KMS XINJIANG INSTITUTE OF ECOLOGY AND GEOGRAPHY，CAS
| 叶尔羌河灌区地下水埋深变化及影响因素分析 | |
| Alternative Title | Spatio-temporal variation of groundwater depth and its influencing factors in the irrigated area of the Yarkant River |
| 白宜斐 | |
| Subtype | 硕士 |
| Thesis Advisor | 王弋 ; 陈亚宁 |
| 2019-06-30 | |
| Degree Grantor | 中国科学院大学 |
| Place of Conferral | 北京 |
| Degree Discipline | 理学硕士 |
| Keyword | 地下水埋深 敏感性分析 相对贡献率 Copula 函数 叶尔羌河灌区 Groundwater depth Sensitivity analysis Relative contribution rate Copula function Irrigation area of Yarkant River basin |
| Abstract | 中国地表水资源相对匮乏，地下水是农业活动、工业生产和居民生活用水的重要来源，尤其是在干旱、半干旱地区，地下水资源的优劣与当地百姓生产生活以及城市的发展休戚相关。 叶尔羌河灌区在过去的几十年里人口增加、 灌溉耕地面积逐年扩大， 导致地表水资源量紧张， 加大了对地下水资源的开采力度，地下水埋深的下降速度加快。地下水埋深下降导致该地区产生植被退化、沙漠化和荒漠化等一系列的生态环境问题。因此分析叶尔羌河灌区地下水埋深与其驱动因子之间的关系，探明地下水位大幅下降的原因，找到导致地下水埋深变化的主要驱动因子， 可以为流域地下水资源的可持续利用和管理提供科技支撑。本研究利用近 10年叶尔羌河灌区内 35眼地下水观测井资料分析叶尔羌河灌区地下水埋深的时空变化特征，利用敏感性法和相对贡献率方法，从人为要素和自然要素两个方面来分析不同驱动因子对叶尔羌河灌区地下水埋深变化的敏感程度并确定影响地下水埋深变化的主导驱动因子。再对驱动因子和地下水埋深数据建立各自的边缘分布函数，通过 2D Copula 函数分析驱动地下水埋深变化的主导因子与地下水埋深之间的相关性，结合条件概率定量分析驱动因子与地下水埋深分布的影响。最后利用 SVM、 ANN 和 ANFIS 模型分别建立基于驱动因子时间序列的地下水埋深模拟模型， 并通过比较选出适合该研究区的地下水埋深模拟模型。研究结论主要如下：（1） 叶尔羌河流域灌区的地下水埋深在沿河道方向上呈现出上游浅（莎车子灌区）、下游深（巴楚子灌区）的趋势；地下水变化最大为巴楚子灌区，莎车子灌区和麦盖提子灌区的地下水位波动较小，农业生产季节对地下水埋深的影响从大到小依次为：巴楚＞麦盖提＞莎车。地下水开采量是造成巴楚和麦盖提地下水埋深加深的主要影响因子。莎车的地下水埋深变化受地下水开采量和地表引水量共同的影响。（2） 研究表明 Copula 函数为地下水埋深分析提供了有效的研究方法。通过选择合适的 Copula 函数，能够更加准确的定量分析地下水埋深变化对各驱动因子的响应。 以巴楚子灌区为例， Frank Copula 函数能够更加精确的表述四个驱动因子与地下水埋深之间的关系。 地下水开采量与地下水埋深呈负相关，渠首径流量、地表引水量和降水与地下水埋深呈正相关，而且地下水开采量对地下水埋深的影响具有一个月的滞后性。同时利用条件概率计算驱动因子对地下水埋深变化的发生概率分析发现，地下水埋深落入较深范围的概率随着地下水开采量的增加而增加，随着地表引水量、径流和降水的增加而减小，并且对分布概率的影响为：地下水开采量>渠首径流>地表引水量>降水。（3） 经过模型检验， 三个模型中 ANFIS 对叶尔羌河灌区地下水埋深验证结果最好， LIBSVM 次之， ANN 的结果最差，因此， ANFIS 模型是三个模型中最合适用来模拟该地区地下水埋深变化的模型。 |
| Other Abstract | Surface water resources are relatively scarce and groundwater play an importantrole in agricultural, industrial and domestic water usage in China, especially in aridand semi-arid regions. Therefore, groundwater is highly related to the agricultureproduction, livelihood of local people and the development of the city. the YarkantRiver Irrigation Area, the largest irrigation area in Xingjiang, had been increasing inpopulation and expanding of irrigated area in the past few decades. The lack ofsurface water resources resulted in the increase of groundwater exploitation and theaccelerated decline in groundwater depth. The decline in groundwater depth has led toa series of ecological and environmental problems including vegetation degradation,desertification and so on. Therefore, the relationship between groundwater depth andits driving factors in the Yarkant River irrigation area is analyzed in this study,identifying the reasons of the significant drop in groundwater level and finding out thedominant driving factors leading to the change of groundwater depth could providescientific support for the sustainable utilization and management of groundwaterresources in the basin.In this paper, the spatial and temporal variation characteristics of groundwaterdepth in the Yarkant River irrigation area are analyzed by using the data of 35groundwater observation wells in the past 10 years. The sensitivity method andrelative contribution rate method are used to analyze the driving factors from the twoviews of human and nature. And the sensitivity of groundwater depth changes and thedominant driving factors were determined. Then, the marginal distribution function ofthe driving factor and groundwater depth data is established. The joint distributionfunction between each driving factor and groundwater depth is calculated by 2DCopula functions to analyze which is the dominant factor driving the change ofgroundwater depth. The joint distribution function combined with conditionalprobability were quantitatively analyzed the effects between driving factors and groundwater depth. Finally, the SVM, ANN and ANFIS models were used to establishthe groundwater depth prediction model based on the driving factor time series, andthe most suitable groundwater depth prediction model can be selected. Theconclusions of the study are as follows:(1) The groundwater depth shows the trend along the river channel, which wasshallow in upstream (Shache Sub-irrigation) and deep in downstream (BachuSubirrigation); the obvious variation of groundwater depth is in Bachu Sub-irrigationArea, whereas less variable in Shache Sub-irrigation and the Maigaiti sub-irrigation.Additionally, the groundwater depth of Bachu and Macgaiti is significantly affectedby the agricultural production season, and Shache is relatively weak. Groundwaterextraction is the main factor of groundwater depth in Bachu and Megaiti. The changein groundwater depth of Shache is affected by both groundwater extraction andsurface water diversion.(2) The copula function provides effective research methods for groundwaterdepth analysis. The response of the various driving factors of the groundwater wasanalyzed more accurately by selecting the appropriate copula function. Taking BachuSub-irrigation Area as an example, the frank copula function can be more accurate toshow the relationship between the four drivers and the groundwater. Groundwaterexploitation is negatively correlated as well as runoff, surface water diversion andprecipitation are positively correlated with groundwater depth, and the influence ofgroundwater exploitation on groundwater depth has a one-month lag. The probabilityof groundwater embedding deep into different ranges were found, the likelihood ofgroundwater drop into a deeper scope increases, and with the increase of surfacewater, runoff and precipitation, the probability of groundwater drop into the lowerrange is large, and the influence of the probability of distribution rank is: thegroundwater extraction quantity of the surface water of >runoff >surface waterdiversion>precipitation.(3) The ANN, LIBSVM and ANFIS models were used to simulate the groundwater depth changes in the Yarkant River irrigation area. The model testshowed that ANFIS is the best groundwater depth verification results, followed byLIBSVM, and the ANN results were the worst. Thus, the ANFIS model is the mostsuitable model for predicting changes in groundwater depth in the region. |
| Subject Area | 自然地理学 |
| Language | 中文 |
| Document Type | 学位论文 |
| Identifier | http://ir.xjlas.org/handle/365004/15308 |
| 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 | 白宜斐. 叶尔羌河灌区地下水埋深变化及影响因素分析[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