It depends, what question are you trying to answer? How is your data distributed? How many site did you collect data? How frequently did you sample (i.e. sample size)? What was collected? How was it collected?

If you are looking at if the "water quality" (i.e. TP, TN, Chl-a, etc.) increased or decreased over the six years you can run a trend analysis.

If you are comparing the sewage confluence zone to unimpacted (pristine) conditions, a simple comparison analysis could be run. Depending on the data distribution, parameteric or non-parametric statistics could be used. 

Hope this is a good start for you.

PJ

That depends on your Parameters an on the Quality of your data. If you send a variable list, and an example of data in an Excel-sheet I can look at it and make a proposal

[email protected]

Thank you to both Sir, Yeah I had worked on different festivals round the year, and in each festivals we collected the sample on pre-festive, on festive and post festive to find the impact and recovery.

In sewage we collected the sample from Sewage, confluence zone and dilution zone for different physico-chemical parameters.

Please GP through one of my paper published in Enviroemt International available on my profile page. You can find some other papers also on this page as I have also spent some time to work in this field. Regards.

Dear Sushil

You should have elaborated on your question. I just can refer you to the following papers. In both papers, you had better pay attention to the statistiacal algorithms for data nalysis I hope they could help.

http://pubs.acs.org/doi/abs/10.1021/jf504480f

http://www.sciencedirect.com/science/article/pii/S0963996913005723

Try to use BIOENV or BEST analysis in PRIMER package.

Big Thanks to all efforts...

The statistical method to be used is dependent on the purpose. If you want to simply know as to which are the primary constituents responsible for water quality, you can go for the PCA.

Dear Sushil:

I agree with those who pointed out, and inquired into, what you are trying to achieve.  What are you looking to assess from your festival samples?

Depending on what your response to the foregoing question is, you can utilize multivariate statistical techniques to determine how each variable impacts others; you may also need to determine if your sampling method was the best to utilize for what you are sampling for.

It would be very helpful for those on the list if you can detail or spell out what it is you are attempting to do.

Best regards,

Itzchak

I agree with Itzchak E. Kornfeld •when he pointed out the use of multivariate algorithms. Here you will look for the joint probability or conditional probability of having a certain value of any element given that other element have occurred. A previous step is to have the correlation matrix of the elements; this will give you an insight of what elements are correlated and how much. If you have the pdf of each element then you may generate stochastically the value of any correlated elements

Please contact me [email protected]. Dont miss please.

I have attached one book chapter in which review on water quality is given. Hope  it will help you.

Pawan.

Pawan:

This chapter is an excellent over view and explanation of the time series subject.

Itzchak

Time series is the best option for water monitoring data. Have you tryied to get an index?