【作者】 曹明亮；

【导师】 周惠成；

【作者基本信息】 大连理工大学 ， 水文学及水资源， 2011， 博士

【摘要】 水资源问题是21世纪全球资源环境的首要问题,直接关系着人类的生存和社会的发展。我国人均水资源量不足、水资源时空分布不均的两大特点,决定了供水安全与防洪安全是我国经济社会发展的重大任务和重要保证。近几十年来,随着人口增长和社会经济的发展,人类活动使水资源的赋存形式发生了深刻变化,使河川径流减少,极端洪水灾害事件增加,供水和防洪问题更加突出。这些变化给我国水资源研究领域带来了许多新问题,如人类活动影响下的径流演变规律与水资源供需平衡问题,及人类活动影响下的洪水预报与洪灾防治问题等。本文以此为切入点,以辽宁省碧流河水库流域为对象,研究气候变化与人类活动对以城市供水为主大型水库来水量的影响；以吉林省丰满水库流域为对象,研究中小水利工程对洪水预报的影响。主要研究内容与成果如下：(1)根据人类活动对径流及洪水的影响特点,将人类活动影响因子分为三类：是影响流域的产汇流条件,主要包括土地利用变化、水保工程、河道整治等影响；二是人类直接取用水,主要包括跨流域调水及灌区等引水工程、地下水开采等；三是影响流域河道的汇流过程,主要包括水库塘坝等蓄水工程的影响。针对上述人类活动影响因子,分析了人类活动对碧流河水库流域和丰满水库流域的径流及洪水的影响特点。(2)采用Kendall秩次相关法、Spearman秩次相关法及线性回归检验法分析两个流域的水文气象要素变化趋势,定性分析引起流域径流变化的原因。在此基础上,根据碧流河水库流域降水、气温的变化,采用总量控制、随机分配的方法对降水序列进行还原计算,该方法通过降水序列的变化趋势计算历年需还原的降水总量,再以代表站丰、平、枯3个典型年各月平均降水天数为条件,将还原降水随机分配到每个月内；采用总量控制、比例放大的方法还原气温序列,该方法通过还原前、后的气温计算还原系数,以此系数对原系列进行放大计算还原气温序列。还原后的降水、气温序列剔除了趋势性特点,仅具有周期性特点。(3)利用碧流河流域的DEM、土壤、土地利用及气象水文等数据建立了碧流河水库流域的SWAT模型。采用以水量平衡为控制条件的多站点多变量的SWAT模型率定方法,在总量控制与年内分配相结合的优化准则下,利用各分站控制流域的属性,从上游到下游逐级确定模型参数,使碧流河水库流域的SWAT模型能反映流域水循环过程。(4)传统GLUE方法采用纳西效率系数Ens作为单一似然判据,常会产生较大的水量模拟误差,影响模型参数的不确定性分析结果。为此本文提出了由纳西效率系数Ens、相对误差Re及相关系数R2三个指标组成多准则似然判据的改进GLUE方法,并采用多目标模糊优选模型确定多准则GLUE方法的后验似然函数值,进而建立评价SWAT模型参数的不确定性分析方法。经验证,对于90%置信区间,改进的GLUE方法获得的区间宽度比传统GLUE方法更窄,SWAT模型的不确定性更小。通过随机扰动分析了降水数据不确定性对模拟结果的影响,发现降水数据的不确定性较小,与参数不确定性分析结论一致。(5)结合气象要素还原计算及SWAT模型情景模拟,提出了气候变化及人类活动对径流影响的定量成因分析法。该方法首先根据GDP及人口变化,确定人类活动影响变点,通过变点前后径流序列及由SWAT模型模拟出的对应还原气象要素序列的还原径流序列,计算出气候变化与人类活动对径流变化的贡献率。在碧流河水库流域的应用表明,对于1981～2005年,气候(降水、气温)变化和人类活动对碧流河水库来水量变化的影响比例分别为39%与61%.针对碧流河水库流域1980s情景及2000s情景的SWAT模型,采用情景模拟法,模拟出不同时段对应不同土地利用及上游是否修建中小水库共3种情景的径流量。对比分析各情景模拟结果,发现碧流河水库来水量减少的主要原因是直接取用水及中小水库的拦蓄,土地利用变化对来水量的影响较小。(6)将Landsat TM/ETM+遥感数据与传统气象水文数据结合,提出基于遥感数据的小水库拦洪计算方法。该方法根据遥感数据提取的水面信息与有资料小水库的实测库容信息,以子流域为基本计算单元,建立不同坡度的各类子流域的小水库塘坝面积库容关系；以包含洪水时段的两景遥感数据为基础,通过分析两景遥感数据的水面面积及对应时段内降水、蒸发的变化,计算出小水库塘坝的逐日水面面积变化,并将其转化为库容的变化,从而计算出小水库塘坝的拦洪量,此拦洪量可用于校正原洪水预报方案。在丰满Ⅱ区流域的应用表明,全流域水利工程对20060826洪水的拦洪量为4389.6万m3,校正后的洪量相对误差由校正前的31.8%降低到10.1%.(7)提出了考虑单景遥感数据的小水库塘坝蓄泄规律分析方法,并将其应用于丰满Ⅱ区流域的7场洪水。该方法首先利用遥感数据获得小水库塘坝的初始蓄水状态,再选用小水库塘坝初始蓄水状态、洪水所处汛期阶段及降水量3个因子,分析不同因子下的小水库塘坝拦洪规律,分析结果可用于洪水实时预报校正。最后对全文做了总结,并对有待于进一步研究的问题进行了展望。更多还原

【Abstract】 Water resources problem, which has a close relationship to human survival and society development, is the most important issue of resource and environment of the world in the 21st century. The safety of water supply and flood control is the graveness task and important guarantee, which is decided by the basic characteristics of the water resources problem in our country:deficiency of water resources quantity per capita and unbalanced spatial-temporal distributing. In the last few decades, along with the population explosion and development of economy and society, the profound variation of water resources mechanism can be found because of human activities. Decreasing the stream flow, increasing the flood frequency, the problem of water supply and flood control will be more stand out. Many new problem of water resources were brought by these variations, such as, runoff change and the balance between water supply and water demand, flood forecast and control under the human-natural driving model. According to these problems, this paper choose Biliu river reservoir basin as study case, analyze the impact of climate change and human activities on large-size reservoir’s runoff which assignment is provide water to cities; choose Fengman river reservoir basin as study case, analyze the impact of middle- and small-size reservoirs on flood forecasting. The study content and result are as follows:(1) Impact factors of human activities are divided into three classes according to the impact of human activities on runoff and flood. The first class of factor is the impact on condition of runoff yield and concentration, such as land use change, runoff yield and concentration, river regulation, and so on. The second class is directly water use, which are mainly contain water diversion project(such as inter-basin water transfer and irrigation district), groundwater exploitation, and so on; The third class is the impact on runoff process, such as the impact of sluice projects like reservoirs and ponds. Analyze the impact characteristic of human activities on Biliu river reservoir basin and Fengman river reservoir basin, aiming at the three classes of factors above.(2) Qualitative analyze the factors of runoff change, by analyzing the variation trend of hydrometeorological data in study area used Kendall rank correlation method, Spearman rank correlation method and linear regression test. On this basis, according to the variation of precipitation and temperature, the method of total amount control and random distribution is used to calculate the reduced precipitation series. This method, first calculate the total amount of reduced precipitation by its variation trend, and then calculate the average precipitation days of each month of representative station in three typical years(wet, normal and dry), finally, random distribute the reduced precipitation to each month according to the calculated precipitation days of each month. The method of total amount control and proportional amplifying is used to calculate the reduced temperature series. This reduced temperature series can be calculated by original temperature series multiplied by reduction coefficient, which is calculated by comparison the temperature variation between before and after reduction. Only periodicity characteristic can be found in the reduced precipitation and temperature series, the trend characteristic is eliminated.(3) The SWAT model of Biliu river reservoir basin is established used DEM, soil, land use and hydrometeorological data. In this study, the author adopted a multi-site and multi-variable approach based on the water balance control to calibration of SWAT model. Under the optimization criterion of total amount control and annual distribution, according to the attribute of each hydrologic station, calibrate the parameter from upstream to downstream successively. The calibrated SWAT model can reflect the water cycle process in Biliu river reservoir basin.(4) The GLUE method use Nash-Sutcliffe coefficient (Ens) as single likelihood criterion, which may lead large simulation error, and then affect the uncertainty analyze. In this study, the author adopted a multi likelihood criterion, which contain three criterions (Nash-Sutcliffe coefficient (Ens), relative error (Re) and determination coefficient (R2)), to improve the GLUE method. And the posterior likelihood function of the improved GLUE method is calculated by multi-objective fuzzy optimization model. Then, the uncertainty analyze method of SWAT model is established. The application shows that, for 90% confidence interval, the width of an interval for improved GLUE method is more narrow than unimproved method, which indicate the smaller uncertainty of SWAT model. The uncertainty of precipitation is found relative less by the method of precipitation randomization change, which got the same conclusion of parameter uncertainty analyze.(5) The method of genesis analysis is adopted to calculate the impact of climate change and human activities on runoff, through the hydrometeorological data reduction method and SWAT model scenario simulation. Firstly, the author find the change-point of human activities by analyze the GDP and population change. The contribution rate of climate change and human activities are calculated by analyze the runoff series pre-and post the change-point and the reduction runoff series simulated by SWAT model. Applying this method to Biliu river reservoir basin, the result shows that, for 1981-2005, the impact of climate (precipitation and temperature) change and human activities on runoff are take the percentage of 39% and 61% respectively. Aiming at scenario 1980s and scenario 2000s of SWAT model in Biliu river reservoir basin, scenario simulation method is used to simulate the runoff of different scenarios which contains different land use and weather consider mid-and small-size reservoirs. It is easy to found that, the main reason of human activities induced runoff decrease is directly water use and detention of the reservoir in up-stream, the impact of land use change is relative less.(6) Combining remote sensing data (Landsat TM/ETM+) and traditional hydrometeorological data, a method for determining the flood detention quantity of ponds and small reservoirs based on remote sensing data was proposed. This method choose sub-watershed as basic calculation unit, and according to the water surface area extracted by remote sensing data and the measured volume of small reservoirs which has measured data, capacity-area relationship of ponds and small reservoirs is established in different category of sub-watershed divided by average slope. Based on the two scenes remote sensing data that contains the flood period, the water surface area change is calculated by analyze the water surface area of two scenes remote sensing data and the variation of precipitation and potential evaporation in corresponding time period. The flood detention quantities can be calculate by convert the water surface variation to volume variation, through the capacity-area relationship. This value can be used to adjust the primary flood forecasting model. The application in Fengman subarea II basin shows that, the flood detention quantity of the entire basin is 4389.6×104m3. Using this value to adjust the primary flood forecasting model, the relative error decreased from 31.8% to 10.1%.(7) The analyze method of detention and discharge mechanism for ponds and small reservoirs based on single scene remote sensing data is proposed, and applied in 7 floods of Fengman subareaⅡbasin. This method, Firstly, the initial volume of ponds and small reservoirs are calculated in the use of remote sensing data, and then choose the initial volume, phase of flood season and precipitation as three factors, analyze the flood detention mechanism in the different condition. The result can be used for real-time flood forecasting adjustment.Finally, a summary is given and some problems to be further studied are discussed.更多还原