水稻土物理性质空间变异性研究

SPATIAL VARIABILITY OF PHYSICAL PROPERTIES OF PADDY SOIL IN FIELD

本文应用时间序列分析方法和随机统计理论研究了水稻土物理性质田间实际观测值的变异特性,着重讨论了它们在二维平面上的空间变异结构。结果表明,各项性质在田间的变异和分布不是完全随机的,在一定的范围内,各测点的观测值之间存在着空间相关现象。因此,根据统计学原理,土壤观测样点的选取,除了需有合适的数目以外,还应确定观测点的合适大小以及观测点之间的合适间距。同时,本文通过土壤大团聚体含量变异原因的分析,表明了各性质的变异是相互关联、相互作用的。对以空间变异结构为基础的Kriging内插技术作了初步的尝试,取得了较好的效果,其內插精度比趋势面法有明显提高,内插值与真值之间的方差平均降低了32%。

Spatial variability of silt and clay contents, macroaggregate and microaggregate contents and organic matter content measurements was studied by using stochastic theory and time series analysis theory.The measurements were made on thirty sites (1×1 meters in size) selected at random in a silty-loamy paddy soil with an area of 127m×71m. Spatial variability of each soil property were characterized statistically by autocorrelation function, semivariance function and their diagram. The results show that the physical properties of paddy soil are not completely disordered in field. There are spatial correlation for measurements at all the observation sites. The limit distance of spatial correlation of silt, clay.macroaggregate, microaggregate and organic matter content are 25, 25, 27, 23 and 57 meters,respectivily. Interrelations among variat■ns of diffirent properties was studied using path analysis. It shows that the variation of macroaggregate more than 70% is caused by three other variations:microaggregate (34%), silt (23%) and organic matter (16%). Finally, Kriging interpolation have been used for spattial prediction of silt content in filed. Comparing with surface interpolation, the average variance estimated by Kriging interpolation is deseased by 32%.