| 我国流域铜的长期水质基准预测模型研究——MLR vs. BLM |
| |
| 引用本文： | 邵美晨,杨晓玲,王梦晓,穆云松,白英臣,吴丰昌.我国流域铜的长期水质基准预测模型研究——MLR vs. BLM[J].环境科学研究,2023,36(6):1236-1244. |
| |
| 作者姓名： | 邵美晨 杨晓玲 王梦晓 穆云松 白英臣 吴丰昌 |
| |
| 作者单位： | 1.中国人民大学环境学院，北京 100872 |
| |
| 基金项目： | 国家自然科学基金面上项目(No.42277235)；中国人民大学新教师启动基金项目(No.2020030262) |
| |
| 摘 要： | 铜在实际水环境中的毒性效应受多种环境理化要素影响，识别典型环境因子进而建立生物有效性模型是获得“原位”水质基准的基础. 国际上广泛应用的生物配位体模型(biotic ligand model, BLM)在实际使用过程中有局限性. 本研究旨在BLM模型的基础上，探究更适用于铜的“原位”毒性校正方法，以更少的环境因子降低环境监管过程的经济和技术成本，为“因地制宜”确定金属水质基准提供理论依据和技术支撑. 同时基于4门22种水生生物的慢性毒性数据，以影响铜生物有效性的关键水化学参数为自变量，即硬度(hardness，H)、pH和溶解有机碳(dissolved organic carbon，DOC)，采用带有交互项的3种逐步多元线性回归(Multiple Linear Regression，MLR)方法建模，从9个模型中筛选出最佳预测模型〔R2=0.553 1，F=61.23，P<0.000 1，AIC(赤池信息准则，Akaike information criterion)=0.9457，BIC(贝叶斯信息准则，Bayesian information criterion)=23.43〕. 并参考我国典型流域水环境特征，正交设计216组水环境条件，对所有物种的慢性毒性值(chronic toxicity value，CTV)进行模型校正. 结果表明:①以预测准确率(RFx,2.0)为评价指标，MLR模型较BLM模型提升了20%，MLR模型残差分析的等级评分也优于BLM模型. ②确定最佳模型为Sigmoidal-Weibull函数(R2=0.986 1，F=1 609，P<0.001)，构建的三环境因子耦合的基准预测模型准确度高(R2=0.990 2，RMSE=0.041 06，F=7 238，P<0.001). ③主要环境因子(H、pH和DOC)只使得SSD曲线水平移动〔与中心(Xc)有关〕，曲线形状无影响〔振幅(A)、指数(d)和协同系数(k)恒定〕. ④在模型的适用范围内，我国流域水体中铜的长期水质基准为4.619~53.75 μg/L. 研究显示，基于BLM模型框架的新MLR模型，具有高稳健性和准确性，有助于体现流域水生态管理的区域性差异.
|
| 关 键 词： | 重金属 水质基准 生物有效性 物种敏感度分布 DOC |
| 收稿时间： | 2022-09-26 |
Predictive Model for Setting Long-Term Water Quality Criteria of Copper in Chinese River Basins: MLR vs. BLM |
| |
| Affiliation: | 1.School of Environment & Natural Resources, Renmin University of China, Beijing 100872, China2.State Key Laboratory of Environmental Criteria and Risk Assessment, Chinese Research Academy of Environmental Sciences, Beijing 100012, China |
| |
| Abstract: | The site-specific toxic effect of copper can be affected by multiple environmental physical and chemical factors. Thus, identifying typical environmental factors and establishing bioavailability models are the basis for deriving ‘in situ’ water quality criteria (WQC). The Biotic Ligand Model (BLM), which is widely applied internationally, has some limitations in practical use. This research is aimed to explore the ‘in-situ’ toxicity and correction methods more suitable for copper based on BLM, reduce the economic and technical costs of environmental supervision process with fewer environmental factors, and provide theoretical and technical support for determining WQC of metals according to local conditions. Based on the chronic toxicity data of 22 aquatic organisms in 4 phyla, the key environmental parameters affecting the bioavailability of copper were taken as independent variables, namely hardness (H), pH and dissolved organic carbon (DOC). In addition, the best prediction model was selected from 9 models by three MLR approaches with interaction terms (R2=0.5531, F=61.23, P<0.0001, AIC=0.9457, BIC=23.43). In view of the water environment characteristics of typical river basins in China, 216 sets of environmental conditions were orthogonally designed, and the chronic toxicity values (CTV) of all species were corrected by environmental properties. According to the results of study: (1) With the prediction accuracy (RFx,2.0) as the evaluation index, the MLR-based model was better than BLM and improved by 20%. Meanwhile, MLR residual analysis rating was also superior to BLM. (2) The Sigmoidal-Weibull function was determined as the best model (R2=0.986 1, F=1609, P<0.001). A prediction model to derive WQC, coupled with three environmental factors, was ultimately constructed (R2=0.9902, RMSE=0.04106, F=7238, P<0.001). (3) The major environmental factors (H, pH and DOC) only changed the horizontal displacement of the SSD curve (related to Xc) but exerted no effect on the curve shape (A, d, and k is constant). (4) Within the application domain of the model, the long-term WQC of copper in the river basins of China is 4.619-53.75 μg/L. Based on the BLM model framework, the findings show that the new MLR model has higher robustness and accuracy, which is conducive to the rationalization of watershed ecological management. |
| |
| Keywords: | |
|
| 点击此处可从《环境科学研究》浏览原始摘要信息 |
|
点击此处可从《环境科学研究》下载全文 |
|
