You need to know what is important to you so you do not confuse statistical and ecological significance;
You need to know the effect size to determine the power of your design.
You must define at the beginning the environmental consequences that are of interest and the extent of variation that requires a response. For a complete analysis of possible errors of inference, the alternative hypothesis must indicate a treatment effect worth detecting.
It is useful to remember that "a difference, to be a difference, must make a difference." Clearly there are differences that are easy to detect, but
are very important ecologically, but not something you can do anything about (e.g. seasonal differences which are beyond your ability to change);
not relevant to the issue of concern (e.g. species variation not affected by the variable you are interested in).
Alternatively, a large enough survey can detect differences that are not of consequence and, therefore, not requiring intervention.
Without any indication of what is of concern to you as a manager it is difficult, if not impossible, to design or implement a study of sufficient relevance or power to detect something that is of interest. Neither approach can be justified scientifically, statistically or socially.