中小尺度地基GPS精细化反演土壤湿度研究

Measurements of soil moisture at various temporal and spatial scales are important for studies of global climate change, weather forecasting, as well as flood or drought prediction and management. While satellite missions have been planned to measure soil moisture at global scales, these missions also need ground-based soil moisture data with high temporal-spatial resolution to validate their observations and retrieval algorithms. Therefore, the main objective of this project is to develop refinement methodology and strategies for soil moisture retrieval using ground-based GPS for meso-micro scale applications. Firstly, the influence, caused by various environment factors, may be different in quantities for different observations or their linear combinations of GPS. However, they might have significant correlation and similar spectrum characteristics. We plan to design a special filter to separate soil moisture component from signal-to-noise data with high accuracy and reliability. Secondly, we will develop a methodology based on spectrum analysis to obtain the two dimensional vegetation distribution map using the GPS reflected signal and then a spatiotemporal synchronization approach will be proposed to remove or reduce the influence about vegetation growth. It will provide a novel way to monitor soil moisture with ordinary geodetic GPS instruments or existing continuously-operating GPS networks. Finally, we plan to employ Geostatistics to analyze retrieval data and then reveal the characterization of soil moisture spatial-temporal variability at small and medium scale region. These results can contribute to time synchronization and geometrical registration of GPS retrieval data and satellite remote sensing image or data. And the scale of GPS soil moisture content measurement will be extremely useful for calibrating and validating soil moisture satellite missions.

不同时空分辨率的土壤湿度观测资料对研究全球气候变化，提高短时天气预报的准确性以及开展洪涝、干旱等灾害的监测与治理都具有重要意义。为弥补卫星遥感土壤湿度时空分辨率的不足以及验证遥感数据与反演算法的有效性，本项目首先从影响GPS信号波动的多种因子在不同类型观测值中的相关特性与频谱特征入手，基于分形理论与小波互相关分析法，研究从SNR信号中精细提取土壤湿度成分的算法；其次利用GPS反射信号中植被生长成分，探索基于谱分析法绘制2维植被分布图的方法，并在时空同步机制下研究去除或减弱植被覆盖对反演精度影响的理论与算法，使常规测量型GPS在单站单天线模式下反演土壤湿度成为现实；最后利用地统计学方法分析反演资料，以揭示中小尺度区域土壤湿度的时空变异规律，为地基GPS反演数据与卫星遥感影像或数据的时间同步与几何配准奠定基础，从而实现对大尺度监测模式的数据及反演算法进行验证与评价。

研究常规测量型地基GPS反演土壤湿度的理论与方法，可充分利用丰富的导航卫星及海量的地面测站资源，实现建立中小尺度地基GPS土壤湿度观测网络。该研究对开展全球气候研究，提高短时天气预报精度以及预防与治理洪涝、干旱、沙尘暴等自然灾害都具有重要的意义。主要研究内容包括4个方面：地基GPS反演土壤湿度时空特性研究、影响观测值波动因素研究、地基观测站优化方法及策略研究、GPS地面反射信号精细化提取方法研究。主要取得的研究成果如下：（1）研究了卫星、反射点与测站间的几何关系，通过在测站与反射点切换观测视角并建立相应坐标系统，实现了卫星地面反射点及其轨迹的精密获取。在此基础上，进一步引入数字地面模型数据，并基于坐标系统转换原理，将4维反射点轨迹由平坦地面扩展到自然起伏地表。同时给出了地基GPS反演土壤湿度的重访周期。（2）研究了GPS信号的极化特性、菲涅尔区分布及GPS天线特性，解释了SNR值受信号入射角度及反射点分布影响而波动的原因，给出了适于GPS反演土壤湿度卫星高度角的选取范围。（3）提出了GPS反演土壤湿度测站优化指标及优化策略，从最佳观测时段、最佳观测卫星及最佳观测位置等方面给出了具体的实施方法与步骤。（4）研究了多路径信号的分形特征，提出了一种基于分形理论及小波变换的地面反射信号提取方法，为丰富GPS反演土壤湿度的信号源提供了新思路。