Toward Forecasting Space Weather Tamas Gombosi The University of Michigan Isradynamics 2010 Conference Ein Bokek (Dead Sea), Israel April 11-16, 2010

Space Weather Space weather: conditions in geospace that can have adverse effects on space-borne and ground-based technological systems and can endanger human life or health. (courtesy of Doug Biesecker, NOAA SWPC)

NRC report: The Economic And Societal Impacts of Space Weather

Modeling the Space Weather System Disparate time scales Disparate spatial scales Disparate physics …… ☞ Model coupling is needed

Coupled Space Weather Models Space Weather Modeling Framework (SWMF) Centered at the University of Michigan Enables flexible Sun-to-Earth model chain Single executable Center for Integrated Space Weather Modeling (CISM) coupled model Centered at Boston University Enables Sun-to-Earth model chain Multiple executables Integrated Magnetosphere-Ionosphere Model Centered at the University of New Hampshire Couples a global magnetosphere, an ionosphere- thermosphere, a radiation belt and a ring current model Single executable

Freely available at

SWMF Capabilities: Physics Fluid Equations Compressible HD Ideal MHD Semi-relativistic MHD Resistive MHD Single-fluid Hall MHD Two-fluid Hall MHD Multi-species (Hall) MHD Multi-fluid (Hall) MHD Anisotropic pressure Heat conduction Additional Physics Multiple materials Non-ideal EOS Radiation Gray diffusion Multigroup diffusion Source terms Gravity, mass loading, chemistry, photo-ionization, recombination, etc… Various resistivity models Semi-empirical coronal heating Alfven wave energy transport and dissipation Self-consistent turbulence

SWMF Capabilities: Numerics Time integration Schemes Local time-stepping for steady state Explicit (with Boris correction) Explicit/implicit Semi-implicit Point-implicit Grids Block-adaptive tree Cartesian Generalized grids including spherical, cylindrical, toroidal TVD Solvers Roe HLLD HLLE Artificial-wind Rusanov Limiters Koren (3 rd order) MC Beta Div(B) control 8-wave Hyperbolic/parabolic scheme Projection Staggered grid (CT) Ray tracing Fast & parallel Synthetic images White light coronagraph EUV LOS 171Å, 195Å, 284Å X-ray radiographs Tomography

SWMF Capabilities: Framework Source code: 250,000 lines of Fortran in the currently used models 34,000 lines of Fortran 90 in the core of the SWMF 13,000 lines of Fortran 90 in the wrappers and couplers User manual with example runs and full documentation of input parameters Fully automated nightly testing on 9 different machine/compiler combinations SWMF runs on any Unix/Linux based system with a Fortran 90 compiler, MPI library, and Perl interpreter

Simulation and Tomography : : : : : : : : : : : : : : :44 n e =10 6 n e =5x10 5 Frazin et al., 2008

Thermodynamic Solar Wind Model (van der Holst, Oran and Sokolov) Sub-photosphere model Multigroup radiation transport MHD Magnetic flux emergence Low corona (1R ☉ r<2.5 R ☉ ) γ=5/3, single temperature Heat conduction Transport and dissipation of total energy of Alfvénic turbulence (E ± ) Wave dissipation represents sources for the plasma momentum and internal energy Additional coronal heating is obtained from the “unsigned flux” model (Abbett 2007) and observed X-ray luminosity (Pevtsov 2003) Corona (2.5R ☉ <r<20R ☉ ) γ=5/3, separate ion and electron temperatures Transport and dissipation of frequency resolved Alfvén wave intensity, I ± (ω) Kolmogorov spectrum is assumed at the inner boundary Observational inputs: Magnetogram driven potential field extrapolation (synoptic maps from GONG or MDI) Density and temperatures near the sun are predicted by the DEMT (Differential Emission Measure Tomography) results of Vasquez and Frazin (2009). Solar X-ray luminosity (EIT, STEREO)

Boundary Conditions for CR2077 with no Free Parameters GONG magnetogramWSA relation STEREO/EUVI Differential Emission Measure Tomography (Vasques & Frazin)

CR2077: Two State Solar Wind In the fast wind Electrons are cold (due to adiabatic cooling) Ions are hot (due to Alfvén wave heating) In the slow wind Electrons are hot (due to heat conduction) Ions are cold (no Alfvén wave heating) Radial velocity Ion temperatureElectron temperature

New Low Corona Model EIT 171ÅEIT 195ÅEIT 284ÅSXT AlMg Old SC model synthesis New LC model synthesis Observation: Aug 27, 1997 CR1913 Downs et al., ApJ, 712, 1219, 2010

Halloween Storm Simulation: Early Phase

Alfvénic Turbulence and SEP Acceleration Recent Hinode observations imply that MHD turbulence might be the source that powers, accelerates and heats the solar wind Today it is possible to use many dimensional models (>3D) We developed a coupled solar wind- turbulence-SEP model Separate transport equation is solved for turbulence Wave stress tensor and energy appear in the solar wind momentum and energy equations Energetic particles are accelerated by interaction with turbulence t=7 m t=80 m t=137 m Sokolov et al., 2009

FTE Simulation (Kuznetsova) Resolution at the dayside magnetoapuse is 1/16 R E (400 km) Total number of cells in the simulation is 20 million Solar wind parameters: n = 2 cm –3 T = 20,000 K Vx = 300 km/s B = 5 nT IMF Turning From Northward (θ = 0 ∘ ) to IMF clock angle θ < 120 ∘

Theta = 120 o Pressure Subsolar FTE Structure Cluster

View from South

Component Reconnection Near Sub-Solar Region (Y ~ 0 – 2 R E ) And Anti-Parallel Reconnection at Flanks (Y ~ R E ) View from the Sun

Z-slice Y-slice X-slice FTE at flanks (Y = 12 Re) is connected to FTE at the subsolar region ( Y = 0 ).