KMS XINJIANG INSTITUTE OF ECOLOGY AND GEOGRAPHY，CAS
| 干旱区气候变化对新疆典型流域水资源的影响研究 | |
| Alternative Title | Study on the climate change in arid regions and its impact on water resources in typical river basins of Xinjiang |
| 罗敏 | |
| Subtype | 博士 |
| Thesis Advisor | 刘铁 ; Philippe De Maeyer ; Amaury Frankl |
| 2019-06-30 | |
| Degree Grantor | 中国科学院大学 |
| Place of Conferral | 北京 |
| Degree Discipline | 理学博士 |
| Keyword | 气候变化 GCM 降尺度 水资源 SWAT Climate change GCM Downscaling Water resources SWAT |
| Abstract | 近几十年来， 在气候变暖背景下， 全球各地温度以及降水模式正发生着快速改变。干旱、半干旱的中亚地区作为“一带一路”建设的重要支点地区， 占全球干旱、半干旱地区总面积的三分之一， 其短缺的水资源是维持该地区脆弱生态环境和制约经济发展的主要因子之一。在全球气候变化背景下，中亚地区水资源安全和生态系统稳定对变化环境的适应能力十分有限。因此，评估气候变化对该地区水资源可能产生的影响十分重要。本研究以中亚地区（包含中国新疆）为研究区，并选取研究区内九个典型流域作为响应区，这九个流域包含了小尺度、中尺度以及大尺度流域，通过基于 GCM（General Circulation Model） 模型降尺度、利用遥感数据进行水文模型的搭建和率定、利用多种偏差校正方法对 RCM（Regional Climate Model） 模型进行误差校正，分析了中亚气候变化的时空分布特征，并进一步开展气候模式与水文模型耦合等方面的工作，研究了气候变化对九个典型流域水资源的影响。 得出的主要结论包括：（1）以 CRU（Climate Research Unit dataset） 格网数据作为参考数据，利用三种降尺度方法，对 18 个 GCM 集进行统计降尺度，研究了中亚降水和温度变化的时空分布特征。研究结果表明，改进的 Delta 降尺度方法和 BMA 降尺度方法对 GCM 输出的降水进行降尺度时，精度明显高于简单的 Delta 降尺度方法；对 GCM 输出的温度进行降尺度时，三种方法的降尺度结果差异并不明显。整体而言，中亚地区降水呈现缓慢的增加趋势，在 RCP（Representative ConcentrationPathway） 4.5 排放情景下， 2021-2060 年相对于 1965-2004 年的平均增长速度为4.63 mm/10a；中亚地区温度呈现明显的上升趋势， 2021-2060 年相对于 1965-2004年的平均升温速度为 0.37 °C/10a。同一时期在 RCP 8.5 排放情景下，暖湿趋势更加明显。降水的变化具有强烈的时空异质性，在中亚的中部和南部部分区域，夏季降水甚至出现了明显的下降趋势。气候变暖在中亚北部地区明显高于南部地区；其中夏季升温速率较其他三个季节最为明显。与此同时，气温升高速率具有强烈的高程和纬度效应，随着海拔（仅在山区）和纬度的增加，升温的速率明显增加。相关性分析结果表明，中亚气候变化与大尺度环流场，如海平面气压、表面温度、径向风、纬向风、垂直风速、位势高度、海温等的异常有较明显的相关关系。（2）采用比例系数法对 TRMM 降水数据进行了误差校正，并将校正后的结果进一步用于 SWAT 和 MIKE SHE 水文模型的驱动。与此同时，本论文利用MODIS 10A2 积雪面积数据对两个水文模型模拟雪盖空间分布的合理性进行了验证。结果表明，利用校正后的 TRMM 降水数据驱动 MIKE SHE 和 SWAT 水文模型，在和田河流域进行水文过程模拟的结果较好，日尺度相关系数 R2 以及纳什效率系数 NSE 均高于 0.6。通过比较 MODIS 10A2 积雪数据与水文模型模拟的雪盖数据，发现两种模型均能较好地反映积雪的时空分布，进一步证明了利用校正后的 TRMM 数据进行水文模拟， 其结果具有合理的时空特征。总体而言，校正后的 TRMM 数据可以代替站点观测数据驱动水文模型，能够在一定程度上解决无资料以及缺资料地区缺少必须的观测气象数据而无法进行水文建模的难题。遥感产品在水文中具有很好的应用前景，遥感产品经过适当的预处理不仅可作为水文建模的输入数据，又可用于水文模型模拟结果精度的验证。（3） 对七种降水和五种温度误差校正方法在典型高寒流域-开都河流域的气候预测和水文模拟中的表现进行了分析、 比较。结果表明，原始的 RCM 输出存在较大误差，且无法通过率定水文模型消除相应的误差。与温度校正方法相比，不同降水校正方法的表现差异较大。基于分布的 DBC 和 EQM 校正方法对降水的校正结果最好，就温度而言，除 LS 方法外，其他四种温度校正方法均得到了较好的结果。在开都河流域利用校正后的降水和温度数据进行水文模拟，结果表明， 降水的频率校正在水文模拟中十分重要。 DT 和 LS 降水校正方法由于没有对降水频率进行合理的校正，在水文模拟中的效果非常差。 EQM 方法校正降水与 VARI 方法校正温度的组合在水文模拟中表现最好。（4）利用 SWAT 模型对新疆地区九个典型高寒流域气候变化对水资源的影响进行了评价。结果表明： 总体上， 相对于 1965-2004 年， 2021-2060 年九个流域的年径流量呈现上升趋势；在 RCP 4.5 排放情景下，年径流量变化介于0.37-40.04%之间，在 RCP 8.5 排放情景下年径流量增加更加明显，增量介于0.38-50.09%之间。蒸散发量和积雪融水量也呈现出上升趋势。但各水文要素的变化具有强烈的时空不均匀特征。不同流域的面积、高程和坡向等要素均会影响气候变化影响的强度。降水增加和温度升高均有利于山区流域水资源总量的增加。但相对于降水变化而言，径流量的变化对温度变化更加敏感； 相对于南疆地区，北疆地区的流域对降水变化的敏感性更加明显。每个流域均存在一个临界温度，一旦温度升高超过该临界温度，升温对该流域水资源总量的影响就会由正效应转化为负效应；在海拔相对较低的流域，临界温度相对较小。气候变化同时还会导致各水文要素比例的调整，主要表现为：降雪率下降，降雨对径流的补给比例增加，蒸发比增加。本文对中亚区域九个典型高寒流域气候变化对水资源的影响进行了评价，主要的创新点包括以下几个方面：首先，改进了 Delta 降尺度方法，对中亚地区未来降水和温度的时空变化特征进行了预测，该方法能够更好地结合多个 GCM 输出结果，对单个 GCM 不合理的结果逐栅格进行剔除。其次， 成功的将遥感数据应用于缺资料的和田河流域进行多个水文模型的搭建和模拟结果验证。 第三，分析了河流的位置、高程、坡向等因素对气候变化影响程度的可能影响。 |
| Other Abstract | In recent decades, global warming has rapidly changed the surface temperatureand precipitation patterns around the world. The arid and semi-arid Central Asia,accounting for one third of the global arid and semi-arid region, has an ecologicallyand economically vulnerable environment and suffers from water shortage, showing alimited resilience to change; therefore, this creates a need for evaluating the possibleclimate change impact on water resources in Central Asian domain. Nine typical riverbasins, representing the small-scale, mesoscale and large-scale basins, were selectedas study area in this work. Based on the downscaling of the general circulation model,the application of remotely sensed data in hydrological modelling, a comparison ofthe bias correction methods and a combination of the RCM and hydrological model,the climate change impact on water resources in the nine selected river basins wereinvestigated. The main conclusions of this dissertation included the following:(1) We assessed the variations in precipitation and temperature in Central Asiabased on a combination of CRU gridded data and eighteen GCMs using threedownscaling methods. The results showed that the Advanced Delta and BMA methodswere distinctly superior to the Delta method when applied to precipitation, whileminor differences existed among the three methods in downscaling temperature. Theprecipitation exhibits an overall increasing trend at an average rate of 4.63 mmdecade-1 in Central Asia during 2021-2060 relative to 1965-2004, while thetemperature increases at a rate of 0.37 °C decade-1 under RCP 4.5 scenario. A strongerwarming and wetting trend is observed under RCP 8.5 scenario than that under RCP4.5 scenario during the same period. The increases in precipitation are expected tohave strong spatiotemporal differences with decreasing trend (even occurring insummer precipitation for the central and southern parts of Central Asia). Higherprojected warming rate appears in the northern part of Central Asia, while thewarming was most conspicuous in summer among all seasons. Meanwhile, themagnitudes of temperature increases were discovered to rise significantly withelevation (only for the mountainous areas) as well as latitude. The changes in localclimate were driven by the anomalies in large-scale circulation fields, such as airpressure at sea level, surface temperature, zonal wind, meridional wind, vertical windvelocity, geopotential height and sea-surface temperature.(2) We applied the proportional coefficient method to correct the TRMM dataand these (data) were further utilized to drive both the SWAT and MIKE SHE models.The remotely sensed snow coverage data were used to validate the simulated results.The results showed that both the MIKE SHE and SWAT models demonstrated anacceptable performance in the Hotan River Basin (when driven by the correctedTRMM rainfall data) and both two evaluation indices R2 and NSE are higher than 0.6on a daily scale. The similarity between the MODIS snow cover and modeled resultsdemonstrated that both models could (re)present the spatial and temporal distributionof the snow cover adequately. The corrected TRMM data can be used to replace thestation-based data, which partly overcome the lack of necessary observedmeteorological data in the ungauged river basins. The remotely sensed products are tobe used in both driving and validating the hydrological models (after appropriatepre-treatments).(3) We compared seven precipitation and five temperature correction methods inclimate projecting and hydrological modelling in the typical alpine Kaidu River Basin.The results showed that the original RCM output data are very biased and thesecannot be corrected by hydrological calibration. The performance of variousprecipitation-corrected methods presented larger differences when compared withvarious temperature-corrected methods. The distribution-based DBC and EQMmethods performed best in the precipitation reproduction while alltemperature-corrected methods (with exception of the LS method), performedextremely well. The correction in wet-day frequency is extremely important for thehydrological projection of the Kaidu River Basin. The DT and LS methods, whichlack a wet-day frequency correction, are not suitable to be applied in dischargesimulations. The combination of the EQM-corrected precipitation and VARI-correctedtemperature performed best in hydrological modeling. The results set out in thisdissertation are of a wider significance and the procedures can be applied to otherregions accordingly.(4) We assessed the impact of climate change on water resources using a SWATmodel in nine typical alpine river basins in Xinjiang. The results showed an overallincreasing rate occurring in the annual total discharge with a range of 0.37-40.04%under RCP 4.5 and 0.38-50.09% under RCP 8.5 across the nine selected river basinsduring 2021-2060 relative to 1965-2004, with a strong spatiotemporal heterogeneity.The ET and snow melting will also present an overall increasing trend in the nearfuture. The area, elevation and slope aspect of the rivers are observed to affect theintensity of the climate change impact. Both δP and △T have a positive influence onthe total discharge, while △T enjoys a higher sensitivity compared with δP. NorthernXinjiang is higher sensitive to δP than southern Xinjiang. Each basin has a criticaltemperature for the sensitivity transformation of warming linked with the shrinkage ofsolid water storage. The critical temperature is smaller in rivers with a relative lowerelevation. The climate change also results in the proportion adjustment of eachhydrologic component. In general, the dependence of the total discharge on rainfallwill increase according to the augmented R/D. The S/P shows a decreasing trend,while ET/P shows a rising trend.This dissertation assesses the impact of climate change on water resources in thenine typical alpine river basins in Central Asia (and some innovations were broughtout to overcome the main challenges). Firstly, the future precipitation and temperaturechange was projected by using the Advanced Delta method to effectively combine theGCM ensemble. Secondly, the remotely sensed products were successfully used todrive and validate the hydrological models in the ungauged Hotan Catchment. Finally,the possible effects of location, elevation, area and slope aspect of rivers on the extentof climate change impact have been investigated. |
| Subject Area | 地图学与地理信息系统 |
| Language | 中文 |
| Document Type | 学位论文 |
| Identifier | http://ir.xjlas.org/handle/365004/15346 |
| Collection | 中国科学院新疆生态与地理研究所 研究系统 |
| Affiliation | 中国科学院新疆生态与地理研究所 |
| First Author Affilication | 中国科学院新疆生态与地理研究所 |
| Recommended Citation GB/T 7714 | 罗敏. 干旱区气候变化对新疆典型流域水资源的影响研究[D]. 北京. 中国科学院大学,2019. |
| Files in This Item: | There are no files associated with this item. | |||||
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment