基于GPR反射波的滨海盐渍土区土壤含水量测定研究

Soil water content is the main factor controlling soil salinization in the coastal saline soil region. As one of the non-destructive detection techniques, ground penetrating radar (GPR) can interpret soil profile structure along the measuring line and inverse soil water content of different layers. For investigating the spatial distribution and variation of soil water content by GPR reflection wave signals, the working mechanism and the precision of the results are worthy of further study. Through single layer and multilayers model experiments with fixed offset (FO) method, the research would reveal the identification ability of GPR for plots with different soil body configurations and soil layers with different dielectric properties, clarify the response characteristics of electromagnetic wave velocity, amplitude values and frequency spectrum to soil water, salt and texture, construct inversion models of soil dielectric constant/soil water content under the different conditions of soil salt content and texture. Through field experiments with FO and common middle point (CMP) methods, the research would verify the inversion models of soil dielectric constant/soil water content, modify model parameters and improve the inversion accuracy of soil water content revision by comparison with the results of oven drying methods. Finally, the research would form the technology integration system combining GPR and 3S to simulate the spatial distribution of soil water..This research aims to reveal the quantitative relationship between GPR signals and soil water content in the coastal saline region and clarify the technical ideas and parameters design of GPR investigation, which would provide a new technical support for real-time monitoring the vertical migration of soil water and improving soil salinization.

土壤水分是滨海盐渍土调控盐碱化的主要因素，无损探测技术GPR可沿测线解析土壤剖面，反演分层土壤含水量，以GPR反射波信号调查土壤水分的空间分布与变异，其工作机理与结果精确度值得深入研究。本项目通过单层次和复合层次地块剖面法模型实验，揭示GPR识别不同土体构型地块以及剖面土壤介电特性差异分层的能力，阐明电磁波传播速度、振幅能值和频谱信号对土壤水分、盐分、质地变化的响应特征，构建不同土壤盐分、质地条件下的土壤介电常数/含水量反演模型；通过剖面法-共中心点法实地实验，验证土壤介电常数/含水量的反演模型，修正模型参数，以烘干法结果为参照提高GPR反演土壤含水量的精确度，最终形成GPR与3S技术相结合模拟土壤水分空间分布的技术集成体系。本项目旨在揭示滨海盐渍土区GPR信号与土壤含水量的定量关系，阐明GPR测定土壤含水量的技术思路和参数设计，为准确实时监测土壤水盐运移、改良土壤盐渍化提供新的技术支撑。

利用GPR（ground penetrating radar，探地雷达）在黄河三角洲滨海盐渍土区实现土壤水分定量监测能够为实时监测土壤水盐运移、改良土壤盐渍化提供新的技术支撑。本研究采用250MHz中心频率GPR，结合FO（fixed offset，固定间距）法、CMP（common middle point，共中心点）法和格网测量，进行典型土壤剖面层位识别与信号提取、土壤水盐探测与模拟、土壤水分纵向迁移路径模拟与剖面三维建模。发现：（1）GPR可以识别130年来不同土地利用方式下典型土壤剖面的土壤层次，耕作层的厚度与分界面清晰可辨，将探地雷达波谱图像与振幅—时间信息进行对比，有助于减小土层厚度的估计误差，0-1m厚度内土层检测误差可达0.1m，光谱图像的形态特征越相似，土壤性质越接近；（2）GPR反演体积含水量与农田表层土壤实测质量含水量的相关系数可达0.8以上，其中Topp模型与巨兆强模型的相关系数最高，朱安宁粉砂黏潮土的θ~Ɛ经验模型的平均反演误差为0.012m3m-3；（3）土壤理化特性综合影响探地雷达的电磁波信号，包络振幅能量的二阶导数值分别与土壤含水量和电导率（EC）值呈负对数函数和幂函数关系；（4）以Matlab为平台进行网络学习和仿真构建的土壤含盐量模型输出准确率达到86.53%，土壤盐分越高，GPR信号反演土壤含水量误差越大；（5）GPR反射波振幅3D模型可反映土壤水分的纵向迁移特征，芦苇、玉米—茅草地块的土壤水分均匀下渗，沙蒿、玉米地块出现水分显著积聚的区域。