another way of putting it, if we look at the DMI 80N the eyeball says the green line starts turning up about day 70
Nice method. Solar insolation is symmetric about the summer solstice but the 80+N graph is not, so it does not pick up onset of waning solar heat in the fall (Neven says forget it from late August on). Around here, it is not unusual when the sun is low for it to get under the cloud deck. The land can take up ~100% of the heat despite 90% cloud cover. However in the Arctic, calm water has a critical angle of ~22º (ice would be different). Below that angle, it's all reflected. But where is the final destination -- absorption by our favorite vaporous greenhouse gas or in a cloud or back out to space? Meanwhile SHEBS and N-ICE2015 were there actually measuring it all.
turn out gradient-delineated graphs of things like temperature that provide visual clues to the error
@ZLabe had some great examples of bad graphs on twitter the other day, both denier-type and bad-scientific. Most of these can be fixed by cropping out unused space and rescaling of uncoupled vertical and horizontal. We don't want to get into the miscommunication business here, deliberate (not even for a good cause) or unintentional.
On 2D and 2D+1T displays, ie maps and animations, the most common communication issue is a muddy palette arisng from lossy compression, resizing in bezier spline mode, re-projecting, or anti-aliasing text and legends.
Those can be fixed by passing from RGB to indexed color which crams all tens of thousands of colors into a dozen or two bins. Returning to RGB, you will find the map contoured without contour line dividers and easily countable map pixels corresponding to each palette bin, 2nd animation shows before and after.
A good place to do this, with sprites shut off, is at nullschool. The palettes there have to have global range which means for the Arctic, the palette spread is too small.
The only difference with time series is that all the frames have to be tiled up into a single image which in effect allows, after re-slicing, operations to be applied to each frame.