Don't think the temperature of the sample has anything to do with the results. The majority of contaminants in water are completely dissolved, many of them are quite easy to sample for the dissolved constituents.
The majority of the errors involved in sampling and measurement occur in the sampling lines. For example, a high temperature sample saturated with dissolved metal or metal oxide may become supersaturated as it is cooled and the contaminant may plate out on the sampling system surfaces.
It is important to keep the sample flow rate and temperature constant. A change may result in a crud burst that can raise the contaminant concentration by an order of magnitude. Silica monitoring will also require that the samples be continously flowing.