[ncl-talk] SPEI (drought index) calculation...

[Dennis Shea]

Hello Qudsia

re: "... if there is a function that transforms from cdfgam_p to the
standard normal distribution. As you have asked 'gmm_transform' is supposed
to acquire the necessary transformation to  'standard normal distribution'.
If you could have a look."

I could not find any function to transform from cdfgam_p to the standard
normal distribution.  Sorry.

On Thu, Apr 17, 2025 at 11:59 PM qudsia zafar <missquaddus at gmail.com> wrote:

> Hi Dennis,
>
> As I couldn't find the direct function for spei calculation just like
> there is for spi in NCL therefore i was trying to do things manually.
>
> Till now I have used the  'dim_gamfit_n' function output (shape and scale
> parameters)  as an input in the *cdfgam_p
> <https://www.ncl.ucar.edu/Document/Functions/Built-in/cdfgam_p.shtml>  *function.
> Now the probabilities (gridded data) acquired have to be transformed into
> the 'standard normal distribution' resulting in the spei values with a mean
> of 0 and standard deviation of 1.
>
> An spei of 0 indicates near normal conditions, positive values indicate
> wetter than average conditions and negative values are drier than average
> conditions.
>
> I was wondering if there is a function that transforms from cdfgam_p to
> the standard normal distribution. As you have asked 'gmm_transform' is
> supposed to acquire the necessary transformation to  'standard normal
> distribution'. If you could have a look.
>
> Thx.
>
> qudsia
>
>
>
>    -
>
>
>
> On Fri, 18 Apr 2025 at 06:41, Dennis Shea <shea at ucar.edu> wrote:
>
>> NCL's *cdfgam_p
>> <https://www.ncl.ucar.edu/Document/Functions/Built-in/cdfgam_p.shtml> *documentation
>> mentions 'gamma' functions from R, Matlab and Excel.
>>
>> What is your 'gmm_transform' supposed to return?
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On Thu, Apr 17, 2025 at 2:32 AM qudsia zafar via ncl-talk <
>> ncl-talk at mailman.ucar.edu> wrote:
>>
>>> Hi i am using cru datasets (ts 4.08) for precipitation and potential
>>> evapotranspiration. I have done (precip_minus_pet = prc - pet) and then
>>> I have used the function 'dim_gamfit_n' to estimate shape and scale
>>> parameters for the gamma distribution fitting.
>>>
>>> However in the next step of spei calculation I need to do the following:
>>>
>>> spei = gmm_transform(precip_minus_pet, shape, scale)
>>>
>>> I have now all the three data (precip_minus_pet, shape, scale) but i
>>> don't know if NCL has a gamma transformation function available? I have
>>> seen 'cdfgam_p' but I don't know if this corresponds to the gamma
>>> transform?. Kindly advise which function can be used in the final step of
>>> spei calculation. Many Thx..
>>>
>>> qudsia
>>>
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> ncl-talk at mailman.ucar.edu
>>> List instructions, subscriber options, unsubscribe:
>>> https://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.ucar.edu/pipermail/ncl-talk/attachments/20250425/46d22a8f/attachment.htm>