With regards to freshwater systems (lakes) we mainly use a 1 L Niskin bottle to collect water samples for chemical (e.g., sulfate) and biological (e.g. microbial diversity) analysis. However, the sampling methods depends strongly on the system (groundwater; running water) and the compounds to be analyzed (e.g., gases, disolved ions).
If you are looking for fecal contamination, targeted sampling for suspected entry points or areas of potential human contact would be appropriate. If you are looking at natural bacterial production, random sampling might be appropriate horizontally, but vertical (water column) structure like a thermocline or hypolimnion should be taken into consideration. Actual sampling methods? Niskin type bottles for depth, some use pump systems, grab samples (bucket science) for the surface. Wear gloves to protect the sample from you. Repeat of Konstantinos's question: what are sampling for? that would determine how you sample.
Thank you everyone for your kind reply. What about for sampling ponds for isolation of organisms? do we still need random sampling for that? is it also a good idea to concentrate the the samples by centrifugation? thank you sirs have a nice day!
Isolation for biodiversity? then target as many different microhabitats as possible. if you are cultivating organisms on agar plates, you will have to dilute not concentrate. Autoclave some filtered pond water or a commercial spring water for a diluent. Make some standard media (it tends to be rich) for copiotrophs. Make some low nutrient medium for oligotrphs. Difco R2A was designed for oligotrphic groundwater bacteria, but even it may be too rich for some.
I can introduce you a better address to find your answer, in this address the standard operating procedures are mentioned , in which the laboratory methods are one of the parts that you can find your answer.