Re: mjo

Regarding the sign in EOF analyses, the point is that EOF analysis can tell where OLR values are varying out of phase -- having opposite signs -- the exact sign is irrelevant. You can simply multiply your results by -1. to change the sign if you so desire as you're not changing the nature of the covariance across your domain or between variables.

Maurice

----- Original Message -----
From: shinn wong <shinnshinnwong@gmail.com>
Date: Sunday, January 3, 2010 11:59 pm
Subject: Re: mjo
To: ncl-talk@ucar.edu

> Thanks. But I still feel a bit puzzled about it. So, if the sign of
> the EOF
> is arbitrary, does it mean that we cannot tell from the sign that
> where is
> the positive and negative OLR anomalies? If yes, how can i modify the
> code
> so that negative OLR anomalies can be distinguished from the positvie
> OLR
> anomalies? Thanks.
>
> 2010/1/4 <shea@ucar.edu>
>
> > The sign of the EOF is arbitrary.
> >
> > Good luck
> >
> > > Hi everyone,
> > > I have a problem concerning with the mjo. I have used the example
> > > mjoclivar_12.ncl<
> > >
> > > to
> > > compute the univariate eof for 20-100 day band-pass filtered olr
> data
> > > from
> > > 1987-2007, but the result seems to be different from that of the typical
> > > one. The conventional EOF1 exhibits negative OLR anomalies (strong
> > > convection) in the tropical Indian Ocean and positive anomalies in
> the
> > > western pacific (see figure 1 in
> > >
> >
> > ),
> > > but the result of mine is just the opposite. Do anyone know the reason?
> > > The
> > > code can be seen in the attachment. Thanks.
> > > shinn
> > > _______________________________________________
> > > ncl-talk mailing list
> > > List instructions, subscriber options, unsubscribe:
> > >
> > >
> >
> >
> >
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Jan 4 06:12:03 2010