| The middle and lower basins of the Yangtze River(MLY)is the region with most freshwater lakes in China.Global warming and intensive human activities in the basin exerted obvious impacts on lakes' ecological environment.Remote sensing(RS)satellites can provide objective,large-scale and long-time datasets for land surface research.Landsat series satellites have been continuously observed for decades since the first satellite was launched in 1972.Due to its high data consistency,Landsat satellite data has great potential in investigating inland lake water environment.Water body has high absorption and low reflection,and thus it is necessary to remove the atmospheric signal as much as possible when using satellite remote sensing data to study water environment.However,Landsat satellite is designed for land application.Owing to the lack of water color band and low signal-to-noise ratio,vicarious calibration and atmospheric correction for Landsat are more difficult than those of water color satellite.In addition,water color is an important indicator of the comprehensive state of water quality.But researches over RS-based water environment in MLY mainly focused on water area,water quantity,traditional water color parameters,etc.Extensive investigation on water color and water quality at the basin scale is relatively rare.Therefore,this thesis firstly evaluated and compared various coefficients of vicarious calibration and atmospheric correction methods of Landsat-8 OLI data so as to select an atmospheric correction method which shows overall optimal performance for Landsat data.Then the Landsat-derived Forel-Ule Index(FUI)model was used to to study the temporal and spatial change of water color of lakes in MLY from 1984 to 2018,as well as the driving factors affecting multi-decadal Water Color dynamics.The corresponding work and the main conclusions of this thesis can be summarized as follows:(1)Verification of vicarious calibration and atmospheric correction methods: AERONET-OC datasets was used to assess the correction effect of three sets of vicarious calibration coefficients and five atmospheric correction methods.After comparison,the atmospheric correction method with overall optimal performance was firstly recommended among all methods in this thesis,namely exponential extrapolation(EXP)algorithms based on NIR/SWIR1.Then the derived Remote sensing reflectance(Rrs)products with a relatively higher precision was applied to obtain water-color retrievals with more accuracy.(Three sets of vicarious calibration coefficients are Pahlevan-R,Pahlevan-L and unity.The AC methods include exponential extrapolation(EXP)algorithms of three kinds of different band combinations embedded in ACOLITE as well as the image-based dark spectrum fitting(DSF)method,Landsat surface reflectance(SR)product,a simple aerosol correction method and iterative NIR algorithm with 2-bands(NIR/SWIR2)model selection in Sea DAS)(2)Analysis of spatiotemporal distribution of lacustrine water color: one of the immediate parameter of the Landsat-based FUI model,i.e.hue angle with continuous value,was applied instead of FUI with discrete values.Based on the cluster of the water color dynamics of 61 large lakes in MLY in 35 years,the interannual changes of water color are classified into three categories.First,water color of lakes generally tends to yellow to brown,and the hue angle tends to decline in the middle years,namely water color tends to green or blue.Second,water color generally tends to yellow to brown as well,but the hue angle had a tendency of increase in recent years which meant the water color has further deepened.Third,the overall water color of lakes tends to be green to blue,but there is no uniform and obvious interannual change trend of water color.According to the results of quantitative analysis of two statistics(change rate and the significance level),41% of the lakes showed significant water color annual changes,among which 16 lakes showed significant negative changes(water color inclined to green to blue and water nutrition level decreased)and 9 ones showed significant positive changes(water color inclined to yellow to brown and water nutrition level increased).(3)Analysis of natural factors as driving factors of lake water color: natural factor datasets(runoff,wind-speed,NDVI)were introduced in this thesis as the driving factors to further explore causes of water color dynamics.The results indicated that quarterly changes of water color of 33 lakes from 1984 to 2018 were significantly related to natural factors,among which lots of lakes were driven by wind speed.Although the maximum Pearson's correlation coefficient(R)was only around 0.5,strong relationships were found between water colors of 10,7 and 22 lakes and runoff,NDVI and wind speed respectively.Because most of the lakes in the basin are shallow lakes,the wind can disturb different water layers in the lake,or even cause the re-suspension of substances at the bottom of the lake.This could render the surface water more turbid,which was displayed as increasing hue angle.(4)Quantifying the influence of reclamation on lacustrine water color: based on the reclamation dataset,151 point pairs(reclamation point,natural water point and lake control point)were selected by sampling in 61 lakes of MLY.By setting up the control experiments,overall spatiotemporal differences of water color were observed between reclaimed areas and natural waters,which was demonstrated by statistics.Taking a reclamation of Lake Beimayang and Lake Yanxi as an example,different change trends of reclamations and natural waters were observed when the time of reclamation segmented with lakes got longer.But the absolute difference of water color between reclamations and natural waters in both lakes became significantly larger.Those indicated spatiotemporal difference of water color happened between certain reclamations and natural waters due to lacustrine flows were blocked.In addition,the In-situ water-quality parameters(concentration of chlorophyll-a and dissolved organic carbon)measured in certain reclaimed areas and corresponding natural waters of lakes proved the influence of reclamation on the water quality again.And water quality might improve or deteriorate. |
PDF Full Text Request