水保所知识产出(1956---)知识库

本文针对近年来黄河泥沙骤减的现状以及发生的原因和机理这一热点问题，
在延安市燕沟流域中部选择一个 2003 年的新建成淤地坝，对小流域近十年的淤
地坝沉积泥沙剖面进行了采样分析，旨在研究退耕还林（草）政策实施四年后成
效进入显现期小流域侵蚀泥沙的演变，进而为解疑黄河泥沙的减少以及退耕还林
（草）政策实施效益评估提供依据。
文中通过淤地坝沉积泥沙剖面沉积旋回和小流域侵蚀性降雨对应关系的分
析，对小流域近十年的土壤侵蚀模数进行了计算；同时利用遥感手段分析了小流
域相应年的植被变化情况，并对植被恢复过程中的小流域的土壤侵蚀变化进行了
分析，明确了两者之间的关系；进一步利用复合指纹识别技术探索了小流域侵蚀
泥沙的来源。得出了以下主要结论：
（1）阐明了小流域 2003-2011 年的土地利用类型的变化，并使用遥感手段
提取分析了小流域植被恢复的时空演变过程。结果表明，小流域主要的土地利用
类型的变化是退耕还林（草）工程造成的，退耕总面积为 1.03 km 2 ，占研究流域
总面积的 19.7%。其中主要的退耕工程在 1999 年，占退耕地总面积的 64.1%，
而 2000-2001 年和 2002 年分别占 12.6%、23.3%。退耕地的植被类型以乔木为主，
且主要是以乔-灌-草相结合的方式出现。而乔木类型以刺槐和苹果树为主；灌木
类型主要是柠条，而草的类型主要是黄蒿和杂草。小流域年平均 NDVI 的平均值
0.61，整体上呈现增加趋势，说明植被覆盖增加，生态环境好转。从变化过程上
看可以分成四个阶段，分别是：①1999-2002 年快速增长阶段，平均增长率 2%；
②2003-2006 年的第二次快速增长阶段，平均增长率也是 2%；③2007-2008 年的
基本稳定阶段；④2009-2011 年的降低阶段，平均减少率 1%。且年尺度上，小流
域 2003-2011 年间年平均降雨量和 NDVI 之间无显著相关性。
（2）分析了淤地坝沉积泥沙剖面中的颗粒组成的含量及分布特征，揭示了
小流域的侵蚀泥沙搬运及沉积过程，结果表明，研究区的沉积泥沙剖面样品中以
粉粒为主，平均含量占总量的 71.83 %，其次为砂粒和粘粒，分别为 12.64%和
15.53%。坝地沉积剖面存在粒径变异的转折点，这个点位可用来衡量水沙动力条件和沟道比降对泥沙粒径分布所起的作用。不同粒径颗粒在沉积泥沙样品中的变
异程度不同，粘粒和砂粒均为中等变异，粉粒为弱变异，并且砂粒的变异性最大。
（3）使用沉积泥沙剖面和降雨资料反演了小流域泥沙侵蚀沉积规律，阐明
了小流域土壤侵蚀历史演变过程。结果表明，小流域 9 年的泥沙侵蚀总量为
6063.91 t，而侵蚀模数 9 年平均值为 137.04 t/km 2 · a，下降至小流域综合治理工程
实施前的 2.3%，平均年下降率 10.86%。并且从 2003 年开始，小流域的土壤侵
蚀模数均不足 200 t/km 2 · a，均为轻度侵蚀，虽然年际之间有波动，但很小可以忽
略不计。这表明在新的侵蚀环境背景下该小流域的侵蚀强度已稳定在轻度以下，
同时证明小流域综合治理的实施对减少区域土壤侵蚀的作用是显著的。
（4）研究了小流域降雨量和植被恢复与土壤侵蚀的相关关系，阐明了小流
域生态恢复过程的泥沙响应机制，结果显示，小流域 2003-2011 年的土壤侵蚀受
小流域年平均降雨量和植被重建的共同影响。小流域的植被重建对土壤侵蚀的减
少作用在 2003 以前已经开始发挥，并且在 2003-2011 年之间发挥了其主导作用。
在这种降雨和植被重建的共同作用下，小流域的土壤侵蚀稳定在了轻度侵蚀以
下。因此，在小流域降雨未发生较大变异的状况下，研究小流域在植被重建的作
用下，4 年后小流域的侵蚀状况达到稳定状态。然而，小流域降雨与人为活动之
间的交互作用对小流域土壤侵蚀的研究需要进一步探索。
（5）运用复合指纹识别技术探索了小流域侵蚀泥沙的来源。通过对研究小
流域内林地、荒草地、灌木地、缓坡耕地、苹果园和支沟道共 6 种泥沙源地土壤
表层样品中的 12 种物质的测定，利用多元判别分析，筛选出 Cu、Fe、Mg、Mn
和 K 五种指纹指标组成最佳复合指纹因子组合。进一步利用这一组合，采用最
优化混合模型计算了 6 种泥沙源地对整个小流域以及每一沉积旋回侵蚀产沙的
贡献率。结果表明每一沉积旋回中来自坡耕地、草地、林地和支沟道的泥沙贡献
比差异较大。2003-2011 年淤地坝的运行阶段，林地、荒草地、灌木地、缓坡耕
地、苹果园和支沟道的泥沙贡献率分别为 29.36%，13.9%，2.43%，11.74%，2.2%
和 40.38%，侵蚀泥沙主要来源于支沟道。
关键词：泥沙响应；侵蚀环境演变；典型流域；植被归一化指数（NDVI）

A phenomenon of sediment discharge was significantly decreasing in the Yellow
River, and research on the causes and mechanism of this hotspot phenomenon is
seriously lacking. A check dam, which was constructed at 2003, was selected from the
middle reaches of the Yangou watershed for analyzing the deposition sediment profile
in recent ten years. This thesis was aimd to analyze the evolution of watershed
sediment feature during the performance period when the‘grain for green project’
(GGP) program has been implemented four years. The erosion modulus from 2003 to
2011 was obtained according to the relationship between the deposition sediment size
diameter and the annual rainfall in the watershed. The annual change of vegetation in
the watershed was analyzed by remote sensing technology, and the relationship
between vegetation and erosion was established. The source of sediment from the
watershed was identified by using composite fingerprint recognition technology.
The results of the study provide a scientific reference to evaluate the implementation
benefits of the GGP program and it may illuminate the significantly decreasing
sediment discharge in the Yellow River. The main conclusions are as follows:
(1) The change of land use type in watershed was analyzed, and the temporal and
spatial evolution of vegetation restoration in small watershed was also analyzed by
remote sensing. The change of land use type in the watershed was caused by the
implementation of the GGP program. The slope farmland decreased 1.03 km 2 , which
account for 19.7% watershed area. And 64.1%, 12.6% and 23.3% of abandoned
farmland was conducted at 1999, 2000-2001 and 2002, respectively. Tree was the
main vegetation type grow in the abandoned farmland, and mainly appeared in the
form of the combination of tree, shrub and grass type. Tree type is dominated by
Robinia pseudoacacia and apple trees, shrub is dominated by Caragana korshinskii,
and grass is dominated by Artemisia annua and weed. The annual NDVI (normalized
vegetation index) showed an increasing trend with an average value of 0.61,
indicating an increasing vegetation cover and improving ecological environment. The
change process can be divided into four stage: 1) the increasing stage with an average  degree of 2% in 1999-2002; 2) the second increasing stage with an average degree of
2% in 2003-2006; 3) stable stage in 2007-2008; 4) the decreasing stage with an
average degree of 1% in 2009-2011. There have no significant difference between
annual rainfall and NDVI during 2003-2011.
(2) The detachment and transport process of watershed was revealed by
analyzing the particle size distribution of the deposition sediment profile. Silt content
was the domination particle in the deposition sediment with an average value of
71.83%, followed by clay and sand content with average value of 15.53% and 12.64%,
respectively. Among the deposition profile, there exist a particle size distribution
turning point, which can be used as discriminating the influence of hydro-sediment
dynamics and gully slope on the particle size distribution in the deposition sediment.
The variation degree in different sediment particles is different. The silt and sand
content are medium variation while clay content belongs to weak variation and the
variability of sand content is the biggest among them.
(3) The historical evolution of soil erosion was clarified by analyzing the law of
sediment deposition according to the relationship between the deposition sediment
size diameter and the annual rainfall in the watershed. The total soil erosion was
6063.91 t and the average erosion modulus was 137.04 t/(km 2 •a), which was only
2.3% of the value before watershed comprehensive control project implementation,
and the average annual decline rate is 10.86%. The watershed erosion modulus was
mild erosion, which was less than 200 t/(km 2 •a) and with slight annual fluctuation.
This indicates that the erosion modulus was mild erosion in the watershed in the new
erosion environment background and be of great significance effect for watershed
comprehensive control project implementation.
(4) The correlation relationship among annual mean rainfall, vegetation and
erosion was established, and the sediment response mechanism was elucidated. The
watershed erosion during 2003-2011 was influence by joint effect of annual mean
rainfall and vegetation restoration, while there was no significant correlation between
soil erosion and annual mean rainfall and average NDVI. The vegetation restoration
for soil erosion reduction effect had been exerted before 2003, and it was the  domination effect during 2003-2011. The watershed erosion modulus was controlled
as mild erosion by the joint effect of annual rainfall and vegetation restoration.
Therefore, the erosion condition would reach a stable stage after 5 years of the GGP
program implementation if the rainfall condition do not have great change. However,
the joint effect of human activity and vegetation restoration on watershed soil erosion
is still need further research.
(5) The source of sediment from the watershed was identified by using
composite fingerprint recognition technology. Forest land, grassland, shrub land,
gentle slope farmland, apple orchard and branch channel were identified as potential
sediment sources. In total, 12 properties of soil surface samples was determined from
six kinds of potential sediment sources and Cu, Fe, Mg, Mn and K were chosen as the
optimum fingerprint factors by using multiple discriminant analysis. The erosion
contribution of each sediment sources was calculated by optimization hybrid model
using these optimum fingerprint factors. The results indicated the calculated
contribution of sediment from slope farmland, grassland, woodland and branch
channel under different deposition layer is of great variation. The contribution of
sediment from forest land, grassland, shrub land, gentle slope farmland, apple orchard
and branch channel were 29.36%，13.9%，2.43%，11.74%，2.2% and 40.38%,
respectively, and sediment was dominated came from the branch channel area.
Keywords: Sediment response;Erosion environment evolution;A typical watershed;
The normalized vegetation index (NDVI)