The problem you're thinking of is when you display two different kinds of data on the same axis and so there's two degrees of freedom in how you choose their re

The problem you're thinking of is when you display two different kinds of data on the same axis and so there's two degrees of freedom in how you choose their relative scaling and offset. This makes it very easy to show "obvious" relationships. But the data here all has a very natural sense in which they share the same axis: they're all temperatures of the world, or simulated values thereof. So the choice in scaling and offset makes little to no difference in how the data is perceived.