Studying Sun–Planet Connections Using the Heliophysics Integrated Observatory (HELIO)
- First Online:
- Cite this article as:
- Pérez-Suárez, D., Maloney, S.A., Higgins, P.A. et al. Sol Phys (2012) 280: 603. doi:10.1007/s11207-012-0110-x
- 693 Downloads
The Heliophysics Integrated Observatory (HELIO) is a software infrastructure involving a collection of web services, heliospheric data sources (e.g., solar, planetary, etc.), and event catalogues – all of which are accessible through a unified front end. In this paper we use the HELIO infrastructure to perform three case studies based on solar events that propagate through the heliosphere. These include a coronal mass ejection that intersects both Earth and Mars, a solar energetic particle event that crosses the orbit of Earth, and a high-speed solar wind stream, produced by a coronal hole, that is observed in situ at Earth (L1). A ballistic propagation model is run as one of the HELIO services and used to model these events, predicting if they will interact with a spacecraft or planet and determining the associated time of arrival. The HELIO infrastructure streamlines the method used to perform these kinds of case study by centralising the process of searching for and visualising data, indicating interesting features on the solar disk, and finally connecting remotely observed solar features with those detected by in situ solar wind and energetic particle instruments. HELIO represents an important leap forward in European heliophysics infrastructure by bridging the boundaries of traditional scientific domains.
A range of phenomena originating from the Sun can produce effects throughout the heliosphere, such as coronal mass ejections (CMEs), solar energetic particle (SEP) events, high-speed solar wind (HSSW) streams, and co-rotating interaction regions (CIRs). One example is the 2003 “Halloween Storm” period (e.g., see the summary article of Gopalswamy et al.2005) in which several CMEs erupted from an active region on the solar surface, resulting in impressive aurorae on Earth, and continuing on to intersect the Voyager spacecraft near the edge of the heliosphere (Intriligator, Rees, and Horbury 2008). Heliospheric phenomena are commonly studied using both remote sensing and in situ instruments, each allowing a unique perspective on the physics of these events. The more complete the picture of these phenomena is, the better we can model and ultimately make predictions of space weather conditions. The beauty of heliophysics is that its understanding requires a diversity of perspectives. However, this causes an inherent problem, because the diversity of disciplines involved results in knowledge gaps between the different disciplines (e.g., coordinate systems, definitions of phenomena, terms and acronyms, data standards, descriptions, and analysis languages).
In astronomy, data standards have been unified with the creation of virtual observatories (VOs) and their associated access tools. These tools allow scientists to obtain public data for any object using only its celestial coordinates. A similar approach has begun in heliophysics with the creation of the Virtual Solar Observatory1 to gather many of the available solar data (N.B. NASA’s Planetary Data System2 is the VO for planetary sciences). The European Grid of Solar Observations (EGSO; Bentley and EGSO Consortium 2002) and more recently the Heliophysics Event Knowledgebase (HEK; Hurlburt et al.2012)3 go a step further, adding event lists and other functionality that is intended to aid users in finding the available data linked to a given event or feature on the Sun. AstroGrid,4 through its HelioScope service, provides tools and services to access data archives and catalogues, find data sets, as well as a limited infrastructure for remote data processing. Finally, CDAWeb5 and Automated Multi Dataset Analysis (AMDA)6 offer data access and basic remote processing capabilities for planetary scientists. Currently solar and planetary data access and analysis are disjointed, although both are necessary to study the inter-disciplinary field of heliophysics.
The HELiophysics Integrated Observatory (HELIO; Bentley et al.2011b) addresses these issues. It combines many features of previous projects and aims to act as a centralised access point for as many heliospheric data as possible and to associate catalogued events with observations in which they can be found. Creating a VO for heliophysics is more complex than for astronomy because of the range of dynamic spatial and temporal scales involved. Instead of the 2D position vector that is generally used to uniquely identify astrophysical objects, time and distance are also required to associate data with heliospheric events. As such, the main interface of HELIO is a step forward by allowing 4D searches and providing automatic association between events, features, and data in time and space using both matching and built-in propagation models.
In this paper, we use HELIO to obtain catalogued event entries and data by running a simple propagation model to perform three heliospheric event case studies. Using the output from specific HELIO services (each described in Section 2), conventional methods are utilised to analyse each event. The results and discussion for each use case is presented in Section 3. Conclusions on the performance and future directions for HELIO are presented in Section 4.
2 The Heliophysics Integrated Observatory
The principal objective of HELIO is to create a collaborative environment where scientists from different communities can discover connections between solar phenomena, interplanetary disturbances, and their effects on the planets. HELIO consists of a set of services that provide access to event and feature catalogues and remote processing services, as well as access to observational data and descriptions of the instruments. All services are available from a centralised web interface (http://hfe.helio-vo.eu) that allows the user to pass information seamlessly from one service to another, without requiring the user to have a deep knowledge of how the system is built. However, the nature of the system architecture allows all of the services to be individually accessible. This enables advanced use of the system through SQL queries or by means of a programming language (e.g., IDL or Python). A complete and detailed description of the HELIO architecture and services can be found in both Bentley et al. (2011a) and the online documentation.7
Identify the occurrence of an interesting event or phenomenon.
Review the availability of suitable observations.
Locate, select, and retrieve the required observations.
A more detailed description of the different services within HELIO and the properties of the propagation model used in this paper are shown in the following subsections.
2.1 HELIO Environment
HELIO consists of a number of services that are combined into a single unified front-end interface, but each service can be independently queried using their own interfaces. This kind of design follows a service oriented architecture, avoiding a single monolithic system and allowing the services to be used individually or combined together through a workflow environment.
This study only makes use of four services provided by HELIO – three Search services (Event, Feature, and Data) and one that runs a propagation model. The other services provide valuable information that is of interest in different workflows. This includes services to search for instruments based on capability or position in the heliosphere, and a service that provides contextual information (e.g., GOES X-ray flux or flare positions on full-disk solar images). In addition to those, HELIO also has a catalog with all non-full-disk instruments. This is aimed at providing more detailed information when searching for instruments that may have observed a particular event. Finally, data mining on in situ measurements is offered through AMDA, while large processing capabilities and storage resources are provided by Grid-Ireland.8 A brief description of the event and feature catalogues, as well as the data provider service is presented.
List of automated feature detection algorithms available in HELIO.
Fuller, Aboudarham, and Bentley (2005)
Bonnin et al. (2012)
Krista and Gallagher (2009)
Higgins et al. (2011)
SDOSS & MDISS
Zharkov, Zharkova, and Ipson (2005)
Watson et al. (2009)
Fuller et al. (2012)
Coronal Radio Emission
Type II radio burst
Bonnin et al. (2011)
Type III radio burst
Lobzin et al. (2009)
2.2 Solar–Heliospheric Event Ballistic Algorithm
The first propagation model implemented in HELIO is the Solar–Heliospheric Event Ballistic Algorithm (SHEBA). Its main purpose is to determine if the Sun, a planet or spacecraft can be associated with an event and, if so, to provide a time interval for this interaction. It has been split into three different functions: CMEs, SEPs, and CIRs. SHEBA can be executed in both a forward and backward sense – i.e., to estimate the time and position of an event on the Sun based on the detection of an event at a certain time and position in the heliosphere and vice versa. SHEBA is a 2D ballistic model, in which events propagate at a constant speed and are confined to propagate in the solar equatorial plane (Burlaga 1984). All objects (spacecraft and planets) are projected onto this plane.
SHEBA inputs and outputs for the three possible scenarios: CME, SEP, and CIR shown in Figure 2. The inputs for these runs come from the 2000 “Bastille Day” event for the CME, a flare from the 2003 “Halloween Storm” period for the SEP, and the edge of one of the CHs shown in Section 3.3. Satellites orbiting planets are approximated by the planets positions.
vsw (km s−1)
1100±200 km s−1
0.58 – 0.84
1.36 – 1.96
26.34 – 38.05
40.25 – 58.14
82 – 120
450±20 km s−1
700±50 km s−1
10 – 10.2
−3.7 – −3.3
1.6 – 1.9
−2.3 – −1.8
−12 – −10.2
−8.1 – −5.2
7.3 – −9.4
4.4 – −8.9
2.4 – −10.8
11 – −11
−9 – −6
The implementation of SHEBA in HELIO is made by a webservice that connects a front-end interface with a set of IDL routines9 that run on the grid system at Trinity College Dublin. SHEBA produces two images as a context, one for the inner solar system and another for the outer (as shown in Figure 2). The input and output parameters are additionally saved as a VOTable that is useful for the interaction with the rest of the HELIO system and other VO tools. The three use cases presented in the following section make use of SHEBA. However, in the future HELIO will incorporate more sophisticated analytical models, or even numerical simulations, which could be chosen by the user as an alternative to SHEBA.
3 Event-Specific Case Studies
HELIO is not designed to carry out scientific data analysis but instead facilitates this goal by providing a central resource to search the various events lists, types of observation, spacecraft locations and capabilities, and to allow these products to be downloaded for analysis. As a result, some of the functions HELIO provides require a priori knowledge or certain user inputs. For example, in the case of automatic detection algorithms, the user should first be familiar with the algorithm limitations as only events or features with characteristics predetermined by the methods or parameters used by the algorithm will be detected (e.g., CACTus relies on Hough transforms and so its algorithm limits detected CMEs to those with constant speed; Robbrecht and Berghmans 2004).
In the following subsections, we investigate three HELIO use cases that highlight each of the applications of the SHEBA propagation system. The first case study investigates CMEs impacting both Earth and Mars (Section 3.1). The second case study tracks a SEP event resulting from a solar flare out through the heliosphere (Section 3.2). The third case study aims to determine if the source of a HSSW stream that sweeps past Earth is due to a CH (Section 3.3).
3.1 CME Use Case
CMEs propagating through the heliosphere can be detected in several forms of remote sensing observations and in situ measurements. Close to the Sun, CMEs are observed by solar imaging instruments such as extreme ultraviolet (EUV) imagers and further out by coronagraphs, heliospheric imagers, space-borne radio instruments and many ground-based radio observatories. CMEs may be detected (directly) through in situ measurements of interplanetary plasma properties or (indirectly) by their effects on planetary magnetospheres. Planetary effects can also be observed through remote observations, e.g., EUV and radio emission from auroral activity triggered by the passage of a CME. Connecting such disparate observations through time and across disciplinary domains is a complex and cumbersome task. HELIO facilitates this effort by providing a central resource to search the various events lists, types of observation, and spacecraft locations and capabilities to allow this mixture of data to be downloaded for analysis.
In this use case the aim is to examine the propagation of CMEs from the Sun into the heliosphere. In particular, we use a study by Falkenberg et al. (2011), which investigates CMEs interacting with Earth and Mars as a template to demonstrate some of the capabilities of the HELIO infrastructure. The study Falkenberg et al. analyses a time frame when the Mars Global Surveyor (MGS; Acuna et al.1992, 1998) was in operation while Earth and Mars were separated by less than 80∘ in longitude.10 This corresponds roughly to the years 2001 and 2003. These time frames could be identified using the HELIO framework to search for this combination of planetary and spacecraft arrangements, but we proceed using the following time ranges: April 2001 – January 2002, and May 2003 – December 2003, based on the events in the paper Falkenberg et al. (2011).
The use case shown in Figure 1 is complicated because the study Falkenberg et al. begins by looking for shock events at Mars. However, no shock event list exists based on data from Mars (other than the list of Falkenberg et al., which would not be a rigorous test of the system) so it is necessary to begin from Earth. Shock events detected at Earth were back-propagated to the Sun and associated with source events. These source events were forward-propagated into the heliosphere, providing time intervals of possible event arrival at Mars.
Extraction of the merged results from the HELIO event catalogue query on the SOHO/CELIAS/MTOF/PM Interplanetary shock list and Falkenberg et al. events for April 2001 – January 2002 where starting times at Earth match. Columns give the time of the event on Earth, comments from the Interplanetary shock list (“?” indicates missing or undetermined measurements as recorded in the catalogue), and the match event number from Falkenberg et al. (2011).
Falkenberg et al. Event Number
Origin: AR 9393, X20, N19W75?, 2 April 21:51 UT; travel time 40.5 hours
Origin: AR 9415, M7.9/2B, S21W04, 9 April 15:34 UT; travel time 45.5 hours; almost overtaken by the above event
Origin: AR 9415, X14.4/2B, S20W85, 15 April 13:50 UT; travel time 58 hours; accompanied by a Ground Level Event
Origin: Halo CME, 24 September 09:12 UT, starting at AR9632, associated with an X2.6 flare at S18E27 observed at 10:38 UT; travel time 33 hours
Origin: 17 November 05:25 UT full halo CME with M2.8/1N flare in AR 9704 at S13E42; travel time 60 hours
Origin: 26 December 05:40 UT partial halo CME with M7.1/1B flare in AR 9742 at N08W54; travel time 71 hours
Origin: 28 December 20:45 UT X3.4 flare in AR 9767 at S20E97 with partial halo CME; travel time 47 hours
Origin?? The time of passage is better determined from one of the rates in the main MTOF sensor
The next task is to forward-propagate both the CMEs from the Sun into the heliosphere. For clarity we will refer as CMEA for the one observed on 1 April, and CMEB for the one on 2 April. The propagation service was employed to estimate the CMEs arrival time at Mars and at Earth (although Earth arrival time could seem redundant for this particular use case, it helps to distinguish between the two candidate CMEs) using the time of the flare, speed of the CME, and the CME source position. Figure 3 shows one possible output of the propagation service using the associated flare start time and a CME speed of 800±300 km s−1. This speed is chosen, as it is already known from the previous step to provide the correct time range at Earth. The estimated times of arrival at Earth are 3 April 2001 01:17 UT – 4 April 2001 22:42 UT for CMEA and 4 April 2001 00:49 UT – 7 April 2001 03:06 UT for CMEB. Comparing both intervals with the ACE measurements of the bulk speed of the solar wind (top panel in Figure 3) indicates that CMEB is more likely to correspond to the one recorded by the SOHO/CELIAS/MTOF/PM Interplanetary shock list on 4 April 2001 14:21 UT (the starting point of the use case). Also, close inspection of the CME and flare catalogues and their data products shows some indications that the CMEA may have lifted up on the east side of the Sun making it less probable to hit Mars. Therefore, the estimated time of arrival at Mars for CMEB is in the range 5 April 2001 07:00 UT – 8 April 2001 03:59 UT and a search for instruments located at Mars in this interval yields MGS Electron Reflectometer (ER) and magnetometer (MAG). The instruments and time range were then used to identify available data and provide download links. These data were subsequently analysed as described in Falkenberg et al. and references therein. The results are displayed in Figure 3, where a shock is detected at Mars in the predicted time range.
Thus, a shock event detected at Earth (L1) was successfully back-propagated to the Sun and associated with two flares and CMEs. These CMEs were forward-propagated first to Earth, which helped us to distinguish which one was the one we tracked back, and consequently find the estimated time of arrival at Mars, leading to the identification of a shock at Mars. Each of the events identified above from consideration of the Falkenberg et al. paper could be treated in a similar manner to obtain all relevant data (e.g., X-ray and coronagraph observations, in situ measurements at Earth and Mars).
3.2 SEP Use Case
SEP events are significant increases (both sudden and gradual) in the flux of high-energy particles following solar flare or CME events. These particles can be accelerated up to near-relativistic speeds and travel out into the heliosphere, spiraling along magnetic field lines that are connected to the source region. When a planet or spacecraft intersects the path of a SEP event, the high-energy of the particles can damage satellite equipment and endanger the health of astronauts. The investigation of SEPs requires knowledge of the time–space locations of active regions, flares, and planets. The SHEBA propagation model (described in Section 2.2) is used to associate the differing spatial and temporal locations, assuming some properties of the solar wind and the accelerated particles.
We present the case study of a combined flare and CME event occurring on 7 June 2011, which also produced a SEP event at Earth. This event was chosen for its extensive press coverage and popularity within the community when the use cases were proposed. This event resulted from a relatively small flare that accompanied one of the most spectacular CMEs so far in Solar Cycle 24. This was well observed by a myriad of instruments including SDO/AIA, which has exceptional spatial and temporal resolutions.
A GOES M2.5 flare is observed at around 7 June 2011 06:00 UT. Searching the GOES soft X-ray Flare List in the event catalogue for flares on the day of interest yields a precise start time (06:16 UT) and provides a region of origin (NOAA AR 11226 at W64 longitude). The time and location of the flare are used as input for the propagation service, with the assumption of a quiet solar wind with a radial speed of 400±20 km s−1 (typical conditions taken from user experience).
3.3 HSSW Stream Use Case
CHs are large-scale density depletions in the corona and are the source of HSSW streams (Altschuler, Trotter, and Orrall 1972). Fast solar wind accelerated in these regions catches up slow solar wind at larger radii as the Sun rotates. The region where the slow and fast streams interact is a CIR (Pizzo 1978). In this case study, we aim to connect an in situ solar wind event observed at Earth (L1) with a CH at the Sun. Here the event in question is the detection of a CIR as it sweeps past a solar wind particle detector. We utilise both a catalogue of in situ events (“Stream Interaction Regions from Wind and ACE Data” as described in Jian et al.2006), and a catalogue of CH detections (data products of CHARM) that relies on EUV images. HELIO services are used to obtain data associated with the event. Conventional analysis techniques are then used to characterise the complete event from the Sun to Earth.
A summary of the Jian et al. (2006) Stream Interaction Region catalogue for February 2004 – April 2004 obtained from HELIO. Events studied in the use case are shadowed.
As there are no alternate EUV instruments observing during the time range of interest, we search for those observing in the He i 10 830 Å wavelength (process 3), which is commonly used to detect CHs using ground-based observations (Henney and Harvey 2007). Although data are available, there are no catalogues confirming the presence of a CH on disk at this time. However, because CHs may persist for multiple solar rotations, we check both 27 days before (process 4−) and 27 days after (process 4+) the propagated time (i.e., 25 February 2004 and 22 April 2004, respectively). Fortunately, CH detections produced by CHARM are available for both days, as shown in Figure 5.
The longitude of the westernmost edge of the CHARM detection is taken to represent the forward edge of the HSSW stream, and hence the location of the CIR. This longitude and detection time are implemented with a user-defined solar wind speed to forward-propagate the CIR Parker spiral (processes 5− and 5+) and generate a predicted travel time for the CIR to reach Earth (i.e., the time taken to catch up with Earth’s orbit due to the rotation of the Sun). A good match is found between the predicted arrival time and the catalogue of Jian et al., indicated by the dark shadowed rows in Table 4 and vertical lines in the ACE/SWEPAM plots of Figure 5.
Users can proceed to download associated data sets and begin detailed analysis now that appropriate time ranges and lists of instruments have been determined using HELIO. External services such as the HEK could have been used to determine the presence of a CH during the SOHO “key-hole” period. In the future, it is expected that the HEK catalogues will be accessible through HELIO.
4 Conclusions and Future Directions
In this paper, we have presented three case studies that rely on the event, feature, and data-search capabilities of HELIO.17 The effort required to perform these case studies was significantly reduced by this system in comparison to relying on the original data sources for each data set. We have focused on the role of the propagation service in connecting the three search services offered by HELIO. However, other important services aided this work, such as the “context service” that allows the visualisation of contextual information at each workflow step and data mining capabilities offered by AMDA through HELIO. This work illustrates how HELIO aids in the location and acquisition of data across the boundaries of traditional scientific domains.
The chosen case studies involve three different workflows that exhibit the potential of the HELIO infrastructure. The CME use case presented in Section 3.1 indicates that a shock event observed at Earth can be used to determine the source location of a CME and predict its arrival time at Mars. The SEP use case, detailed in Section 3.2, highlights that it can be used to determine which planets or spacecraft were hit by high-energy particles by taking the spiral into account. Finally, the HSSW use case outlined in Section 3.3 shows the connection between HSSW streams and CHs over different solar rotations. This example additionally showcases the flexibility of HELIO to find supplementary data sources or alternate times when the desired data would be available.
Development of new visualisation functions – e.g., the ability to preview data.
Addition of more sophisticated propagation models – e.g., semi-empirical and magnetohydrodynamic simulations that are physically more complete.
Addition of new algorithms to detect the same features in different data – e.g., the code described in Henney and Harvey (2007) to detect CHs in He i 10 830 Å images.
Addition of algorithms to detect the same features in the same data to allow inter-comparison of algorithm outputs – e.g., the Spatial Possibilistic Clustering Algorithm (SPoCA; Barra et al.2009), which detects CHs in EUV images (for comparison with CHARM).
The primary point of access to HELIO is intended to be the web front-end interface, but it can also be accessed by other means, e.g., the HELIO branch in the SolarSoftWare18 (SSW; Freeland and Handy 1998) tree of the IDL environment and workflow tools such as Taverna.19 Using HELIO through the Taverna workflow application allows users to share their workflows with the rest of the community using the MyExperiment infrastructure.20 The operations described in the CME,21 SEP,22 and HSSW23use cases have been converted into Taverna workflows and are accessible online.
It is important that the consortia engaged in the development of heliophysics VOs collaborate, in order to accelerate the progression of the field and avoid replication of work. A goal of HELIO is to incorporate knowledge from additional projects, e.g., the Solar-Terrestrial Investigations and Archives (SOTERIA), the Europlanet Research Infrastructure, the Space Weather European Network (SWENET), and the HEK.24 This will be achieved by creating inter-catalogue links and utilising algorithms that have been developed for these projects.
The Taverna workflow that retrieves the time ranges where two objects were located in the heliosphere within a number of longitudinal degrees is accessible from: http://www.myexperiment.org/workflows/2524.html.
The CME on the 2 April 2001 was recorded by the CACTus catalogue one hour later (i.e. 23:06 UT), however, closer inspection of their data shows that they recorded are the same event.
“Key-hole” is a period of reduced communication bandwidth. See: http://soho.nascom.nasa.gov/hotshots/2004_01_04.
All data and scripts used to generate the results of this paper are available at: http://dx.doi.org/10.6084/m9.figshare.92577.
HEK is exploring using crowd-sourcing to populate some catalogues, seeking to emulate the success of the Galaxy Zoo, adding another layer of knowledge (Showalter et al.2009).
HELIO is a research infrastructure funded under the Capacities specific programme within the European Commission’s Seventh Framework Programme (FP7; Project No. 238969). The use cases described in this paper are the result of work carried out during splinter sessions at the HELIO Coordinated Data Analysis Workshop 2 25 hosted at ICTP in Trieste. PAH acknowledges support from ESA/PRODEX and HELIO. DSB is supported by a Marie Curie Intra-European Fellowship within the 7th European Community Framework Programme. MD is supported by funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement No. 263252 [COMESEP].
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.