基于成像光谱技术的滨海土壤关键属性演变过程定量研究

The soil’s properties and functions and genesis in time and space under the conditions of global change are fundamental challenges of today’s soil research. The characteristics of spatial and temporal evolution in coastal soil are the prerequisite and foundation of coastal resource exploitation and utilization. In current study of depth change of soil property in profile, the deficiencies include large sample interval and uncertainty of fitting results, which can’t display the continous depth curve of soil property. .In the most decades, the potential of soil spectroscopy as a new rapid and cost-efficient technique used to predict soil properties has been proved. The soil samples of soil reflectance spectroscopy used for soil profiles are the mixed samples of specfic horizen. These samples can not reflect the changes in the horizen. Imaging spectroscopy can collect spectroscopy for every pixels which can be used to mapping soil properties for the whole profile and deriving continuous depth functions. Therefore, it should be possible to apply imaging spectroscopy technique in chronosequence analysis..The objectives of this project are to: (i) determine the feasibility of using image spectroscopy for soil chronosequence mapping; (ii) illuminate the evolution processes in the millennial chronosequences developed from typical parent materials in the north of Jiangsu under the effect of soil-forming factors and human activity using imaging spectroscopy. In this study, we collect the soil profiles developed from typical parent materials in different historical periods and construct the soil chronosequences after determining the parent material uniformity and soil age. Then we analyze the evolution characteristics of the chemical, physical, and mineralogical soil properties and select the spectral characteristics or index indicating the degree of soil evolution. The multiple regression, partial least squares regression and support vector machine methods will be used to relate the spectral characteristics or index to predict the soil properties. The models are used to map these soil profiles and then the continue depth functions of properties are extracted. The characteristics of spatial and temporal evolution are analyzed according to the continue depth functions. This progect is in a cross area of soil science and remote sensing. It is a supplement of our soil spectral library and has theoretical significance of revealing the soil evolution processes and clarifying the mechanism of soil chronosequence analysis using imaging spectroscopy. It also has practical significance of broadening the utility of imaging spectroscopy technology to pedogesis.

定量研究土壤发生学、土壤属性和功能的时空变化是全球变化背景下土壤科学研究面临的挑战。目前，土壤剖面属性的深度变化研究存在采样间隔大、属性模拟结果不确定等不足，无法反映成土属性连续深度分布特征；成像光谱技术实现了点状反射光谱向面状图像光谱的转变，为获取剖面属性连续深度分布提供了重要技术支持。本研究选择苏北滨海地区千年以来长江与黄河沉积物母质、不同成土年龄土壤剖面构建土壤时间序列，建立该地区典型母质条件下土壤剖面成像光谱数据库，分别进行关键成土属性剖面制图、属性连续深度分布曲线提取，从而分析千年以来关键物理、化学、矿物属性的时间演变规律，定量确定其时间响应函数，阐明在自然成土因素和人为利用影响下典型母质滨海土壤发生过程和演化规律。本研究是土壤科学与遥感科学的交叉性综合研究，对于揭示滨海土壤属性演变规律、阐明土壤剖面成像光谱遥感分析机理、实现遥感制图具有重要实践意义，为土壤发生学研究提供新技术。

土壤发生学研究的核心是土壤剖面，它记录了土壤发生过程与历史，但在过去几十年中研究土壤剖面的技术和工具并没有发生质的变化，限制了土壤发生学研究与应用的快速发展。因此，研究快速测定、分析土壤剖面的新技术，实现土壤演化过程的定量分析，对于全球变化和加速人为活动背景下土壤的动态变化过程研究具有重要现实意义。本研究结合历史海岸线变化，分别在苏北地区（南通通州、射阳-滨海、东台）和浙江慈溪四个地区采集1米深完整土壤剖面，构建土壤时间序列，对比分析研究区土壤主要属性演化过程、演化规律；利用成像光谱技术进行剖面属性（有机碳及氮、磷、钾等）制图，提取属性连续深度变化曲线，分析其垂向变化特征及时间演变规律，为土壤剖面研究提供新的技术支持。研究结果表明，成像光谱技术可以较好进行土壤有机碳、氮、磷、钾及土壤肥力等属性制图。由于土壤有机碳的积累过程，有机碳垂向分布表现出随深度增加而减少的规律，随成土年龄增大而增加（通州除外）；由于碳酸钙的淋溶过程，其含量随深度增加而增大，随成土年龄增大而减少。但受样点数量限制，不同地区有机碳及碳酸钙的时间演变规律表现有线性或非线性。如东台、慈溪两个地区的有机碳及碳酸钙为非线性变化过程，但射阳-滨海地区的有机碳及碳酸钙为线性变化过程。总体来讲，土壤形成过程是一个固碳过程，滨海地区土壤目前还具有非常大的固碳潜力。本研究发现一个重要现象，土壤有机碳含量与土壤有机碳密度之间存在非常好相关关系，可以用于直接通过有机碳含量预测有机碳密度，这对于缺少容重的土壤数据而言非常重要。本研究建立的土壤时间序列剖面库、成像光谱数据库及对应数据属性库，将为该地区其它属性演变研究提供很好的数据基础。