This study illustrates how aerosol affects the top of atmosphere (TOA)radiation flux in the NUIST model,using a hypothetical anthropogenic aerosol forcing determined by Coupled Model Intercomparison Project Phase 6.There are three ways by which anthropogenic aerosol affects TOA radiation flux.First,anthropogenic aerosol affects TOA radiation flux through the changes in aerosol optical properties,known as the aerosol radiative effect.This effect mainly depends on anthropogenic aerosol radiative forcing,and is not related to the Twomey effect.Second,anthropogenic aerosol affects TOA radiation flux through the changes in cloud optical properties,referred to as aerosol effects on cloud radiative forcing.These effects include the Twomey effect and semi-direct effect.Finally,anthropogenic aerosol affects TOA radiation flux via the changes in other optical factors.Compared with the first two ways,this third one is negligible.This study focuses on the second way,and discusses the methods used for estimating anthropogenic aerosol effects.In order to distinguish contributions to aerosol effects on cloud radiative forcing from semi-direct and indirect effects,we added two experiments:Rad and Tmy.The Rad experiment only considers the anthropogenic aerosol optical properties,which the Tmy experiment only considers the anthropogenic aerosol Twomey effect.Including the experiments used in Part 1,there are four experiments in total:Base,Rad,Tmy and Both.The results of the Rad and Base experiments show that anthropogenic aerosol semi-direct effect is estimated at 0.21 W·m^-2.Semi-direct effect could also be obtained as the difference between the aerosol effects on cloud radiative forcing from Both and Base experiments (i.e.semi-direct effect and Twomey effect)and the aerosol effects on cloud radiative forcing from Tmy and Base experiments (i.e.Twomey effect).This semi-direct effect is 0.09 W·m^-2,significantly less than those from the Bad and Base experiments.The difference between the Tmy and Base experiments shows that the effective radiative forcing (ERF)caused by anthropogenic aerosol Twomey effect is -0.22 W·m^-2.Meanwhile,the difference between the Both and Rad experiments shows that the ERF caused by the Twomey effect is -0.30 W·m^-2,which is stronger (more negative)than that caused by the difference between the Tmy and Base experiments.It is clear that the assessment results depend on experimental design.One likely reason for this is that aerosol semi-direct and indirect effects interact in a very complex manner,and the assumption of linear superposition is not reasonable.It is worth noting that the diagnosed Twomey effect not only includes the Twomey effect itself,but also cloud-relevant adjustments caused by the Twomey effect.Aerosol could make an obvious impact on quantifying cloud forcing,as aerosol also can scatter and absorb solar radiation.The short wave cloud forcing (SWCF)from Base experiments is -47.54 W·m^-2.Under no-aerosol conditions,the SWCF* (i.e.the SWCF calculated as a diagnostic with aerosol scattering and absorption neglected)from the Base experiment is -49.51 W·m^-2.The difference caused by the impact of aerosol is 1.98 W·m^-2.Compared with the Base experiment,this difference from the Rad experiment is increased by 0.38 W·m^-2,because anthropogenic aerosol is considered.In other words,for the same cloud optical properties,the SWCF diagnosed with aerosol optical properties may be shown to increase by~0.38 W·m^-2 after considering anthropogenic aerosol optical properties.Due to this obvious positive bias,the aerosol effects on cloud radiative forcing must be estimated from the change in SWCF* (diagnosed with aerosol scattering and absorption neglected),rather than SWCF (diagnosed using the traditional method).