【作者】 王盛萍；

【导师】 谢宝元； 张志强； 孙阁；

【作者基本信息】 北京林业大学 ， 生态环境工程， 2007， 博士

【摘要】 黄土高原区域长期气候干旱、区域水资源短缺、土壤侵蚀严重。在气候变暖以及区域退耕还林还草工程实施的背景下，有效揭示区域土地利用与植被变化以及气候变化的生态水文响应具有重要意义。通过MIKESHE模型的生态水文模拟，比较分析了生态水文条件不同的流域生态水文响应特性，目的旨在深刻区别认识不同流域生态-水文关系，为黄土高原区域生态环境恢复、以及最优水资源规划及管理等提供重要理论依据。在应用Zhang模型以及Budyko模型等经验模型探讨流域土地利用与气候变化生态水文响应的基础上，转而借助MIKESHE建立黄土高原吕二沟流域和奥地利Seitengraben流域模型对二者进行模拟和评价；通过设置极端土地利用情景、空间配置情景，以及降水减少20％、气温升高2℃的气候变化情景等模拟分析典型小流域土地利用与气候变化的生态水文响应；针对黄土高原吕二沟流域侵蚀严重等独特的生态水文现象，结合MIKESHE模型与MUSLE模型进行土壤侵蚀模拟评价及情景分析。采用Zhang模型经验模型水文响应分析表明，经验模型模型并未能有效揭示吕二沟流域土地利用变化水文响应；Zhang模型与Budyko模型虽然揭示了气候变化影响的水文响应灵敏度信息，但观测序列气候变异不能完全解释水文变化。定量确定土地利用与气候变化生态水文响应有待于基于物理过程的分布式模型的应用。MIKESHE模型在黄土高原吕二沟流域与奥地利Seitengraben流域的模拟应用中表现有一定差异：模型基本模拟反映吕二沟流域的降雨-产流过程，校正期与检验期相关系数R分别为0.833和0.630，但精度较低，流量偏差F_Bal分别为-40.0％和-17.9％；模型较好反映了Seitengraben流域蓄满产流过程，R分别为0.711和0.785，流量偏差F_Bal分别为-1.97％、-3.56％。吕二沟流域约23.5％、18.78％、24.91％、14.8％径流来自于梯田、草地、村庄及非生产用地、以及灌木林，径流过程涨水段主要由上述区域地表径流组成；Seitengraben流域总地表径流中61.7％和37.5％来自于冬作物区域和林地区域；林地较其余土地利用基流补给持续时间长，各各土地利用地表径流与基流季节分布特征基本一致。极端情景分析表明，Seitengraben流域极端土地利用转变可能引起生态水文响应灵敏度大于吕二沟流域：极端土地利用情景转变引起吕二沟流域和Seitengraben流域蒸发散（占降水量百分比）变化最大分别为2.97％和10.14％，地表径流变化最大分别为3.42％和6.88％，总径流变化最大分别为3.46％和12.6％。空间配置土地利用情景表明：吕二沟流域林地覆盖进一步逐级增加并无显著减少流域产水量；Seitengraben流域林地覆盖每增加10％，模拟径流减少9.5mm。流域二者对降水变化的生态水文响应均较大：降水减少引起吕二沟和Seitengraben流域总径流减少约66.59％～89.40％。吕二沟流域对气温升高水文响应较Seitengraben流域灵敏：气温升高使吕二沟和Seitengraben流域总径流分别减少41.69％～49.83％和3.69％～8.04％。土壤侵蚀空间动态分布特征表明：流域沟道和沟坡侵蚀产沙占全流域90％左右。流域村庄厂矿和居民用地侵蚀产沙占33％，坡耕地与草地分别占27％和24％。流域无明显侵蚀和轻度侵蚀面积占42％和31.87％。不同情景模拟中用材林较原土地利用的减少效益最显著，约70～1398t／km²，其次为草地，果园和灌木林减沙效益相当约70～517t／km²。更多还原

【Abstract】 The Loess Plateau has experienced long time of drought and severe soil erosion. With the proceeding of afforestation and reforestation in the region and the trend of climate warming, it is important to explore the eco-hydrological response to land use change and climate variation. We aim to provide the water resource management, plannging, and eco-environment restoration with theoretical foundation by understanding the various eco-hydrological responses for different catchment system.On the basis of the understanding of the deficiency of empirical model applied in the research of effects of land use change and climatic fluctuation, we built the models of Lv’ergou watershed in the Loess plateau and Seitengraben catchment of Austria, respectively, with the MIKESHE model code. With the validated models, we discussed the potential effects of land use change and climate variation by the simulation of "extreme land use scenarion", "spatial deployment land use scenario" and climatic scenarios which decreased 20% of precipitation and raise the temperature of 2℃. Furthermore by couping MIKESHE modeling results and MUSLE equation we have performed evaluation of spatial distribution of soil erosion of Lv’ergou watershed and scenario simulation.Hydrological effect analysis by Zhang model has shown that the impact of land use change can not been detected with empirical model as much of model error occred with the prediction. With either Zhang model or Budyko model, hydrological response of both watersheds has displayed consistent sensitivity to climate change. However, the attempt of explaning the inter-annual runoff anomy with the climate variation was failed, which indicate the existence of the impact of land use change in the runoff observation to some degree.MIKESHE has generally represent the runoff process of Lv’ergou watershed with the correlationcoefficient of 0.833 and 0.630, and F_Bal of -40% and -17.9%, respectively for calibration and validation period. And the performance of MIKESHE in the application of Seitengraben catchment was well, with R of 0.711 and 0.785, and F_Bal of-1.97% and -3.56%, respectivelly for calibration andvalidation period.Scenario analysis has demonstrated that eco-hydrological response of Seitengraben catchment was more sensitive than that of Lv’ergou watershed. Conversion between various land uses for Lv’ergou watershed would cause the potential maximum change of 2.97% in evapotranspiration and 3.46% in total streamflow, and 10.14% for Seitengraben catchment in evapotranspiration and 12.6% in streamflow. The 20% of increase of forest cover for lv’ergou watershed has reduced the total water yield by 14mm （around 18%）, but the conversion from either sloping filed or grass land to forest didn’t shown significant impact. While for Seitengraben catchment it was found that every 10% of increase of forest cover would reduce the total water yiled by 9.5mm. Both lv’ergou watershed and Seitengraben catchment has shown similar response to precipitation change. With the decrease of 20% of precipitation, simulation of evapotranspiration, overland, as well as total streamflow was found decreased by 6.16% tol2.4%,74.69% to 95.16%, and 66.59% to 89.40%, respectively. Hydrological response of lv’ergou waterhed to temperature rise was more sensitive than that of Seitengraben catchment. With 2 degree of temperature rise, simulation of evapotranspiration for lv’ergou watershed was decreased by 6.4% to 8.7%, runoff by 42.82% to 51.82%, and total streamflow by 41.69% to 49.38%. On the contrary, for Seitengraben catchment, evapotranspiration was decreased by 0.84% to 1.04%, runoffby 4.78% to 10.53%, and total streamflow by 3.69% to 8.04%.The coupling of MIKESHE and MUSLE has displayed that the channel of lv’ergou watershed accounted for 90% of the total sediment yield. 33% of sediment yield was attributed to village and non-agricultural land use, and 27% and 24% to sloping field and grass land, respectively. Scenario simulations for various land use has indicated that the conversion to timber forest has most significant effect on controlling soil erosion with 70 to 1398t/km² of decrease in sediment yield, which was followed by the conversion to grassland. While the conversion to shrub or orchard has shown the least effect with 70 to 517t/km2 of decrease in sediment yield.更多还原