Generally, most of the homogeneity tests examine temporal variations in variance and average of a time series. So, in most of the cases, applying a statistical test such as Levene or Bartlett can be sufficient. For more details :

http://www.itl.nist.gov/div898/handbook/eda/section3/eda35a.htm

Hi Befikadu Esayas Amphune,

Before you start applying standard tests it is convenient to look at the data carefully, a simple scatter-plot of the data could reveal out-layers, recording errors, seasonal variations, change in instrument location, etc.

Then you can try to understand the source of these characteristics. In wind speed, for instance, records from stations include cyclonic, synoptic and stormy winds.

These winds are generated by different physical phenomena and they behave differently so to have a homogeneous data-set it is important to separate each type of wind before you start processing them, see for instance:

https://www.researchgate.net/publication/253712084_A_Statistical_Model_of_Severe_Winds

Hope it helps,

Augusto

Hi Befikadu,

I agree with previous discussion to look the data more carefuly one mor etime.

In case you already sure with the time series, in doing homogeneity test, I have experiences that may be good to share it here.

I did a homogeneity tests to detect data inhomogeneity in Jakarta's precipitation series by using four absolute homogeneity tests

(Wijngaard et al., 2003) including Pettitt’s test (1979); Alexandersson’s SNHT test (1986), The Buishand’s test (1982), and von Neumann’s ratio tests. From these test you may make a data classification as useful, potentially useful, suspect and reject. For further reading here is my published paper containing this homogeneity test.

http://onlinelibrary.wiley.com/doi/10.1002/joc.4548/suppinfo

or

Article Temperature, extreme precipitation, and diurnal rainfall cha...

Thank you so much for the suggestions and references. I will consider them for further check-up. As for the data quality, it

has been carried out in Climpact2 in R. It is after the data quality control that I wanted to know for my curiosity on the differences among the different homogenity tests for the same dataset in Xlstat.