iPATH

The iPATH model is a 2-D MHD SEP model that simulates diffusive shock acceleration (DSA) at CME-driven shocks and follows the subsequent transport of energetic particles through the inner heliosphere. iPATH models the background solar wind and CME-driven shocks at the ecliptic plane starting at 0.05 AU using the 2D MHD ZEUS code (http://www.ap.smu.ca/~dclarke/zeus3d/version3.6/home.html) and produces time profiles of SEP intensity spectra and pitch angle distributions as outputs at selected vantage points (e.g., at Earth or Mars). It considers both perpendicular and parallel diffusive factors of energetic particles, which come from nonlinear guiding center theory (NLGC) and quasi-linear theory (QLT) respectively. The transport module is a Monte Carlo code which follows test particles through space described by the focused transport equation (FTE) and is set up for parallel computations. This model is improved over the original PATH model which was 1-D.

iPATH first creates the shock perturbation along the inner boundary, and propagates the CME outward with the forward shock region tracked with a 2-D onion-shell model. For each time step, a new outer shell is created based on the shock speed and all previous shells convect and adiabatically expand with the solar wind. Then accelerated particle distributions are calculated along the whole shock front, based on the diffusive shock acceleration. Accelerated particles are then allowed to diffuse back to the shock complex, and between each parcel behind the shock via parallel and perpendicular diffusion. This gives the distribution function in each shock parcel at each time step, which is important for the ESP phase when the shock arrives at the observer. Once a particle has moved a certain distance during a single time step, it escapes the shock and is transported through the unperturbed solar wind via a focused transport scheme, which includes terms for weak scattering and cross-field diffusion from the random walk of magnetic field lines. The FTE is solved using a backward stochastic differential equation approach until a steady state is found where ensemble averages of many test particle paths give the full particle distribution function.

A complete description of most of the algorithms used in ZEUS-3D and many of its tests may be found in Clarke (1996, doi:10.1086/176730) and Clarke (2010, doi:10.1088/0067-0049/187/1/119).

Like other models using CME information as inputs, iPATH’s performance as a prediction model relies on how early we can acquire good quality CME parameters from observations. As the inner boundary is set at 0.05 AU (10 Rsun) and the shock is formed even further in, it cannot capture the SEP acceleration happening in the low corona and does not produce output for the first couple of hours of the events. iPATH is not able to fully capture the effects of previous CMEs without modifications on the model so the operational version is best suited for single CME events. The background solar wind used in iPATH is a simple homogeneous Parker spiral model, which cannot capture some complex solar wind geometries in a specific event. But this simplification is efficient for general prediction purposes.

Continuous/RT runs (nowcast module):

• A background solar wind is simulated every 8 hours, with initial conditions (solar wind speed and density, magnetic field intensity, turbulence level, and seed population injection rate) derived from ACE and/or DSCOVR real-time plasma data averaged over the preceding 8 hours.
• CME info are downloaded from DONKI every 15 minutes: any new/updated CME will trigger a CME simulation followed by an SEP transport simulation to Earth, Mars, Venus, STEREO-A, Parker Solar Probe, and Bepi Colombo.
• The CME and SEP simulations both lasts 72 hours of in-simulation time.
• Since the CME simulation is 2D, the 3D speed from DONKI is projected on the solar equatorial plane.
• The CME shock finding algorithm is optimized to work also for slow CMEs.
• ESPs are not simulated when the shock arrives at an observer.

• Background solar wind parameters based on 1 AU observations (solar wind density, speed, temperature, and magnetic field strength)
• CME parameters (CME speed, width, location, and perturbation duration)
• Turbulence parameters (turbulence level, turbulence spectral parameters, and radial dependencies of these parameters, for advanced users)
• Suprathermal seed particle input (suprathermal particle energy spectral parameters)
• Observer locations (radius, longitude)

• Time profiles for energetic proton/heavy ion intensities (differential fluxes) at a wide energy range (from hundreds of keVs to GeVs) at the chosen observer locations
• Event-integrated fluences (in MeV-1cm-2)
• Time profiles for integral flux above certain energies (in pfu)
• Figures for flux/fluence results and CME/shock configurations.

Initial release: version 2.1.0 was deployed as Continuous/RT Run on December 22, 2023.

Version 2.2.0 was deployed as Continuous/RT Run on March 17, 2024, with the following updates:

• CME simulations try to use the previous background if current one is missing/failed.
• Added more checks on input data for solar wind background simulation.

Version 2.2.1 was deployed as Continuous/RT Run on June 25, 2024, with the following updates:

• Replaced PSP with BepiColombo.
• Added support for LE (leading edge) vs SH (shock) CME feature codes in CME checking (skip non-LE and non-SH CMEs) and trigger block in JSON files for the SEP scoreboard. These are features used by M2M when tracking the CME for measuring speed and location.

Version 2.2.2 was deployed as Continuous/RT Run on August 2, 2024, with the following updates:

• Added default solar wind plasma values in case of missing input DONKI data.
• Use 1.5 protons/cc as minimum solar wind density and 275 km/s as minimum solar wind speed.