Assessing the factors influencing water quality using environment water quality index and partial least squares structural equation model in the Ebinur Lake Watershed, Xinjiang, China

Surface water quality deterioration is commonly associated with environmental changes and human activities. Although some research has been carried out to evaluate the relationship between various influencing factors and water quality, there is still very little scientific understanding on how to accurately define the key factors of water quality deterioration. This study aims to quantify the impact of environmental factors and land use land cover (LULC) changes on water quality in the Ebinur Lake Watershed, Xinjiang, China. A total of 20 water parameters were used to calculate the Environment Water Quality Index (CWQI). Meanwhile, the partial least squares-structural equation model (PLS-SEM) was used to quantify the impact of eleven factors influencing water quality in the watershed. About 33.3% of the monitoring points that located mostly in the downstream region with dominant anthropogenic activities were detected as poor quality. There were no obvious temporal changes in water quality from 2016 to 2019. The PLS-SEM simulation shows that the latent variable “land use/cover types” (path coefficient = − 0.600) and “Environmental factor” (path coefficient = − 0.313) are two major factors affected water quality in the Ebinur Lake Watershed, with a strong explanatory power to water quality change (R² = 0.727). In the latent variable “Environmental factors”, the “NDVI” and “night light brightness value” have a great influence on water quality, with the weights of 0.451 and 0.427, respectively. Correspondingly, the “farmland” and “forest land” within the latent variable of “Land use/cover type” have a considerable impact water quality, with the weights of 0.361 and − 0.340, respectively. In conclusion, the influence of anthropogenic activities on surface water quality of the Ebinur Lake Watershed is greater than that of environmental factors. Compared with the traditional multivariate statistical method, PLS-SEM provides a new insight for quantifying the complex relationship between different influencing factors and water quality.