[ncl-talk] Segmentation fault
Dennis Shea
shea at ucar.edu
Sat May 23 13:09:07 MDT 2015
[1] If your data has missing values, you can not use either specx_anal or
the wavelet tool.
[2] http://www.ncl.ucar.edu/Document/Functions/Built-in/dtrend.shtml
states: "Missing values (_FillValue) are not allowed."
[3] If there are **isolated** missing values you could likely do a simple
linear interpolation
to fill them. However, **any** interpolation will affect the spectra.
If you have many _FillValue .... I would not even try to fill them in.
The resulting spectra will reflect the effects of the interpolation as
much as the 'rea;' data.
---
Punch line: I don't think you can get what you want ... given your data.
--
Your statement:
"My script is from this link:
http://paos.colorado.edu/research/wavelets/
There is no NCL script at that web site.
In fact, NCL call the fortran code presented at the website.
---
The interactive code there is IDL ... not NCL
Regards
On Sat, May 23, 2015 at 9:54 AM, Amadou Coulibaly <mpapin24 at gmail.com>
wrote:
> Hi Dennis,
>
> My script is from this link:
> http://paos.colorado.edu/research/wavelets/
>
> My Temp look like this:
>
> Variable: Temp
> Type: float
> Total Size: 1051968 bytes
> 262992 values
> Number of Dimensions: 1
> Dimensions and sizes: [time | 262992]
> Coordinates:
> time: [410227200..1356994800]
> Number Of Attributes: 7
> coordinates : lon lat
> grid_type : unstructured
> comment : Air temperature is the bulk temperature of the air, not the
> surf
> ace (skin) temperature.
> shortname : ta
> code : 11
> units : K
> _FillValue : 9.96921e+36
> (0) 299.15
> (1) 9.96921e+36
> (2) 9.96921e+36
> (3) 296.35
>
> It contains a lot of missing data. That's why I used the following command:
> ; Remove mean and linear trend (recommended, not required)
> ; Missing values are not allowed.
> ;************************************
> Temp = dtrend(Temp,False)
>
> After that I printed Temp, but look stange for me with negative values:
>
> Variable: Temp
> Type: float
> Total Size: 1051968 bytes
> 262992 values
> Number of Dimensions: 1
> Dimensions and sizes: [time | 262992]
> Coordinates:
> time: [410227200..1356994800]
> Number Of Attributes: 7
> coordinates : lon lat
> grid_type : unstructured
> comment : Air temperature is the bulk temperature of the air, not the
> surf
> ace (skin) temperature.
> shortname : ta
> code : 11
> units : K
> _FillValue : 9.96921e+36
> (0) -8.256236e+36
> (1) 1.713005e+36
> (2) 1.713035e+36
> (3) -8.256145e+36
>
> I think the problem is from that command trying to remove mean and linear
> trend. Because without that I can get the wavelet function (w), but it
> contains only zero as values.
>
> How to handle this issue?
>
> Best
>
>
> On 23 May 2015 at 14:17, Dennis Shea <shea at ucar.edu> wrote:
>
>> [1] I have no idea why the script should seg fault. Maybe your parameter
>> settings are not appropriate for the input data.
>>
>> [2] What does "Temp" look like? printVarSummary(Temp)
>>
>> [3] Are there missing values?
>>
>> [4] If you want the "power spectrum", why not look at
>>
>> http://www.ncl.ucar.edu/Document/Functions/Built-in/specx_anal.shtml
>> http://www.ncl.ucar.edu/Applications/spec.shtml
>>
>> Good luck
>>
>> On Sat, May 23, 2015 at 8:00 AM, Amadou Coulibaly <mpapin24 at gmail.com>
>> wrote:
>>
>>> Dear ncl users,
>>>
>>> When I tried ti run my script (wavelet1.ncl), I am getting a strange
>>> error which is:
>>> Segmentation fault (core dumped)
>>>
>>> I am using ncl.6.2.1 and my system is: Linux amadou-Satellite-C855
>>> 3.19.0-16-generic #16-Ubuntu SMP Thu Apr 30 16:09:58 UTC 2015 x86_64 x86_64
>>> x86_64 GNU/Linux
>>>
>>> My dataset contains hourly data (1983-2012) of wind(speed/direction),
>>> air_temperature, ... for 46 sites. I just wanted to plot the power spectrum
>>> of only one site and for the all 30 years. I think the error comes from
>>> line 45 when I tried to compute the wavelet function (w). Please find
>>> attached my script.
>>>
>>> How can I solve this issue?
>>>
>>>
>>> --
>>> * COULIBALY AMADOU *
>>> PhD Student on West African Climate System (WACS)
>>> FUTA - Federal University of Technology of Akure, Nigeria
>>>
>>> *Visiting Student - *
>>> *University of Cologne, Germany**Institute of Geophysics and
>>> Meteorology*
>>> Pohligstr. 3 / Office 3.102
>>> D-50969 Köln
>>>
>>> *Project*: WASCAL (West African Science Service Centre on Climate Change
>>> and Adapted Land Use)
>>> Phone:(+234) 810 795 2836 /(+223) 65 67 27 57 /(+226) 64 57 37 27 /+49
>>> 15218352574
>>>
>>> E-mail: mpapin24 at gmail.com / coulibalya68 at yahoo.com
>>>
>>> "*The time is always right to do right": Nelson Mandela*
>>>
>>> *"Character is like a tree and reputation like a shadow. The shadow is
>>> what we think of it; the tree is the real thing" : Abraham Lincoln*
>>>
>>> *"Do what you can, with what you have, where you are" Theodore Roosevelt*
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>>
>>
>
>
> --
> * COULIBALY AMADOU *
> PhD Student on West African Climate System (WACS)
> FUTA - Federal University of Technology of Akure, Nigeria
>
> *Visiting Student - *
> *University of Cologne, Germany**Institute of Geophysics and Meteorology*
> Pohligstr. 3 / Office 3.102
> D-50969 Köln
>
> *Project*: WASCAL (West African Science Service Centre on Climate Change
> and Adapted Land Use)
> Phone:(+234) 810 795 2836 /(+223) 65 67 27 57 /(+226) 64 57 37 27 /+49
> 15218352574
>
> E-mail: mpapin24 at gmail.com / coulibalya68 at yahoo.com
>
> "*The time is always right to do right": Nelson Mandela*
>
> *"Character is like a tree and reputation like a shadow. The shadow is
> what we think of it; the tree is the real thing" : Abraham Lincoln*
>
> *"Do what you can, with what you have, where you are" Theodore Roosevelt*
>
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150523/c9e2014b/attachment.html
More information about the ncl-talk
mailing list