Hey @stitch, sorry to be so neglectful of updates! I'm super busy with a research job at the New Zealand MetService right now, but that'll be over in 2 weeks and I plan to get back to some dev work. I can see the issue you mention when I use chrome, but hadn't noticed it before since I use firefox for dev. I'll see what I can do, gotta learn about cross-browser compatibility sometime :) I want to add an option to zoom the skewT near the surface too. Thanks for the heads up, I'll post here when I push the next version.
I use the sfc temp +3 degree for now as an approx - probably too much in the morning or when the cloud cover is important.
The surface temperature reported by the model does not seem to work better (I may want to give it an another try later).
Does it answer your question or is the question more about the temperature gradient ?