【作者】 王皓；

【导师】 魏鸣；

【作者基本信息】 南京信息工程大学 ， 大气遥感科学与技术， 2013， 博士

【摘要】 水汽是大气的一种主要成分,也是一种温室气体,尽管在大气中的含量很少,但是其在大气中的变化却十分剧烈。其空间分布极不均匀,时间变化也极其迅速,是大气中变化最大的一种成分,并且其变化尺度比风速、气温要精细得多。气象学和天气预报的基本问题之一就是要较为准确地测量大气中水汽的分布及其变化情况。运用GPS技术估算大气中的水汽含量是20世纪90年代兴起的一种极有潜力、实用价值很大的一种新方法或监测技术,由于其在获取大气水汽时具有高精度、高容量、高时空分辨率、全天候、近实时等诸多优点,因此受到了气象工作者的广泛重视。国家、部门和地方为此投入了大量的人力和物力,并已获取、积累了大量的GPS原始资料,但目前我国地基GPS水汽监测网建设还处于起步阶段,加之该探测技术涉及测量学与气象学知识的交叉融合,存在着很多需要解决及进一步研究的问题。通过GPS延迟量反演大气可降水量具有处理环节多、技术较复杂等特点,它们制约了地基GPS资料的有效使用；在GPS反演水汽过程中对于加权平均温度的建模以及应用缺乏统一的标准；地基GPS数据处理过程中,存在GPS原始数据或地面气象数据中有一方数据缺失,就无法计算出GPS可降水量(GPS-PWV)的问题；利用GPS-PWV在天气气候分析中多局限于PWV自身变化的研究,对水汽变化的深层次缘由研究较少。针对以上这些问题,本文开展了深入的分析与研究,主要得到了以下一些结论：(1)通过自身的实践与应用,较全面地提出了解决目前气象业务部门运用地基GPS反演水汽技术中存在的两个关键性问题,即：GPS数据处理软件的使用问题以及业务化运行中的一体化处理问题；(2)以地基GPS反演水汽的整个流程为主线,定性地分析了其中存在的主要误差源,定量地计算出了允许误差的范围,进一步提升了GPS反演水汽的精度问(3)利用数学方法和智能算法分别提出了两种可降水量缺测时的简便插补方案,通过主成分分析法解决了拟合因子的选择问题,通过敏感性试验得到了拟合可降水量神经网络模型的结构组成；(4)对于加权平均温度存在的拟合公式不统一,各区域有不同的Tm计算模型的问题,提出了适用于我国的通用Tm计算模型,并达到了气象应用的精度要求；(5)对江苏和重庆地区两次强降雨过程中GPS-PWV的变化特征进行了细致分析,并结合NCEP再分析资料、常规探空资料、卫星云图资料、地面气象资料以及wrf数值模拟结果等对过程中的动力、热力过程进行了详细剖析,揭示出了水汽变化的深层次缘由,深化了GPS-PWV作为预报指标的意义；(6)利用GPS-PWV资料首次对成都地区秋季可降水量的空间分布与循环周期进行了分析,研究得到了PWV与海拔高度间存在负相关关系,PWV季节内存在1/4季的变化特征,十月中旬是PWV由多到少的转折期,成都地区秋季降水过程主要集中在夜间,大气水汽总量的累积或释放与地面实际降水量有着较好的对应关系,可降水量的峰值出现的时间一般提前于降水强度极大值出现的时间等结论。更多还原

【Abstract】 Water vapor is a major component of the atmosphere and is also a greenhouse gas. Although it is a minor component of the atmosphere, water vapor has considerable impact on local atmospheric conditions due to its uneven spatial distribution and rapidly shifting concentrations. The water vapor is a component that changes most in the atmosphere, whose changing scale is finer than the wind and temperature. One of the basic problems of meteorology and weather forecasting is obtaining accurate measurements of the distribution and variation of water vapor in the atmosphere.First appearing in the1990s, global positioning system (GPS) techniques are new monitoring tools that provide mechanisms to estimate water vapor concentrations in the atmosphere. Due to high precision, high capacity, high spatial and temporal resolution, and ability for all-weather operation, and in obtaining water vapor data from the atmosphere at near real-time speeds, it has received much attention. National, sector, and local have invested substantial manpower and material resources in GPS techniques, and as a result, have accumulated a great quantity of original GPS data. However, the ground-based GPS water vapor monitoring network in China is still in its rudimentary stages and its detection techniques involve cross-integration of surveying and meteorological knowledge; several problems are yet to be further solved and studied.Using GPS delay to determine precipitable water vapor (PWV) in the atmosphere entails several problems:too many processing links and complicated techniques, which restrain the effective use of the ground-based GPS data; the inversion process of water vapor data by GPS; it lacks a unified standard of modeling and application of weighted average temperature; in the processing of ground-based GPS data, GPS-PWV cannot be determined if either the original GPS data or the ground meteorological data is missing; GPS-PWV in weather analysis is restricted to research on changes of PWV itself; and researches on the deep-seated reasons for water vapor change is rare. In response to these questions, this paper carries out a deep analysis and study, and arrives at the following conclusions.(1) The analysis determines that the two key problems in meteorological operation department’s current techniques of water vapor inversion by ground-based GPS data are the usage of software in processing GPS data and integrated treatment in the operation;(2) It focuses on the entire process of water vapor inversion by ground-based GPS data, qualitatively and quantitatively analyzes the main error sources, presents a solution to this error source, and calculates the range of error source reduction required to meet the accuracy demands of meteorological applications;(3) The article uses mathematical methods and intelligent algorithms to propose two complementary schemes for addressing the lack of measurement of precipitable water vapor. Through principle component analysis, the selection problems of fitting factors are resolved, and through a sensitivity test, the structural composition of the neural network model that best fits precipitable water volume is obtained;(4) As the fitting formula for the weighted average temperature is not unified, each region has its own problem of Tm calculation model. The article puts forward a Tm calculation model that can be applied in China, and this model achieves the requirement of having sufficient accuracy for use in meteorological applications;(5) It conducts a systematic analysis of the changing characteristics of GPS-PWV in two strong rainfalls in Jiangsu and Chongqing, assiduously researches the dynamic and thermodynamic processes in line with NCEP reanalysis data, conventional sounding data, ground meteorological data, and WRF numerical simulation results, and reveals the deep-seated reasons for the water vapor change that is the internal reason for PWV change, as well as deepens the significance of GPS-PWV as forecast index;(6) It uses GPS-PWV data for the first time to analyze the spatial distribution and cycle of GPS at Chengdu plain in autumn. The research establishes a negative correlation between PWV and the altitude, and reveals that PWV changes features in one quarter of a season, i.e., that mid-autumn is a turning point where PWV changes from more to less, where rainfall of Chengdu in autumn occurs mainly at night, and there is sufficient correspondence between the accumulation or release of the total amount of water vapor in atmosphere and the actual precipitation on ground, and where the precipitation peak usually appears earlier than the maximum of precipitation intensity.更多还原