【作者】 张旭斌；

【导师】 黄炎和； 林金石；

【作者基本信息】 福建农林大学 ， 土壤学， 2012， 硕士

【摘要】 崩岗侵蚀是一种特殊的侵蚀形式，了解崩岗侵蚀产沙过程是研究崩岗侵蚀机理所必须解决的关键问题。本研究在福建省安溪县选择3条典型崩岗，在每条崩壁上从下到上采集土样，分析其在不同粒径下的元素含量，筛选出含量显著差异的元素作为最佳指纹因子；同时分析降雨后崩壁侵蚀泥沙中各指纹因子的浓度，通过多元判别法和主成分分析法确定各土层对侵蚀泥沙的贡献率，确定适合南方花岗岩崩岗泥沙来源的组合指纹识别方法；利用该方法分析崩岗系统各部位沉积泥沙来源。主要得到以下结论：1.对南方花岗岩区选定的崩岗土体样品不同粒径下64种测定指标的分析表明，崩岗A通过Kruskal-Wallis H-test筛选出来不同粒径下的指纹因子共有36种，崩岗B有22种，崩崩C有31种。2.应用逐步判别分析的方法对Kruskal-Wallis检验筛选出具有差异的指纹因子进行组合得到最佳组合指纹因子，崩岗A的最佳组合指纹因子有6种，崩岗B的有5种，崩岗C的有5种。3.通过对研究区一次崩岗崩坍的监测，采样，根据多元混合模型计算，得到了这次崩塌来自3种源地的的泥沙贡献率，对于崩岗A来说，红土层、砂土层和碎屑层对崩积堆泥沙的相对贡献分别为66.24%、16.06%、17.7%，相对平均误差为2.70%，对冲积扇泥沙的相对贡献分别为10.25%、38.99%、50.76%，相对平均误差为3.29%；对于崩岗B来说，红土层、砂土层和碎屑层对崩积堆泥沙的相对贡献分别为25.00%、42.08%、39.92%，相对平均误差为5.32%，对冲积扇泥沙的相对贡献分别为30.08%、42.31%、27.61%，相对平均误差为6.37%；对于崩岗C来说，红土层、砂土层和碎屑层对崩积堆泥沙的相对贡献分别为31.38%、50.31%、18.32%，相对平均误差为7.64%，对冲积扇泥沙的相对贡献分别为29.13%、40.54%、30.33%，相对平均误差为9.82%。4.根据计算结果，本次试验得到崩岗A的崩积堆和冲积扇的GOF值为0.86和0.90；崩岗B的崩积堆和冲积扇的GOF值为0.82和0.91；崩岗C的崩积堆和冲积扇的GOF值为0.92和0.88；可见所得到的GOF的值都大于0.8，再次证明由该混合模型所得出的结果可以接受。更多还原

【Abstract】 Slope collapse erosion is a special kind of erosion. Understanding the sedimentformation process of slope collapse erosion is the key to the research on themechanism of slope collapse erosion.In this thesis,three typical eroded slopes werechosen in Anxi county in Fujian province and the soil samples were collected fromeach of the three slopes from the bottom up.Various element contents of the soilsamples of different grain sizes were analyzed and the element contents which wereobviously different from each other were selected as the best fingerprintfactors;Meanwhile, the concentrations of various fingerprint factors in the sedimentforming after the rainfall erosion were analyzed as well and the contribution rates ofvarious soil layers to sediment were determined through the mltivariate discriminantmethod and the principal component analysis. Besides,the composite fingerprintidentification method which was suitable for the analysis on the sediment source ofthe slope collape erosion in the granite area in the south was determined; Thesediment source of every parts of the slope collape system was analyzed through thismethod and the results were as follows:1.Through analysis on the69kinds of the determined indices of the soil samplesof different grain diameters in the granite area in the south,it was showed that throughKruskal-Wallis H-test,the eroded slope A had36fingerprint factors under differentgrain sizes while the eroded slope B and C had22factors and31factors respectively.2.Through the stepwise discriminant analysis method,the fingerprint factorswhich were selected by Kruskal-Wallis H-test and different from each other arecombined to gain the best composite factors.The number of the best compositefingerprint factors of slope A,slope B and slope C was6,5and5respectively.3.Through monitoring and sampling toward the slope collapse in the researcharea and according to the multiplex mixed model,the contribution rate of the threesource regions to the sediment was obtained.As for slope collapse A,the relativecontribution rates of the red soil layer,the sandy soil layer and the fragment layer tothe sediment of the collapse heap were66.24%,16.06%and17.7%respectively and the relative average error was2.70%;the relative contribution rates of the red soillayer,the sandy soil layer and the fragment layer to the sediment of the alluvial fanwere10.25%,38.99%and50.76%respectively and the relative average error was3.29%.As for slope collapse B,the relative contribution rates of the red soil layer,thesandy soil layer and the fragment layer to the sediment of the collapse heap were25%,42.08%and32.92%respectively and the relative average error was5.32%;therelative contribution rates of the red soil layer,the sandy soil layer and the fragmentlayer to the sediment of the alluvial fan were30.08%,42.31%and27.67%respectively and the relative average error was6.37%.As for slope collapse C,therelative contribution rates of the red soil layer,the sandy soil layer and the fragmentlayer to the sediment of the collapse heap were31.38%,50.31%and18.32%respectively and the relative average error was7.64%;the relative contribution rates ofthe red soil layer,the sandy soil layer and the fragment layer to the sediment of thealluvial fan were29.13%,40.54%and30.33%respectively and the relative averageerror was9.82%.4.According to the computing result,the GOF values of the collapse heap and thealluvial fan of the slope collapse A obtained by this experiment were0.86and0.90;theGOF values of the collapse heap and the alluvial fan of the slope collapse B were0.82and0.91;the GOF values of the collapse heap and the alluvial fan of the slope collapseC were0.92and0.88;it was showed that all the GOF values were more than0.8,which again proved that the result obtained by this mixed model was acceptable.更多还原