一种非线性系统集员辨识算法

针对带有未知有界噪声的非线性动态系统的鲁棒辨识问题,提出了一种新的非线性动态系统的集员辨识算法.利用径向基函数神经网络的逼近能力,根据系统的输入输出数据,选用径向基函数神经网络对未知非线性系统建模.径向基函数神经网络的中心被确定之后,考虑到建模误差与系统噪声有界,利用径向基函数神经网络为参数线性模型的特点,使用参数线性集员辨识算法辨识径向基函数神经网络的输出权值.由于集员辨识算法所得到的是网络输出权值的集合估计,在系统运行过程中,可以很方便地利用所建模型预测实际系统的输出范围.仿真表明,集员辨识算法辨识网络的输出权值比最小二乘法较少的受未知动态系统噪声分布的影响.     

Advances in space research nonlinear dependence of Dst and AE indices on the electric field of magnetic clouds

Using the Dst and AE geomagnetic index values and parameters of interplanetary magnetic field and solar wind we have examined the geoeffectiveness of transient ejections in the solar wind, namely, magnetic clouds and high-speed streams. It is found that for magnetic clouds the dependences of indices on the solar wind electric field are nonlinear of different kind. In contrast to magnetic clouds, the dependence of Dst and AE geomagnetic index values on the solar wind electric field agrees closely with the linear one for high-speed streams. We suggest approximating formulas to describe dependences obtained taking into account the relation of the electric field transpolar potential to the electric field and dynamic pressure of the solar wind. We suppose that the interplanetary magnetic field fluctuations also contribute to these dependences.     

Interplanetary shocks and geomagnetic activity during solar maximum (2000) and solar minimum (1995–1996)

Plasma and magnetic field parameter variations through fast forward interplanetary shocks were correlated with the peak geomagnetic activity index Dst in a period from 0 to 3 days after the shock, during solar maximum (2000) and solar minimum (1995–1996). Solar wind speed (V) and total magnetic field (B_t) were the parameters with higher correlations with peak Dst index. The correlation coefficients were higher during solar minimum (r² = 56% for V and 39% for B_t) than during solar maximum (r² = 15% for V and 12% for B_t). A statistical distribution of geomagnetic activity levels following interplanetary shocks was obtained. It was observed that during solar maximum, 36% and 28% of interplanetary shocks were followed by intense (Dst  −100 nT) and moderate (−50  Dst < −100 nT) geomagnetic activity, whereas during solar minimum 13% and 33% of the shocks were followed by intense and moderate geomagnetic activity. It can be concluded that the upstream/downstream variations of V and B_t through the shocks were the parameters better correlated with geomagnetic activity level, and during solar maximum a higher relative number of interplanetary shocks can be followed by intense geomagnetic activity than during solar minimum. One can extrapolate, for forecasting goals, that during a whole solar cycle a shock has a probability of around 50% to be followed by intense/moderate geomagnetic activity.     

AL index dependence on the solar wind input revisited

Estimates of the geomagnetic indices made with real-time solar wind measurements form the basis of many space weather forecast techniques. We analyze 20 years of hourly AL and OMNI solar wind data to determine geomagnetic importance of various solar wind and IMF parameters. Besides the solar wind driver of primary importance (VB_s), the first-order contributions, significantly increasing the quality of the model are: solar wind velocity, 2 h of solar wind prehistory and 1 h of AL history. The factors of secondary importance, marginally contributing to overall statistical quality, are IMF B_y, solar wind density, and IMF fluctuations. The dynamic pressure is geomagnetically effective only if the pressure is lower than the average. Modelling of the same data set with an artificial neural network (ANN) confirmed our selection of important factors. Statistically the ANN model was just marginally better than our analytic expression . The AU index dependence is principally different from AL in several respects; therefore modelling of the AE composite index is physically misleading.     

Interplanetary shocks and sudden impulses during solar maximum (2000) and solar minimum (1995–1996)

In this work a study is performed on the correlation between fast forward interplanetary shock parameters at 1 Astronomical Unit and sudden impulse (SI) amplitudes in the H-component of the geomagnetic field, for periods of solar activity maximum (year 2000) and minimum (year 1995–1996). Solar wind temperature, density and speed, and total magnetic field, were taken to calculate the static pressures (thermal and magnetic) both in the upstream and downstream sides of the shocks. The variations of the solar wind parameters and pressures were then correlated with SI amplitudes. The solar wind speed variations presented good correlations with sudden impulses, with correlation coefficients larger than 0.70 both in solar maximum and solar minimum, whereas the solar wind density presented very low correlation. The parameter better correlated with SI was the square root dynamic pressure variation, showing a larger correlation during solar maximum (r = 0.82) than during solar minimum (r = 0.77). The correlations of SI with square root thermal and magnetic pressure were smaller than with the dynamic pressure, but they also present a good correlation, with r > 0.70 during both solar maximum and minimum. Multiple linear correlation analysis of SI in terms of the three pressure terms have shown that 78% and 85% of the variance in SI during solar maximum and minimum, respectively, are explained by the three pressure variations. Average sudden impulse amplitude was 25 nT during solar maximum and 21 nT during solar minimum, while average square root dynamic pressure variation is 1.20 and 0.86 nPa^1/2 during solar maximum and minimum, respectively. Thus on average, fast forward interplanetary shocks are 33% stronger during solar maximum than during solar minimum, and the magnetospheric SI response has amplitude 20% higher during solar maximum than during solar minimum. A comparison with theoretical predictions (Tsyganenko’s model corrected by Earth’s induced currents) of the coefficient of sudden impulse change with solar wind dynamic pressure variation showed excellent agreement, with values around 17 nT/nPa^1/2.     

地球磁场对太阳风扰动响应的研究

分别对行星际激波、太阳风动压增大事件和减小事件的地球磁场响应进行了比较. 分析结果表明, 同步轨道磁场对太阳风扰动在向阳面产生较强的正响应, 在背阳面 响应较弱且有时会出现负响应, 地磁指数SYM-H对太阳风扰动的响应为正响应. 同时还得出, 向阳侧同步轨道磁场响应幅度d Bz与地磁指数响应幅度d SYM-H、上下游动压均方差均具有较好的相关性. 地磁指数响应幅度与同步轨道磁场响应幅度相关关系在激波和动压增大事件中具有一致性, 动压减小事件出 现明显差异, 这说明激波和动压增大事件在影响地球磁场方面具有某种共性.      

太阳风扰动对极光电急流和环电流指数的影响

利用WIND卫星的太阳风观测数据和地磁活动指数, 研究了太阳风扰动对环电流SYM-H指数, 西向极光电急流AL指数和东向极光电急流AU指数的影响. 结果表明, 太阳风动压增长和减少能够同步或延迟引起地磁活动指数的强烈扰动, 其包括环电流指数的上升, 西向极光电急流指数的下降和东向极光电急流指数的上升. 太阳风动压的突然剧烈增加还能够触发超级亚暴和大磁暴. 太阳风动压脉冲引起的地磁效应具有复杂的表现形式, 这说明太阳风动压脉冲的地磁效应不仅与太阳风动压脉冲大小和持续时间有关, 还与磁层本身所处的状态有关. 时间尺度较长, 消耗能量较大的磁暴只有大的持续时间较长的太阳风动压脉冲才能激发.      

Classification of solar wind structures and intense geomagnetic storm alarms with self-organizing maps

Intense geomagnetic storms (D_st < −100 nT) usually occur when a large interplanetary duskward-electric field (with Ey > 5 mV m⁻¹) lasts for more than 3 h. In this article, a self-organizing map (SOM) neural network is used to recognize different patterns in the temporal variation of hourly averaged Ey data and to predict intense storms. The input parameters of SOM are the hourly averaged Ey data over 3 h. The output layer of the SOM has a total of 400 neurons. The hourly Ey data are calculated from solar wind data, which are provided by NSSDC OMNIWeb and ACE spacecraft and contain information on 143 intense storms and a fair number of moderate storms, weak storms and quiet periods between September 3, 1966 and June 30, 2002. Our results show that SOM is able to classify solar wind structures and therefore to give timely intense storm alarms. In our SOM, 21 neurons out of 400 are identified to be closely associated with the intense storms and they successfully predict 134 intense storms out of the 143 ones selected. In particular, there are 14 neurons for which, if one or more of them are present, the occurrence probability of intense storms is about 90%. In addition, several of these 14 neurons can predict big magnetic storm (D_st  −180 nT). In summary, our method achieves high accuracy in predicting intense geomagnetic storms and could be applied in space environment prediction.     

Forecasting of DST variations from polar cap indices using neural networks

The 15-min averaged polar cap (PC) index was used as an input parameter for the Dst variation forecasting. The PC index is known to describe well the principal features of the solar wind as well as the total energy input to the magnetosphere. This allowed us to design a neural network able to forecast the Dst variations from 1 to 4 h ahead. 1998 PC and Dst data sets were used for training and testing and 1997 data sets was used for validation proposes. From the 15 moderate and strong geomagnetic storms observed during 1997, nine were successfully forecasted. In three cases the observed minimum Dst value was less than the predicted one, and only in three cases the neural network was not able to reproduce the features of the geomagnetic storm.     

基于机器学习相似度算法的Kp指数预报

基于机器学习中的相似度算法,建立了在历史太阳风数据中寻找与当前太阳风特征相近事例的推荐模型,用来预报地磁Kp指数.使用1998-2019年间随机选择的120个太阳风事例作为测试数据集,该模型能够推荐得到历史上与输入太阳风造成相似地磁影响的太阳风事例,最优事例的Kp指数与实际值的均方根误差为0.79,相关系数为0.93....     

“Fine structure” of the storm-substorm relationship: Ion injections during DST decrease

Following a long period of consensus on the storm-substorm relationship, a dispute on this topic has emerged in recent years. The importance of substorms for the buildup of the terrestrial ring current, which is the major element of magnetic storms, has been questioned in several studies. This paper is an effort to assess the “fine structure” of the storm-substorm relationship, by investigating the correlation between the changes in Dst and the substorm-associated O⁺ enhancements in the inner magnetosphere during the storm main phase. For this purpose we use energetic ion measurements from the Magnetospheric Ion Composition Spectrometer (MICS) on board the Combined Release and Radiation Effects Satellite (CRRES), and the newly produced high-resolution (5-min) Dst index for the intense storm of June 5, 1991. Substorm signatures from both MICS measurements and ground magnetometers correlate well with changes in the Dst decrease rate. This implies a significant influence of substorm occurrence on storm dynamics.     

Solar wind dependence of the Pc1 wave activity

There are a host of factors influencing the excitation of Pc1 geomagnetic pulsations, which are ULF waves in the frequency range between 0.2 and 5 Hz. We have studied carefully the dependence of the pearl-type Pc1 activity at Sodankylä, Finland (L = 5.1) on the plasma density N in front of the magnetosphere, the bulk velocity V of the solar wind, and the intensity B of the IMF. The result is as follows: high values of N and reduced values of V are favorable to appearance of Pc1, whereas the dependence of Pc1 activity on B is practically absent. We also show that the probability of Pc1 occurrence decreases with the interplanetary electric field, and increases with solar wind impact pressure and with the plasma to magnetic pressure ratio “beta”.     

冕洞特征参数与地磁暴强度及发生时间统计

地磁暴是空间天气预报的重要对象.在太阳活动周下降年和低年，冕洞发出的高速流经过三天左右行星际传输到达地球并引发的地磁暴占主导地位.目前地磁暴的预报通常依赖于1AU处卫星就位监测的太阳风参数，预报提前量只有1h左右.为了增加地磁暴预报提前量，需要从高速流和地磁暴的源头即太阳出发，建立冕洞特征参数与地磁暴的定量关系.分析了2010年5月到2016年12月的152个冕洞-地磁暴事件，利用SDO/AIA太阳极紫外图像提取了两类冕洞特征参数，分析了其与地磁暴期间ap，Dst和AE三种地磁指数的统计关系，给出冕洞特征参数与地磁暴强度以及发生时间的统计特征，为基于冕洞成像观测提前1～3天预报地磁暴提供了依据.      

Solar terrestrial effects of two distinct types

The occurrence frequencies or fluxes of most of the solar phenomena show a 11-year cycle like that of sunspots. However, the average characteristics of these phenomena may not show a 11-year cycle. Among the terrestrial parameters, some related directly to the occurrence frequencies of solar phenomena (for example, ionospheric number densities related to solar EUV fluxes which show 11-year cycle like sunspots) show 11-year cycles, including the double-peak structures near sunspot maxima. Other terrestrial parameters related to average characteristics may not show 11-year sunspot cycles. For example, long-term geomagnetic activity (Ap or Dst indices) is related to the average interplanetary solar wind speed V and the total magnetic field B. The average values of V depend not on the occurrence frequency of ICMEs and/or CIRs as such, but on the relative proportion of slow and high-speed events in them. Hence, V values (and Ap values) in any year could be low, normal or high irrespective of the phase of the 11-year cycle, except that during sunspot minimum, V (and Ap) values are also low. However, 2–3 years after the solar minimum (well before sunspot maximum), V values increase, oscillate near a high level for several years, and may even increase further during the declining phase of sunspot activity, due to increased influence of high-speed CIRs (corotating interplanetary regions). Thus, Ap would have no fixed relationship with sunspot activity. If some terrestrial parameter shows a 11-year cycle, chances are that the solar connection is through the occurrence frequencies (and not average characteristics) of some solar parameter.     

冕洞特征参数与重现型地磁暴关系的统计研究

在提取冕洞特征参数的基础上,利用1996年到2005年8月近十年来对地磁扰动有影响的356个冕洞事例,定量分析了冕洞特征参数(包括冕洞的面积比、经纬度跨度等)与冕洞高速流特征、重现型地磁扰动特征(包括扰动大小和持续时间等)之间的相关性,研究发现,从引起地磁扰动的冕洞在整个太阳活动周的分布来看,在地磁扰动峰年中冕洞影响同样具有重要的贡献;冕洞高速流太阳风速度与地磁扰动强度之间存在较强的相关性,而高速流中太阳风速度与冕洞面积比关系不大,与冕洞亮度存在一定相关性;冕洞的经度跨度与地磁扰动持续时间存在很强的正相关性.      

On the time lag between solar wind dynamic parameters and solar activity UV proxies

The solar activity displays variability and periodic behaviours over a wide range of timescales, with the presence of a most prominent cycle with a mean length of 11 years. Such variability is transported within the heliosphere by solar wind, radiation and other processes, affecting the properties of the interplanetary medium. The presence of solar activity–related periodicities is well visible in different solar wind and geomagnetic indices, although their time lags with respect to the solar cycle lead to hysteresis cycles. Here, we investigate the time lag behaviour between a physical proxy of the solar activity, the Ca II K index, and two solar wind parameters (speed and dynamic pressure), studying how their pairwise relative lags vary over almost five solar cycles. We find that the lag between Ca II K index and solar wind speed is not constant over the whole time interval investigated, with values ranging from 6 years to $～$1 year (average 3.2 years). A similar behaviour is found also for the solar wind dynamic pressure. Then, by using a Lomb-Scargle periodogram analysis we obtain a 10.21-year mean periodicity for the speed and 10.30-year for the dynamic pressure. We speculate that the different periodicities of the solar wind parameters with respect to the solar 11-year cycle may be related to the overall observed temporal evolution of the time lags. Finally, by accounting for them, we obtain empirical relations that link the amplitude of the Ca II K index to the two solar wind parameters.     

The ion energy spectra in the ring current during the geomagnetic storm of February, 1986

Using data from the CHEM instrument on the AMPTE/CCE spacecraft, we follow the development of the ring current energy spectra (1–300 keV/e) of the ion species H⁺, O⁺, He⁺, and He⁺⁺ in the post-noon and pre-noon local time sectors during the geomagnetic storm of February 1986. By comparing displays of phase space density, f, vs. magnetic moment, μ, we can distinguish between enhancements due to newly injected ions and those due to adiabatic energization of a pre-existing population. In both the local time sectors, the initial drop in Dst is associated with enhanced phase space densities of all species. The spectra observed during the pass when the Dst dropped to a minimem of −312 nT show a strong local time asymmetry. In the post-noon sector, the spectra showed the influx of a new population of ions, rich in O⁺ and He⁺⁺. In the pre-noon sector, the flux increase was consistent with adiabatic energization of the ion population injected earlier in the storm. This local time difference is consistent with a greatly enhanced convection electric field which brings a new population from the magnetotail to the post-noon, but not the pre-noon local time sector.     

利用行星际监测数据进行地磁暴预报

利用全连接神经网络方法应用于地磁Dst指数的预报中.对ACE卫星探测的太阳风和行星际磁场及其变化对未来几小时的Dst指数的影响进行了统计分析,发现在这些行星际实测参数中,对Dst指数作用较为明显的是太阳风速度、太阳风质子密度和行星际磁场南向分量,同时,当前Dst指数实测值对今后几小时的Dst指数已有很强的制约作用.在统计分析的基础上,建立了全连接神经网络预报模型.由于采用了全连接神经网络结构,模式能够反映出太阳风、行星际磁场等参数与地磁Dst指数参数的复杂联系,可以自动建立输入参量的最佳组合方式,提高了预报精度.通过利用大量实测数据对神经网络模式进行训练,最终建立了利用优选的ACE卫星行星际监测数据提前2 h对Dst指数进行预报.通过检测,预报的误差为14.3%.      

Statistical characteristics of low latitude Pc4s: Influence of solar wind and IMF parameters

Pc4 signatures for the year 2013, extracted from geomagnetic north–south and east–west components of induction coil magnetometer (LEMI 30) from low latitude station Desalpar (DSP), operated by Institute of Seismological Research (ISR), India have been investigated vis-à-vis the prevalent interplanetary parameters (IMF) as well as the geomagnetic activity indices. A clear dominance of Pc4-5 (467 events) over Pc3 (17 events) is observed. Local time variation of Pc4 shows a peak in the noon sector in both X and Y components. Our investigations show that the dominant peak frequency is 10 mHz at low latitude region. Correlations with solar wind and IMF parameters illustrate highest occurrence of Pc4 for a solar wind speed of 300–400 km/s and average IMF B field of 3–6 nT. The amplitude of Pc4s at DSP shows an increase with increasing solar wind speed, plasma density, solar wind dynamic pressure and average B field which is also reflected in the trend of frequency variation of these pulsations. We report that IMF clock angle at low latitude does not have influence on Pc4 occurrence. Based on the characteristics of these events, detected in latitudinally distributed stations from low and mid-latitudes from northern and southern hemisphere, we infer that modes were compressional, which could be driven by K-H instability or solar wind dynamic pressure, as compressional modes can propagate to low latitude with little attenuation.