Implementation and Evaluation of Physics‐Driven Dynamic Entrainment‐Mixing Parameterization in a Climate Model and Its Impact on Low‐Cloud Simulation

Abstract The turbulent entrainment‐mixing process in the Community Earth System Model version 1.2 (CESM1.2) is assumed to follow the extremely inhomogeneous entrainment‐mixing. However, different entrainment‐mixing scenarios can occur in real clouds. To address this deficiency, a unifying parameterization that represents different entrainment‐mixing processes is implemented and evaluated in CESM1.2. The results indicate that the homogeneous mixing degree values simulated by the new parameterization in CESM1.2 are predominantly greater than 50%, suggesting a tendency toward homogeneous mixing. Compared to the extremely inhomogeneous mixing mechanism, the new parameterization increases the cloud droplet number concentration ( N c ). More importantly, the new parameterization improves low‐cloud fraction (CLDLOW) simulation in Northwest Pacific (NWP) and Southeast Pacific (SEP) regions, with relative improvements of 2.95% and 4.17%, respectively. Furthermore, the improvements reach up to 44.6% and 16.2% in the NWP and SEP regions, respectively, when considering the relationship between N c and CLDLOW. Further analysis reveals that the new parameterization enhances cloud optical depth, longwave radiative cooling effect, net condensation rate, cloud water mixing ratio, lower‐troposphere stability, and CLDLOW by increasing N c . These results underscore the importance of improving entrainment‐mixing parameterization in climate models.