# Plotting the wind direction in 3D?

How can I plot the wind direction in 3D, using the 3 wind vectors u, v and w? Normally, I create 2D wind rose plots in the traditional way, using R or MATLAB.

Which ways do you know for plotting this in 3D, using the additional w vector?
Do you have code examples or tips for packages?

• This is an excellent question. I don't know of any source that records "updrafts" and "downdrafts", but it would be great to find one.
– user967
Jan 21, 2016 at 17:16
• A relevant observation is that with the exception of thunderclouds, winds are almost exclusively horizontal. This is certainly entirely true for anything close to the ground. In those cases, you may simply not have to worry about the vertical component. Jan 25, 2016 at 1:19
• In polar and/or alpine environments, there are some local wind systems with important vertical components, such as the katabatic winds in the Antarctic.
– Arne
Jan 30, 2016 at 0:37
• it's really depend of OP's context: subsidience, thermals / thunderstorm / tropical storms, frontology, orography, all needs the vertical component :-) Feb 21, 2016 at 1:26
• How about a vector field plot? stackoverflow.com/questions/7130474/… Jul 12, 2016 at 16:24

The context I often see the u,v,w components shown in one plot are forward and backward trajectory plots like those from the NOAA HYSPLIT Model, which provides separate panels for horizontal and vertical movement. There is also 3-D imagery like what's available from UCAR in the cross sections examples for WRF, shown below. • it is better and simpler than VAPOR ! Great to know +1 Oct 17, 2016 at 4:28
• The rainbow colour scheme makes me cringe Mar 13, 2017 at 8:25

Try quiver3 function in Matlab. For example, this is the example in Matlab's website for this function: Try Mayavi - Mayavi 3D Python plotting. You can supply the u,v,w components of the wind vector as input and plot a 3D diagram How about using the same 2D wind rose plot for an additional figure, but color it with respect to velocity in the vertical? So you will have two of the similar looking wind rose plots, they will both show the directions in the horizontal 2D, but the second plot will have a colorbar that ranges ,say, from -1 m/s to 1m/s. I think that would be a very useful plot.

You will need to store the indices used in the calculation of the horizontal directions and calculate the histogram for the vertical flow that corresponds to them. Or just keep the part of the code where horizontal directions are calculated but use the vertical velocity to calculate the magnitude. Keep the sign to indicate upward/downward.