کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
4457482 | 1620925 | 2013 | 15 صفحه PDF | دانلود رایگان |

• Portugal has a very large number of mineral waters with a very variable chemistry.
• The diversity is related to geology which defines the composition of these waters.
• Three different classification methods are compared to assess this variability.
• Maps are prepared that show variability between samples and classification methods.
Three different water classification techniques were applied to classify all mineral waters in Portugal from which chemical analysis were found in the literature. The three techniques were hierarchical clustering as can be applied from the Statistical Package for the Social Sciences and two empirical techniques, one based primarily on chemical properties of the waters and one based primarily on hydrological properties. The three methods were compared and it was found that there was a reasonable overlap between the hierarchical clustering and the classification method based primarily on chemical properties, while the hydrological technique showed systematic differences. The results from all classifications are plotted on block maps to indicate areas where different water types are dominant. It is concluded that for the general classification of mineral waters in Portugal the empirical techniques are most usable as newly discovered mineral waters can easily be fitted into the classification, while a new calculation, inferring the risk of samples moving from one to another cluster, needs to be made if hierarchical clustering is applied. Considering the fact that often chemical characterization of mineral waters is used in discussions it is suggested that the classification method based on chemical characteristics is perhaps the most useful to apply to Portuguese mineral waters. The empirical classification techniques can also be applied easily to mineral water sets from other regions than Portugal.
Journal: Journal of Geochemical Exploration - Volume 132, September 2013, Pages 125–139