An Application of Landsat-5TM Image Data for Water Quality Mapping in Lake Beysehir, Turkey

The main goal of this study was to investigate spatial patterns in water quality in Lake Beysehir, which is the largest freshwater reservoir in Turkey, by using Landsat-5TM (Thematic Mapper) data and ground surveys. Suspended sediment (SS), turbidity, Secchi disk depth (SDD), and chlorophyll-a (chl-a) data were collected from 40 sampling stations in August, 2006. Spatial patterns in these parameters were estimated using bivariate and multiple regression (MR) techniques based on Landsat-5TM multispectral data and water quality sampling data. Single TM bands, band ratios, and combinations of TM bands were estimated and correlated with the measured water quality parameters. The best regression models showed that the measured and estimated values of water quality parameters were in good agreement (0.60 < R ² < 0.71). TM3 provided a significant relationship (R ² = 0.67, p < 0.0001) with SS concentration. MR between chl-a and various combinations of TM bands showed that TM1, TM2, and TM4 are strongly correlated with measured chl-a concentrations (R ² = 0.60, p < 0.0001). MR of turbidity showed that TM1, TM2, and TM3 explain 60% (p < 0.0001) of the variance in turbidity. MR of SDD showed a strong relationship with measured SDD, with R ² = 0.71 (p < 0.0001) for the ratio TM1/TM3 and TM1 band combinations. The spatial distribution maps present apparent spatial variations of selected parameters for the study area covering the largest freshwater lake and drinking water reservoir in Turkey. Interpretation of thematic water quality maps indicated similar spatial distributions for SS, turbidity, and SDD. A large area in the middle portion of the lake showed very low chl-a concentrations as it is far from point and nonpoint sources of incoming nutrients. The trophic state index values were calculated from chl-a and SDD measurements. Lake Beysehir was classified as a mesotrophic or eutrophic lake according to chl-a or SDD parameters, respectively.