非参数核密度估计模型预测双酚A的物种敏感度分布规律  

Application of Non-Parametric Kernel Density Estimation Model in P rediction of Species Sensitivity Distribution of Bisphenol A

作　　者：杨瑞君 张楚 杨评 冯承莲[3] 李丹 陶建伟 叶璟 Yang Ruijun;Zhang Chu;Yang Ping;Feng Chenglian;Li Dan;Tao Jianwei;Ye Jing(School of Computer Science&Information Engineering,Shanghai Institute of Technology,Shanghai 201418,China;School of Chemical and Environmental Engineering,Shanghai Institute of Technology,Shanghai 201418,China;State Key Laboratory of Environmental Criteria and Risk Assessment,Chinese Research Academy of Environmental Sciences,Beijing 100012,China;School of Ecological Technology and Engineering,Shanghai Institute of Technology,Shanghai 201418,China)

机构地区：[1]上海应用技术大学计算机科学与信息工程学院,上海201418 [2]上海应用技术大学化学与环境工程学院,上海201418 [3]中国环境科学研究院环境基准与风险评估国家重点实验室,北京100012 [4]上海应用技术大学生态技术与工程学院,上海201418

出　　处：《生态毒理学报》2024年第4期120-130,共11页Asian Journal of Ecotoxicology

基　　金：国家重点研发计划项目(2021YFC3201001);国家自然科学基金青年项目(21307082);国家自然科学基金面上项目(42277274);上海市自然科学基金面上项目(18ZR1438000);上海应用技术大学协同创新基金-跨学科、多领域合作研究专项“新污染物PFAS的水生态风险评估体系构建”(XTCX2024-03)。

摘　　要：双酚A(bisphenol A,BPA)已被证实是内分泌干扰物,可干扰生物体正常的激素功能,对生殖、发育和免疫系统产生不良影响。BPA进入环境后可能会对水生生物和陆地生物造成毒性效应,破坏生态平衡。针对BPA对水生生物的毒性特点,利用Python语言,基于非参数核密度估计模型构建BPA淡水水生生物的敏感度分布曲线,选用4种传统参数模型进行对比,并推导出保护淡水水生生物的BPA水质基准建议值。结果表明,Python构建模型简单高效,相较于传统的参数模型,非参数核密度估计方法在推导BPA的水质基准建议值中更加稳健和精准,其中3种参数模型和非参数核密度估计模型计算得到的雌激素效应毒性的HC 5值分别为4.175、5.096、3.888和1.179μg·L^(-1);其他毒性的HC 5值为7.139、7.452、7.533和5.869μg·L^(-1)。非参数核密度估计的方法能够更好地构建物种敏感度分布曲线,为进一步研究BPA的水质基准和更好地保护淡水水生生物提供了有力支持,同时研究成果以期为我国地表水环境质量标准的制修订做出贡献。Bisphenol A(BPA)has been proven to be an endocrine disruptor that interferes with normal hormonal functions in organisms,adversely affecting reproduction,development,and the immune system.Once BPA enters the environment,it may cause toxic effects on aquatic and terrestrial organisms,disrupting ecological balance.This study focuses on the toxicity characteristics of BPA to aquatic organisms.Using the open-source Python language and its extensive libraries,a non-parametric kernel density estimation model was constructed to develop a sensitivity distribution curve for BPA’s impact on freshwater aquatic species.Four traditional parametric models were s elected for comparison,and the derived water quality criteria for BPA aimed to protect freshwater aquatic life.The results demonstrated that the Python model is simple and efficient to construct.Compared to traditional parametric models,the non-parametric kernel density estimation method proved to be more robust and accurate in deriving BPA water quality criteria.The HC 5 values for estrogenic toxicity calculated by three parametric models and the n on-parametric kernel density estimation model were 4.175,5.096,3.888,and 1.179μg·L^(-1),respectively.For other toxicities,the HC 5 values were 7.139,7.452,7.533,and 5.869μg·L^(-1)for the four models,respectively.Therefore,the n on-parametric kernel density method can provide a better way to construct species sensitivity distribution curves,providing strong support for further research on BPA water quality criteria and improved protection of freshwater a quatic organisms.The findings are expected to contribute to the formulation and revision of surface water e nvironmental quality standards for BPA in China.

关 键 词：双酚A 淡水生物 雌激素效应 非参数核密度估计 物种敏感度分布 水质基准阈值 

分 类 号：X171.5[环境科学与工程—环境科学]