According to http://sac.csic.es/spei/home.html
to calculate SPEI we use precipitation and potential evapotranspiration (PET):
"the difference between the precipitation (P) and PET for the month i is calculated:
D_{i} = P_{i} - PET_{i}
which provides a simple measure of the water surplus or deficit for the analyzed month.
The calculated Di values are aggregated at different time scales"
and after this aggregation for N number of months and additional computations, SPEI for N months is calculated.
My questions are:
http://sac.csic.es/spei/home.html
Santiago Begueria and Sergio Vicente-Serrano of CSIC produced a R-project package with explatory cases. Different intervals produce different value of SPEI but often it is scaled to a sensitive portion of crop growth, https://cran.r-project.org/web/packages/SPEI/SPEI.pdf
Santiago Begueria and Sergio Vicente-Serrano of CSIC produced a R-project package with explatory cases. Different intervals produce different value of SPEI but often it is scaled to a sensitive portion of crop growth, https://cran.r-project.org/web/packages/SPEI/SPEI.pdf
Thanks Guido for nice attachment, I agree with you, rather the value outputs are a kind of moving average type, if we take 3 months interval say, the first value comes on the third month itself (1-3), 4th month correspond to (2-4 months) so on, and if crop growing season is june-august, we take value of 8th month, which covers 6-8,
if crop is of 4 months duration, june-sept, then we have to take 4 months interval for computation of SPEI, and accordingly we have to choose the value.
If one month before the season is rainfall period, that is imp in SPEI, accordingly we have to plan our interval
regards
Thank you for help, but I am not sure if I understands properly your answers.
Nevertheless, I checked "A Multiscalar Drought Index Sensitive to Global Warming..." article. It seems that SPEI is not a moving average type.
Correct me if I am wrong, but calculation algorithm for SPEI is as follows:
SPEI = W - \frac{C_{0} + C_{1}W + C_{2}W^{2}}{1 + d_{1}W + d_{2}W^{2} + d_{3}W^{3}} \\ \\
\text{where W = } \sqrt{-2ln(P)} \text{ for } P \leqslant 0.5 \\
\text{and P is the probability of exceeding a determined D value, P = 1 - F(x)} \\
\text{C are constant coefficients}
Article A Multiscalar Drought Index Sensitive to Global Warming: The...
Dear Mateusz:
I also meet some questions when i want to write a programme with Matlab about SPEI. My question is that in your 3rd step: fit log-logistic probability density function. Before you do this step, the time series of D is ordered from the smallest value to the biggest value???
In you step 3, the model parameter F(x) in Page 1705 of "A multiscalar drought ... index" can't be obtained in some cases. Therefore, i can't obtain the SPEI value.
But the programme is based on the above paper. I want to know whether you solve your question?
Thanks
Dear Zengyun Hu,
I'm usig an R! package to calculate SPEI. My question was about way of comparing different time-scale drought indices, not about implementation details.
I think that you can try to check code of the mentioned package to find a solution for your issue or ask another question on Research Gate
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Soil Science
- Soil Analysis