Abstract A computationally efficient semi-Lagrangian advection (SLA) scheme designed for microphysical variable advection was implemented into the Weather Research and Forecasting (WRF) Model with spectral bin microphysics (SBM). The primary goal was to reduce the CPU time for advection of the scalar SBM variables and demonstrate its applicability to simulation of a real-data case study in the eta vertical coordinate system. A mesoscale convective system (MCS) in Midlatitude Continental Convective Clouds Experiment (MC3E-0520) was selected for simulations. We compared the SLA and high-order, nonlinear monotonic advection schemes and tested the sensitivity of the simulated radar reflectivity, microphysical, and dynamic properties of the MCS to the choice of microphysical schemes, aerosol concentration, and grid spacing. Simulations using the SLA and monotonic advection schemes were similar, and the differences between them were much smaller than those between the SBM and bulk microphysical schemes tested. Improvement of the grid resolution had a larger impact on the vertical velocity field than did the choice of aerosol concentration. The total computational time in simulations with SLA was about 25% shorter than that with monotonic advection, which resulted from a reduction of more than 50% in computational time required for advection of the microphysical variables. Significance Statement Clouds comprise particles of various sizes and types: aerosols, liquid droplets, ice crystals, hail, and snow. Two major approaches describe the time and spatial evolution of these particle species: solving basic microphysical (bin microphysics) and simplified parameterization (bulk microphysics) equations. Bin microphysics is time-consuming as it operates with size distributions described by multiple mass bins. Among the microphysical processes, the advection of the microphysical variables requires significant CPU time. We tested a new numerical advection scheme that requires several times less computer time than the standard high-order nonlinear schemes being of similar accuracy. Due to the introduced change, the advection of droplet size distributions became less of a factor affecting the total computer time of bin-type schemes. This is an important step toward making the bin approach computationally efficient without losing accuracy. It is shown that the sensitivity of the results to spatial grid spacing and to the choice of microphysical schemes is significant and may be larger than that to variations in the aerosol concentration.