It would help if you stated what sort of region you were dealing with. With three samples, you will  have a starting point for comparison of the results. How representative they will be depends on the variability of the each environment, and whether you have looked at the full range of micro-environments present. By comparing the results, you will get an idea of the overlap in the species of micro-organisms, but only by adding new sites and seeing whether the degree the past results match your new data can you obtain a reasonable idea of whether you have adequately sampled that area. Remember that there are a great number of variables involved including past history, substrates present, opportunities for contamination, etc..

Stuart.

While studying bacterial diversity, you may not get diversity by selecting sites at equal distance (100 m) or getting three samples from each site unless you are studying in detail a particular site. You have to select variants with change in habitat conditions such as pH, soil conditions, or other ecosystem variables.

The objective of the study, variations in the ecosystem characteristics, vegetation, contaminants etc are important considerations.

Consider the potential complexity of the situation.  Some bacterial sources such as wildlife, cattle or pets are intermittent or increased with stormwater.  One of my co-authored reports deals with coliform in Stekoa Creek drainage (about 1998).  Perhaps it will give you some ideas as well as attached collection of papers on drinking water quality from forests and grasslands. Variability may be high on abundance, but unsure if the suggested sampling might give an idea of presence, absence considerations for a specific flow or wetness regime.

Thank you sir William F. Hansen