Investigating Surface Water Pollution by Integrated Remotely Sensed and Field Spectral Measurement Data: A Case Study

Abstract

Water quality assessment using remote sensing and its terrestrial components is carried out in short time for larger areas. Another issue that is as important regarding water availability is access to quality water. It is important to investigate the availability of the analysis of remotely sensed data instead of environmental and chemical analysis that determines water quality and usability. To examine the detection of water qualities without taking water samples in situ, spectral library data was used in the Hafik Region. In this context we used spectral measurement data of water samples previously taken from Imranli, where the Kizilirmak River originates, and used for spectral classification of water quality. Matched filtering was used for integrating spectral data and CHRIS Proba image as the spectral classification method. To conduct an accuracy analysis, chemical oxygen demand measurement was carried out at 10 points determined as 1st and 2nd water quality in the study area on the river and lakes according to the Ministry of the Environment and Urbanization. The overall accuracy of the classification was calculated as 70%. The results of this study have shown the importance of spectral classification of satellite imagery in evaluating water quality and monitoring water resources.