Natural organic matter in drinking water -: the "NOM-typing project", background and basic characteristics of original water samples and NOM isolates

The complex nature of natural organic matter (NOM), and the impact of this matter on drinking water quality, necessitate international co-operative efforts. After several decades of struggle, many experienced NOM scientists world wide recognize the need for international sample references, or better yet, to work together on the same set of samples. The aim of the "NOM-typing project" was a multi-method characterization of a limited number of NOM samples, isolated from different water sources. The background of the project and the nature of the sampling sites are described here, and some data comparing the composition of the samples, before and after the isolation, are presented. The techniques used for isolation were reverse osmosis (RO) and evaporation (EVA). The NOM samples were isolated from eight different locations in the southern part of Norway. The nature of the corresponding catchments differs, however, all samples were from areas with no influence of agriculture or local industry. The key issue of the project was that all samples were isolated with exactly the same methods and by the same team. In spite of the uniformity of the isolation technique used for the nine samples, the recoveries of the material differ from sample to sample. Generally, the loss of NOM is in the range of 10%, however, for one sample, the loss was as high as 35%. These discrepancies are probably partly because of differences in the nature of the NOM and partly due to differences in the general composition of the ambient water. There are a few remarkable differences between the nine samples. There was a high percentage of ash in isolates from acidified areas. These are also the samples where the loss of NOM during the isolation process was highest. The density of the RO-isolates differs 10-fold between the samples. The lipophilisity of the NOM samples, expressed as relative solubility in octanol, differs more than 30-fold. (C) 1999 Elsevier Science Ltd.