Predictive Science Inc.
predsci.bsky.social
Predictive Science Inc.
@predsci.bsky.social
PSI is an employee-owned research company delivering predictive models in space weather, solar physics, and communicable diseases; powered by globally recognized scientists, advanced magnetohydrodynamic models, and high-performance computing.
This week, we're showcasing another component of our in-house -- but open source -- suite of modeling tools...the Solar Wind Generator (SWiG)! If you need to rapidly generate WSA-like solar wind outputs, SWiG is a must-have.
github.com/predsci/swig
February 12, 2026 at 9:49 PM
Introducing spherical harmonics to analyze the dominant spatial scales of the magnetic field in each run allowed Caroline to quantify the structure and draw concrete, compelling correlations. This work is part of Caroline's PhD thesis work; more of her fascinating is coming soon!
January 30, 2026 at 9:16 PM
The team also showed that a highly structured surface magnetic field meant that more heat was injected into the corona overall. This is related to the total unsigned flux, as well as changes in flux-tube expansion that attenuates the waves that reflect within coronal loops.
January 30, 2026 at 9:16 PM
By using a range of resolutions for the input magnetic field, the team found that fine-scale structure in the photosphere led to much more structured heating and connectivity in the middle corona (a region where many physical transitions occur, critical for solar wind formation).
January 30, 2026 at 9:16 PM
In @caroline-evans, Cooper Downs, and Don Schmit's new #ApJ paper, a series of simulations at increasing resolution uncover the correlation between magnetic complexity and coronal heating. The MAS model acted as a lab to probe difficult-to-observe scales.
buff.ly/ls2p29Z
January 30, 2026 at 9:16 PM
🚨📄 New paper alert! 📄🚨
In their new paper published in Space Weather, PSI's Pete Riley and Michal Ben-Nun find that the chances of an extreme #spaceweather event depends strongly on the strength of the #solar cycle.

What does this mean? Read on below!
agupubs.onlinelibrary.wiley.com/doi/full/10....
January 21, 2026 at 8:24 PM
Using #MPI to run from 1-1000s of CPU cores, it can also run on 1+ #NVIDIA/#Intel #GPUs for fast local calculations.
GitHub repository examples help you get started.
For easy PFSS+CS solutions try running our SWiG pkg that uses POT3D.
Information on SWiG coming soon - stay tuned!
January 14, 2026 at 9:42 PM
Need to calculate potential field solutions of the solar corona, really fast, for grids up to billions of cells (doi.org/10.3847/1538...), or on nonlinear grids (to refine the #solution in ROIs)?
Check out PSI's POT3D code at github.com/predsci/pot3d!
#github #modeling #computing
January 14, 2026 at 9:42 PM
Ryder Davidson's poster "MHDweb: An Interactive Visualization Toolkit" showcases PSI's #data #visualization suite whose web-based interface analyzes our Coronal and Heliospheric MAS model solutions and compares the results with in-situ #spacecraft data: buff.ly/2or5UXJ
January 7, 2026 at 9:05 PM
Emily Mason's talk "Statistical Study of Long-Lived Active Regions" details a newly-identified pool of active regions that last for several solar rotations, and which produce roughly 4-7 times more significant flares than their counterparts: buff.ly/99S2e51
January 7, 2026 at 9:05 PM
Jon Linker's talk "Dynamics of the Time-Dependent Solar Corona and Inner Heliosphere" reviews our ground-breaking near-real-time, data-driven eclipse simulation and compares its results with steady-state MHD and PFSS models using the same boundary data: buff.ly/99S2e51
January 7, 2026 at 9:05 PM
Cooper Downs' poster, "The Unique Prospects of High-Latitude Observations:A Modeling Perspective" illustrates the ways in which remote-sensing and in situ instruments far from the ecliptic can improve our theoretical understanding of the Sun-Earth system. buff.ly/z8Ve24Q
January 7, 2026 at 9:05 PM