福建省水土流失强度的地统计分析

Geostatistical Analysis on Soil and Water Loss in Fujian Province

结合"3S"技术与野外调查,提取土地利用、植被覆盖度和坡度等信息,生成1995、2000和2007年水土流失现状图,采用地统计模型对3个年份水土流失强度进行计算。结果表明:福建省水土流失强度可用球状的半方差函数模型进行模拟,其方向约呈北北东—南南西走向,与地形走向大体一致;水土流失强度的基台值不断减少,整体向轻度化和均质化方向发展。1995年福建省水土流失强度的结构方差所占比例为12.31%,块金方差所占比例为87.69%,2000年分别为3.51%和96.49%,2007年分别为2.94%和97.06%,块金方差占主导地位且呈上升趋势,表明水土流失强度受多种因素综合影响且人为活动等随机性因素的影响不断加强,小尺度上人为活动等随机性因素为主导因素而中尺度上反之。今后研究中应进一步探讨结构性因素和随机性因素的内涵,并加强小尺度水土流失空间变异中人为活动影响的研究。

The information which includes land use, vegetation cover and slope, was extracted by combining ＂3S＂ techniques with field investigation in the research region of Fujian Province, and the distribution maps of soil and water loss in the year of 1995, 2000 and 2007 were compiled. Moreover, the soil and water loss intensities for 1995, 2000 and 2007 were analyzed respectively by means of Geostatistics. The results show that the soil and water loss intensities can be exactly simulated by the semivariograms of spherical models. The soil and water loss intensities are nearly NNE - SSW directions which are consistent with the topographic orientation. The sill values about the soil and water loss intensities decrease, which means the soil and water loss intensities become weaker and evener. The proportions of structural variance values and nugget variance values are 12.31% and 87.69% （1995）, 3.51% and 96.49% （2000）, and 2.94% and 97.06% （2007） respectively. The proportions of stochastic factors are bigger than structural factors and increase with the time, which means the soil and water loss intensities are closely connected with many factors, and anthropogenic activity is one of the most important factors. Furthermore, the stochastic factors are predominant in small scale, whereas the structural factors are predominant in medium scale. The connotations of structural factors and stochastic factors need further analysis, and natural factors and anthropogenic activity must be considered simultaneously for the spatial variability of soil and water loss in small scale.