Dear All,

i have a 50 years rainfall data, i would like know a proper method to find the maximum probable rainfal. Past data verying randomly, there is no uniform variation in a perticular month in each year.

Well I always tend to opt for the simpler approach: did you try to calculate the empiric cumulative distribution of your 50 years data?

Cf. for instance http://homepages.cae.wisc.edu/~ie642/content/Techniques/EmpiricalDistributions/empirica.htm on how to do it.

This will only calculate the percentage of occurence of the maximum observed rainfall in your 50 year data, it won't extrapolate to a more extreme probability (such as 1 in 100 years, 1 in 1000 years), for this you'd have to use a statistical model.