I have 64 years of monthly rainfall data for three stations out of which each station has a month of serially correlated data (Station A: Jan, Station B and C: Apr). I use the modified mann-kendall test package in R studio and use the syntax where x is time series and ci is confidence interval. When I run it for data of Station A for ci = 0.90, 0.95 and 0.99, I get different results (z stat, p-value etc.) for different ci. But in case of Station B, I get the same results for different ci. The z stat and p-values obtained are same at all confidence interval. This happens in case of Station C too. What is the issue here?