毛竹林竹鞭生长特征和空间结构的探地雷达无损探测  

Non-Destructive Detection by Ground Penetrating Radar of Growth Characteristics and Spatial Structure of Rhizomes in Moso Bamboo Forest

作　　者：熊雨露 周宇峰[1] 李平衡[1] 童亮 周国模[1] 施拥军[1] 杜华强[1] Xiong Yulu;Zhou Yufeng;Li Pingheng;Tong Liang;Zhou Guomo;Shi Yongjun;Du Huaqiang(State Key Laboratory of Subtropical Silviculture,Key Laboratory of Forest Ecosystem Carbon Cycle,Sequestration and Emission Reduction in Zhejiang Province,School of Environmental and Resource Sciences,Zhejiang A&F University,Hangzhou 311300)

机构地区：[1]浙江农林大学省部共建亚热带森林培育国家重点实验室,浙江省森林生态系统碳循环与固碳减排重点实验室,浙江农林大学环境与资源学院,杭州311300

出　　处：《林业科学》2020年第12期19-27,共9页Scientia Silvae Sinicae

基　　金：国家自然科学基金项目(41671411)。

摘　　要：【目的】发展毛竹林地下竹鞭空间结构的探地雷达无损探测方法,解决因毛竹竹鞭深入地下和观测不便而使空间结构及地下生物量无法准确预测的难题,为毛竹林地下竹鞭研究提供新思路。【方法】设置毛竹林样地并采用探地雷达进行探测,在雷达数据预处理后提取竹鞭信息,同时基于双曲线模型估计竹鞭直径,并利用直径(D)与长度(L)拟合竹鞭生物量,再用实测值进行模型检验,最后分析竹鞭的垂直与水平空间结构。【结果】探地雷达能有效探测毛竹林地下竹鞭的位置信息,双曲线模型对竹鞭直径和长度估计误差分别为-14.45%~20.66%和0.53%~8.51%,对空间位置3个方向(X,Y,Z)的估计误差分别为0.13%~6.65%、1.23%~6.55%和2.42%~7.41%;对于生物量,以DL为变量的多项式模型和以D2L为变量的指数模型拟合结果较好,确定系数R2为0.93~0.95,均方根误差RMSE为18.3~22.4 kg·hm^-2;从模型检验结果来看,以D2L为变量的指数和多项式模型模拟效果显著好于以D、D2和DL为变量的指数和多项式模型;以D2L为变量的指数模型是竹鞭生物量(BR)的最佳拟合模型,具体形式为BR=65.17e0.002D2L;对竹鞭空间结构分析表明,竹鞭主要垂直分布在0~40 cm土层,该土层竹鞭条数占竹鞭条数总量的91%,竹鞭鞭长和生物量分别占总量的95%和93%,且竹鞭直径在下层(>20 cm)土壤显著大于表层(0~20 cm)土壤;竹鞭在地下的走势较为曲折,空间分布相对均匀,同一竹鞭的直径差异不大,但不同竹鞭的直径差异显著(P<0.01),样地内单位面积竹鞭的长度和生物量分别为54080 m·hm^-2和1001.17 kg·hm^-2。【结论】利用探地雷达能成功探测地下竹鞭的位置信息,可得到竹鞭垂直和水平方向的空间结构,同时竹鞭直径模型与生物量模型的预测结果也较好,表明探地雷达技术可为研究毛竹林竹鞭结构及其动态与地上立竹的相互关系等提供技术支撑。【Objective】A method of non-destructive detection for the spatial structure of underground bamboo rhizomes in moso bamboo forest by using ground penetrating radar(GPR)was developed,in order to solve the problem that the spatial structure and underground biomass could not be accurately predicted due to the deep underground and inconvenient observation of moso bamboo rhizomes,providing a new idea for the study of underground bamboo rhizomes in moso bamboo forest.【Method】The GPR was used to detect the bamboo rhizomes in the sample plots of moso bamboo forest.The information of bamboo rhizomes was extracted after radar data pretreatment.At the same time,the diameters of bamboo rhizomes were estimated based on hyperbolic model,and the biomass of bamboo rhizomes was fitted by diameter(D)and length(L).Then the measured values were used to test the model.Finally,the vertical and horizontal spatial structure of bamboo rhizomes was analyzed.【Result】The GPR could effectively detect the position of the underground bamboo rhizomes.The error range of the hyperbolic model for the diameter and length of the bamboo rhizomes were-14.45%-20.66%and 0.53%-8.51%.The estimated error range of the space position in X,Y and Z directions were 0.13%-6.65%,1.23%-6.55%,and 2.42%-7.41%,respectively.The estimated values were close to the measured values.As for biomass,polynomial model and exponential model using DL and D2L as variables had the best fitting results.The coefficient of determination(R2)ranged from 0.93 to 0.95,and the root mean square error(RMSE)was from 18.3 to 22.4 kg·hm^-2.The simulated results of the exponential and polynomial models with D2L as variables were significantly better than that of the exponential and polynomial models with D,D2 and DL as variables.Generally,the exponential model with D2L as the variable was the best fitting model of bamboo rhizomes biomass(BR),while the specific expression was BR=65.17e0.002D2L.Analysis of bamboo rhizomes spatial structure showed that the vertical distribution of bambo

关 键 词：毛竹林 探地雷达 竹鞭系统 空间结构 无损探测 

分 类 号：S757.29[农业科学—森林经理学]