IUAP Project P7/08 CHARM

Contemporary physical challenges for Heliospheric and AstRophysical Models

Coordinator:Prof. Rony Keppens,Centre for mathematical Plasma Astrophysics, Department of mathematics, KU Leuven

Annual Scientific Report for the period 1­10­2012 till 30­11­2013.

Table of Contents.

List of abbreviations. General Information about partner teams. Introduction. Description of the research completed. Network organisation and operation. Publications.

List of abbreviations.

AU: Astronomical UnitAGU: American Geophysical UnionBATS’R’US: Block Adaptive Tree solver ‘Roe’ Upwind Scheme (Michigan)BISA: Belgian Institute for Space AeronomyBOF: Bijzonder Onderzoeks FondsCACTus: a software package for Computer­Aided CME tracking (ROB)CAstLe: Computational Astrophysics LeidenCHARM: Contemporary physical challenges for Heliospheric and AstRophysical ModelingCME: Coronal Mass EjectionCmPA: Centre for mathematical Plasma Astrophysics (KU Leuven)DNS: Direct Numerical SimulationESA: European Space AgencyEUV: Extreme UltraVioletFWO: Fonds Wetenschappelijk OnderzoekICC: Institute for Computational Cosmology (Durham)IIA: Indian Institute for AstrophysicsIMAGE (space mission)KH: Kelvin­HelmholtzKUL: KU LeuvenLCS: Low Coronal SignaturesLES: Large Eddy SimulationMHD: MagnetoHydroDynamicsMPI­AMRVAC: Message Passing Interface, Adaptive Mesh Refinement Versatile Advection CodeOpenGGCM: Open Geospace General Circulation ModelPCTR: Prominence Corona Transition RegionPIC : Particle­In­CellPROBA2: Belgian space mission, instruments are:PROBA2/LYRA; PROBA2/SWAPQPO: Quasi­Periodic OscillationROB: Royal Observatory BelgiumSDO: Solar Dynamics Observatory, instruments are:SDO/AIA; SDO/EVE; SDO/HMI; …SOHO: Solar mission, instruments are:SOHO/ACE; SOHO/LASCO; SOHO/SUMERSolar Orbiter: future solar mission, instruments are: Solar Orbiter EUISPH: Smooth Particle HydrodynamicsSPP: Statistical and Plasma Physics unit (ULB)SWMF: Space Weather Modeling Framework (Michigan)UCLA: University of California at Los AngelesULB: Universite Libre de BruxellesUT: Universal TimeVDF: Velocity Distribution FunctionVSC: Vlaams Supercomputing CentreWP: Work Package

General Information about the partner teams.

Five Belgian research teams, strengthened by two renowned international partners, join forces to investigate Contemporary physical challenges in Heliospheric and AstRophysical Models. The participating research teams are Centre for mathematical Plasma Astrophysics at KU Leuven, Astronomical Observatory at UGent, Fluid and plasma dynamics Research Unit at ULB, Solar Physics Research Department of the Royal Observatory of Belgium, Solar Wind Research Unit of the Belgian Institute for Space Aeronomy, Computational Astrophysics at Leiden University, Institute for Computational Cosmology at Durham University. An overview of the network partners and their research unit is given below.

Network Coordinator & Partner 1 team: KU Leuven Centre for mathematical PlasmaAstrophysicsName: Keppens Rony and the Centre for mathematical Plasma AstrophysicsInstitution: Department of Mathematics, KU Leuven

The research activities of the Centre for mathematical Plasma Astrophysics are in applied mathematics, with an emphasis on theoretical and computational plasma physics, relevant for solar physics, astrophysics and laboratory (fusion) plasmas. CmPA staff members target the coupling of continuum plasma physics descriptions with particle­based treatments, to address small­scale and turbulent reconnection phenomena in fully self­consistent global model efforts.

Partner 2 team: UGent Astronomical ObservatoryName: Baes Maarten and the `Sterrenkundig Observatorium'Institution: Department of Physics and Astronomy, UGent

The `Sterrenkundig Observatorium' in Ghent is a centre for extragalactic research and its focus has evolved over the past years into two (related) directions: the study of the cosmic evolution of galaxies through a combination of observations and numerical N­body/SPH simulations, and the observation and modeling of the interstellar medium of galaxies, with a special focus on the interstellar dust. This research is both observational and theoretical in nature.

Partner 3 team: ULB Statistical and Plasma Physics groupName: Knaepen Bernard and the Statistical and Plasma Physics groupInstitution: Department of Physics, ULB

The Statistical and Plasma Physics (SPP) group at ULB is specialized in the modelling and numerical simulations of turbulent (magneto­)hydrodynamical flows. All codes available in the SPP group run on high performance computer architectures with high scalability. Several group members have a strong expertise in the optimization of codes.

Partner 4 team: ROB Solar Physics and Space Weather directorateName: Berghmans David and the `Solar Physics and Space Weather' directorateInstitution: ROB

The Operational Directorate "Solar Physics and Space Weather" is one of the 4 scientific operational directorates of the Royal Observatory of Belgium (ROB). ROB specialized in automated data analysis by feature recognition software and in the physics of solar eruptions from corona till measured in situ at L1. The developed insights were put to use in the ROB space weather forecast center, now at the forefront of European space weather operations.

Partner 5 team: BISA Solar Wind unitName : Pierrard Viviane and the Solar Wind unitInstitution : BISA (BIRA­IASB)

The unit Solar Wind belongs to the Space Physics department of the Belgian Institute for Space Aeronomy. The research activities of this unit are mainly dedicated to the development of kinetic models relevant for the solar wind, for the planetary polar winds, aurora, the dynamics of the plasmasphere­ionosphere system and space radiations. Our activities also include the analysis of satellite observations (Cluster and IMAGE) for comparison with physics­based models.

International Partner 1: Universiteit Leiden CAstLe research groupName: Portegies Zwart Simon and the CAstLe research groupInstitution: Leiden Observatory, Universiteit Leiden

The research group for Computational Astrophysics Leiden (CAstLe) aims at studying the universe by means of simulation. The specific areas of research in astrophysics include the evolution of binary (and higher order multiple) stars, the dynamical evolution of dense stellar systems and of galactic nuclei. From a computational point of view, the research group aims at simulation environments for solving the equations for gravitational dynamics, stellar structure and evolution, hydrodynamics and radiative transfer.International Partner 2: Durham University Institute for Compational CosmologyName: Theuns Tom and the Institute for Computational Cosmology (ICC)Institution: Ogden Centre for Fundamental Physics, Durham University

The ICC is a world­leading group of cosmologists, focused on studying the formation and evolution of galaxies, and the cosmological aspects of dark matter and dark energy, observationally, using models, and using high­performance computing. The ICC is the main UK node of Virgo, a world leading consortium in the subject of cosmological simulations. It includes most groups in the UK active in cosmological simulation, with many international partners.

Introduction.

The CHARM network has started active collaborations on a variety of topics. As researchhighlights of the first year, we mention:

the realization of multi­dimensional models for radiatively mediated dynamics in the solarcorona, in particular the modeling of coronal rain as a result of thermal instability inrealistically structured coronal arcades (WP5)

the direct tool coupling established between the UGent Skirt Monte Carlo code and othernetwork tools, such as the KU Leuven MPI­AMRVAC software, allowing to perform virtualviews on the dust distributions in molecular clouds (WP1), as well as on realistic galaxiesas simulated in the Eagle project (WP5).

the breakthrough reached in Particle­in­cell (PIC) based modeling of turbulenceassociated with the Kelvin­Helmholtz instability (WP3).

Networking activities by the CHARM collaboration has utilized optimally its many links with international networks. In particular, we turned both of our training events into well­attended international schools: one was held in Leiden and jointly organized and co­financed by the ITN network Cosmocomp, a second was organized jointly with the FP7 program eHeroes and took place in Leuven. More on these events is listed in the section on network operation below.

Thanks to the CHARM visibility, we also secured additional funding through various channels. We can e.g. mention that postdoc Chun Xia, originally hired on CHARM funding at KU Leuven, was awarded a BOF/F+ half­year fellowship, and also is recipient for a Pegasus short 1­year fellowship. Similarly, PhD candidate Sofia Moschou, jointly supervised between KU Leuven and BISA teams, working on CHARM research, got awarded a FWO fellowship. At KU Leuven, Prof. Van Doorsselaere used the CHARM framework to apply for a Topping Up grant with Indian colleagues, and this was granted and strengthens our research. Using this Topping Up grant, we have hosted S. Krishna Prasad and D. Banerjee (from the Indian Institute for Astrophysics) at KU Leuven. As a scientific result from the Topping Up collaboration started between KU Leuven and Indian colleagues, we compared the observed damping dependence for omni­present long­period slow magneto­acoustic waves (that were discovered last year) with different theoretical models (WP2).

Description of the research completed.

Results within WP1: Challenges in coupling models and toolsWP Leader: Prof. Giovanni Lapenta, KU LeuvenTeams: KUL/UGent/ULB/BISA/Leiden/Durham

Modern numerical approaches for gas and plasma dynamics may employ radically different discretizations and application­specific representations. In this work package, our guiding principle is: Collaboration in concrete efforts to couple proven strategies, in order to address physical applications beyond the capabilities of our current suite of software tools.

WP1­A. Direct Tool Coupling.

Here we mention as an example the currently fully functional data exchange between the 3D block­tree adaptive simulations as produced with the massively parallel MPI­AMRVAC tool (KU Leuven), and the UGent Skirt code for postprocessing views on the dust distribution in simulated astrophysical scenarios. Specific applications focus on shear­flow related Kelvin­Helmholtz roll­up events in 3D molecular cloud environments where the gas distribution is enriched by multiple dust species. The first virtual views on the complex vortical patterns resulting from the Kelvin­Helmholtz nonlinear development have been demonstrated at the network meetings. Recent work concentrates on ensuring that the gas­dust simulations as performed with MPI­AMRVAC can also provide the full dust distribution as computed to the Skirt code, so that virtual views generated by assuming dust to gas density proportionalities can be contrasted with the actual dynamically evolved dust distribution. This is important as it shows rather different filamentary morphologies due to the pressureless evolution of the dust species, as influenced by drag forces. Joint work is in preparation, exploiting the Flemish Supercomputer Facilities at Ghent and Leuven (Tier1 and VIC3) to perform the computations, data analysis and visualization.

WP1­B Dynamic Coupling

Astronomical phenomena are governed by processes on all spatial and temporal scales, ranging from days to 13.8 Gyr as well as from km size up to the size of the Universe. This enormous range in scales is contrived, but as long as there is a physical connection between the smallest and largest scales it is important to be able to resolve them all, and for the study of many astronomical phenomena this governance is present. Although covering all these scales is a challenge for numerical modelers, the worst message is the equally broad and complex range in physics, and the way in which these processes propagate through all scales. In our recent effort to cover all scales and all relevant physical processes on these scales we continued the design of the Astrophysics Multipurpose Software Environment (AMUSE). AMUSE is a Python­based

framework with production quality community codes and provides a specialized environment to connect this plethora of solvers to a homogeneous problem solving environment. This provides opportunities to couple N­body solvers with large scale gas­dynamical evolutions, incorporating radiative transfer treatments.

For astrophysical plasma modeling involving charged constituents, we have investigated the physics and mathematical aspects of coupling large scale fluid models with local kinetic models. By their nature, the two approaches focus on different physical space scales and in consequence temporal scales. We focus especially on the coupling of magnetohydrodynamic (MHD) models with Particle­in­cell (PIC) codes for plasma physics. The former are used to study the global Earth space environment on scales of hundreds of Earth radii (order million km, one Earth radius is 6371 km) and considers hours or days of evolution. The latter works on a few tens Earth radii and spans seconds to a few minutes. By this clear separation of scales it is quite convenient and physically justified to spin off the PIC simulation from a MHD simulation. If one is interested in local effect on the kinetic scales, the feedback to the MHD run from the PIC run is irrelevant because the MHD model does not have the scales to react to such fast processes. To prove this one way coupling we have collaborated with UCLA to consider the specific event of 15 February of 2008 modelled by OpenGGCM and considered in particular the instant of 03:48 UT when a dipolarization event is observed to take off. This is the instant in the MHD run just before the event happens. We have then used the MHD state as initiation for an iPIC3D study. iPIC3D is our in house KU Leuven tool for PIC simulation, particularly suited for coupling with fluid models thanks to its implicit formulation based on the implicit moment method that presents more overlap with the fluid concepts and scales. We considered a subdomain of 20x12x12 Earth radii and compared the evolution of the fields in the MHD run alone and in the PIC run coupled to MHD. The latter not only shows the onset of many physical processes not present in MHD and allowing the energy release to progress at a much enhanced pace but it also allows us to study how particles are accelerated. This work has just been presented at the AGU meeting in San Francisco on Monday 8 Dec, 2013 and will be published shortly. The figure below shows an example of the extra physics produced by the PIC code coupled one way to MHD: the Hall quadrupolar magnetic field is displayed. The published material has of course the full information of all the new physics absent in MHD.

The top panel shows the state of the MHD global run done with openGGCM relative at 03:48 UT on February 15, 2008. The two bottom panels are the MHD state in the inset considered and the kinetic state spun off in a one way coupling approach. The extra physics due to the Hall effect leads to a quadrupolar component for the dawn­dusk magnetic field and to faster reconnection.

WP2­C. Integrated Coupling

While the time scales separation as advocated above is a means to obtain new physics insight on short term processes on the scales of minutes, to investigate the long term effect of kinetic­fluid coupling a full two­way coupling is needed. We have undertaken this task as well, in collaboration with the premier US space weather modelling effort: the SWMF of University of Michigan. With their help, we have worked at coupling the BATS’R’US MHD code (based in Michigan but available open source) with our iPic3D code. The coupling has been completed and applied first in benchmark cases to study the ability of the two codes to communicate with each other. The iPic3D code covers a limited portion of the domain where small scale physics is developing in a reconnection site. The two codes communicate without producing spurious effects and show the ability to feedback with each other. The example below compares the results from BATS’R’US used alone in the Hall MHD mode with the fully two ways coupled

Hall­MHD plus iPic3D codes, on the right. The iPic3D code covers the area within the box outlined in black. Reconnection is allowed to progress more swiftly by kinetic processes ignored in the pure Hall MHD run.

Left: Pure Hall MHD. Right: Hall MHD coupled with iPic3D (active only within the outlined box).

Results within WP2: Challenges for model validation through observationsWP Leader: Dr. David Berghmans, ROB.Teams: ROB/BISA/KUL

Numerical simulations and theoretical models are not ivory towers but need to be confronted with relevant observations. Irrespective of the physical domain (galaxy formation, solar eruptions, magnetospheric activity), this confrontation is not simply about constraining and validating simulations and models but is a complex, mutual interaction at various levels.

WP2­A.Case Studies.

In the framework of her PHD, Elke D'Huys (ROB), in collaboration with Dan Seaton (ROB) and Stefaan Poedts (KUL) initiated a study on the properties and initiation mechanisms for CMEs without distinct coronal signatures. Though easily visible in coronagraph observations, these so­called stealth CMEs do not obviously exhibit any of the low­coronal signatures typically associated with solar eruptions (changes in magnetic configuration, flows, solar flares, the formation of post­ flare loop arcades, EUV waves, erupting filaments, or coronal dimmings). We focus on what the presence or absence of these signatures can tell us concerning the mechanisms by which these stealth CMEs are initiated and driven.

CMEs without low coronal signatures (LCS) can have important implications on space weather, since many early warning signs for significant space weather activity are not present in these events. A better understanding of their characteristics and initiation mechanism will significantly improve our ability to assess their potential geo­effectiveness. We identified CMEs without LCS from the CACTus LASCO CME detections for 2012 by excluding CMEs associated with flares or other EUV variability, as well as back­sided events. Visual inspection revealed weak coronal signatures for many of the remaining candidate stealth CMEs and allowed to confirm 40 stealth CME detections. The general properties of stealth CMEs could then be characterized. Stealth CMEs are rather slow events, typically with a median velocity between 100 km/s and 500 km/s, although faster stealth CMEs were also identified. All CMEs with a width larger than 80 degrees were associated with low coronal signatures of an eruption. Most stealth CMEs occurred near the north pole. For CMEs with LCS the position angle is more evenly spread across the solar disk (Fig. below, left panel).

Left: Position Angle distribution for stealth CMEs (green) compared to CMEs with LCS (red, number of occurrences divided by 20). Right: CME angular width distributions for CMEs with and without low coronal signatures.

We investigated the frequency distributions of CMEs with and without LCS as a function of width (Fig. above, right panel). These distributions show a linear behavior over a large range of angular widths, indicating scale invariance. This implies that there is no typical size for a CME. We find that the distributions for stealth CMEs and CMEs with LCS have a different slope, suggesting a different initiation mechanism may be at work for each class of events. We also compared the height­time evolution of stealth CMEs to published results for different eruption mechanisms. The best fits to our measurements are exponential and parabolic profiles, corresponding to ideal MHD instabilities and breakout, respectively. Through numerical simulations of selected events, we will test if these models can explain the lack of coronal signatures observed with stealth CMEs. This lack suggests that these eruptions are not driven by impulsive reconnection near the solar surface, which is consistent with the evidence from our height­time profiles.

Daniel Seaton (ROB) studied the causes of an filament eruption that occurred on 2011 August 4 in conjunction with Francesco Zuccarello (KUL). This eruption took place in a complex active region that had experienced several eruptive flares in the days and hours leading up to the large eruption that we studied. We showed that flux cancellation at the neutral line and small reconnective events at the base of the erupting filament facilitated the slow rise of the filament to the point where it erupted via the torus instability. This result adds to mounting evidence that, although many processes can facilitate the destabilization of flux ropes in the corona, their actual eruptions are triggered by catastrophic loss of equilibrium due to ideal MHD instabilities. Additional energy release due to magnetic reconnection below the eruption powers the associated solar flare and contributes to the acceleration of the outgoing coronal mass ejection.

The associated figure shows the reconstructed three­dimensional position of the erupting filament, superimposed on a SDO/HMI magnetogram obtained at the time of the eruption. The reconstructed points are colored by the value of the decay index (N). The eruption began in the region where the decay index is around 1.4, the critical value for the onset of the torus instability.

WP2­B. Solar Flares and (coronal) seismology.

As a result of the Topping Up grant (Tom Van Doorsselaere), a collaboration was set up with the Indian Institute for Astrophysics (Dipankar Banerjee). The mutual exchanges have resulted in a the observational testing of theoretical damping mechanisms that have been put forward for explaining the decay of omni­present long­period slow magneto­acoustic oscillations in the solar corona. We have found that the damping time dependence is different in the polar regions than in the active regions. It could be that different damping mechanisms operate, and further investigations are ongoing.

Marie Dominique (ROB) started to investigate the quasi­periodic oscillations (QPO, see image below) that have been reported during the rising phase of a few solar flares (Dolla et al., ApJL, 2012). This analysis is performed in the frame of a PhD that she is doing under the supervision of Andrei Zhukov (ROB) and Giovanni Lapenta (KUL), and to which Laurent Dolla (ROB) and Tom Van Doorsselaere (KUL) are actively collaborating.

The mechanism that produces those QPO is not fully understood yet, but a few hypotheses have been proposed:

The reconnection process at the origin of the solar flare might exhibit a periodic behavior MHD waves could imprint oscillatory movements to the flare loops, possibly inducing

periodic variations of the flare signal. The same MHD waves could modulate the beam of electrons accelerated during the

reconnection process, so that they would reach the lower layers of the solar atmosphere as “packets”.

In the past, Tom Van Doorsselaere used two sets of QPO with different periodicities to perform coronal seismology (Van Doorsselaere et al., ApJ, 2011). Here, we focus on the observation of short­period oscillations (~10­20 sec), and confront them to the above­mentioned scenarios as well as to flare models, with the idea of being able to disentangle between the three mechanisms. Methods to detect QPOs based on wavelets and Fourier transforms have been

implemented and compared. A wide­range analysis, covering all flares above M5.0 that have been observed by the instruments PROBA2/LYRA and SDO/EVE has been initiated to determine the percentage of flares exhibiting QPO, and try to highlight common characteristics of the populations with and without QPO.

WP2­C. Computer­Aided observations.

The soFast processing pipeline to automatically detect flares in PROBA2/SWAP data has been further developed, (see Fig. below, and at http://sidc.be/sofast for realtime results). This work is part of the Phd of Katrien Bonte (KUL) and is conducted in collaboration with David Berghmans (ROB, co­promotor) and Dan Seaton (ROB). The work in 2013 consisted of building a complete catalogue of all flares ever observed with PROBA2/SWAP. The catalogue has been validated and checked for possible biases (Bonte, Berghmans, Poedts et al, 2014 in preparation). Meanwhile, a spin­off of this work is implemented in the EUI instrument on the future Solar Orbiter mission of ESA to automatically identify the most interesting data sequences that are worthy to send to the ground through the limited telemetry bandwidth.

Fig. Example of automated flare detection in SWAP data by the soFast algorithm. The flare is thesmall, bright patch in the top left corner of the red square on the left.

The same Solar Orbiter mission was extensively discussed as a WP2­C splinter during the first CHARM annual meeting (ROB, April 19 2013): “how can numerical models & large scale simulations help the Solar Orbiter science planners to determine the optimal solar pointing target?”

Results within WP3: Turbulence and particle acceleration in astrophysical plasmasWP Leader: Prof. Bernard Knaepen, ULB.Teams: ULB/KUL/ROB/BISA/UGent/Durham/Leiden

Hydro and magnetohydrodynamic turbulence is known to play an important role in many astrophysical systems possessing high kinetic and magnetic Reynolds numbers. As it is central to all applications of interest across our network collaboration, we formulate a major work package with guiding principle: Progress in both fundamental research on turbulent flows and induced particle accelerations, and place these insights directly in specific astrophysical contexts.

WP3­A. DNS to LES simulations with particle tracking.

In the general context of transport of charged particles, we considered two different approaches and related studies.

First, we started to study the role of small scale turbulent structures on the acceleration of charged tracer particles in MHD flows in relevant conditions for different plasma systems by means of particle trajectories and statistics obtained from high resolution direct numerical simulations (DNS) and those obtained in filtered versions of these DNS fields. By starting from a 32^3 numerical domain containing a well­developed turbulent plasma we feed higher and higher resolution simulations up to a numerical box of 1024^3. As constant of all simulations, we maintain a fixed product of the maximum resolved wave­number with the Kolmogorov length scale (measure of smallest scales in a given turbulent flow). We have considered so far two kinds of configurations: an isotropic initial configuration with random magnetic and velocity fields and the same but with a guide field. The purpose is to relate the particle diffusion­coefficient properties to the characteristics of the electromagnetic fields and to study the influence of small­scale structures on particle transport and to mimic their role in more realistic models. First attempts to filter out the small scales and observe particle trajectories were tested in low­resolution simulations of well­developed plasma turbulence.

Second, we considered a new approach to study fractional diffusion of tracer particles in plasma turbulence to understand velocity distribution function “fat tails”, thus, implying larger probabilities of finding a particle with a high velocity than with a regular Gaussian distribution. This characteristic has consequences for the description of certain phenomena in plasma physics, particularly for systems such as space plasmas or magnetically confined plasmas in thermonuclear fusion. The most standard approaches to the statistical dynamical description of these systems, being based on a perturbative treatment of the interaction strength, also known as the weak deflection limit, cannot take into account these high energy processes that are crucial to explain high energy tails of velocity distributions. On the other hand, the nonperturbative treatment of the interaction (the fact that the force acting on a given particle of the system can be arbitrarily strong) is also only valid for spatially homogeneous systems and in the short time limit.

The former feature is known to be unrealistic, as plasmas can exhibit large density inhomogeneities, especially in a state of turbulence. The latter limitation is simply “not practical”, as in an experimental context one is usually concerned with longer time regimes. The new approach we have developed overcomes these shortcomings by introducing hypotheses on the spatial and time coarse graining of the distribution function, .(r, , )f v t

.< ax d < m f(r,v,t)∇f(r,v,t)| |

In other words, the distance d is much smaller than the typical distance under which we can observe a significant density variation. And so, we assume that the interaction force felt by a given particle of the system due to any other particle further away than the distance d, can be treated perturbatively. The interaction due to particles within a distance d, on the other hand, will have to be given a full nonperturbative treatment. The second hypothesis concerns the time coarse graining of . We will assume that there exists a nonvanishing quantity that is (r, , )f v t tΔ much smaller than the typical time under which we can observe a significant variation of (r, , )f v tduring which the following approximation of the microscopic N­body dynamics is valid

F (x ) Δtv j = u j + 1m ∑

N

k=j/ j − x k

(t t) Δtr j + Δ = x j + u j

, corresponding respectively to the random velocity and position of the particle j at time t. u j x j Furthermore, we have theoretically estimated the time under which the above approximation is valid, and which is of the order of,

t Δ ~ √ mn α

where is the coupling constant ( for a plasma of identical charges ) and is the α /4πε e 20 e n

average particle density, and is the mass of the identical particles. It is certainly not a m coincidence that this time also corresponds to the inverse of the plasma frequency, indeed, this quantity is known to be the fastest oscillation occurring in a neutral plasma. So, in this new approach, this time merely corresponds to a small variation of the distribution function. Combining the two hypotheses on the spatial and time coarse graining, we finally get the following equation of evolution for the distribution function,

f(r, , ) − f(r, , ) ∂ f(r, , ) (r) n(x , ) (− ) f(r, , ) ∂t v t = v ∙ ∇ v t − 1m v v t ∙ Fm −C( mn α)

1/41 t Δv

3/4 v t

where

(r) r d v f(r , , ) F (r )Fm = ∫

d3 ′ 3 ′ ′ v′ t − r′

is the mean field, or Vlasov term, and is a constant. The fractional Laplacian 4/15 (2πα/m)C = 3/2 term in velocity space is the consequence of the nonperturbative treatment of the interaction. It describes the effect on the Vlasov dynamics of local field fluctuations caused by the fluctuation in positions of nearby particles. As it represents a correction, it is also much smaller in order of magnitude than the Vlasov term. Its proportionality to the local particle density indicates (r, )n t that this effect, as expected, will be stronger in regions where the concentration of particles is higher. The order of the fractional velocity derivative is, just as the index of the Holtsmark distribution describing the fluctuations of the electric field in a neutral plasma, also directly related to the behavior of the Coulombian interaction force over short distances./r1 2

WP3­B. Turbulence in the solar wind.

In collaboration with ROB, ULB performed a series of simulation studies by means of adapted shell models to model and study the effect of the expansion on solar wind turbulence. Different expansion regimes were considered: non­expanding, Alfvénic (strong coupling between magnetic field and velocity fluctuations, b and u respectively, or WKB) and non­Alfvénic regimes (weak coupling between magnetic and velocity fluctuations, or NWKB). Furthermore, for each regime we considered two possible polarisations (the wave­numbers can be orthogonal or parallel to the radial direction (R). The results show the existence of a frozen part of the energy spectrum around the first decades on the perpendicular simulations, a dominance of the linear expansion in the Alfvénic simulation with a R^(­1) energy decrease and an enhancement of turbulent effect in the non­Alfvénic simulation. The expansion could have for impact of slowing down the non­linear energy transfer (Alfvénic case). The decoupling of u and b in the non­Alfvénic case can cause an enhancement of turbulent process in the flow. The expanding shell model could retrieve observed wind properties like the existence of a f^(−1) part in the spectrum. However, those observations are for the the parallel component of the wave number while our results show it for the perpendicular component, this suggests the necessity of a coupling mechanism between those two directions to reproduce observational data. In the figure below, we show the spectral energy density compensated for a ­5/3 spectrum as a function of the wave­number (k) for different helio­distances (R) in presence of expansion (measured by the parameter ε here equal to 10, high expansion): on the (bottom) top panel you observe a (N)WKB expansion with perpendicular (left panel) and parallel (right panel) polarisation.

Other shell models will be used in the next future to see the model impact on the expansion dynamics, to take into account the effect of a (actually observed) average magnetic field and as intermediary step to full MHD simulations of solar wind turbulence with expansion effects. These results can be generalised to study opposite dynamics as well, that is with compressive effects, usually considered in the astrophysical context.

WP3­C. Turbulence in astrophysical (galactic to extragalactic) conditions.

The plasma conditions in space and astrophysical plasmas are often kinetic. We have carried out a heroic simulation requiring millions of CPU­hours for the evolution of the Kelvin­Helmholtz (KH) instability in conditions typical of the Earth magnetopause. The simulation evolves producing turbulent scales that range from the MHD scales to the smaller kinetic scales. All scales are covered thanks to the use of the implicit moment method of iPic3D that resolves the large scales due to its increased stability, stepping over the smallest Debye length scales that are not of interest in turbulence studies. The result below shows for the first time a turbulent cascade spanning both MHD and kinetic scales, showing a change in pace of the power law index at smaller scales where kinetic effects become important.

The turbulent state reached in the iPic3D evolution of the KH instability is shown on the left. Above the two power law indices developing to the left and to the right of the characteristic scale of the ion gyroradius (vertical line).

Results within WP4: From large to small with charged constituents: MHD to kinetictreatmentsWP Leader: Prof. Viviane Pierrard, BISA.Teams: BISA/ROB/KUL

Kinetic and MHD are complementary approaches to model space plasmas. In this work package, we will address topics where continuum MHD and kinetic aspects meet up, and are closely intertwined. Our guiding principle here is: Improve our understanding of space plasma phenomena (especially for the solar wind and magnetospheric realm), by correlating our MHD to kinetic models, ultimately leading up to full multi­scale treatments.

WP4­A. Kinetic/MHD complementarity.

As a first step in the improvement of the kinetic approach, we studied how the proton velocity distribution function (VDF) is modified by Alfvénic turbulence in the solar wind (Pierrard and Voitenko, 2013). We used Fokker­Planck diffusion terms to calculate the Alfvénic turbulence and solved numerically the kinetic evolution equation. Assuming a displaced Maxwellian for the protons at 14 solar radii, we showed that the turbulence leads to a fast development of an anti­sunward tail. These results provide a natural explanation for the nonthermal tails in the proton VDF observed in situ in the solar wind beyond 0.3 AU.

The kinetic approach was also used to study the links between the solar coronal characteristics and the solar wind at 1 AU as observed by SOHO, ACE, Stereo A and B. As a first approximation, we consider an exospheric model taking into account the effects of the external forces (gravitational and electromagnetic) in a magnetic field decreasing as r^­2 where r is the radial distance. The rotation of the Sun deforms the solar streams into the Parker's spiral shape. Using typical solar observations in the solar corona or at 1 AU as boundary conditions, we determine the solar wind characteristics in the meridian and the ecliptic plane from the solar corona to larger distances (Pierrard and Pieters, 2013).

WP4­B. Kinetic effects with Alfvén waves.

We studied the velocity­space quasi­linear diffusion of the solar wind protons driven by oblique Alfvén turbulence at proton kinetic scales (Voitenko and Pierrard, 2013). Turbulent fluctuations at these scales possess the properties of kinetic Alfvén waves that are efficient in Cherenkov­resonant interactions. The proton diffusion forms a quasi­linear plateau corresponding to the nonthermal proton tail in the VDF.

In collaboration with the Royal Observatory of Belgium, BISA researchers compared coronal temperature, density and heat flux profiles obtained from brightness measurements during solar eclipses and from kinetic models (Pierrard et al., 2013). This comparison indicates similar

distributions at large radial distances (> 7 Rs) in the collisionless region, but rather important differences in the acceleration region. A maximum of temperature is observed around 2 Rs in the polar regions. The exospheric heat flux based on Kappa VDF for the electrons is directed away from the Sun. This could indicate that the source of coronal heatings extends well above the transition region.

WP4­C. Multi­scale physics for reconnection.

In a code comparison effort, we addressed (Keppens et al, 2013) the long­term evolution of an idealised double current system entering reconnection regimes where chaotic behavior plays a prominent role. We quantified the energetics in high magnetic Reynolds number evolutions, enriched by secondary tearing events, multiple magnetic island coalescence, and compressive versus resistive heating scenarios. Our study paid particular attention to the required numerical resolutions achievable by modern (grid­adaptive) computations, and commented on the challenge associated with resolving chaotic island formation and interaction. We used shock­capturing, conservative, grid­adaptive simulations for investigating trends dominated by both physical (resistivity) and numerical (resolution) parameters, and confronted them with (visco­)resistive MHD simulations performed with very different, but equally widely used discretization schemes. This allowed us to comment on the obtained evolutions in a manner irrespective of the adopted discretization strategy.

Our findings demonstrate that all schemes used (finite volume based shock­capturing, high order finite differences, and PIC­like methods) qualitatively agree on the various evolutionary stages, and that resistivity values of order 0.001 already can lead to chaotic island appearance. However, none of the methods exploited demonstrates convergence in the strong sense in these chaotic regimes. At the same time, non­perturbed tests for showing convergence over long time scales in ideal to resistive regimes are provided as well, where all methods are shown to agree. This effort prepares for future work where multi­physics approaches will be deployed to study the physics at scales no longer allowing pure MHD treatments.

Results within WP5: Radiation & dynamics in astrophysical applicationsWP Leader: Prof. Maarten Baes, UGent.Teams: UGent/Leiden/Durham/KUL/ROB

Our guiding principle is here to cross­fertilize astrophysical models focusing on vastly different scales, through improved treatments of the radiative­dynamical feedback loop, improved physics modelling, and code validation. We strive for more realistic modeling efforts (1) in the solar atmosphere, connected to the formation and dynamics of solar prominences, (2) on the structure of the obscuring tori around active galactic nuclei, and (3) to the calculation of the observable properties of simulated dusty galaxies in large­scale cosmological simulations and for the multiphase interstellar medium.

WP5­A. Formation and dynamics of solar prominences.

In earlier work (Xia et al, 2012), we demonstrated the in­situ formation of a quiescent prominence in a sheared magnetic arcade by chromospheric evaporation and thermal instability in a multi­dimensional MHD model. At an IUA Symposium (Paris, 2013) we improved our setup and reproduced the formation of a curtain­like prominence from first principles, while showing the coexistence of the growing, large­scale prominence with short­lived dynamic coronal rain in overlying loops. When the localized heating is gradually switched off, the central prominence expands laterally beyond the range of its self­created magnetic dips and falls down along the arched loops. The dipped loops recover their initially arched shape and the prominence plasma drains to the chromosphere completely.

KU Leuven research has also realized a breakthrough in our understanding of solar rain, a much more dynamic and small­scale process related to solar prominences. Solar rain is quite a bit different from its wet variant on Earth: in the solar atmosphere, raindrops measure up to 400 kilometer in width, and 800 kilometer in length. They also do not fall straight down, but follow arched pathways through the hot solar corona. They sometimes even go up, instead of falling down to the solar surface. Coronal rain manifests itself in the million degree hot solar atmosphere, where all matter is in the fourth physical state: plasma. The rain is a consequence of a local, sudden and catastrophic loss of thermal equilibrium, as ever more energy escapes the plasma by radiation (light). As a consequence, much cooler and heavier plasma blobs are formed continuously, which are forced to follow the curious bends traced out by the solar magnetic fields. Sometimes, pressure differences build up, causing the blobs to go up, once in a while. In a publication in the Astrophysical Journal Letters, PhD student Xia Fang, together with Chun Xia and Rony Keppens at the CmPA, explain how these results represent a worldwide scoop for solar plasma­astrophysics.

One essential element to improve modeling and make them to be more realistic in describing prominences formation, stability and disappearance, is to have good observational constraints. ROB is contributing to this effort by working to understand the properties of the radiative emission of the prominence­corona transition region (PCTR). This thin layer is important as it interfaces the structure with the environment corona, and thus may have a role in the prominence thermal and pressure stability. It also contributes to the whole prominence radiative losses which, under stable conditions, are almost constant in time, suggesting the presence of a continuous heating injection into the structure. The PCTR, as the prominence cool core, has a dynamic nature, with observed flows that may have a role in the mechanical stability of the whole structures. All these observational elements, once opportunely defined, are good inputs for prominences modeling. One of the recent results of ROB work has been to interpret the SDO/AIA images of prominences in terms of their thermal properties. The figure below shows a prominence observed with the AIA band centered at 171 Å, that is optimized to collect the emission at around log T = 5.8. Beside this band contains contribution from cooler emission, the work of Parenti et al. 2012 showed that the emitting plasma had at least a temperature of 4.5 105 K. This result suggests that the PCTR contains more plasma at high temperature than previously thought.

Left: SDO/AIA image of a prominence in the Fe IX 171 Å (log T = 5.8) band. Around the cool dark core is visiblesome emission of the PCTR. Right: a radial profile of the AIA intensity in the 171 band inside the prominence(solid line) and in the background corona (dashed). Emission of the prominence above the corona is seenaround pixel 200. From Parenti et al 2012.

This result has stimulated further investigations, as it was achieved using imager data, having the properties of integrating multiple­temperature information. The next step is to confirm and better quantify this result using spectroscopic data, which provide unambiguous temperature information. For this goal a quiescent prominence was observed in 2012 with the SOHO/SUMER spectrometer which collected data from spectral lines emitted by the plasma of the PCTR. The figure below (left) shows the prominence as observed by SDO/AIA imager in the 171 band. Here the off­limb prominence is partially seen in emission. The figure below (right) shows the variation of the Mg X line intensity along the SUMER slit, which crosses the prominence. This line is emitted by plasma at a temperature of 1MK. The pixels lower than 240 collected the photons from the off­limb emission (the step in intensity marks the limb). The low signal region between pixels 150 and 200 is due to the absorption of these photons by neutral hydrogen present in the dense core of the prominence. Above this region (pixels 120­150) the Mg X line is still present slightly in emission above the background corona, indicating that it originates from the PCTR and contributes to its radiative losses. Here the plasma is less dense and warmer than in the core, so that the neutral hydrogen absorption is less effective. The work is still in progress, as we need to better quantify these finding, but such preliminary results seems to confirm the Parenti et al. 2012 results.

SDO/AIA 171 Å image of theprominence observed in 2012. Thewhite vertical line marks the positionof the SOHO/SUMER slit.

Mg X line intensity along the SUMER slit. This line is emitted byplasma at 1MK. A faint emission above the background corona isseen above the limb in the PCTR (pixels 120­150).

WP4­B. Radiative transfer in large­scale cosmological simulations & WP5­D. Multi­phaseinterstellar medium.

New IUAP PhD student James Trayford, who is 50 per cent funded by a Durham grant, has been hired to apply the SKIRT radiative transfer code developed in Gent to the Eagle simulations. Eagle is a large simulation project that aims to generate a realistic galaxy population in a cosmologically representative volume. Its key design requirement is to use a physically motivated sub­grid model for galaxy formation to mimic processes below the resolution scale, such as star formation, and the feedback from supernovae and accreting black holes, and reproduce the observed galaxy stellar mass function.

The international Eagle collaboration (nodes Durham, Leiden, Heidelberg and Munich) has been successful in securing a large (40 M hours) CPU time allocation on the Prace Curie facility in Paris (PI J Schaye, Leiden), as well as on the Dirac 2 facility in Durham. An example image showing the gas density in a thin slice centred on a group of galaxy is shown below. Colours refer to gas temperature (hot: red, warm: green, cold:blue), with zooms onto the main and satellite galaxies as insets.

Gas distribution in a slice through a galaxy group in the Eagle simulation, colour coding denotesgas temperature.

Trayford has developed new software to enable an in depth comparison of the galaxies to data. Using stellar ages and metallicities as computed by the simulation, Bruzual­Charlot isochrones are used in population synthesis modelling. We found that the coarse sampling of star formation in the simulations unduly affects the blue colours of galaxies, with a few recent star formation events potentially dominating the colour of a star forming galaxy. We therefore use the star formation rate rather than actual young stars to compute the colour­magnitude above.

G­R versus R galaxy colour magnitude diagram for the Eagle galaxies (dots) overplot on thebimodal distribution observed for nearby galaxies (contours)

The next step involves using SKIRT for computing dust obscuration and in the longer time dust emission. This will enable us to look at submillimeter galaxies as well. The UGent group has been developing the SKIRT code to make the post­processing of SPH generated galaxy models possible. In the first place, this necessitates a framework to run simulations with complex 3D geometries and large dynamic ranges. PhD students Waad Saftly and Peter Camps (the latter funded by the IUAP grant) has developed different hierarchical grid structures in the SKIRT code to make such simulations possible. The use of hierarchical octrees in radiative transfer simulations has been quite common, but Saftly et al. (2013, A&A, 554, A10) presented the first in­depth investigation on construction and traversal algorithms on such grids in the framework of Monte Carlo radiative transfer. Beyond these “classic” hierarchical grids, we have also explored the use of more experimental grid structures for radiative transfer. Camps et al. (2013, A&A, 560, A35) for the first time used unstructured 3D Voronoi grids in the context of radiative transfer, and demonstrated that, in spite of the geometrical complexity, these grids have interesting advantages (see figure below). In fact, the benefits of using a Voronoi grid in radiative transfer simulation codes will often outweigh the somewhat slower performance. Finally, Saftly et al. (2013b, A&A, in press) explored hierarchical k­d tree grids as an alternative to octree grids. They found that k­d trees are generally superior to octrees for both discretization efficiency and grid

traversal, and their use is strongly advocated.

Illustration of the use of unstructured Voronoi grids in 3D dust radiative transfer simulations. The top row shows the dust density distribution of a simple spiral galaxy model, the bottom row shows the calculated image. The left panels are calculated using a hierarchical octree, the right correspond to a Voronoi grid. From Camps et al. (2013).

Thanks to these advancements in the radiative transfer code, it is now feasible to efficiently deal with complex 3D geometries with arbitrary density gradients. In particular, the implemented changes enable to post­process the results one obtains from cosmological N­body/SPH simulations. As an example, the figure below shows a true colour image of an edge­on Eagle galaxy below generated using SKIRT. The dust lane is clearly visible.

B­V­R SKIRT image of an edge­on Eaglegalaxy. Assuming dust abundance dependson local gas metallicity results in the dustlane.

WP5­C. Obscuring Tori and AGN jet dynamics.

Relativistic hydro and magnetohydrodynamics (MHD) provide a continuum fluid description for plasma dynamics characterized by shock­dominated flows approaching the speed of light. Using our open­source software MPI­AMRVAC code (Keppens et al, PPCF, 2013), we used hydrodynamical models to quantify how energy transfer from Active Galactic Nuclei (AGN) jets to their surrounding interstellar/intergalactic medium (ISM/IGM) gets mediated through shocks and various fluid instability mechanisms. With jet parameters representative for Fanaroff­Riley type­II jets with finite opening angles, we quantiied the ISM volumes affected by jet injection and distinguished the roles of mixing versus shock­heating in cocoon regions. This provides insight in energy feedback by AGN jets, usually incorporated parametrically in cosmological evolution scenarios. While relativistic hydro models allow to better constrain dynamical parameters like the Lorentz factor and density contrast between jets and their surroundings, the role of magnetic fields in AGN jet dynamics and propagation characteristics needs full relativistic MHD treatments. Then, self­stabilization mechanisms related to the detailed magnetic pitch variations are at play.

Result within WP6: Training and dissemination activitiesWP Leader: Prof. Rony Keppens, KU LeuvenTeams: All partner teams

This work package is discussed in the following section, as part of the network organisation andoperation.

Network organisation and operation.

Activities organized between 1 October 2012 and 30­11­2013 as part of the IAP network CHARMare:

Our CHARM kickoff meeting, serving as our first IAP initiative, was held at KU Leuven onOctober 8­9, 2012. With about 50 participants, and all network nodes present to showcase theirexpertise to the network peers, a strong programme was put together. It was mixed withdemonstration from example simulation tools, ongoing joint PhD works, and scientific aspirationsfor collaborative work. The final day also dedicated time to management aspects, and the plansto organize joint schools during our first year were kick­started.

Our annual meeting was held at ROB, on 18­19 April 2013. We organized the meeting withmore extensive talks about the various institutes expertise, and continued to organize theremainder of the meeting with work package relevant presentations. For each, a tutorial lecturesketched the broad scientific theme and challenges, followed up by ongoing research betweenthe teams. On friday, specific splinter sessions served to brainstorm on work package follow­upin parallel sessions. These culminated in slideless reports by session representatives to thewhole network. A further part discussed the management, and the upcoming school planning.

As a concrete network training event, in Leiden, May 27­31 2013, our IAP organized, inconjunction with the ITN Cosmocomp, a spring school on Radiative transfer treatments forastrophysical applications. We had to close the registration early, as we reached maximalcapacity (30 participants+) before the deadline. The format was well­received by all students andlecturers: extensive tutorials in the morning parts, followed by hands­on exercises with codeslike Skirt, RamsesRT, Enzo, Urchin, or frameworks like AMUSE. The poster session was a veryinteractive and lively event, as was the dinner in downtown Leiden.

A second network training event was held in Leuven, Sept 16­19, 2013, entitled Space Science Training Week 2013: Data Driven Modeling and Forecasting. This summer school introduced a generation of young researchers (advanced master students, PhDs, and junior postdoctoral researchers) to the diverse aspects of space weather related research. It covered theoretical approaches to space weather and its drivers, presented modern solar data analysis

tools, and covered state­of­the­art solar and space science simulations. Participants learned about forecasting aspects and their quality control for space weather events, but also experienced hands­on training in scientific proposal writing and received do­and­don't tips for scientific presentations. The scientific program was enriched by a public evening lecture on the solar influence on our climate, given by Prof. Sami K. Solanki, Director Max Planck Institute for Solar System Research, Lindau, on Wednesday, September 18, 2013. The school could benefit from its embedding within CHARM, and a European FP7 Project eHeroes [http://www.eheroes.eu] with 15 different partner institutes.

Next to these network wide events, we can mention visits between the various teams and WPbrainstorm sessions such as:

1. October 12, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

2. October 19, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

3. October 26, 2012: Lapo Bettarini (ULB) visits ROB to discuss about co­supervisedMaster Thesis of Thibault Le Polain on turbulence in expanding solar wind.

4. October 26, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

5. November 09, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

6. November 16, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

7. November 19, 2012: Elke D’Huys (ROB) visits KUL to discuss research progress8. November 23, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress9. November 30, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress10. December 07, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress11. December 13, 2012: V. Pierrard (BISA) visits KUL for co­supervised PhD work of Sofia

Moschou.12. December 12­14, 2012: Peter Camps (UGent) visits Durham to discuss the SKIRT

component of the EAGLE simulation project.13. December 14, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress14. December 18, 2012: Andrea Verdini (ROB) visits ULB for co­supervised Master Thesis of

Thibault Le Polain.15. December 21, 2012: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress16. January 14, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress

17. February 5, 2013: Andrea Verdini (ROB) visits ULB for co­supervised Master Thesis ofThibault Le Polain.

18. March 06, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress19. March 11, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress20. March 11, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress21. March 13, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress22. March 13, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress23. March 18, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress24. March 18, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress25. March 20, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress26. March 20, 2013: Peter Camps (UGent) visits KUL to discuss SKIRT­MPI­AMRVAC

coupling.27. March 20, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress28. March 25, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress29. March 25, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress30. March 27, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress31. March 27, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress32. March­July, 2013: Andrea Verdini (ROB) weekly (on tuesday) visits ULB for

co­supervised Master Thesis of Thibault Le Polain.33. April 8 and 26, 2013: BISA (Vivianne Pierrard, Yuri Voitenko) and ROB (Andrei Zhukov,

Laurent Dolla, Luciano Rodriguez) teams discuss solar wind models and observations.34. April 22, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress35. April 24, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress36. April 25 and 26, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress37. April 26, 2013: Marie Dominique (ROB) visits KUL for PHD course and discussion on

research progress38. May 16, 2013: Lapo Bettarini (ULB) visits ROB for co­supervised Master Thesis of

Thibault Le Polain.

39. May 21, 2013: Lapo Bettarini (ULB) visits KUL to discuss WP1 plans.40. May 23, 2013: Rony Keppens (KUL) visits UGent to take part in the PhD jury for Dr. Joris

Verstappen.41. May 24, 2013: Lapo Bettarini (ULB) visits ROB for co­supervised Master Thesis of

Thibault Le Polain.42. May 24, 2013: Michael Pieters (BISA) visits KULeuven to discuss kinetic/MHD

approaches of the solar wind and prepare a possible joint PhD.43. May 30, 2013: BISA and ROB organized an Alfven workshop44. June 17­20: Dr. Qingmin Zhang (Nanjing, China, Purple Mountain Observatory) visits

KUL, to collaborate with Chun Xia and other CmPA team members.45. June 10­28, 2013: S. Krishna Prasad (India) visited KUL using a topping up grant

connected to CHARM.46. June 24­28, 2013: Prof. Dipankar Banerjee (India) visited KUL using a topping up grant

connected to CHARM. He also gave a Capita Selecta course in the Master ofMathematics.

47. July 2, 2013: David Berghmans (ROB) visits KULeuven to discuss progress of commonPhD student Katrien Bonte.

48. July 29, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress49. August 05, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress50. August 12, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress51. August 22, 2013. Vivianne Pierrard visits KU Leuven, to take part in the ADS progress

report meeting for Sofia Moschou.52. September 05, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress53. September 16­17, 2013: D. Seaton (ROB) visits KUL to give lectures at the Space

Science Training Week54. October 7, 2013. Kris Borremans (BISA) gives a seminar on his research to the ADS

doctorate committee at KU Leuven.55. October 8, 2013: Lapo Bettarini (ULB) visits KUL to discuss with Giovanni Lapenta the

setting­up of simulation of turbulence for WP3.56. September 23 ­ October 4, 2013: Marko Stalevski (Astronomical Observatory of

Belgrade) visits UGent to discuss the inclusion of polarization in the SKIRT radiativetransfer code and the modeling of AGN.

57. September 30 ­ October 4, 2013: René Goosmann (Strasbourg) visits UGent tocollaborate with Maarten Baes on polarized radiative transfer.

58. October 03, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

59. October 10, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion onresearch progress

60. October 16 and 17, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussionon research progress

61. October 23 and 24, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussionon research progress

62. October 24, 2013. Rony Keppens visits UGent to give a public outreach lecture for the

VVN on Campus de Sterre.63. October 24, 2013. Marie Dominique (ROB), David Berghmans (ROB) and Andrei Zhukov

(ROB) visit KUL to discuss with Giovanni Lapenta the progress of common PhD MarieDominique.

64. October 30, 2013: Elke D’Huys (ROB) visits KUL to discuss research progress65. November 07, 2013: Elke D’Huys (ROB) visits KUL for PHD course and discussion on

research progress66. November 13 and 14, 2013: Elke D’Huys (ROB) visits KUL for PHD course and

discussion on research progress67. November 14, 2013: Jo Raes (UGent) visits KUL to discuss joint PhD under CHARM68. November 20, 2013: During European Space Weather Week (http://sidc.be/ESWW10)

Stefaan Poedts (KULeuven) and David Berghmans (ROB) discuss the progress ofcommon PhD Katrien Bonte.

69. November 27 and 28, 2013: Elke D’Huys (ROB) visits KUL for PHD course anddiscussion on research progress

70. November 28, 2013: Tom Van Doorsselaere (KUL) visits UGent to discuss joint PhD ofJo Raes.

71. December 6, 2013: Katrien Bonte (KUL) visit ROB to discuss PhD work with D. B.Seaton.

72. September 9­11, 2013: James Trayford (Gent) visits M Baes (Gent) to discuss PhD work73. August 17­24, James Trayford attended Monte Carlos RT summer school in St Andrews

A list of current co­supervised PhD students is:

Sofia Moschou (KU Leuven­BISA) Promotors: Keppens, Pierrard, counselor BerghmansKris Borremans (KU Leuven­BISA) Promotors: Pierrard, Lapenta, counselor BerghmansMarie Dominique (KU Leuven­ROB) Promotors: Zhukov, Lapenta, counselor PierrardK. Bonte (KULeuven­ROB) Promotors: Poedts, BerghmansElke D’Huys (ROB­KU Leuven) Promotors: Seaton, Poedts, counselor PierrardPeter Camps (UGent­Durham) Promotors: Baes, Theuns, counselor KeppensJo Raes (Ugent­KU Leuven) Promotors: Baes, Van Doorsselaere, counselor TheunsJames Trayford (UGent ­ Durham) Promotors Theuns & Bower, co­promotor Baes

In preparationGraciela Lopez Rosson (BISA­KULeuven) Promotors: Pierrard, KeppensPieters Michael (BISA­KULeuven) Promotors: Pierrard, Poedts

Publications.

a) Publications from each team.

CmPA, KU Leuven:

Q.M. Zhang, P.F. Chen, C. Xia, R. Keppens, H.S. Ji, 2013, A&A 554, A124. Parametric survey oflongitudinal prominence oscillation simulations.

X. Fang, C. Xia, R. Keppens, 2013, ApJ Letters 771, L29. Multidimensional modeling of coronal raindynamics.

R. Keppens, O. Porth, K. Galsgaard, J.T. Frederiksen, A.L. Restante, G. Lapenta, C. Parnell, 2013, Phys.of Plasmas 20, 092109 (17pp). Full paper, doi: 10.1063/1.4820946 Resistive MHD reconnection: resolvinglong­term, chaotic dynamics.

R. Keppens, O. Porth, R. Monceau­Baroux, & S. Walg, 2013, Plasma Phys. Control. Fusion 55, 124038(7pp). Full paper, doi: 10.1088/0741­3335/55/12/124038 Relativistic HD and MHD modelling for AGN jets

Jess D.B., Reznikova V.E., Van Doorsselaere T., Keys P.H., Mackay D.H., 2013, ApJ 779, 168 (11pp),accepted, The influence of the magnetic field on running penumbral waves in the solar chromosphere

Moreels M.G., Goossens M., and Van Doorsselaere T., 2013, A&A, vol. 555, pp. A75, Cross­sectional areaand intensity variations of sausage modes. DOI: 10.1051/0004­6361/201321545

Srivastava, A.K., and Goossens, M.: 2013, ``X6.9 Flare induced Vertical Kink Oscillations in a Large­ScalePlasma Curtain as observed by SDO/AIA '', The Astrophysical Journal, The Astrophysical Journal, 777:17(9pp), 2013 November 1

Soler, R., Goossens, M., Terradas, J., and Oliver, R.: 2013, ``The behavior of transverse waves in nonuniformsolar flux tubes. I. Comparison of ideal and resistive results'', The Astrophysical Journal, 777:158 (17pp),2013 November 10.

ROB:

Zuccarello, F. P., Seaton, D. B., Mierla. M., Poedts, S., Rachmeler, L. A., Romano, P., Zuccarello, F.,Observational evidence of torus instability as trigger mechanism for coronal mass ejections: the 2011August 4 filament eruption, ApJ, 2013, Submitted.

BISA:

Pierrard V. and Y. Voitenko, Modification of the proton velocity distributions by Alfvénic turbulence in thesolar wind, Solar Phys., 288, 355­368, 2013, doi: 10.1007/s11207­013­0294­8.

Voitenko Y. and V. Pierrard, Velocity­space proton diffusion in the solar wind turbulence, Solar Physics,

288, 369­387, 2013. doi: 10.1007/s11207­013­0296­6

Pierrard V., K. Borremans, K. Stegen and J. Lemaire, Coronal temperature profiles obtained from kineticmodels and from coronal brightness measurements obtained during solar eclipses, Solar Physics, 289,183­192, doi: 10.1007/S11207­013­0320­x, 2014

Pierrard V. and M. Pieters, Toward a 3D kinetic model of the solar wind, accepted in 12th AnnualInternational Astrophys. Conf. series, 2013.

ULB:

UGent:

De Geyter G., Baes M., Fritz J., Camps P., 2013, A&A, 550, A74. FitSKIRT: genetic algorithms toautomatically fit dusty galaxies with a Monte Carlo radiative transfer code.

Saftly W., Camps P., Baes M., Gordon K. D., Vandewoude S., Rahimi A., Stalevski M., 2013, A&A, 554,A10. Using hierarchical octrees in Monte Carlo radiative transfer simulations.

Camps P., Baes M., Saftly W., A&A, 560, A35. Using 3D Voronoi grids in radiative transfer simulations

Saftly W., Baes M., Camps P., A&A, in press (arXiv:1311.0705). Hierarchical octree and k­d tree grids for3D radiative transfer simulations

Durham­Leiden.

Crain R A, McCarthy I G, Schaye J, Theuns T, Frenk C., 2013 MNRAS, 432, 3005. Enriching the hotcircumgalactic medium

Sawala T, Frenk C S, Crain R A, Jenkins A, Schaye J, Theuns T, Zavala J., 2013 MNRAS, 431, 1366. Theabundance of (not just) dark matter haloes

Altay G, Theuns T, Schaye J, Booth C, Dalla Vecchia C, 2013 MNRAS, 436, 2689. The impact of differentphysical processes on the statistics of Lyman­limit and damped Lyman α absorbers

Haas M R, Schaye J, Booth C M, Dalla Vecchia C, Springel V, Theuns T, Wiersma R, Robert P C, 2013MNRAS, 435, 2391. Physical properties of simulated galaxy populations at z = 2 ­ I. Effect of metal­linecooling and feedback from star formation and AGN

Rollinde E, Theuns T, Schaye J, Paris I, Petitjean P, 2013 MNRAS, 428. 540. Sample variance and Lymanα forest transmission statistics

van Elteren A, Pelupessy I, Portegies Zwart S, 2013 submitted to Phil Trans of the Royal Society.Multi­scale and multi­domain computational astrophysics

b) Co­publications. (publications with at least 2 different research teams, and contractuallyassigned to the IAP project).

Several joint works are under revision or in preparation, so we expect this list to start up andgrow in the coming year.