Knowledge Resource Center for Ecological Environment in Arid Area
| 腾格里沙漠南缘风积物光释光年代学 | |
| 其他题名 | Optically Stimulated Luminescence Chronology of Aeolian Deposits from the South Margin of the Tengger Desert |
| 彭俊 | |
| 出版年 | 2017 |
| 学位类型 | 博士 |
| 导师 | 董治宝 |
| 学位授予单位 | 中国科学院大学 |
| 中文摘要 | 光释光（OSL）测年法是测定第四纪沉积物年代的一种常用方法，被广泛用于河流、湖泊、冰川、沙漠沉积物定年。近50年来，释光测年技术在测年方法、技术手段、实验仪器上均得到了迅速发展。中国是世界上沙漠与沙地分布较多的国家，沙漠化与沙尘暴等自然灾害是制约中国西北干旱区经济与生态环境发展的重要因素。研究沙漠与沙地的演化规律及其与气候环境之间的关系是开发利用干旱区资源、控制自然灾害与保护生态环境的重要手段。本文对腾格里沙漠南缘的风积物剖面进行了系统的地层采样，并利用OSL测年技术分析了这些风积物地层年代数据。使用实测与模拟数据评价了贝叶斯信息准则（BIC）在OSL信号衰减曲线组分数估计中的可行性，使用BIC法对腾格里沙漠南缘风积物样品OSL信号组分结构进行了分析。尝试了一类马尔科夫链蒙特卡罗采样算法（切片采样）在释光年龄模型参数优化中的应用，使用模拟与实测De数据对切片算法可靠性进行了理论验证，并根据腾格里沙漠南缘风积物样品实测De分布对该算法参数估计进行了检验。开发了石英释光动力学模型模拟的开源R程序KMS，使用该程序对释光测年中常用的SGC与gSGC法的可靠性进行了模拟评价，并对两种方法在腾格沙漠南缘风积物中的适用性进行了比较分析。结合本文OSL年代与前人发表的地层年代数据，重建了腾格里沙漠南缘20 ka以来的风沙活动历史。本文得出如下主要结论：（1）基于不同情景的模拟结果表明，BIC法信号组分数估计的可靠性与测量时间与信号通道数、不同信号组分衰减常数的差异、及信号组分的信噪比有关。BIC法在估计中、高信噪比OSL信号衰减曲线组分数中有巨大潜力。腾格里沙漠南缘风积物样品的释光信号衰减曲线一般可分解为2–3个信号组分，沙质黄土比风成沙样品包含更多的信号组分。对2333条曲线分析表明，快、中信号组分的光电离截面分别为2.57×10-17 cm2、1.35×10-18 cm2，占释光信号衰减曲线0–0.3 s信号总量的比重分别为88.16%、9.84%。（2）切片采样算法能根据测量数据与模型的联合似然函数进行随机采样获得参数的样本分布。基于模拟与实测De数据的分析表明，切片采样算法参数估计与极大似然估计法是一致的，该算法具有对参数初值依赖性低、误差（置信区间）估计更准确的特点，提供了释光年龄模型参数估计的一种新方法。对腾格里沙漠南缘风积物样品的切片采样算法分析表明，这些样品De分布集中，晒退性较好。（3）模拟结果表明，在额外剂量Dr接近给定剂量Dn的情况下，gSGC法De估计的精度要显著高于SGC法。在Dn<120 Gy范围内，SGC与gSGC法De估计的精度与准确度差别均小于4%。尽管两种方法估计的De准确度在Dn<210 Gy范围内不显著，对于更高剂量，gSGC法De估计具有更高的准确度。对于腾格里沙漠南缘风积物样品分析表明，两种方法估计的De与单片再生剂量法在De<50 Gy范围内具有很好的一致性，对于De<50 Gy的风积物样品，SGC与gSGC法De估计的精度与准确度差别均不显著。（4）结合本文OSL年代地层数据及前人发表的数据综合分析表明，腾格里沙漠南缘沙丘活化阶段为20–10 ka，固定阶段为10–6 ka，3 ka之后再次活化。沙丘活动历史与东亚季风关系密切，冬季风强盛时沙丘发生大规模移动，夏季风强盛时沙丘被大范围固定。3 ka之后的风沙活动可能与夏季风强度减弱及人类活动有关。腾格里沙漠南缘20 ka 以来的风沙活动历史与中国北方其它沙漠演化历史大体一致，都表现出末次盛冰期至全新世早期沙丘活化、全新世早至中期沙丘固定、晚全新世沙沙丘再次活化的特征。（5）北半球夏季太阳辐射量与东亚夏季风均在12 ka之后逐渐增强，而腾格里沙漠南缘沙丘在10 ka左右由流动逐渐向固定过渡，两者存在明显滞后。北半球夏季太阳辐射强度峰值期间，青藏高原东北部的下沉气流及高温导致的蒸发作用造成12–10 ka出现大范围风成沙堆积。腾格里沙漠南缘沙丘活动主要受地表有效湿度控制，而地表有效湿度变化又与季风降水量变化及局地对流与蒸发密切相关。沙丘由流动向固定阶段的转化与有效湿度阀值密切相关，当有效湿度水平达到某特定值时，源区风积物被大范围固定，从而降低了沙丘的流动性。 |
| 英文摘要 | Optically stimulated luminescence (OSL) dating techniques is a routinely used method for determining the burial ages of Quaternary sediments，and has been widely used for dating sediments from fluvials, lacustrines, glaciers, and deserts. The past fifty years has seen the fast development of luminescence dating techniques in dating methods, laboratory protocols, and dating instruments.Chinahas large area of deserts and sand fields. Desertification and sand storm have been major factors that impede the development of economic and improvement of ecological environment in arid region of northwesternChina. Studying desert and sand filed evolution as well as their intrinsic relationship with climate change provides scientific basis for exploiting and utilizing resources in arid regions, regulating natural calamities, and protecting ecological environment.This study conducted systematic multiple-site sampling and OSL dating along the south margin of the Tengger Desert. The applicability of the Bayesian Information Criterion (BIC) for estimating the numbers of components in OSL decay curves was investigated using simulated and measured decay curves. The BIC method was applied to study structure characteristics of OSL decay curves of aeolian deposits from the south margin of the Tengger Desert. A Markov chain Monte Carlo method (Slice sampling) was used to optimize parameters of statistical age models. The validate of Slice sampling was tested using simulated and measured equivalent dose (De) distributions and further checked using measured data sets of aeolian deposits from the south margin of the Tengger Desert. An open source R program KMS was developed to simulated quartz luminescence using comprehensive kinetic models. The program was used to assess the reliability of Standardised growth curve (SGC) and global-SGC (gSGC) methods for De estimation. The applicability of SGC and gSGC methods in estimating De of aeolian deposits from the south margin of the Tengger Desert was investigated and compared. OSL dates obtained from this study were conbined with previously published stratigraphy and chronology data to reconstruct history of aeolian activity in the south margin of the Tengger Desert. The following conclusions were drawn from this study: (1) Various simulated scenarios indicate that the measurement time and the number of channels, the variability of the decay constants, and the signal-to-noise ratios of a decaying component were major factors influencing the reliability of the BIC method for number of components identification. It suggests that the BIC method has great potential to estimate the number of components in an OSL decay curve with a moderate to high signal-to-noise ratio. OSL decay curves from the south margin of the Tengger Desert can be decomposed into 2–3 decaying components using the BIC method. Sandy loess samples contain more number of components compared to aeolian sand samples. Photoionisation cross-section for the fast and medium decaying components were respectively estimated at 2.57×10-17 cm2 and 1.35×10-18 cm2 according to analysis result from a total of 2333 OSL decay curves. Proportions of fast and medium decaying components account for respectively 88.16% and 9.84% of the initial 0–0.3 s OSL signals. (2) The Slice sampling method enables the sampling distributions of parameters to be obtained from the joint likelihood function that is determined by observations and the specified model. Both simulated and measured De sets demonstrate that parameters estimated from the Slice sampling method were consistent with those obtained using the maximum likelihood estimation method. The Slice sampling method is insensitive to initial states of parameters, and the standard errors (or confidence intervals) of parameters assessed using this method are more reliable. It suggests that the Slice sampling method provides an alternative for age model optimization. Analysis of De distributions of aeolian deposits from the south margin of the Tengger Desert using Slice sampling indicates that most of these samples have very narrow De distributions and were well-bleached before being buried. (3) The simulation demonstrate that the gSGC method is almost always intrinsically more precise than the SGC method given that the size of additional regenerative dose Dr used for re-normalizing the sensitivity-corrected natural dose signal is close in size to the given dose Dn. For Dn<120 Gy, differences in precision and accuracy of De estimates between SGC and gSGC methods were less than 4%. Though difference in accuracy of De estimates between the two methods is relatively insignificant for Dn<210 Gy, De estimates obtained from gSGC is obviously more accurate for higher doses. For aeolian deposits from the south margin of the Tengger Desert, De estimates obtained from SGC and gSGC methods were consistent with each other for De<50 Gy, and were comparable to those obtained from the single aliquot regenerative-dose protocol. It demonstrates that differences in precision and accuracy of De estimates between SGC and gSGC methods were insignificant for Holocene samples. (4) OSL dating results obtained from this study, in combination with previous aeolian sediment chronologies, show that aeolian activity occurred extensively between 20 and 10 ka, sand dunes were stabilized between 10 and 6 ka, and sand dunes were remobilized again after 3 ka. It suggested that the history of aeolian activity is closely related to the evolution of the East Asian Monsoon. Sand dunes mobilized when stronger winter monsoon prevailed and sand stabilization occurred when stronger summer monsoon prevailed. Sand dunes mobilization from 3 ka onwards may have been resulted from decreased summer monsoon and human activity. This question deserves further investigation. History of aeolian activity in the south margin of the Tengger Desert since 20 ka was in general consistent with evolution pattern of other deserts and sand fields in northernChina. The pattern can be summarized as: sand dunes mobilized from the last glacial maximum to the early Holocene, sand dunes stabilized from early to middle Holocene, and sand dunes reactivated again in late Holocene. (5) Both Northern Hemisphere summer insolation and Asia summer monsoon precipitation increased gradually since 12 ka, while sand dunes from the south margin of the Tengger Desert started to stabilized at 10 ka, suggesting that obvious lags exists between dune stabilization and meliorated climate. Increased atmospheric subsidence northeastern edge of the Qinghai-Tibetan Plateau and increased evaporation due to high summer insolation between 12–10 ka resulted in mobilization of sand dunes during this period. Sand dune activity in the south margin of the Tengger Desert is mainly controlled by surface effective moisture level. The variation of moisture is determined by monsoon precipitation and local convection and evaporation effects. The transition between sand mobilization and stabilization is mainly controlled by the surface effective moisture threshold. The supply of aeolian sand from source areas would be largely suppressed if moisture levels were higher than a certain threshold value. |
| 中文关键词 | 腾格里沙漠 ; 光释光测年 ; 贝叶斯 ; 切片采样 ; SGC与gSGC ; 风沙活动 |
| 英文关键词 | Tengger Desert OSL dating Bayesian Slice sampling SGC and gSGC Aeolian activity |
| 语种 | 中文 |
| 国家 | 中国 |
| 来源学科分类 | 自然地理学 |
| 来源机构 | 中国科学院西北生态环境资源研究院 |
| 资源类型 | 学位论文 |
| 条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/287842 |
| 推荐引用方式 GB/T 7714 | 彭俊. 腾格里沙漠南缘风积物光释光年代学[D]. 中国科学院大学,2017. |
| 条目包含的文件 | 条目无相关文件。 | |||||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [彭俊]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [彭俊]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [彭俊]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
