Remote Sensing Monitoring Of Algal Blooms In Hulun Lake Based On A Novel Spectral Index ABDI

With the global warming,the development of social economy and the increasing intensity of human activities,the ecological environment of China and even the whole world has been damaged.Among them,algal bloom caused by water eutrophication has become one of the most concerned problems in modern society.Large-scale algal blooms will damage the lake landscape,affect the development of fishery and tourism,and endanger the health of residents.Compared with traditional methods,remote sensing technology has become an important means of algal bloom detection and early warning,which can quickly and accurately obtain the information of algal bloom outbreak.In this study,Hulun Lake in Inner Mongolia was taken as the research area.The spectral characteristics of different water types were analyzed based on Sentinel-2 MSI image,and a new algal bloom detection index(ABDI)was constructed.Otsu algorithm was used to distinguish algal bloom in Hulun Lake.Then,the accuracy validation,sensitivity analysis and general applicability test of ABDI were carried out,and compared with other spectral indexes(FAI,AFAI,NDVI and EVI).Finally,the temporal and spatial characteristics of algal bloom in Hulun Lake from 2016 to 2020,the effects of meteorological factors(temperature,precipitation,sunshine duration,wind speed and pressure)on algal bloom area,and the meteorological conditions on the day of large-scale algal bloom outbreak were analyzed.(1)Based on the spectral reflectance of Sentinel-2 MSI image,the spectral characteristics of algal bloom,turbid water and clear water in Hulun Lake were analyzed.The results show that the reflectance of algal blooms is low in the red band,but high in the red edge band and near infrared band.The reflectance of turbid water is the same as that of some algal blooms from red edge to near infrared,but the reflectance of turbid water is higher than that of algal blooms and clear water.Therefore,based on the above characteristics,this study constructed the ABDI,which can effectively remove the influence of turbid water and accurately identify algal blooms.(2)The classification results of different spectral indices(FAI,AFAI,NDVI and EVI)were validated based on the visual interpretation results of algal blooms in Hulun Lake on August 24 and 31,2019.The results show that the overall accuracy(99.6%and 98.59%)and Kappa coefficient(0.95 and0.89)of ABDI are the highest,and the classification error is the lowest.Combined with the scatter plot,the extraction accuracy of algal blooms from ABDI was the highest(R~2=0.87,RMSE=0.67 km~2),which was better than other common spectral indexes such as FAI.The sensitivity of each index in turbid water and thin cloud was discussed.The results show that the sensitivity of ABDI is the lowest to the thin cloud and turbid water.In addition,algal blooms in Taihu Lake,Dianchi Lake,Chaohu Lake and Yellow Sea were successfully extracted by ABDI(R~2≥0.97).(3)The algal blooms in Hulun Lake occur in summer(July to September).In July,small algal blooms are dominant,while large algal blooms are occurred in late August to early September.Algal blooms are mainly distributed in the coastal waters,especially in the north,and less frequently in the center of the lake.The results show that higher temperature can promote the recovery and growth of algal blooms,but there is a lag due to the slow change of water temperature;Precipitation into the lake will bring in a lot of nutrients and promote the growth of algal blooms;Sunshine duration is closely related to overcast precipitation,so it has indirect effect on algal bloom area;Higher wind speed and high pressure will affect the vertical distribution of algal blooms and inhibit the formation of algal blooms.In addition,large-scale algal blooms often occur in the environment with the average temperature higher than 15℃,wind speed lower than 3 m/s,sunshine duration higher than 10 h,and air pressure lower than 945 h Pa.