Time-series monitoring and prediction of cyanobacteria blooms based on FY3D/MERSI data

ZHANG Ruo-lin; WANG Meng; MENG Qing-yan; SUN Yun-xiao; YUAN Xi-tun; ZHANG Lin-lin; WU Han-tian; SUN Zhen-hui; WANG Jia-long

PDF(2117 KB)

China Environmental Science ›› 2026, Vol. 46 ›› Issue (3) : 1509-1521.

Environmental Ecology

Time-series monitoring and prediction of cyanobacteria blooms based on FY3D/MERSI data

ZHANG Ruo-lin¹, WANG Meng², MENG Qing-yan^3,4, SUN Yun-xiao⁵, YUAN Xi-tun¹, ZHANG Lin-lin³, WU Han-tian⁶, SUN Zhen-hui⁵, WANG Jia-long⁵

Author information +

History +

Abstract

Conventional remote sensing platforms like Landsat and Sentinel face inherent limitations in achieving continuous dynamic monitoring of cyanobacterial blooms due to their inadequate temporal resolution. Similarly, existing FY satellite-based monitoring and prediction approaches are constrained by unidimensional analyses and oversimplified models dependent on single-variable inputs. To overcome these challenges, this study developed a comprehensive multidimensional monitoring and prediction system utilizing high-temporal-resolution FY3D/MERSI data for Taihu Lake cyanobacterial blooms. Our methodology integrated spectral indices with an optimized Otsu algorithm for accurate bloom detection, coupled with innovative dual-scale (pixel and sub-pixel) quantification techniques. Through this framework, we conducted systematic multidimensional analysis of Taihu Lake's cyanobacterial blooms (2019~2023), examining critical parameters including bloom intensity, frequency patterns, and outbreak severity dynamics. Concurrently, integrating meteorological and water quality factors, we constructed a BP neural network model to predict bloom areas. Validation analyses demonstrated the strong performance of FY3D/MERSI data in cyanobacterial bloom extraction, showing excellent agreement with in-situ measurements (R² = 0.98, RMSE = 21.6km²). Cyanobacterial bloom areas in Lake Tai exhibited seasonal double-peak patterns, with peaks occurring in May and October. Bloom severity predominantly ranged from mild to moderate, while severe blooms remained at low levels throughout the year. Outbreak frequency was significantly higher in the western, southern coastal, and central lake areas than in the eastern region, and higher frequencies occurred during summer and autumn. Inter annual outbreak intensity showed a gradual decreasing trend, with the proportion of “no cyanobacterial blooms” rising to 89% in 2023. Correlation analysis indicated that factors such as temperature and relative humidity promote bloom outbreaks, while turbidity and fluorescent dissolved organic matter exerted inhibitory effects. The BP neural network model demonstrated robust predictive capability, achieving a root mean square error (RMSE) of 27.07km² and a Pearson correlation coefficient (r) of 0.94 on the validation set. Notably, 83% of predictions fell within ±25km² of observed values. These metrics collectively confirm the model's effectiveness in capturing the dynamic spatiotemporal trends of cyanobacterial bloom areas.

Key words

FY3D/MERSI / cyanobacteria blooms / time series monitoring / prediction model / BP neural network

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

ZHANG Ruo-lin, WANG Meng, MENG Qing-yan, SUN Yun-xiao, YUAN Xi-tun, ZHANG Lin-lin, WU Han-tian, SUN Zhen-hui, WANG Jia-long. Time-series monitoring and prediction of cyanobacteria blooms based on FY3D/MERSI data[J]. China Environmental Science. 2026, 46(3): 1509-1521

References

[1] 刘一良,张景,王丝丝,等.“全球生态环境遥感监测年度报告”回顾:2012~2021 [J]. 遥感学报, 2022,26 (10):2106-2120. Liu Y L, Zhang J, Wang S S, et al. Review of the 'global ecosystem and environment remote sensing monitoring annual report': 2012~ 2021 [J]. Journal of Remote Sensing, 2022,26(10):2106-2120.
[2] 侯润琳,文军,雷磊,等.基于风云三号气象卫星数据的太湖水体表层蓝藻水华特征研究 [J]. 农业灾害研究, 2024,14(6):232-234. Hou R L, Wen J, Lei L, et al. Research on the characteristics of blue-green algae blooms on the surface of taihu lake based on data from fengyun-3 meteorological satellite [J]. Journal of Agricultural Catastrophology, 2024,14(6):232-234.
[3] 郑超,童旭东,祝善友,等.基于多源影像数据与Otsu-RF方法的太湖蓝藻水华识别及监测 [J]. 测绘通报, 2025,(5):1-7. Zheng C, Tong X D, Zhu S Y, et al. Identification and monitoring of cyanobacterial blooms in taihu lake based on multi-source image data and otsu-rf algorithm [J]. Bulletin of Surveying and Mapping, 2025, (5):1-7.
[4] 金晓龙,邓学良,戴睿,等.基于多源数据的巢湖蓝藻水华时空分布及驱动因素分析 [J]. 环境科学, 2024,45(5):2694-2706. Jin X L, Deng X L, Dai R, et al. Analysis of the spatiotemporal distribution of algal blooms and its driving factors in chaohu lake based on multi-source datasets [J]. Environmental Science, 2024,45(5): 2694-2706.
[5] 龚铭,王百群,赵方凯,等.滇池蓝藻水华时空变化特征及环境影响因素 [J]. 水土保持研究, 2024,31(4):419-429. Gong M, Wang B Q, Zhao F K, et al. Spatiotemporal variations of cyanobacteria blooms in lake dianchi and the relationship with environmental factors [J]. Research of Soil and Water Conservation, 2024,31(4):419-429.
[6] 马洁莹.基于多源遥感数据的巢湖水华时空变化特征研究 [D]. 南京:南京信息工程大学, 2021. Ma J Y.Research on the spatio-temporal variation characteristics of algal blooms in chaohu lake based on multi-source remote sensing data [D]. Nanjing: Nanjing university of information science technology, 2021.
[7] 祁国华,马晓双,何诗瑜,等.基于多源遥感数据的巢湖水华长时序时空变化(2009~2018年)分析与发生概率预测 [J]. 湖泊科学, 2021, 33(2):414-427. Qi G H, Ma X S, He S Y, et al. Long-term spatiotemporal variation analysis and probability prediction of algal blooms in lake chaohu (2009~2018) based on multi-source remote sensing data [J]. Wetland Science, 2021,33(2):414-427.
[8] 潘佳敏,冯禹昊,谢平,等.基于小波—神经网络耦合模型对云南星云湖富营养化气象驱动因子的分析 [J]. 湖泊科学, 2021,33(2): 428-438. Pan J M, Feng Y H, Xie P, et al. Meteorological driving factor of eutrophication in lake xingyun,yunnan based on the coupling model of wavelet analysis neural network [J]. Wetland Science, 2021,33(2): 428-438.
[9] 张虎军,宋挺,朱冰川,等.太湖蓝藻水华暴发程度年度预测 [J]. 中国环境监测, 2022,38(1):157-164. Zhang H J, Song T, Zhu B C, et al. Annual forecast of the extent of cyanobacteria bloom in taihu lake [J]. Environmental Monitoring in China, 2022,38(1):157-164.
[10] Rousso B Z, Bertone E, Stewart R, et al. A systematic literature review of forecasting and predictive models for cyanobacteria blooms in freshwater lakes [J]. Water Research, 2020,182:115959.
[11] Liu X, Zhang Y, Shi K, et al. Mapping aquatic vegetation in a large,shallow eutrophic lake:a frequency-based approach using multiple years of modis data [J]. Remote Sensing, 2015,7(8):10295- 10320.
[12] Huang C, Li Y, Yang H, et al. Detection of algal bloom and factors influencing its formation in taihu lake from 2000 to 2011 by modis [J]. Environmental Earth Sciences, 2014,71:3705-3714.
[13] 王成林,潘维玉,韩月琪,等.全球气候变化对太湖蓝藻水华发展演变的影响 [J]. 中国环境科学, 2010,30(6):822-828. Wang C L, Pan W Y, Han Y Q, et al. Effect of global climate change on cyanobacteria bloom in taihu lake [J]. Chinese Environmental Science, 2010,30(6):822-828.
[14] Zhang Q, Xiao C. Cloud detection of rgb color aerial photographs by progressive refinement scheme [J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52(11):7264-7275.
[15] 汪雨豪.基于哨兵2A影像的太湖蓝藻水华提取方法研究及应用 [D]. 北京:中国地质大学(北京), 2020. Wang Y H.The research and application of the method of extracting cyanobacteria bloom in taihu lake based on sentinel 2a image [D]. Beijing: China University of Geosciences (Beijing), 2020.
[16] 全国卫星气象与空间天气标准化技术委员会(SAC/TC347).卫星遥感监测技术导则蓝藻水华:GB/T45424-2025 [S]. 北京:中国标准出版社, 2025. National technical committee 347 on satellite meteorology and space weather monitoring of standardization administration of china.Technical guidelines for satellite monitoring—cyanobacterial bloom:gb/t45424-2025 [S]. Beijing: Standards press of China, 2025.
[17] 张悦,陈冰,李旭文,等.基于Sentinel-2与随机森林算法的太湖水生植被分布监测 [J]. 环境监控与预警, 2023,15(6):42-49. Zhang Y, Chen B, Li X W, et al. Monitoring aquatic vegetation distribution of taihu lake from sentinel-2 and random forest algorithm [J]. Environmental Monitoring and Forewarning, 2023,15(6):42-49.
[18] 沙林伟,刘阁,温志丹,等.基于MODIS数据的太湖蓝藻水华时空变化规律研究 [J]. 湿地科学, 2018,16(3):432-437. Sha L W, Liu G, Wen Z D, et al. Spatio-temporal variation rules of cyanobacteria blooms in taihu lake using modis data [J]. Wetland Science, 2018,16(3):432-437.
[19] 李若楠,王琦,刘书明.基于典型相关系数和随机森林的水质预警方法 [J]. 中国环境科学, 2021,41(9):4457-4464. Li R N, Wang Q, Liu S M. Water quality warning method based on canonical correlation coefficient and random forest [J]. Chinese Environmental Science, 2021,41(9):4457-4464.
[20] 嵇晶晶,佘冬立,阿力木・阿布来提,等.基于优化BP神经网络的盐渍化农田排水沟水体溶存N₂O浓度预测 [J]. 中国环境科学, 2024,44(10):5839-5846. Ji J J, She D L, Alimu Abulati, et al. Predicting dissolved n₂o concentrations from drainage ditches in salt-affected farmlands based on optimized backpropagation neural network [J]. Chinese Environmental Science, 2024,44(10):5839-5846.
[21] 罗成明,卢文喜,王梓博,等.地下水污染模拟模型的不确定性分析 [J]. 中国环境科学, 2022,42(7):3224-3233. Luo C M, Lu W X, Wang Z B, et al. Uncertainty analysis of groundwater pollution simulation model [J]. Chinese Environmental Science, 2022,42(7):3224-3233.
[22] 张峤,邓贵仕.Levenberg-Marquardt神经网络在煤矿作业人员人因可靠性评价中应用研究 [J]. 大连理工大学学报, 2015,55(4): 424-430. Zhang Q, Deng G S. Investigation on application of levenberg- marquardt neural networks in human reliability evaluation of coal mine operators [J]. Journal of Dalian University of Technology, 2015, 55(4):424-430.
[23] 宋挺,徐超,严飞,等.基于Sentinel-2数据的湖泊蓝藻水华时空动态监测 [J]. 中国环境监测, 2025,41(1):214-224. Song T, Xu C, Yan F, et al. Spatio-temporal dynamics monitoring of lake cyanobacterial blooms based on sentinel-2data [J]. Environmental Monitoring in China, 2025,41(1):214-224.
[24] 王思琪.基于多源数据的太湖蓝藻水华风险评估及日内变化时空规律挖掘研究 [D]. 武汉:武汉大学, 2023. Wang S Q. Research on risk assessment of cyanobacterial bloom and spatio-temporal law mining on diurnal bloom changes in lake taihu based on multi-source data [D]. Wuhan: Wuhan University, 2023.
[25] 张娣.基于分级标准的太湖水域蓝藻气象环境影响因素与预测模型研究 [D]. 南京:南京信息工程大学, 2014. Zhang D. The study of meteorological environmental factors and predicting models of algal bloom in lake tai based on the grading standards [D]. Nanjing: Nanjing University of Information Science Technology, 2014.
[26] 罗晓春,杭鑫,曹云,等.太湖富营养化条件下影响蓝藻水华的主导气象因子 [J]. 湖泊科学, 2019,31(5):1248-1258. Luo X C, Hang X, Cao Y, et al. Dominant meteorological factors affecting cyanobacterial blooms under eutrophication in lake taihu [J]. Wetland Science, 2019,31(5):1248-1258.
[27] 何云川.基于贝叶斯网络的太湖叶绿素建模及预测 [D]. 北京:中国矿业大学, 2022. He Y C. Modeling and prediction of chlorophyll a in lake taihu based on bayesian network [D]. Beijing: China University of Mining and Technology, 2022.
[28] 韦洁琳.南水北调东线调蓄湖泊洪泽湖蓝藻时空分布特征及水华发生机制 [D]. 济南:山东大学, 2023. Wei J L.Spatio-temporal distribution characteristics of cyanobacteria and bloom mechanism in hongze lake on the east route of south-to- north water diversion project [D]. Jinan: Shandong University, 2023.
[29] 覃苗.福建平潭三十六脚湖蓝藻水华预警研究 [D]. 福州:福建师范大学, 2020. Qin M. Research on early warning of cyanobacteria blooms in sanshiliujiao lake of pingtan, fujian [D]. Fuzhou: Fujian Normal University, 2020.
[30] 付娆,张海洋,衣葵花,等.水体盐度对藜麦生长、离子积累及吸收氮磷营养盐的影响 [J]. 应用与环境生物学报, 2024,(9):88-95. Fu R, Zhang H Y, Yi K H, et al. Effects of salinity on growth,ion accumulation and absorption of nitrogen and phosphorus nutrients in quinoa under hydroponic conditions [J]. Chinese Journal of Applied and Environmental Biology, 2024,(9):88-95.
[31] 丁鹏,邹晓荣,许回,等.基于BP神经网络的长鳍金枪鱼渔获量与气候因子关系研究 [J]. 海洋学报, 2024,46(9):88-95. Ding P, Zou X R, Xu H, et al. Study on the relationship between catch of thunnus alalunga and climatic factors based on bp neural network [J]. Haiyang xuebao, 2024,46(9):88-95.
[32] 张晓忆,景元书,陈飞,等.基于RDALR模型分析气象条件对太湖蓝藻水华发生的影响及预报 [J]. 环境工程学报, 2016,10(10): 5721-5729. Zhang X Y, Jing S Y, Chen F, et al. Effect and forecast of weather conditions on cyanobacterial bloom outbreaks based on rdalr model in taihu lake [J]. Chinese Journal of Environmental Engineering, 2016,10 (10):5721-5729.
[33] 许明,刘伟京,白永刚,等.太湖蓝藻水华期可溶有机物的生物降解 [J]. 中国环境科学, 2018,38(9):3494-3501. Xu M, Liu W J, Bai Y G, et al. Biodegradation of dissolved organic matter in lake taihu during cyanobacterial blooms [J]. Chinese Environmental Science, 2018,38(9):3494-3501.
[34] 胡月红,张屹,赵芳,等.基于BP神经网络的湘江蓝藻水华定量遥感反演模型 [J]. 中国环境监测, 2024,40(6):239-248. Hu Y H, Zhang Y, Zhao F, et al. Research on the quantitative sensing images inversion model of cyanobacteria bloom in the xiangjiang river based on back propagation neural network [J]. Environmental Monitoring in China, 2024,40(6):239-248.
[35] 罗华军,刘德富,黄应平.基于遗传算法-支持向量机的水库叶绿素a浓度短期预测非线性时序模型 [J]. 水利学报, 2009,40(1):46-51. Luo H J, Liu D F, Huang Y P. Genetic algorithm-support vector machine model for short-term prediction of chlorophyll a concentration nonlinear time series [J]. Journal of Hydraulic Engineering, 2009,40(1):46-51.
[36] 吴娟,朱跃龙,金松,等.三种机器学习模型在太湖藻华面积预测中的应用 [J]. 河海大学学报(自然科学版), 2020,48(6):542-551. Wu J, Zhu Y L, Jin S, et al. Area prediction of cyanobacterial blooms based on three machine learning methods in taihu lake [J]. Journal of Hohai University (Natural Sciences), 2020,48(6):542-551.
[37] 欧阳添,闪锟,周博天,等.基于LSTM网络的在线藻类时序数据预测研究:以三峡水库为例 [J]. 湖泊科学, 2021,33(4):1031-1042. Ouyang T, Shan K, Zhou B T, et al. Research on the online forecasting of algal kinetics based on time-series data and lstm neural network:taking three gorges reservoir as an example [J]. Wetland Science, 2021,33(4):1031-1042.