Measurement of any type of existing ecological interaction that might help to understand changes of the system with respect to time i.e. evolution of the system.
better not to classify factors like biotic & abiotic, because cannot differentiate carefully. Hot of body, or wet (liquids) from organism are biotic or abiotic? Poison (toxins, venom) biotic or abiotic ? These all forms of co-actions that are the object of biocoenology.
So, you should determine inner and outer factors. I’m explaining my students on examples – you boy if you are feeling hot from pretty girl, is it biotic factor, or oxygen that is producing by plants? And sulfur producing by some bottom microbiota?
But when the dean is entering our lecture room he is outer influence against our temporary community, but inner for the faculty.
If you have to do with saturated community, inner relations would be much more intensive. To identify acting factors you should use traditional statistical and multydimensional method, but you should know that regression or correlation existance do not mean relations between species. But can use for some purposes (look fish productivity relation with water transparency that is dependent on the development of algae !!! - can find on my RG site). To have real working models of relations should right differential equations.