黄土高原小流域洪水泥沙来源的复合指纹分析法研究
文献类型：学位论文
| 学位类别 | 硕士 |
| 答辩日期 | 2008 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 陕西 |
| 导师 | 杨明义 |
| 学位专业 | 土壤学 |
| 中文摘要 | 流域泥沙来源分析对流域治理及水土保持措施效益评价有重要意义。本研究以典型黄土丘陵沟壑区燕沟流域内一支沟——庙沟作为研究区，通过采集研究区内主沟道沟壁、果园、农地及支沟道土壤样品，分析了采集样品中多种土壤性质，结合洪水泥沙过程监测，利用复合指纹识别技术，研究了洪水携沙过程中不同泥沙源地对洪水泥沙贡献比的动态变化，主要取得如下结论：1 筛选出土壤指纹识别因子测定采集的土壤样品性质，本试验共测定21种土壤性质。对测定的土壤性质利用非参数Kruskal-Wallis H-test的统计分析检验，土壤全氮等12种土壤性质其检验P值低于0.05，表明这12种土壤指标在四种泥沙源地间有显著差异，可以作为土壤指纹识别因子。2 得到了最佳指纹识别因子组合利用多元逐步判别分析，在12种可用土壤指纹识别因子中，筛选出土壤全氮、低频质量磁化率、Mg、Ca、Cs-137组成最佳组合。利用这5种土壤性质组合，主沟道土壤样品正确判别率达到100%，果园为90.9%，农地为78.1%，支沟道为92.3%，总判别正确率达到84.7%。3 研究了不同源地洪水泥沙贡献比以筛选出的5种土壤性质组成最佳指纹识别因子组合，利用多元混合模型，对一次洪水携带的泥沙中各源地的泥沙贡献比进行了动态分析，尤其对来自主沟道的泥沙贡献比的变化分析，隐含了可能发生重力侵蚀的信息，对今后定量化研究黄土区的重力侵蚀提供了手段。通过加权平均，总体上，本次洪水携带的总泥沙有26.4%来自主沟道沟壁，65.5%来自果园，4.8%来自小块的农地以及3.3%来自支沟道。4 证明了在黄土高原可以应用复合指纹识别方法研究洪水泥沙来源通过对泥沙来源研究结果相对平均误差分析，其最大值为18.7%，最小值为2.7%，表明利用复合指纹识别技术研究黄土高原泥沙来源可信度非常高。另外，在土壤质地较为均一的黄土丘陵区能通过严格的数学统计分析，找到最佳指纹识别因子组合，证明复合指纹识别方法可以在黄土高原地区用来进行泥沙来源的研究。 |
| 英文摘要 | Information on flood suspended sediment sources is an important requirement for the catchment control and the assessment of soil conservation measurements. The soil samples were collected on the surface of the orchard and cultivated slopes, and the gully walls on the main gully and sub-gullies in a small catchment-Miaogou catchment in the north of Shaanxi area. The concentration values of 21 fingerprint properties were analyzed using Kruskal-Wallis H-test and multivariate discriminant function analysis. Combining the monitoring of water and flood sediment flux, we studied the dynamic contribution changes of different sediment sources during flood event using composite fingerprinting technique. The main conclusions take follow: 1 Assessing the ability of the tracer properties to discriminate the sources types In this study we analyzed twenty-one kinds of soil properties. All analyzed soil properties were tested for their ability to distinguish individual source types, using non-parametric Kruskal-Wallis H test. Totally, 12 properties which contain soil total nitrogen and so on yielded P-values lower than 0.05. It shows that these properties have significant difference between the four sediment source areas and they can use as fingerprinting factors.2 Identifying the best composite fingerprint for discriminating source materialsStepwise multivariate discrimiant function analysis was employed to identify an optimum composite fingerprint from the properties selected. Soil total nitrogen, Low frequency quality magnetic susceptibility, Mg, Ca and Cs-137 make the optimum composite fingerprint. Using the composite fingerprint, 100% of main gully source material samples were classified correctly, 81.8% of the branch gullies, 78% of the cultivated land and 92.3% of the orchard source material samples was classified correctly, and 84.7% of the total source material samples was classified correctly.3 Establishing the relative contributions from the individual source groupsUsing a multivariate mixing model with the composite fingerprint to dynamic analyze sediment contribution of the different sediment source during individual flood event. Especially the information of the contribution changes of main gully sediment implied the occurrence possibility of gravitational erosion. It provided one method of study gravitational erosion in loess region. Totally, the contribution in this flood from gully wall, orchard land, farmland and branch channel were estimated to be c. 26.4, 65.5, 4.8 and 3.3%.4 Documenting the availability of using composite fingerprinting technique to study flood sediment source on the loess plateau.The maximum and minimum value of relative mean error to the result of sediment source is 18.7% and 2.7%. The result indicated the confidence level is very high when using of the composite fingerprinting technique to study sediment source on the loess plateau. |
| 公开日期 | 2011-07-01 |
| 源URL | [http://ir.iswc.ac.cn/handle/361005/4105]  |
| 专题 | 水土保持研究所_水保所知识产出(1956-2013) |
| 推荐引用方式 GB/T 7714 | . 黄土高原小流域洪水泥沙来源的复合指纹分析法研究[D]. 陕西. 中国科学院研究生院. 2008. |
入库方式： OAI收割
来源：水土保持研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。