Hi,
I am using XLSTAT to check the trend in precipitation and temperature series using Mann Kendall test. My only aim is to check whether a trend is present or not.Is it necessary to pre whiten the time series or I can use xlstat to run mann kendall test directly?
Dear Saima,
Pre-whitening is required if you need to investigate the correlation between two time series, when both include a trend: in that case the bulk of the correlation may be coming from the trends rather than from the statistical coincidence between fluctuations. The following web pages and documents may be useful in this context (among many others):
https://onlinecourses.science.psu.edu/stat510/node/75
http://www.astro.cornell.edu/~cordes/A6523/Prewhitening.pdf
https://www.stat.berkeley.edu/~bartlett/courses/153-fall2010/lectures/23.pdf
From the limited description of your project, I would not advise you to use this procedure: you should determine the eventual presence and strength of a trend in your (single) time series, and then assess its statistical significance.
Good luck! Michel.
- Badges
- Science topic
- Similar topics
- Geoscience
- Hydrology