应用遥感技术监测丹江口水库氨氮分布研究

Application of Remote Sensing Technology to Monitor NH3N Distribution in the Danjiangkou Reservoir

摘要: 丹江口水库作为南水北调中线工程的水源地，其水质状况直接影响到受水区经济发展和居民用水安全。本研究针对丹江口水库现状水质监测存在的耗时耗力、成本高、监测点有限等现象，基于遥感技术和智能算法的基本理论，利用高分辨率遥感数据和实测水质数据研究典型水质参数氨氮的遥感反演技术。将构建的人工神经网络模型应用于丹江口水库，获得2007~2009年丹江口水库氨氮浓度时空分布图。结果表明，人工神经网络模型能有效地反演丹江口水库氨氮的时空分布情况，率定期和验证期的平均相对误差分别为3.59%和19.28%，相较于多元回归模型的8.3%和22.4%而言，人工神经网络模型反演效果更为精准，更能模拟实际情况；研究时段内，丹江口水库水质总体上趋好，氨氮浓度均低于地表水环境质量II类标准限值，丰水期明显高于枯水期。水质遥感反演作为一种非传统性的监测手段，具有空间全覆盖、快速和成本低等传统水质监测难以比拟的优点，能够作为实测数据的有益补充，极具有深入研究和推广的价值和必要。

Abstract: Aiming at the disadvantages of high-cost and limited sites of water quality monitoring, this paper develops artificial neural network (ANN) models to estimate NH3N concentrations from high spatial resolution remote sensing imagery and in situ water samples collected concurrently with overpassing satellite. The artificial neural network models constructed were applied to Danjiangkou reservoir to get the spatial and temporal distribution of ammonia nitrogen concentration from 2007 to 2009. It turned out that, the average relative errors at calibration and validation of ANN are respectively 3.59% and 19.28% for NH3N, which is better than multiple regression model’s 8.3% and 22.4%. It shows the models are effective for monitoring water quality state and spatial distribution of reservoir with the value of in-depth research and extension.