耦合两阶段梯形模式和再分析资料的微波遥感土壤湿度降尺度研究

It is significant and difficult to enhance the spatial resolution of microwave remotely sensed soil moisture. Disaggregation method by combining optical/thermal and microwave remote sensing data is a great way to attain that goal. However, this disaggregation method are facing some limitations such as depending on traditional triangle and trapezoid space and not working for non-clear sky conditions. Motivated by these limitations, we dedicate to study disaggregating microwave remotely sensed soil moisture by coupling a two-stage trapezoid space model and atmospheric reanalysis. The aims of this study are demonstrating effect mechanism between soil moisture and soil evaporative efficiency, constructing coupling mode between two-stage trapezoid and atmospheric reanalysis, and developing an all-weather disaggregation method by combining optical/thermal and microwave data. Specific contents of this study area: 1) Analyzing the effect mechanism between soil moisture and soil evaporative efficiency and building an improved connection between them taking into account the atmosphere condition, soil properties, and topography. 2) Constructing a coupling mode between two-stage trapezoid and atmospheric analysis that has clear sky and non-clear sky conditions and aims to estimate all-weather soil evaporative efficiency or soil moisture index. 3) Developing an all-weather method for disaggregating microwave remotely sensed soil moisture based on Taylor series expansion, improved UCLA, and machine learning methods. This study benefits to provide multiscale soil moisture data for many applications such as agricultural production, drought monitoring, and water resource management, etc.

提升微波遥感土壤湿度的空间分辨率具有重要意义也面临诸多困难，其中融合光学/热红外和微波数据的降尺度方法是实现这一目标的重要途径。然而，此类方法面临依赖传统的三角形和梯形模式、易受云雨天气影响等诸多限制。为此，本项目开展耦合两阶段梯形模式和再分析资料的微波遥感土壤湿度降尺度研究，旨在阐明土壤湿度对土壤蒸发效率的驱动机制，建立两阶段梯形模式和再分析资料的耦合方式，构建融合光学/热红外与微波数据的土壤湿度全天候降尺度算法。研究内容包括：1）梳理土壤蒸发效率与土壤湿度的影响机理，构建顾及大气状态、土壤属性、或地形地貌的土壤湿度与土壤蒸发效率/湿度指数关联模型；2）针对晴空与非晴空两种状态，以估算全天候的土壤蒸发效率/湿度指数为目的，建立两阶段梯形和再分析资料的耦合方式；3）基于泰勒级数展开、改进的UCLA、和机器学习算法等，设计微波遥感土壤湿度的全天候降尺度算法。

提升微波遥感土壤湿度的空间分辨率具有重要意义也面临诸多困难，其中融合光学/热红外和微波数据的降尺度方法是实现这一目标的重要途径。然而，此类方法面临依赖传统的三角形和梯形模式、易受云雨天气影响等诸多限制。为此，本项目开展耦合两阶段梯形模式和再分析资料的微波遥感土壤湿度降尺度研究，旨在建立两阶段梯形模式和再分析资料的耦合方式，构建融合光学/热红外与微波数据的土壤湿度降尺度算法。本项目计划执行期为 2019-2022 年，预期成果为发表学术论文 6 篇，其中 SCI 收录论文不少于 3 篇；培养博士生 1~2 名，硕士生 2~4 名。执行结果表明，本项目建立了DSCALE_mod16 微波土壤湿度降尺度模型、提出了基于 LST/FVC 空间的土壤蒸发效率计算方法、构建了耦合机器学习与物理模型的降尺度策略、提出基于新型 Priestly-Taylor 公式确定LST/FVC 空间干湿边界的方法、构建中国区域ESA CCI 土壤湿度数据空隙的填补策略、提出微波和气象数据融合的方法，解决云污染条件下无有效光学和热红外遥感观测的问题等。在本项目资助下，合计发表 SCI论文13 篇（其中一篇网络首发）、EI论文 3 篇（其中2篇网络首发）、中文核心5 篇、已授权发明专利 2 项、出版专著 1 部、从 2019 年到 2022 年共培养硕士毕业生15 名，培养在读博士生 4 名。本项目预期目标已全部完成。