Re: [ferret_users] lagged spatial correlations

[William.S.Kessler@xxxxxxxx]

First of all, when you find such a result, it is always best to plot the data directly as a check. For example, pick a location where the lag correlation magnitude is high, and plot its time series against the ENSO time series. The lags should usually be visible (test it by shifting the time series with @SHF). Going back to the original data usually provides a resolution to confusion.

But in this case I think the answer is simple: you are misinterpreting the meaning of @SHF. When you compare, for example, ensots with ssts[l=@shf:-3], you are comparing present ENSO with SST from 3 month *previously*. That is, for this case, SST *leads* ENSO, not lags as you say. 

(The meaning of @SHF is frequently confusing! One way to remind yourself is to list, say ssts and ssts[l=@shf:-3] for a few timesteps.)

Billy K

----- Original Message -----
From: Fabian Lienert <cccma-student-003@xxxxxxxx>
Date: Wednesday, September 9, 2009 9:32 pm
Subject: [ferret_users] lagged spatial correlations
To: ferret <oar.pmel.ferret_users@xxxxxxxx>

> Hi Ferret crowd,
> 
> I am calculating lagged spatial correlations between a monthly ENSO time series and SSTs 
> (in space and time) using l=@shf:X and the go variance script.  
> From a lagged correlation analysis using a different software (R), I know 
> that in the tropical Indian Ocean (10S-10N, 60E-100E), mean SSTs are lagging 
> ENSO by 3 months when the lagged correlation is maximal.
> 
> Plotting the lagged spatial correlations in Ferret gives apparently the opposite result.
> 
> Case 1 when SSTs lead ENSO by 3 months. Correlation is max in Indian Ocean:
> let q = ensots; let p = ssts[l=@shf:3]; go variance; fill correl
> 
> Case 2 when SSTs lag ENSO by 3 months. Correlation is much smaller in 
> Indian Ocean than Case 1:
> let q = ensots; let p = ssts[l=@shf:-3]; go variance; fill correl
> 
> Anybody knows what I am doing wrong?
> 
> Many thanks,
> Fabian
> -- 
> Fabian Lienert                  |  Canadian Centre for Climate 
> Modelling and Analysis
> PhD Candidate                   |  Environment Canada
> phone: +1(250)721-7211          |  University of Victoria
> cccma-student-003@xxxxxxxx      |  P.O. Box 1700
>     |  Victoria, BC,  V8W 2Y2, Canada