| Lake is the important water resourse for human being. Due to both natural factors and human activities effects, facing with severe decline for water quality and eutrophic,which is directly affect the lake's ecological environment and human's production and life.Poyang Lake is China's largest freshwater lake,its water areas is wide and large amplitude,the sampling of the physical and chemical analysis using fixed-point detection method to monitor water quality is time-consuming,laborious and uneconomic, and it is difficult to reflect water quality of whole lake. Dynamic monitoring applied in water quality by remote sensing is undoubtedly the huge potential in modern water environmented monitoring, which can comprehensive,timely,fast and dynamic obtain water quality information from remote sensing data, and it's an important supplement of conventional water quality monitor, and it can provide scientific basis for Poyang Lake Eco-economic Region.Water quality monitors in Poyang Lake through Yangtze River and at the entrance of five rivers, its water spectral reflectance characteristics are revealed; Spectral response of water quality parameters such as total suspended solids (TSS),chlorophyll a (Chl-a),total phosphorus (TP),total nitrogen (TN),transparency (SD) is analyzed to identify the sensitivity of the measured hyperspectral remote sensing, then establishs retrieval models; On the lake's main water pollution parameters water quality retrieval models are established by sensitive bands of MODIS and TM/ETM+ data, this paper systematically studied its temporal-spatial distribution and variation and evaluated nutritional status of Poyang Lake. From several factors of hydrology change,wetland sediment covering types and human activities reseach to impact on water quality parameters, and finally put forward some suggestions. The main contents and conclusions are as follows:Researching field measurement and laboratory spectral measurement data of Poyang Lake. Above-water spectrum measurement method was employed to measure the in-situ spectrum of surface water. TSS spectrum experiment in lab shows the spectral reflectance was an increasing tread with TSS concertration increasing, spectral characteristics appear two peaks, reflectance change in first peak is greater amplitude, two reflection peaks appered a "red shift" phenomenon with the increase of suspended matter. Field spectrum measurement shows that a stable spectrum can obtain at 400-850nm, TSS and SD have the highest response at red and near infrared band, and suspended solids is the dominant contribution to the spectral reflectance of Poyang Lake.Owing to Chl-a concentration is low, spectral reflectance of suspended solids cover up water Chl-a's information,Chl-a and spectral reflectance appeared unrelatable, correlation of first-order derivative,peak ratio and Chl-a concentration significantly improved within the Chl-a fluorescence peak scope. TP,TN and Chl-a have high correlation, TP and first-derivative in red-band was significantly correlated,but TN spectral response is unstable. hyperspectral inverse regression models are established by bands sensitivity analysis.Inverse models of main contaminated water quality parameters are founded by MODIS and TM/ETM+ data.According to MODIS,TM/ETM+ band settings, correlation between TSS,TP,SD and band or bands combinations reflectance is analysised; Each statistical regression models for water quality parameters were established by high period, average period and low period; The results demonstrate that MODIS bandl-band4 are sensitive bands of water quality parameter,especially band1,band3 are the main characteristic bands,and result of statistical regression model is ideal.Except band2 of TM/ETM+ data sensitivity is weak, bands from bandl to band4 are the best bands of water quality parameter inversion; Regress model of TSS showed the highest correlation, Correlation between SD and TSS are negatively,TP and TSS have good correlation on the basis of data in 2009, TP estimated value with indirect inversion algorithm is better than direct inversion algorithm; Statistical regression retrieval models for each period reflects the optical properties of lake water in different period, the accuracy of models is better, but still less than universal. The RBF neural network model combined with the spectral reflectance data and water quality parameters data during whole periods, through a lot of training and testing to optimize adjustable parameters such as the number of nodes of neurons, the target error, SPREAD constants, the results show that strong universal model, in inversion accuracy to some extent than the traditional regression model. For analysing temporal and spatial characteristics of water quality in whole Poyang Lake, water quality parameters are retrievaled by using retrieval models and multi-temporal images. TSS concentration value take on increased trend for high water period to low water period,TSS concentration is perennial the highest in Duchang,Xingzi to Hukou,TSS concentration is lower in master lake area. From the high water period to low water period SD value of whole lake showed gradually smaller trend.TP content was declining from low water period to high water period,At low water period the lowest TP content is Xiuhe river at the entrance of five rivers,next is Xinjiang river,other rivers have almost same TP content.Adopting synthesis score method to obtain water nutrition state in Poyang Lake results indicate that eutrophication descending from low water period to high water period,and nutrition state in three periods is quite different, trophic state index trend to increase year by year.Considering influencing factors such as hydrological characteristics,wetland sediment cover types and human activity analyze furtherly the spatial and temporal distribution and variation of water quality parameters. water level changes of Poyang Lake greatly impact on water quality, TSS,TP are higher in low water level,with the water level increasing TSS and TP decreases, however SD is opposite, water quality has the rapid changes process characteristics in a certain range of water level;Five rivers flow is greater, TSS and TP are smaller,however SD is opposite,TSS,TP values in low water period are greater than high water period,when water level changed slightly, correlation between sediment concentration and TSS,TP is positive,then correlation with SD is negative. Research finds that wetland sediment cover in Poyang Lake has influence on the water quality parameters.and in upcovered water of dirrerent sediment type TSS,TP content change disciplinarian:sand> mudflats> grassland,and SD is opposite,With the rapid development of economy, human activities increased the deterioration of water quality in Poyang Lake.In order to realize to harmonize socio-economic and environmental development, and to achieve the sustainable development of Poyang Lake,this paper put forward the corresponding countermeasures and proposal:We should improve awareness of environmental protection; Adjust the industrial structure to appropriate industrial policy;Strengthen the pollution comprehensive management of five rivers;Control human activities which influence water quality in the lake area; Strengthening scientific research on wetland protection, Construct scientific decision-making and improve environmental supervised ability.
|
PDF Full Text Request