In analysis of hydro-meteorologic variables, many times we encounter count based series exceeding certain threshold value. For example, number of days in a year having rainfall more than 100 mm or number of days in a year having temperature more than 45 degrees Celsius, etc. While analyzing a fair length of such time series (say 30 years or more), we may encounter some years wherein the aforesaid thresholds might not have been exceeded. This may induce 'zero' values in the time series. Does presence of considerable zero values make widely used non-parametric trend analysis approaches such as Mann-Kendall and Sen's slope tests inappropriate for application to such data?

If so, kindly please suggest any suitable non-parametric trend analysis technique which is more appropriate in case the data contains considerable 'zero' values?

Thanks in advance!

Priyank Sharma

More Priyank J. Sharma's questions See All
Similar questions and discussions