Identification Of Water Colour Anomaly And Analysis Of Temporal And Spatial Variation Of Water Colour Typical Inland Waters Based On Sentinel-2 Satellite

Recently,the water pollution problem in China has become increasingly serious,especially in large lakes and reservoirs,such as algal blooms,river and pit pollution.Monitoring the changes in aquatic environment and recognition of water colour anomaly using satellite-based remote sensing technology has the advantages of reduced costs,high efficiency,wide application range,and inclusion of real-time dynamic monitoring.Meanwhile,it has scientific and social significance for monitoring the change trends and determining the driving factors of inland water quality.Due to seasonal and regional differences,the inland water composition and aerosol characteristics are complex.Additionally,the inland water is not concentrated and widely distributed and thus will be affected by the land proximity effect and other related problems.Therefore,it is a challenge that what establishes a universal monitoring model for inland water quality based on remote sensing.To conduct remote sensing-based monitoring of inland water quality over a wide range,Forel-Ule Index(FUI),which is not affected by the changes in the water composition and light field,is an important optical parameter.FUI has a good correlation with water clarity.In this study,Baiyangdian Lake,Qiqihar City,Xiong’an New Area,and Yangtze River were selected as study areas,and the FUI,hue angle,and water clarity(Secchi disk depth,透明度)were selected as water quality evaluation indexes to analyse the temporal and spatial variations in water colour typical inland water bodies and to identify anomaly water bodies.(1)The study was conducted in Qiqihar City of the Heilongjiang Province and Xiong’an New Area of the Hebei Province.The measured spectral differences and water colour of the normal and anomaly water bodies in the two study sites were analysed to investigate the influence of the external environment on water colour.A method for screening wide-ranged water bodies with anomaly water colour was proposed in this study based on the water colour and hue angle.Moreover,the method constructed a model based on the extraction of hue anglefrom Sentinel-2 MSI(multi-spectral instrument,MSI)data and calibration of linear fitting formula and accuracy improvement of the hue angle.Subsequently,the R~2,the root mean square error(RMSE),mean relative error(MRE)values of the model were 0.9029,4.397°,and 1.744%,respectively.The study indicated that when the screening threshold of anomaly water colour was set at a hue angleof 230.958°,the proposed model exhibited good stability and acceptance and thus can be used as a reference to screen anomaly water colour.(2)As for Baiyangdian Lake,it is an important water resource in Beijing city,Tianjin city and Hebei Province,and FUI is taken as the water quality evaluation parameter.Based on the Sentinel-2 MSI data,an extraction method for the FUI was developed.Subsequently,the MRE and RMSE values of the method were3.54%and 0.57,respectively.Furthermore,the FUI is inversely proportional to the degree of water clarity.Using this method,the study calculated the FUI and time series data of the Baiyangdian Lake from April to November during 2016-2020 on local and Google Earth Engine(GEE)cloud platform.And we analysed the spatial distribution patterns,seasonal variations,and inter-annual variations of the lake.To precisely analyse the spatiotemporal variation in the FUI of the Baiyangdian Lake,a partition statistical method suitable for remote sensing of water colour was proposed.The Baiyangdian Lake was divided into seven sub regions by water depth and functional area.The spatial distribution results indicated that the FUI value was high in the south and low in the north.The seasonal variations in the Baiyangdian Lake and seven sub-regions were consistent,with turbid water in spring and autumn,and clear water in summer.Inter-annual variations analyses for 2016-2020 indicated that the zone of A became progressively turbid,whereas the B、C、D、E、F and G zones exhibited slow and gradually decreasing trends.Our findings suggest that the overall water quality of Baiyangdian Lake may be better,which may be related to the governace polices of the region.(3)The study was conducted in Yangtze River.And it is the longest river in Asia.Furthermore,based on the Sentinel-2 MSI data,a water clarity retrieval algorithm was constructed,based on which a water clarity retrieval model was developed using the FUI and hue angle.Subsequently,the accuracy of the model results was evaluated using the water clarity retrieval results of the MODIS data as true reference values.The R~2,RMSE,MRE and Pearson correlation coefficient values of the model were 0.8854,0.07m,and 14.0%,respectively,while the correlation coefficient was 0.93.The method was also used to calculate the water clarity and time series data of the Yangtze River from January to December during 2017-2020 on local and GEE platform.Additionally,the FUI spatial distribution pattern,seasonal variations,and inter-annual variations of the region were analysed.Consequently,the spatial distribution pattern results of the SDD of Yangtze River indicated a high SDD in the west and a low SDD in the east.The main driving factors affecting the Yangtze River SDD was sediment runoff,water level,and precipitation.The upstream and downstream Yangtze River SDD was negatively correlated with the change in precipitation,water level and sediment runoff,whereas the midstream Yangtze River SDD was positively correlated with the change in water level.The upper and lower reaches of the Yangtze River and overall SDD showed a weak downward trend,and the middle reaches of the Yangtze River remained almost unchanged.The study constructed a model based on the extraction of hue angleand FUI from Sentinel-2 MSI data and calibration of linear fitting formula and accuracy improvement of the hue angleand FUI.And the study indicated that when the screening threshold of anomaly water colour was set at a hue angle.Taking the FUI as the water quality evaluation index,the temporal and spatial variation of water color in Baiyangdian water area is analyzed;A water clarity retrieval algorithm was constructed,based on which a water clarity retrieval model was developed using the FUI and hue angle.And the model is applied to the main stream of the Yangtze River,and the spatio-temporal change analysis of SDD was carried out.There are 49 figures,11 tables and 211 references.