Based on hydroinformatics and machine learning: Evolutionary characteristics of water quality in the Yangtze River Basin over the past two decades

The Yangtze River Basin, as the largest in China, assumes pivotal significance in comprehending its evolutionary characteristics through the examination of long-term hydroinformatics analyses. The rapid advancement of machine learning has significantly facilitated the elucidation of patterns in the evolution of water quality within the basin. This investigation relies on water quality data from 55 monitoring stations in the Yangtze River Basin, covering the period from 2004 to 2022. To address missing data, seasonal supplementation was undertaken using Kalman Smoothing and State Space Models algorithms. Subsequently, the Random Forest framework was applied to analyze the temporal evolution and spatial distribution characteristics of water quality parameters. Employing the Gower dissimilarity method, clustering analysis was conducted on mixed-type water quality data, resulting in the identification of distinct clusters, constructing water quality index models under different clusters. The findings indicate an overall enhancement in water quality within the Yangtze River Basin, with winter showcasing superior water quality and the main stem exhibiting significantly higher water quality than its tributaries. Notably significant total phosphorus pollution was observed throughout all seasons, particularly concentrated in the confluence area of the Minjiang and the main stem of the Yangtze. The comprehensive clustering results unveil the existence of three distinct water quality states in the Yangtze River Basin, further subdivided into four periods: 2004-2007, 2008-2014, 2015-2017, and 2018-2022, elucidating the evolving characteristics of water quality in the basin. The entire subset regression model underscores ammonia nitrogen and total phosphorus as common water quality parameters across various seasons and key tributaries, emphasizing the pollution characteristics of water quality in the Yangtze River Basin. This research furnishes essential data and scientific evidence, enriching our understanding and contributing to the efficient management and protection of the aquatic ecosystem in the Yangtze River Basin.