Hi Sina,

There seems to be a few packages in R that will directly calculate the SPI including

SPI: https://cran.r-project.org/web/packages/spi/spi.pdf

and

SPEI: https://cran.r-project.org/web/packages/SPEI/SPEI.pdf

A quick look at the SPI package doesn't seem to specify how the SPI is calculated but the SPEI package function spi() by default uses the gamma distribution.

This post: https://www.r-bloggers.com/visualizing-drought/

has code with a bit of explaination that you may find helpful, but if your more interested in just getting to the spi, then I'd stick with the SPEI package.

If your more interested in fitting a gamma distribution, you may want google fitting distributions with R. The MASS package has a function that will try to fit a gamma distribution.

I haven't tested any of these packages so yu may want to test them agains data you know the spi for.

Hope this helps.

Cheers,

Justin

Hi Sina,

I would use the R package "fitdistrplus". This package has a series of functions to diagnose a given distribution, fit a series of theoretical distributions and then test their goodness of fit through different tests to make a final decision.

In practice, it is as simple as fitting the theoretical distribution to which your data seem to fit best. Let´s say that your data are stored in X.

You can carry out the diagnostics by typing:

plotdist(X, hist=T, demp=T)

descdist(X, boot=1000)

To fit the distribution:

x.gamma