Analysis of water quality influencing factors under multi-source data fusion based on PLS-SEM model: An example of East-Liao River in China.

Owing to alterations in the environment and human activities, the quality of surface water is declining. Despite a substantial number of studies on the factors that impact water quality, there is still a need for a better understanding of the major causes of water quality degradation. This study fused multi-source data using partial least squares structural equation modeling to evaluate the effects of weather, soil composition, and geographical features on the water quality of the East Liao River (ELR), Jilin Province, China. The impacts of land-use practices on water quality at different buffer scales were analyzed. The most significant correlation between land use and water quality was observed at a distance of 4 km. The severity of water pollution was significantly influenced by soil type, with a path coefficient of 0.689 (p < 0.001). Conversely, landscape factors exhibited a notable adverse effect, indicated by a path coefficient of -0.608 (p < 0.001). Additionally, meteorological factors exhibited a significant impact, with a path coefficient of 0.463 (p < 0.001). The indirect effects of landscape elements on water quality were also examined. Water quality could be indirectly influenced by landscape through soil factors, as evidenced by a path coefficient of -0.572 (p < 0.01). In this study, new ideas for studying water quality drivers using multi-source data fusion are introduced. Managers can leverage the findings of this study to improve their decision-making and effectively address water quality issues in ELR located in Jilin Province, China.