天山地区GPS连续站高程时间序列分析

发布时间：2018-03-05 09:31

本文选题：GPS　切入点：高程时间序列　出处：《中国地震局地震研究所》2013年硕士论文　论文类型：学位论文

【摘要】：二十世纪90年代初,IGS(International GPS Service for Geodynamics)的建立标志着GPS全球定位系统应用于地球物理研究工作的开始。世界上各个国家的地学研究机构开始在不同的国家、不同的地区建立起GPS连续跟踪站。随着GPS连续跟踪站的大量涌现,积累了大量的数据,使得GPS对于地球物理研究的可行性与可靠性大幅提高。 UNAVCO Facility(美国卫星导航系统与地壳形变观测研究大学联合体)在全世界范围内建立了大量GPS连续跟踪站,本论文利用该机构在天山地区的若干连续站观测数据进行研究。本文前面部分介绍了时间序列分析的基本理论,各种时间序列模型；介绍了时间序列分析常用的各种方法,例如最小二乘线性拟合、功率谱密度的计算以及小波多分辨率分析等。在本文的后半部分,先筛选了数据连续性较好,缺值情况较轻的11个连续跟踪站。通过高精度GNSS数据处理软件-Bernese解算得到了这些连续站十年左右的高程方向的原始坐标时间序列。然后利用最小二乘线性拟合的方法,从原始坐标时间序列中提取出线性运动速度,并将线性运动趋势从原始坐标时间序列中予以剔除,获得了无线性趋势的残差时间序列。通过计算发现,唯一一个位于天山地区以外的作为参考而选择的KRTV站在十余年间站线性运动的趋势不明显,其他10个连续观测站的线性拟合的斜率均为正值,并且在大小上呈现出良好的区域一致性,说明该地区在过去十余年间存在整体的隆升趋势。随后,对残差时间序列进行了谱分析,利用Welch方法计算功率谱的谱指数。分析结果表明,除少数站(3个)的噪声模型可以表示为白噪声+闪烁噪声外,其他站均需要用白噪声+闪烁噪声+随机漫步噪声的模型描述(即其谱指数均小于-1)。最后对时间序列进行了小波多分辨率分析,利用Coiflets小波滤波器,提取出各个跟踪站的年周期项和半年周期项。并统计了各个测站周年项与半周年项的振幅以及最大值出现的月份。通过分析发现,各个跟踪站均具有明显的季节性变化项,并且这种变化表现出一定的区域性。这种季节性的变化趋势是由各种复杂的地球物理因素引起的。因此在进行高精度的地壳形变监测时,应当选择相同季节的进行GPS重复观测,以排除季节性误差的影响。
[Abstract]:In 1990s, the establishment of International GPS Service for Geodynamic marked the beginning of the application of the GPS Global Positioning system to geophysical research. GPS continuous tracking stations have been established in different regions. With the emergence of GPS continuous tracking stations a large number of data have been accumulated and the feasibility and reliability of GPS for geophysical research have been greatly improved. A large number of GPS continuous tracking stations have been established in the United States Satellite Navigation system (UNAVCO) and crustal deformation observation Research University (USSCS). In this paper, the data of several continuous stations in Tianshan region are used to study the data. In the first part of this paper, the basic theory of time series analysis and various time series models are introduced, and various methods of time series analysis, such as least square linear fitting, are introduced. Power spectral density calculation and wavelet multi-resolution analysis. In the second half of this paper, the data continuity is better. The original coordinate time series of the elevation direction of these continuous stations about ten years are obtained by using the high-precision GNSS data processing software-Bernese, and then the least square linear fitting method is used. The linear motion velocity is extracted from the original coordinate time series, and the linear motion trend is removed from the original coordinate time series. The trend of linear motion of the only KRTV station selected as a reference outside Tianshan area is not obvious in more than a decade, and the slope of linear fitting for the other 10 continuous observation stations is all positive. It also shows good regional consistency in size, which indicates that there is an overall uplift trend in this area in the past ten years. Then, the spectral analysis of the residual time series is carried out, and the spectral exponent of the power spectrum is calculated by using the Welch method. The results show that the noise model of a few stations (3) can be expressed as white noise flicker noise. The other stations need to be described by the model of white noise flicker noise random walk noise (that is its spectral index is less than -1). Finally, the wavelet multi-resolution analysis of time series is carried out, and the Coiflets wavelet filter is used. The annual and semi-annual periodic items of each tracking station are extracted, and the amplitude and the month of the maximum value of the annual and semi-annual items of each station are counted. Through analysis, it is found that each tracking station has obvious seasonal variation term. The seasonal variation is caused by a variety of complex geophysical factors. Therefore, GPS repeated observation should be carried out in the same season for high precision crustal deformation monitoring. To rule out seasonal errors.
【学位授予单位】：中国地震局地震研究所
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P228.4

【参考文献】