Abstract:

China’s water ecological protection has entered a new stage of integrated and coordinated management of water resources, water environment, and water ecology (“Three Waters”), with water ecology emerging as the central focus of protection efforts. However, current practices predominantly rely on surface water quality standards as the primary protection criteria, whose core function is more oriented towards the protection of the resource uses of water body, ignoring the causal relationship between the impacts of environmental stressors on aquatic ecosystems. This singular management framework results in imbalance or degradation of aquatic ecosystems even when water quality standards are strictly enforced. In contrast, environmental stressor-biological response relationships capture the specific response characteristics and carrying potential of ecosystems to pollutants. Analyzing these relationships enables the detection of ecosystem shifts along environmental gradients, providing critical early warnings of potential structural and functional changes within ecosystems. This approach is particularly important for the protection of aquatic biodiversity and ecosystem stability. In addition, most of the current studies on the water ecological and environmental carrying capacity use the basin as the estimation unit. Yet, the water ecological carrying capacity is jointly influenced by basin-scale factors such as basin climate, hydrology, topography and geomorphology, slope, soil type, and land use, which have significant spatial heterogeneity, making it necessary to carry out basin ecological management in a more refined manner. Benthic algae serve as the main primary producers in small to medium rivers. Their fixed growth characteristics enable them to be highly effective indicators of environmental changes, making them a widely used tool for assessing river ecological status. This study was conducted in the Chishui River, a major tributary of the upper reaches of the Yangtze River. The Chishui River basin is located in a karst landscape area with unique red sandstone and laterite layers. Recognized for its rich biodiversity, the basin constitutes a critical ecological barrier zone in the upper Yangtze. A comprehensive basin-wide survey covering 121 sampling sites was undertaken during the low-flow period in April 2023, when the physical scouring of benthic algae by water flow was minimized, and the relationship between nutrients and algae was most significant. In this study, a total of 332 species of 95 genera and 3 phyla of benthic algae were identified. The algal community was dominated by Bacillariophyta (diatoms, 273 species), followed by Chlorophyta (green algae, 42 species) and Cyanophyta (blue-green algae, 29 species). Achnanthidium minutissimum and Cocconeis placentula were the absolute dominant species. For each sample site, the species composition, richness and biomass of benthic algae were assessed, and the concentrations of total nitrogen (TN) and total phosphorus (TP) were measured concurrently. TN and TP ecological thresholds were estimated for different aspects of algal community, including species richness (α-diversity), community composition (β-diversity) and biomass (ecological function), using both parametric (piecewise linear regression) and non-parametric (nonparametric changepoint analysis, thresholds indicator taxa analysis, and gradient forest) methods. The results showed different ecological thresholds reflecting the different sensitivities of bio-metrics to environmental stresses. For TN, the lowest threshold (representing the most conservative estimate) was 1.119 mg·L⁻¹, while for TP, the corresponding minimum threshold was 0.013 mg·L⁻¹. A comparison with China's National Water Quality Standards (Class II: TN=0.5 mg·L⁻¹; TP=0.1 mg·L⁻¹) showed that the ecological TN threshold was more than twice the Class II limit, compared to the ecological thresholds for TP, which were an order of magnitude lower than the Class II standard. By integrating these ecological thresholds with flow data simulated by the Soil and Water Assessment Tool (SWAT) model, we estimated the annual water ecological carrying capacity for TN and TP of the Chishui River Basin to be 9033.42 t and 104.95 t, respectively. The TN ecological carrying capacity is approximately twice the water environmental carrying capacity (4570.45 t). Conversely, the TP ecological carrying capacity represented only about one-eighth of the water environmental carrying capacity (914.15 t). The notably lower TP carrying capacity observed in this study likely stems from fundamental biogeochemical differences in nitrogen and phosphorus cycling. This stark contrast highlights a critical advantage of the water ecological carrying capacity approach: unlike water quality standards which are often adapted from international experiences, water ecological carrying capacity avoids disconnection from ecological endpoints. It thereby provides an objective and realistic reflection of an ecosystem’s true load-bearing limits, enabling a more scientifically grounded identification of the aquatic ecosystem's actual response level to environmental stress. Consequently, this approach helps prevent both excessive and insufficient management interventions. Temporal analysis of the water ecological carrying capacity revealed a unimodal seasonal pattern. The carrying capacity peaked between July and September, while the lowest capacity occurred during the dry season (January to April). This pattern likely arises because high river runoff during wet periods enhances pollutant assimilation capacity, whereas insufficient hydrodynamic conditions in the dry season constrain load-bearing capability. We also observed significant spatial heterogeneity in water ecological carrying capacity across the basin, prompting investigation into underlying drivers of this variation. Through recursive feature elimination and random forest analysis, we identified four soil types—Luvisols, Cambisols, Anthrosols, and Alisols—as the most important factors influencing the spatial variability of water ecological carrying capacity. These soils are characterized by low water retention and infiltration capacity, which promotes surface runoff generation and thus increases local carrying capacity. Despite the pronounced impact of impervious surfaces in suppressing precipitation interception, urban land was not selected as a significant environmental factor influencing water ecological carrying capacity in this study. This absence likely reflects that the spatial share of urban land in the Chishui River basin is very small (only 0.5%), rendering its hydrological effects insufficiently influential on basin-scale ecological carrying capacity to be included among key drivers. Our findings enable the implementation of finer-scale management strategies that account for basin heterogeneity. Regions exhibiting high ecological carrying capacity may support regulated development of agriculture and industry to foster economic growth, whereas low-capacity areas should prioritize protection, pollution control, and ecological restoration. Future work should focus on integrating pollutant source inventories with nutrient load models to quantify location-specific exceedance magnitudes and identify priority intervention areas. This integration will provide a scientific foundation for ecological protection, evidence-based industrial planning, and optimized pollution load allocation within the Chishui River Basin, ultimately informing ecologically rational decision-making to enhance regional sustainability.

Key words: ecological carrying capacity, ecological thresholds, water quality standards, machine learning, hydrological modeling, catchment management

摘要： 中国水生态环境保护进入了水资源、水环境、水生态“三水”统筹、协同治理的新阶段，水生态成为保护的核心内容。然而，实际工作中仍多以地表水环境质量标准为保护依据，缺乏基于环境胁迫-生物响应关系的生态保护标准，这种单一的管理模式难以对水生态系统进行有效保护，更无法科学量化水生态承载能力。以长江上游一级支流赤水河为例，应用非参数阈值分析方法识别着生藻类对总氮（TN）、总磷（TP）变化的响应阈值，并以氮磷生态阈值为依据估算赤水河的生态承载力。此外，进一步对赤水河氮磷生态承载力的空间变化进行分析并利用机器学习法筛选主要影响因子。结果显示，赤水河TN生态阈值为1.119 mg·L⁻¹，高于地表水II类水质标准（0.5 mg·L⁻¹）；TP生态阈值为0.013 mg·L⁻¹，低于地表水II类水质标准（0.1 mg·L⁻¹）。赤水河TN生态承载力为9033.42 t，约为水环境承载力（4570.75 t）的2倍；而TP生态承载力为104.95 t，约为水环境承载力（914.15 t）的1/8。赤水河TN、TP生态承载力存在较大的空间异质性，高活性淋溶土、雏形土、人为土和高活性强酸土等土壤类型是影响生态承载力空间变化的关键因子。依据环境胁迫-生物响应关系确定的水生态承载力可作为赤水河氮磷营养物负荷的管理标准，为赤水河流域生态环境保护、科学调控产业布局以及优化污染负荷分配提供理论依据。

关键词: 生态承载力, 生态阈值, 水环境质量标准, 机器学习, 水文模拟, 流域管理