基于区域剖分算法的土壤碳通量空间采样策略

由于土壤碳通量在空间分布上具有很强的异质性,传统的采样方法难以对区域土壤碳通量进行精确估算,因此确定适当的采样策略对区域土壤碳通量的估算具有重要意义.本文提出一种逐点递增式采样的区域剖分部署策略(RDPG):设定初始采样点,使用改进的凸包插值算法构造Delaunay三角网,根据邻近已知采样点插值计算三角形各边垂直平分线的交点的离散度,选择离散度最大的点作为新增采样点.采用该方法对变异系数为0.42～0.59的仿真试验区域进行多次试验,结果表明:在相同试验条件下,RDPG布局策略能够获得比随机采样和均匀采样策略更高的区域土壤碳通量估算准确度.RDPG方法考虑了区域土壤碳通量的空间异质性,提高了区域土壤碳通量拟合精度.

A spatial sampling strategy for soil carbon flux estimation based on the region division

Regional soil carbon flux can not be accurately estimated using traditional sampling methods due to the strong spatial heterogeneity of its spatial distribution, so it is significant to determine appropriate sampling strategies for the estimation of regional soil carbon flux. In this paper, a point-by-point incremental sampling strategy, region division based on point-growing (RDPG), was proposed. According to RDPG, initial sampling points were set firstly, the Delaunay triangulations were constructed using an improved convex hull interpolation algorithm, the discrete degree values at the intersection points of each side perpendicular bisectors of triangles were calculated with interpolation based on known values of sampling points, and then the point with the biggest discrete degree was chosen as a new sampling point. Using the RDPG method, multiple experiments were carried out in some simulation fields with coefficients of variation from 0.42 to 0.59. The results showed that, based on the same experimental conditions,compared with random sampling and uniform sampling strategies, the RDPG strategy could lead to higher accuracy of regional soil carbon flux estimation. The RDPG strategy took the spatial heterogeneity of soil carbon flux into account, which improved the fitting accuracy of regional soil carbon flux.