Arid
青藏高原永冻土活动层厚度预测指标集的建立与制图
其他题名Construction of predictive index set and predictive mapping on active layer thickness of permafrost in Qinghai-Tibet plateau
陈吉科
出版年2014
学位类型硕士
导师赵玉国
学位授予单位中国科学院大学
中文摘要青藏高原多年冻土区是世界上中低纬度海拔最高、面积最大的多年冻土区,也是典型的高原冻土区。但是,随着全球气候变暖,青藏高原多年冻土区出现了退化,表现为活动层厚度增大、多年冻土厚度减小乃至多年冻结层消失。其中,活动层是指在多年冻土之上每年寒季冻结、暖季融化的岩土层。活动层在能量平衡、水文循环、大气与陆地表面的碳交换、生态系统以及在寒冷地区的人类基础设施方面起着重要的作用,是气候变化的直接指示器,在全球变化系统中起着十分重要的作用。准确确定永冻土活动层厚度在全球变化研究和模拟、水资源评估与利用、多年冻土的生态保护和治理等方面都具有重要作用。同时,研究永冻土活动层厚度可将自然状态下的地质、地理因素改变或者将自然变化控制条件转化为人为控制因素,这在寒区工程设计和施工等方面具有重要意义。 由于青藏高原气候恶劣、环境艰苦,获得大量代表性较好的样点非常困难;同时,考虑到物理模型难以空间扩展以及半经验、经验模型对局地因素考虑不全的局限性,本研究尝试基于少量代表性较差的调查样点构建用于模拟活动层厚度的预测指标集;然后,通过建立景观因子与活动层厚度之间的土壤景观模型模拟得到了研究区永冻土活动层厚度空间分布图。 因此,本研究通过收集研究区永冻土活动层厚度数据、数字高程模型、MODIS标准陆地产品等数据,利用经典统计学方法、GIS技术,研究了永冻土活动层厚度与其各个影响因子之间的相关关系;利用主成分分析、相关分析以及影响因子频率分析,去除冗余变量,获得了影响因子集中起主要作用的预测指标集;然后,在利用专家经验法等对预测指标集中的各个影响因子赋予权重的基础上,采用样点个体代表性的方法获取了研究区永冻土活动层厚度的分布图,并进行了结果比较。本研究主要结论如下: (1)利用采样点活动层厚度与不同时间段的地表反照率的相关性分析,选取2009?2011连续三年春季平均地表反照率数据作为模拟永冻土活动层厚度的一个影响因子。 (2)不同植被类型采样点的永冻土活动层厚度分析表明:沼泽草甸的永冻土活动层厚度的平均值最小,高寒草原的永冻土活动层厚度的离散程度最大;高寒草甸、高寒草原、高寒荒漠的永冻土活动层厚度与沼泽草甸的永冻土活动层厚度呈显著性差异,但高寒草原、高寒荒漠永冻土活动层厚度没有显著性差异。 (3)永冻土活动层厚度与土壤属性和环境因子(坡度、坡向、海拔、地表反照率、夜间地表温度、白天地表温度、地表昼夜温差、NDVI)的相关分析表明:永冻土活动层厚度与土壤表层粘粒、粉粒呈显著负相关,与土壤有机质、极细砂粒、细砂粒、中砂粒无显著相关;与坡度相关性最强,与Albedo、海拔、夜间地表温度呈极显著相关,而与NDVI、坡向、纬度、白天地表反照率的相关性不显著。 (4)利用影响因子随着训练集次数的增加出现的频率变化,选择出现频率较高且趋于稳定的影响因子作为模拟研究区永冻土活动层厚度分布的指标集:地表昼夜温差、海拔、坡度、坡向、NDVI、母岩类型,涵盖了气候、地形、母质、植被四个方面的因素,能够较为有效地反映永冻土活动层厚度的影响因素。 (5)通过主成分分析等确定预测指标集中各因子对永冻土活动层厚度的权重大小,采用加权法和最小限制因子法推算两点之间的相似度函数,获得了研究区的永冻土活动层厚度分布图,结果表明:加权法精度高于最小限制因子法;其取得最佳精度的权重组合为:气候变量为0.5,地形变量为0.3,植被为0.2;验证精度表明:AC为0.75,MAE为72.35cm,RMSE明显小于SD。模拟得到的永冻土活动层厚度没有呈现大片状分布,在较小的区域内仍具有变异,说明所选的预测指标集在反映活动层厚度大的空间变化趋势的同时,也能够反映永冻土活动层厚度的局地变化特征。 (6)通过对不同母岩、坡向、海拔梯度活动层厚度空间分布特征分析发现:母岩为沉积岩的区域活动层厚度的平均值最大;母岩为深成岩的区域活动层厚度平均值最小。母岩为沉积岩、疏松沉积物、深成岩的区域活动层厚度范围和变异程度依次减小。在研究区域内坡向类型对活动层厚度的影响程度较小;较低海拔梯度活动层厚度平均值较大,高海拔梯度区域活动层厚度平均值较小且差别不大。 总之,本文基于少量调查样点利用样点个体代表性方法实现了研究区永冻土活动层厚度分布制图,取得了较高的精度和分辨率。另外,针对调查样点全局代表性较差以及数量有限的局限,本研究没有采用常规的一次性全样本最优建模方式,而采取多次抽样分别建模方式,获得预测指标集和制图结果,这对相似的研究具有借鉴意义。 关键词:青藏高原;永冻土;活动层厚度;指标筛选;数字土壤制图;因子权重
英文摘要Qinghai-Tibet plateau is the highest and most extensive middle-low latitude permafrost region in the world. However, with the continuously increasing of global warming, permafrost has occurred extensive degradation over the Qinghai-Tibet plateau such as the increasing of active layer thickness, the diminishing of permafrost thickness, even the disappearance of permafrost. Active layer is defined as the top layer of ground subject to annual thawing in the warm seasons and annual freezing in the cold seasons in areas underlain by permafrost. Active layer plays important roles in energy balance, hydrology cycle, carton change between the atmosphere and land-surface, ecological system and human infrastructures in the cold regions.It is important for global change research and stimulation, water resource evaluation and utilization, and ecological conservation and governance to accurately determine the active layer thickness. By studying the active layer thickness, it is conducive to change the geological and geographical factors in the natural state, and convert controlled conditions of spontaneous change into artificially controlled factors which are important in engineering design and construction in the cold regions. On the on hand, there has many difficulties in acquiring a large number of sample sites with high representativeness due to bad weather and harsh environment in Qinghai-Tibet Plateau; on the other hand, in view of the limit of physical model, semi-experience model and experience model, this study attempted to construct the prediction index set by using a small number of sample sites with low representativeness; then the spatial distribution of the active layer thickness was obtained by constructing soil-landscape model between influencing factors and the active layer thickness. Based on the collected data including active layer data, DEM, MODIS standard land products, this study researched correlation coefficients between active layer thickness and influencing factors using classical statistical methods and GIS. Firstly, principal component analysis, correlation analysis and changing trends of influencing factors were used to remove redundant variables. In consequence, the predictive index set comprised of factors which play major roles on the active layer thickness was constructed. Secondly, every influencing factor of the predictive index set was endowed with the corresponding weight respectively by principal component analysis and so on. Based on this, spatial distribution of active layer thickness were finally achieved by using the method of based on individual representativeness of sample sites in this study and mapping results achieved in different situations of weight combination were compared. We concluded that: 1. The correlations between the active layer thickness and albedo data for different time periods of three years (2009-2011) were analyzed. The results showed that active layer thickness was significantly correlated with albedo data in March, April, and spring. The three-year average albedo data in spring were chosen as an influencing factor to predict the spatial distribution of active layer thickness with regard to the quality of remote sensing data. 2. The active layer thickness of sample sites for different vegetation types was researched, the results showed that: the average active layer thickness of sample sites for swamp meadow was the minimum. Discrete degree of the active layer thickness was the maximum as the vegetation type of sample sites is alpine steppe. There were significant differences in active layer thickness of sample sites for different types: the active layer thickness of samples for swamp meadow demonstrated significant differences with the active layer thickness for other types of vegetation. The active layer thickness of sample sites for alpine meadow appeared no differences with the active layer thickness of sample sites for alpine desertification. 3. The correlation ana
中文关键词青藏高原 ; 永冻土 ; 活动层厚度 ; 指标筛选 ; 数字土壤制图 ; 因子权重
英文关键词The Qinghai-Tibet plateau Permafrost Active layer thickness Factor selection Predictive index set Soil mapping
语种中文
国家中国
来源学科分类地图学与地理信息系统
来源机构中国科学院南京土壤研究所
资源类型学位论文
条目标识符http://119.78.100.177/qdio/handle/2XILL650/287410
推荐引用方式
GB/T 7714
陈吉科. 青藏高原永冻土活动层厚度预测指标集的建立与制图[D]. 中国科学院大学,2014.
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
[陈吉科]的文章
百度学术
百度学术中相似的文章
[陈吉科]的文章
必应学术
必应学术中相似的文章
[陈吉科]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。