I have 9 samples of rainfall data, which all are average of time periods (columns) at given station (rows). Samples are all equal in number of variables. My question is - how to test, if any one of those 9 samples are statistically significant?
There are nine samples, each representing the average of decade (as in period of 10 days) over the timeframe of 1935-2015 (during June-August) in all the stations measured.
Like one sample is collection average of June's first decade (dates 1-10) in all stations between 1935 and 2015.
Are the samples from the same geographical area? (climate classification)
What is the distance between the stations?
What is your null assumption (H0)?
If it is from a different geographical area, there probably should be some change, so with a large enough sample size, you will be able to prove this change ...
In this case, you should be more interested in effect size and on the confidence interval.
Generally, if the null assumption is that the amount of rainfall is equal (the same geographical area) you can use the one way ANOVA and/or Tukey HSD.
Geographically - it's Estonia. Small enough country to not have very distinctive climatic differences, but I am rather interested to prove its difference on time scale (as in one decade has less rainfall than others in question)
I leave here one dummy example of dataset. In short, the aim is to determine, if one (or more) of those columns has big enough difference to prove it statistically...or which test to use for it?
EDIT 1: In addition question - which test would be good choice for testing if rainfall is dependant on altitude (which normally would be, but Estonia's overall altitude is fairly low, so just for fun testing it)?