水-气界面汞通量动态预测模型构建——以长寿湖为例

白薇扬; 柳睿; 江豪浩

PDF(1420 KB)

中国环境科学 ›› 2026, Vol. 46 ›› Issue (1) : 20-28.

岩溶关键带重金属污染成因

水-气界面汞通量动态预测模型构建——以长寿湖为例

白薇扬^1,2, 柳睿¹, 江豪浩¹

作者信息 +

Dynamic prediction model for mercury flux at the water-air interface: A case study of Changshou Lake

BAI Wei-yang^1,2, LIU Rui¹, JIANG Hao-hao¹

Author information +

文章历史 +

摘要

针对湖泊系统汞通量估算受多环境变化因素协同影响,不确定性明显,以喀斯特岩溶的人工水库长寿湖为研究对象,综合运用野外采样,通径分析与机器学习方法来揭示汞通量交换机制.研究发现:汞通量受多因子直接/间接的协同调控.其中,水体汞浓度(0.897)、紫外线强度(0.463)和光照强度(0.446)的直接效应主导着水体溶解气态汞Hg⁰(DGM)生成,而风速通过光照强度(0.226)的间接影响强于其直接扩散作用(0.197).基于机器学习构建的汞通量预测混合模型(0.9GBR(梯度提升机)+0.05RF(随机森林)+0.05SVR(支持向量机回归)),在保持预测精度的同时提升鲁棒性和稳定性,实现湖泊多参数耦合下汞通量的动态预测,为区域汞环境风险评估与管控提供决策支持.

Abstract

Quantifying Hg flux in lake systems remains challenging due to the synergistic influences of multiple environmental drivers. To address this, we investigated Changshou Lake, an artificial reservoir situated in a karst region, through an integrated approach that included field sampling, path analysis, and machine learning, to elucidate the mechanisms of Hg flux exchange. The results indicated that Hg flux was jointly regulated by multiple factors via both direct and indirect pathways. Water Hg concentration (0.897), UV radiation (0.463), and solar radiation (0.446) were identified as exerting dominant direct effects on the production of dissolved gaseous mercury (DGM, Hg⁰). In contrast, the influence of wind speed on Hg flux was observed to be stronger through its indirect enhancement of solar radiation (0.252) than through direct diffusion (0.197). A hybrid machine-learning prediction mode [0.9 Gradient Boosting Regression (GBR) + 0.05 Random Forest (RF) + 0.05 Support Vector Regression (SVR)] was developed. This model was shown to improve generalization ability and stability while maintaining predictive accuracy, enabling dynamic prediction of Hg flux under the coupling of multiple lake parameters. The approach provides decision support for regional Hg risk assessment and management.

导出引用

白薇扬, 柳睿, 江豪浩. 水-气界面汞通量动态预测模型构建——以长寿湖为例[J]. 中国环境科学. 2026, 46(1): 20-28

BAI Wei-yang, LIU Rui, JIANG Hao-hao. Dynamic prediction model for mercury flux at the water-air interface: A case study of Changshou Lake[J]. China Environmental Science. 2026, 46(1): 20-28

中图分类号： X142

参考文献

[1] Stumm W, Morgan J J. Aquatic chemistry: Chemical equilibria and rates in natural waters[M]. 3rd ed. New York: Wiley, 1996.
[2] 阴永光,李雁宾,蔡勇,等.汞的环境光化学[J]. 环境化学, 2011, 30(1):84-91. Yin Y G, Li Y B, Cai Y, Jiang G B. Environmental photochemistry of mercury[J]. Environmental Chemistry, 2011,30(1):84-91.
[3] Schroeder W H, Anholt B R. Global mercury observation system: latitudinal gradient of atmospheric mercury speciation[J]. Environmental Science & Technology, 2017,51(12):6813-6822.
[4] 何天容,冯新斌,郭艳娜,等.红枫湖沉积物中汞的环境地球化学循环[J]. 环境科学, 2008,29(7):1768-1774. He T R, Feng X B, Guo Y N, et al. Geochemical cycling of mercury in the sediment of Hongfeng Reservior[J]. Environmental Science, 2008, 29(7):1768-1774.
[5] 姚珩,冯新斌,仇广乐,等.贵州洪家渡水库水体不同形态汞的分布特征[J]. 生态学杂志, 2011,30(5):961-968. Yao H, Feng X B, Qiu G L, et al. Distribution of different mercury forms of Hongjiadu Reservoir from Guizhou province[J]. Chinese Journal of Ecology, 2011,30(5):961-968.
[6] 方志青,陈秋禹,尹德良,等.三峡库区支流河口沉积物重金属分布特征及风险评价[J]. 环境科学, 2018,39(6):2607-2614. Fang Z Q, Chen Q Y, Yin D L, et al. Distribution characteristics and risk assessment of heavy metals in the sediments of the estuary of the tributaries in the Three Gorges Reservoir, SW China[J]. Environmental Science, 2018,39(6):2607-2614.
[7] 张成,宋丽,王定勇,等.干湿交替条件下三峡水库消落带土壤汞形态变化[J]. 应用生态学报, 2013,24(12):3531-3536. Zhang C, Song L, Wang D Y, et al. Mercury speciation transformation in soil of the water-level-fluctuating zone in the Three Gorges Area under alternative dry-wet condition[J]. Chinese Journal of Applied Ecology, 2013,24(12):3531-3536.
[8] Zhou C, Tada A, Yano S, et al. Mercury dynamic simulation of Minamata Bay by using a three-dimensional numerical model[J]. Environmental Science and Pollution Research, 2019,26(36):36732-36747.
[9] USEPA. Method 1631: Mercury in water by oxidation, Purge and trap, and cold vapor atomic fluorescence spectrometry[S]. Washington DC: United States Environmental Protection Agency, 2002.
[10] TES Electrical Electronic Corp. TES-II digital illuminometer: product specifications[M]. Taiwan, China: TES Electrical Electronic Corp., 2017.
[11] DeltaTRAK Inc. FlashCheck pocket probe digital thermometer 11000: Specifications and calibration guide[M]. California, USA: DeltaTRAK, 2008.
[12] Nielsen-Kellerman. Kestrel 4000Pocket weather tracker: User manual[M]. Commonwealth of Pennsylvania, USA: Nielsen- Kellerman Co. 2012.
[13] Sartorius AG. PT-10pH meter: Technical documentation[M]. Göttingen, Germany: Sartorius Group, 2015.
[14] YSI Incorporated. YSI 550A dissolved oxygen instrument: User manual[M]. Washington, USA: YSI Environmental, 2005.
[15] GE Analytical Instruments. Sievers innovOx laboratory TOC analyzer: Operation manual[M]. Colorado, USA: General Electric Company, 2010.
[16] 闫海鱼,冯新斌,商立海,等.天然水体中痕量汞的形态分析方法研究[J]. 分析测试学报, 2003,22(5):10-13. Yan H Y, Feng X B, Shang L H, et al. Research on the morphological analysis method of trace mercury in natural water bodies[J]. Journal of Instrumental Analysis, 2003,(5):10-13.
[17] Lopez S A, Travnikov O, Sonke J E, et al. Photochemistry of oxidized Hg(I) and Hg(II) species suggests missing mercury oxidation in the troposphere[J]. Proceedings of the National Academy of Sciences, 2020,117(49):30949-30956.
[18] Vette A F, Landis M S, Keeler G J. Mercury photochemistry in the marine boundary layer: Modeling and experimental evidence of rapid destruction[J]. Environmental Science & Technology, 2002,36(1):5-9.
[19] 何天容,冯新斌,郭艳娜,等.红枫湖水体中活性汞和溶解气态汞的分布特征及其控制因素[J]. 环境科学研究, 2008,21(2):14-17. He T R, Feng X B, Guo Y N, et al. Distributions of reactive and dissolved gaseous mercury and controlling factors in Hongfeng Reservoir[J]. Research of Environmental Sciences, 2008,21(2):14-17.
[20] Janssen S E, Murray R K, Dewild J F, et al. Chemical and physical controls on mercury source signatures in stream fish from the Northeastern United States[J]. Environmental Science & Technology, 2019,53(17):10110-10119.
[21] Atangana B M S, Ngoupayou N J R, Deliege J F. Hydrogeochemistry and mercury contamination of surface water in the Lom Gold Basin (East Cameroon): Water quality index, multivariate statistical analysis and spatial interpolation[J]. Water. 2023,15(13):2502.
[22] Crocker W C, Zhang H. Seasonal and diurnal variation of air/water exchange of gaseous mercury in a southern reservoir lake (Cane Creek Lake, Tennessee, USA)[J]. Water, 2020,12(8):2102.
[23] Sullivan C J, Vokoun J C, Perkins C R. Spatiotemporal changes in largemouth bass mercury concentrations from Connecticut waterbodies, 1995~2021[J]. Environmental Monitoring and Assessment, 2023,195(6):780.
[24] Huynh T M T, Ni C F, et al. Predicting heavy metal concentrations in shallow aquifer systems based on low-cost physiochemical parameters using machine learning techniques[J]. International Journal of Environmental Research and Public Health, 2022,19(19).12180.
[25] 陈亚松,刘家雯,赵云鹏,等.基于机器学习的人工湿地出水水质预测与影响因素[J]. 中国环境科学, 2025,45(6):3161-3170. Chen Y S, Liu J W, Zhao Y P, et al. Prediction of effluent water quality and analysis of influencing factors in constructed wetlands based on machine learning[J]. China Environmental Science, 2025,45(6):3161-3170
[26] 成鹏,汪西莉.SVR参数对非线性函数拟合的影响[J]. 计算机工程, 2011,37(3):189-191. Cheng P, Wang X L. Influence of SVR parameter on non-linear function approximation[J]. Computer Engineering, 2011,37(3):189-191.
[27] 钟原,张泰,李平,等.随机森林的融合模型在压井方法分类中的研究[J]. 西南石油大学学报(自然科学版), 2022,44(1):165-173. Zhong Y, Zhang T, Li P, et al. Research on classification of well killing method based on Random Forest fusion model[J]. Journal of Southwest Petroleum University(Science & Technology Edition), 2022, 44(1):165-173.
[28] 刘丽贞,黄琪,迟殿委.基于随机森林回归模型和高频数据的鄱阳湖子湖电导率预测[J]. 水电能源科学, 2023,41(10):50-53. Liu L Z, Huang Q, Chi D W, Prediction of electrical conductivity in sub-lake of Poyang Lake based on random forest regression model and high-frequency Data[J]. Water Resources and Power, 2023,41(10):50-53.
[29] 张瑞杰,周春艳,陈辉,等.基于GF-5B卫星的随机森林模型反演京津冀地区PM_2.5[J]. 中国环境科学, 2024,44(11):5961-5970. Zhang R J, Zhou C Y, Chen H, et al. Inversion of PM_2.5 in the Beijing- Tianjin-Hebei region using a random forest model based on GF-5B satellite[J]. China Environmental Science, 2024,44(11):5961-5970.
[30] 吴志明,李建超,王睿.基于随机森林的内陆湖泊水体有色可溶性有机物(CDOM)浓度遥感估算[J]. 湖泊科学, 2018,30(4):979-991. Wu Z M, Li J C, Wang R. Estimation of CDOM concentration in inland lake based on random forest using Sentinel3A OLCI[J]. Journal of Lake Sciences, 2018,30(4):979-991.
[31] 孔宪喻,苏荣国.基于支持向量机的黄东海富营养化快速评价技术[J]. 中国环境科学, 2016,36(1):143-148. Kong X Y, Su R G. A support vector machine-based technology for rapidly assessing trophic status of the Yellow Sea and the East China Sea[J]. China Environmental Science, 2016,36(1):143-148.