[Met_help] Information about Spearmnn rank correlation

[Tressa Fowler]

Francesco,

Thank you for calling our attention to this error in our  
documentation. The omission of the 1- piece is a typo.

However, it turns out that the actual formula used for the calculation  
is not this one, as it fails to handle ties in the ranks. To handle  
ties, we use the formula for the Pearson correlation, but with the  
ranks as input rather than the raw data. The most accessible reference  
for this is on Wikipedia ( we use the second formula listed ).

http://en.wikipedia.org/wiki/Spearman's_rank_correlation_coefficient

We will update the documentation to reflect the correct formula for  
this statistic.

Thanks much,

Tressa


On Apr 20, 2009, at 4:55 AM, Francesco Piani wrote:

> Good morning.
> I'm a weather forecaster working for Tuscany Region, Italy. We use  
> WRF-NMM as an operational model and MET to verify its results. In  
> the MET user guide, appendix C, the Spearman rank correlation is  
> indicated as 6/n(n2-1)*... instead of 1-6/n(n2-1)*..., as indicated  
> in many articles (Schultz, 1995, An explicit cloud physics  
> parameterization for operational numerical weather prediction, Mon.  
> Wea. Rev., 123, 3331-3343).
> I would like what is the method used in MET to calculate this index,  
> in order to interpret statistical results in a reasonable way.
>
> Thanks in advance.
>
> Francesco Piani
>
> -------------------------------------------------------------
> Dott.Ing. Francesco Piani
> Weather Forecaster
> Tuscany Region - Hydrological Service - Centro Funzionale
> Laboratory for Meteorology and Environmental Modelling - National  
> Council of Research
> Via Madonna del Piano, 10 - Building D
> 50019 Sesto Fiorentino (Fi)
>  Italy
>
> Tel.  +39 055 44 830.23
> Fax  +39 055 44 40 83
> email: piani at lamma.rete.toscana.it
> email: francesco.piani at regione.toscana.it
> _______________________________________________
> Met_help mailing list
> Met_help at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/met_help

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/met_help/attachments/20090427/6ec54508/attachment.html