[ncl-talk] Wind vector question

[Leslie Hartten (CU)]

Hi Debasish,

   If you want to do statistics with vector wind properly, you will need 
to be using special functions.  This sub-specialty is called "circular 
statistics" or "directional statistics".  I have found N. I. Fisher's 
book (_Statistical Analysis of Circular Data_, Cambridge University 
Press, 1993) to be the most readable of the few that are out there.  It 
comes with lots of examples drawn from geology, meteorology, and animal 
movement studies, which helps make sense of the new term names and 
uses.  There are some errata collected over the last 30 years, and they 
can be found at https://valuemetrics.com.au/resources/#other  There is 
pretty solid MATLAB code available for circular statistics (see 
https://www.mathworks.com/matlabcentral/fileexchange/10676-circular-statistics-toolbox-directional-statistics), 
and there may also be some for other languages.

   I don't have access to my copy of the book right now, but as I recall 
RMSE is not actually an appropriate concept with directional data.  
However, Fisher's book can show you what would be an appropriate substitute.

   Note that I've got no "ties" to Fisher's work other than I've found 
it clear and useable.

Cheers,
   Leslie
--
Dr. Leslie M. Hartten, retired
formerly of
Univ. of Colorado's
   Cooperative Institute for Research in Environmental Sciences (CIRES)
at the NOAA Physical Sciences Laboratory
--
Disclaimer:  The contents of this message are mine personally and do not 
necessarily reflect any position of the University of Colorado.

On 9/22/23 9:09 PM, Debasish Hazra via ncl-talk wrote:
> Hi,
> Is there a way to plot wind vector wind RMSE instead of  U and V 
> separately using NCL. I have u and v wind from model and analysis.
>
> Thanks
> Debasish
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at mailman.ucar.edu
> List instructions, subscriber options, unsubscribe:
> https://mailman.ucar.edu/mailman/listinfo/ncl-talk