海河流域河流氮素变化趋势及关键因子识别

汪昊; 高凯璐; 尹高飞; 孙学莹; 刘一航; 张旭琳; 王凤; 翟梦瑜; 王佳宝; 李昭慧; 文宏达; 李文超

PDF(1428 KB)

中国环境科学 ›› 2025, Vol. 45 ›› Issue (7) : 3968-3977.

环境生态

海河流域河流氮素变化趋势及关键因子识别

汪昊¹, 高凯璐¹, 尹高飞¹, 孙学莹², 刘一航¹, 张旭琳¹, 王凤³, 翟梦瑜⁴, 王佳宝¹, 李昭慧¹, 文宏达¹, 李文超¹

作者信息 +

Identification of river nitrogen trends and key factors in the Haihe River

WANG Hao¹, GAO Kai-lu¹, YIN Gao-fei¹, SUN Xue-ying², LIU Yi-hang¹, ZHANG Xu-lin¹, WANG Feng³, ZHAI Meng-yu⁴, WANG Jia-bao¹, LI Zhao-hui¹, WEN Hong-da¹, LI Wen-chao¹

Author information +

文章历史 +

摘要

基于2021~2024年海河流域69个水质监测站的监测数据,通过耦合随机森林模型(RF)和长短时记忆神经网络(LSTM)构建了TN浓度预测的RF-LSTM数据驱动模型,并探究了TN浓度的时空变化规律及主要影响因素.结果表明:2021年TN浓度较高,2022年TN浓度较低,年均值变化在4.97~5.22mg/L;5个子流域TN浓度年内呈U型的季节变化特征,TN浓度最高值出现在1月和2月,最低值出现在8月;水温和溶解氧浓度是影响海河流域水体TN浓度的关键因子,水温与TN浓度呈极显著负相关,溶解氧浓度与TN浓度呈极显著正相关;RF-LSTM数据驱动模型的预测效果(R²:0.830~0.939)优于传统单一LSTM (R²:0.818~0.931)模型,且RF-LSTM数据驱动模型具有更强的泛化性能.

Abstract

This study constructed a data-driven RF-LSTM model to predict Total Nitrogen (TN) concentrations in the Haihe River, based on monitoring data from 69 water quality stations collected between 2021 and 2024. By coupling the Random Forest (RF) algorithm with a Long Short-Term Memory (LSTM) neural network, the model also explored the spatiotemporal variation and key influencing factors of TN. Results showed that TN concentrations were relatively high in 2021 and lower in 2022, with annual averages ranging from 4.97 to 5.22mg/L. Across the five watersheds, TN levels displayed a U-shaped seasonal pattern, peaking in January and February and reaching the lowest levels in August. Temperature and dissolved oxygen were identified as the main influencing factors. TN concentration showed a highly significant negative correlation with temperature and a highly significant positive correlation with dissolved oxygen. The RF-LSTM model outperformed the traditional standalone LSTM model in prediction accuracy (R²:0.830~0.939, R²:0.818~0.931) and demonstrated stronger generalization performance.

导出引用

汪昊, 高凯璐, 尹高飞, 孙学莹, 刘一航, 张旭琳, 王凤, 翟梦瑜, 王佳宝, 李昭慧, 文宏达, 李文超. 海河流域河流氮素变化趋势及关键因子识别[J]. 中国环境科学. 2025, 45(7): 3968-3977

WANG Hao, GAO Kai-lu, YIN Gao-fei, SUN Xue-ying, LIU Yi-hang, ZHANG Xu-lin, WANG Feng, ZHAI Meng-yu, WANG Jia-bao, LI Zhao-hui, WEN Hong-da, LI Wen-chao. Identification of river nitrogen trends and key factors in the Haihe River[J]. China Environmental Science. 2025, 45(7): 3968-3977

中图分类号： X522

参考文献

[1] Zhang Y, He X, Lian G, et al. Monitoring and spatial traceability of river water quality using Sentinel-2satellite images [J]. Science of the Total Environment, 2023,894:164862.
[2] Wang A, Yang D, Tang L. Spatiotemporal variation in nitrogen loads and their impacts on river water quality in the upper Yangtze River basin [J]. Journal of Hydrology, 2020,590(1):125487.
[3] 徐建锋,尹炜,闫峰陵,等.农业源头流域景观异质性与溪流水质耦合关系 [J]. 中国环境科学, 2016,36(10):3193-3200. Xu J F, Yin W, Yan F L, et al. The coupling relationship between landscape heterogeneity and stream water quality in an agricultural catchment [J]. China Environmental Science, 2016,36(10):3193-3200.
[4] Kaveh K, Kaveh H, Bui M D, et al. Long short-term memory for predicting daily suspended sediment concentration [J]. Engineering with Computers, 2021,37:2013-2027.
[5] Nourani V, Behfar N. Multi-station runoff-sediment modeling using seasonal LSTM models [J]. Journal of Hydrology, 2021,601:126672.
[6] Fang L, Shao D. Application of long short-term memory (LSTM) on the prediction of rainfall-runoff in karst area [J]. Frontiers in Physics, 2022,9:790687.
[7] 杜尚海,古成科,张文静.随机森林理论及其在水文地质领域的研究进展 [J]. 中国环境科学, 2022,42(9):4285-4295. Du S H, Gu C K, Zhang W J. A review on the progresses in random forests theory and its applications in hydrogeology [J]. China Environmental Science, 2022,42(9):4285-4295.
[8] Xue J, Yuan C, Ji X, et al. Predictive modeling of nitrogen and phosphorus concentrations in rivers using a machine learning framework: A case study in an urban-rural transitional area in Wenzhou China [J]. Science of the Total Environment, 2024,910: 168521.
[9] Hasani S S, Arias M E, Nguyen H Q, et al. Leveraging explainable machine learning for enhanced management of lake water quality [J]. Journal of Environmental Management, 2024,370:122890.
[10] Zhi W, Baniecki H, Liu J, et al. Increasing phosphorus loss despite widespread concentration decline in US rivers [J]. Proceedings of the National Academy of Sciences, 2024,121(48):e2402028121.
[11] Zhang W, Jin X, Liu D, et al. Assessment of the sediment quality of freshwater ecosystems in eastern China based on spatial and temporal variation of nutrients [J]. Environmental Science and Pollution Research, 2017,24:19412-19421.
[12] 冯爱萍,吴传庆,王雪蕾,等.海河流域氮磷面源污染空间特征遥感解析 [J]. 中国环境科学, 2019,39(7):10. Feng A P, Wu C Q, Wang X L, et al. Spatial character analysis on nitrogen and phosphorus diffuse pollution in Haihe River Basin by remote sensing [J]. China Environmental Science, 2019,39(7):10.
[13] Yao T, Liu S, Hu S, et al. Response of vegetation ecosystems to flash drought with solar-induced chlorophyll fluorescence over the Hai River Basin, China during 2001~2019 [J]. Journal of Environmental Management, 2022,313:114947.
[14] 成星露,孙永华,张王宽,等.气候干湿冷暖化复合类型的识别与量化方法 [J]. 地球信息科学学报, 2023,25(11):2204-2217. Cheng X L, Sun Y H, Zhang W K, et al. Method to identify and quantify compound dry(wet)-cool(warm) climate trend [J]. Journal of Geo-information Science, 2023,25(11):2204-2217.
[15] 赵安周,李子洋,承达瑜.海河流域饱和水汽压差时空演变及影响因素分析 [J]. 生态学报, 2025,45(1):385-394. Zhao A Z, Li Z Y, Cheng D Y. Temporal and spatial variation and influencing factors of saturated water vapor pressure difference in the Hai River Basin [J]. Acta Ecologica Sinica, 2025,45(1):385-394.
[16] 王庆明,赵勇,王浩,等.海河流域地表水资源衰减归因与规律 [J]. 中国科学:地球科学, 2024,54(5):1573-1587. Wang Q M, Zhao Y, Wang H, et al. Surface water resource attenuation attribution and patterns in Hai River Basin [J]. Scientia Sinica(Terrae), 2024,54(5):1573-1587.
[17] Yua X, Jun Z. Water resource risk assessment based on non-point source pollution [J]. Water, 2021,13(14):1907.
[18] 李海林,周文浩,吴炳毅,等.基于云模型理论的数据预处理及校准分析方法研究 [J]. 系统工程理论与实践, 2025:1-19. Li H L, Zhou W H, Wu B Y, et al. Research on data preprocessing and calibration method based on cloud model [J]. Systems Engineering-Theory and Practice, 2025:1-19.
[19] Zou T, Yang K, Pan M, et al. Analysis of the temporal and spatial evolution of turbidity in Tonle Sap Lake and its influencing factors [J]. Science of The Total Environment, 2024,943(000):16.
[20] 郝玉莹,赵林,孙同,等.基于RF-LSTM的地表水体水质预测 [J]. 水资源与水工程学报, 2021,32(6):41-48. Hao Y Y, Zhao L, Sun T, et al. Surface water quality prediction based on RF-LSTM [J]. Journal of Water Resources and Water Engineering, 2021,32(6):41-48.
[21] Wan D. Distribution, sources and main controlling factors of nitrate in a typical intensive agricultural region, northwestern China: Vertical profile perspectives [J]. Environmental Research, 2023,237(1): 116911-116911.
[22] Podgorski J E, Labhasetwar P, Saha D, et al. Prediction modeling and mapping of groundwater fluoride contamination throughout India [J]. Environmental Science and Technology, 2018,52(17):9889-9898.
[23] Wei W, Nghiem A, Ma R, et al. Factors controlling iodine enrichment in a coastal plain aquifer in the North Jiangsu Yishusi Plain, China [J]. Journal of Contaminant Hydrology, 2021,243:103894.
[24] Tesoriero A J, Gronberg J A, Juckem P F, et al. Predicting redox‐sensitive contaminant concentrations in groundwater using random forest classification [J]. Water Resources Research, 2017,53(8):7316-7331.
[25] Luo L, Zhang Y, Dong W, et al. Ensemble empirical mode decomposition and a long short-term memory neural network for surface water quality prediction of the Xiaofu River, China [J]. Water, 2023,15(8):1625.
[26] Yu Y, Si X, Hu C, et al. A review of recurrent neural networks: LSTM cells and network architectures [J]. Neural computation, 2019,31(7): 1235-1270.
[27] 林天宙,彭杨,罗诗琦,等.基于RF-LSTM模型的长江上游汛期日含沙量预测 [J]. 中国农村水利水电, 2024,(10):32-39. Lin T Z, Peng Y, Luo S Q, et al. Prediction of daily sediment concentrations in the upper reaches of Yangtze River during flood season based on RF-LSTM model [J]. China Rural Water and Hydropower, 2024,(10):32-39.
[28] Adamowski J, Chan H F. A wavelet neural network conjunction model for groundwater level forecasting [J/OL]. Journal of Hydrology, 2011,407(1-4):28-40.
[29] Dayev Z A. Application of artificial neural networks instead of the orifice plate discharge coefficient [J], Flow Measurement and Instrumentation, 2020,71:101674.
[30] 熊伟光,彭嘉玉,李晓光,等.2018~2022年海河流域总氮浓度时空变化特征及驱动因素分析 [J/OL]. 环境科学, 2025,46(1):184-192. Xiong W G, Peng J Y, Li X G, et al. Spatial-Temporal Variations and Driving Factors Analysis of Total Nitrogen Concentration in the Haihe River Basin from 2018 to 2022 [J/OL]. Environmental Science, 2025,46(1):184-192.
[31] Li S, Wang X, Qiao B, et al. First flush characteristics of rainfall runoff from a paddy field in the Taihu Lake watershed, China [J]. Environmental Science and Pollution Research, 2017,24(9):1-16.
[32] 韦英怀,胡敏鹏,陈丁江.我国主要河流水系硝态氮污染特征及定量源解析 [J]. 环境科学, 2024,45(02):755-767. Wei H Y, Hu M P, Chen D J. Nitrate pollution characteristics and its quantitative source identification of major river systems in China. Environmental Science [J], 2024,45(2):755-767.
[33] Zhao Z M L. Agricultural nitrogen and phosphorus emissions to water and their mitigation options in the Haihe Basin, China [J]. Agricultural Water Management, 2019,212:262-272.
[34] 徐华山,徐宗学,唐芳芳,等.漳卫南运河流域水质时空变化特征及其污染源识别 [J]. 环境科学, 2012,33(2):359-369. Xu H S, Xu Z X, Tang F F, et al. Spatiotemporal variation analysis and identification of water pollution sources in the Zhangweinan River Basin [J]. Environmental Science, 2012,33(2):359-369.
[35] 孙雪,于格,刘汝海,等.海河南系子牙河水系湿地生态系统健康评价研究 [J]. 中国海洋大学学报, 2019,49(11):120-132. Sun X, Yu G, Liu R H, et al. Wetland ecosystem health assessment in Ziya River Basin, Haihe River [J]. Periodical of Ocean University of China, 2019,49(11):120-132.
[36] 张厌非.试谈永定河上游水土流失和水土保持 [J]. 中国水土保持, 1984,(2): 44-46. Zhang Y F, Discussion on soil and water loss and conservation in the upper reaches of Yongding River [J]. Soil and Water Conservation in China, 1984,(2):44-46.
[37] 温燕华.密云水库上游潮白河流域典型小流域治理研究 [D]. 邯郸:河北工程大学, 2020. Wen Y H. Study on governance of typical small watersheds in Chaobai River Basin in the upper reaches of Miyun Reservoir [D]. Handan: Hebei University of Engineering, 2020.
[38] 李琳琳,王国清,秦攀,等.白洋淀水环境状况与治理保护对策 [J]. 科技导报, 2019,37(21):14-25. Li L L, Wang G Q, Qin P, et al. Water environmental quality of Baiyangdian Lake and its treatment and protection countermeasures [J]. Science and Technology Review, 2019,37(21):14-25.
[39] Huang J, Zhang Y, Bing H, et al. Characterizing the river water quality in China: Recent progress and on-going challenges [J]. Water Research, 2021,201(4):117309.
[40] Liu L, Yang J, Yu X, et al. Patterns in the composition of microbial communities from a subtropical river: Effects of environmental, spatial and temporal factors [J]. PLOS ONE, 2017,8(11):81232.
[41] Lei Y, Dong F, Liu X, et al. Short-term variations and correlations in water quality after dam removal in the Chishui river basin [J]. Journal of Environmental Management, 2023,327:116917.
[42] Wang Y T, Xu H, Shen Z, et al. Effects of dissolved oxygen on water quality and biofilms in the raw water distribution pipeline for a reserved water source [J]. Environmental Science: Water Research and Technology, 2022,8(10):2277-2288.
[43] Thach T T, Harada M, Oniki A, et al. Experimental study on the influence of dissolved organic matter in water and redox state of bottom sediment on water quality dynamics under anaerobic conditions in an organically polluted water body [J]. Paddy and Water Environment, 2017,15(4):889-906.
[44] 张朋超,赵有明,刘翔,等.麻雀搜索算法优化极端梯度提升模型的岩石爆破块度预测 [J]. 科学技术与工程, 2024,24(24):10212-10219. Zhang P C, Zhao Y M, Liu X, et al. Prediction for blasting fragmentation of rocks using extreme gradient boosting optimized bysparrow search algorithm [J]. Science Technology and Engineering, 2024,24(24):10212-10219.
[45] 麦嘉琪.基于机器学习的偶氮染料最大吸收波长规律研究 [D]. 上海:上海大学, 2023. Mai J Q. Research on the Maximum Absorption Wavelength of Azo Dyes based on Machine Learning [D]. Shanghai: Shanghai University, 2023.
[46] 李振今,王志勇,刘晓彤,等.互花米草治理对黄河三角洲湿地利用及景观格局的影响——利用随机森林和改进的CA-Markov模型 [J]. 生态学报, 2024,44(18):8366-8382. Li Z J, Wang Z Y, Liu X T, et al. Using random forest and improved CA-Markov model to evaluate the effect of Spartina alterniflora clearing on wetland use and landscape pattern in the Yellow River Delta wetland [J]. Acta Ecologica Sinica, 2024,44(18):8366-8382.
[47] 向玲,陈锦鹏,付晓孟婷,等.基于ViT和LSTM的风速多步预测 [J]. 太阳能学报, 2024,45(9):525-533. Xiang L, Chen J P, Fu X M T, et al. Multi-step wind speed forecasting based on VIT and LSTM [J]. Acta Energiae Solaris Sinica, 2024,45(9): 525-533.
[48] 汪敏,荣腾飞,李茜,等.基于可学习小波自注意力模型的海上风电功率超短期预测 [J]. 高电压技术, 2025,51(3):1422-1433. Wang M, Rong T F, Li Q, et al. Ultra-short term prediction of offshore wind power based on learnable wavelet self-attention model [J]. High Voltage Engineering, 2025,51(3):1422-1433.
[49] Ling X, Lu G, Xue C. Environmental and anthropogenic factors affect bacterial community and nitrogen removal in the Yarlung Zangbo River [J]. Environmental science and pollution research international, 2022,29(56):84590-84599.
[50] Feng S, Wang M, Xu W, et al. The impact of emissions controls on atmospheric nitrogen inputs to Chinese river basins highlights the urgency of ammonia abatement [J]. 2024,10(37):2558.
[51] Ferrier R C, Whitehead P G, Sefton C, et al. Modelling impacts of land use change and climate change on nitrate-nitrogen in the River Don, North East Scotland [J]. Water Research, 1995,29(8):1950-1956.
[52] Ballard T C. Past and future changes in river temperatures and nitrogen export [M]. Stanford: Stanford University, 2021:1–100.