丘塘景观土壤养分的空间变异

Spatial variability of soil nutrients in a hill-pond landscape

理解土壤养分空间分布的异质性对于评价和管理土地资源具有重要意义。利用地理信息系统和地统计学方法定量研究了丘塘景观土壤养分的空间异质性特征。结果表明 ,土壤全氮的有效变程最大 (为 16 5 m) ,有机碳次之 (10 2 m) ,而全磷的变程最小(90 m)。土壤有机碳含量由高到低 5个不同级别的土壤面积与丘塘景观整个面积的比值的变化范围较小 (10 .5 4 %～ 2 3.15 % ) ,土壤全氮含量比值变化范围较大 (5 .79%～ 32 .73% ) ,土壤全磷的比值的变化范围最大 (1.80 %～ 4 2 .0 6 % )。土壤有机碳和土壤全氮的分布情况较为一致 ,不同级别斑块分布也很相似 ,土壤有机碳含量高的地方土壤全氮也高。表层土壤有机碳和全磷的空间异质分布用球状模型拟合最佳 ,而全氮的空间分布规律更宜用指数模型来拟合。景观尺度的半方差拟合总体上优于斑块尺度。相对有机碳和全氮 ,全磷的空间异质性更多由随机因素 (如人类施肥活动 )引起和决定。土壤全氮的取样尺度应大于 16 5 m,而有机碳、全磷的取样距离则分别可大于 10 3m和 90 m。

Understanding the spatial variability of soil nutrients and its relationship with ecosystem biogeochemical cycles is important for managing and utilizing land resources. The objective of this study was to characterize the spatial variation of soil nutrients in a hill-pond landscape. This landscape is composed of four patch types: forest in the upper slope, a fish pond at the bottom of the watershed, a fruit garden and grassland in between the forest and fish pond. Soil samples at 0～10 cm and 10～20 cm depths were collected from each of the four patch types. Total carbon content (TC), total nitrogen (TN), and total phosphorus (TP) were analyzed. The spatial variability of the nutrients at the two soil layers was analyzed using geostatistical methods and Geographical Information System (GIS). Semivariance analysis showed that the effective range of TN (164.1 m) was larger than TC (102.4 m) and TP (89.2 m) at 0～10 cm depth. The effective range of TN increased from 164.1 m at 0～10cm to 1242.7m at 10～20cm, whereas those of TC and TP remained almost unchanged. The ratios of nugget to sill for TC, TN, and TP were 5%, 6%, and 49%, respectively. In terms of the ratio of nugget to sill, the distribution pattern of TC was similar to TN. The content of each of the three types of soil nutrients were classified into 5 levels, each of which occupied a certain amount of landscape area. TC showed less variation (10.54%～23.15%) than TN (5.79%～32.73%) and TP (1.80%～42.06%). In general, the semi-variogram for the models three soil nutrients fit better at the landscape than at the patch type scales. TC and TP were better fitted by spherical model, or exponential model was better for TN. The spatially structured variance (the variance due to the location of sampling sites) accounted for a large proportion of the sample variance for OC(95%), TN(94%) and TP(53%). According the effective range, the efficiency of the sampling can be improved. The scale of the sampling is 165m(TN), 103m(OC) and 90m(TP).To enhance analyzing precision we need increase the sample number, especially for the orchard.