Abstract
This study aims to explain the spatial patterns of surface water pollution and investigate water quality for irrigation purposes in the Kızılırmak and Yeşilırmak basins in Turkey. Multivariate statistical techniques such as cluster analysis (CA), principal component analysis/factor analysis (PCA/FA), and multiple regression analysis (MRA) were employed to optimize statistical information and modeling of selected water quality parameters based on the Water Quality Index (WQI) and CORINE land use. The CA grouped the more than ten water quality observation stations within those basins into three groups (G1, G2, and G3) based on their quality properties and pollution levels. Backward stepwise mode discriminant analysis (DA) suggested that the two clusters better explained the spatial similarities. The PCA/FA applied to data sets of two special groups calculated three or four factors for both basins, capturing 81.07%, 72.16%, 73.83%, and 84.49% of the total variance, respectively. WQI values ranging from 34.61 to 63.87 showed the irrigation water quality of the sampling stations. Applying MRA to WQI and deriving the main parameters from PCA/FA demonstrated the efficiency of combining WQI and related irrigation water quality parameters analyzed by multivariate statistical techniques in this study. Based on the results of the WQI dissemination map, 48.59% of the Kızılırmak basin and 98.98% of the Yeşilırmak basin fall within the “high restriction” category. In the Yeşilırmak basin, 2.21% of the total basin area is categorized as “severe restriction”. The remaining areas of both basins need “moderate restrictions”. Maps of the WQI dissemination on CORINE land use classes illustrate that the surface water of basins requires some extent of treatment before consumption. The results of this study have been used to identify major problems in terms of irrigation water quality in both basins and contribute to planning processes for decision-makers.