| 摘要: | 電波掩星(radio occultation)的技術一開始是運用在行星科學上，用於探測行星的大氣並得到其參數剖面。隨著全球定位系統(Global Positioning System, or GPS)的出現，並使用地球低軌道(low Earth orbit)衛星來接收全球定位系統衛星的訊號，使得電波掩星也可用於觀測全球的大氣與電離層。在2006年，台灣發射了六顆地球低軌道衛星用於電波掩星的任務，名為福衛三號星系計畫(Constellation Observing System for Meteorology, Ionosphere and Climate, or FORMOSAT-3/COSMIC)。在本論文中，福衛三號星系計畫的電波掩星資料將被用來發展與改進反演與模擬技術，以及分析電離層散塊E層(sporadic-E layer, or Es layer)的活動。本論文的主要動機在於提出能夠了解在訊號路徑上發生的大氣異常現象，對於所觀測到的訊號以及使用觀測訊號的反掩結果所造成的影響的方法，以便於在未來能夠提高反掩的精確度。本論文內容主要分為三階段來達到此目標： (1) 發展反掩程序來了解電波掩星的反掩過程； (2) 發展路徑覓跡法來了解訊號在大氣中的傳播過程； (3) 使用電離層散塊E層為一個大氣的異常現象的例子，使用所發展的路徑覓跡法來模擬電離層散塊E層對於電波掩星資料振幅的影響，並與實際觀測資料比較。 為了能夠了解大氣異常現象對於反掩結果的影響，首先要先發展電波掩星的反掩程序，本論文的第二章主要是描述自行開發的中央大學電波掩星(National Central University Radio Occultation, or NCURO)反演模式。此模式的主要重點在於經過開環迴路(open loop)之後超額相位的修正技術，以及訊號品質標準的評估。當使用開環迴路時，在地球低軌道衛星所接收到的超額相位會被訊號中的衛星星曆資訊所影響。除此之外，當掩星訊號穿過越低層大氣時，訊號的品質也隨著穿透高度遞減而變差，最終將會因為品質太差而無法反掩，為了解決這個問題，訊號的離散度(degree of unclearness)將被建立為新的品質標準來取代訊號的訊噪比。最後的反掩結果將與美國大氣科學大學聯盟(University Corporation for Atmospheric Resarch, or UCAR)的福衛三號資料分析及檔案中心(COSMIC Data Analysis and Archival Center, or CDAAC)的結果以及屏東探空氣球的觀測結果作比較。 除了要了解大氣異常現象對於反掩結果的影響之外，大氣異常現象對於訊號傳播所造成的影響也需要了解，本論文的第三章描述一個新的三維路徑覓跡法(ray-tracing method)，此路徑覓跡法能夠使模擬訊號在大氣中的傳播過程更接近真實的狀況。在這個方法中，包含兩個瞄準模式(aiming algorithm)，來確保訊號的路徑的起點與終點能夠落在預定的範圍內。在之前的研究中，路徑覓跡法主要是用來驗證反掩模式以及對資料同化的結果作評估，為了簡化模擬的過程，地球的大氣常常假設為球型對稱，並且確切的全球定位系統與地球低軌道衛星的位置也沒被考慮在內，這兩個假設使電波掩星的模擬不符合真實的狀況。在本論文中的路徑覓跡法中，地球的外型假設為?球體，大氣結構的部分則是引入歐洲中尺度天氣預報中心(European Centre for Medium-Range Weather Forecasts, or ECMWF)的分析模式，而兩個瞄準模式將用來決定訊號路徑的起始方向，使其起始於預定的全球定位系統衛星的位置，並結束於地球低軌道衛星位置的附近。最後，此路徑覓跡法藉由三個不同的大氣結構來進行測試及驗證: 1. 理想的球形對稱結構； 2. 使用歐洲中尺度天氣預報中心的分析模式並假設地球外型為?球體； 3. 外加重力波與對流層頂的擾動於歐洲中尺度天氣預報中心的分析模式。使用理想的球形對稱大氣結構時，模擬結果與實際的狀況間的誤差小於0.6%；使用第二種大氣結構時，所有的訊號路徑結束點最後都能夠收斂在預設的範圍內；使用第三種大氣結構時，模擬結果都顯示外加擾動所呈現的特徵。 本論文的第四章以電離層散塊E層為一個大氣異常現象的例子，使用第三章的路徑覓跡法來模擬訊號傳播經過電離層散塊E層之後所接收到的振幅，得到振福與傳播路徑中的電子濃度之間的關係，最後與觀測資料比較。除此之外，並使用從2008年中到2011年中的福衛三號電波掩星資料進行分析。根據模擬結果，訊號經過散塊E層的訊噪比可分為多層型(multiple-layer-type)與單層型(single-layer-type)，並且可以由三年的資料得到兩種不同種類的散塊E層的全球性分布，其中四個季節的多層型散塊E層將與從水平風模式(Horizontal Wind Model, or HWM07)所得到的四季風切作比較。除此之外，多層型與單層型散塊E層四季的全球性高度變化相似，但是電子濃度強度的分布不同，而單層型散塊E層的電子濃度強度與E層的全球電子濃度極大值(NmE)分布相似，而E層的全球電子濃度極大值則與太陽的天頂角有直接的關係。 在本論文中，以電離層散塊E層為一個大氣異常現象的例子，使用路徑覓跡法來模擬訊號傳播經過電離層散塊E層時在觀測資料上會看到的現象，並與觀測資料比較，藉此了解對於掩星訊號所造成的影響，在日後藉由反掩模式便可了解對於反掩結果的影響，在未來便可藉由這些資訊發展並改進反掩模式來增加反掩結果的精確度。 Radio occultation (RO) technique, which has been used in planetary science, is a method to obtain the parameter profiles of the atmosphere. With the advent of Global Positioning System (GPS), to observe the global atmosphere and ionosphere of the Earth becomes possible by using low Earth orbit (LEO) satellites to receive the signal of GPS. In 2006, Taiwan launched six LEO satellites as a RO constellation mission, known as Constellation Observing System for Meteorology, Ionosphere and Climate (FORMOSAT-3/COSMIC, or F-3) and the data of F-3 is used in this thesis. The motivation of this thesis is bring up the methods to understand the influence of atmospheric anomaly, which occurs on the signal trajectory, on the observational and retrieval results in order to increase the retrieval accuracy in the future. In order to reach the motivation, three steps are contained in this thesis: (1) A retrieval algorithm is developed to understand the process of the RO data retrieval; (2) A ray tracing model for GPS signal is proposed to understand the signal propagating process in the atmosphere; (3) The ionospheric sporadic-E (Es) layer is used to be an example of atmospheric anomaly to simulate its influence on the amplitude profile by using ray tracing model and compare with the observational data. In order to know the influence of atmospheric anomaly on the observational and retrieval results, a retrieval algorithm should be developed first. A retrieval algorithm, NCURO (National Central University Radio Occultation), is developed in this thesis to obtain the information of atmospheric parameter profiles. The focus of the algorithm development is on the correction of the excess phase of the signal received with open loop (OL) technique, and the criteria for assessment of the data quality. When the OL is activated, the excess phase of the GPS signal is modulated with navigation messages of satellites. In our algorithms, two methods are incorporated to recover the excess phase. Moreover, as the altitude of the received signal decreases, the quality of the GPS signal generally deteriorates, and eventually the signal is too noisy to be processed. In order to assess the quality of the signal, instead of the signal-to-noise ratio (SNR), the degree of unclearness is defined and used in the algorithm. In this thesis, the algorithm including the phase correction methods and the criteria for the quality assessment will be described. The data retrieval using the algorithm will be compared with those obtained from CDAAC at UCAR and Pingtung radiosonde measurement. Some intermediate results of the NCURO algorithm will also be demonstrated. Except considering the atmospheric anomaly influenced on the observational and retrieval results, the influence of atmospheric anomaly on the signal propagation should be known. A three-dimensional ray tracing model for GPS signal is proposed to make simulation conform to the realistic RO signal propagation. In the model, two aiming algorithms are developed to ensure the initial and end points of the ray trajectory located in the prescribed region. In past studies, the ray tracing algorithms are often used to support the retrieval algorithms and assess the impact of data assimilation. The ray tracing techniques applied to the RO signal simulation usually assumed a spherically symmetric atmosphere for simplicity. Also, the exact locations of of GPS and LEO satellites are not considered in the simulation. These two assumptions make the simulation unrealistic for the GPS signal propagation in RO technique. In the proposed model, the shape of the Earth is assumed as an ellipse. The information from European Centre for Medium-Range Weather Forecasts (ECMWF) analysis is used to setup the atmosphere in the simulation. Two aiming algorithms are developed to determine the initial propagating direction of GPS signal. The aiming algorithms will also make the simulated signal start from the prescribed GPS satellite position and end in the close vicinity of the LEO satellite position. The proposed model is examined and demonstrated in the designed simulation using three atmospheric structures: the ideal structure of spherical symmetry, the ECMWF analysis with the consideration of the Earth’s flattening, and the artificial perturbation added in ECMWF analysis which allows consideration of gravity waves and the tropopause. For the ideal atmospheric structure, the fractional difference between real and simulated refractivity results is less than 0.6%. For the ECMWF analysis and the consideration of the Earth’s flattening, all the simulated end points are located in the prescribed region. And for the artificial perturbation added in ECMWF, the simulated results show the corresponding characteristics of the artificial perturbation. The signal propagating through the ionospheric Es layer, which is an example of atmospheric anomaly, is simulated by using the ray tracing model. The relation between the amplitude of RO signals and the electron density profiles of the ionosphere is simulated and compare with the observational data. Furthermore, the RO data recorded in the time period from mid-2008 to mid-2011 are used for the analysis. Based on the simulation results, the multiple-layer-type (MLT) and the single-layer-type (SLT) Es layers which are defined by the shape of SNR, are used to analyze the global distribution of Es layer. The seasonal MLT Es layer is compared with the seasonal wind shear, which is obtained from the Horizontal Wind Model (HWM07). Furthermore, the seasonal MLT Es layer is compared with the SLT Es layer, and the global altitude distributions of MLT and SLT Es layers are similar while the magnitude distributions are different. Unlike the MLT Es layer, the global distribution of the SLT Es layer is similar to the distribution of E region peak electron density (NmE), which is related to the solar zenith angle. In this thesis, the ionospheric Es layer is an example to show the influence of atmospheric anomaly on RO data by using ray tracing model. And the influence of the atmospheric anomaly on retrieval results also can be anticipated by using the retrieval algorithm. With the information of the influence caused by atmospheric anomaly, the retrieval algorithm can be corrected and developed to retrieve atmospheric parameter profiles with more accuracy in the future. |
