Application of Novel Interplanetary Scintillation Visualisations using LOFAR A Case Study of Merged CMEs from September 2017_2

2025-04-27 0 0 3.11MB 17 页 10玖币

侵权投诉

https://dx.doi.org/10.1016/j.jasr.xxxx.xx.xxx

Available online at www.sciencedirect.com

Advances in Space Research xx (2022) xxx-xxx

www.elsevier.com/locate/asr

Application of Novel Interplanetary Scintillation Visualisations using

LOFAR: A Case Study of Merged CMEs from September 2017

R.A. Fallowsa,e,∗,K.Iwaib, B.V. Jacksonc,P.Zhangd,a, M.M. Bisie,P.Zuccaa

aASTRON - the Netherlands Institute for Radio Astronomy, Oude Hoogeveensedijk 4, 7991 PD Dwingeloo, the Netherlands

bInstitute for Space-Earth Environmental Research, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8601, Japan

cCenter for Astrophysics and Space Sciences, University of California, San Diego, LaJolla, California, 92093-0424, USA

dInstitute of Astronomy and National Astronomical Observatory,

Bulgarian Academy of Sciences, Soﬁa 1784, Bulgaria

eRAL Space, United Kingdom Research and Innovation – Science &Technology Facilities Council – Rutherford Appleton Laboratory, Harwell Campus,

Oxfordshire, OX11 0QX, UK

Received 19 April 2022; Received in ﬁnal form 22 August 2022; Accepted 29 August 2022;

Available online 19 September 2022

Abstract

Observations of interplanetary scintillation (IPS – the scintillation of compact radio sources due to density variations in the solar wind) enable

the velocity of the solar wind to be determined, and its bulk density to be estimated, throughout the inner heliosphere. A series of observations

using the Low Frequency Array (LOFAR - a radio telescope centred on the Netherlands with stations across Europe) were undertaken using this

technique to observe the passage of an ultra-fast CME which launched from the Sun following the X-class ﬂare of 10 September 2017. LOFAR

observed the strong radio source 3C147 at an elongation of 82 degrees from the Sun over a period of more than 30 hours and observed a strong

increase in speed to 900 km s−1followed two hours later by a strong increase in the level of scintillation, interpreted as a strong increase in density.

Both speed and density remained enhanced for a period of more than seven hours, to beyond the period of observation. Further analysis of these

data demonstrates a view of magnetic-ﬁeld rotation due to the passage of the CME, using advanced IPS techniques only available to a unique

instrument such as LOFAR.

Keywords: Interplanetary scintillation; Coronal mass ejection; Solar wind

1. Introduction

Observations of interplanetary scintillation (IPS - Clarke

(1964); Hewish et al. (1964)) have been used for several

decades to observe the solar wind and Coronal Mass Ejections

(CMEs) throughout the inner heliosphere. Such observations

are typically used to estimate solar wind velocity (e.g. Coles,

1996; Manoharan & Ananthakrishnan, 1990; Kojima & Kak-

inuma, 1990) and/or “g-level”, a normalised measure of the

strength of scintillation, related to density (e.g. Jackson et al.,

1998; Tappin, 1986). The IPS array operated by the Insti-

∗Corresponding author: email: rafallows@gmail.com

tute for Space-Earth Environmental Research (ISEE), Japan,

in operation since the 1980s, is used to take regular measure-

ments of these quantities (Tokumaru et al., 2011). These, in

turn, are fed into a tomographic model to provide 3-D recon-

structions of solar wind velocity and density throughout the in-

ner heliosphere every six hours, and a ﬁve-day prediction (see

https://ips.ucsd.edu/high resolution predictions - although he-

liospheric conditions resulting, e.g., the launch of a CME can

be predicted at best only about two days ahead of time, co-

rotating structures can be forecast at least ﬁve days ahead) of

these values at the Sun-Earth L1 Lagrangian point (e.g. Jack-

son et al., 2020, and references therein). Such views incorpo-

rate both the background solar wind and any CMEs present, so

arXiv:2210.02135v1 [astro-ph.SR] 5 Oct 2022

2Richard Fallows et al. /Advances in Space Research xx (2022) xxx-xxx

long as these were detected in the observations of IPS used as

input. Highly detailed views of the inner heliosphere are pos-

sible with this technique, given many more daily observations

of suﬃcient quality than are regularly available at present (Bisi

et al., 2009).

As a transit instrument reliant upon Earth rotation to per-

form successive observations of the same set of radio sources,

the ISEE IPS array is limited to only short-duration observa-

tions of each radio source every 24 hours, requiring the assis-

tance of MHD simulations to reconstruct the fast propagating

CMEs (e.g. Iwai et al., 2021). Observing stations with radio

source tracking capabilities can make multiple passes during a

single day, or dwell on individual sources for a longer period of

time, and simultaneous observation from such system(s) with

stations several hundred or more kilometres apart are capable of

getting much more detail from single observations. Such obser-

vations enable multiple solar wind streams to be detected cross-

ing the observing station to radio source lines of sight, such as

the “fast and faster” solar wind streams detected by the Ulysses

spacecraft and observed in IPS using the combined European

Incoherent Scatter (EISCAT) and Multi-Element Radio-Linked

Interferometer Network (MERLIN) systems (Bisi et al., 2007).

Furthermore, longer-duration observations are possible which

enable changes in solar wind structure (e.g. due to the onset

and/or the passage of a CME) to be tracked across a single line

of sight. For example, the onset of a CME from May 2005 was

detected and part of its structure tracked in an observation taken

using the EISCAT and MERLIN systems (Bisi et al., 2010;

Chang et al., 2021), and micro-scale structure in the slow so-

lar wind observed in measurements by EISCAT (e.g. Hardwick

et al., 2013, and references therein). It has also proved pos-

sible to detect an oﬀ-radial component to the fast solar wind,

where Dorrian et al. (2013) demonstrated that the polar solar

wind shows a slight equatorwards expansion, and Breen et al.

(2008) noted that a fast stream adjacent to the May 2005 CME

was deviated 8-15◦oﬀ-radial by the CME itself.

LOFAR (the low-frequency array, van Haarlem et al. (2013))

is Europe’s largest and most ﬂexible radio telescope, with capa-

bilities which enable much more information to be extracted

from multi-station observations of IPS. The wide bandwidth

enables any change in the scintillation pattern with frequency,

e.g. between weak and strong scintillation, to be directly ob-

served and features seen which would be invisible in a single-

frequency measurement (as observed in ionospheric scintilla-

tion measurements taken using LOFAR hardware, for exam-

ple, Fallows et al., 2014, 2020). Furthermore, the international

array contains 14 stations (at the time of writing - 13 were

available at the time of the observations described here) out-

side the Netherlands with baselines of ∼200 km to >∼2000 km,

in addition to the Dutch array containing a dense “core” of sta-

tions and 14 “remote” stations scattered across the north-east

of the Netherlands. All stations are connected via dedicated

high-speed data links to correlation and processing facilities in

Groningen, Netherlands. The array as it was in September 2017

is depicted in Figure 1 (it has since gained a new station near

Ventspils in Latvia, a further station will be built near Medic-

ina in Italy in 2023, and there are plans for a further station in

Fig. 1: Map showing the distribution of LOFAR stations over Europe (6 in

Germany, 3 in Poland, 1 each in France, Ireland, Sweden, and the UK, plus the

Dutch array of 38 stations) at the time of observation in September 2017. All

stations are connected via dedicated high-speed data links to correlation and

processing facilities in Groningen, Netherlands.

Bulgaria). This enables the spatial extent of the IPS correlation

between stations to be investigated, leading to information on

the density structure giving rise to the IPS to be studied, as will

be detailed later in this paper.

September 2017 was the most active period of solar cycle 24,

with three X-class ﬂares, numerous M-class ﬂares, and multi-

ple CMEs. The early arrival of the CME associated with the

6 September X-9 ﬂare produced severe geomagnetic storming

on 7 and 8 September; a further CME, thought at the time to

be ultra-fast with a speed of ∼3000 km s−1, launched on 10

September associated with a further X-8.2 ﬂare, as the active

region responsible rotated around the solar limb. A ﬂurry of

activity ensued following this latter event, as groups around the

world attempted to ﬁnd the CME in the heliosphere. A Direc-

tor’s Discretionary Time (DDT) proposal was quickly submit-

ted and approved, which enabled LOFAR to take observations

of IPS for ∼30 hours from late morning on 11 September with

the aim of ﬁnding the CME and tracking its passage across one

or more lines of sight.

This paper details the observations taken, introduces analy-

sis techniques which make full use of IPS data taken with LO-

FAR, and demonstrates the possibility for IPS to show magnetic

ﬁeld orientation as the CME passes across the line of sight. Full

MHD modelling of this event incorporating a comparison with

the LOFAR results presented here is given in a companion pa-

per by Iwai et al. (in press, 2022).

2. LOFAR Observations

At 15:35 UT on 10 September 2017 an X8.2 ﬂare was

observed as active region AR12673 (then at S08, W88) ro-

tated around the west limb of the Sun. This was associ-

ated with an ultra-fast CME ﬁrst observed in the LASCO C2

coronagraph at 16:00 UT, with a velocity measured through

the LASCO C3 ﬁeld of view of 3,212 km s−1. Full details

of this event as seen by LASCO can be found via the Halo

CME alert at https://umbra.nascom.nasa.gov/lasco/

Richard Fallows et al. /Advances in Space Research xx (2022) xxx-xxx 3

Fig. 2: Segment of a polar plot giving the elongations and position angles (clock

angle anti-clockwise from solar north) in the sky plane of the radio sources

observed by LOFAR. A LASCO C3 diﬀerence image taken at 16:54 UT on 10

September 2017 is superposed for reference (not displayed to scale).

observations/halo/2017/170910/. Later analyses, un-

available at the time, showed that this ultra-fast CME merged

during its passage through the inner heliosphere with two slow

CMEs which had launched on 9 September 2017 and them-

selves merged whilst still within the LASCO C3 ﬁeld of view

(e.g. Guo et al., 2018; Lee et al., 2018). The LOFAR observa-

tions detailed here therefore observed the result of the merger

of all three CMEs, rather than the single event originally envis-

aged, and at a later time than expected.

LOFAR observations were carried out from 11:30 UT on 11

September to 14:00 UT on 12 September 2017 to observe this

event, and alternated between four radio sources, chosen such

that it was considered likely that the CME would pass across

the line of sight to at least one of them. Figure 2 gives a plot

of the locations of the sources used, along with a reference

LASCO C3 diﬀerence image from the time. Observations were

of 9 minutes duration, with an obligatory 1 minute gap in be-

tween, alternating between sources during the periods when any

source was above an elevation of 25◦as seen from the LOFAR

core. The exception to this was 3C147 which is circumpolar

from LOFAR latitudes and for which observations continued

throughout. The observing scheme is detailed in Table 1.

All observations recorded Stokes-I dynamic spectra for 400

subbands, each 195.3125 kHz wide, covering contiguously the

frequency range 110-190 MHz, with an integration time of

0.01 s. Dynamic spectra were recorded individually for each

international and Dutch remote station included in the obser-

vation (only the station near Potsdam, Germany, was unavail-

able due to a fault at the time) and for a single tied-array beam

from the coherently-combined core stations (which is assumed

in subsequent analysis as being equivalent to a single station lo-

cated at the centre of the core, with the coordinates of station

CS002LBA).

Before detailing the initial processing of these data it is help-

ful to describe the coordinate system used with reference to the

line of sight between the radio source and an observing station

on Earth, as illustrated by the schematic in Figure 3. Although

the line of sight cuts through an extended portion of the in-

ner heliosphere, the majority of scattering typically comes from

around the point of closest approach (the so-called “P-point”) of

the line of sight to the Sun (as detailed in Section 3). Therefore,

Date Period Sources Observed

2017-09-11 11:30 - 14:29 UT 3C186,3C159,3C147

2017-09-11 14:30 - 16:19 UT 3C196,3C147

2017-09-11 16:20 - 23:59 UT 3C147

2017-09-12 00:00 - 13:59 UT 3C159,3C147

Table 1: Periods of observation of each radio source. All observations were

of 9 minutes duration with a 1 minute gap between. The sources are noted

in the order of observation so, for example, on 11 September 3C186 was ob-

served 11:30-11:39 UT, 3C159 11:40-11:49 UT, 3C147 11:50-11:59 UT, back

to 3C186 etc.

Sun

Prad

Ptan

Earth

VLOS

VBaseline

Fig. 3: Schematic plot of coordinates for an observation of IPS. The red vector

marks the line of sight, the blue vector marks the direction of a baseline between

a pair of observing stations, and the green vectors mark the components of the

projection of this baseline onto a plane perpendicular to the line of sight at the

P-point - the point of closest approach of the line of sight to the Sun.

the location of the P-point is usually used to deﬁne the coordi-

nates of the observation relative to the Sun (e.g., a heliographic

latitude and distance from the Sun).

When an observation uses more than one station, the physi-

cal baseline between each pair of stations can be projected onto

the sky-plane in the direction of the radio source from the Earth

and expressed in terms of components in the radial direction

from the Sun and tangential to it (green vectors in Figure 3).

The values of these components naturally change during the

course of an observation as the Earth rotates (see descriptions

of the geometry given in Dorrian et al. (2013) and Moran et al.

(1998)), so mean values are used in the analysis of each obser-

vation.

Initial processing consisted of mitigating the eﬀects of radio-

frequency interference (RFI), time series calculation, and cal-

culation of auto- and cross-correlation functions from each 9-

minute observation. This was carried out following a similar

procedure to that of Fallows et al. (2020) for ionospheric scin-

tillation, but with some parameter diﬀerences:

•RFI mitigation: A median ﬁlter was applied to the dy-

namic spectra using a window of (1.95 MHz ×0.525 s)

and then the original data divided by the median-ﬁltered

version to ﬂatten out the scintillation pattern. RFI were

then identiﬁed as absolute values greater than 10 σ, where

σis the median absolute deviation (MAD) of the ﬂattened

dataset. The MAD is used because the RFI can manifest

as extreme outliers in the data, making this measure more

4Richard Fallows et al. /Advances in Space Research xx (2022) xxx-xxx

robust than the standard deviation. Data points identiﬁed

as RFI are ﬂagged and not used in further processing.

•Time series’ of intensity received by each station are cal-

culated by averaging the intensities for each time sample

over the full frequency band of 110–190 MHz; this is rea-

sonable since the scintillation pattern remains highly cor-

related over the band in this set of observations.

•Calculate auto- power spectra using the intensity time se-

ries’ from each station.

•Apply a high-pass ﬁlter at 0.2 Hz to exclude the DC-

component and any obvious ionospheric scintillation or

slow system variation at the low spectral frequencies, and a

low-pass ﬁlter ( fc) at 5 Hz to cut out white noise at the high

spectral frequencies. The white noise is also subtracted us-

ing an average of spectral power over the high frequencies

above the low-pass ﬁlter value. An example is shown in

Figure 4.

•Calculate auto-correlation functions using the ﬁltered

power spectra.

•Cross- power spectra and cross-correlation functions were

calculated for time series’ from every pair of stations (a

total of 351 combinations) following the same methods.

•The baseline between every pair of stations was projected

onto the sky-plane and components in the radial direction

from the Sun and tangential to it calculated.

The processing described above can only mitigate the eﬀects

of short-duration and/or narrow-band spikes of RFI, so a further

selection process is necessary to try and exclude stations more

strongly aﬀected by interference in any given observations, or

suﬀering from weak signal-to-noise for other reasons (e.g., the

remote stations contain half the number of high-band antennas -

high-band refers to the frequency range, which covers that used

here - compared to the international stations, and the station in

Ireland was newly-built and not yet well-calibrated at the time

of observation). These eﬀects are most obvious in the auto-

correlation functions which should be more or less the same

for simultaneous data taken from diﬀerent stations. The auto-

correlation functions of unreliable data tend to exhibit a much

narrower peak than those of reliable data. Hence datasets whose

normalised auto-correlation function values close to the peak

deviated by more than one standard deviation from the median

for all auto-correlation functions in the observation were ex-

cluded from further analysis.

The cross-correlation function typically exhibits one or more

peaks at one or more time-lags, corresponding to the velocity(s)

of material crossing the lines of sight and so dependent on the

length of the projected baseline between the pair of stations cor-

related. Since this is only sensitive to material crossing perpen-

dicularly to the lines of sight, velocity estimates presented here

are in the sky plane and thus represent foreshortened versions

of the true values for locations along the line of sight away from

the P-point.

Fig. 4: Example raw power spectrum and spectrum after ﬁltering and noise sub-

traction, calculated using SE607 data from the 9-minute observation of 3C147

taken at 05:30 UT on 12 September 2017.

A set of example cross-correlation functions (CCFs) from

the observations presented here is given in Figure 5, which il-

lustrate the eﬀect of increasing baseline length (see also the

model CCFs given in Coles, 1996). Two diﬀerent solar wind

velocities were detected in this observation; a regular slow solar

wind stream and that of the faster CME. The CCF from the pair

of stations with the shortest baseline displayed here (212 km

radial, blue solid curve in Figure 5 top) registers the second,

slow, stream only as a bump at a time lag of just over 1 s, but

this bump becomes a second distinct peak in the CCFs as the

baseline length increases, allowing a more direct measurement

of the velocity it corresponds to.

Since very few almost purely radial baselines are available,

Figure 5 shows two sets of baselines, each following very ap-

proximately a diﬀerent oﬀ-radial direction (no baselines are ex-

actly radial in the dataset), one tending a few degrees polewards

(bottom) and the other equatorwards (top). This illustrates a fur-

ther interesting aspect: In the lower set of plots, the slow stream

part of the cross-correlation functions (covering time-lags 2-3 s)

appears to broaden into a further “bump”, suggesting the pres-

ence of a third, slightly slower, stream. This is less apparent in

the upper set of plots, possibly indicating that this third stream

is slightly oﬀ-radial in its direction.

3. Analysis

IPS is the result of an integral of scattering taking place

along an extended line of sight. Typically, scattering is assumed

to be “weak”, an assumption which is valid into an approximate

elongation from the Sun which is dependent on observing fre-

quency (Coles, 1978) and solar wind conditions (e.g., the weak

scattering assumption remains valid closer to the Sun for the

less-dense fast solar wind above polar coronal holes than the

denser slow wind, as demonstrated in, e.g., Manoharan (1993)

and Fallows et al. (2002)). Closer to the Sun this assumption

breaks down, although it has been demonstrated to remain valid

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

https://dx.doi.org/10.1016/j.jasr.xxxx.xx.xxx0273-1177/©2022COSPAR.PublishedbyElsevierLtdAllrightsreserved.Availableonlineatwww.sciencedirect.comAdvancesinSpaceResearchxx(2022)xxx-xxxwww.elsevier.com/locate/asrApplicationofNovelInterplanetaryScintillationVisualisationsusingLOFAR:ACaseStudyofMergedCM...

展开>> 收起<<

Application of Novel Interplanetary Scintillation Visualisations using LOFAR A Case Study of Merged CMEs from September 2017_2.pdf

共17页,预览4页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Application of Novel Interplanetary Scintillation Visualisations using LOFAR A Case Study of Merged CMEs from September 2017_2

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: