Dynamics And Impact Factors Of Aquatic Vegetation In Inland Lakes Using MODIS Data

Aquatic vegetation is an important part of the lake ecosystem,which plays an important role in inhibiting sediment resuspension,improving water transparency,absorbing nutrients,inhibiting the growth of phytoplankton,absorbing heavy metals,and restoring the lake wetland ecosystem.Therefore,monitoring aquatic vegetation is of great significance for monitoring lake ecosystem security.At present,using remote sensing to monitor aquatic vegetation is mostly for a single lake based on a single scene image,which is difficult to meet the long-term monitoring needs of regional lakes.Given this,this study uses the phenological characteristics of vegetation to develop a recognition model of aquatic vegetation,which can be applied to multiple inland lakes,solving the threshold problem needed to determine for single scene image recognition,reflecting the growth information of aquatic vegetation within the year,and meeting the long-term monitoring needs of regional lakes.In this study,MODIS-Aqua images were used to construct NDVI(Normalized Difference Vegetation Index)long time series images.After comparing different time series curve smoothing methods and selecting one suitable for all kinds of surface features in the lake,analyzing the phenological characteristics of aquatic vegetation,and determining the standard curve,a recognition model of aquatic vegetation based on spectral matching was constructed by combining spectral angle with Euclidean distance.The temporal and spatial distribution of aquatic vegetation in Lake Taihu and Lake Hongze from 2003 to 2019 was obtained,and the influence of meteorological factors and nutrient concentrations on aquatic vegetation was analyzed.The main conclusions are as follows:(Note:Floating leaf vegetation in this study includes floating leaf vegetation,emergent vegetation and floating vegetation.)(1)Asymmetric Gaussian function fitting(A-G filter)can be applied to smooth the NDVI time series curves of floating leaf vegetation,submerged vegetation,cyanobacteria bloom and water body in inland lakes,and can effectively distinguish four types of surface features.(2)There are 4 types of typical NDVI time series curves for floating leaf vegetation,3 types of typical submerged vegetation,10 types of typical cyanobacteria bloom and 10 types of typical water body in Lake Taihu.The phenological differences among the floating leaf vegetation are obvious.The beginning time of the growing season is between 54.22 days and 111.7 days,and the length of the growing season is between 165.9 days and 305.5 days.The shape of the NDVI time series curve of submerged vegetation is similar to that of floating leaf vegetation,but the value is much lower than that of floating leaf vegetation.The start time of the growing season is between 90.24 days and 146.1 days,and the length of the growing season is between117.9 and 265 days.The NDVI time series curve of cyanobacteria bloom is in a peak state,and its peak may appear at any time of the year.The NDVI time series curve of the water body is relatively flat,and the peak value of 60%curve is lower than-0.15.(3)Taking the typical NDVI time series curve as the standard curve,the spectral matching model was constructed by using spectral angle and Euclidean distance,and the original NDVI data of time series image(NDVI-MVC data after MVC synthesis)was introduced to improve the classification accuracy of some cyanobacteria bloom and submerged vegetation.The accuracy of the developed spectral matching model is87.62%,and the Kappa coefficient is 0.829.(4)From 2003 to 2019,the aquatic vegetation in Lake Taihu and Lake Hongze changed significantly.The aquatic vegetation in Lake Taihu showed a downward trend before 2007,picked up in 2008,and showed a relatively stable state from 2008 to 2014.After a sharp decline in 2015,the aquatic vegetation showed an upward trend until 2019.However,the current coverage area of aquatic vegetation is still small,only 355.44km~2,including 219km~2 floating leaf vegetation,accounting for 61.615%,and 136.44km~2submerged vegetation,accounting for 38.39%.The aquatic vegetation of Lake Hongze showed an upward trend from 2003 to 2014 and decreased to 284.5km~2 in 2015,and then increased from 2015 to 2019.Through multiple linear regression analysis and redundancy analysis,it was found that on the inter-annual scale,meteorological factors such as annual average wind speed,annual average temperature and accumulated precipitation from 20 to 20 o’clock had no significant impact on the aquatic vegetation of Lake Taihu and Lake Hongze,and the influences of nutrient concentrations such as total nitrogen and total phosphorus on the aquatic vegetation of Lake Taihu were not significant.