initial commit of meps updraft helicity

[tackandr]

Code review for lua script related to updraft helicity. This is somewhat unconventional so I wanted to have some review on it. It's on the edge of the lua limitations but I wanted to give it a try this way first as it is quick to write but if needed I could do it as a full plugin as well.

[mpartio]

I think I got the basic idea but the actual definition of updraft helicity is still a bit of a mystery to me. Could you provide some documentation for example a link to a page where UH is explained?

Some variables are not set, like dx and mode. Also you have hard coded grid properties in the file but I guess those are only for testing.

The code itself is quite easy to follow. How is the performance? I would guess running a luatool over all hybrid levels isn't exactly fast? If the performance is sufficient, I think this lua script is nice and short. If the performance is bad, maybe some could could be implemented in c++ and/or cuda?

[tackandr]

I think this paper describes the idea in detail
https://www.spc.noaa.gov/publications/milne/vert-uh.pdf

Missing variables got eaten by the copycat. But yes, probably need to be read from the configuration in the production version.

Regarding performance:
Configuration that precomputes hybrid levels 21-40 and then calls the hitool took 3500ms on pansuola for a single time step and about 2min for a 67 step forecast.

[mpartio]

Thanks, I think I got the picture what UH is about. I think this Lua script is good because even I could understand it 😄

With those timings I don't think performance will be an issue. Production machines are probably also a bit faster than pansuola. And if it turns out to be too slow, some parts can be moved to c++ (or the whole code).

👍