利用支持向量机预测大磁暴期间Dst指数的变化

利用支持向量机（SVM）模型对大磁暴期间Dst指数进行预报研究.以1995-2014年期间的80次大磁暴（Dst≤-100nT）事件共2662组观测数据为研究对象,以对应时间的太阳风参数为模型输入参数,同时建立了神经网络模型和线性机模型进行对比,并利用交叉验证提高预测结果的可靠性.为比较不同模型的预测效果,选用相关系数（CC）、均方根误差（RMS）、磁暴期间Dst指数最小值预测结果的平均绝对误差以及Dst指数最小值出现时间预测结果的平均绝对误差等统计量作为对比参数.结果显示SVM模型的预测效果最好,其中相关系数为0.89,均方根误差为24.27nT,所有磁暴事件的最小Dst值预测平均绝对误差为17.35nT,最小Dst值出现时间的预测平均绝对误差为3.2h.为进一步检验模型对不同活动水平磁暴预报效果的可能差异,将所有磁暴事件分为大磁暴（-200 相似文献   

行星际扰动与不同级别磁暴强度关系的研究

利用1997-2004年间ACE卫星太阳风观测的时均值数据和相应的Dst指数,针对Dstmin≤-50 nT的磁暴,分析了行星际参数(Bz,Ey,v,Pk,|B|,ε'=vxB2zsin4(θ/2))与Dst指数的相关关系.验证了Ey,Bz与Dst指数的良好相关性;按磁暴强度的不同,发现磁暴强度越大,行星际参数与磁暴强度(Dstmin)的相关性就越好.对于中等磁暴(-100 nT＜Dstmin≤-50 nT),行星际参数与磁暴强度的相关系数不高.如果把磁暴分为两个档次,即-150 nT＜Dstmin≤-50 nT的磁暴和Dstmin≤-150 nT的磁暴,计算结果表明,ε'与Dst指数的相关性是最好的.在诸多行星际参数中,就单一因素来说,Ey对磁暴强度影响最大,Bz对磁暴强度影响次之.      

磁暴期间中国中低纬电离层不规则体与扰动分析

2017年9月8日发生了一次强磁暴,Kp指数最大值达到8.利用区域电离层格网模型（Regional Ionosphere Map,RIM）和区域ROTI（Rate of TEC Index）地图,分析了磁暴期间中国及其周边地区电离层TEC扰动特征和低纬地区电离层不规则体的产生与发展情况,同时利用不同纬度IGS（International GNSS Service）测站BJFS（39.6°N,115.9°E）,JFNG（30.5°N,114.5°E）和HKWS（22.4°N,114.3°E）的GPS双频观测值,获取各测站的ROTI和DROT（Standard Deviation of Differential ROT）指数变化趋势.结果表明:此次磁暴发生期间电离层扰动先以正相扰动为主,主要发生在中低纬区域,dTEC（differential TEC）最大值达到14.9TECU,随后电离层正相扰动逐渐衰减,在低纬区域发生电离层负相扰动,dTEC最小值达到-7.2TECU;在12:30UT-13:30UT时段,中国南部低纬地区发生明显的电离层不规则体事件;相比BJFS和JFNG两个测站,位于低纬的HKWS测站的ROTI和DROT指数变化更为剧烈,这表明电离层不规则体结构存在纬度差异.      

一种基于白谱法的电离层天气扰动指数

基于一种电离层扰动提取方法——白谱法,利用IGS提供的电离层TEC网格数据,获得电离层J_s指数、J_r指数和J_p指数,分别反映单站、纬度圈（沿经度积分）及行星际尺度下的电离层天气扰动状态.在2015年3月的一次磁暴过程中,J_s指数、J_r指数及J_p指数均很好地反映出电离层响应地磁暴的过程,磁暴前后J_p指数与Dst指数相关系数达到-0.72;J_s图从二维角度很好地表征了电离层天气的扰动过程.在此基础上,统计分析了2011——2014年J_p指数与Dst指数的相关性,结果表明:限定J_p≥2,J_p指数与对应时间Dst指数的相关系数为-0.67;限定J_p≥3,二者相关系数更高,达到-0.87.通过分析不同J_p指数阈值下不同等级磁暴的次数,发现J_p指数可以很好地反映磁暴下的电离层整体扰动,为指示电离层天气状态提供了可能的参数.      

2000年4月磁暴期间武汉地区F电离层突然抬升现象

2000年4月6-7日的大磁暴(Dst最小值达到-317nT),急始(SSC)在6日1640UT左右出现,随后磁暴主相开始,约在7日0013UT进入恢复相.磁暴主相前期武汉地区F区电离层出现突然抬高现象,在2h内h′F和h_mF₂分别较暴前日增加约200km.此期间台湾中沥也几乎同时出现了F区电离层突增现象.磁暴主相前期f₀F₂较暴前日下降1.6MHz,其变化幅度在f₀F₂逐日变化起伏范围内,但7日f₀F₂最大值明显低于4-6日暴前水平,并伴随着波动特征.认为此次磁暴主相前期武汉地区F区电离层突增现象,可能和夜间磁暴常出现的扰动东向电场有关.      

太阳活动下降年昆明地区TEC变化特征与IRI-2020模拟比较分析

利用昆明低纬度测站（24.7°N,102.9°E，磁纬15.1°N）2016-2019年的观测数据和最新版的国际参考电离层（IRI-2020）模拟结果，对昆明地区电离层总电子含量（TEC）在太阳活动下降年期间的变化特征及与模型输出进行对比研究。结果表明,昆明TEC存在明显的春秋高值、夏冬低值的半年异常；白天高值、夜间低值的日变化特点突出，日峰值出现在06:30-08:00 UT（约13:00-15:00 LT）;TEC随太阳活动减弱而明显下降，年平均峰值在2016-2019年分别为48,33,27,24 TECU；日峰值TEC与F10.7存在显著相关，月均值相关系数达到0.86，而与Ap指数则表现为弱相关；IRI-2020能较好地模拟昆明地区TEC的季节变化，但与观测值存在较大差异；均方根偏差值多集中在2～15 TECU，相对偏差百分比值主要在–85%～50%范围变化。对比结果表明IRI-2020的预测精度仍有待提高。     

基于IGS电离层TEC格网的扰动特征统计分析

电离层总电子含量（TEC）是研究空间天气特性的重要参量,通过分析电离层TEC,可以了解空间环境的变化特征.利用IGS提供的1999—2016年全球电离层TEC格网数据,按照地磁纬度将全球划分为高、中、中低、低磁纬四个区域,计算不同区域的电离层扰动;利用大量统计数据选取电离层扰动事件的判定阈值,分析电离层扰动与太阳活动、时空之间的关系;计算电离层扰动指数与地磁活动之间的相关系数.结果显示:电离层扰动与太阳活动变化具有较强的正相关特性.在太阳活动低年,电离层扰动事件发生的概率约为1.79%,在太阳活动高年发生扰动的概率约为10.18%.在空间分布上,无论是太阳活动高年还是低年,高磁纬地区发生扰动事件的概率均大于其他磁纬出现扰动事件的概率.计算得到的中磁纬和中低磁纬地区电离层扰动指数与全球地磁指数Ap的相关系数分别为0.57和0.56,说明电离层扰动指数与Ap具有较好的相关关系;高磁纬电离层扰动指数与Ap的相关系数为0.44;低磁纬扰动指数与Ap的相关系数为0.39.以上结果表明,不同区域电离层扰动与全球地磁指数Ap的相关性不同,测定区域地磁指数可能会提高与电离层扰动的相关性.      

海南地区强磁暴期间强区域扩展F事件研究

利用2004年海南DPS-4数字测高仪观测到的强区域扩展F(SSF)数据,分析研究了5个强磁暴(Dst＜-100 nT)事件期间海南SSF的响应特征.结果发现,在海南地区,5个强磁暴事件中有3个磁暴Dst最小值位于2200-0200 LT之间,在磁暴主相及恢复相初期均出现了SSF现象,这种触发作用可能源于磁层直接渗透电场的作用,而另两个磁暴Dst最小值均发生在白天,一个SSF现象出现在磁暴的恢复相晚间,另一个SSF现象出现在超强磁暴的初相晚间,后者可能由该超强磁暴的急始造成的直接渗透电场所触发;5个强磁暴期间发生的SSF现象或者仅出现在午夜前,或者先出现在午夜前并持续到午夜后;同时,还就这些观测结果与Dabas等人有关磁暴对ESF影响的结论进行对比和讨论.      

第23太阳活动周太阳极紫外辐射与磁暴相关性分析

统计第23个太阳活动周内中等及以上强度（Dst_min<-50nT）的磁暴事件,线性拟合分析磁暴主相DDst_min和达到DDst_min前一个表征太阳极紫外辐射强度的F_10.7之间的相关性.结果表明:随着太阳极紫外辐射增强,DDst_min<-50nT的磁暴出现的总数增多,在弱、中等和强太阳极紫外辐射条件下,其数量分别为56,84和85;随着太阳极紫外辐射增强,强磁暴（-200nT ≤ Dst_min<-100nT）和大磁暴（Dst_min<-200nT）发生的数量和相对发生率呈增长趋势,尤其是大磁暴数目（1,4,12）和相对发生率（1.79%,4.76%,14.12%）明显呈增长趋势;大磁暴（|Dst_min|）与太阳极紫外辐射（F_10.7）之间存在中度正相关关系,其相关系数为0.532,并且主要体现在大磁暴（|Dst_min|）与强太阳极紫外辐射（F_10.7）之间的中度正相关性,其相关系数为0.582.大磁暴与强太阳极紫外辐射之间的相关性可为空间天气预报提供参考依据.      

2008—2009年电离层对重现型地磁活动的响应

利用2008—2009年的GPS TEC数据,分析了电离层对冕洞引起的重现型地磁活动的响应. 结果表明,在太阳活动低年,电离层TEC表现出与地磁 ap指数（采用全球3h等效幅度指数ap来表征）和太阳风速度相似的9天和13.5天短周期变化,表明TEC的这种短周期特性主要与重现型地磁活动相关. 地磁纬度和地方时分析表明,夜间高纬地区正负相扰动明显,中低纬地区则以正相扰动为主,较大的TEC变幅主要发生在南北半球高纬地区,夜间南半球高纬地区TEC变化相对ap指数变化有相位延迟. 白天中低纬地区正负相扰动明显,TEC短周期变化与ap指数变化相位基本一致. 2008年TEC的9天和13.5天周期变化幅度大于2009年.      

Investigation of ionospheric response to two moderate geomagnetic storms using GPS–TEC measurements in the South American and African sectors during the ascending phase of solar cycle 24

The responses of the ionospheric F region using GPS–TEC measurements during two moderate geomagnetic storms at equatorial, low-, and mid-latitude regions over the South American and African sectors in May 2010, during the ascending phase of solar cycle 24, are investigated. The first moderate geomagnetic storm studied reached a minimum Dst value of −64 nT at 1500 UT on 02 May 2010 and the second moderate geomagnetic storm reached a minimum Dst value of −85 nT at 1400 UT on 29 May 2010. In this paper, we present vertical total electron content (VTEC) and phase fluctuations (in TECU/min) from Global Positioning System (GPS) observations from the equatorial to mid-latitude regions in the South American and African sectors. Our results obtained during these two moderate geomagnetic storms from both sectors show significant positive ionospheric storms during daytime hours at the equatorial, low-, and mid-latitude regions during the main and recovery phases of the storms. The thermospheric wind circulation change towards the equator is a strong indicator that suggests an important mechanism is responsible for these positive phases at these regions. A pre-storm event that was observed in the African sector from low- to the mid-latitude regions on 01 May 2010 was absent in the South American sector. This study also showed that there was no generation or suppression of ionospheric irregularities by storm events. Therefore, knowledge about the suppression and generation of ionospheric irregularities during moderate geomagnetic storms is still unclear.     

Geomagnetic storm effects on the occurrences of ionospheric irregularities over the African equatorial/low-latitude region

The study investigated the effects of intense geomagnetic storms of 2015 on the occurrences of large scale ionospheric irregularities over the African equatorial/low-latitude region. Four major/intense geomagnetic storms of 2015 were analyzed for this study. These storms occurred on 17th March 2015 (?229?nT), 22nd June 2015 (?204?nT), 7th October 2015 (?124?nT), and 20th December 2015 (?170?nT). Total Electron Content (TEC) data obtained from five African Global Navigation Satellite Systems (GNSS) stations, grouped into eastern and western sectors were used to derive the ionospheric irregularities proxy indices, e.g., rate of change of TEC (ROT), ROT index (ROTI) and ROTI daily average (ROTI_AVE). These indices were characterized alongside with the disturbance storm time (Dst), the Y component of the Interplanetary Electric Field (IEFy), polar cap (PC) index and the H component of the Earth’s magnetic field from ground-based magnetometers. Irregularities manifested in the form of fluctuations in TEC. Prompt penetration of electric field (PPEF) and disturbance dynamo electric field (DDEF) modulated the behaviour of irregularities during the main and recovery phases of the geomagnetic storms. The effect of electric field over both sectors depends on the local time of southward turning of IMF Bz. Consequently, westward electric field inhibited irregularities during the main phase of March and October 2015 geomagnetic storms, while for the June 2015 storm, eastward electric field triggered weak irregularities over the eastern sector. The effect of electric field on irregularities during December 2015 storm was insignificant. During the recovery phase of the storms, westward DDEF suppressed irregularities.     

Study of TEC changes during geomagnetic storms occurred near the crest of the equatorial ionospheric ionization anomaly in the Indian sector

GPS satellites data obtained at Bhopal (23.16° N, 77.36° E, geomagnetic latitude 14.23° N) India were analyzed to study the TEC changes during several geomagnetic storms (−300 nT < Dst < −50 nT) occurred in 2005–2007. We had segregated the storms according to the Dst value, i.e. moderate storms (−100 nT < Dst ? −50 nT), strong storms (−150 nT < Dst < −100 nT), and severe storms (Dst less than −150 nT). Total of 21 geomagnetic storms (10 moderate, 9 strong, 2 severe) are considered for the present study. Deviation in vertical total electron content (VTEC) during the main phase of the storm was found to be associated with the prompt penetration of electric field originated due to the under-shielding and over-shielding conditions for almost all geomagnetic storms discussed in this paper. For most of the storms VTEC shows the positive percentage deviation during the main phase while it shows positive as well as the negative deviation during the recovery phase of the storms. The −80% deviation in VTEC was found for geomagnetic storm occurred on July 17, 2005 and the negative trend continued for recovery phase of the storm. This was mainly due to the thermospheric composition changes by Joule heating effect at auroral latitudes that generate electric field disturbance at low latitudes. Traveling ionospheric disturbances (TIDs) were responsible for the formation of wave like nature in VTEC for the storms occurred on May 15, 2005, whereas it was not observed for storm occurred on August 24, 2005.     

Impact of geomagnetic storm of September 7–8, 2017 on ionosphere and HF propagation: A multi-instrument study

The present study reveals the features of ionospheric parameters variations during the geomagnetic storm of September 7–8, 2017. In particular, parameters of vertical (foF2, foEs) and oblique ionospheric sounding (MOF, modes), absorption level, Total Electron Content (TEC) and particle fluxes at high altitudes were under analysis. The storm was characterized by two Dst-index mimima and can be considered as a sequence of two storms: first - with Dstmin?=??142?nT at 02 UT on September 8th and second - with Dstmin?=??122?nT and at 15 UT on September 8th. It was found that these two storms had different impacts on the ionosphere and HF propagation at mid- and high-latitudes of Northern Hemisphere. The signals of vertical and oblique ionospheric sounding were present in all ionograms before the first storm. Further, at the maximum of the first storm these signals were totally absorbed. Then, before the second storm and during its maximum the signals were detected again in the ionograms due to the low absorption. GOES satellite data showed the significant burst of electrons and protons only during the first storm and small particle fluxes - during the second storm. This feature was also confirmed with GPS data: TEC increased during the first storm and decreased during the second storm.     

Effect of severe geomagnetic storm conditions on atomic oxygen greenline dayglow emission in mesosphere

Severe geomagnetic storms and their effects on the 557.7 nm dayglow emission are studied in mesosphere. This study is primarily based on photochemical model with the necessary input obtained from a combination of experimental observations and empirical models. The model results are presented for a low latitude station Tirunelveli (8.7°N, 77.8°E). The volume emission rates are calculated using MSISE-90 and NRLMSISE-00 neutral atmospheric models. A comparison is made between the results obtained from these two models. A positive correlation amongst volume emission rate (VER), O, O₂ number densities and Dst index has been found. The present results indicate that the variation in emission rate is more for MSISE-90 than in NRLMSISE-00 model. The maximum depletion in the VER of greenline dayglow emission is found to be about 30% at 96 km during the main phase of the one of the geomagnetic storms investigated in the case of MSISE-90 (which is strongest with Dst index −216 nT). The O₂ density decreases about 22% at 96 km during the main phase of the same geomagnetic storm.The NRLSMSISE-00 model does not show any appreciable change in the number density of O during any of the two events. The present study also shows that the altitude of peak emission rate is unaffected by the geomagnetic storms. The effect of geomagnetic storm on the greenline nightglow emission has also been studied. It is found that almost no correlation can be established between the Dst index and variations in the volume emission rates using the NRLMSISE-00 neutral model atmosphere. However, a positive correlation is found in the case of MSISE-90 and the maximum depletion in the case of nightglow is about 40% for one of the storms. The present study shows that there are significant differences between the results obtained using MSISE-90 and NRLMSISE-00.     

Midlatitude ionospheric F2-layer response to eruptive solar events-caused geomagnetic disturbances over Hungary during the maximum of the solar cycle 24: A case study

In our study we analyze and compare the response and behavior of the ionospheric F2 and of the sporadic E-layer during three strong (i.e., Dst?<??100nT) individual geomagnetic storms from years 2012, 2013 and 2015, winter time period. The data was provided by the state-of the art digital ionosonde of the Széchenyi István Geophysical Observatory located at midlatitude, Nagycenk, Hungary (IAGA code: NCK, geomagnetic latitude: 46.17° geomagnetic longitude: 98.85°). The local time of the sudden commencement (SC) was used to characterize the type of the ionospheric storm (after Mendillo and Narvaez, 2010). This way two regular positive phase (RPP) ionospheric storms and one no-positive phase (NPP) storm have been analyzed. In all three cases a significant increase in electron density of the foF2 layer can be observed at dawn/early morning (around 6:00 UT, 07:00 LT). Also we can observe the fade-out of the ionospheric layers at night during the geomagnetically disturbed time periods. Our results suggest that the fade-out effect is not connected to the occurrence of the sporadic E-layers.     

Latitudinal features of Total Electron Content over the African and European longitude sector following the St. Patrick’s day storm of 2015

GNSS TEC values have been obtained from 18 stations distributed from the magnetic equator to nearly 80°N magnetic dip in the African and west-European longitude sector corresponding to the March 17–18, 2015 geomagnetic storm. Significantly depleted ionosphere have been observed at stations north of 50°N geographic on March 18, 2015 following the above storm over a longitude swath 11.9°–21°E covering the Eastern Africa and Western European longitude sector. High ROTI values were noted on March 17^th at locations around 80°N magnetic dip. Two prominent peaks in PCN were noted around 09:00 UT and 14:00 UT on March 17, 2015 and around 15:00 UT on March 18, 2015. Daytime thermospheric (O/N₂) ratio was markedly less on March 18th at latitudes above 60°N geographic which is suggested to be the major driver behind depleted high latitude ionosphere during the recovery phase of the storm on March 18, 2015.     

磁暴数统计特性研究

Dst是一个表征磁暴强度的空间天气指数. 通过统计1957-2008年 发生的中等磁暴(-100<Dst≤ -50nT)和强磁暴(Dst ≤ -100nT)在太阳活动周上升年、极大年、下降年和极小年的时间分布情 况, 分析其随季节变化的统计特性, 进而讨论了引起磁暴的原因. 结果表明, 对于同一太阳活动周, 极大年地磁暴发生次数远大于极小年地磁暴的发生次数, 这与太阳黑子数的变化趋势是一致的; 通常太阳活动周强磁暴出现双峰结构, 而第23周中等磁暴出现双峰结构, 强磁暴则出现三峰结构, 这可能与1999 年强 磁暴发生次数异常少, 使1998年凸显出来的现象有关; 磁暴主要发生在分季, 随着Dst指数的增加, 磁暴发生次数明显增加.      

磁暴期间电离层扰动的GPS台网观测分析

给出了一种利用GPS台网观测获取TEC快速变化的计算方法，并将该方法用于东亚一澳大利亚扇区的GPS台网观测数据，分析了2000年7月14—18H大磁暴期间的电离层响应，揭示出电离层暴期间赤道异常峰的压缩和移动等特性．计算结果表明，在站点分布不均匀、原始观测数据不足且随时间跳变等多种不利因素的影响下，这种新的算法仍能保持很好的计算稳定性，并能快速地提取给定时空范围内的三维TEC短时变化的特征，适用于研究电离层暴等情况下引起的TEC扰动．