Examine wet bulb temperature implementation

[dopplershift]

There was a poster at AGU comparing our implementation (among others) of wet bulb temperature and comparing to Davies-Jones 2008 and a new "NEWT" implementation--both of the latter showing better error characteristics. At some point, it would be worth seeing if there's low hanging fruit to improve ours.

[alioacar]

Hi @dopplershift ,
I've been tinkering the wet bulb temperature implementation for a bit. Here is the comparison of MetPy and the other two (Davies-Jones(Colin Raymond's DJ08) and NEWT).

It seems like MetPy has a bias compared the other two while they have a good agreement with each other. I think the difference comes from the temperature dependence of the latent heat of vaporization.(it should decrease if T increases and increase if T decreases). So I directly implemented it (Lv0 - (cpv-cpl)*(T-T0)) into to the moist lapse rate calculation and now it's really close what DJ08 and NEWT produce. Note that the different scales for the figure above and below.

However most tests related to the pseudo-adiabats(moist lapse) have failed now. So I'm not entirely sure it is correct.

[dopplershift]

I'm not shocked that updating the formula causes test failures; you are producing different values, and thus plots.

[alioacar]

I'm asking because I have no experience in this field—are there any validated datasets or measurements against which these calculations can be compared? (Not only for WBT but also like LCL or LFC ...) Also, I'm not sure if this level of accuracy is even desired but if it's needed, I can put in some effort here with some metrics and validation plots maybe.

[dopplershift]

None that I'm aware of. You could try comparing curves to e.g. those from the National Weather Service, but it's usually really hard to match since you don't know what set of assumptions is going into those curves.