利用中尺度数值气象模式预报大气光学湍流

发布时间：2018-11-23 16:30

【摘要】：光学湍流效应是制约光电工程应用的重要影响因素,定量描述湍流效应与折射率结构常数(C_n~2)有关。对于具体的光电工程应用,一般可以用多种测量仪器进行C_n~2的实时实地测量,以定量分析湍流效应。然而,在许多光电工程的设计及可能的应用场景中,需要对应用场景的C_n~2进行大范围、长期系统地测量,这是测量仪器难以胜任的。近十多年来,利用中尺度数值气象模式获取大气光学湍流参数逐渐成为国际上较为关注的研究热点,为此我们尝试开展了利用数值气象模式(WRF)预报C_n~2廓线及其随时间变化特征的研究。本文总结了国内外C_n~2建模以及预报模型研究的基础上,开展了利用WRF模式预报丽江高美古天文观测站、茂名博贺海洋观测站、新疆库尔勒地区的C_n~2廓线以及不同下垫面(南海近海面、南极泰山站冰雪面、成都内陆等)C_n~2时间演变特征的研究。本文主要围绕WRF模式预报C_n~2的技术特点和难点以及可行性,开展了以下几方面的研究工作:1.详细介绍了 WRF模式基本情况包括模式框架、坐标方程、物理参数过程等,同时介绍了 WRF模式在PC机上的安装流程、初始场数据的使用以及调试运行等。2.用常规气象参数估算C_n~2廓线的方法主要是依据Tatarskii模式,而Tatarskii模式中外尺度是关键参数,但难以直接测量。研究对比了四种外尺度参数化模式(Dewan模式、Coulman模式、Sterenborg模式和HMNSP99模式)。用探空气球实测的常规气象参数估算的C_n~2廓线与湍流气象探空仪实测的C_n~2廓线进行对比。发现采用四种外尺度模式估算的C_n~2廓线无论在变化趋势上还是量级上,四种模式之间的差异都很大。发现在与外尺度有关的如温度梯度、风速梯度、Richardson数等几个参数中,加入了温度梯度和风速梯度的HMNSP99外尺度模式估算的C2与测量的C_n~2在变化趋势和量级上最为接近。3.基于WRF模式,结合Tatarskii模式和HMNSP99模式分别预报了高美古、茂名以及库尔勒等三个典型地区的温度、风速和C_n~2廓线,并用湍流气象探空仪实测的相应廓线作为对比。结果表明:WRF预报的温度和风速廓线与三个地区的探空实测结果非常接近,相关性可分别达到90%、80%以上。预报的C_n~2廓线基本满足C_n~2随高度变化的特征,相关性在75%左右,但C_n~2廓线的变化细节与实测值差异稍大。三个典型地区气候类型各异,表现出的C_n~2廓线分布特征也具有明显的气候类型差异。总体来说WRF预报的高美古地区的C_n~2廓线要好于在茂名和库尔勒地区的预报值。4.以Monin-Obukhov相似理论为依据并结合Bulk空气动力学方法,利用WRF模式预报了中国南海近海面上、中国南极泰山站冰雪面上以及中国内陆成都地区近地面层常规气象参数(如温度、湿度、风速、风向,等)和C_n~2。用自动气象站和温度脉动仪实测的近地面层常规气象参数以及C_n~2作为对比验证,结果显示WRF预报的近地面层常规气象参数以及C_n~2与实测值吻合的较好,而且预报值能够准确地反映出近海面上,冰雪面上的常规气象参数以及C的日变化特征。使用了如平均偏差(Bias)、均方根误差(RMSE)、纠正偏差(σ)、相关系数(Rxy)、列联表等统计工具分析了预报值的稳定性和可靠性,其相关统计结果是令人满意的。成都、南海以及南极泰山站在一定程度上代表了陆地、海面以及冰雪面等典型的下垫面类型,从对比结果可以看出C_n~2的日变化特征具有显著的地域性差异。总体来看,WRF模拟冰雪面上的C_n~2与实测结果吻合的最好,海面上次之,陆地上最差。陆地上的环境更加复杂多变,测量点受周围影响较大,并且模式的水平分辨率有限,这些因素可能是陆地上预报值与测量值差异较大的原因。通过在高美古、茂名、库尔勒三个典型地区以及陆地、海面、冰雪面等不同下垫面的C_n~2的估算预报和测量对比表明,利用WRF模式预报的C_n~2廓线和近地面层C_n~2在变化趋势和量级上与实测数据基本相符合,但在WRF模式预报的常规气象参数的空间分辨率和精度、不同下垫面的光学湍流参数化方法等方面还需要进一步的改进和完善。
[Abstract]:The optical turbulence effect is an important factor that restricts the application of the photoelectric engineering, and the quantitative description of the turbulence effect is related to the refractive index structure constant (C _ n ~ 2). In this paper, the real-time field measurement of C _ n ~ (2) can be carried out with a variety of measuring instruments for the specific photoelectric engineering application, so as to quantitatively analyze the turbulence effect. However, in many photoelectric engineering design and possible application scenarios, it is necessary to measure the C _ n ~ 2 of the application scene for a long time, which is difficult for measuring instruments. In recent ten years, using the mesoscale numerical weather model to acquire the atmospheric optical turbulence parameters has become a hot topic of international interest, and we have tried to use the numerical weather pattern (WRF) to forecast the C _ n ~ 2 profile line and its time-changing characteristics. On the basis of the study of the models of C _ n ~ 2 and the research of the forecasting model, this paper has carried out the study on the model of WRF for the prediction of the high-quality ancient astronomical observatory of Lijiang, the Ocean Observatory of Maoming Boga, the C _ n ~ 2 contour lines in the Korla region of Xinjiang, and the different subsides (the near-sea surface of the South China Sea and the ice and snow surface of the Mount Tai Station, Study on the characteristics of the time evolution of c _ n ~ 2 in the inland of chengdu). In this paper, the technical characteristics, difficulties and feasibility of C _ n ~ (2) are predicted mainly around WRF mode, and the following research work is carried out: 1. The basic situation of WRF mode is described in detail, including the mode frame, the coordinate equation, the physical parameter process, etc. The installation process of WRF mode on the PC, the use of the initial field data and the operation of the debugging are also introduced. The method for estimating the C _ n ~ 2 contour line by the regular meteorological parameters is mainly based on the Tatiarskii mode, while the Chinese and foreign dimensions of the Tatiareskii mode are the key parameters, but it is difficult to measure directly. Four out-of-scale parametric modes (Dewan mode, Coulman model, Sterigenborg mode, and HMNSP99 mode) are compared. The C _ n ~ 2 contour line estimated by the conventional meteorological parameters measured by the sounding balloon is compared with the measured C _ n ~ 2 contour line of the turbulent meteorological sounding instrument. It is found that the C _ n ~ 2 contour lines, which are estimated by the four out-scale modes, are both on the order of change or on the order of magnitude, and the difference between the four modes is very large. It is found that in several parameters, such as temperature gradient, wind speed gradient, Richardson number and so on, C2 and measured C _ n ~ (2) of the HMNSP99 external-scale model with temperature gradient and wind speed gradient is the closest to the variation trend and magnitude. Based on WRF model, the temperature, wind speed and C _ n ~ 2 contour line of the three typical areas, such as the high, the Maoming and the Korla, are predicted in combination with the Tatoskii mode and the HMNSP99 mode, and the corresponding contour lines measured by the turbulence meteorological sounding instrument are used as the comparison. The results show that the temperature and the profile line of WRF are very close to that of the three areas, and the correlation can reach 90% and more than 80%, respectively. The predicted C _ n ~ 2 profile basically satisfies the characteristics of the height variation of C _ n ~ (2), the correlation is about 75%, but the variation of the C _ n ~ 2 profile line is slightly different from the measured value. The climate types of the three typical regions are different, and the distribution characteristics of the C _ n ~ 2 profile also have obvious climatic types. In general, the C _ n ~ 2 contour line in the high-US-old region of WRF forecast is better than the forecast value in Maoming and Kuril area. The general meteorological parameters (such as temperature, humidity, wind speed, wind direction, and the like) and C _ n ~ (2) of the near-surface layer of the South China Sea, on the ice and snow surface of the south China Sea, and in the inland Chengdu area of China, are predicted with the method of the Monin-Oukhov similarity theory and combined with the Bulk air dynamics method. The conventional meteorological parameters and C _ n ~ 2 of the near-ground level measured by the automatic weather station and the temperature pulse instrument are used as the comparative verification. The results show that the conventional meteorological parameters of the near-ground level of WRF prediction and the good agreement between the C _ n ~ (2) and the measured value are good, and the forecast value can be accurately reflected on the offshore surface. the conventional weather parameters on the ice and snow surface and the diurnal variation characteristics of the c. The stability and reliability of the forecast value are analyzed by statistical tools such as the mean deviation (Bias), the mean square error (RMSE), the correction deviation (Rxy), the coefficient of correlation (Rxy), and the column-linked table, and the relevant statistical results are satisfactory. The typical lower surface types, such as land, sea surface and ice and snow surface, are represented in Chengdu, the South China Sea and the Antarctic Mountain Tai Station. From the results of the comparison, it can be seen that the diurnal variation of C _ n ~ 2 has significant regional difference. In general, WRF simulated the best of the C _ n ~ 2 on the ice and snow surface and the measured results, the last of the sea and the worst on the land. The environment of the land is more complex and changeable, the measuring point is affected by the surroundings, and the horizontal resolution of the mode is limited, and these factors may be the cause of the difference between the forecast value on land and the measured value. It is shown that the C _ n ~ 2 contour line and the near-ground layer C _ n ~ (2) in WRF mode are basically in accordance with the measured data by the comparison of the estimation and the measurement of the C _ n ~ (2) in the three typical areas of the high-US-old, Maoming and Korla and the land, sea surface and ice-snow surface. However, the spatial resolution and precision of the conventional meteorological parameters in WRF mode and the method of optical turbulence parametrization of different subplanes also need to be further improved and improved.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：O357.5;O436

【参考文献】