A common cause of error in water quality analysis is improper sampling. The results of a water quality analysis of a sample show only what is in the sample. For the results to be meaningful, the sample must contain essentially the same constituents as the body of water from which it was taken. If it does, then it is a representative sample.
For planning purposes, the definition of representative sample would be the smallest number of data points required to yield a statistically accurate description of the body of water being analysed.
Some general principles for collecting representative water samples include:
- Do not include large, non-homogeneous particles, such as leaves and detritus in the sample.
- If wading in flowing waters, place the sampling apparatus upstream of yourself to avoid contamination.
- Collect a sufficient volume to permit replicate analyses and quality control testing.
Collecting representative samples from flowing water
When sampling smaller streams, usually one sample per location is sufficient. For deep streams, a single depth-integrated sample is usually adequate, and for shallow streams, a single grab sample at the centre of flow is usually all that is required. If a stream is really wide, two or more grab samples could be taken and mixed together to form a composite.
Don't worry if you have to sample a large river, there are many methods to sample them; which one you choose depends on the river's depth, width and flow velocity. One common method used on large rivers is to collect a composite sample from a cross section of the river. Samples can be collected at 60% depth, or depth-integrated over a number of points along that cross section and pooled together. Generally the more points sampled along the cross section, the more representative the composite will be.
When taking more than one sample in flowing water always start sampling downstream and working your way upstream. This will reduce the likelihood of a contaminated sample.
Since the fluvial characteristics of a sampling location can change with the season, annual maximum and minimum flows and year-round accessibility should be considered when establishing a sampling station on a river or stream.
Unless you are particularly looking to assess the impact of a disturbance on the water quality, it is best to avoid sampling near:
- Effluent discharges or the entrance of tributaries
- Marked changes in characteristics of the stream channel, for example immediately downstream of a slump
- Artificial and physical structures such as dams, weirs, and wingwalls
Collecting representative samples from standing water
Larger water bodies like lakes and reservoirs present a challenging sample-point selection task. If the purpose of the sampling program is to determine the quality of the water leaving the lake or reservoir, a sample should be selected from the outlet of the water body. However, if the purpose of the sampling program is to determine the quality of the water body, then choosing sampling points become more difficult because of spatial variability.
Vertically, water in lakes and reservoirs is divided into limnetic zones and photic zones that are often sampled individually. The zones are dependent on light, temperature, and in special cases, dissolved oxygen.
Sampling limnetic zones involves first determining where they are. This is achieved by taking oxygen and temperature profiles with electronic meters. Probes for each parameter are lowered into the water at determined depths: just below the surface and at regular intervals thereafter until just above the lake bottom. Once their depths have been calculated, discrete samples can be taken within each zone. The samples must be collected with discrete samplers that "cut out" individual pieces of water. Alternatively you could also collect samples at specific depth intervals. This usually entails more samples than just taking one sample from each limnetic zone. Generally, discrete samples at the midpoint of the various limnetic zones or at regular depth intervals are taken over the deepest point of the lake. This ensures that you have a representative sample from the hypolimnion.
Sampling from the euphotic zone involves first determining its lower limit. This is achieved either by taking a light profile with a light meter or estimated the lower limit depth by using a secchi disk. The euphotic zone is usually sampled by taking depth-integrated samples of the zone at several locations in the lake and combined to form a composite.
Lake sampling is always done from the top down. This is to prevent mixing deep water with surface water, or from mixing water between different zones. Samples are usually only collected to within 1 m of the lake bottom. The collection of water close to the bottom risks disturbing sediments that will distort the results of the sample.
Small ponds are prone to stagnation and stratification; the lack of mixing from inflow, wind, and water currents produces these conditions. The most common way to take a representative sample from a small pond is to obtain a depth integrated sample taken over the entire depth of the water column to within one meter of the bottom in the middle of the water body. Sampling for specific reasons may require the use of a discrete grab sampler, which allows a specific zone or depth to be targeted, such as when groundwater influence is suspected.
Some key factors when locating sampling stations at lakes and other non-flowing water bodies include:
- Access to desirable sampling points
- The bathymetry or, depth profile, of the water body
- Non-uniformity of the lake shape
- Artificial and physical structures such as docks.