About Lanczos filter in NCL

From: waqar younas <vickyqau_at_nyahnyahspammersnyahnyah>
Date: Thu Sep 06 2012 - 16:55:47 MDT

Hi guys,
I am trying to filter 20-90 days variabilities from the forecast data of
GEM model in NCL (version 6.1.0). This data consists of 90 days forecast. I
am trying to use the MJO-Clivar codes on forecast data (these codes are set
for observation which is continuous for year, 365 days). My main problem is
that if i want to filter the 20-90 days variabilities from 90 days, the
maximum period i can get is 45 days (statistically the maximum range is n/2
if you n days). But i found a paper, in which the author did the same thing
but using FORTRAN. He also mentioned that Fourier filter is applied to
these 90 days of data, which is wave number 1-5. Now i do not know how to
control these wave numbers in the filter? Moreover, the lanczos filter is
set for observed data which have 365 days so that wgt for lanczos filter is
201. When i use it for 90 days, it says the size of wgt cannot be greater
than 90 so i use 90 instead of 201. The filter still running but i do not
know what is significance of wgt = 90 instead of 201? Is it still retaining
the 20-90 days variabilities?
Thanks.

-- 
Best Regards
Waqar Younas
PhD Candidate
Natural Resources and Environmental Studies (NRES)
University of Northern British Columbia (UNBC)
Canada

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu Sep 6 16:55:57 2012

This archive was generated by hypermail 2.1.8 : Tue Sep 11 2012 - 15:30:42 MDT