顾及水汽时空变化参数的GNSS对流层层析建模方法研究

GNSS-based tropospheric tomography is a promising technique to acquire multi-dimensionally dynamic variation of atmospheric water vapor which plays a crucial role in the monitoring and forecasting of mesoscale disastrous weather. However, the tomographic technique still faces some problems: 1) the spatiotemporal distribution information of water vapor has not been fully considered in the traditional tomographic modeling, leading to an imprecise tomography model; and 2) unsatisfactory spatiotemporal resolution of the tomographic products caused by nonideal spatial distribution of the observational data and singleness of the modeling data source. Based on multi-source and multi-period data, this project aims to investigate the high-accuracy and high-spatiotemporal tomography modeling method considering the spatiotemporal variation parameters of water vapor. Specifically, this study will start with tomography grid discretization with an aim to construct high-accuracy numerical integration parameterized tomography model by considering the spatiotemporal variation parameters of water vapor. Then, this project will study the method of the water vapor tomography using multi-observational and multi-period data. This study will try to improve the ill-posed problem of the tomography inversion equation and to make the reconstruction of high-spatiotemporal water vapor field achievable. This research will also apply the tomographic products on mining the migration and transportation characteristics of the water vapor during heavy precipitation events. Heavy rain prediction model will be constructed using tomographic water vapor products combined with terrain, pressure, temperature and wind field data. The outcomes of the project will improve the accuracy and spatiotemporal resolution of the tomographic results. Besides, it also will help to develop the GNSS meteorology and expand the range of its applications.

GNSS对流层层析技术可获取水汽的多维动态变化信息，从而为监测和预报中尺度灾害天气提供重要科学依据。但当前层析技术还存在以下问题：1）常规层析建模中未充分考虑水汽时空分布信息，导致参数化模型不够精细；2）观测值空间分布不理想以及建模数据源单一致使层析产品时空分辨率不足。为此本项目拟基于多源多时段数据研究顾及水汽时空变化参数的高精度高时空分辨率层析建模方法。具体地，本项目将从层析网格离散化入手，顾及水汽时空分布参数，建立高精度数值积分参数化层析建模理论；研究融合多源多时段水汽数据的数值积分参数化层析算法，改善观测值空间分布结构以实现高时空分辨率水汽场的反演；利用层析产品挖掘短期临近暴雨预报中水汽运移规律，结合地形、气压、温度和风场数据构建暴雨预报模型并评估其可靠性。本项目成果将提高层析水汽产品的精度和时空分辨率，促进GNSS气象学的发展与应用。

水汽是地球大气的重要组成部分，在天气动力系统、大气环境科学、测绘科学与技术以及水文学等诸多领域发挥着至关重要的作用。大气中的水汽主要集中在对流层内，其高动态变化特性使得准确及时获取高时空分辨率水汽分布成为一项十分困难的任务。GNSS对流层层析技术可获取水汽的多维动态变化信息，拥有众多传统水汽探测手段所无可比拟的优势，从而为监测和预报中尺度灾害天气提供重要科学依据。但当前层析技术还存在以下问题：1）常规层析建模中未充分考虑水汽时空分布信息，导致参数化模型不够精细；2）观测值空间分布不理想以及建模数据源单一致使层析产品时空分辨率不足。本项目重点研究高精度数值积分参数化水汽层析理论和顾及水汽时空变化参数的高时空分辨率层析模型精化方法，具体研究内容包括：考虑水汽时空分布的数值积分参数化层析模型精细构建；高时空分辨率水汽层析建模中多源观测数据的融合；评估层析水汽产品在短期临近暴雨预报中的可靠性。研究成果表明：1）顾及水汽时空变化参数的数值积分参数化层析技术可为GNSS气象学、GNSS定位、数值天气预报等提供高精度大气水汽场分布信息；2）融合FY-4A等静止气象卫星观测数据可有效克服GNSS对流层层析方程的不适定问题，进而提升三维水汽产品反演精度；3）水汽层析产品在GNSS定位、大气河监测中的应用证实了层析技术的可靠性，拓展了层析技术的应用范围；4）基于湖南省层析水汽产品的神经网络降雨预报模型能够正确预测超过94%中雨及以上等级的降雨事件，验证了层析产品在大范围降雨落区预报的可行性。本项目对提升参数化对流层层析建模精度和多源水汽数据融合的层析技术具有重要的理论价值；也对中尺度极端天气中水汽多维变化规律的挖掘、数值天气模式同化、灾害天气监测预报能力的提升、卫星对地观测技术的对流层延迟改正等具有重要的应用价值和社会意义。