Abstract The cloudâbase mass flux closure is a critical component of convective parameterizations and usually consists of scalings for the cloud(âcore) area fraction and vertical velocity. Here we evaluate if observations from the EUREC4A campaign support ideas behind two common closures. All closure parameters are diagnosed at the mesoscale from dropsonde data and turbulence measurements. The closure models are compared to the observed mass flux estimated as a residual of the subâcloud layer mass budget from the same dropsonde arrays. Both closures capture the magnitude of the reference mass flux. However, the closure based on the surfaceâbased convective velocity scale strongly underestimates mass flux variability, whereas the closure using a convective inhibition based area fraction and turbulence kinetic energy (TKE) based vertical velocity scale strongly overestimates it. TKE alone explains nearly 80% of mass flux variability, suggesting that TKE aggregates information of area fraction and vertical velocity of cloudâbase updrafts.