[ncl-talk] bandpass filter olr data expressed as pentads

[Dennis Shea]

The averaging process is a low-pass filter. A pentad average is a 5-day
average.  (I assume just a running average.) The frequency response at
f=0.2=1/5=(1/N=5) is zero. At frequencies 0.2 to 0.4, the response is
negative. Hence, I am not sure what you would be 'band-passing'.

I suggest you use the band-pass on the original daily data.



On Sun, Dec 25, 2016 at 11:04 PM, Lyndon Mark Olaguera <
olagueralyndonmark429 at gmail.com> wrote:

> Dear all,
>
> I have daily interpolated olr data from NOAA (https://www.esrl.noaa.gov/
> psd/data/gridded/). I already expressed these data by pentads (73
> timesteps). I wanted to do a 20-72 days bandpass filter (butterworth).
>
> I am confused as to the correct cutoff frequencies should I use and the
> idea behind this.
>
> The time step of the data sets is 1 (1 to 73). Is this the correct way of
> doing this?
>
>
> *fca = dt/t1 = 1./72.   (low frequency cutoff)
> fcb = dt/t2 = 1./20.   (high frequency cutoff)*
>
>
> Or this:
>
>
> *fca = dt/t1 = 5./72. (low frequency cutoff)*
>
> *fcb = dt/t2 = 5./20. (high frequency cutoff)*
>
> I hope anyone can help me.
>
>
> Many thanks in advance,
>
> Lyndon
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20161226/d9ae3b47/attachment.html