Identification Of Algal Bloom And Aquatic Vegetation In China Based On Multi-source Remote Sensing Images

China has a diverse climate and many lakes,which not only perform ornamental,recreational and ecological functions,but are also an important source of drinking water for people.However,with the development of social and economic development,more and more lakes have been seriously polluted.In the China Ecological Environment Bulletin released by the Ministry of Ecology and Environment of the People’s Republic of China in 2020,28.1% of China’s lakes are in eutrophic state,and the outbreak of algal blooms is a manifestation of the highest stage of eutrophication in water bodies.The outbreak of algal blooms seriously affects the functions of lakes as drinking water sources,tourism and fishery farming,and also poses a serious threat to the health of the surrounding residents if algal blooms occur in drinking water sources,so effective monitoring of algal blooms is a matter of urgency.Remote sensing is a new monitoring method with the advantages of wide range,real time,fast and low cost,and is widely used.Since the proposal of remote sensing technology and its continuous development,researchers have proposed various model algorithms to enrich the algal bloom identification methods and improve the accuracy of algal bloom inversion.There are also some problems,firstly,when using the Floating Algal Index（FAI）alone to identify algal blooms,misclassification may occur in highly turbid water bodies;secondly,as algal blooms and aquatic vegetation are difficult to distinguish between them using remote sensing means due to similar spectral characteristics,these may reduce the accuracy of algal bloom identification,which is not conducive to our effective monitoring of algal blooms.In this paper,five different types of lakes were selected and MODIS and Sentinel 2images were used to distinguish algal blooms from aquatic vegetation in turbid water bodies based on the different phenological characteristics of algal blooms and aquatic vegetation.The accuracy of the MODIS images was verified to be above 76.6% and up to 97.48%,which can effectively distinguish between algal blooms and aquatic vegetation.At the same time,this paper uses the advantages of high temporal resolution of MODIS images to conduct statistical analysis of the spatial and temporal dynamics of algal blooms and aquatic vegetation,effectively reflecting the temporal and spatial characteristics of algal bloom outbreaks in lakes,changes in the area of algal blooms and the growth of aquatic vegetation in lakes.To address the problem of interference caused by turbid water when extracting algal blooms with a single index,this paper uses multiple spectral indices to effectively remove the interference of highly turbid water to extract vegetation information.To address the problem of interference caused by turbid water when extracting algal blooms with a single index,this paper uses multiple spectral indices to effectively remove the interference of highly turbid water to extract vegetation information.The Sentinel 2images with high spatial resolution were selected and a decision tree was constructed by using FAI,Normalized Difference Vegetation Index（NDVI）and Normalized Difference Water Index（NDWI）to find the sum of multiple indices to extract the information of vegetation,which effectively removed the interference caused by high turbidity.The accuracy of the vegetation information extracted from the images was significantly improved,with an accuracy of 96.1%.Using the improved vegetation presence frequency index（VPF）method,all four lakes involved in the accuracy verification achieved an accuracy of over 80%,enabling more accurate identification of algal blooms and aquatic vegetation.The five lakes selected in this paper are located in different climatic zones,in different lake areas and represent different lake types.Two kinds of images were used to identify algal blooms and aquatic vegetation in five typical lakes in China,respectively,to achieve the application of the algal bloom and aquatic vegetation differentiation method in different regions of the country.Two new lakes are also added to verify the wide applicability of the method.The best identification method for algal blooms and aquatic vegetation in several typical lake areas in China is also discussed,and provides a reference for the differentiation of other similar features.Tests of the spatial transferability of the method in separate lakes with different optical properties show its potential application in other turbid water bodies.This study provides strong methodological and theoretical support for future water quality monitoring in turbid water bodies.It is of practical relevance to the water environment management and governance sector,especially in the absence of practical measurement data.