【作者】 李森；

【导师】 周清波；

【作者基本信息】 中国农业科学院 ， 作物信息科学， 2007， 博士

【摘要】 作为陆面生态系统水循环的重要组成，土壤水分是植物生长发育的基本条件，也是研究植物水分胁迫、进行旱情监测、农作物估产等的一个重要指标。因此，大面积监测土壤水分在水文、气象和农业科学领域具有较大的意义，土壤水分监测一直是人们十分关注的问题。应用遥感技术进行土壤水分等地表参数的反演在过去的几十年间取得了显著进展，微波遥感尤其是主动微波遥感具有全天时、全天候观测的优点，并且对地表具有一定的穿透能力，能够弥补其他遥感方式在土壤水分监测应用中的不足，为流域尺度土壤水分监测提供了新的方法和途径。目前的星载SAR系统在向多极化、多角度方向发展，因此采用不同极化方式及不同入射角条件下的后向散射系数组合类研究雷达后向散射系数与地表参数之间的关系成为可能。本研究在比较三种典型的地表微波散射模型(Oh模型、Dubois模型、IEM模型)的基础上，用理论模型IEM模型对地表的微波散射特征进行了模拟和分析，提出了用多极化SAR后向散射系数数据来反演地表粗糙度参数和土壤水分的经验模型。本研究的主要研究成果包括以下几方面：(1)对Dubois模型、Oh模型和IEM模型的比较结果表明，Dubois模型对介电常数、入射角变化的响应与实际情况呈截然不同，因此在用雷达数据进行土壤水分反演研究时不推荐选择Dubois模型。Oh模型能够模拟交叉极化的后向散射特征，而IEM模型只能模拟同极化的后向散射数据，因此如果需要模拟交叉极化微波后向散射特征时，可以选择Oh模型。理论模型IEM模型对同极化后向散射特征的模拟能够比Oh模型更好地刻画地表微波散射特征，因此在进行同极化模拟时应选择IEM模型。(2)分析了均方根高度、相关长度、土壤含水量等地表参数以及入射角、极化方式等系统参数对雷达微波后向散射特征的影响，揭示了雷达后向散射系数随这些参数的变化而变化的规律，对雷达反演地表土壤水分的理想系统参数设置进行了探讨。(3)提出了利用同极化(VV或HH)后向散射系数在小入射角和大入射角情况下的差来反演地表组合粗糙度参数Z_s的经验模型；基于该模型，建立了利用雷达后向散射系数反演地表土壤水分的经验模型。(4)提出了利用VV和HH极化相同角度的后向散射系数差来反演地表均方根高度的经验模型，该模型可以在较大的地表粗糙度范围内实现地表均方根高度的反演，从而使建立雷达后向散射系数与地表粗糙度之间像元尺度的对应关系成为可能。(5)首次对S波段地表微波散射特征进行了模拟与分析，并针对该波段提出了相应的地表粗糙度参数及土壤含水量反演经验模型，从而为将来S波段星载SAR系统发射升空后在环境监测及防旱减灾方面的应用提供技术支持。(6)利用获取的ENVISAR-ASAR数据，对建立的经验模型在河北平原南部进行了应用与验证，结果表明本研究建立的地表粗糙度参数反演模型和土壤水分反演模型均能取得较好的应用效果。更多还原

【Abstract】 Soil moisture content plays a critical role in the surface energy balance at the soil-atmosphere interface and can be considered as a key state variable that influences the distribution of the radiant energy, the runoff generation and the percolation of water into the soil. Soil moisture monitor is always a focus in hydrology, meteorology and agriculture researches. In the past several decades, large scale soil moisture monitor by remote sensing, such as optical, infrared, and microwave methods have developed rapidly. Common methods and optical remote sensing have their own shortcoming in soil water monitor, thus cannot reach the need of soil water monitor. Active microwave techniques, particularly Synthetic Aperture Radar (SAR), take advantage of their all-weather and night-and-day measurement capacities, so it offers a new way to monitor soil moisture in drainage scale.Multi-polarization and multi-incidence angle is the tendency of space-borne SAR, so it is possible to reveal the relationship between radar backscatter coefficient and surface parameters by using different backscatter coefficient under different polarization mode and incidence angle. First three classical surface microwave backscatter model, i.e. Oh model, Dubois model and IEM model were compared and the IEM model was chosen in the following research. Empirical inversion model of surface roughness and soil moisture content were proposed on the simulation and analysis of surface backscatter characteristic under different input parameters (different bands and different polarizations) by using the IEM model.The results and innovations of this dissertation are mainly in the following aspects:(1) The comparison of Dubois model, Oh model and IEM model indicates that the Dubois model cannot describe the real backscatter character of surface under different dielectric constant or incidence angle, so it was not preferred in soil moisture inversion. Oh model can be used in cross polarization backscatter simulation. IEM model is preferred in co-polarization simulation because it can describe the backscatter character perfect.(2) On the basis of analysis of backscatter character under different surface roughness, soil water content, incidence angle and polarization conditions, the response of radar backscatter coefficient to these parameters was revealed.(3) An empirical model to inverse mixed surface roughness Z_s was proposed by using the backscatter coefficient difference between little and large incidence angles; then the empirical soil moisture inversion model was proposed based on the mixed surface roughness inversion model.(4) An empirical model to inverse RMS height was proposed by using the difference between VV and HH polarization backscatter coefficients under the same incidence angle. This model can inverse RMS height in a large roughness ranges. The use of the new empirical model can avoid tedious ground measurements to a great extent, and makes it possible to represent soil surface in pixel scale.(5) The backscatter characteristic at S band was simulated and analyzed for the first time, and the empirical modes for surface roughness and soil moisture were proposed. This research work provides technical support for the applications of S band space borne SAR of our country in the near future. (6) The surface roughness and soil moisture inversion models were applied in south Hebei Plain by using the ENVISAT-ASAR data and in situ experimental data. The inversion results of both parameters fitted the in situ data well.更多还原