KMS XINJIANG INSTITUTE OF ECOLOGY AND GEOGRAPHY，CAS
| 基于游程理论和 Copula 函数的中亚干旱风险评估 | |
| Alternative Title | Drought Risk Assessment in Central Asia Using Run Theory and a Probabilistic Copula Function Approach |
| 张乐园 | |
| Subtype | 硕士 |
| Thesis Advisor | 王弋 ; 陈亚宁 |
| 2020-06-30 | |
| Degree Grantor | 中国科学院大学 |
| Place of Conferral | 北京 |
| Degree Discipline | 理学硕士 |
| Keyword | 干旱风险 游程理论 Copula 重现期 中亚 Drought Risk Run Theory Copula Return Period Central Asia |
| Abstract | 中亚地处干旱半干旱地区，降水较少而蒸散发较大，干旱已经成为限制五个中亚国家发展和区域稳定的重要因素。 准确地描述中亚干旱的时空变化特征，量化研究干旱，评估干旱的风险和重现期，对保证当地人民正常的生产实践活动具有十分重要的作用。 本研究基于 CRU 数据集计算了中亚五国标准化降水蒸散指数（SPEI） ， 研究了 1961-2017 年间中亚五国干旱的时空变化特征， 全面分析了中亚地区SPEI的年际、年内变化变化特征，基于SPEI和游程理论定义了干旱历时、干旱强度和干旱峰值，通过 Copula 函数构造联合分布模型，计算联合概率密度和联合重现期，对1961-2017年中亚五国的干旱风险进行了具体地统计与分析， 旨在为维护中亚地区地稳定发展，合理有效地进行水资源的管理和分配，减轻干旱对于当地的不利影响提供一些参考。论文主要分析结果如下：1. 中亚五国 SPEI 时空变化趋势： 年际变化上， 1961-2017 年间中亚的干旱年际波动较大； 1961-2008 年间中亚地区的气候呈持续变干趋势。 1975、 1995 和2008 年中亚分别经历了中度干旱，轻度干旱和中度干旱； 1994 年中亚干旱发生了突变； 年内变化上，春季和冬季干旱变化缓慢，夏秋两季干旱加重明显；四季突变发生情况具有异质性。空间上， 中亚不同季节的干旱分布空间差异性较大，总体上中西部干旱化趋势较明显。春季中部及西南部地区干旱显著；夏秋季中亚大部分地区干旱有加重趋势，其中西部干旱更明显；秋季吉尔吉斯斯坦、乌兹别克斯坦的临界处变干趋势显著；冬季以哈萨克斯坦南部、 土库曼斯坦以及乌兹别克斯坦干旱化趋势显著。2. 中亚五国气候变量与 SPEI 变化趋势： 1961-2017 年中亚地区的降水除在哈萨克斯坦中部、乌兹别克斯坦中部和土库曼斯坦小部分地区呈减少趋势外，在其他大多数地区呈现缓慢增加趋势；温度在整个中亚地区呈现持续上升态势，潜在蒸散发增长速率以咸海流域为中心，呈环形向四周递减； SPEI 值的变化说明，土库曼斯坦、乌兹别克斯坦中部、哈萨克斯坦中南部等地的干旱趋势加重，其他区域则有所缓解。 通过计算 1、 3、 6、 9、 12 个月尺度的 SPEI，发现在 1961-1974年和 1979-1995 年间整个中亚地区 SPEI 值波动不大，反映气候较稳定；而1974-1979 年和 1995-2017 年间中亚地区 SPEI 值相对较低，表明地区经历了较长时间的干旱期。 并且随着时间尺度增加，干旱事件的发生频次和强度开始变低，但其干旱历时逐渐延长。3. 中亚五国干旱变量特征和 Copula 函数的选择：不同的干旱变量从不同的角度说明了中亚干旱的特征。中亚的干旱强度西部大于东部；干旱历时与干旱强度具有显著的空间对照性，即干旱强度高（低）的区域，干旱历时也高（低）；干旱峰值反映除去轻度干旱，中亚干旱主要以重度干旱为主，中度干旱较少，极端干旱在哈萨克斯坦中部有发生。干旱历时的值大致以环形分布从外围向中心递减。通过 Pearson product-moment， Pearman’s rank-order 和 Kendall rank 相关系数衡量干旱变量之间的相关性，发现中亚五国的干旱强度（ Ds）、干旱历时（Dd）与干旱峰值（Dp）三组变量之间显著相关。 本研究采用 MVCAT 软件依次对变量进行拟合，其中 Generalized pareto 和 Generalized extreme value 两种函数出现频次最高。通过计算 RMSE、NSE 和 AIC 值，发现五国的二维 Copula 函数以 Frank和 Gumbel 为主； 通过反 Kendall 参数法计算参数， 发现五国均适宜以三维 Gumble对称性函数构建相关性结构。4. 中亚五国干旱等级情况和干旱风险分析：不同等级的干旱在中亚的发生频次不同。一方面在干旱频次发生的基数上，轻旱>中旱>重旱>极端干旱；另一方面同一等级的干旱类型在不同国家的发生频次也存在较大的差异性，其中极端干旱发生的频次空间差异性最大。 时间尺度上，三维 Copula 函数反映的干旱危险性在 1961-2017 年经历了中等强度到低强度再到高强度的变化过程，其中2000-2017 年间中亚五国干旱风险较高。空间尺度上，哈萨克斯坦的干旱风险波动较大，最高值出现在 1990-1999 年间，最低值出现在 1980-1989 年间；吉尔吉斯斯坦和塔吉克斯坦两国的干旱风险变化相对稳定在 0.254-0.5 之间。 由三维干旱强度、干旱历时和干旱峰值定义的1961-2017年中亚五国所有类型干旱事件中，重现期为 2 年的约占 80％， 2-10 年的约占 15％， 10 年以上的约占 5％。 |
| Other Abstract | Central Asia is located in arid and semi-arid areas, with less precipitation andgreater evapotranspiration. Accurately describing the temporal and spatial variationcharacteristics of drought in Central Asia, quantitatively studying drought, andevaluating the risk and return period of drought have a very important role in ensuringthe normal production practices of local people. This study calculated thestandardized precipitation evapotranspiration index (SPEI) of the five Central Asiancountries based on the CRU dataset. The characteristics of drought's spatio-temporalchanges in the five Central Asian countries from 1961 to 2017 were studied, and theinterannual and intra-annual changes of SPEI in Central Asia were comprehensivelyanalyzed. Based on SPEI and Run theory, drought duration, drought intensity anddrought peak were defined. The Copula function was used to construct a jointdistribution model, and the joint probability density and joint return period werecalculated. The drought risks of the five Central Asian countries from 1961 to 2017were specifically analyzed. This study aims to provide some reference for maintainingthe stable development of Central Asia, rationally and effectively managing andallocating water resources, and reducing the adverse impact of drought on the localarea.The main results of the research are as follows:1. The spatial and temporal trends of SPEI in the five Central Asian countries: Interms of interannual changes, the drought in Central Asia fluctuated greatly from 1961to 2017, and the climate in Central Asia continued to dry out from 1961 to 2008.Central Asia experienced moderate drought, mild drought, and moderate drought in1975, 1995, and 2008; the drought in Central Asia was abrupt in 1994. In terms ofannual changes, the drought in spring and winter changes slowly, drought in summerand autumn increases significantly, and the occurrence of mutations in four seasons isheterogeneous. In terms of spatial changes, the spatial distribution of drought inCentral Asia varies greatly in different seasons, and the overall trend of drought in thecentral and western regions is more obvious. The drought in the central andsouthwestern regions in spring is significant, and the drought in most parts of CentralAsia in summer and autumn tends to increase, with the drought in the western regionbeing more pronounced. In the autumn, the critical part of Kyrgyzstan and Uzbekistan tends to dry out significantly. In winter, the trend of drought in southern Kazakhstan,Turkmenistan, and Uzbekistan is significant.2. The change trend of climate variables and SPEI in the five Central Asiancountries: The precipitation in Central Asia from 1961 to 2017 showed a slowlyincreasing trend in most regions, but decreased in central Kazakhstan, centralUzbekistan and a small part of Turkmenistan. The temperature in the entire CentralAsia region continues to rise, and the growth rate of potential evapotranspiration iscentered on the Aral Sea Basin and decreases in a circle to the surroundings. Thechanges in SPEI values indicate that drought has increased in Turkmenistan, centralUzbekistan, central and southern Kazakhstan, and other regions, while it hasdecreased in other regions. By calculating the SPEI on the 1, 3, 6, 9, 12 month scale, itwas found that the SPEI value of the entire Central Asia region did not fluctuategreatly between 1961-1974 and 1979-1995, reflecting a relatively stable climate。However, the SPEI value of Central Asia was relatively low between 1974-1979 and1995-2017, indicating that the region has experienced a longer period of drought. Andas the time scale increases, the frequency and intensity of drought events begin todecrease, but the duration of droughts gradually increases.3. The characteristics of drought variables and the choice of Copula function inthe five countries of Central Asia: different drought variables illustrate thecharacteristics of drought in Central Asia from different angles. The drought intensityin Central Asia is greater in the west than in the east. The drought duration anddrought intensity have a significant spatial contrast, which is manifested in areaswhere the drought intensity is high (low), and the drought duration is also high (low).The peak drought reflects the fact that regardless of the mild drought, the drought inCentral Asia is mainly dominated by severe drought, moderate drought is less, andextreme drought occurs in central Kazakhstan. The value of the drought durationdecreases roughly from the periphery to the center in a circular distribution. In thisstudy, the correlation between drought variables was measured by Pearsonproduct-moment, Pearman's rank-order and Kendall rank correlation coefficients, andit was found that the drought intensity (Ds), drought duration (Dd) and drought peak(Dp) in the five Central Asian countries Group variables are significantly correlated.In this study, the MVCAT software was used to fit the variables in turn. The resultsshowed that the generalized pareto and generalized extreme value functions had the highest frequency. By calculating the RMSE, NSE and AIC indicators, it is found thatthe two-dimensional Copula functions of the five countries are dominated by Frankand Gumbel. Based on the anti-Kendall parameter method to calculate the parameters,it is found that the five countries are suitable to construct the correlation structure withthe three-dimensional Gumble symmetry function.4. Drought level situation and drought risk analysis in the five Central Asiancountries: different levels of drought occur in Central Asia at different frequencies. Onthe one hand, on the basis of the frequency of droughts, light drought> mediumdrought> heavy drought> extreme drought; on the other hand, the frequency ofdrought types of the same level in different countries also has great differences,among which the frequency of extreme drought The spatial difference is the largest.On the time scale, the drought risk reflected by the three-dimensional Copula functionexperienced a medium-intensity to low-intensity to high-intensity change from 1961to 2017, and the risk of drought in the five Central Asian countries was high from2000 to 2017. On the spatial scale, the drought risk of Kazakhstan fluctuated greatly,with the highest value appearing in 1990-1999 and the lowest value appearing in1980-1989. The drought risk changes in Kyrgyzstan and Tajikistan are relativelystable between 0.254-0.5. The return period of all types of drought in the five CentralAsian countries from 1961 to 2017 defined by the three-dimensional drought intensity,drought duration and drought peak: about 80% in 2 years, about 15% in 2-10 years,more than 10 years About 5%. |
| Subject Area | 自然地理学 |
| Language | 中文 |
| Document Type | 学位论文 |
| Identifier | http://ir.xjlas.org/handle/365004/15472 |
| 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 | 张乐园. 基于游程理论和 Copula 函数的中亚干旱风险评估[D]. 北京. 中国科学院大学,2020. |
| 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