Re: why low pass filter produced NaN

fca = 1/90 returns an integer 0

The documentation states:

fca - A scalar indicating the cut-off frequency of the ideal high
       or low-pass filter: (0.0 < fca < 0.5).

Use

fca = 1/90. or 1/0/90.0 or 1.0/90

On 10/14/13 9:46 AM, Xi Chang wrote:
> Hallo NCl,
>
> I have a daily data:
>
> Variable: data
> Type: float
> Total Size: 2012636160 bytes
> 503159040 values
> Number of Dimensions: 4
> Dimensions and sizes: [time | 20256] x [level | 23] x [lat | 15] x [lon |
> 72]
> Coordinates:
> time: [ 18..486474]
> level: [ 1..1000]
> lat: [ 20.. 90]
> lon: [ 0.. 355]
> Number Of Attributes: 2
> typeConversion_op_ncl : double converted to float
> table : 128
>
>
> I did a low pass filter as follow:
>
> ncl 38> nwt = 201
> ncl 39> fca = 1/90
> ncl 40> ihp = 0
> ncl 41> nsigma = 1.
> ncl 42> wgt = filwgts_lanczos (nwt, ihp, fca, -999., nsigma)
> ncl 43>
> ncl 44> filter = wgt_runave_n_Wrap ( data, wgt, 0, 0 )
> ncl 45> printVarSummary(filter)
> ncl 46> printMinMax(filter,True)
>
>
> why all the results are NaN
>
> *(0) min=-nan max=-nan*
>
> Thank you
>
>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Oct 14 10:00:13 2013