Creating a circulation vector om vertical cross section.

[MUHAMMAD YUNUS BIN AHMAD MAZUKI]

Hello.

I would like to plot circulation vector, using u,v, and w. It will be something similar to Figure 1 in https://blogs.reading.ac.uk/weather-and-climate-at-reading/2018/how-the-hadley-cells-work/. The figure is as below.

There was a discussion about this in https://github.com/NCAR/wrf-python/issues/143. No solution was provided however.

If the desired circulation is along a longitude line, one would simply make a vector plot of w and u (for example Walker circulation). Similarly if the desired circulation is along a latitude line, one would simply make a vector plot of w and v (fro example Hadley Circulation). But what if the desired circulation vector is at angle?

I would like to approach this using wind-speed and wind direction.

I'm using the wrf-python tutorial data from https://github.com/NCAR/wrf_tutorial_data for this question. I have also attached an ipynb file showing the steps I have taken . There is a transect line showing cross-section I would like to plot.

Below is the plots using crossvert.

I would like to make a vector plot using w and wind speed after this. However wind speed is not a vector. I would like to turn it into a pseudo-vector. If you look at the cross section of wind direction, the left side is from one direction, and the right side is from other position. I would like to calculate the angles of the transect against the lat-lon at every point in the crossvert, to be used as a threshold. If the wind direction at a point in the cross vert is below the threshold, I would apply negative to the wind speed at that point. In opposite situation, I would apply positive. This would turn one side of the wind speed cross-section into positive and the other into negative. This, when vectored/quivered with w, will make a circulation vector.

My question is how do I calculate the angle of the transect line aginst lat/lon axis? There is a lat/lon coordinate information inside the crossvert which I believe should be helpful.

I'm actually working on South Pole projection data, so the threshold angle I'm looking for is different at each point in the transect line.

A simpler question would be how do I plot a circulation vector along a transect line using u,v, and w data. I searched around, and there are questions and discussions about is using wrf-python, but I could not find any solution posted.

Yunus.

[Marco Miani]

Hi Yunus,

I am very interested in this very topic, too. I had already tried extracting a vertical profile for scalar variables (dust, or temp) and that was no problem.

Now, when it comes to vector quantities (like wind), and the extraction profile is at an angle, things get a bit nasty.

I couldn't figure out yet if wind components extracted on the cross section plane are already projected on that plane,

or if, instead,

they are a "mere" interpolation at that height and position.

Also, I am trying to "play with input data" to shed some light on this and wipe away ambiguities. I have set up a little test bench, but couldn't get my hands on it just yet. I will follow this topic with great interest to see what comes out of it and, if possible, contribute with some ideas and findings.

Bye

Marco

[MUHAMMAD YUNUS BIN AHMAD MAZUKI]

Hello.

Marco,

From what I understand in discussion about vertcross is that, vertcross merely interpolates zonal-wind at the location along the transect line. The zonal and meridional wind values calculated in the vertcross is still pointing at the original direction.

All,

I have decided to manually rotates the ua, and va, denoted as ua' and va'

Below is the original ua and va on a pressure level.

And below is the rotated ua' and va'.

Below is the rotated vertcross

Below is an attemp at vertical circulation along the transect line.

Again, I have attached updated ipynb for this question. I have to confess, I'm not well-informed in hurricane, and thus I'm not certain whether my calculation is correct or not. I believe I have made sure that all zonal and meridional differentiation are properly taken care of. I will leave this here, and validate my methods with other simulation that is more one-sided wind flow, such as strong wind event.

Yunus