I would suggest that you rather conduct multivariate statistical analysis for your study.

Sincerely

Well,

In general yes you can; in details/ multiple and also different you need as Romain answer, it is necessary to use statistical analysis especially correlation.

Sincerely

Belkacem Bekkoussa using average concentrations is not a good idea. For this purpose, we should take help from statistical methods and burrow them to do multivariate statistical analysis and perform correlations to know about the association of elements with other elements which also considers the environment/host as a key factor.

Thanks

Come on folks, think about this. Averages, means, and standard deviations only have significance when estimating the properties of a single normal population. By definition, "multiple and different campaigns" implies multiple populations. Secondly, groundwater concentration data tend to be lognormally distribute, so geometric means, etc. may be more appropriate after testing for normality. Maybe I don't understand the question.

Normally Clark values are taken as reference concentration for specific Geochemical Sampling media. You may accordingly define the normal population, Background, and Anomalies. There are several statistical methods to define the Anomalies and population distribution, standardization, and Uni-variate and multivariate statistical analysis.

No, if the parameters are variable and time-dependant.

NO

Global Geochemical Continuum Model you can follow

thank you for all for your contribution. maybe we should give some precision. We would like to carry out a complete hydrochemical study on a region and we have physico-chemical analysis carried out on boreholes with different dates for the same point (sometimes up to 20 years apart). it has been noticed that there not a very large variations in concentrations for the same element. we wondered if we could not work with the average of the concentrations of the boreholes in order to valorize all data.

I would be very grateful if you can tell me similar works. Sincerely and thank you again for all.

Terrific, now we know you are dealing with water wells and not monitoring wells. Interesting, your findings are the exception not the rule. Most water well samples collected over large periods show large variations in water composition due to vertical aquifer mixing dynamics arising from poor well construction practices. Accordingly, I will assume that the study wells in your region are reliably tapping a single confined aquifer. I will also assume that one of your central goals is to map chemical groundwater evolution from recharge to discharge over a substantial area. That being the case, mapping mean values for selected parameters may appropriate. You should still first test for normality and make the appropriate transformations as needed. I would first test and compare trends in conservative vs. reactive parameters (eg. chloride vs bicarbonate, or stable isotopes of water vs. stable isotopes of DIC). If you have sufficient temporal data from individual sites, you may find that measurements of variance quite instructive. These are the simplest parameters to examine first before continuing to more complicated factor/cluster type analyses. While you are at it, take a look at the work of Craig Bethke, Joseph Toth, Francis Chapelle and Derek Loveley for different approaches to examining regional GW transport in confined aquifers.

thank you so much Anthony W Gorody. the explanations you have provided are very instructive and very relevant.

Sincerly.