You may use the variation of Standarized Precipitation Index (SPI), to assess the  variation of rainfall easily. The method is easy and often proven to be effective. Furthermore, it requires only precipitation data. In addition you could also use frequency analysis methods such as Gumbel frequency analysis or trend analysis techniques such as  Mann-Kendalman analysis. Wish you would be able to utilized these tools to fulfill your cause.

I am grateful to you, Lahiru Adayanga.

I have actually used trend and break point detection analysis, specifically, Mann-Kendall, Pettitt test and Hubert Segmentation test. No trend and break points were detected. However, I want to believe those were not enough to detect variations since trend is different from variation, in my view. I shall try other ones (SPI & Gumbel) as you suggested. Once again, I am grateful.

Some useful papers has been attached, all the best

Here you have two nice Investigators to ask, follow, and work...

https://www.researchgate.net/profile/Sara_Del_Rio2 

https://www.researchgate.net/profile/A_Penas

an extremely simple approach to test for an increasing (or decreasing) variance trend, would be to divide your data arbitrarily into 2 parts and use an F-test (ratio of variances). The ratio follows a chi-squared distribution. This ignores the serial dependence in your data but would provide an initial indication as to whether the variance in rainfall has increased or decreased over time. More complicated (but more appropriate) methods to assess how rainfall variance may be changing over time include ARIMA and GARCH time-series modelling approaches (e.g.. see Harvey, A.C. 1993. Time Series Models, 2nd ed. MIT Press

I am most grateful to you all for your invaluable ideas and concern.

You can refer to the suggestion by Ian D Jonsen. You can also examine the difference of CV (CV=σ/μ) between two equally long intervals of rainfall data series to assess change in rainfall variability...

Guoyu Ren, thank you very much for your wonderful idea.

Dear Andrew,

very well placed by colleagues some methodologies, here we are working with the uncertainty of the observations, for this, we use neural networks as a tool, where we identified which the spatial and temporal variability of rain on a particular station compared to other observational stations along a spatial distribution. It could be an alternative.

best,

David

I am very grateful to you, David Mendes, for you additional idea.

Hi

if ever you still are interested

here is a free software to detect changes in time series

http://khronostat.software.informer.com/1.0/

for free