannot select your scenario and a little wheel is spinning to the right of it, then the runs are still to be performed (grey wheel) or currently running in the background (blue wheel). Once they are ready, the runs will automatically be selectable for plotting.

Select the output variable

Select the climate variable that you would like to plot. For example, global-mean surface air temperatures, CO2 concentrations, radiative forcing etc. that you would like to plot.

Note: Depending on whether your runs were performed in the "standard" setup or "probabilistic" setup, you will either see either a single line per scenario or colored ranges. The colored ranges for the "probabilistic runs" denote the distribution of all the 171 or 600 runs that were performed with dark shading denoting the 50% percentile region and light shading the 66% percentile region (from 17% to 83%)