采样点布置和空间预测对揭示红壤区土壤有机碳变异性的影响

Understanding the spatial variability of soil organic carbon is very necessary to carry out the soil carbon sequestration program. The allocation and account of soil sampling points and the method of spatial prediction have vital impact on detection of soil organic carbon. This study will be conducted in Yujiang County, Jiangxi Province, China's hilly red soil areas. Based on large sample size, the sampling suitability evaluation on four most frequently used sampling modes, including grid, soil type, land use pattern, and land use pattern-soil type based modes, will be conducted by classical statistics and geostatistics. And the relationships between prediction accuracy and sampling size will be quantitived based on various sampling sizes. Moreover, a new spatial prediction method will be proposed conforming to the spatial variation character of soil organic carbon in red soil areas. The results of this study will provide theoretical reference and indicate which sampling points allocation is more efficient, how many sampling points are needed for special prediction accuracy request, and how to achieve efficient spatial prediction of soil organic carbon at county scale in China's hilly red soil areas.

明确土壤有机碳空间变异是实施土壤固碳的基础。而采样点的布置模式、数量和空间预测方法直接影响到土壤有机碳空间变异的揭示程度。本研究以红壤区-江西省余江县高密度采样点数据为基础，利用经典统计和地统计学方法，对大多数研究者常用的规则网格、土地利用类型、土壤类型及土地利用-土壤类型等4种样点布置模式揭示土壤有机碳空间变异程度进行了评价，进而量化样点数量与有机碳预测精度的关系，并基于红壤区土壤有机碳空间变异特点，提出高效空间预测方法。本研究结果可为红壤区县级尺度哪种样点布置模式的采样效率更高？在特定精度要求下应布设多少个采样点？以及何种方法可实现土壤有机碳的高效空间预测等科学问题的解决提供理论依据。

针对中国红壤丘陵区土壤有机碳（SOC）空间变异性的准确揭示相对困难，本课题研究从采样点布置模式、数量及点面拓展方法三个方面对揭示红壤区县域SOC空间变异的影响。在对比分析了已有的几种采样点布置模式的采样效率，提出了按土地利用—土壤类型法进行样点布置的采样效率较高，而广泛使用的网格法的效率较低；基于不同的采样点数量，分析了采样点数量变化对揭示SOC空间变异性精度的影响，并给出了研究区采样点数量与SOC空间预测精度的量化关系，以及各土壤类型和土地利用方式的SOC预测精度与样点数量之间的关系；在对比分析了多种不同的点面拓展方法对SOC空间预测精度的基础上，明确了各种点面拓展方法的优劣，指出预测效果最好的点面拓展方法是结合土地利用方式信息的克里格和图斑连接方法（KLu和PLu），而OK和PGr方法的预测精度均较差。本课题研究结果对制定区域合理SOC调查采样设计和高效获取红壤区SOC空间分布具有重要的指导价值。