基于Sentinel-2的香港近海浊度遥感监测研究

孟仟; 黄珏

doi:10.12284/hyxb2025023

基于Sentinel-2的香港近海浊度遥感监测研究

doi: 10.12284/hyxb2025023

孟仟^1,,
黄珏^1, ,

山东科技大学测绘与空间信息学院，山东青岛 266590

基金项目: 国家自然科学基金（42076185）。

详细信息

作者简介:
孟仟（2001—），女，山东省泰安市人，研究方向为水色遥感。E-mail：mengqian@sdust.edu.cn

通讯作者:
黄珏，教授，研究方向为水环境遥感。E-mail：huangjue@sdust.edu.cn

中图分类号: TP79
计量
- 文章访问数: 162
- HTML全文浏览量: 45
- PDF下载量: 65
- 被引次数: 0
出版历程
- 收稿日期: 2024-10-29
- 修回日期: 2025-01-22
- 网络出版日期: 2025-02-21
- 刊出日期: 2025-04-30

A Study on remote sensing monitoring of nearshore turbidity in Hong Kong based on Sentinel-2

Meng Qian^1
,,
Huang Jue^{1
, ,}

College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, China

摘要

摘要: 浊度是评估水质状况的可靠指标之一，通过浊度监测能够有效反映水体健康状况，保障生态系统的可持续发展和水资源的安全利用。本文利用2016−2023年的Sentinel-2 MSI影像，通过构建遥感定量反演模型，分析8 a间香港近海浊度的时空分布特征和变化规律，并探究其主要影响因素。对比传统经验模型、随机森林（RF）模型、梯度提升决策树（GBDT）模型、K最近邻（KNN）模型，基于RF的浊度反演模型精度最高（R² = 0.71，RMSE = 1.77 NTU，MAE = 1.44 NTU）。结果表明，年均水体浊度的变化范围为4.02～4.16 NTU，近8 a呈波动下降趋势（−0.0243 NTU/a），且空间分布呈西北高东南低的特点；季均水体浊度从高到低依次为冬季（4.54 NTU）、秋季（4.03 NTU）、春季（3.86 NTU）、夏季（3.76 NTU）。本文利用气象数据及香港地区的污水处理投资数据，从自然环境和人类活动两方面分析影响浊度时空分布的因素。香港近海水体浊度与入海径流量、气温呈负相关关系，并受到香港污水处理这一人为因素的影响，此外热带气旋活动期内的降水量和风速与水体浊度变化显著相关。
- 香港近海 /
- 浊度 /
- Sentinel-2 /
- 遥感反演 /
- 时空分布
Abstract: Turbidity is a reliable indicator for assessing water quality conditions. Turbidity monitoring can effectively reflect the health status of water bodies and guarantee the sustainable development of ecosystems and the safe utilization of water resources. In this study, Sentinel-2 MSI images from 2016 to 2023 were employed in the construction of a quantitative inversion model based on measured data. The temporal and spatial distribution characteristics and variation rules of water turbidity in the Hong Kong coastal waters over the past eight years were analyzed, and the main influencing factors were explored. In comparison to the traditional empirical model, random forest (RF) model, gradient boosted decision tree (GBDT) model, and K Nearest Neighbor (KNN) model, the RF-based turbidity inversion model had the highest accuracy (R² = 0.71, RMSE = 1.77 NTU, MAE = 1.44 NTU). The results demonstrate that the annual average turbidity of the water body fluctuates between 4.02 and 4.16 NTU, exhibiting a downward trend over the past eight years (−0.0243 NTU/a). Additionally, the spatial distribution is high in the north-west and low in the south-east. The seasonal average water turbidity, in descending order, was as follows: winter (4.54 NTU), autumn (4.03 NTU), spring (3.86 NTU) and summer (3.76 NTU). Utilizing meteorological data and investment data on sewage treatment in Hong Kong, we analyzed the factors affecting the spatial and temporal distribution of turbidity in terms of the natural environment and human activities. Turbidity in Hong Kong’s offshore waters exhibits a negative correlation with inlet runoff and air temperature. Additionally, it is influenced by the anthropogenic factor of sewage treatment in Hong Kong. Furthermore, there is a significant correlation between precipitation and wind speed during the period of tropical cyclone activity and the change in turbidity in the water column.
- Hong Kong offshore waters /
- turbidity /
- Sentinel-2 /
- remote sensing retrieval /
- spatial and temporal distribution

HTML全文

图 1 研究区域及监测站点示意图

Fig. 1 Study area and schematic diagram of monitoring stations

下载: 全尺寸图片幻灯片

图 2 验证数据集反演模型性能比较

Fig. 2 Comparison of inversion model performance for validation dataset

下载: 全尺寸图片幻灯片

图 3 2016−2023年研究区域浊度年均变化

Fig. 3 Annual mean change of turbidity in the study area from 2016 to 2023

下载: 全尺寸图片幻灯片

图 4 2020年11月5日MODIS真彩色影像（a）、Sentinel-2 MSI真彩色影像（b）及MODIS浊度空间分布反演图（c）、Sentinel-2浊度空间分布反演图（d）

Fig. 4 MODIS true-color imagery (a), Sentinel-2 MSI true-color imagery (b) and inversion of MODIS spatial distribution of turbidity (c) and Sentinel-2 spatial distribution of turbidity (d) for November 5, 2020

下载: 全尺寸图片幻灯片

图 5 2016−2023年香港近海浊度逐年均值分布

Fig. 5 Annual mean distribution of turbidity in the offshore waters of Hong Kong from 2016 to 2023

下载: 全尺寸图片幻灯片

图 6 2016−2023年香港近海浊度月均分布

Fig. 6 Monthly mean distribution of turbidity in the offshore waters of Hong Kong from 2016 to 2023

下载: 全尺寸图片幻灯片

图 7 香港近海浊度季均值（a、b、c、d）和年均值（e）分布

Fig. 7 Seasonal mean (a, b, c, d) and annual mean (e) distribution of turbidity in the offshore waters of Hong Kong

下载: 全尺寸图片幻灯片

图 8 香港近海月均浊度与入海径流量（a）、降水量（b）、气温（c）、风速（d）的关系

Fig. 8 Relationship between monthly mean turbidity and inlet runoff (a), precipitation (b), air temperature (c) and wind speed (d) in the Hong Kong offshore waters

下载: 全尺寸图片幻灯片

图 9 自然环境因素贡献度（a）及香港近海年均浊度与污水处理投资的关系（b）

Fig. 9 Contribution of natural environmental factors and relationship between annual mean turbidity and investment data on sewage treatment in the Hong Kong offshore waters

下载: 全尺寸图片幻灯片

表 1 Sentinel-2 MSI影像光谱波段信息

Tab. 1 Sentinel-2 MSI image spectral band information

波段号	波段	中心波长/μm	空间分辨率/m	波段号	波段	中心波长/μm	空间分辨率/m
B1	气溶胶	0.443	60	B8	近红外	0.842	10
B2	蓝	0.490	10	B8A	窄波近红外	0.865	20
B3	绿	0.560	10	B9	水蒸气	0.945	60
B4	红	0.665	10	B10	卷云	1.375	60
B5	红边1	0.705	20	B11	短波红外	1.610	20
B6	红边2	0.740	20	B12	短波红外	2.190	20
B7	红边3	0.783	20

下载: 导出CSV

表 2 经验模型表达式及机器学习模型最优参数组合

Tab. 2 Empirical model expressions and optimal parameter combinations for machine learning models

模型	变量	表达式/参数
多项式模型	LH₄₅₆	y = 666.14x² + 76.2x + 4.11
指数函数模型	B3-B2，(B3+B5)/(B2/B3)	${ {y}=1.95{\mathrm{e}}^{(3.56{\mathrm{x}}_{1}+0.68{\mathrm{x}}_{2})} }$
RF	B4， B5， B3-B2， LH₄₅₆， LH₅₆₇， (B3+B5)/(B2/B3)	n_estimators：25 min_samples_split：4 max_depth：15
GBDT		n_estimators：100 learning_rate：0.08 min_samples_split：3 max_depth：5
KNN		n_neighbors：15 algorithm：kd_tree weights：uniform
注：在多项式模型中，变量x为LH₄₅₆；指数函数模型中变量x₁、x₂分别为B3-B2，(B3+B5)/(B2/B3)。

下载: 导出CSV

表 3 Sentinel-2典型波段及波段组合与浊度的相关性分析

Tab. 3 Correlation analysis of typical Sentinel-2 bands and band combinations with turbidity

波段组合	相关系数	波段组合	相关系数
B4	0.42^**	LH₄₅₆	0.76^**
B5	0.42^**	LH₅₆₇	−0.69^**
B3-B2	0.60^**	(B3+B5)/(B2/B3)	0.57^**
注：**表示相关性在0.01水平上显著。

下载: 导出CSV

表 4 建模数据集各模型性能对比

Tab. 4 Comparison of the performance of each model in the modelled dataset

模型	R²	RMSE/NTU	MAE/NTU
多项式模型	0.67	1.97	2.36
指数函数模型	0.59	2.07	1.52
RF	0.91	1.11	0.68
GBDT	0.80	1.55	1.28
KNN	0.63	1.20	1.27

下载: 导出CSV

参考文献(33)

[1]	Dogliotti A I, Ruddick K G, Nechad B, et al. A single algorithm to retrieve turbidity from remotely-sensed data in all coastal and estuarine waters[J]. Remote Sensing of Environment, 2015, 156: 157−168. doi: 10.1016/j.rse.2014.09.020
[2]	Nienhuis J H, Ashton A D, Edmonds D A, et al. Global-scale human impact on delta morphology has led to net land area gain[J]. Nature, 2020, 577(7791): 514−518. doi: 10.1038/s41586-019-1905-9
[3]	McCarthy M J, Muller-Karger F E, Otis D B, et al. Impacts of 40 years of land cover change on water quality in Tampa Bay, Florida[J]. Cogent Geoscience, 2018, 4(1): 1422956. doi: 10.1080/23312041.2017.1422956
[4]	Li Jian, Chen Xiaoling, Tian Liqiao, et al. Improved capabilities of the Chinese high-resolution remote sensing satellite GF-1 for monitoring suspended particulate matter (SPM) in inland waters: radiometric and spatial considerations[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 106: 145−156. doi: 10.1016/j.isprsjprs.2015.05.009
[5]	Liu Ge, Li Lin, Song Kaishan, et al. An OLCI-based algorithm for semi-empirically partitioning absorption coefficient and estimating chlorophyll a concentration in various turbid case-2 waters[J]. Remote Sensing of Environment, 2020, 239: 111648. doi: 10.1016/j.rse.2020.111648
[6]	Peterson K T, Sagan V, Sloan J J. Deep learning-based water quality estimation and anomaly detection using Landsat-8/Sentinel-2 virtual constellation and cloud computing[J]. GIScience & Remote Sensing, 2020, 57(4): 510−525.
[7]	Zheng Lufei, Qiu Zhongfeng, Zhou Yan, et al. Comparisons of algorithms to estimate water turbidity in the coastal areas of China[J]. International Journal of Remote Sensing, 2016, 37(24): 6165−6186. doi: 10.1080/01431161.2016.1256510
[8]	Wang Shengqiang, Mao Ying, Zheng Lufei, et al. Remote sensing of water turbidity in the Eastern China Seas from Geostationary Ocean Colour Imager[J]. International Journal of Remote Sensing, 2020, 41(11): 4080−4101. doi: 10.1080/01431161.2020.1714775
[9]	El-Alem A, Chokmani K. A machine learning-based regional hybrid model for remote retrieving turbidity from Landsat imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 8021605.
[10]	Drusch M, Del Bello U, Carlier S, et al. Sentinel-2: ESA’s optical high-resolution mission for GMES operational services[J]. Remote Sensing of Environment, 2012, 120: 25−36. doi: 10.1016/j.rse.2011.11.026
[11]	Tian Liqiao, Wai O W H, Chen Xiaoling, et al. Assessment of total suspended sediment distribution under varying tidal conditions in deep bay: initial results from HJ-1A/1B satellite CCD images[J]. Remote Sensing, 2014, 6(10): 9911−9929. doi: 10.3390/rs6109911
[12]	Nazeer M, Nichol J E. Combining Landsat TM/ETM+ and HJ-1 A/B CCD sensors for monitoring coastal water quality in Hong Kong[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(9): 1898−1902. doi: 10.1109/LGRS.2015.2436899
[13]	Hafeez S, Wong M S. Measurement of coastal water quality indicators using Sentinel-2; an evaluation over Hong Kong and the Pearl River Estuary[C]//IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium. Yokohama: IEEE, 2019: 8249−8252.
[14]	Kuhn C, De Matos Valerio A, Ward N, et al. Performance of Landsat-8 and Sentinel-2 surface reflectance products for river remote sensing retrievals of chlorophyll-a and turbidity[J]. Remote Sensing of Environment, 2019, 224: 104−118. doi: 10.1016/j.rse.2019.01.023
[15]	Warren M A, Simis S G H, Selmes N. Complementary water quality observations from high and medium resolution Sentinel sensors by aligning chlorophyll-a and turbidity algorithms[J]. Remote Sensing of Environment, 2021, 265: 112651. doi: 10.1016/j.rse.2021.112651
[16]	Wilson R T. Py6S: a Python interface to the 6S radiative transfer model[J]. Computers & Geosciences, 2013, 51: 166−171.
[17]	Nazeer M, Nichol J E, Yung Y K. Evaluation of atmospheric correction models and Landsat surface reflectance product in an urban coastal environment[J]. International Journal of Remote Sensing, 2014, 35(16): 6271−6291. doi: 10.1080/01431161.2014.951742
[18]	Ouma Y O, Noor K, Herbert K. Modelling reservoir chlorophyll-a, TSS, and turbidity using Sentinel-2A MSI and Landsat-8 OLI satellite sensors with empirical multivariate regression[J]. Journal of Sensors, 2020, 2020: 21.
[19]	Murphy S, Hård J. Atmospheric correction of Sentinel 2 imagery in Google Earth Engine using Py6S[EB/OL]. (2020−10−16) [2024−10−29]. https://github.com/samsammurphy/gee-atmcorr-S2. Murphy S, Hård J. Atmospheric correction of Sentinel 2 imagery in Google Earth Engine using Py6S[EB/OL]. (2020−10−16) [2024−10−29]. https://github.com/samsammurphy/gee-atmcorr-S2.
[20]	Kwong I H Y, Wong F K K, Fung T. Automatic mapping and monitoring of marine water quality parameters in Hong Kong using Sentinel-2 image time-series and Google Earth Engine cloud computing[J]. Frontiers in Marine Science, 2022, 9: 871470. doi: 10.3389/fmars.2022.871470
[21]	Gons H J. Optical teledetection of chlorophyll a in turbid inland waters[J]. Environmental Science & Technology, 1999, 33(7): 1127−1132.
[22]	Gower J, King S, Borstad G, et al. Detection of intense plankton blooms using the 709 nm band of the MERIS imaging spectrometer[J]. International Journal of Remote Sensing, 2005, 26(9): 2005−2012. doi: 10.1080/01431160500075857
[23]	Topp S N, Pavelsky T M, Jensen D, et al. Research trends in the use of remote sensing for inland water quality science: moving towards multidisciplinary applications[J]. Water, 2020, 12(1): 169. doi: 10.3390/w12010169
[24]	Ma Yue, Song Kaishan, Wen Zhidan, et al. Remote sensing of turbidity for lakes in northeast China using Sentinel-2 images with machine learning algorithms[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 9132−9146. doi: 10.1109/JSTARS.2021.3109292
[25]	Wang Jun, Tong Yan, Feng Lian, et al. Satellite-observed decreases in water turbidity in the Pearl River Estuary: potential linkage with sea-level rise[J]. Journal of Geophysical Research: Oceans, 2021, 126(4): e2020JC016842. doi: 10.1029/2020JC016842
[26]	李盈盈. 辽河口浊度遥感反演及时空变化分析[D]. 大连: 大连理工大学, 2022. Li Yingying. Water turbidity retrieval and spatial-temporal analysis based on remote sensing data in the Liao River Estuary[D]. Dalian: Dalian University of Technology, 2022.
[27]	黎栩霞, 王裕东, 肖佑鹏, 等. 深圳近岸海域水质遥感监测及时空变化[J]. 环境工程, 2024, 42(1): 243−252. Li Xuxia, Wang Yudong, Xiao Youpeng, et al. Quality monitoring of Shenzhen’s coastal waters by satellite and its spatiotemporal variation[J]. Environmental Engineering, 2024, 42(1): 243−252.
[28]	Hou Xuejiao, Feng Lian, Duan Hongtao, et al. Fifteen-year monitoring of the turbidity dynamics in large lakes and reservoirs in the middle and lower basin of the Yangtze River, China[J]. Remote Sensing of Environment, 2017, 190: 107−121. doi: 10.1016/j.rse.2016.12.006
[29]	Lin Jie, Zou Xinqing, Huang Faming. Effects of the thermal discharge from an offshore power plant on plankton and macrobenthic communities in subtropical China[J]. Marine Pollution Bulletin, 2018, 131: 106−114. doi: 10.1016/j.marpolbul.2018.04.005
[30]	Jiang Cong, Xiong Lihua, Wang Dingbao, et al. Separating the impacts of climate change and human activities on runoff using the Budyko-type equations with time-varying parameters[J]. Journal of Hydrology, 2015, 522: 326−338. doi: 10.1016/j.jhydrol.2014.12.060
[31]	Zheng Zhubin, Li Yunmei, Guo Yulong, et al. Landsat-based long-term monitoring of total suspended matter concentration pattern change in the wet season for Dongting Lake, China[J]. Remote Sensing, 2015, 7(10): 13975−13999. doi: 10.3390/rs71013975
[32]	Lu Jian, Jiang Jingbo, Li Anchun, et al. Impact of Typhoon Chan-hom on the marine environment and sediment dynamics on the inner shelf of the East China Sea: in-situ seafloor observations[J]. Marine Geology, 2018, 406: 72−83. doi: 10.1016/j.margeo.2018.09.009
[33]	Liu Dong, Bai Yan, Wei Xiaodao, et al. Sewage treatment decreased organic carbon resources in Hong Kong waters during 1986−2020[J]. Environmental Pollution, 2023, 335: 122219. doi: 10.1016/j.envpol.2023.122219