Application of MHD numerical modelling in space weather : Accuracy issues and condition
Application of MHD numerical modelling in space weather : Accuracy issues and condition
Monday, 10 July 2017: 14:40
Furong Room (Cynn Hotel)
Abstract:
Time-dependent simulations in magnetohydrodynamic numerical models are often used to study space weather phenomena of relatively long time period and large size space, which include many examples from the solar origin to the Earth impact in the heliosphere.
In this study, we firstly introduce a class of exact analytic solutions of MHD when the boundary is driven by certain impulsive impacts.
By comparing MHD simulations with this analytical solution, which allows arbitrary source motion, it is shown that our exact solutions are consistent with numerical solutions when both spatial and temporal resolution gradually increase in the numerical model. When we apply the MHD model to the Sun-Earth space weather issues such as CME propagation, it is found that numerical errors in such long distance simulations are unavoidable, and full attention to the model parameters is required to maintain simulations with a certain level of accuracy. Our MHD simulation tests for CMEs, based on the grid resolution of N=100~100,000 over the interval of 1AU, explicitly show how there occur significant computational errors such as up to 5-15 hrs in arrival timing at the Earth, in the model of the similar resolution used in current operation models (e.g., N=128~512 over 1AU). It is suggested how we can avoid serious errors by optimizing the model parameters in advance.
In this study, we firstly introduce a class of exact analytic solutions of MHD when the boundary is driven by certain impulsive impacts.
By comparing MHD simulations with this analytical solution, which allows arbitrary source motion, it is shown that our exact solutions are consistent with numerical solutions when both spatial and temporal resolution gradually increase in the numerical model. When we apply the MHD model to the Sun-Earth space weather issues such as CME propagation, it is found that numerical errors in such long distance simulations are unavoidable, and full attention to the model parameters is required to maintain simulations with a certain level of accuracy. Our MHD simulation tests for CMEs, based on the grid resolution of N=100~100,000 over the interval of 1AU, explicitly show how there occur significant computational errors such as up to 5-15 hrs in arrival timing at the Earth, in the model of the similar resolution used in current operation models (e.g., N=128~512 over 1AU). It is suggested how we can avoid serious errors by optimizing the model parameters in advance.