| To maintain the healthy water ecosystem of East Lake Taihu and ensure the safety of its water source is of great significance to the drinking water safety of Wujiang District of Suzhou City and the water safety of the Yangtze River Delta eco-green integrated development demonstration zone.At present,there have been a lot of studies on chlorophyll-a concentration inversion models of algal lakes with high concentration of chlorophyll-a in Taihu Lake,and the chlorophyll-a inversion models generally directly remove the growth areas of aquatic vegetation.The chlorophyll-a inversion models of grassy lakes such as East Lake Taihu with low concentration of chlorophyll-a in Taihu Lake still need to be further studied.The method system of monitoring chlorophyll-a concentration in lake aquatic vegetation and water body is not complete.To solve the above problems,in order to reveal the temporal and spatial variation characteristics of chlorophyll-a concentration of aquatic vegetation and water source in East Lake Taihu,this study obtained the spectral data of aquatic vegetation and water through monitoring based on the satellite image of Sentinel-2,and constructed a classification extraction model of aquatic vegetation and water by using NDVI index to extract the distribution area of aquatic vegetation and water.An inversion model of chlorophyll-a in open water was established to analyze the spatio-temporal variation of aquatic vegetation and chlorophyll-a concentration in East Lake Taihu from 2017 to 2022.The main conclusions are as follows:(1)The aquatic vegetation extraction model of East Lake Taihu was established by using NDVI index.When the NDVI threshold was 0.24,that is,when NDVI≥0.24,the image element was determined to be aquatic vegetation;otherwise,it was determined to be water.The overall accuracy based on test set data is 100%.The results of accuracy verification based on 6 landscape images showed that the producer accuracy of aquatic vegetation reached more than 88%,the overall accuracy reached 96.10%,the root mean square error(RMSE)was 0.01km~2,and mean absolute percentage error(MAPE)was 3.90%.The overall classification accuracy was 96.56%,RMSE was 0.01km~2,and MAPE was 3.44%.(2)The distribution of aquatic vegetation in East Lake Taihu has obvious spatiotemporal heterogeneity.In terms of time scale,the area of aquatic vegetation fluctuated and decreased,reaching the lowest value of 41.49 km~2 in 2022,which decreased by 33.55%compared with the maximum of 62.43 km~2 in 2018.In terms of spatial distribution,the area of aquatic vegetation in the former East Jiaozui girth network decreased most significantly.Almost all the aquatic vegetation in area A disappeared,while the aquatic vegetation in areas B and C gradually shrank to the inner bank.(3)Multiple linear regression analysis method was used to establish the chlorophyll-a concentration inversion model in East Lake Taihu,and the accuracy and universal evaluation of the model results were carried out,and the optimal inversion model for Sentinel-2 satellite was selected.The results showed that the wavelength combination model had a better inversion accuracy for chlorophyll-a concentration of Sentinel-2 satellite images.Chlorophyll-a inversion model of Sentinel-2A is"Chla=69.701×(1/B4-1.2/B5+1.6/B6)×B7-94.312",In accuracy verification,RMSE was 2.17μg/L,MAPE was 22.12%,and R~2 was 0.8046.In universality evaluation,R~2 was 0.6315 for comparison between inversion and measured Chlorophyll-a concentration.Chlorophyll-a inversion model of Sentinel-2B is"Chla=63.434×(1/B4-1.6/B5+1.8/B6)×B8-68.632",In accuracy verification,RMSE was 2.64μg/L,MAPE was 19.10%,and R~2 was 0.6282.In the general evaluation,R~2 was 0.632 for comparison between inversion and measured Chlorophyll-a concentration.(4)From 2017 to 2022,the mean annual Chlorophyll-a concentration in the open water of East Lake Taihu was 8.05μg/L,and that in Miaogang and Tingzi Port intakes was 11.11μg/L and 10.30μg/L,respectively.In the period of cyanobacteria outbreak in the East Lake Taihu,cyanobacteria gathered from the center of the lake to the shore.The concentration of Chlorophyll-a in the southwest water was higher than that in the northeast lake,and that in the south of the horn mouth of Taipu River was higher than that in the north.On August 17,2019,the concentration of Chlorophyll-a in the intake of Miaogang and Tingzi Port water source exceeded the"major event"limit(60μg/L),and the maximum concentration of Chlorophyll-a in the intake reached 108.93μg/L and 95.41μg/L,respectively.In this study,a classification model of aquatic vegetation and water bodies in East Lake Taihu was proposed,and chlorophyll-a inversion model was established to form a complete set of remote sensing monitoring system of aquatic vegetation and chlorophyll-a in East Lake Taihu,providing scientific basis and technical support for water ecological management,water source protection and cyanobacteria bloom prevention and control in East Lake Taihu.It is of guiding significance to construct chlorophyll-a inversion model in grassy lakes or chlorophyll-a water bodies with low concentration. |
PDF Full Text Request