Re: [ferret_users] Correlation space-point

["William S. Kessler" <william.s.kessler@xxxxxxxx>]

This should work fine: 

let p=precip[d=1,x=110e:145e@ave,y=22n:45n@ave]   ! p varies in time only
let q=ua[d=2,x=110e:145e,y=22n:45n]                         ! q varies in (x,y,t)

==>> correl [defined by go variance] will be an (x,y) field that can be contoured.

BK

> On Feb 5, 2015, at 11:35 PM, Nicolas Freychet <nicolas.freychet@xxxxxxxxx> wrote:
> 
> Dear Ferret users,
> 
> I am trying to compute a correlation between mean precipitation over a region and winds over a larger scale region. What I want to obtain actually is a correlation between [each grid point of the wind data] and [the precipitation of one special location]
> 
> So I tried something like:
> 
> let p=precip[d=1,x=110e:145e@ave,y=22n:45n@ave]
> let q=ua[d=2,x=110e:145e,y=22n:45n]
> go variance
> 
> But obviously, because there is a 2D variable and a 3D variable, Ferret cannot compute it directly. Do you have any suggestion?
> 
> (I'm thinking maybe to create a file on the same grid as Wind, and just copy the Precip on each grid point, and the just compute the spatial correlation over space, but maybe there is a more simple solution?)
> 
> Best,
> Nicolas
> 
> 
> 
> -- 
> ----------------------------------------------------
> Nicolas Freychet
> Postdoc Reseach Fellow at RCEC
> Academia Sinica
> ----------------------------------------------------