Investigation of Temporal Trends in Hydrochemical Quality of Surface Water in Western Turkey Hu ¨ lya Boyacioglu Æ Hayal Boyacioglu Received: 24 July 2007 / Accepted: 14 April 2008 / Published online: 18 May 2008 Ó Springer Science+Business Media, LLC 2008 Abstract This study comprises application of non-para- metric trend analysis (Mann-Kendall test and Sen’s Slope estimator techniques) and multidimensional scaling method to water quality data sets. Water samples analyzed (for chloride, nitrate-nitrogen, sodium, sulfate and total dissolved solid parameters) semi-monthly at seven river monitoring stations in Tahtali Basin, Turkey for 6 years were evaluated. The results revealed that agricultural discharges caused spatial differences in terms of water quality in the basin. Moreover, variable concentrations generally decreased or did not change over time. This study showed that, trend detection and data grouping methods help decision makers to judge effectiveness of management programs. Keywords Mann-Kendall test Á Multidimensional scaling Á Sen’s Slope estimator Á Tahtali Basin In water resources management, the quality of surface waters has recently became as significant as their quantity since the former directly affects the amount of water that can be used for various purposes such as drinking, agri- cultural, recreational and industrial uses etc. Water quality assessment encompasses monitoring, data evaluation, reporting, and dissemination of the condition of the aquatic environment. Major objectives of water quality assessment are describing water quality at regional or national scales, investigating spatial-temporal trends and determining if the water quality meets previously defined objectives for des- ignated uses etc (World Bank 2003; Ouyang 2005). Long term ambient water quality monitoring provides historical database that can be used by the institutions at all levels of society to evaluate water quality (Yake 1979). Non-parametric statistical methods including trend analysis and multidimensional scaling are increasing in use to transfer of complex data to information, which is useful to water managers to assess water quality and develop object oriented management plans. Non-parametric tests are applied when the data is incomplete or a significant amount of the data is missing (Naddeo et al. 2007). This study attempts to evaluate effectiveness of water quality management programs in Tahtali River Basin, Turkey by applying statistical methods to multidimensional data set comprising chemical water quality parameters. In this scope, box plots were plotted to provide a visual impression of the location and shape of the underlying distributions. ‘Mann- Kendall’ test was used to evaluate whether a significant increase or decrease for each water quality parameter occurred. When a significant trend was found for a specific parameter, a second statistic- ‘Sen’s Slope estimator’ was applied to esti- mate the slope (unit change per time period) or the magnitude of the trend. Finally, ‘multidimensional scaling’ was per- formed to reduce the dimensionality of the data set to a sufficiently small value to allow visual inspection of the set. Materials and Methods Tahtali Basin is located in the southern part of the City of Izmir in western part of Turkey and has about 550 km 2 H. Boyacioglu (&) Faculty of Engineering, Department of Environmental Engineering, Dokuz Eylul University, Tinaztepe Campus, Buca 35160, Izmir, Turkey e-mail: hulya.boyacioglu@deu.edu.tr H. Boyacioglu Faculty of Science, Department of Statistics, Ege University, Bornova 35100, Izmir, Turkey e-mail: hayal.boyacioglu@ege.edu.tr 123 Bull Environ Contam Toxicol (2008) 80:469–474 DOI 10.1007/s00128-008-9439-0