面向区域土壤碳通量估算的多向插值空间采样策略

区域土壤碳通量的准确测量对陆地生态系统碳循环过程分析具有十分重要的作用。由于土壤碳通量空间异质性强,采用随机抽样的方法无法对区域土壤碳通量进行准确估算,而大范围的多点采样则需要大量的人力及设备成本。基于一种自制的仪器,提出了一种递增式采样的多向插值采样策略(MDI Multiple Directional Interpolation):在设定初始采样点的基础上,通过对已有采样点的测量,通过径向插值的方法计算采样点连线交点,将不同径向计算值差异最大的点作为新增采样点,以此逐步增加。通过对20幅的50×50网格区域仿真,结果表明(1)MDI布局策略能够针对土壤碳通量的变化情况而反馈采样点的疏密。(2)误差分析得出采样点数量(n=10)较少,MDI布局策略对碳通量的估算误差比随机布局策略低,比平均布局策略稍高;随采样点增多,3种布局策略误差均降低;采样点数量n=40,MDI布局策略对碳通量的估算误差(0.028)比平均布局策略的误差(0.32)降低了12.5%,比随机布局策略的误差(0.04)降低了30.0%。MDI布局策略根据土壤碳通量的变化梯度合理分配采样点,降低区域土壤碳通量监测误差。

Spatial sampling layout strategy for regional soil carbon flux estimation based on multiple directional interpolation

Soil carbon flux is a vital part of the terrestrial ecosystem carbon cycle. Carbon dioxide released from soil is 10% to 20% of the total amount of that in atmosphere, equal to carbon assimilation by photosynthesis in plant canopies. Thus accurate measurement of regional soil carbon flux is significant for research on carbon cycling in terrestrial ecosystems. Remote data combined with models to analyze regional carbon balance is important but unreliable. The data from direct measurement of soil carbon flux turns to be more accurate. During the measuring process, random sampling results in inaccurate estimates of the regional soil carbon flux because of strong spatial heterogeneity. Another method, multi-point sampling in a wide area, requires much more human labor and equipment costs. However, a wireless sensor net (WSN) is a new method that has been applied to the measurement of soil carbon flux. Its advantages are low expense, wide coverage, multi-point synchronization, continuous monitoring, and be applicability to a wide range of areas. By using a self-regulating instrument based on WSN, Lr100GE-6400, to measure regional soil carbon flux, this paper presents a new layout strategy, Multiple Directional Interpolation (MDI). The strategy is as follows: (a) Set up the original well-distributed sampling sites and measure the carbon flux of these sites. (b) Calculate the intersections of different connection lines between sampling points by the radial point interpolation method and regard the points with maximum radial differences as new sampling points. Consequently, points are gradually increased based on this method to achieve accurate estimates of regional soil carbon flux. MDI layout strategy considers the spatial heterogeneity of soil carbon flux. It determines the sampling points based on dispersion, so that MDI can provide more accurate and comprehensive spatial information. Based on simulations in 20 fields of the 50×50 mesh region, layout is set up by MDI, random layout strategy, and uniform layout strategy, respectively. Keeping the same sample points, the three strategies were compared, and results were as follows: (1) MDI layout strategy can reflect the density of sampling points in terms of the variation gradient of soil carbon flux, decreasing use of sampling points and increasing the accuracy of measurement. (2) Error analysis of the experiments verified that the MDI layout strategy is more accurate than the average and random layout strategies. If there are fewer sampling points (n=10), errors in the MDI layout strategy are less than the random and more than the average strategies; if sampling points increase, the three types of layout strategies tend to be more accurate. If there are 40 sampling points, errors in MDI (0.028) are 12.5% less than that of the average (0.32),and 30.0% less than that of the random strategy. In conclusion, MDI layout strategy generally allocates the sampling points more reasonably based on the variation gradient of soil carbon flux. More sampling points are selected in the region with a large variation gradient in soil carbon flux, which matches the variation of soil carbon flux accurately, while fewer are located in the region with a small variation gradient in soil carbon flux but these do not affect the accuracy of soil carbon flux measurement. This method maximizes the effect of every sampling point and minimizes errors.