Yan Jun-Xia, Li Jun-Jian, Li Hong-Jian, Zhang Yi-Hui
Institute of Loess Plateau, Shanxi University, Taiyuan 030006, China.
Huan Jing Ke Xue. 2013 Oct;34(10):3992-9.
We measured soil respiration (R) and the environmental factors influencing Rs including soil temperature ( T, ) , soil water content (SWC) and soil organic carbon (SOC) in a subalpine meadow in Yundin mountain of Shanxi province, and performed an analysis of the heterogeneity of Rs, T10 and SWC and SOC as well as their relationships using both traditional statistics and geostatistics methods. The results from traditional statistics showed that the measured data of R and the environmental factors exhibited a normal distribution with variation coefficients ranging between 12% and 24%. The variation of all the measured factors was in the middle range. The fact that the correlation coefficient between Rs, and SOC (r =0. 61) was larger than that between Rs, and T15(r =0. 27) and SWC (r = 0. 26) indicates that the heterogeneity of Rs was controlled mainly by SOC. The results from geostatistics analyses showed that linear model could reflect the spatial characteristics of Rs and the environmental factors. The C0/(C0 + C) values for Rs, T10, SWC and SOC were 41% , 3% , 77% , and 57% , respectively, indicating that the spatial heterogeneities of both RB and T10 were resulted mainly from structural factor, but the spatial heterogeneities of SWC and SOC were controlled mainly by random factor. The range of semi-variogram function was 53. 2 m for Rs, T, and SWC, and 52. 1 m for SOC. The fractal dimension value of Rs, T10, SWC, and SOC was 1. 85, 1. 60, 1. 96, and 1. 95, respectively, indicating that SWC had the weakest spatial dependence on scale and the most complicated spatial distribution pattern, while T10 had the simplest spatial distribution pattern. The spatial distribution of Rs showed a similar distribution character to both SWC and SOC, but had its own regularity. With the decrease of the confidence level and the estimated accuracy, the required sampling number of the Rs and the environmental factors measurements declined substantially.
我们测定了山西省云顶山亚高山草甸的土壤呼吸（R）以及影响土壤呼吸速率（Rs）的环境因子，包括土壤温度（T）、土壤含水量（SWC）和土壤有机碳（SOC），并运用传统统计学和地统计学方法对Rs、T、SWC和SOC的空间异质性及其关系进行了分析。传统统计学结果表明，R和环境因子的实测数据呈正态分布，变异系数在12%至24%之间。所有实测因子的变异处于中等范围。Rs与SOC之间的相关系数（r = 0.61）大于Rs与T（r = 0.27）和SWC（r = 0.26）之间的相关系数，这表明Rs的空间异质性主要受SOC控制。地统计学分析结果表明，线性模型能够反映Rs和环境因子的空间特征。Rs、T、SWC和SOC的C0/(C0 + C)值分别为41%、3%、77%和57%，这表明Rs和T的空间异质性主要由结构因子引起，而SWC和SOC的空间异质性主要受随机因子控制。Rs、T和SWC的半变异函数范围为53.2 m，SOC为52.1 m。Rs、T、SWC和SOC的分形维数值分别为1.85、1.60、1.96和1.95，这表明SWC在尺度上的空间依赖性最弱，空间分布模式最复杂，而T的空间分布模式最简单。Rs的空间分布与SWC和SOC具有相似的分布特征，但有其自身规律。随着置信水平和估计精度的降低，Rs和环境因子测量所需的采样数量大幅下降。
