利用钻孔温度梯度重建过去地表温度变化研究

发布时间：2018-07-05 20:47

  本文选题：钻孔温度方法 + 多年冻土　； 参考：《兰州大学》2015年博士论文

【摘要】：对过去气候变化的研究有助于我们理解现在气候变化和预测未来气候。作为地球“第三极”的青藏高原,具有独特的地理位置和热力系统,并对气候变化非常敏感。了解青藏高原过去气候变化历史、变化特征,对于准确评估目前气候变化与预测未来青藏高原气候变化提供了重要依据。然而青藏高原的气象观测资料观测时间短,观测站点少且覆盖不均,因此仅靠气象观测不足以让我们了解青藏高原长时间尺度的气候变化。过去地表温度变化的气候信息可通过分析近期观测的钻孔温度剖面进行重建。青藏高原作为多年冻土大区,多年冻土内部土壤冻结,地温热传递主要以热传导方式进行,适宜利用钻孔温度方法的开展。钻孔温度方法较其他代用指标方法是基于地温剖面和地表温度变化的物理联系之上的,具有更强的物理意义。本文利用青藏高原多个钻孔温度剖面和地热梯度,重建青藏高原不同地区过去地表温度变化历史。地温垂直温度剖面受地中热流和地表温度变化影响。地球内部热流通过地热梯度影响地温形成稳态温度,在此基础上地表温度变化以热传导方式向地下传播使稳态地温发生扰动而产生偏离,偏离稳态温度的这部分地温作为瞬时温度记录了过去地表温度变化的信息。钻孔温度方法利用近期观测到的钻孔温度剖面,根据地温梯度分离出稳态温度,进而利用一维热传导模型分析瞬时温度剖面,重建过去地表温度变化。由于年际地表温度变化随深度明显衰减,而十年或更长时期的温度变化(“古气候信号”)向多年冻土更深处传播,多年冻土温度分布是气候变化和地表能量平衡长期变化的敏感指示。陆地作为天然的气候低通滤波器,使得多年冻土区的地温剖面可用来重建过去地表温度低频变化趋势。在研究钻孔温度重建过去地表温度变化的反问题之前,首先需要理解正向问题,地温如何响应地表温度变化。对于简单一次气候变化情形,根据一维热传导模型,对于不同类型地表温度变化地下温度场具有显式的解析解。对于多年冻土区域热传导且考虑相变的复杂过程,仅能通过数值方法求解。本文利用控制体积方法数值模拟多年冻土地温相变问题。控制体积方法在离散程度上介于有限差分和有限元方法之间,具有更直接的物理解释。计算时考虑未冻水含量及相变潜热,并随时间重新计算各深度热物理参数进而准确计算地温变化。通过地表周期变化情形阐明控制体积方法数值模拟多年冻土相变问题的过程及未冻水含量对热参数、地温的影响。基于对正问题的理解,本文对钻孔温度重建过去地表温度变化的反问题提出改进的tikhonov方法。基于过去地表温度变化和近期钻孔地温剖面间的物理联系,在对问题参数化后我们利用tikhonov正则化方法来重建过去地表气候变化。此方法是基于奇异值分解(svd)方法之上的,具有相同的参数化,都将问题转化为求解不适定的矩阵方程。本文利用两个数值例子模拟地表升温和复杂气候事件来验证方法有效性以及与奇异值分解方法相比的改进效果。由于钻孔温度观测具有无法避免的观测误差,我们对模拟的地温剖面添加随机扰动误差来模拟观测误差。利用扰动的地温剖面重建地表温度变化并与假设的地温剖面比较从而验证方法有效性。两个实验例子中重建的地表温度及相关不确定性分析表明tikhonov方法可较好地重建地表温度,并且改进方法可成功压制噪音导致的不稳定性得到更平滑的地表温度变化。通过比较地表温度误差可选出更适合tikhonov正则化的正则化参数选取方法。此外,本文利用tikhonov方法分析了钻孔温度方法的求解能力,可更好地理解重建的地表温度。钻孔温度重建地表温度变化的结果是依赖于所选取的反演方法,我们基于热传导方程反边界值问题给出创新的基本解方法。基本解方法根据热传导方程的基本解将问题参数化之后转化为待定线性系统,由于反问题的不适定性,此方程组受钻孔温度观测误差影响无法直接求解。利用tikhonov正则化和广义交叉核实方法选取正则化参数求解待定参数,进而同时重建地表温度变化和地表温度热流变化。数值模拟例子表明基本解方法是可行且稳定的,并且对模拟钻孔温度剖面添加不同水平随机误差扰动后仍能有效降低误差扰动带来的不适定性,得到精确的地表温度变化。利用钻孔温度重建地表温度变化与其他地球物理反问题相同,最大的求解难点在于观测误差导致的结果不稳定性。由于不同方法采用不同的参数化和优化方法,因此钻孔温度问题的结果依赖于所选取的方法。本文综合比较应用较广泛的泛函空间反演(fsi)、奇异值分解方法(svd)、改进的tikhonov方法以及创新的基本解方法(mfs)。通过五类不同类型地表温度变化的模拟例子来比较各方法数值结果:(1)阶梯变化;(2)线性升温;(3)光滑线性升温;(4)周期变化;(5)复杂周期变化,并在模拟例子中添加不同水平模拟观测误差。重建的地表温度结果比较表明,在钻孔温度剖面具有较小误差扰动下,所有方法均能给出较精确的地表温度变化重建结果。尽管四种方法具有不同的参数化方式和正则化方法选取,重建地表温度变化具有相似结果,仅在气候时间和温度幅度上有细微差别。方法的有效性是依赖于地表温度变化类型的。基本解方法更适用于重建周期变化和复杂周期变化信号。对于其他类型地表温度变化,Tikhonov方法在较小钻孔温度误差0.001℃和0.01℃情形下,结果最精确。泛函空间反演方法较在钻孔温度剖面误差较大时仍能重建地表温度变化趋势,但对初始地表温度重建较其他方法相比有较大误差。并且泛函空间反演在重建近期地表温度变化时具有更高分辨率,更精确。基于钻孔温度方法研究,本文根据青藏高原不同地区钻孔温度剖面对各研究点进行过去地表温度变化的单点重建研究,利用钻孔温度方法反演得到不同时间区间古气候信息:1)黑河上游100米钻孔PT1钻孔由于气候变暖导致进入多年冻土的长期净热流约为0.014 Wm-2,深处稳态热流约0.0247 Wm-2。PT1钻孔1952年至2012年地表温度由-2.7℃线性升高约0.5至0.65℃;2)黑河上游150米钻孔PT9钻孔地热梯度为2.25℃/100m,1895年至2015年地表温度由-2.3℃升温至-1.5℃;3)奇异值分解方法和Tikhonov方法根据五道梁120米钻孔温度剖面重建地表温度结果表明在过去1930年至2013年间地表温度升温1.8(±0.2)℃,且剧烈升温过程开始于1980年代。五道梁气象观测站的气温观测结果验证了Tikhonov方法重建2008年至2012年间的地表温度波动,且在时间重合阶段气温和重建的地表温度具有相同趋势。4)根据昆仑山钻孔220米钻孔温度剖面,奇异值分解方法和Tikhonov方法重建地表温度结果表明,1700年至2013年地表温度由-6.5(±0.8)℃升高至-2.8(±0.2)℃。两方法重建的地表温度变化具有相同趋势,具体升温时间和幅度略有差别。根据五道梁观测站的气温观测数据对比表明,Tikhonov方法重建的地表温度更可靠。5)柴达木盆地7个钻孔(最大深度220米至400米)温度反演表明此区域过去514年地表温度由6.1℃升高了1.2℃(-0.11~2.21℃),并表现出1500年至1900年间的小冰期寒冷信号。最冷时期发生在1780至1790年间,当时的地表温度为5.4℃。在19世纪和20世纪间,重建的地表温度具有升温趋势,且在20世纪末达到最高值,随后开始降温。重建的地表温度变化幅度已由EdGCM模式模拟的地表平均气温所验证,细节温度特征得到代用指标结果验证。基于钻孔温度方法研究,本文给出钻孔温度方法的改进方法和创新方法,并进行方法比较。本论文根据青藏高原中部不同地区钻孔温度剖面进行过去地表温度变化的单点重建研究,并利用地温梯度反演得到各钻孔位置的过去地表温度变化信息。
[Abstract]:The study of past climate change will help us understand the present climate change and predict the future climate. As the "third pole" of the earth, the Qinghai Tibet Plateau has unique geographical location and thermal system, and is very sensitive to climate change. It provides an important basis for climate change in the Qinghai Tibet Plateau in the future. However, the observation time of the meteorological observation data in the Qinghai Tibet Plateau is short, the observation site is few and the coverage is uneven. Therefore, the meteorological observation alone is not enough to let us understand the climate change of the long time scale of the Qinghai Tibet Plateau. The Qinghai Xizang Plateau, as a large area of permafrost, freezes the soil in permafrost, and the heat transfer is mainly carried out in the heat conduction mode, which is suitable for the development of the borehole temperature method. The borehole temperature method is the physical relation of the geothermal profile and the surface temperature change compared with the other substitution index methods. In this paper, the history of surface temperature changes in different regions of the Qinghai Tibet Plateau is rebuilt by using the temperature profiles and geothermal gradients in the Qinghai Tibet Plateau. The vertical temperature profiles of the Qinghai Tibet Plateau are influenced by the changes of the heat flow and surface temperature in the earth. On this basis, the change of surface temperature changes by heat conduction way to the ground to cause the deviation of the steady state temperature, and the part of the ground temperature deviating from the steady state temperature as the instantaneous temperature records the information of the change of the surface temperature in the past. The transient temperature profile is analyzed by one dimensional heat conduction model, and the change of surface temperature in the past is rebuilt. The temperature variation of the interannual surface temperature changes with the depth obviously, and the temperature change ("paleoclimate signal") of ten years or longer ("paleoclimate signal") propagates deeper into the permafrost, and the temperature distribution of permafrost is the climate change and the surface energy. The land temperature profile of the permafrost region can be used to reconstruct the low frequency trend of the surface temperature in the permafrost region as a natural climate low pass filter. Before the inverse problem of the surface temperature changes in the reconstruction of the borehole temperature, the first need to solve the forward problem and how the geotemperature respond to the surface temperature change. For a simple one time climate change, according to the one-dimensional heat conduction model, the underground temperature field of different types of surface temperature changes has an explicit analytical solution. For the complex process of heat conduction in permafrost regions and considering the phase transformation, the numerical method can only be solved by numerical method. This paper uses the method of controlling volume to simulate the perennial frozen land. The control volume method is between the finite difference and the finite element method in the discrete degree, and has a more direct physical explanation. The calculation of the unfrozen water content and the latent heat of the phase change is taken into account, and the thermal physical parameters of each depth are recalculated and then the ground temperature change is accurately calculated with time. The control volume is clarified through the change of the surface period. The process of the phase transformation of permafrost and the influence of the content of unfrozen water on the thermal parameters and the ground temperature are numerically simulated. Based on the understanding of the positive problem, an improved Tikhonov method is proposed for the inverse problem of the reconstruction of the surface temperature in the past. Based on the physical relations between the surface temperature changes and the recent borehole geothermal profiles, After the problem is parameterized, we use the Tikhonov regularization method to reconstruct the surface climate change in the past. This method is based on the singular value decomposition (SVD) method and has the same parameterization. All the problems are transformed into an ill posed matrix equation. Two numerical examples are used to simulate the surface temperature rise and complex climate events. The effectiveness of the method and the improved effect compared with the singular value decomposition method. As the observation error of the borehole temperature is unavoidable, we add random perturbation error to the Simulated Geothermal profile to simulate the observation error. Method effectiveness. The reconstruction of surface temperature and correlation uncertainty in two experimental examples shows that the Tikhonov method can better reconstruct the surface temperature, and the improved method can successfully suppress the noise induced instability to get a more smooth surface temperature change. By comparing the surface temperature error, it is more suitable for Tikhonov regularization. In addition, the Tikhonov method is used to analyze the solving ability of the borehole temperature method, which can better understand the surface temperature of the reconstructed surface. The result of the change of the surface temperature in the reconstruction of the borehole temperature depends on the selected inversion method. Based on the inverse boundary value problem of the thermal conduction equation, we give the basic solution. According to the basic solution of the heat conduction equation, the basic solution transforms the problem into a undetermined linear system. Due to the discomfort of the inverse problem, the equation can not be solved directly by the influence of the borehole temperature observation error. Tikhonov regularization and the generalized cross verifying method are used to select the regularized parameters to solve the undetermined parameters. At the same time, the surface temperature change and surface temperature heat flux are rebuilt. The numerical simulation example shows that the basic solution method is feasible and stable. And after adding different horizontal random error disturbance to the simulated borehole temperature profile, it can effectively reduce the discomfort caused by the error disturbance, and get the exact surface temperature change. The variation of ground surface temperature is the same as other geophysical inverse problems. The most difficult problem is the instability of the results caused by the observation error. Because different methods are used for different methods of parameterization and optimization, the result of the temperature problem depends on the selected method. FSI), singular value decomposition method (SVD), improved Tikhonov method and innovative basic solution (MFS). Numerical results of five different types of surface temperature variations are compared: (1) ladder change; (2) linear temperature rise; (3) smooth linear heating; (4) periodic variation; (5) complex periodic changes, and added in simulation examples Compared with the simulated observation errors at different levels, the results of the rebuilt surface temperature show that, under the small error disturbance of the borehole temperature profile, all the methods can give more accurate results of the surface temperature change reconstruction. Although the four methods have different parameterization methods and the regularized square method, the reconstruction of the surface temperature change has a similar knot. The effectiveness of the method is dependent on the type of surface temperature change. The basic solution is more suitable for the reconstructive cycle and the complex periodic change signals. For the other types of surface temperature changes, the Tikhonov method results in the case of a smaller drill hole temperature error of 0.001 and 0.01 degrees centigrade. Most accurate. The functional space inversion method can still reconstruct the trend of surface temperature change when the error of the borehole temperature profile is large, but the initial surface temperature reconstruction has a larger error compared with other methods. And the functional space inversion has higher resolution and more accurate in the reconstruction of the surface temperature in the near future. In this paper, a single point reconstruction of the surface temperature changes in the past is carried out on the basis of the borehole temperature profiles in different areas of the Qinghai Tibet Plateau. The paleoclimate information of different time intervals is retrieved by the method of borehole temperature: 1) the long-term net heat flow of the PT1 boreholes in the 100 meter borehole in the upper reaches of Heihe is about 0.014 due to the warming of the climate. Wm-2, deep steady heat flow about 0.0247 Wm-2.PT1 boreholes from 1952 to 2012, the surface temperature rises from -2.7 C to 0.5 to 0.65 degrees C; 2) the geothermal gradient of 150 m boreholes in the upper reaches of Heihe is 2.25 centigrade /100m, the surface temperature from 1895 to 2015 is heated from -2.3 to -1.5, 3) and 3) the singular value decomposition method and the Tikhonov method are based on the five beam 120 meters. The results of surface temperature reconstruction from the pore temperature profile show that the surface temperature increased by 1.8 (+ 0.2) C in the past 1930 to 2013, and the intense heating process began in 1980s. The temperature observation results of the five road meteorological observation station verified that the Tikhonov method rebuilt the ground surface temperature fluctuations from 2008 to 2012, and the temperature and weight in the time reclosing stage The ground surface temperature has the same trend.4) according to the 220 meter borehole temperature profile of the borehole in Kunlun Mountains. The results of the surface temperature reconstruction by the singular value decomposition method and the Tikhonov method show that the surface temperature rises from -6.5 (+ 0.8) C to -2.8 (+ 0.2) C from 1700 to 2013. The surface temperature change of the two method has the same trend and the specific heating time. According to the temperature observation data of the five beam observation station, the surface temperature rebuilt by the Tikhonov method is more reliable.5). The temperature inversion of the 7 boreholes in the Qaidam Basin (the maximum depth of 220 to 400 m) shows that the surface temperature in this area has increased by 1.2 (-0.11~2.21 C) at 6.1 C in the past 514 years, and shows from 1500 to 1900. The cold signal of the small ice period of the year. The coldest period occurred between 1780 and 1790, at that time the surface temperature was 5.4. In 19th Century and 20th Century, the rebuilt surface temperature had a warming trend, and reached the peak at the end of twentieth Century, and then began to cool. The surface temperature change of the reconstructed surface has been measured by the average surface temperature simulated by the EdGCM model. It is verified that the characteristics of the detail temperature are verified by the substituting index. Based on the study of the borehole temperature method, this paper gives the improved method and innovation method of the borehole temperature method, and compares the methods. Temperature gradient inversion is used to get the information of past surface temperature changes in each borehole.
【学位授予单位】：兰州大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P467
，

本文编号：2101730