HAO Colloquium - Binzheng Zhang, University of Hong Kong

Doing more with Moore: Harnessing next-generation supercomputers to move beyond global magnetospheric modeling in the fluid limit

It has been more than three decades since the first global MHD model of the Earth’s magnetosphere was develop. As a backbone of global geospace models, magnetospheric models now couple with various non-ideal MHD physics codes such as ionosphere-thermosphere codes (e.g., NCAR-TIEGCM), inner magnetospheric codes (e.g., RCM, CRCM), ionospheric outflow codes (e.g., IPWM, PWOM) and particle codes (e.g., CHIMP, ePIC), etc. The one used during the CISM era – the Lyon-Fedder-Mobarry (LFM) global magnetosphere model has been transitioned towards a new code base named Gamera (Grid Agnostic MHD for Extended Research Applications), with significant upgrades in both numerical schemes and software implementations. Recent development of the Gamera magnetospheric codes (Gameraspheres) enables resolutions approaching ion-scale especially in the magnetotail. In this talk, I will show preliminary results from comparisons of Gamerasphere simulations of the Earth’s magnetosphere-ionosphere system, with three different grid resolutions (median, high res) including the highest one approaching ion scale (extra-high res). Preliminary results indicate that although grid convergence is not yet (or may never) achieved in magnetospheric problems, high- and extra-high resolution simulations does provide results in approximately the same regime, e.g., in both magnetotail dynamics and auroral precipitation, suggesting that the high-res simulation model may be suitable for most global magnetospheric physics research and space weather applications. With the development of next-generation supercomputers, these high- and extra-high resolution Gamerasphere simulations may provide much better descriptions of magnetospheric dynamics, which improves the modeling of global geospace dynamics significantly through the coupling to the ionosphere-thermosphere-atmosphere modules.

Date and time: 
Wednesday, July 24, 2019 - 2:00pm to 3:00pm