The story behind NSF WQI is well explained in Larry Canter's book: Environmental Impact Assessment (1996), in chapter 5 about environmental indicators. Basically, a Delphi survey was conducted in 1970 by NSF. 142 experts were interviewed regarding parameters that could affect wáter quality. Through 3 questionnaires, experts concluded that only 9 parameters could defined water quality for surface water.
I am replying to your question assuming that you are interested in finding out different tools/techniques for selecting parameters to be included in an Environmental Indicator.
First step would be the literature review on the physics/mechanism of the phenomenon for which the indicator is being developed. Afterwards, the review of similar kinds of environmental indicators existing in literature would help you conceive a coarse picture of the procedure for refining the existing indicators or developing a new one.
Now from this two-stage review you would be able to arrive at a tentatively comprehensive list of parameters for the indicator. Then to compare parameters with each other, you can use various techniques such as Delphi Technique, Analytical Hierarchy Process and Ration-wise comparison.
Not necessary only 9 parameters for calculation of WQI. Its upon your research objective. you can take more the 9 parameters for the calculation of WQI.