This issue is meant for a discussion on how to represent model variation.
When showing climate change projections, the ensemble mean isn't really enough. it's pretty important to show the variation between models in the ensemble, as well as the range of interannual variability in the model projections. We toyed with this during the beta version development, but the "spaghetti plots" we came up with were too messy:
Here's a different way to show the range of ensemble variation. the light-colored full range is the minimum and maximum model projection in each year. the medium colored range covers 4 of the 6 models, and the dark range covers 2 of the 6 models. this graph is pretty easy to make in R using polygons, but perhaps is harder to do for the web?
can you show 2 RCPs using this approach? yes, though some would argue the graph gets too busy. if the user has the option to turn the second RCP on or off, then this approach is probably ok.
The same approach could be used in the scatterplot.
This issue is meant for a discussion on how to represent model variation.
When showing climate change projections, the ensemble mean isn't really enough. it's pretty important to show the variation between models in the ensemble, as well as the range of interannual variability in the model projections. We toyed with this during the beta version development, but the "spaghetti plots" we came up with were too messy:
Here's a different way to show the range of ensemble variation. the light-colored full range is the minimum and maximum model projection in each year. the medium colored range covers 4 of the 6 models, and the dark range covers 2 of the 6 models. this graph is pretty easy to make in R using polygons, but perhaps is harder to do for the web?
can you show 2 RCPs using this approach? yes, though some would argue the graph gets too busy. if the user has the option to turn the second RCP on or off, then this approach is probably ok.
The same approach could be used in the scatterplot.
what do you guys think?