基于遥感反演的缺资料流域非点源污染模型研究

With the development of China's environmental protection work, point source pollution has been gradually brought under control, non-point source pollution become the main source of water pollution. China's current hydrology and water quality monitoring network is still not perfect, hydrology and water quality data in non-key areas is in a lack, monitoring data of the non-point source is scarce, then large-scale environmental treatment is restricted. Therefore, how to estimate the non-point source pollution caused by the rainfall-runoff in these areas, and establish a non-point source model has become a pressing problem. This project intends to use the correlation of suspended solids, dissolved organic matter and total phosphorus, total nitrogen in rainfall runoff, through remote sensing retrieval to calculate suspended solids, dissolved organic matter, and calculate the total phosphorus, total nitrogen indirectly, estimate the runoff pollution in no-data areas. Subsequently, with further study on rainfall runoff and pollutant enrichment, erosion, migration, transfer of reaction mechanisms, on the basis of high-resolution remote sensing inversion to calculate basin evapotranspiration, soil moisture, vegetation cover and soil erosion, as the model input parameters, this project develop a based on remote sensing retrieval of the lack of information on watershed runoff pollution model, and use the remote sensing retrieval, directly or indirectly calculate the suspended solids, total phosphorus and total nitrogen then calibrate and validate the model. This make a contribution to the in-depth treatment of China's water environment.

随着我国环境保护工作的推进，点源污染已逐步得到控制，非点源污染逐渐成为水体污染的主要来源。由于我国目前的水文水质监测站网还不健全，非重点区域的水文水质数据较少，非点源污染实测数据更是稀缺，制约了大规模的环境治理。因此，如何估算这些区域由降雨径流引起的非点源污染，并据此建立模型，成为亟待解决的难题。本项目拟通过遥感直接反演降雨径流中的固态悬浮物、可溶性有机物，并根据这两项指标与总磷、总氮的相关关系和回归算法，计算总磷、总氮。随后，在深入研究流域降雨径流和污染物富集、冲刷、迁移转移转化机理的基础上，建立非点源污染模型，并使用高分辨遥感反演流域蒸散发、土壤含水量、植被覆盖和土壤侵蚀状况，作为模型的输入参数，最终构建较少依赖地面实测资料的，基于遥感反演的缺资料流域非点源污染模型，并运用遥感反演的固态悬浮物、总磷和总氮等指标对该模型进行率定和验证，为我国的水环境深入治理作出贡献。

我国目前的水文水质监测站网还不健全，非重点区域的水文水质数据较少，非点源污染实测数据更是稀缺，制约了大规模的环境治理。因此，如何估算这些区域由降雨径流引起的非点源污染，并据此建立模型，成为亟待解决的难题。本项目利用水体中固态悬浮物、可溶性有机物与总磷、氨氮的相关关系，基于多元回归算法、人工神经网络和支持向量机等算法，遥感反演水体中的总磷、总氮。随后，在深入研究流域降雨径流和污染物富集、冲刷、迁移转移转化机理的基础上，建立非点源污染模型，并使用高分辨遥感反演流域蒸散发、土壤含水量、植被覆盖，作为模型的输入参数，最终构建较少依赖地面实测资料的，基于遥感反演的缺资料流域非点源污染模型。同时，也运用遥感反演的固态悬浮物、总磷和总氮等指标对该模型进行了率定和验证。基于数据条件，在沙颍河、温瑞塘河和磁湖进行了水质反演，在沙颍河建立了非点源污染模型，均取得了满意的成果。