Simulation And Retrieval To Water Quality Parameters By Remote Sensing In Shitoukoumen Reservoir Of Changchun

China's water pollution is a serious problem. Remote monitoring of water quality is by the relationship between water body reflectance spectral characteristics and concentration of water quality indicators, establish water quality indicators inversion algorithm. Methods commonly used are mainly empirical, semi-empirical and analytical methods. As the relationship between remote sensing images data and water quality parameters is nonlinear, the neural network model commonly uses hidden nonlinear regression model. Artificial neural networks with its adaptive, self-organizing and fault-tolerant exactly suited for simulation of complex relationship. Therefore, this research uses neural network model.At present the water quality model uses the Streeter-Phelps model, the QUAL model, the WASP model, the BASINS model, the OTIS model, the MIKE model and so on. WASP model can simulate a lot of elements of water quality to meet the needs of research, which is applicable to the steady process and unsteady water process of rivers, lakes, reservoir and other water bodies. It has been successfully exploited by many scholars at home and abroad to get high precision simulation of water, and has withstood the test of practice.The shitoukoumen reservoir is the most important water supply reservoir for Changchun. From the land use, its water quality is unceasingly developing to the bad direction. ASTER data which has the high space, spectrum and radiation resolution is the rare data resources for geology and the ecologicalenvironment research.Using the atmospheric correction FLAASH in ENVI, to realize the atmospheric correction to the band 1,band 2 and band 3 of ASTER data, and eliminate the radiation distortion of remote sensing digital images caused by atmosphere, so as to access the real features of radiation information. After correction, abstracting the DN value of pixel, in accordance with the calibration information of ASTER data, transfer DN value into radiance values. Then import to a well-trained two-tire BP network with a hidden layer. Each input node representative a band of visible light radiation brightness value of ASTER data. The data of input layer are distributed to the node of each hidden layer for computation. The output of hidden layer becomes the input of output layer for computation again. The output of output layer is the water quality indicator dissolved oxygen(DO), biochemical oxygen demand(BOD5)and ammonia nitrogen(NH3-N)on the distribution of pixel concentration, which is what we will interested. Finally, the map shows Shitoukoumen reservoir area regional water quality objective concentration distribution.This paper introduced the WASP model, and it determines the simulation to the water quality objectives for the target are dissolved oxygen(DO), biochemical oxygen demand(BOD5)and ammonia nitrogen(NH3-N).EUTRO which is the description of DO balance has four levels of complexity, and I choose the simplest S-P equation simulation. The study area will be divided into 7 paragraphs, following the importation documents establishment. Using trial-and-error method, according to the seven actual sampling data for the needed first three model parameters, rates: Reaeration rate k 2=1.028 d?1 ,Nitrification rate @ 20°C k1 2=0.11 d?1 , Deoxygenation rate @ 20°C k d=0.17 d ?1 , Oxygen to carbon ratio aO C=32/12, non-predacity Phytoplankton death velocity k1 D=0.005 d?1 , Half saturation constant for oxygen limitation K BOD=0.5 mg /L ,Half saturation constant for oxygen limitation K NO3 ? N=0.1 mg /L , Sediment Oxygen demand S SOD=0.8 g ? m ?2 ? d?1, Phytoplankton growth rate GP 1=0.4 d?1 , the algal respiration rate constant at 20°C k1 R=0.125 d?1 .Running the model, comparing the simulation data and measured results matched well, the relative error of 10% is less than 61.9%.The relative error less than 15% for 95.2%,which the simulated average relative error of DO is 10.05%,7.18% for BOD5, and 8.09% for NH3-N.Using the water concentration simulated from WASP model, based on surface water quality standards, it makes a comprehensive, integrated evaluation to the water quality within the scope of Shitoukoumen reservoir water catchment's basin, and gives out recommendations for improvement.