Authors: Gang Li (University of Alabama in Huntsville), Meng Jin (SETI), Zheyi Ding (KU LEUVEN)

In forecasting SEP events, confidence level is an essential aspect.  However, it did not receive as much attention as the forecasting itself. To some extend, this is because that in modelling SEP events,  uncertainty of the modelling results was hard to assess or in some cases, it was not considered. Nevertheless,  uncertainty can be as important as the modelling results (e.g. fluence or time profile of SEP).  There are various sources of uncertainty in SEP modelling.   A few of these include, uncertainty in the background solar wind, uncertainty in the background IMF field (which affects the magnetic connection) and the MHD turbulence level, uncertainty of the shock properties (shock geometry and compression ratio),  uncertainty in the seed population,  and uncertainty in the diffusion coefficient (especially the cross field diffusion coefficient), etc.  We discuss the effects of some of these uncertainties and argue that using an ensemble approach in SEP modelling can provide a natural procedure in quantifying uncertainty of SEP modelling, and provide a means of quantifying the confidence level of SEP forecast.