基于探地雷达的典型喀斯特坡地土层厚度估测

土层厚度是坡地水文过程的一个重要控制因素。本研究使用MALA公司的探地雷达，分别采用100和500MHZ频率的天线对土层厚度进行探测，然后通过开挖探槽实测土层厚度，同时调查了影响探地雷达结果的因素—土石交界面的基岩风化度，并建立了不同输入下的线性回归模型和GEP模型，对探地雷达在喀斯特坡地土层厚度估测的适用性进行了探讨。结果表明，喀斯特坡面土层浅薄且含有较多碎石，使用频率较高的天线可以提高探测精度；同时将不同频率天线探测的结果以及基岩风化度作为输入，均能提高对土层厚度的模拟精度；相同输入条件下，总体来看，基于GEP算法建立的土层厚度模型模拟精度较经典统计的线性回归和多元回归模型高，R2最大能提高7.6%

Estimation of Thickness of Soil Layer on Typical Karst Hillslopes Using a Ground Penetrating Radar

Soil thickness acts as an important control for hillslope hydrologic processes. This study use the ground penetrating radar produced by the company of MALA to detect the soil thickness in hillslope of karst region, the frequency were 100 and 500 MHZ, respectively. At the same time, the degree of weathering of bedrock was explored. Then the classical linear regression models and GEP models were established under different input variables to verify the feasibility of ground penetrating radar in soil thickness detected. The results showed that the use of high frequency of the antenna can improve detection precision, and the use of different frequency of the antenna at the same time and introduce the degree of weathering of bedrock as input can improve the simulation accuracy. Overall, the GEP models were consistently more effective than linear regression models, the coefficient of determination can reach 0.742.