Check the longitudes when loading data
After some discussion with @cprodhomme and @rfernand, I open this discussion to point out one issue that I found when loading data with Load with specific longitudes. I have found that most of the issues that we had in the past long time ago in IC3) have been solved. I did some tests playing with the longitudes with one CNRM-CM5 experiment. Load provide consistent results when playing with negative or positive longitudes, crossing or not the Greenwich meridian. Except in one configuration when I use (-180;+180), where I found a bug:
(0;360) or (-180;+179.9) give correct and consistent results. That's not the case for (-180;+180) that cannot be used. You can see here the script and the data that I did to have a look at this issue: /esnas/scratch/mmenegoz/MORDICUS/36_members/ENSO/test_load/test_load_longitudes.R
And the plot that show the difference of temperature in the Tropics between data loaded with (0;360) and (-180;+180):Temperature_tropic_1991_diff2.eps
I think we should have to check also that there is no more issues when loading data from other models, where the longitudes and the latitudes are not ordered as in CNRM-CM5.