Abstract
ABSTRACT MS Thesis DETECTION OF WATER QUALITY TRENDS IN THE RIVERS OF THE SUSURLUK BASIN Serdar KALAYCI Selçuk University Graduate School of Natural and Applied Sciences Department of Civil Engineering Supervisor : Assoc. Prof. Dr. Yüksel BURDURLU 1997, 61 Pages Jury : Assoc. Prof. Dr. Yüksel BURDURLU Assist Prof. Dr. Saim KOÇAK Assist. Prof. Dr. Ercan KAHYA In recent years, some non-parametric tests to determin trends of surface water quality have been proposed. Assumptions in the classical parametric tests (i.e., normality, linearity, independence) are usually not met by typical surface water quality data. At the same time, statistical tests for trends of water quality are commonly confounded by several of the following problems: missing values, censored data (i.e., values reported as less than a specified quantity), flow relatedness and seasonality. For the reasons, several non-parametric tests (Sen's T, Spearman's Rho, Mann-Kendall, Seasonal Kendall, Van Belle and Hughes, Wilcoxon Paired-Replicate, Mann-Whitney and Kruskall-Wallis'H Tests) have been developed. In this study, surface water quality data of the rivers of the Susurluk Basin collected by EIEI were used. Four different non-parametric trend tests were selected to detect linear trends of surface water quality. These tests are the Sen's T test, the Spearman's Rho test, the Mann-Kendall test and the Seasonal Kendall test. The linear slopes (change per unit time) of trends are calculated by using a non-parametric estimator developed by Sen. In addition, homogeneities in monthly trends are tested by using a method developed by Van Belle and Hughes. Key Words : Sen's T, Spearman's Rho, Mann-Kendall, Seasonal Kendall, Van Belle and Hughes, water quality, non-parametric tests, seasonality. 11
