Monitoring the spatial changes of river water quality through the fusion of Landsat ۸ images and statistical models (A case study: Sefidroud River, Northwest Iran)

The present study determined the relationship between spectral reflectance of satellite images and water quality parameters (WQPs) to monitor water quality of Sefidroud River in Northwest Iran. The data of ۱۲ WQPs were obtained from the regional water department in four hydrometric stations during ۲۰۱۳-۲۰۱۸. Principal component analysis (PCA) was used to reduce the number of WQPs. After pre-processing ۱۹ Landsat ۸ images simultaneously with WQPs sampling and fusing panchromatic band with spectral bands and increasing the band resolution to ۱۵ × ۱۵ m, the findings were employed to determine the relationship between spectral reflectance and band ratios. The PCA results revealed that Sum A (sum of anions), Sum C (sum of cations), TDS and EC were in the first principal component and HCO۳- in the second principal component. The relationship of these five WQPs was statistically investigated with seven single spectral bands and ۲۱ band ratios using regression models and curve estimation, thereby choosing the best regression model. Sum A, TDS and EC was correlated with band ۵, Sum C with band ratio ۴/۳, and HCO۳- with band ۶ (R = ۷۰%). Mapping spatial changes in WQPs using regression models and spectral data indicated higher values in the upstream areas prior to Manjil Dam, decreased values in the dam inflow, while increased those in its outflow. The research results highlighted the usefulness of Landsat ۸ images for water quality monitoring.

کلیدواژه ها:

factor analysis ، Landsat ۸ ، Sefidroud River ، Water Quality ، WQPs