【作者】 唐丽霞；

【导师】 谢宝元； 张志强；

【作者基本信息】 北京林业大学 ， 工程绿化， 2009， 博士

【摘要】 黄土高原地区面临着水土流失严重和水资源短缺等关键的生态问题,在气候变化和黄土高原以植被恢复、保持水土为主要内容的生态恢复与治理工程实施的背景下,黄土高原区河川径流和输沙演变规律及其成因的研究越来越受到广泛的关注和重视。本文以黄土高原区典型中尺度流域清水河流域(面积436km~2)为研究对象,采用非参数统计(Mann-Kendall)秩检验方法结合滑动T检验和跃变参数分析法对流域内径流、输沙及气候因素进行趋势分析和突变点检验;同时基于流域内土地利用变化特点,分析土地利用和气候因素变异对径流和输沙变化的贡献率;最后结合SWAT模型和情景模拟的手段分析清水河流域土地利用/植被变化下流域径流和输沙的变化,旨在为黄土高原中尺度流域建设高效水资源利用和有效防治水土流失的植被配置方式提供理论依据。对航片和TM影像解译土地利用数据及土地利用现状图分析,清水河流域1959-2007年间土地利用/植被覆盖变化特征为乔木林面积增加,灌木林面积增加,草地面积减少,农地面积减少,园地面积增加,居民区面积迅速增加。流域内针叶林和阔叶林分布在高海拔、204～0°的阴坡和沟谷;农地和园地分布于低海拔、缓坡的阳坡;灌木林和草地分布在海拔850～1550m的范围内。结合Fragstats软件分析,整个清水河流域景观斑块数目减少,斑块平均面积扩大,面积加权平均形状因子减小,反映出景观的破碎度降低,景观连接性增强,景观异质性减弱;景观多样性方面,原有的优势景观类型优势度降低,各种类型斑块在景观中的分布向着均匀化方向演变。MK趋势性分析和突变点检验表明:降雨量及极端降雨指数均无趋势性变化;平均气温、最高气温、最低气温有显著的上升趋势,突变点位于1997年、1994年和1998年;潜在蒸发散有极显著的上升趋势,突变点在1997年;径流量和输沙量都有极显著的下降趋势,突变点同在1980年。根据水量平衡原理分析,气候因素对径流减少的贡献率为48.17%。而土地利用的贡献率为51.83%。根据通用土壤流失方程分析,降雨因素对输沙量减少的贡献率为9.89%。而土地利用的贡献率为90.11%。土地利用变化中主要是林地的变化,可见,清水河流域林地面积的增加是导致径流和输沙减少的原因之一。SWAT模型模拟结果表明:平均径流、输沙在校准期与验证期的模型效率系数Ens都在0.5以上,相对误差小于20%,R~2达到0.7以上,表明模型在模拟清水河流域的径流和输沙方面具有良好的适用性。对降雨、径流深和输沙量的空间分布分析表明,降雨对径流和泥沙的影响有直接和间接两个方面,间接影响表现在其对植被分布的影响,另外地形因素起着重要的作用,坡度大的子流域,其径流和输沙都高于其它子流域。不确定性分析结果表明,径流和输沙对DEM的分辨率响应显著,两者都表现出随DEM分辨率的增大而增加的变化趋势。不同潜在蒸发散计算径流量高低次序为Priestley-Taylor＞Penman-Monteith＞Hargreaves。极端土地利用情景下,径流的高低次序依次为灌木林地＞草地＞阔叶林地＞园地＞针叶林,与原地类相比灌木林地和草地均使模拟径流增加,其余的情景都使模拟径流减少。不同情景的输沙量依次为草地＞灌木林地＞阔叶林地＞园地＞针叶林地,与原地类比较草地使模拟输沙量增加,其余的情景都使模拟输沙量减少;不同森林覆盖率情景下年均径流和输沙量都随森林覆盖率的增大呈下降趋势,森林覆盖率在40-70%之间的输沙量下降速率最大。根据极端土地利用情景模拟结果和土地利用的地形分异特点,为高效利用水资源同时减少土壤侵蚀,设计土地利用/植被覆盖的最优情景,并模拟径流和输沙,结果表明:最优情景下径流深增加16.34%,输沙模数减小94.71%,可见,最优情景的植被配置起到了高效利用水资源和减少土壤侵蚀的作用。同时,最优情景下径流的空间分布与原地类相比没有太大的改变,输沙模数的空间分布呈现出高值区由集中到分散的变化趋势,说明在地类发生改变的同时,影响输沙因素的敏感性也发生了变化,植被因素的敏感性降低,径流和降雨的敏感性增加,同时地形因素依然是最敏感的因素。更多还原

【Abstract】 The decreasing Yellow River streamflow during the past fifty years has been attributed to land use/land cover change and climate variability in addition to the uptake by water projects.Large scale re-vegetation based soil and water conservation deployed to combat the severe soil erosion since 50s of last century on the Loess Plateau of China have resulted in the dramatic change of land use and land cover during the past six decades or so.On the other hand,the region is considered one of the most sensitive aera for globle climate variability.Therefore,it is critically important to understand and predict the coupling effects of land use and cliamate variability on the runoff and sediment characteristics for integrated watershed management and ecological restoration.The study examined the trend of annual streamflow,sediment,precipitation,temperature and PET and their change points for a typical mesoscale Qingshui River watershed located on the Loess Plateau by using non-parametric Mann-Kendall test,Moving t-test technique,and hopped parameter analysis.A simple water balance and mass balance based methodology was used to separate the landuse change effects on the runoff and sediment yields from that of climate variability.To predict the future land use/land cover change effects on the runoff and sediment yield,the calibrated and validated at both annual and monthly scales SWAT was used to analyze the potential effects of land use/ land cover scenarios on the sediment and streamflow.The ultimate objective of model excercises was to provide the sound theoretical basis for vegetation restoration and management aiming at optimized configuration of vegetation across the watershed to improve the soil erosion control while reduced water use. Based on the land use data and the land use map which interpretated by aerial photo and TM images,the main transfer characters of land use/land cover is woodland increased,shrubland increased, grassland decreased,cropland decreased,orchard increased and urban rapidly increased.In the basin broed-leaved forests and coniferous forests are located in the high-altitude,shadow slope and bottom valley with slope of 20～40°;cropland and orchard are located in the low-altitude of gentle and sunny slope;shrubland and grassland are located in the altitude of 850～1550m.By using fragstats software analysis,the number of landscape patches decreased,the average area of patches increased and AWMSI decreased,which indicated that the fragmentation of landscape become lower,the landscape connectivity become better and the landscape heterogeneity become weaken.On the side of landscape diversity,the dominance of original dominance landscape reduced and the distribution of various patches were apt to homogenization.The results of MK trend analysis and change point test showed that precipitation and extreme precipitation had no significant trend,;while mean temperature,maximum temperature and minimum temperature exhibited significant increasing trend,the change points occurred in 1997、1994 and 1998. There was a significant upward trend in PET and its change point was 1997.Both runoff and sediment discharge exhibited downward trend and they shared the same change point of 1980.Based on the principle of water balance,it was estimated that climatic variables accounted for 48.17%of the reduction in mean annual streamflow,and the other 51.83%contribution to the change of land-use. Then by using USLE,it was found that precipitation contributed 9.89%to the sediment discharge relief and the contribution of land-use was 90.11%.Among the land-use changes,the woodland’s change play an important role,so the increase of forest area appeared to be the main cause of the reduced streamflow and sediment discharge.The simulated results by SWAT model showed that:in the calibration and verifcation periods the efficient coefficient Ens of the average runoff and sediment discharge were over 0.5,the relative error were less than 20%and R~2 reached more than 0.7,which indicated that the model had good applicability in Qingshui River Watershed.According to the distribution of the precipitation,depth of runoff and sediment discharge,it was known that precipitation effected runoff and sediment discharge on both direct and indirect,the indirect showed the effects on the distribution of vegetation.On the other hand terrain factors played an important role especially in the slope,runoff and sediment discharge of sub-basins with the steep slope were higher than others.Uncertainty analysis results showed that runoff and sediment discharge were significantly in response to the resolution of the DEM, and exhibited increasing trend with the DEM resolution increased.The order of runoff calculated by different PET method was Priestley-Taylor＞Penmon＞Hargreaves.The runoff followed the order of shrubland＞grassland＞broad-leaved woodland＞orchard＞coniferous woodland under the extreme scenarios.Compared with the original landuse type,the shrub land and grassland made the runoff increase,and the rest of scenarios led runoff reduce.With different scenarios sediment discharge followed by grassland＞shrubland＞broad-leaved woodland＞orchard＞coniferous woodland,among the scenarios grassland made sediment diacharge increase while the others made sediment discharge reduce.Along with the forest cover increased,the runoff and sediment discharge decreased;meanwhile the sediment discharge had the maximum rate of descent with the forest coverage rate between 40-70%.For efficient using of water resources and reducing the soil erosion,the optimal landuse scenario was designed according to the simulated results of extreme landuse sceniaros and the characters of landuse with differential terrain.The simulated results of runoff and sediment diacharge with optimal scenario showed that:the depth of runoff and sediment modulus decrease16.34%and 94.71%.It seemed that the optimal scenario played an active role in efficient use water resources and lessen soil erosion.At the same time,the spatial distribution of runoff of optimal scenario had no change,but the spatial distribution of sediment modulus had a great change that high-value areas was from concentration to scatter which illustrated that the sensitivity of the factors affected sediment had changed,the sensitivity of vegetation was decreased and the sensitivity of precipitation and runoff was increased,while the terrain factor is still the most sensitive factor.更多还原