# [ncl-talk] Partial derivative along a coordinate

Dennis Shea shea at ucar.edu
Mon Apr 17 08:59:14 MDT 2017

```Please look at the documentation for:

http://www.ncl.ucar.edu/Document/Functions/Built-in/center_finite_diff_n.shtml

Cheers

On Mon, Apr 17, 2017 at 8:54 AM, gibies george <gibiesgeorge at gmail.com>
wrote:

>
> But the SST case I discussed is just an example.
>
> My actual problem is to get zonal, meridional and vertical gradient of
> ocean temperature profile (4D data with time).
>
> I also have to get time tendency of temperature.
>
> That is why I put a generalized mail initially.
>
> Thank and regards
>
> On Apr 17, 2017 6:46 PM, "Dennis Shea" <shea at ucar.edu> wrote:
>
>> NCL has functions to calculate gradients (derivatives).
>>
>> Your original post did not show that  you wanted gradients of SST.
>> Presumably, these have missing values (land).
>>  *ftcurvd *will compute gradients but, as noted in the documentation, it
>> does not allow missing values.
>>
>> NCL has other functions for gradients:
>>
>>
>> Some use spherical harmonics which are "highly accurate"  but, again, do
>> not allow for missing values.
>>
>> --
>> The following 6.4.0 function does allow for missing values.
>>
>> latlon_cfd.shtml
>>
>> Read the documentation. This returns a variable of type 'list.' The list
>> (container) consists of two variables the zonal and meridional gradients.
>> ---
>> The following is essentially Example 2:
>>
>>    sst  = f->SST                        ; assume a rectilinear grid
>>
>>
>> ; for clarity, explicitly extract the variables within the list
>>
>> accessing elements of a list
>>
>>    print("---")
>>
>>
>> Good luck
>>
>> On Mon, Apr 17, 2017 at 4:01 AM, gibies george <gibies at tropmet.res.in>
>> wrote:
>>
>>>
>>> Suppose if I have to calculate zonal gradient of SST(time|:,lat|:,lon|:)
>>>
>>> xcoord=confirm(*dimsizes*
>>> <https://www.ncl.ucar.edu/Document/Functions/Built-in/dimsizes.shtml>
>>> (SST),SST&lon,2)
>>> dsst_dx = *ftcurvd*(xcoord,SST,SST&lon)
>>>
>>> Is this correct ?
>>>
>>>
>>>
>>>
>>> *Dr. Gibies George,*
>>> *Project Scientist C,*
>>> *Monsoon Mission,*
>>>
>>> *Indian Institute of Tropical Meteorology,Dr. Homi Bhabha Road, *
>>> *NCL Post Office, *
>>>
>>>
>>>
>>>
>>>
>>>
>>> *Pashan, Pune 411008, India.Office Teleph  : +912025904423
>>> <+91%2020%202590%204423>Office mail      : gibies at tropmet.res.in
>>> <gibies at tropmet.res.in>____________________________________________Vettikuzhiyil
>>>
>>>
>>>
>>> *Puthencruz-682308, Kochi, India.Mobile number :  +919028523910
>>> <+91%2090285%2023910>, +919420407392 <+91%2094204%2007392>, Personal mail :
>>> gibiesgeorge at gmail.com
>>>
>>>
>>> print this e-mail unless you really need. Because every 3000 pages cost a
>>> tree. *
>>> *Save Trees, Save mother Earth." *
>>>
>>>
>>> On 13 April 2017 at 20:31, Andy <awmiller at mit.edu> wrote:
>>>
>>>> Yes, you can use ftcurvd.
>>>>
>>>> Andy
>>>>
>>>> Andy Miller
>>>> MIT PAOC PhD Candidate
>>>>
>>>> 77 Massachusetts Ave | 54-1715
>>>> Cambridge, MA 02139
>>>>
>>>> Phone: 617-253-3715 <(617)%20253-3715>
>>>> Email: awmiller at mit.edu
>>>>
>>>> On 04/13/2017 02:46 AM, gibies george wrote:
>>>>
>>>> Dear friends,
>>>>
>>>> Is there a function available to find partial derivative of a variable
>>>> along one of its coordinate (say, lonx, laty, levz, time, etc ) ?
>>>>
>>>> thanks and regards
>>>>
>>>>
>>>> *Dr. Gibies George,*
>>>> *Project Scientist C,*
>>>> *Monsoon Mission,*
>>>>
>>>> *Indian Institute of Tropical Meteorology, Dr. Homi Bhabha Road, *
>>>> *NCL Post Office, *
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> *Pashan, Pune 411008, India. Office Teleph  : +912025904423
>>>> <020%202590%204423> Office mail      : gibies at tropmet.res.in
>>>> <gibies at tropmet.res.in> ____________________________________________
>>>> Vettikuzhiyil House, Ezhakkaranadu South **Post Office,*
>>>>
>>>>
>>>>
>>>> * Puthencruz-682308, Kochi, India. Mobile number :  +919028523910
>>>> <090285%2023910>, +919420407392 <094204%2007392>,  Personal mail :
>>>> gibiesgeorge at gmail.com <gibiesgeorge at gmail.com>
>>>>
>>>>
>>>> print this e-mail unless you really need. Because every 3000 pages cost a
>>>> tree. *
>>>> *Save Trees, Save mother Earth." *
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing listncl-talk at ucar.edu
>>>> List instructions, subscriber options, unsubscribe:http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> ncl-talk at ucar.edu
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>>>
>>>
>>> _______________________________________________
>>> 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/20170417/ecf4b0d9/attachment.html
```