Getting extreme values for SPI gamma mle calculation

[ThomasOtt314]

Hi, I was working on calculating SPI using the gamma mle distribution and noticed some extreme values in the example data when I filtered for only data before 2011-01-01. I did this to have each year be a full year since the Wichita dataset cuts off in October 2011.

There are some pretty extreme values when using the gamma mle distribution. I also tried the gamma lmom distribution which didn't show those extreme values.

I also looked at the difference between them which showed April consistently opposite signed from each other.

Any help or guidance on this would be greatly appreciated!