Nonlinear Associations and Spatiotemporal Responses between the Evolution of Blue-Green Space Landscape Patterns and Water-Environment Health in Highly Urbanized Plain River-Network Regions: A Case Study of Suzhou City, China.

Under rapid urbanization, the blue-green space configurations of plain river-network cities are undergoing continuous restructuring, profoundly affecting water-environment health by regulating runoff pathways, pollutant transport, and aquatic self-purification processes. However, the nonlinear, scale-dependent, and stage-specific feedback mechanisms between landscape pattern evolution and water-quality response remain poorly quantified. Using Suzhou City-a representative plain river-network city in the Taihu Lake Basin-as a case study, this research integrates long-term land-use data spanning 2004-2024 with routine water-quality records from 66 monitoring sections. Water quality was evaluated using the CCME-WQI; blue-green landscape metrics were computed within multiscale buffers of 500-2000 m; and an XGBoost-SHAP interpretable machine-learning framework was applied to compare the evolution of landscape-water-quality coupling relationships across two stages: 2004-2012 and 2016-2024. The results indicate that: (1) water-environment quality in Suzhou improved continuously from 2004 to 2024, with the proportion of Poor and Marginal sections declining from approximately 75% to below 15%, and Good or better sections rising from approximately 10% to 45%, accompanied by a pronounced structural turning point around 2016; (2) across both stages, the SHAP importance of the structural blue-space metric (LPI_BlueSpace) consistently exceeded that of the compositional metric (PLAND_BlueSpace), indicating that the retention of core water-body patches explains the spatial heterogeneity of water quality more effectively than total blue-space area; meanwhile, ED_BlueSpace exhibited a stable negative response, suggesting that fragmentation of the water-network boundary suppresses water quality; (3) the most sensitive spatial scale of blue-space influence contracted markedly, migrating from the macro-scale in the earlier stage toward meso- and micro-scales in the later stage. This study characterizes landscape-water-quality associations in highly urbanized plain river-network cities, offering exploratory empirical evidence for structural blue-green space conservation and multi-scale coordinated governance.