Final Report Summary - F-CHROMA (Flare Chromospheres: Observations, Models and Archives)
Executive Summary:
The F-CHROMA project was a dedicated multi-mode, multi-wavelength study of solar flares as observed in the lower solar atmosphere, or chromosphere. F-CHROMA scientists led nu-merous successful campaigns to observed flares from ground-based solar telescopes, co-ordinated with space-based facilities. This campaign data, and other data, consisting of flare images and spectroscopy, were analysed in combination with state-of-the-art numerical simulations to deduce the structure and evolution of the flare chromosphere, and understand the mechanisms by which energy is transported through the solar atmosphere in a flare, and dissipated in the form of heat, ionisation, and radiation.
Studies of the flare chromosphere enjoyed a renaissance with F-CHROMA, which succeed-ed in placing the importance of observing and understanding the lower-atmosphere energy input in flares front-and-centre for space missions such as IRIS, and the forthcoming major ground-based DKIST telescope. Our experiences in planning and executing flare observa-tions have revealed optimal ways to obtain data on the all-important flare energy-release phase. Discoveries and advances from F-CHROMA included measurements of hot, fast up-flows of hundreds of km/s from ‘elementary excitations’ in the chromosphere, each with an identical velocity profile. These are accompanied by cooler downflows of some tens of km/s, which appear to have an optically-thin component, leading to the exciting prospect of directly mapping the entire flow field in the chromosphere. F-CHROMA studies have demonstrated the critical role of hydrogen and helium ionisation in the temperature and dynamic structure of the chromosphere, and its central role in generating the broad-band UV, IR and ‘white light’ continuum emission by which flares were first identified. We showed areas of agree-ment and – perhaps more interestingly – disagreement with the 45-year-old ‘collisional thick target’ model of electron beam energy transport and deposition, and also made the first radi-ation hydrodynamic simulations of a flare atmosphere excited by a flux of Alfvén waves, showing that this is a viable alternative for generating flare heating and radiation. We have also embarked on the ambitious tasks of spectropolarimetry of the flare chromosphere, and on 2- and 3-D flare radiation transfer modelling; though at early stages they usefully point the way for future studies. And for the first time ever we have successfully cross-validated two independent flare radiation hydrodynamics codes, giving us confidence that – within the ap-proximations of both codes – they are capturing the evolution of a flare atmosphere correctly.
The F-CHROMA project produced an archive of the output of these flare models, which are tricky and time-consuming to run, along with the software necessary to interrogate and inter-pret the model output. An archive of high-quality ground-based flare observations has also been published, along with a catalogue of most ground-based flare observations in Cycles 23 and 24, and their space-based counterparts. We involved amateur astronomers in our flare observations, in two “F-HUNTERS” pro-am observing campaigns, with considerable success, and we disseminated the results of our project through more than 50 refereed papers, more than 100 presentations, and a training school for the next generation of solar flare physicists.
At the end of F-CHROMA, we have a vision of the flare atmosphere, and of flare energy transport, that emphasises excitation of the deep chromosphere, the central role of ionisa-tion, and the great power of optically-thick radiation, such as will be observed in great detail by DKIST, to test different models of flare energy transport and solve long-standing puzzles in flare physics.
Project Context and Objectives:
The F-CHROMA project was a dedicated multi-mode, multi-wavelength study of solar flares as observed in the lower solar atmosphere, or chromosphere. F-CHROMA scientists led numerous successful campaigns to observed flares from ground-based solar telescopes, co-ordinated with space-based facilities. This campaign data, and other data, consisting of flare images and spectroscopy, were analysed in combination with state-of-the-art numerical simulations to deduce the structure and evolution of the flare chromosphere, and understand the mechanisms by which energy is transported through the solar atmosphere in a flare, and dissipated in the form of heat, ionisation, and radiation.
Studies of the flare chromosphere enjoyed a renaissance with F-CHROMA, which succeeded in placing the importance of observing and understanding the lower-atmosphere energy input in flares front-and-centre for space missions such as IRIS, and the forthcoming major ground-based DKIST telescope. Our experiences in planning and executing flare observations have revealed optimal ways to obtain data on the all-important flare energy-release phase. Discoveries and advances from F-CHROMA included measurements of hot, fast upflows of hundreds of km/s from ‘elementary excitations’ in the chromosphere, each with an identical velocity profile. These are accompanied by cooler downflows of some tens of km/s, which appear to have an optically-thin component, leading to the exciting prospect of directly mapping the entire flow field in the chromosphere. F-CHROMA studies have demonstrated the critical role of hydrogen and helium ionisation in the temperature and dynamic structure of the chromosphere, and its central role in generating the broad-band UV, IR and ‘white light’ continuum emission by which flares were first identified. We showed areas of agreement and – perhaps more interestingly – disagreement with the 45-year-old ‘collisional thick target’ model of electron beam energy transport and deposition, and also made the first radiation hydrodynamic simulations of a flare atmosphere excited by a flux of Alfvén waves, showing that this is a viable alternative for generating flare heating and radiation. We have also embarked on the ambitious tasks of spectropolarimetry of the flare chromosphere, and on 2- and 3-D flare radiation transfer modelling; though at early stages they usefully point the way for future studies. And for the first time ever we have successfully cross-validated two independent flare radiation hydrodynamics codes, giving us confidence that – within the approximations of both codes – they are capturing the evolution of a flare atmosphere correctly.
The F-CHROMA project produced an archive of the output of these flare models, which are tricky and time-consuming to run, along with the software necessary to interrogate and interpret the model output. An archive of high-quality ground-based flare observations has also been published, along with a catalogue of most ground-based flare observations in Cycles 23 and 24, and their space-based counterparts. We involved amateur astronomers in our flare observations, in two “F-HUNTERS” pro-am observing campaigns, with considerable success, and we disseminated the results of our project through more than 50 refereed papers, more than 100 presentations, and a training school for the next generation of solar flare physicists.
At the end of F-CHROMA, we have a vision of the flare atmosphere, and of flare energy transport, that emphasises excitation of the deep chromosphere, the central role of ionisation, and the great power of optically-thick radiation, such as will be observed in great detail by DKIST, to test different models of flare energy transport and solve long-standing puzzles in flare physics.
2. Project context and objectives
Our nearest star, the Sun, controls the solar system first and foremost by its gravitational presence and by the steady radiation output on which our existence on this planet relies completely. However, it also controls its environment in less obvious ways. The phenomenon of ‘space weather’ is increasingly in the public consciousness. Space weather encompasses many effects, but is primarily due to the influence of the Sun on the electromagnetic environment of the Earth, planets, and heliosphere. Sources of space weather are both solar flares, which are dramatic changes in the solar radiation at high energies in particular, and coronal mass ejections (CMEs) which are expulsions of magnetised, electrified gas (plasma) from the atmosphere of the Sun. The results of space weather at Earth are, primarily, disturbances in our magnetic field and heating of our upper atmosphere. Some decades ago this was of little relevance to human existence, but now our lives are highly dependent on, for example, spacecraft orbiting in that heated upper atmosphere and electrical power distribution systems that are sensitive to those magnetic disturbances. Therefore understanding the physical causes and effects of space weather, and working towards predicting its occurrence, is of central importance to our technological society.
Solar flares are dramatic and short-lived flashes of radiation in the solar atmosphere. The total energy released in the form of flare radiation can be on the order of 1025Joules, the energy equivalent to 2.5 billion megatons of TNT. This is provided by restructuring of stressed magnetic field in the outer atmosphere of the Sun (the corona) to a lower energy state. Solar flares are very often closely associated with CMEs, and the most intense radiation – which occurs in the optical and UV parts of the spectrum – is associated with, and likely caused by, substantial numbers of highly accelerated non-thermal particles. The centrally important question of flare physics is to understand how the energy stored in stressed coronal magnetic field is released so rapidly, transported through the atmosphere of the Sun, and converted – with very high efficiency – into the kinetic energy of the non-thermal particles and thus, or otherwise, into the flare’s radiation output. This is a question that must be answered by putting together a broad range of observational and theoretical information. In the F-CHROMA project we focused on understanding how the transport and conversion of the flare energy can be deduced from observed flare radiation.
Solar flares were first observed in the optical in 1859 as rapid brightenings in white light from the solar surface, or photosphere. The development in the late 19th century of the spectroheliograph, which allows images to be made in isolated regions of the spectrum, revealed major disturbances in spectral lines now recognised to come from the narrow, complex interface between the photosphere and corona, the chromosphere. In particular, the 3-2 transition of hydrogen, Hα, characteristic of a plasma at ~10,000 K was extremely bright, and indeed analysis of this radiation is one of the main ways that we can understand what is happening in a flare. At that time flares were assumed to be entirely a chromospheric phenomenon, based solely on the energy implied by the enormous optical increases. However, with the associated sudden ionospheric disturbances detected by ground-based magnetometers, it became clear that flares involve substantial increases in ionising radiation at UV to X-ray wavelengths, verified by the dramatic, dynamical evolution of the coronal portion of the flare as seen from space in extreme UV and soft X-ray. Substantial advances in understanding of the magnetic and thermal evolution of flare plasma have been made by focusing on these energy ranges and on the coronal behaviour.
Still, the solar chromosphere remains of primary interest during flares. It is here that the majority of the flare’s radiative output originates, making it the main source of flare diagnostic information. Using chromospheric radiation we can deduce flare heating and ionisation, and subsequent cooling and de-excitation of different layers of the atmosphere. We can also measure plasma flows and non-thermal velocities, and infer character of accelerated particle distributions. The energy deposition and its variation in space and time can in principle be understood from the chromosphere and its observed evolution. However the chromosphere is a very complicated environment, so doing this successfully requires forward modeling – using first principles physical simulations to predict the radiation signatures arising from energy input in different forms. Novel developments in radiation hydrodynamical modeling address this problem, calculating the evolution of the chromospheric environment, and from this predicting the radiation signatures that allow a direct link from observables to physical conditions. Besides providing crucial flare diagnostics, the chromosphere may even be intimately related to the acceleration of the non-thermal electrons and ions that are a defining flare characteristic.
Our progress in this area has also been accelerated greatly by the acquisition of excellent datasets from both space-based and ground-based observatories. These facilities are observing at sub-arcsecond scales, corresponding to physical sizes on the Sun of 500km or less. This level of detail is probably still insufficient to capture the elementary processes happening on the Sun, but we do begin to have hints that important structures in space and time are not too far below our current capabilities. The flaring chromosphere emits across the entire electromagnetic spectrum, with different wavelengths characteristic of different temperature and locations (or heights, in a one-dimensional model) so that combining radiation signatures across wavelengths allows us, painstakingly, to build up a complete picture of its structure. No less important, but more difficult to interpret, is the information about the magnetic field in a solar flare, at the level of the photosphere and chromosphere, which can be deduced from the emitted radiation. We are just at the beginning of measuring and interpreting this emission,
The Sun is not the only star that produces flares. Indeed it seems to be a rather inactive star compared to many others of its own and other spectral classes (colour). However, it is the only star where we can directly image the flare and hope to understand how it works. By doing so, we may also extrapolate and apply our knowledge to other stars. Stellar flares have become a subject of great interest in the last few years as it becomes clear first of all how many stars have planetary systems around them, and secondly what the influences of the short wavelength radiation emitted during flares are both for the creation and destruction of conditions favourable to the production of pre-biotic molecules.
Following a long gap during which studies of the flaring corona were predominant, the central importance of the chromosphere in the energetics and dynamics of a flare (and the Sun as a whole) is being reasserted. Our project has capitalized on the renewed interest and investment worldwide in this fascinating region, leading to a multi-mode, multi-wavelength study of the solar flare chromosphere, supported by dedicated theoretical modeling. This concerted international effort in joint science exploitation of space-based and ground-based data, theory and numerical simulation has led to significant advances in our understanding of the physics of solar flares, and a roadmap for further developing this in the future.
Project Results:
(Please note - Figure numbers found in this report refer to the pdf version uploaded separately)
1 Main Scientific Results
1.1 Overview
Solar flares are sudden and dramatic outbursts of intense radiation from the solar atmosphere, caused by the release of enormous amounts of energy from a stressed magnetic field. Our project’s top-level scientific goals were to use observations and theory to understand how energy moves through the solar atmosphere following its release, to predict the spectrum of radiation that is emitted during a flare, and to identify how the next generation of solar observatories can best be used to further explore the basic physics of solar flares and, by extension, flares on other stars. We also aimed to promote the use of the best available models and observations in the scientific community, to grow the number of scientists, especially early career scientists, who are expert in the analysis of flares, and to involve amateur astronomers in the observation of these exciting and – so far – unpredictable events. We have made significant progress in all of these areas.
We concentrated on the chromosphere as this is the layer in which the majority of the solar flare radiation originates. The chromospheric spectrum can be probed in detail to reveal the changes, in time and space, of properties such as temperature, density and ionisation state that result from flare excitation. Thereby, the physical processes that lead to these changes can be picked apart. The non-flaring solar chromosphere is already a complex environment, and the flaring chromosphere presents additional complications in that it is (i) difficult to ob-serve due to the rarity and unpredictability of flare events, and (ii) in a highly non-equilibrium condition because of the intense and transient input of energy, which makes the physics challenging to model. Therefore we assembled a team with experience in planning and exe-cuting flare observations, in analysing all kinds of solar observations from the infrared to the X-rays, and in designing and running state-of-the-art numerical simulations, in an integrated approach. A focussed effort on the physics of the flaring solar chromosphere had never been undertaken before at this level. F-CHROMA has also prepared European and international flare physicists for the exploitation of major upcoming facilities such as ESA’s Solar Orbiter Mission with a launch foreseen in 2018, the Daniel K. Inouye Solar Telescope (DKIST) which sees first light in 2019, and now the European Solar Telescope (EST) which was adopted onto the European Research Infrastructures roadmap in 2016, with planned start of operations in 2026.
Based on research that has taken place in F-CHROMA, and before, we can formulate a top-level view of the evolution of the chromosphere during a flare.
We have established that energy can be delivered to the flare chromosphere either by electron beams or by Alfvén waves. Alfvén waves are capable of penetrating deeper into the chromosphere and exciting the temperature minimum region (Russell & Fletcher 2013), while electron beams of some 10s of keV energy deposit their energy higher up. Energy is also transported by thermal conduction from the upper chromosphere, which is heated to around 10MK (Graham et al. & Fletcher 2013). The main sink of energy in the flare chromosphere is the latent heat of ionisation of both hydrogen and helium, processes that act as temporary ‘thermostats’. Hydrogen ionisation is never complete in the lower atmosphere, since the high density leads to rapid recombination producing Balmer and Paschen radiation, and continu-ous re-ionisation occurs because of ongoing heating. However, in the upper chromosphere recombination is not so rapid, and when He ionisation is complete there are no other effec-tive radiation loss channels in this optically-thin medium, so the plasma starts to expand causing both upflows and (slower) downflows. Above the He ionisation layer, optically-thin line radiation of incompletely ionised metals, coupled with conduction, is also insufficient to rid the plasma of its energy locally, resulting in rapid upwards expansion. All of the hot plasma is expanding: there is no ‘stationary component’ at 10MK
In the moving plasma, optically thin lines are formed (including, surprisingly, redshifted opti-cally thin components of cool lines at ~8,000K) and provide a tool to reliably derive the amplitude and direction of the flows. Often, especially in slightly later phases of the flare, there is also sufficient absorbing material above the layer at which the main chromospheric line core forms to produce either red or blue wing absorption, according to whether this material is up-flowing or downflowing. The flare continuum is primarily due to hydrogen recombination emission. This will also lead to some backwarming of the photosphere, but models show this is unlikely to be significant. In more detail, some of the most significant advances and discoveries we have made are as follows:
• Discovered that the heated (10MK) chromospheric plasma is entirely upflowing, and the flows evolve in an identical way across numerous sources, apparently irrespective of energy input rate (Graham & Cauzzi 2015);
• Made new observations of the flare continuum, and guided by modeling, identified these as predominantly due to recombination emission. This emphasises the importance of ionisation in a flare (Heinzel & Kleint 2014; Simões et al. 2017 in prep.);
• Demonstrated that optically thick spectral line profiles that appear to show redshifted emission may in fact result from blueshifted absorption, indicating that naïve Doppler interpretations of lineshifts can be misleading (Kuridze et al. 2015, 2016; Kerr et al 2015).
• Made the first ever simultaneous observations of a flare in Hα and Hβ (Capparelli et al. 2017 in prep)
• Cross-validated, for the first time, two completely independent flare radiation hydro-dynamics codes, RADYN and Flarix (Kasparova et al. 2017 in prep), and promoted the use of the former as a mainstream analysis tool.
• Demonstrated the viability of Alfvén wave energy transport model as an alternative to the long-standing electron beam model for producing chromospheric flare radiation (Kerr et al. 2016)
• Found a flare where the absence of hydrogen Balmer lines may rule out any significant role for electrons in impulsive phase energy transport (Procházka et al. 2017)
In addition to these scientific advances, our success overall can be assessed by revisiting the stated scientific objectives of the F-CHROMA proposal:
(i) Perform joint observations and analysis of space-based and ground-based observations
of flares, delivering improved diagnostics for conditions in the flare chromosphere;
A dedicated and pro-active campaign of proposing for every ground-based opportunity resulted in 31 new ground-based flare datasets, most with accompanying space-based data. Analysis has commenced on many of these datasets, with 14 papers published and in preparation. The observing plans were constructed by F-CHROMA members with the goal of capturing, with both imaging and spectroscopy, the impulsive energy release phase of the flare.
(ii) Carry out detailed modeling of these diagnostics for a range of flare energy inputs
Our observations have stimulated targeted modelling efforts of both line and continuum radiation in a beam-heated atmosphere. Examples of this are Mg II h & k lines, the Ca II 8542 Å line, hydrogen and helium lines and continua, and the infrared free-free continuum. The range of flare energy inputs is reflected both in modelling electron beam energy input with a broad range of particle beam parameters, and exploring different energy in-put mechanisms, i.e. particle beams versus Alfvén waves. To facilitate comparisons, we have also worked on improved atomic data for Mg II and other ions.
(iii) Utilise forward modeling techniques to determine the physical properties of the solar
chromosphere during solar flares;
We have addressed this in two ways. We have generated a suite of numerical models covering a range of beam input parameters to investigate the range of physical properties that occur. We have used these models on their own to understand the likely conditions pertaining in the chromosphere. An example of this is our work on the locations at which the optical and infrared continua are formed in flares (Simões et al. 2017 in prep). We have also made comparisons between model predictions and observations, which have led to as much understanding about why the two fail to agree, as where they do agree. For example, we see that to reproduce the observed Mg II line profiles, with a non-reversed core, requires a higher-density upper chromosphere than is predicted by electron beam models. But the models have additionally shed light on unexpected ob-served properties, such as the two-component line profiles in some cool chromospheric lines, which are interpreted as due to an optically thick, stationary source, and an overlying optically thin downward moving chromospheric condensation.
(iv) Use models and observations to optimize targeted flare observing sequences for present
and future missions and observatories, hence maximizing their scientific return;
Based on our experiences with F-CHROMA observing campaigns, and on our existing analyses and modeling we can make some strong recommendations for different future observations targeted at understanding flare energy transport: (i) we have demonstrated the importance of ‘sit-and-stare’ with a single active region target, rather than target-hopping. This optimises the chances of capturing observations jointly with space-missions, and provides a continuous record before, during and after an event, allowing the rapid evolution of the chromosphere to be tracked; (ii) it is clear that, in general, tem-poral evolution is key, as the radiation from the transient initial phase of flares encodes the most information about the evolution of the chromosphere; (iii) that the spectropolar-imetric instruments on DKIST should have frequent use as flare spectrometers only, sac-rificing the polarimetry for rapid spectroscopic line scans to capture the line profile evolution – on timescales of a second – at more than one spectral line: (iv) that to progress on understanding flare chromospheric dynamics, linking the rapid upflows seen in hot lines from the Interface Region Imaging Spectrometer (IRIS) and cool lines from both IRIS and the ground, we need to focus on lines formed at temperatures of a few x 105 K to 106 K, which will be present in the long passband for the SPICE spectrometer on Solar Orbiter.
(v) produce a catalogue of existing and future ground-based solar flare observations, linked
to available space-based counterparts;
The catalogue of ground-based solar flare observations was published and announced within the first year of the F-CHROMA project on 01 December 2014, and continues to be updated. The 534 records contain information on coverage, contacts who will provide data, and any known publications on the ground-based data.
(vi) populate an archive of ground-based flare observations and make it available to the community;
Many of the new ground-based datasets obtained during F-CHROMA campaigns have been placed in our public archive, as a resource for the community. The link is availa-ble from the F-CHROMA catalogue web page, and data in FITS format can be down-loaded. The public archive at present contains 33 flare records, 22 of which arise from F-CHROMA campaigns. This reflects the rarity of good ground-based flare observations (even with our concerted efforts to improve this!) and it is to our knowledge the first attempt to make ground-based flare data systematically available for analysis. We hope that this resource will encourage and facilitate familiarisation with the next generation of ground-based solar data, for example by providing datasets for students to work on.
(vii) produce a flare spectral model database and make it available to the community.
Following testing of the RADYN and RH codes as part of F-CHROMA, and the im-provement of some aspects of the RADYN computation, we have populated a data-base of beam-driven radiation hydrodynamics simulations, representing a range of total energy, electron spectral index and low-energy cutoff. The model database is linked from our F-CHROMA website. Instead of producing the flare spectral model database, we have made available the full simulation results, and the code, and guide, needed to interrogate those data to produce spectra, as it is the most flexible approach to making the output available to users.
It is impossible to cover all of the activities and results of F-CHROMA, and this report is therefore selective, focusing on the areas where perhaps the most surprising or substantial advances have been made. The numerous outputs arising from the project throughout its duration and beyond can be seen in the 55 refereed publications and 100 presentations linked at http://www.fchroma.org/?page_id=33 and http://www.fchroma.org/?page_id=219.
1.2 New observations
High quality solar flare observations from the ground are rare. Therefore we had a highly pro-active programme of proposing for every ground-based observing opportunity, including with the newest facilities (GREGOR). This resulted in 31 new ground-based flare data sets many of which can be found at https://star.pst.qub.ac.uk/webdav/public/fchroma/ Most have ac-companying space-based data. Analysis has commenced on many of these datasets, with 14 papers in preparation. The observing plans constructed by F-CHROMA team members had the goal of capturing, with both imaging and spectroscopy, the impulsive energy release phase of the flare.
1.2.1 The F-CHROMA flare campaigns
In total there were 23 F-CHROMA organised observing periods throughout 2014-2016. The majority of these were allocated in a competitive process on major facilities such as the Dunn Solar Telescope and the Swedish Solar Telescope. These were run both in service mode, and in observer mode. The remainder of the observing periods were carried out at facilities controlled by F-CHROMA team members, namely the Multichannel Flare Spectrograph at the Ondrejov Observatory (ASU), and the Multichannel Subtractive Double Pass Spectrograph at the Large Coronagraph and Horizontal Telescope at Bialkow (UWR). In total 31 flares were observed from these facilities, in whole or in part. A list of campaigns and flares captured can be found in Appendix 1. Particularly encouraging was the enthusiastic support from space-based observatories for our efforts, with a large number of events observed also by the Interface Region Imaging Spectrometer (IRIS) spacecraft, which focuses on the chromosphere. Several of the observing periods covered the two F-CHROMA F-HUNTERS pro-am observing campaigns.
Around 20 new high-quality flare datasets from the campaigns have been formatted and placed in the F-CHROMA data archive, which is currently at a volume of over 400GB of flare data. However, this is an on-going process, which will continue over the coming year. This resource for the community will, we hope, encourage and facilitate familiarisation ground-based solar data in preparation for DKIST and the European Solar Telescope.
1.2.2 New types of flare observations
The campaigns resulted in new types of flare observations, using facilities which had not previously been used for flares, such as the Facility InfraRed Spectrometer (FIRS) infrared imaging spectropolarimeter at the Dunn Solar Telescope, and the Crisp Imaging SpectroPolarimeter (CRISP) optical imaging spectropolarimeter at the Swedish Solar Telescope. A new instrument was also developed at Ondrejov, to capture the (spatially-integrated) flare optical spectrum in the range 350-440nm (Figure 3). This includes the Balmer jump at 364.6nm cor-responding to the head of the free-bound hydrogen recombination continuum onto level 2. The Image Selector allows a spectrum to be taken over a field-of-view corresponding to an active region, thus increasing signal-to-background compared with ‘Sun-as-a-Star’ instruments, and improving the chances of observing a flare compared to slit spectrometers. The instrument has sub-second temporal resolution, and was used with success (Kotrč et al 2016) to capture the X 1.0 flare SOL2014-06-11T09:06.
It is very surprising to see in these observations that in some flares neither the Balmer jump nor the low-order Balmer are visible (nor are the Lyman lines, observable by EVE). This is problematic to explain with the standard electron-beam heated model (Procházka et al. 2017). The instrument is now being upgraded to detect the hydrogen Paschen continuum.
Existing instruments were also used in original ways to optimise flare science, and – as is usual when novel setups are tried – to make discoveries. For example, on the Dunn Solar Telescope the ROSA high-speed imager ran in Hβ, and the IBIS imaging spectrometer in Hα, obtaining to the best of our knowledge the only joint observations in these two important flare lines (see Section 3.4.2.3). They reveal the exceptionally high degree of temporal correlation between these lines, apart from during the critical first few seconds of the event. The IRIS spectrometer was run as fast as possible in ‘sit-and-stare’ mode, obtaining the highest temporal resolution of flare evaporation and condensation flows yet made, in individual flare pixels (Graham & Cauzzi 2015). Dedicated flare spectropolarimetric observations were also ob-tained in the early decay phase of the X1.6 event SOL2014-10-22T14:02 using IBIS and these are still being analysed (Guglielmino et al. 2017 in prep).
1.2.3 Lessons learned and recommendations for future observations
In planning and carrying out these observations we learned what is most important for ad-vancing our understanding of the physics of the chromosphere during the impulsive energy release phase of a flare. This does not address all aspects of impulsive-phase flare physics of course – one can easily imagine scenarios where the large-scale spatial evolution is needed to address the physical question of interest. But focusing on the energy transport and dissipation we know now that the following are vital:
(i) pre-flare coverage: this is necessary to characterise the state of the chromosphere before energy input begins, as well as to capture the very earliest stages of the impulsive phase. ‘Flare triggers’, to switch telescope mode or pointing, are unlikely to be useful in this respect.
(ii) high temporal resolution: Almost everything else should be sacrificed to obtain the highest-possible temporal resolution, because the chromosphere responds on timescales of se-conds to energy input, and observationally many of the strong spectral lines seem to reach a ‘saturated’ state after a few seconds. Theoretically, this may be because the level popula-tions approach their local thermodynamic equilibrium values because of high collisionality. At this point, discriminating information about the energy input is harder to obtain.
(iii) spectroscopy: resolving the details of the spectral line profiles is important for under-standing plasma flows at different layers in the chromosphere. In the optical, good sampling (~0.1Å) for around ± 1Å in the line core is required, as the core moves during flares. In the UV and EUV, spectral windows around strong, hot lines such as Fe XXI should be extended into the blue to capture the earliest phases of the flow dynamics, where blueshifts corresponding to more than 270km/s are detected.
The above constraints point us inevitably in the direction of “sit-and-stare” type observations, where a target is chosen and tracked for a long period. The spectrometer slit (if applicable) should be placed as close as possible to perpendicular to the polarity inversion line, and rasters, if any, should be small to optimise the time cadence. It is recognised that it may be diffi-cult to convince the operators of large and competitive facilities such as DKIST and Solar Orbiter to run what appears to be restrictive programmes of this type, with the risk of accu-mulating a large amount of data that is not useful for flare studies. But we are convinced that this strategy offer the best scientific return for impulsive phase studies
1.2.4 Analysis arising from the F-CHROMA flare observations
In Table 1 in the uploaded pdf we give a list of papers published or in prep that use data gathered during F-CHROMA organised campaigns, or other campaigns involving F-CHROMA scientists. As can be seen, these are primarily from the 2014 observing campaigns so far. The successful Sep-tember/October 2015 campaigns, and a couple of flares from 2016, will also be analysed in the future.
1.3 New flare data analysis
There are numerous new multi-wavelength data analysis results, contained in (to date) 35 published or accepted journal articles, with 12 more manuscripts in a late stage of prepara-tion. We focus here on a few important to have emerged from this analysis.
1.3.1 Hydrogen continuum emission from IRIS and from the ground
The hydrogen Balmer continuum is the spectral range between the Balmer jump at 3646Å, and (nominally) the Lyman jump at 912Å. The true Balmer continuum is the recombination emission onto level 2 from hydrogen atoms transiently ionised during a flare. In this broad range one will also find blackbody emission and other continua such as the optically-thick silicon and carbon continua. However, in the near UV channel of IRIS these metal continua are weak, and, by sampling the spectrum between the strong emission lines Heinzel & Kleint (2014) were able to measure the Balmer continuum in the well-observed flare SOL2014-03-29T17:48 and demonstrate that it is consistent with the expected recombination intensity. Kleint et al. (2016) extended this analysis to the IRIS far UV channel, the optical channel on the Helioseismic and Magnetic Imager (HMI) and the FIRS infrared spectrometer on the Dunn Solar Telescope (DST), extending the range to include the Paschen continuum (re-combination onto level 3) showing that the spectra thus generated are not consistent with an enhanced photospheric blackbody alone (Figure 4). That the flare continuum emission in-volves recombination in this large event now seems conclusive, and points to the central role of hydrogen ionisation in the flare energetics. There is, nonetheless, a requirement for an enhanced (by a few 100K) photospheric blackbody to explain the radiation intensity, and this would require a very intense heating of the lower atmosphere.
It has long been known that the Balmer recombination edge is often ‘missing’ in solar flare observations. This was reported by Kowalski et al (2015) and may be interpreted as the edge being masked by a ‘piling up’ of broadened, high-order Balmer lines into a bump. A very unusual observation of an X1.0 flare was made with the novel flare broadband spectrometer described in Section 1.2.2 in which the continuum enhancement was strong, with the ‘pile up’ feature present, but with little to no detectable low order Balmer lines (Procházka et al. 2017). These should be at 434.1nm 410.2 nm, 397.0nm 388.9nm and 383.5nm in Figure 5. This can only be explained by heating of the deep atmosphere without the overlying atmosphere being excited sufficiently to generate hydrogen line enhancement. This is not possible with ‘normal’ electron beams, i.e. with low-energy cutoffs in the 10s of keV range. It suggests that we should be looking for explanations in terms of proton beams or Alfven wave heating which may be able to heat the deep atmosphere, leaving the upper chromosphere relatively unaffected.
1.3.2 Hot plasma upflows
The response of the flare atmosphere to intense heating is to expand upwards, with a momentum-balancing downwards counterpart. One of the most interesting F-CHROMA results was on this phenomenon from Graham & Cauzzi (2015). They analysed IRIS spectra including the hot line of Fe XXI at 1354Å and found very high blueshifts, and very strong (non-thermal) broadening. The measurements were made at 86 individual flare ribbon pixels which were excited in sequence and observed as the flare ribbon moved down the spectrometer slit.
The velocity profile as a function of time in each pixel was obtained (Figure 6) and perhaps the most surprising aspect of the observation is that if the velocity profiles as a function of time are stacked in time, with the base being the time corresponding to the first point at which the enhanced Fe XXI signature becomes detectable, the majority of the velocity pro-files overlay one another very closely. This suggests two main things: that the evolution of the evaporative flow is rather insensitive to the details of the atmosphere and to the strength of the energy input – within factors of a few at least. This might be consistent with a picture where the Fe XXI signature arises from heating only of the rather tenous upper part of the pre-flare chromosphere, and expands into a much lower density, “undisturbed” overlying co-rona. Upflows of over 100km/s are present for around 5-10 minutes though the input of energy is much shorter-lived. Numerical simulations of beam energy input tend to show flow velocities reducing to below 50km/s within a 60 seconds of the beam switching off (Kerr 2017), so the observed behaviour may imply continuing low-level energy input.
1.3.3 Line shapes and asymmetries in cool lines
Line shapes and asymmetries in lines formed at low temperatures have turned out to be extremely interesting for understanding the dynamics of the low atmosphere. Different wave-lengths in the spectral lines are formed in different regions, with the line core generally formed high up in the chromosphere (being more optically thick, the escaping photons come from higher levels), and the wings from low down. Work on Hα, (Kuridze et al. 2015), Na I (Kuridze et al. 2016) and Mg II h & k (Kerr et al. 2015, Liu et al. 2015) show a variety of line profiles, which evolve during the first few seconds of a flare from being (briefly) centrally re-versed to being non-reversed, and then back to centrally reversed again. The variety and evolution of these line profiles have prompted a concerted modelling effort, described in Section 1.4. A non-reversed emission line profile in these lines may indicate that the source function in the line core is strongly coupled to the Planck function due to a high density at the line formation height. The reversed profiles indicate decoupling of Planck and source functions, and also can have emission and absorption contributions from moving components, both optically thin and optically thick, overlying the optically thick part of the spectral line. Observed line shifts and asymmetries have tended to be interpreted in terms of flows at different atmospheric levels in a rather simplistic way, e.g. the ‘bisector’ method, in which the flow speed is inferred from the location of the line bisector at e.g. the 50% level in a flare line. However, the wavelength-dependent opacity effects, made obvious by RADYN modelling, make this approach potentially very misleading.
The high cadence of IRIS observations has been particularly useful in revealing the velocity evolution of the lower chromosphere. Corresponding to the high-speed upflows described in Section 1.3.2 the Mg II lines were used to identify downflows associated with chromospheric condensation; applying a bisector method, average downflows as high as 30km/s were ob-served, decelerating to zero in around 60s (Graham & Cauzzi 2015). The onset of these downflows appears to precede the onset of the upflows by around 60s, though this may be an issue of detectability of the upflow line profiles (e.g. starting very weak and blueshifted outside the IRIS passband). Of course the bisector method might be a too simplistic approach; examining the line profiles in greater detail it appears that the profiles in fact have two components, one almost stationary and narrow, and the other substantially redshifted and broadened. These start off as two quite well distinguished components, with the red-shifted component slowing and merging into the stationary component as time proceeds (Kowalski et al. 2017, Graham et al. 2017 in prep). The interpretation of this is discussed in Section 1.4.2 but it seems to give a potential direct window onto the flow field of the upper chromosphere.
1.4 New modelling
1.4.1 Comparison of RADYN and Flarix simulations
One of the most important activities and major activities carried out by the F-CHROMA team was a comparison of two different flare radiative hydrodynamics codes (Kasparova et al. 2017 in prep). It is important to recognise that such cross-validations are rarely done in any kind of computations in solar physics; perhaps the authors of codes prefer not to confront the codes’ possible shortcomings. An attempt to cross compare flare hydrodynamics codes was carried out in the 1980s during the SMM days (Hudson, private communication) but was not carried through. We believe that our comparison of the RADYN code (Carlsson & Stein 1992, Abbett & Hawley 1999), originating primarily from the UIO and the Flarix code (Heinzel et al, 2016) originating from the ASU, is the first of its kind. The two codes have completely different numerical schemes, and different approaches to solving the hydrodynamics of the plasma and the non-LTE level populations, though they do have a common approach to radiative transfer. The two codes were initiated with the same model atmosphere, and driven by the same input beam. Happily, the output model atmospheres track one another very closely (Figure 7), thus giving us confidence in using both codes. In practice the RADYN code is currently better documented, and has an output format which is well integrated into the RH package for computing spectral lines in detail, so we have used this in our model archive.
1.4.2 Interpretation of model output and discoveries arising from this
The careful comparison of observations and model outputs has been one of the most valuable exercises during the F-CHROMA project, not only for what the models can explain or help us understand, but also for what they currently fail to predict. Below are described a few common themes that have emerged from our analyses.
1.4.2.1 Understanding the line profiles in terms of flows
The shifts and asymmetries observed in the line spectrum of solar flares in Hα and Ca II 8542 Å by Kuridze at al. (2015) and in Mg II h & k by Kerr et al. (2015, 2017), Liu et al (2015) have been compared with synthesised line profiles produced with RADYN and RH. To obtain reliable simulations, the flare parameters used in the model were estimated from the HXR data. In the Kuridze et al. study, the temporal evolution of the observed and simulated Hα line profiles show excess emission in the red wing (red asymmetry) before flare maximum and excess in the blue wing (blue asymmetry) after maximum. The multicomponent velocity field in the chromosphere creates differences in the opacity between the red and blue wings of Hα, which produces the asymmetric chromospheric emissions. At early phases of the flare, material upflows absorb emission in the blue wing, creating the red asymmetry, while at later phases, emission in the red wing is absorbed by plasma downflows, reversing the asymmetry to blue, confirming an idea proposed by Heinzel et al. (1994). This indicates that Hα line shifts as determined by bisectors need to be carefully analysed and contrasted with simulations to properly derive flow speed and directions
The work by Graham et al (2017, in prep) and Kowalski et al (2017) reveal another interest-ing feature; low-temperature spectral lines such as Fe II 2815.45Å, Si II 1348.55Å, CI 1354.28Å, C II 1335Å as well as the Mg II subordinate line at 2791.6Å are each composed of a narrow, almost static component and a broader redshifted component (See Figure 8.) The interpretation, backed up by modeling, is that the narrow, stationary component is optically thick and originates from the deep atmosphere, whereas the broader redshifted component originates from an optically thin, overlying chromospheric condensation, at temperatures be-tween around 8,000 and 25,000K. This is supported qualitatively by RADYN modeling, though some of the details differ. If the broader redshifted component is indeed optically thin, it offers a completely new possibility to construct the flow and turbulence field at line formation temperatures in the few to 10 kK range.
1.4.2.2 Beam driven versus wave driven simulations and comparison with Mg II lines
In the Kerr et al (2015, 2016) studies of Mg II h & k the line profiles throughout most of the flare have a small red asymmetry, and no clear central reversal. RADYN simulations using a beam energy input always produce a rather symmetric line profile with no asymmetry, whereas running RADYN with an Alfven wave energy input gives a line profile that is more similar to the observations, due to a different profile of energy deposition through the atmosphere (Figure 9).
1.4.2.3 Evolution and saturation of the hydrogen lines in flares
Observationally, one finds an initial transient behaviour in the ‘Lyman decrements’ (ratios of Lyman line intensities) at the beginning of a flare, but thereafter they become rather constant, even as the energy input changes substantially (M. Kennedy, PhD Thesis). This behaviour is seen in several flares, but there is a small flare-to-flare variation in the average ratios. A small grid of RADYN models with quite extreme ranges of beam parameters showed a similar pattern of transient evolution followed by flat ratios (though the ratios were systematically below the observed ratios, possibly because the events are a superposition of short-lived, high ratio values in the initial transient phases of a series of injections, and longer-lasting low ratio values). The reason for the constant ratio values seems to be that the high collisionality, due to high densities and temperatures at the Lyman line formation heights, drives the level populations close to their LTE values. This is also the case for the ratio of Hα and Hβ (Capparelli et al. 2017, in prep, Figure 10). It was hoped that these line ratios would track the evolution of mean electron temperatures around where the line cores form, but it seems that this will only be possible, if at all, during the initial few seconds of a flare.
1.4.2.4 Interpretation of the continuum
FLARIX simulations were also used to synthesize the hydrogen recombination continua. The Balmer continuum was compared to IRIS observations (note that the first detection of the Balmer continuum in IRIS flare spectra was one of important results of F-CHROMA, see first period report). The Paschen continuum was then compared to SDO/HMI observations of both disk and limb flares. It was also shown that at the flare peak, the radiation losses are dominated by hydrogen and namely by the Balmer continuum and thus Balmer-continuum detection in the IRIS spectra represents a critical constraint for the modelling.
Recent high-cadence observations in the infrared at 5.2μm and 8.2μm, made at the McMath-Pierce telescope at Kitt Peak National Observatory showed well-defined, high contrast chromospheric footpoints. This stimulated work on the physical processes leading to the IR emission, RADYN was used to examine in detail the formation of optical and IR continuum, by calculating the contribution function in a number of spectral bands, including for the first time the electron free-free and the H- free-free opacity (Figure 10). The flare continuum enhancement is mostly formed in the chromosphere due to hydrogen recombination: Balmer in NUV, Paschen in optical, Brackett/Pfund in the near-IR (λ<2µm), and the evolution of the ionisation so the ‘total electron content’ of the chromosphere turns out to be critical. For all wavelengths, the contribution function has an excess in the chromosphere during the simulated flare and the emission is optically thin, while the photosphere is mostly undisturbed, so continuum enhancement from H- opacity is unimportant. The same occurs for all simulated cases, for different properties of the electron beam. Only the chromospheric Balmer continuum is strong enough to cause a minor photospheric temperature increase – consistent with the findings of Kleint et al. (2016) but, in the beam models used, this is still not sufficient to produce a significant increase in the continuum emission. These results strongly indicate that NUV, optical (white light) and NIR continua are primarily a result of hydrogen recombination in the chromosphere (Simões et al. 2017 in prep)
1.4.2.5 Where models and data do not match
There are areas where the modeling and the data seem still to be in disagreement, and the-se common features will guide our thinking and motivate developments in future modeling. These common features are: (i) the ‘filling in’ of normally reversed line cores, which requires a higher pressure corona/transition region than is predicted in the RADYN electron-beam-driven simulations; (ii) enhanced far wing and continuum emission requiring heating of deep layers of the atmosphere, again more so than predicted by the RADYN electron-beam simulations; (iii) line broadening, which requires an increased turbulent broadening, probably one that varies through the atmosphere. In addition to more work on beam-driven simulations, models involving Alfvén wave energy transport may offer solutions to some of these discrepancies.
1.4.3 2D and 3D models
Papers have been published on 2D radiation transfer in MHD models with the FLASH code, and with the MALI code. Adding a second dimension to radiative transfer naturally produces fibril-like structures in cool flare loops (Falewicz et al. 2015) and can produce effects in the line profiles – such as the appearance of the second, shifted component that is seen in flare loop observations (Heinzel et al. 2017 in prep.) The PENCIL code was used to model the evolution of a delta spot in 3D (Chatterjee et al. 2016, Figure 11), and very localised, rapidly evolving regions of high temperature were identified, very suggestive of flaring energy re-lease. It is too early to draw general conclusions about the importance of 2D or 3D radiation transfer, but the fact that these highly-developed MHD codes show evidence of flare-like activity is very encouraging for the future development of flare modeling.
1.4.4 Provision of grid of models
The grid of models uses the RADYN code, which has been validated against Flarix. It is now published at https://star.pst.qub.ac.uk/wiki/doku.php/public/solarmodels/start. These models are already being used by PhD students in Glasgow, and attendees at the F-CHROMA train-ing school were also introduced to their use. The RADYN code used incorporates a Fokker-Planck beam solver, a new starting model atmosphere resembling the VAL3C atmosphere, better resolution of gradients, and a more finely-resolved frequency grid across the lines. The injected beam has a triangular, symmetric time profile of 20s duration, to represent an ‘elementary flare burst’. The beam power law index, low energy cutoff, and integrated flux are all varied, resulting in a grid of 96 models. The output is in Common Data Format (CDF) and can be examined to look at the evolution of atmospheric properties (temperature, density, velocity, ionisation and atomic excitation levels) and also the formation and evolution of spectral line profiles and continuum for several strong lines and continua including several energetically important lines and continua of H and He, and Mg II h & k, Ca II lines. A full list is given in Allred et al. (2015). For lines not calculated by RADYN the model atmosphere output can be ingested into full radiation transfer codes such as the RH code.
Documentation and analysis software is available at http://folk.uio.no/matsc/radyn/tools.tar
1.5. References
1. Abbett, W. P. & Hawley, S. L. Dynamic Models of Optical Emission in Impulsive Solar Flares, Ap. J. 521, 906 (1999)
2. Allred, J. C., Kowalski, A. F., Carlsson, M., A Unified Computational Model for Solar and Stel-lar Flares, ApJ 809, 104 (2015)
3. Capparelli, V. Zuccarello, F., Romano, P., Simoes, P. J. A., Fletcher L. et al. Hα and Hβ emission in a C3.3 solar flare: Comparison between observations and simulations, ApJ 2017 (in prep).
4. Carlsson, M. & Stein, R. F., Non-LTE radiating acoustic shocks and CA II K2V bright points, ApJ Letts. 397, L59 (1992)
5. Chatterjee, P., Hansteen, V., Carlsson, M., Modeling Repeatedly Flaring δ Sunspots Phys.Rev.Letts. Volume 116, Issue 10, id.101101
6. Falewicz, R., Rudawy, P. Murawski, K. Srivastava, A.K. 2D MHD and 1D HD models of a solar flare – a comprehensive comparison of the results, ApJ 813, 70 (2015)
7. Graham, D. R., Fletcher, L., Hannah, I. G. H., Milligan, R. O., The Emission Measure Distribution of Impulsive Phase Flare Footpoints ApJ 676, 83 (2013)
8. Graham, D. R., Cauzzi, G., Temporal Evolution of Multiple Evaporating Ribbon Sources in a So-lar Flare. ApJL, 807, 22 (2015).
9. Graham, D. R. et al., Spectral analysis and modelling of solar flare chromospheric condensation, (2017 in prep.)
10. Fletcher, L., Berlicki, A., Awasthi, A. K., Gronkiewicz, D. Recruiting flare hunters for citizen science, Astronomy & Geophysics 57 6.21 (2016).
11. Heinzel, P., Karlicky, M., Kotrc, P, Svestka, Z. 1994, On the occurrence of blue asymmetry in chromospheric flare spectra Sol. Phys. 152, 393
12. Heinzel, P., Kleint, L., Hydrogen Balmer Continuum in Solar Flares Detected by the Interface Region Imaging Spectrograph (IRIS), Astrophysical Journal Letters, 794, L23, (2014)
13. Heinzel, P., Kasparova, J., Varady, M. et al, Numerical RHS simulations of the flaring chromosphere with Flarix, arXiv:1602.00016 (2016)
14. Heinzel, P. et al.: 2D optically-thick radiation losses in cool flare loops (2017, in prep)
15. Kasparova, J. et al.: Mutual comparison between two RHD codes FLARIX and RADYN (2017 in prep)
16. Kerr, G. S., Simões, P. J.A. Qiu, J., Fletcher, L., IRIS observations of the Mg II H and K lines during a Solar Flare, Astronomy & Astrophysics 583, A50 (2015)
17. Kerr, G.S. Fletcher, L., Russell, A. J. B., Allred, J. C. Simulations of the Mh II k and Ca II 8542 lines from an Alfven Wave-Heated Flare Chromosphere, ApJ 827, 101 (2016)
18. Kerr, G., 2017, PhD Thesis, University of Glasgow.
19. Kleint, L.; Heinzel, P.; Judge, P. et al Continuum Enhancements in the Ultraviolet, the Visible and the Infrared during the X1 Flare on 2014 March 29, ApJ 816, 11 (2016)
20. Kotrč, P.; Procházka, O.; Heinzel, P. New Observations of Balmer Continuum Flux in Solar Flares. Instrument Description and First Results. Solar Physics 291, 229 (2016)
21. Kowalski, A. F., Cauzzi, G., Fletcher, L., Optical spectral observations of a flickering white-light kernel in a C1 solar flare, Astrophysical Journal 798, 107 (2015).
22. Kowalski, Adam F.; Allred, Joel C.; Daw, Adrian et al. The Atmospheric Response to High Non-thermal Electron Beam Fluxes in Solar Flares. I. Modeling the Brightest NUV Footpoints in the X1 Solar Flare of 2014 March 29 ApJ 836, 12 (2017)
23. Kuridze, D., Mathioudakis, M., Simoes, P. J. A. et al, H-alpha line profile asymmetries and the chromospheric flare velocity field, ApJ 813, 125 (2015)
24. Kuridze, D.; Mathioudakis, M.; Christian, D. J. et al, Observations and Simulations of the Na i D1 Line Profiles in an M-class Solar Flare ApJ 832, 147 (2016).
25. Liu, W.; Heinzel, P.; Kleint, L. Kasparova, J. Mg ii Lines Observed During the X-class Flare on 29 March 2014 by the Interface Region Imaging Spectrograph, Solar Physics 290, 3525 (2015)
26. Procházka, Ondřej; Milligan, Ryan O.; Allred, Joel C et al. Suppression of Hydrogen Emission in an X-class White-light Solar Flare (2017) accepted. arXiv170200638P
27. Ranjan, S. & Sasselov, D., Influence of the UV Environment on the Synthesis of Prebiotic Mole-cules Astrobiology 16, 68 (2016)
28. Russell, A. J. B. & Fletcher, L. Propagation of Alfvénic Waves from Corona to Chromosphere and Consequences for Solar Flares, ApJ 765, 81 (2013)
29. Simões, P. J. A. et al., Formation of the thermal infrared continuum in solar flares (2017, in prep)
Potential Impact:
2 Impact of the Project
2.1 Promoting the importance of the flaring chromosphere
We have worked hard to disseminate both the science of F-CHROMA, with a list of 55 (and growing) published, refereed papers. These have been cited by over 180 other papers so far, with a median number of downloads per published paper of over 100. These numbers are very healthy given the short lifetime of the project. Full metrics on the refereed papers can be monitored at https://ui.adsabs.harvard.edu/#/public-libraries/ANK_KO_BRuCqeYop3ZOFlw
Members of the F-CHROMA team gave (at least) 100 scientific talks and posters about the flare chromosphere and closely related physics at national and international events: a list of these is given in the “Use and Dissemination of Foreground” section of this report. In addition, dissemination activities specifically dedicated to advertising F-CHROMA and its outputs took place at the following events:
• ‘Solar and Stellar Flares’ Prague, Czech Republic, June 2014 (Poster, Fletcher et al)
• AAS Solar Physics Division Meeting, Boston, USA, June 2014 (Poster, Cauzzi et al)
• ‘RadioSun-5’, Ceske Budejovice, Czech Republic, 2016 (Poster, Berlicki et al)
• ‘CoolStars 19’, Uppsala, Sweden, 2016 (Talk, Berlicki et al)
• ‘Hinode 10’ Nagoya, Japan September 2016 (Carlyle & Carlsson)
Via our roles on the Science Working Group of the DKIST we have also emphasised the types of observations that can be carried out to probe the physics of the flaring chromo-sphere, and we have recently proposed to hold one of the first DKIST Community Work-shops with the aim of developing chromospheric flare science use cases.
2.2 Impact beyond solar physics
The Sun is our nearest active star, but observations, particularly from the Kepler satellite, show that magnetically-driven flaring is a common phenomenon in solar-type stars and stars of other spectral classes, particularly the M stars. The spectrum of stellar flares can be used, as in the case of solar flares, to diagnose the properties of the flaring atmosphere of stars, and our work will be useful here too in interpreting stellar observations, as the degree of spatial resolution allowed in solar observations (e.g. measuring velocities in individual pixels) breaks the degeneracy between space and time in a flare evolution in a way that is not accessible for stars. The irradiance spectrum is also significant for exoplanet studies. On the Earth, flare-related fluctuations in the solar irradiance in the UV-soft X-ray causes small transient ionisation changes in our ionosphere, which can temporarily change the upper atmosphere chemistry. But in exoplanets much closer to their hyper-active host stars than we are the changes to the planetary atmosphere might be very much more dramatic and long-lived, with implications for the stability of the atmosphere. Interestingly, the wavelength region be-tween 200nm and 300nm is thought to be critical for the formation of pre-biotic molecules and accessible through an early-Earth-like atmosphere (Ranjan & Sasselov 2016), so the use and development of RADYN simulations to quantify this irradiance could have influence in this exciting field.
2.3 Providing resources for the community
The major resources provided for the community have been the flare catalogue, the ground-based flare data archive, and the archive of RADYN flare models, as described in Section 1.4.4. The number of unique loads (unique users) of the start page of the flare catalogue and database in the first 20 months after its release was 1032, indicating a good level of interest. Top hits were from the UK, USA (reflecting, we believe, the announcements made in the national Solar Physics Newsletters) followed by Germany, Japan, India and China. The model archive is already being used by PhD students and postdocs within the F-CHROMA team members, greatly facilitating their research by avoiding the need to repeatedly run RADYN simulations. We anticipate many more users once the model archive is announced.
2.4 Training young scientists
A core dissemination activity was the organisation of a very successful F-CHROMA training school in the late autumn of 2016. It was dedicated to PhD students and young researchers to provide training on the physics of solar flares, flare data analysis, flare diagnostics and the use of the main modelling codes. The workshop was organized in Wroclaw (Poland). The full name of the workshop was: "F-CHROMA Training Workshop: Observations and Modelling of Solar Flares". Workshop information can be found at: http://school2016.astro.uni.wroc.pl/
We received 40 applications, which were carefully analysed by the SOC taking into account their scientific profile and their level. In the end we were able to accept 23 participants (14 female and 9 male) from 13 countries: Argentina, China, Czech Republic, Georgia, Greece, Ireland, Norway, Poland, Russia, Slovakia, Slovenia, United Kingdom, and the USA.
The workshop program reflected the F-CHROMA team expertise and F CHROMA team members, including some junior members, delivered all lectures and practical classes. All topics were divided into lectures and practical classes. The students also had the possibility to present themselves. There were some observational classes about data analysis and processing and more theoretical classes on the stationary and time-dependent modelling of the solar flaring atmosphere, including short introductions on the usage of main modelling codes.
At the end of the workshop, all participants, including lecturers, received the questionnaire to fill, to rate the workshop’s scientific level, organisation and other issues. The survey revealed that opinions about the workshop and its organisation were very positive and participants were satisfied. There was no clear difference between the opinions of male and female participants.
2.5 Involving amateurs
One of the most exciting parts of the F-CHROMA project was the organisation of two ‘professional-amateur’ (pro-am) flare observing campaigns. These were called "F-HUNTERS" and took place in 2015 and 2016:
• 1st F-HUNTERS Campaign: September 19-27, 2015
• 2nd F-HUNTERS Campaign: July 14-23, 2016.
The campaign periods were chosen to coincide with CHROMA team observing runs at ground-based observing facilities: 2015 F-HUNTERS#1: Dunn Solar Telescope at NSO (USA), Swedish Solar Telescope on La Palma (Spain), the MSDP at Bialkow Observatory (Poland) and the Multicamera Spectrograph, Ondrejov (Czech Republic); 2016 F HUNTERS#2: GREGOR Solar Telescope, Multi-channel Sunbtractive Double Pass Spectrograph, Bialkow Observatory (Poland) and Multichannel Flare Spectrograph, Ondrejov (Czech Republic). Observations by amateurs could therefore support the data obtained by professionals. In addition, the IRIS satellite dedicated some part of its time for a "flare watch" program giving additional spectral data in the UV range. Other space instruments (e.g. SDO, RHESSI) also contributed to these observations. Both campaigns were announced in a dedicated F-CHROMA dissemination site at http://fchroma.astro.uni.wroc.pl/index.php/f-hunters.html and were also announced on Facebook, Twitter and in many websites of amateur astronomers’ associations in many countries. There were also presentations and press announcements in local media (Figure 13).
The F-HUNTERS campaigns were coordinated via the F-CHROMA dissemination website. We prepared all information necessary for amateurs to observe and submit their data, including manuals with various levels of technical detail. All those interested in the F-HUNTERS campaign could register in the dissemination website in order to subscribe to the latest information with the observing solar targets during the campaign. For each day of the campaign, after analysis of solar images, the F-CHROMA team decided on the target for the flare watch on the following day. A special Information Sheet with the target details was prepared and sent to all subscribers by email (Figure 14). The same target info was published in Facebook, Twitter and in the F-CHROMA website. In total, we prepared 19 information sheets for all days of both campaigns. Professional observatories also observed the same target.
More than 150 participants registered on our F-CHROMA dissemination website and re-quested to receive the Information Sheet with solar observing target. Before the campaigns we also created a special form for observers who wished to send us their observational data. This form was available under the link: http://fchroma.astro.uni.wroc.pl/index.php/data-submission.html. Data came from more than 20 observers from many countries (e.g. Spain, Italy, Germany, Poland, UK, USA, The Netherlands, Czech Republic, Slovakia). The quality of the data was varied: from single JPEG images to the time series of FITs or TIFF sequences. Unfortunately, in most cases the amateur observers did not follow our request to calibrate the data using flat fields and dark current. In the F-HUNTERS#1 campaign, much of the data is low quality and very often the images don't contains the flare images but only active regions. However, the quality of the amateurs’ data was much higher for F HUNTERS#2 campaign. Examples of this are shown in Figure 12. These images are of high quality and may be suitable for scientific analysis.
We were very pleased with the enthusiasm and engagement of the amateurs, and of course also to the professional observatories who dedicated valuable observing time to the campaign. The campaign was promoted to fellow professional astronomers at two scientific meetings, the 37th Annual meeting of the Polish Astronomical Society, Posnan, Poland, 2015 (Talk & Poster, Berlicki) and the 12th Thinkshop Potsdam, Germany, 2015, (Poster, Berlicki). We have written an article about the campaigns and the lessons learned for the journal Astronomy & Geophysics (Fletcher et al. 2016)
List of Websites:
www.fchroma.org
The F-CHROMA project was a dedicated multi-mode, multi-wavelength study of solar flares as observed in the lower solar atmosphere, or chromosphere. F-CHROMA scientists led nu-merous successful campaigns to observed flares from ground-based solar telescopes, co-ordinated with space-based facilities. This campaign data, and other data, consisting of flare images and spectroscopy, were analysed in combination with state-of-the-art numerical simulations to deduce the structure and evolution of the flare chromosphere, and understand the mechanisms by which energy is transported through the solar atmosphere in a flare, and dissipated in the form of heat, ionisation, and radiation.
Studies of the flare chromosphere enjoyed a renaissance with F-CHROMA, which succeed-ed in placing the importance of observing and understanding the lower-atmosphere energy input in flares front-and-centre for space missions such as IRIS, and the forthcoming major ground-based DKIST telescope. Our experiences in planning and executing flare observa-tions have revealed optimal ways to obtain data on the all-important flare energy-release phase. Discoveries and advances from F-CHROMA included measurements of hot, fast up-flows of hundreds of km/s from ‘elementary excitations’ in the chromosphere, each with an identical velocity profile. These are accompanied by cooler downflows of some tens of km/s, which appear to have an optically-thin component, leading to the exciting prospect of directly mapping the entire flow field in the chromosphere. F-CHROMA studies have demonstrated the critical role of hydrogen and helium ionisation in the temperature and dynamic structure of the chromosphere, and its central role in generating the broad-band UV, IR and ‘white light’ continuum emission by which flares were first identified. We showed areas of agree-ment and – perhaps more interestingly – disagreement with the 45-year-old ‘collisional thick target’ model of electron beam energy transport and deposition, and also made the first radi-ation hydrodynamic simulations of a flare atmosphere excited by a flux of Alfvén waves, showing that this is a viable alternative for generating flare heating and radiation. We have also embarked on the ambitious tasks of spectropolarimetry of the flare chromosphere, and on 2- and 3-D flare radiation transfer modelling; though at early stages they usefully point the way for future studies. And for the first time ever we have successfully cross-validated two independent flare radiation hydrodynamics codes, giving us confidence that – within the ap-proximations of both codes – they are capturing the evolution of a flare atmosphere correctly.
The F-CHROMA project produced an archive of the output of these flare models, which are tricky and time-consuming to run, along with the software necessary to interrogate and inter-pret the model output. An archive of high-quality ground-based flare observations has also been published, along with a catalogue of most ground-based flare observations in Cycles 23 and 24, and their space-based counterparts. We involved amateur astronomers in our flare observations, in two “F-HUNTERS” pro-am observing campaigns, with considerable success, and we disseminated the results of our project through more than 50 refereed papers, more than 100 presentations, and a training school for the next generation of solar flare physicists.
At the end of F-CHROMA, we have a vision of the flare atmosphere, and of flare energy transport, that emphasises excitation of the deep chromosphere, the central role of ionisa-tion, and the great power of optically-thick radiation, such as will be observed in great detail by DKIST, to test different models of flare energy transport and solve long-standing puzzles in flare physics.
Project Context and Objectives:
The F-CHROMA project was a dedicated multi-mode, multi-wavelength study of solar flares as observed in the lower solar atmosphere, or chromosphere. F-CHROMA scientists led numerous successful campaigns to observed flares from ground-based solar telescopes, co-ordinated with space-based facilities. This campaign data, and other data, consisting of flare images and spectroscopy, were analysed in combination with state-of-the-art numerical simulations to deduce the structure and evolution of the flare chromosphere, and understand the mechanisms by which energy is transported through the solar atmosphere in a flare, and dissipated in the form of heat, ionisation, and radiation.
Studies of the flare chromosphere enjoyed a renaissance with F-CHROMA, which succeeded in placing the importance of observing and understanding the lower-atmosphere energy input in flares front-and-centre for space missions such as IRIS, and the forthcoming major ground-based DKIST telescope. Our experiences in planning and executing flare observations have revealed optimal ways to obtain data on the all-important flare energy-release phase. Discoveries and advances from F-CHROMA included measurements of hot, fast upflows of hundreds of km/s from ‘elementary excitations’ in the chromosphere, each with an identical velocity profile. These are accompanied by cooler downflows of some tens of km/s, which appear to have an optically-thin component, leading to the exciting prospect of directly mapping the entire flow field in the chromosphere. F-CHROMA studies have demonstrated the critical role of hydrogen and helium ionisation in the temperature and dynamic structure of the chromosphere, and its central role in generating the broad-band UV, IR and ‘white light’ continuum emission by which flares were first identified. We showed areas of agreement and – perhaps more interestingly – disagreement with the 45-year-old ‘collisional thick target’ model of electron beam energy transport and deposition, and also made the first radiation hydrodynamic simulations of a flare atmosphere excited by a flux of Alfvén waves, showing that this is a viable alternative for generating flare heating and radiation. We have also embarked on the ambitious tasks of spectropolarimetry of the flare chromosphere, and on 2- and 3-D flare radiation transfer modelling; though at early stages they usefully point the way for future studies. And for the first time ever we have successfully cross-validated two independent flare radiation hydrodynamics codes, giving us confidence that – within the approximations of both codes – they are capturing the evolution of a flare atmosphere correctly.
The F-CHROMA project produced an archive of the output of these flare models, which are tricky and time-consuming to run, along with the software necessary to interrogate and interpret the model output. An archive of high-quality ground-based flare observations has also been published, along with a catalogue of most ground-based flare observations in Cycles 23 and 24, and their space-based counterparts. We involved amateur astronomers in our flare observations, in two “F-HUNTERS” pro-am observing campaigns, with considerable success, and we disseminated the results of our project through more than 50 refereed papers, more than 100 presentations, and a training school for the next generation of solar flare physicists.
At the end of F-CHROMA, we have a vision of the flare atmosphere, and of flare energy transport, that emphasises excitation of the deep chromosphere, the central role of ionisation, and the great power of optically-thick radiation, such as will be observed in great detail by DKIST, to test different models of flare energy transport and solve long-standing puzzles in flare physics.
2. Project context and objectives
Our nearest star, the Sun, controls the solar system first and foremost by its gravitational presence and by the steady radiation output on which our existence on this planet relies completely. However, it also controls its environment in less obvious ways. The phenomenon of ‘space weather’ is increasingly in the public consciousness. Space weather encompasses many effects, but is primarily due to the influence of the Sun on the electromagnetic environment of the Earth, planets, and heliosphere. Sources of space weather are both solar flares, which are dramatic changes in the solar radiation at high energies in particular, and coronal mass ejections (CMEs) which are expulsions of magnetised, electrified gas (plasma) from the atmosphere of the Sun. The results of space weather at Earth are, primarily, disturbances in our magnetic field and heating of our upper atmosphere. Some decades ago this was of little relevance to human existence, but now our lives are highly dependent on, for example, spacecraft orbiting in that heated upper atmosphere and electrical power distribution systems that are sensitive to those magnetic disturbances. Therefore understanding the physical causes and effects of space weather, and working towards predicting its occurrence, is of central importance to our technological society.
Solar flares are dramatic and short-lived flashes of radiation in the solar atmosphere. The total energy released in the form of flare radiation can be on the order of 1025Joules, the energy equivalent to 2.5 billion megatons of TNT. This is provided by restructuring of stressed magnetic field in the outer atmosphere of the Sun (the corona) to a lower energy state. Solar flares are very often closely associated with CMEs, and the most intense radiation – which occurs in the optical and UV parts of the spectrum – is associated with, and likely caused by, substantial numbers of highly accelerated non-thermal particles. The centrally important question of flare physics is to understand how the energy stored in stressed coronal magnetic field is released so rapidly, transported through the atmosphere of the Sun, and converted – with very high efficiency – into the kinetic energy of the non-thermal particles and thus, or otherwise, into the flare’s radiation output. This is a question that must be answered by putting together a broad range of observational and theoretical information. In the F-CHROMA project we focused on understanding how the transport and conversion of the flare energy can be deduced from observed flare radiation.
Solar flares were first observed in the optical in 1859 as rapid brightenings in white light from the solar surface, or photosphere. The development in the late 19th century of the spectroheliograph, which allows images to be made in isolated regions of the spectrum, revealed major disturbances in spectral lines now recognised to come from the narrow, complex interface between the photosphere and corona, the chromosphere. In particular, the 3-2 transition of hydrogen, Hα, characteristic of a plasma at ~10,000 K was extremely bright, and indeed analysis of this radiation is one of the main ways that we can understand what is happening in a flare. At that time flares were assumed to be entirely a chromospheric phenomenon, based solely on the energy implied by the enormous optical increases. However, with the associated sudden ionospheric disturbances detected by ground-based magnetometers, it became clear that flares involve substantial increases in ionising radiation at UV to X-ray wavelengths, verified by the dramatic, dynamical evolution of the coronal portion of the flare as seen from space in extreme UV and soft X-ray. Substantial advances in understanding of the magnetic and thermal evolution of flare plasma have been made by focusing on these energy ranges and on the coronal behaviour.
Still, the solar chromosphere remains of primary interest during flares. It is here that the majority of the flare’s radiative output originates, making it the main source of flare diagnostic information. Using chromospheric radiation we can deduce flare heating and ionisation, and subsequent cooling and de-excitation of different layers of the atmosphere. We can also measure plasma flows and non-thermal velocities, and infer character of accelerated particle distributions. The energy deposition and its variation in space and time can in principle be understood from the chromosphere and its observed evolution. However the chromosphere is a very complicated environment, so doing this successfully requires forward modeling – using first principles physical simulations to predict the radiation signatures arising from energy input in different forms. Novel developments in radiation hydrodynamical modeling address this problem, calculating the evolution of the chromospheric environment, and from this predicting the radiation signatures that allow a direct link from observables to physical conditions. Besides providing crucial flare diagnostics, the chromosphere may even be intimately related to the acceleration of the non-thermal electrons and ions that are a defining flare characteristic.
Our progress in this area has also been accelerated greatly by the acquisition of excellent datasets from both space-based and ground-based observatories. These facilities are observing at sub-arcsecond scales, corresponding to physical sizes on the Sun of 500km or less. This level of detail is probably still insufficient to capture the elementary processes happening on the Sun, but we do begin to have hints that important structures in space and time are not too far below our current capabilities. The flaring chromosphere emits across the entire electromagnetic spectrum, with different wavelengths characteristic of different temperature and locations (or heights, in a one-dimensional model) so that combining radiation signatures across wavelengths allows us, painstakingly, to build up a complete picture of its structure. No less important, but more difficult to interpret, is the information about the magnetic field in a solar flare, at the level of the photosphere and chromosphere, which can be deduced from the emitted radiation. We are just at the beginning of measuring and interpreting this emission,
The Sun is not the only star that produces flares. Indeed it seems to be a rather inactive star compared to many others of its own and other spectral classes (colour). However, it is the only star where we can directly image the flare and hope to understand how it works. By doing so, we may also extrapolate and apply our knowledge to other stars. Stellar flares have become a subject of great interest in the last few years as it becomes clear first of all how many stars have planetary systems around them, and secondly what the influences of the short wavelength radiation emitted during flares are both for the creation and destruction of conditions favourable to the production of pre-biotic molecules.
Following a long gap during which studies of the flaring corona were predominant, the central importance of the chromosphere in the energetics and dynamics of a flare (and the Sun as a whole) is being reasserted. Our project has capitalized on the renewed interest and investment worldwide in this fascinating region, leading to a multi-mode, multi-wavelength study of the solar flare chromosphere, supported by dedicated theoretical modeling. This concerted international effort in joint science exploitation of space-based and ground-based data, theory and numerical simulation has led to significant advances in our understanding of the physics of solar flares, and a roadmap for further developing this in the future.
Project Results:
(Please note - Figure numbers found in this report refer to the pdf version uploaded separately)
1 Main Scientific Results
1.1 Overview
Solar flares are sudden and dramatic outbursts of intense radiation from the solar atmosphere, caused by the release of enormous amounts of energy from a stressed magnetic field. Our project’s top-level scientific goals were to use observations and theory to understand how energy moves through the solar atmosphere following its release, to predict the spectrum of radiation that is emitted during a flare, and to identify how the next generation of solar observatories can best be used to further explore the basic physics of solar flares and, by extension, flares on other stars. We also aimed to promote the use of the best available models and observations in the scientific community, to grow the number of scientists, especially early career scientists, who are expert in the analysis of flares, and to involve amateur astronomers in the observation of these exciting and – so far – unpredictable events. We have made significant progress in all of these areas.
We concentrated on the chromosphere as this is the layer in which the majority of the solar flare radiation originates. The chromospheric spectrum can be probed in detail to reveal the changes, in time and space, of properties such as temperature, density and ionisation state that result from flare excitation. Thereby, the physical processes that lead to these changes can be picked apart. The non-flaring solar chromosphere is already a complex environment, and the flaring chromosphere presents additional complications in that it is (i) difficult to ob-serve due to the rarity and unpredictability of flare events, and (ii) in a highly non-equilibrium condition because of the intense and transient input of energy, which makes the physics challenging to model. Therefore we assembled a team with experience in planning and exe-cuting flare observations, in analysing all kinds of solar observations from the infrared to the X-rays, and in designing and running state-of-the-art numerical simulations, in an integrated approach. A focussed effort on the physics of the flaring solar chromosphere had never been undertaken before at this level. F-CHROMA has also prepared European and international flare physicists for the exploitation of major upcoming facilities such as ESA’s Solar Orbiter Mission with a launch foreseen in 2018, the Daniel K. Inouye Solar Telescope (DKIST) which sees first light in 2019, and now the European Solar Telescope (EST) which was adopted onto the European Research Infrastructures roadmap in 2016, with planned start of operations in 2026.
Based on research that has taken place in F-CHROMA, and before, we can formulate a top-level view of the evolution of the chromosphere during a flare.
We have established that energy can be delivered to the flare chromosphere either by electron beams or by Alfvén waves. Alfvén waves are capable of penetrating deeper into the chromosphere and exciting the temperature minimum region (Russell & Fletcher 2013), while electron beams of some 10s of keV energy deposit their energy higher up. Energy is also transported by thermal conduction from the upper chromosphere, which is heated to around 10MK (Graham et al. & Fletcher 2013). The main sink of energy in the flare chromosphere is the latent heat of ionisation of both hydrogen and helium, processes that act as temporary ‘thermostats’. Hydrogen ionisation is never complete in the lower atmosphere, since the high density leads to rapid recombination producing Balmer and Paschen radiation, and continu-ous re-ionisation occurs because of ongoing heating. However, in the upper chromosphere recombination is not so rapid, and when He ionisation is complete there are no other effec-tive radiation loss channels in this optically-thin medium, so the plasma starts to expand causing both upflows and (slower) downflows. Above the He ionisation layer, optically-thin line radiation of incompletely ionised metals, coupled with conduction, is also insufficient to rid the plasma of its energy locally, resulting in rapid upwards expansion. All of the hot plasma is expanding: there is no ‘stationary component’ at 10MK
In the moving plasma, optically thin lines are formed (including, surprisingly, redshifted opti-cally thin components of cool lines at ~8,000K) and provide a tool to reliably derive the amplitude and direction of the flows. Often, especially in slightly later phases of the flare, there is also sufficient absorbing material above the layer at which the main chromospheric line core forms to produce either red or blue wing absorption, according to whether this material is up-flowing or downflowing. The flare continuum is primarily due to hydrogen recombination emission. This will also lead to some backwarming of the photosphere, but models show this is unlikely to be significant. In more detail, some of the most significant advances and discoveries we have made are as follows:
• Discovered that the heated (10MK) chromospheric plasma is entirely upflowing, and the flows evolve in an identical way across numerous sources, apparently irrespective of energy input rate (Graham & Cauzzi 2015);
• Made new observations of the flare continuum, and guided by modeling, identified these as predominantly due to recombination emission. This emphasises the importance of ionisation in a flare (Heinzel & Kleint 2014; Simões et al. 2017 in prep.);
• Demonstrated that optically thick spectral line profiles that appear to show redshifted emission may in fact result from blueshifted absorption, indicating that naïve Doppler interpretations of lineshifts can be misleading (Kuridze et al. 2015, 2016; Kerr et al 2015).
• Made the first ever simultaneous observations of a flare in Hα and Hβ (Capparelli et al. 2017 in prep)
• Cross-validated, for the first time, two completely independent flare radiation hydro-dynamics codes, RADYN and Flarix (Kasparova et al. 2017 in prep), and promoted the use of the former as a mainstream analysis tool.
• Demonstrated the viability of Alfvén wave energy transport model as an alternative to the long-standing electron beam model for producing chromospheric flare radiation (Kerr et al. 2016)
• Found a flare where the absence of hydrogen Balmer lines may rule out any significant role for electrons in impulsive phase energy transport (Procházka et al. 2017)
In addition to these scientific advances, our success overall can be assessed by revisiting the stated scientific objectives of the F-CHROMA proposal:
(i) Perform joint observations and analysis of space-based and ground-based observations
of flares, delivering improved diagnostics for conditions in the flare chromosphere;
A dedicated and pro-active campaign of proposing for every ground-based opportunity resulted in 31 new ground-based flare datasets, most with accompanying space-based data. Analysis has commenced on many of these datasets, with 14 papers published and in preparation. The observing plans were constructed by F-CHROMA members with the goal of capturing, with both imaging and spectroscopy, the impulsive energy release phase of the flare.
(ii) Carry out detailed modeling of these diagnostics for a range of flare energy inputs
Our observations have stimulated targeted modelling efforts of both line and continuum radiation in a beam-heated atmosphere. Examples of this are Mg II h & k lines, the Ca II 8542 Å line, hydrogen and helium lines and continua, and the infrared free-free continuum. The range of flare energy inputs is reflected both in modelling electron beam energy input with a broad range of particle beam parameters, and exploring different energy in-put mechanisms, i.e. particle beams versus Alfvén waves. To facilitate comparisons, we have also worked on improved atomic data for Mg II and other ions.
(iii) Utilise forward modeling techniques to determine the physical properties of the solar
chromosphere during solar flares;
We have addressed this in two ways. We have generated a suite of numerical models covering a range of beam input parameters to investigate the range of physical properties that occur. We have used these models on their own to understand the likely conditions pertaining in the chromosphere. An example of this is our work on the locations at which the optical and infrared continua are formed in flares (Simões et al. 2017 in prep). We have also made comparisons between model predictions and observations, which have led to as much understanding about why the two fail to agree, as where they do agree. For example, we see that to reproduce the observed Mg II line profiles, with a non-reversed core, requires a higher-density upper chromosphere than is predicted by electron beam models. But the models have additionally shed light on unexpected ob-served properties, such as the two-component line profiles in some cool chromospheric lines, which are interpreted as due to an optically thick, stationary source, and an overlying optically thin downward moving chromospheric condensation.
(iv) Use models and observations to optimize targeted flare observing sequences for present
and future missions and observatories, hence maximizing their scientific return;
Based on our experiences with F-CHROMA observing campaigns, and on our existing analyses and modeling we can make some strong recommendations for different future observations targeted at understanding flare energy transport: (i) we have demonstrated the importance of ‘sit-and-stare’ with a single active region target, rather than target-hopping. This optimises the chances of capturing observations jointly with space-missions, and provides a continuous record before, during and after an event, allowing the rapid evolution of the chromosphere to be tracked; (ii) it is clear that, in general, tem-poral evolution is key, as the radiation from the transient initial phase of flares encodes the most information about the evolution of the chromosphere; (iii) that the spectropolar-imetric instruments on DKIST should have frequent use as flare spectrometers only, sac-rificing the polarimetry for rapid spectroscopic line scans to capture the line profile evolution – on timescales of a second – at more than one spectral line: (iv) that to progress on understanding flare chromospheric dynamics, linking the rapid upflows seen in hot lines from the Interface Region Imaging Spectrometer (IRIS) and cool lines from both IRIS and the ground, we need to focus on lines formed at temperatures of a few x 105 K to 106 K, which will be present in the long passband for the SPICE spectrometer on Solar Orbiter.
(v) produce a catalogue of existing and future ground-based solar flare observations, linked
to available space-based counterparts;
The catalogue of ground-based solar flare observations was published and announced within the first year of the F-CHROMA project on 01 December 2014, and continues to be updated. The 534 records contain information on coverage, contacts who will provide data, and any known publications on the ground-based data.
(vi) populate an archive of ground-based flare observations and make it available to the community;
Many of the new ground-based datasets obtained during F-CHROMA campaigns have been placed in our public archive, as a resource for the community. The link is availa-ble from the F-CHROMA catalogue web page, and data in FITS format can be down-loaded. The public archive at present contains 33 flare records, 22 of which arise from F-CHROMA campaigns. This reflects the rarity of good ground-based flare observations (even with our concerted efforts to improve this!) and it is to our knowledge the first attempt to make ground-based flare data systematically available for analysis. We hope that this resource will encourage and facilitate familiarisation with the next generation of ground-based solar data, for example by providing datasets for students to work on.
(vii) produce a flare spectral model database and make it available to the community.
Following testing of the RADYN and RH codes as part of F-CHROMA, and the im-provement of some aspects of the RADYN computation, we have populated a data-base of beam-driven radiation hydrodynamics simulations, representing a range of total energy, electron spectral index and low-energy cutoff. The model database is linked from our F-CHROMA website. Instead of producing the flare spectral model database, we have made available the full simulation results, and the code, and guide, needed to interrogate those data to produce spectra, as it is the most flexible approach to making the output available to users.
It is impossible to cover all of the activities and results of F-CHROMA, and this report is therefore selective, focusing on the areas where perhaps the most surprising or substantial advances have been made. The numerous outputs arising from the project throughout its duration and beyond can be seen in the 55 refereed publications and 100 presentations linked at http://www.fchroma.org/?page_id=33 and http://www.fchroma.org/?page_id=219.
1.2 New observations
High quality solar flare observations from the ground are rare. Therefore we had a highly pro-active programme of proposing for every ground-based observing opportunity, including with the newest facilities (GREGOR). This resulted in 31 new ground-based flare data sets many of which can be found at https://star.pst.qub.ac.uk/webdav/public/fchroma/ Most have ac-companying space-based data. Analysis has commenced on many of these datasets, with 14 papers in preparation. The observing plans constructed by F-CHROMA team members had the goal of capturing, with both imaging and spectroscopy, the impulsive energy release phase of the flare.
1.2.1 The F-CHROMA flare campaigns
In total there were 23 F-CHROMA organised observing periods throughout 2014-2016. The majority of these were allocated in a competitive process on major facilities such as the Dunn Solar Telescope and the Swedish Solar Telescope. These were run both in service mode, and in observer mode. The remainder of the observing periods were carried out at facilities controlled by F-CHROMA team members, namely the Multichannel Flare Spectrograph at the Ondrejov Observatory (ASU), and the Multichannel Subtractive Double Pass Spectrograph at the Large Coronagraph and Horizontal Telescope at Bialkow (UWR). In total 31 flares were observed from these facilities, in whole or in part. A list of campaigns and flares captured can be found in Appendix 1. Particularly encouraging was the enthusiastic support from space-based observatories for our efforts, with a large number of events observed also by the Interface Region Imaging Spectrometer (IRIS) spacecraft, which focuses on the chromosphere. Several of the observing periods covered the two F-CHROMA F-HUNTERS pro-am observing campaigns.
Around 20 new high-quality flare datasets from the campaigns have been formatted and placed in the F-CHROMA data archive, which is currently at a volume of over 400GB of flare data. However, this is an on-going process, which will continue over the coming year. This resource for the community will, we hope, encourage and facilitate familiarisation ground-based solar data in preparation for DKIST and the European Solar Telescope.
1.2.2 New types of flare observations
The campaigns resulted in new types of flare observations, using facilities which had not previously been used for flares, such as the Facility InfraRed Spectrometer (FIRS) infrared imaging spectropolarimeter at the Dunn Solar Telescope, and the Crisp Imaging SpectroPolarimeter (CRISP) optical imaging spectropolarimeter at the Swedish Solar Telescope. A new instrument was also developed at Ondrejov, to capture the (spatially-integrated) flare optical spectrum in the range 350-440nm (Figure 3). This includes the Balmer jump at 364.6nm cor-responding to the head of the free-bound hydrogen recombination continuum onto level 2. The Image Selector allows a spectrum to be taken over a field-of-view corresponding to an active region, thus increasing signal-to-background compared with ‘Sun-as-a-Star’ instruments, and improving the chances of observing a flare compared to slit spectrometers. The instrument has sub-second temporal resolution, and was used with success (Kotrč et al 2016) to capture the X 1.0 flare SOL2014-06-11T09:06.
It is very surprising to see in these observations that in some flares neither the Balmer jump nor the low-order Balmer are visible (nor are the Lyman lines, observable by EVE). This is problematic to explain with the standard electron-beam heated model (Procházka et al. 2017). The instrument is now being upgraded to detect the hydrogen Paschen continuum.
Existing instruments were also used in original ways to optimise flare science, and – as is usual when novel setups are tried – to make discoveries. For example, on the Dunn Solar Telescope the ROSA high-speed imager ran in Hβ, and the IBIS imaging spectrometer in Hα, obtaining to the best of our knowledge the only joint observations in these two important flare lines (see Section 3.4.2.3). They reveal the exceptionally high degree of temporal correlation between these lines, apart from during the critical first few seconds of the event. The IRIS spectrometer was run as fast as possible in ‘sit-and-stare’ mode, obtaining the highest temporal resolution of flare evaporation and condensation flows yet made, in individual flare pixels (Graham & Cauzzi 2015). Dedicated flare spectropolarimetric observations were also ob-tained in the early decay phase of the X1.6 event SOL2014-10-22T14:02 using IBIS and these are still being analysed (Guglielmino et al. 2017 in prep).
1.2.3 Lessons learned and recommendations for future observations
In planning and carrying out these observations we learned what is most important for ad-vancing our understanding of the physics of the chromosphere during the impulsive energy release phase of a flare. This does not address all aspects of impulsive-phase flare physics of course – one can easily imagine scenarios where the large-scale spatial evolution is needed to address the physical question of interest. But focusing on the energy transport and dissipation we know now that the following are vital:
(i) pre-flare coverage: this is necessary to characterise the state of the chromosphere before energy input begins, as well as to capture the very earliest stages of the impulsive phase. ‘Flare triggers’, to switch telescope mode or pointing, are unlikely to be useful in this respect.
(ii) high temporal resolution: Almost everything else should be sacrificed to obtain the highest-possible temporal resolution, because the chromosphere responds on timescales of se-conds to energy input, and observationally many of the strong spectral lines seem to reach a ‘saturated’ state after a few seconds. Theoretically, this may be because the level popula-tions approach their local thermodynamic equilibrium values because of high collisionality. At this point, discriminating information about the energy input is harder to obtain.
(iii) spectroscopy: resolving the details of the spectral line profiles is important for under-standing plasma flows at different layers in the chromosphere. In the optical, good sampling (~0.1Å) for around ± 1Å in the line core is required, as the core moves during flares. In the UV and EUV, spectral windows around strong, hot lines such as Fe XXI should be extended into the blue to capture the earliest phases of the flow dynamics, where blueshifts corresponding to more than 270km/s are detected.
The above constraints point us inevitably in the direction of “sit-and-stare” type observations, where a target is chosen and tracked for a long period. The spectrometer slit (if applicable) should be placed as close as possible to perpendicular to the polarity inversion line, and rasters, if any, should be small to optimise the time cadence. It is recognised that it may be diffi-cult to convince the operators of large and competitive facilities such as DKIST and Solar Orbiter to run what appears to be restrictive programmes of this type, with the risk of accu-mulating a large amount of data that is not useful for flare studies. But we are convinced that this strategy offer the best scientific return for impulsive phase studies
1.2.4 Analysis arising from the F-CHROMA flare observations
In Table 1 in the uploaded pdf we give a list of papers published or in prep that use data gathered during F-CHROMA organised campaigns, or other campaigns involving F-CHROMA scientists. As can be seen, these are primarily from the 2014 observing campaigns so far. The successful Sep-tember/October 2015 campaigns, and a couple of flares from 2016, will also be analysed in the future.
1.3 New flare data analysis
There are numerous new multi-wavelength data analysis results, contained in (to date) 35 published or accepted journal articles, with 12 more manuscripts in a late stage of prepara-tion. We focus here on a few important to have emerged from this analysis.
1.3.1 Hydrogen continuum emission from IRIS and from the ground
The hydrogen Balmer continuum is the spectral range between the Balmer jump at 3646Å, and (nominally) the Lyman jump at 912Å. The true Balmer continuum is the recombination emission onto level 2 from hydrogen atoms transiently ionised during a flare. In this broad range one will also find blackbody emission and other continua such as the optically-thick silicon and carbon continua. However, in the near UV channel of IRIS these metal continua are weak, and, by sampling the spectrum between the strong emission lines Heinzel & Kleint (2014) were able to measure the Balmer continuum in the well-observed flare SOL2014-03-29T17:48 and demonstrate that it is consistent with the expected recombination intensity. Kleint et al. (2016) extended this analysis to the IRIS far UV channel, the optical channel on the Helioseismic and Magnetic Imager (HMI) and the FIRS infrared spectrometer on the Dunn Solar Telescope (DST), extending the range to include the Paschen continuum (re-combination onto level 3) showing that the spectra thus generated are not consistent with an enhanced photospheric blackbody alone (Figure 4). That the flare continuum emission in-volves recombination in this large event now seems conclusive, and points to the central role of hydrogen ionisation in the flare energetics. There is, nonetheless, a requirement for an enhanced (by a few 100K) photospheric blackbody to explain the radiation intensity, and this would require a very intense heating of the lower atmosphere.
It has long been known that the Balmer recombination edge is often ‘missing’ in solar flare observations. This was reported by Kowalski et al (2015) and may be interpreted as the edge being masked by a ‘piling up’ of broadened, high-order Balmer lines into a bump. A very unusual observation of an X1.0 flare was made with the novel flare broadband spectrometer described in Section 1.2.2 in which the continuum enhancement was strong, with the ‘pile up’ feature present, but with little to no detectable low order Balmer lines (Procházka et al. 2017). These should be at 434.1nm 410.2 nm, 397.0nm 388.9nm and 383.5nm in Figure 5. This can only be explained by heating of the deep atmosphere without the overlying atmosphere being excited sufficiently to generate hydrogen line enhancement. This is not possible with ‘normal’ electron beams, i.e. with low-energy cutoffs in the 10s of keV range. It suggests that we should be looking for explanations in terms of proton beams or Alfven wave heating which may be able to heat the deep atmosphere, leaving the upper chromosphere relatively unaffected.
1.3.2 Hot plasma upflows
The response of the flare atmosphere to intense heating is to expand upwards, with a momentum-balancing downwards counterpart. One of the most interesting F-CHROMA results was on this phenomenon from Graham & Cauzzi (2015). They analysed IRIS spectra including the hot line of Fe XXI at 1354Å and found very high blueshifts, and very strong (non-thermal) broadening. The measurements were made at 86 individual flare ribbon pixels which were excited in sequence and observed as the flare ribbon moved down the spectrometer slit.
The velocity profile as a function of time in each pixel was obtained (Figure 6) and perhaps the most surprising aspect of the observation is that if the velocity profiles as a function of time are stacked in time, with the base being the time corresponding to the first point at which the enhanced Fe XXI signature becomes detectable, the majority of the velocity pro-files overlay one another very closely. This suggests two main things: that the evolution of the evaporative flow is rather insensitive to the details of the atmosphere and to the strength of the energy input – within factors of a few at least. This might be consistent with a picture where the Fe XXI signature arises from heating only of the rather tenous upper part of the pre-flare chromosphere, and expands into a much lower density, “undisturbed” overlying co-rona. Upflows of over 100km/s are present for around 5-10 minutes though the input of energy is much shorter-lived. Numerical simulations of beam energy input tend to show flow velocities reducing to below 50km/s within a 60 seconds of the beam switching off (Kerr 2017), so the observed behaviour may imply continuing low-level energy input.
1.3.3 Line shapes and asymmetries in cool lines
Line shapes and asymmetries in lines formed at low temperatures have turned out to be extremely interesting for understanding the dynamics of the low atmosphere. Different wave-lengths in the spectral lines are formed in different regions, with the line core generally formed high up in the chromosphere (being more optically thick, the escaping photons come from higher levels), and the wings from low down. Work on Hα, (Kuridze et al. 2015), Na I (Kuridze et al. 2016) and Mg II h & k (Kerr et al. 2015, Liu et al. 2015) show a variety of line profiles, which evolve during the first few seconds of a flare from being (briefly) centrally re-versed to being non-reversed, and then back to centrally reversed again. The variety and evolution of these line profiles have prompted a concerted modelling effort, described in Section 1.4. A non-reversed emission line profile in these lines may indicate that the source function in the line core is strongly coupled to the Planck function due to a high density at the line formation height. The reversed profiles indicate decoupling of Planck and source functions, and also can have emission and absorption contributions from moving components, both optically thin and optically thick, overlying the optically thick part of the spectral line. Observed line shifts and asymmetries have tended to be interpreted in terms of flows at different atmospheric levels in a rather simplistic way, e.g. the ‘bisector’ method, in which the flow speed is inferred from the location of the line bisector at e.g. the 50% level in a flare line. However, the wavelength-dependent opacity effects, made obvious by RADYN modelling, make this approach potentially very misleading.
The high cadence of IRIS observations has been particularly useful in revealing the velocity evolution of the lower chromosphere. Corresponding to the high-speed upflows described in Section 1.3.2 the Mg II lines were used to identify downflows associated with chromospheric condensation; applying a bisector method, average downflows as high as 30km/s were ob-served, decelerating to zero in around 60s (Graham & Cauzzi 2015). The onset of these downflows appears to precede the onset of the upflows by around 60s, though this may be an issue of detectability of the upflow line profiles (e.g. starting very weak and blueshifted outside the IRIS passband). Of course the bisector method might be a too simplistic approach; examining the line profiles in greater detail it appears that the profiles in fact have two components, one almost stationary and narrow, and the other substantially redshifted and broadened. These start off as two quite well distinguished components, with the red-shifted component slowing and merging into the stationary component as time proceeds (Kowalski et al. 2017, Graham et al. 2017 in prep). The interpretation of this is discussed in Section 1.4.2 but it seems to give a potential direct window onto the flow field of the upper chromosphere.
1.4 New modelling
1.4.1 Comparison of RADYN and Flarix simulations
One of the most important activities and major activities carried out by the F-CHROMA team was a comparison of two different flare radiative hydrodynamics codes (Kasparova et al. 2017 in prep). It is important to recognise that such cross-validations are rarely done in any kind of computations in solar physics; perhaps the authors of codes prefer not to confront the codes’ possible shortcomings. An attempt to cross compare flare hydrodynamics codes was carried out in the 1980s during the SMM days (Hudson, private communication) but was not carried through. We believe that our comparison of the RADYN code (Carlsson & Stein 1992, Abbett & Hawley 1999), originating primarily from the UIO and the Flarix code (Heinzel et al, 2016) originating from the ASU, is the first of its kind. The two codes have completely different numerical schemes, and different approaches to solving the hydrodynamics of the plasma and the non-LTE level populations, though they do have a common approach to radiative transfer. The two codes were initiated with the same model atmosphere, and driven by the same input beam. Happily, the output model atmospheres track one another very closely (Figure 7), thus giving us confidence in using both codes. In practice the RADYN code is currently better documented, and has an output format which is well integrated into the RH package for computing spectral lines in detail, so we have used this in our model archive.
1.4.2 Interpretation of model output and discoveries arising from this
The careful comparison of observations and model outputs has been one of the most valuable exercises during the F-CHROMA project, not only for what the models can explain or help us understand, but also for what they currently fail to predict. Below are described a few common themes that have emerged from our analyses.
1.4.2.1 Understanding the line profiles in terms of flows
The shifts and asymmetries observed in the line spectrum of solar flares in Hα and Ca II 8542 Å by Kuridze at al. (2015) and in Mg II h & k by Kerr et al. (2015, 2017), Liu et al (2015) have been compared with synthesised line profiles produced with RADYN and RH. To obtain reliable simulations, the flare parameters used in the model were estimated from the HXR data. In the Kuridze et al. study, the temporal evolution of the observed and simulated Hα line profiles show excess emission in the red wing (red asymmetry) before flare maximum and excess in the blue wing (blue asymmetry) after maximum. The multicomponent velocity field in the chromosphere creates differences in the opacity between the red and blue wings of Hα, which produces the asymmetric chromospheric emissions. At early phases of the flare, material upflows absorb emission in the blue wing, creating the red asymmetry, while at later phases, emission in the red wing is absorbed by plasma downflows, reversing the asymmetry to blue, confirming an idea proposed by Heinzel et al. (1994). This indicates that Hα line shifts as determined by bisectors need to be carefully analysed and contrasted with simulations to properly derive flow speed and directions
The work by Graham et al (2017, in prep) and Kowalski et al (2017) reveal another interest-ing feature; low-temperature spectral lines such as Fe II 2815.45Å, Si II 1348.55Å, CI 1354.28Å, C II 1335Å as well as the Mg II subordinate line at 2791.6Å are each composed of a narrow, almost static component and a broader redshifted component (See Figure 8.) The interpretation, backed up by modeling, is that the narrow, stationary component is optically thick and originates from the deep atmosphere, whereas the broader redshifted component originates from an optically thin, overlying chromospheric condensation, at temperatures be-tween around 8,000 and 25,000K. This is supported qualitatively by RADYN modeling, though some of the details differ. If the broader redshifted component is indeed optically thin, it offers a completely new possibility to construct the flow and turbulence field at line formation temperatures in the few to 10 kK range.
1.4.2.2 Beam driven versus wave driven simulations and comparison with Mg II lines
In the Kerr et al (2015, 2016) studies of Mg II h & k the line profiles throughout most of the flare have a small red asymmetry, and no clear central reversal. RADYN simulations using a beam energy input always produce a rather symmetric line profile with no asymmetry, whereas running RADYN with an Alfven wave energy input gives a line profile that is more similar to the observations, due to a different profile of energy deposition through the atmosphere (Figure 9).
1.4.2.3 Evolution and saturation of the hydrogen lines in flares
Observationally, one finds an initial transient behaviour in the ‘Lyman decrements’ (ratios of Lyman line intensities) at the beginning of a flare, but thereafter they become rather constant, even as the energy input changes substantially (M. Kennedy, PhD Thesis). This behaviour is seen in several flares, but there is a small flare-to-flare variation in the average ratios. A small grid of RADYN models with quite extreme ranges of beam parameters showed a similar pattern of transient evolution followed by flat ratios (though the ratios were systematically below the observed ratios, possibly because the events are a superposition of short-lived, high ratio values in the initial transient phases of a series of injections, and longer-lasting low ratio values). The reason for the constant ratio values seems to be that the high collisionality, due to high densities and temperatures at the Lyman line formation heights, drives the level populations close to their LTE values. This is also the case for the ratio of Hα and Hβ (Capparelli et al. 2017, in prep, Figure 10). It was hoped that these line ratios would track the evolution of mean electron temperatures around where the line cores form, but it seems that this will only be possible, if at all, during the initial few seconds of a flare.
1.4.2.4 Interpretation of the continuum
FLARIX simulations were also used to synthesize the hydrogen recombination continua. The Balmer continuum was compared to IRIS observations (note that the first detection of the Balmer continuum in IRIS flare spectra was one of important results of F-CHROMA, see first period report). The Paschen continuum was then compared to SDO/HMI observations of both disk and limb flares. It was also shown that at the flare peak, the radiation losses are dominated by hydrogen and namely by the Balmer continuum and thus Balmer-continuum detection in the IRIS spectra represents a critical constraint for the modelling.
Recent high-cadence observations in the infrared at 5.2μm and 8.2μm, made at the McMath-Pierce telescope at Kitt Peak National Observatory showed well-defined, high contrast chromospheric footpoints. This stimulated work on the physical processes leading to the IR emission, RADYN was used to examine in detail the formation of optical and IR continuum, by calculating the contribution function in a number of spectral bands, including for the first time the electron free-free and the H- free-free opacity (Figure 10). The flare continuum enhancement is mostly formed in the chromosphere due to hydrogen recombination: Balmer in NUV, Paschen in optical, Brackett/Pfund in the near-IR (λ<2µm), and the evolution of the ionisation so the ‘total electron content’ of the chromosphere turns out to be critical. For all wavelengths, the contribution function has an excess in the chromosphere during the simulated flare and the emission is optically thin, while the photosphere is mostly undisturbed, so continuum enhancement from H- opacity is unimportant. The same occurs for all simulated cases, for different properties of the electron beam. Only the chromospheric Balmer continuum is strong enough to cause a minor photospheric temperature increase – consistent with the findings of Kleint et al. (2016) but, in the beam models used, this is still not sufficient to produce a significant increase in the continuum emission. These results strongly indicate that NUV, optical (white light) and NIR continua are primarily a result of hydrogen recombination in the chromosphere (Simões et al. 2017 in prep)
1.4.2.5 Where models and data do not match
There are areas where the modeling and the data seem still to be in disagreement, and the-se common features will guide our thinking and motivate developments in future modeling. These common features are: (i) the ‘filling in’ of normally reversed line cores, which requires a higher pressure corona/transition region than is predicted in the RADYN electron-beam-driven simulations; (ii) enhanced far wing and continuum emission requiring heating of deep layers of the atmosphere, again more so than predicted by the RADYN electron-beam simulations; (iii) line broadening, which requires an increased turbulent broadening, probably one that varies through the atmosphere. In addition to more work on beam-driven simulations, models involving Alfvén wave energy transport may offer solutions to some of these discrepancies.
1.4.3 2D and 3D models
Papers have been published on 2D radiation transfer in MHD models with the FLASH code, and with the MALI code. Adding a second dimension to radiative transfer naturally produces fibril-like structures in cool flare loops (Falewicz et al. 2015) and can produce effects in the line profiles – such as the appearance of the second, shifted component that is seen in flare loop observations (Heinzel et al. 2017 in prep.) The PENCIL code was used to model the evolution of a delta spot in 3D (Chatterjee et al. 2016, Figure 11), and very localised, rapidly evolving regions of high temperature were identified, very suggestive of flaring energy re-lease. It is too early to draw general conclusions about the importance of 2D or 3D radiation transfer, but the fact that these highly-developed MHD codes show evidence of flare-like activity is very encouraging for the future development of flare modeling.
1.4.4 Provision of grid of models
The grid of models uses the RADYN code, which has been validated against Flarix. It is now published at https://star.pst.qub.ac.uk/wiki/doku.php/public/solarmodels/start. These models are already being used by PhD students in Glasgow, and attendees at the F-CHROMA train-ing school were also introduced to their use. The RADYN code used incorporates a Fokker-Planck beam solver, a new starting model atmosphere resembling the VAL3C atmosphere, better resolution of gradients, and a more finely-resolved frequency grid across the lines. The injected beam has a triangular, symmetric time profile of 20s duration, to represent an ‘elementary flare burst’. The beam power law index, low energy cutoff, and integrated flux are all varied, resulting in a grid of 96 models. The output is in Common Data Format (CDF) and can be examined to look at the evolution of atmospheric properties (temperature, density, velocity, ionisation and atomic excitation levels) and also the formation and evolution of spectral line profiles and continuum for several strong lines and continua including several energetically important lines and continua of H and He, and Mg II h & k, Ca II lines. A full list is given in Allred et al. (2015). For lines not calculated by RADYN the model atmosphere output can be ingested into full radiation transfer codes such as the RH code.
Documentation and analysis software is available at http://folk.uio.no/matsc/radyn/tools.tar
1.5. References
1. Abbett, W. P. & Hawley, S. L. Dynamic Models of Optical Emission in Impulsive Solar Flares, Ap. J. 521, 906 (1999)
2. Allred, J. C., Kowalski, A. F., Carlsson, M., A Unified Computational Model for Solar and Stel-lar Flares, ApJ 809, 104 (2015)
3. Capparelli, V. Zuccarello, F., Romano, P., Simoes, P. J. A., Fletcher L. et al. Hα and Hβ emission in a C3.3 solar flare: Comparison between observations and simulations, ApJ 2017 (in prep).
4. Carlsson, M. & Stein, R. F., Non-LTE radiating acoustic shocks and CA II K2V bright points, ApJ Letts. 397, L59 (1992)
5. Chatterjee, P., Hansteen, V., Carlsson, M., Modeling Repeatedly Flaring δ Sunspots Phys.Rev.Letts. Volume 116, Issue 10, id.101101
6. Falewicz, R., Rudawy, P. Murawski, K. Srivastava, A.K. 2D MHD and 1D HD models of a solar flare – a comprehensive comparison of the results, ApJ 813, 70 (2015)
7. Graham, D. R., Fletcher, L., Hannah, I. G. H., Milligan, R. O., The Emission Measure Distribution of Impulsive Phase Flare Footpoints ApJ 676, 83 (2013)
8. Graham, D. R., Cauzzi, G., Temporal Evolution of Multiple Evaporating Ribbon Sources in a So-lar Flare. ApJL, 807, 22 (2015).
9. Graham, D. R. et al., Spectral analysis and modelling of solar flare chromospheric condensation, (2017 in prep.)
10. Fletcher, L., Berlicki, A., Awasthi, A. K., Gronkiewicz, D. Recruiting flare hunters for citizen science, Astronomy & Geophysics 57 6.21 (2016).
11. Heinzel, P., Karlicky, M., Kotrc, P, Svestka, Z. 1994, On the occurrence of blue asymmetry in chromospheric flare spectra Sol. Phys. 152, 393
12. Heinzel, P., Kleint, L., Hydrogen Balmer Continuum in Solar Flares Detected by the Interface Region Imaging Spectrograph (IRIS), Astrophysical Journal Letters, 794, L23, (2014)
13. Heinzel, P., Kasparova, J., Varady, M. et al, Numerical RHS simulations of the flaring chromosphere with Flarix, arXiv:1602.00016 (2016)
14. Heinzel, P. et al.: 2D optically-thick radiation losses in cool flare loops (2017, in prep)
15. Kasparova, J. et al.: Mutual comparison between two RHD codes FLARIX and RADYN (2017 in prep)
16. Kerr, G. S., Simões, P. J.A. Qiu, J., Fletcher, L., IRIS observations of the Mg II H and K lines during a Solar Flare, Astronomy & Astrophysics 583, A50 (2015)
17. Kerr, G.S. Fletcher, L., Russell, A. J. B., Allred, J. C. Simulations of the Mh II k and Ca II 8542 lines from an Alfven Wave-Heated Flare Chromosphere, ApJ 827, 101 (2016)
18. Kerr, G., 2017, PhD Thesis, University of Glasgow.
19. Kleint, L.; Heinzel, P.; Judge, P. et al Continuum Enhancements in the Ultraviolet, the Visible and the Infrared during the X1 Flare on 2014 March 29, ApJ 816, 11 (2016)
20. Kotrč, P.; Procházka, O.; Heinzel, P. New Observations of Balmer Continuum Flux in Solar Flares. Instrument Description and First Results. Solar Physics 291, 229 (2016)
21. Kowalski, A. F., Cauzzi, G., Fletcher, L., Optical spectral observations of a flickering white-light kernel in a C1 solar flare, Astrophysical Journal 798, 107 (2015).
22. Kowalski, Adam F.; Allred, Joel C.; Daw, Adrian et al. The Atmospheric Response to High Non-thermal Electron Beam Fluxes in Solar Flares. I. Modeling the Brightest NUV Footpoints in the X1 Solar Flare of 2014 March 29 ApJ 836, 12 (2017)
23. Kuridze, D., Mathioudakis, M., Simoes, P. J. A. et al, H-alpha line profile asymmetries and the chromospheric flare velocity field, ApJ 813, 125 (2015)
24. Kuridze, D.; Mathioudakis, M.; Christian, D. J. et al, Observations and Simulations of the Na i D1 Line Profiles in an M-class Solar Flare ApJ 832, 147 (2016).
25. Liu, W.; Heinzel, P.; Kleint, L. Kasparova, J. Mg ii Lines Observed During the X-class Flare on 29 March 2014 by the Interface Region Imaging Spectrograph, Solar Physics 290, 3525 (2015)
26. Procházka, Ondřej; Milligan, Ryan O.; Allred, Joel C et al. Suppression of Hydrogen Emission in an X-class White-light Solar Flare (2017) accepted. arXiv170200638P
27. Ranjan, S. & Sasselov, D., Influence of the UV Environment on the Synthesis of Prebiotic Mole-cules Astrobiology 16, 68 (2016)
28. Russell, A. J. B. & Fletcher, L. Propagation of Alfvénic Waves from Corona to Chromosphere and Consequences for Solar Flares, ApJ 765, 81 (2013)
29. Simões, P. J. A. et al., Formation of the thermal infrared continuum in solar flares (2017, in prep)
Potential Impact:
2 Impact of the Project
2.1 Promoting the importance of the flaring chromosphere
We have worked hard to disseminate both the science of F-CHROMA, with a list of 55 (and growing) published, refereed papers. These have been cited by over 180 other papers so far, with a median number of downloads per published paper of over 100. These numbers are very healthy given the short lifetime of the project. Full metrics on the refereed papers can be monitored at https://ui.adsabs.harvard.edu/#/public-libraries/ANK_KO_BRuCqeYop3ZOFlw
Members of the F-CHROMA team gave (at least) 100 scientific talks and posters about the flare chromosphere and closely related physics at national and international events: a list of these is given in the “Use and Dissemination of Foreground” section of this report. In addition, dissemination activities specifically dedicated to advertising F-CHROMA and its outputs took place at the following events:
• ‘Solar and Stellar Flares’ Prague, Czech Republic, June 2014 (Poster, Fletcher et al)
• AAS Solar Physics Division Meeting, Boston, USA, June 2014 (Poster, Cauzzi et al)
• ‘RadioSun-5’, Ceske Budejovice, Czech Republic, 2016 (Poster, Berlicki et al)
• ‘CoolStars 19’, Uppsala, Sweden, 2016 (Talk, Berlicki et al)
• ‘Hinode 10’ Nagoya, Japan September 2016 (Carlyle & Carlsson)
Via our roles on the Science Working Group of the DKIST we have also emphasised the types of observations that can be carried out to probe the physics of the flaring chromo-sphere, and we have recently proposed to hold one of the first DKIST Community Work-shops with the aim of developing chromospheric flare science use cases.
2.2 Impact beyond solar physics
The Sun is our nearest active star, but observations, particularly from the Kepler satellite, show that magnetically-driven flaring is a common phenomenon in solar-type stars and stars of other spectral classes, particularly the M stars. The spectrum of stellar flares can be used, as in the case of solar flares, to diagnose the properties of the flaring atmosphere of stars, and our work will be useful here too in interpreting stellar observations, as the degree of spatial resolution allowed in solar observations (e.g. measuring velocities in individual pixels) breaks the degeneracy between space and time in a flare evolution in a way that is not accessible for stars. The irradiance spectrum is also significant for exoplanet studies. On the Earth, flare-related fluctuations in the solar irradiance in the UV-soft X-ray causes small transient ionisation changes in our ionosphere, which can temporarily change the upper atmosphere chemistry. But in exoplanets much closer to their hyper-active host stars than we are the changes to the planetary atmosphere might be very much more dramatic and long-lived, with implications for the stability of the atmosphere. Interestingly, the wavelength region be-tween 200nm and 300nm is thought to be critical for the formation of pre-biotic molecules and accessible through an early-Earth-like atmosphere (Ranjan & Sasselov 2016), so the use and development of RADYN simulations to quantify this irradiance could have influence in this exciting field.
2.3 Providing resources for the community
The major resources provided for the community have been the flare catalogue, the ground-based flare data archive, and the archive of RADYN flare models, as described in Section 1.4.4. The number of unique loads (unique users) of the start page of the flare catalogue and database in the first 20 months after its release was 1032, indicating a good level of interest. Top hits were from the UK, USA (reflecting, we believe, the announcements made in the national Solar Physics Newsletters) followed by Germany, Japan, India and China. The model archive is already being used by PhD students and postdocs within the F-CHROMA team members, greatly facilitating their research by avoiding the need to repeatedly run RADYN simulations. We anticipate many more users once the model archive is announced.
2.4 Training young scientists
A core dissemination activity was the organisation of a very successful F-CHROMA training school in the late autumn of 2016. It was dedicated to PhD students and young researchers to provide training on the physics of solar flares, flare data analysis, flare diagnostics and the use of the main modelling codes. The workshop was organized in Wroclaw (Poland). The full name of the workshop was: "F-CHROMA Training Workshop: Observations and Modelling of Solar Flares". Workshop information can be found at: http://school2016.astro.uni.wroc.pl/
We received 40 applications, which were carefully analysed by the SOC taking into account their scientific profile and their level. In the end we were able to accept 23 participants (14 female and 9 male) from 13 countries: Argentina, China, Czech Republic, Georgia, Greece, Ireland, Norway, Poland, Russia, Slovakia, Slovenia, United Kingdom, and the USA.
The workshop program reflected the F-CHROMA team expertise and F CHROMA team members, including some junior members, delivered all lectures and practical classes. All topics were divided into lectures and practical classes. The students also had the possibility to present themselves. There were some observational classes about data analysis and processing and more theoretical classes on the stationary and time-dependent modelling of the solar flaring atmosphere, including short introductions on the usage of main modelling codes.
At the end of the workshop, all participants, including lecturers, received the questionnaire to fill, to rate the workshop’s scientific level, organisation and other issues. The survey revealed that opinions about the workshop and its organisation were very positive and participants were satisfied. There was no clear difference between the opinions of male and female participants.
2.5 Involving amateurs
One of the most exciting parts of the F-CHROMA project was the organisation of two ‘professional-amateur’ (pro-am) flare observing campaigns. These were called "F-HUNTERS" and took place in 2015 and 2016:
• 1st F-HUNTERS Campaign: September 19-27, 2015
• 2nd F-HUNTERS Campaign: July 14-23, 2016.
The campaign periods were chosen to coincide with CHROMA team observing runs at ground-based observing facilities: 2015 F-HUNTERS#1: Dunn Solar Telescope at NSO (USA), Swedish Solar Telescope on La Palma (Spain), the MSDP at Bialkow Observatory (Poland) and the Multicamera Spectrograph, Ondrejov (Czech Republic); 2016 F HUNTERS#2: GREGOR Solar Telescope, Multi-channel Sunbtractive Double Pass Spectrograph, Bialkow Observatory (Poland) and Multichannel Flare Spectrograph, Ondrejov (Czech Republic). Observations by amateurs could therefore support the data obtained by professionals. In addition, the IRIS satellite dedicated some part of its time for a "flare watch" program giving additional spectral data in the UV range. Other space instruments (e.g. SDO, RHESSI) also contributed to these observations. Both campaigns were announced in a dedicated F-CHROMA dissemination site at http://fchroma.astro.uni.wroc.pl/index.php/f-hunters.html and were also announced on Facebook, Twitter and in many websites of amateur astronomers’ associations in many countries. There were also presentations and press announcements in local media (Figure 13).
The F-HUNTERS campaigns were coordinated via the F-CHROMA dissemination website. We prepared all information necessary for amateurs to observe and submit their data, including manuals with various levels of technical detail. All those interested in the F-HUNTERS campaign could register in the dissemination website in order to subscribe to the latest information with the observing solar targets during the campaign. For each day of the campaign, after analysis of solar images, the F-CHROMA team decided on the target for the flare watch on the following day. A special Information Sheet with the target details was prepared and sent to all subscribers by email (Figure 14). The same target info was published in Facebook, Twitter and in the F-CHROMA website. In total, we prepared 19 information sheets for all days of both campaigns. Professional observatories also observed the same target.
More than 150 participants registered on our F-CHROMA dissemination website and re-quested to receive the Information Sheet with solar observing target. Before the campaigns we also created a special form for observers who wished to send us their observational data. This form was available under the link: http://fchroma.astro.uni.wroc.pl/index.php/data-submission.html. Data came from more than 20 observers from many countries (e.g. Spain, Italy, Germany, Poland, UK, USA, The Netherlands, Czech Republic, Slovakia). The quality of the data was varied: from single JPEG images to the time series of FITs or TIFF sequences. Unfortunately, in most cases the amateur observers did not follow our request to calibrate the data using flat fields and dark current. In the F-HUNTERS#1 campaign, much of the data is low quality and very often the images don't contains the flare images but only active regions. However, the quality of the amateurs’ data was much higher for F HUNTERS#2 campaign. Examples of this are shown in Figure 12. These images are of high quality and may be suitable for scientific analysis.
We were very pleased with the enthusiasm and engagement of the amateurs, and of course also to the professional observatories who dedicated valuable observing time to the campaign. The campaign was promoted to fellow professional astronomers at two scientific meetings, the 37th Annual meeting of the Polish Astronomical Society, Posnan, Poland, 2015 (Talk & Poster, Berlicki) and the 12th Thinkshop Potsdam, Germany, 2015, (Poster, Berlicki). We have written an article about the campaigns and the lessons learned for the journal Astronomy & Geophysics (Fletcher et al. 2016)
List of Websites:
www.fchroma.org
