The causes of algal blooms exist significant scale effect in tributary of the Three Gorges Reservoir

Abstract

Algal blooms of rivers becomes a global issue, while there is controversy over the attribution of blooms due to the differences of research scales. This study utilized random forest and spatiotemporal geographically weighted models to explore the triggers and mechanisms of blooms from the different spatial resolutions (river, local and grid scales) in tributary bay of the Three Gorges reservoir based on Sentinel-2 data and Environmental Fluid Dynamics Code model. The results indicated that blooms in the Daning River were more severe during the spring and summer, with the average chlorophyll about 10.15 ug/L and 9.55 ug/L. The high-risk areas primarily located in the middle reaches (10.65 ug/L). The causes of blooms existed significant scaling effect. The water temperature recognized at the river scale was not a sufficient condition for blooms in the early stage of drainage. Compared with temperature, the relative importance of Zme ranked in the top three, which highlighted the influences of Zme at the local and grid scales. The macro-analysis identified the importance of water level fluctuations during drainage stage and the low level period, with ranking in the top two. The effects indirectly altered Zme and Zm, thus highlighting triggering effect of Zme and Zm at local scales. During the stage, thermal stratification and water circulation patterns determined the heterogeneity of blooms. The lag time between chlorophyll and these key factors ranged from 1 to 5 days. This diagnostic to the spatiotemporal scale contributes to understand of the formation mechanism of algal blooms in the tributary bay of the Three Gorges reservoir, providing reference for early warning and prevention strategy.