草地植被根系对土壤水文特性的影响机理及其光谱学特征研究

This proposal aims to further the hyperspectral remote sensing application research in soil erodibility estimation in a large scale based on the research achievement of soil properties research supported by the previous fund. In this project, the Hulunbeier grassland will be used as the typical research area, the applicants want to amplify the application of hyperspectral remote sensing in detecting soil erodibility over plant covered area. To realize this objective, a four-step analysis procedure is designed. Firstly, by selecting several typical plants, the mechanism of plant root effects on the soil erodibility will be studied, and important parameters of both soil properties and plant root will be identified. Through building the correlation equations between the plant root and soil erodibility, the influence of root system on soil erodibility over plant covered area can be quantified, which improves the understanding of soil physical and chemical properties and their relationship with plant cover. Secondly, the coordinative relationships between the plant canopy and root in the changing environments are analyzed through field quadrat survey and indoor plant cultivation experiments. This coordinative relationship will then be summarized using multiply regression methods. Thirdly, through analyzing the relationships between the above ground and blow ground biomass, the important canopy parameters can be found out that alter soil erodibility values indirectly. New hyperspectral experiments will be designed both indoor and over field standard runoff plot to find out ways to calculate these parameters using remote sensing data. Hyperspectral reflectance characteristics are recorded to compare the soil erodibility values over plant covered area and non-plant covered area. Based on the previous three steps, the influence of plant root on controlling soil erosion can be analyzed using hyperspectral remote sensing techniques. And a quantitative analysis method will be found out to facilitate the determination of soil parameter over the plant covered area in a large scale region.

土壤水文特性是控制土壤水文过程的重要因素，其中可蚀性是反映土壤受水蚀过程影响的控制性指标。本项目以呼伦贝尔草原为研究区，从完善草地植被根系对改善土壤理化性质的试验研究入手，通过野外调查采样与室内外模拟试验等手段，探索几种典型草地根系对土壤可蚀性影响机理，筛选不同类型草地区对土壤可蚀性有显著影响关系的土壤及植被参数，并建立植被根系生物量与土壤可蚀性的定量关系。在系统研究典型草地植被地上-地下生物量异速生长关系的基础上，识别变化环境下草地植被地上-地下生物量分配比（R/S），从而实现植被地上生物量与地下生物量指标的转化互算。通过测试植被地上生物量的反射光谱特征，组合其特征光谱波段构建植被地上生物量的高光谱反演方法。基于此，结合无人机高光谱影像数据，凝练研究区内草地植被覆盖下的土壤可蚀性高光谱定量反演方法，为大尺度土壤水文参数预测研究提供理论基础和技术手段。

高光谱遥感反演植被根系对土壤属性的影响是水土保持学科领域的重点也是难点，更是优化土壤属性模型的精准度和适配度的前沿方法，由于土壤属性的变化会导致流域出现严重的水土流失，严重制约了区域的经济发展。本项目：利用野外调研实验结果研究的地下与地上生物量的实测数据，构建地上生物量与植被之间的关系。①结果表明，地下地上生物量之间关系存在Logistic关系。②植被地上生物量与LAI与土壤可蚀性存在动态相关关系。③通过室内光谱试验和室外光谱验证以及人工降雨侵蚀试验，利用主成分分析法提取了几个对土壤侵蚀影响较大的土壤理化参数，建立了单要素土壤参数的高光谱特征光谱库。④通过单相关分析，建立了单特征光谱的土壤单要素一元反演模型；通过偏最小二乘回归法（PLSR），建立了多特征光谱的土壤单要素多元反演模型，结果发现，利用PLSR分析建立的反演模型效果较好。间接性的用高光谱信息解释了土壤可蚀性，半定量的建立了土壤可蚀性的高光谱反演模型。.