PNNL

Abstract

Parameterizing convective microphysics in Earth system models (ESMs) is a quickly evolving field compared to the relatively well-established stratiform counterpart. Highly variable levels of sophistication for representing the interaction of ESM convection schemes with cloud processes compounds complexity of both interpretation and evaluation. Improving upon current implementations requires evaluating ESMs with well-constrained cases of convection across modes including shallow cumulus, cumulus congestus, and deep convection. Here, we propose a benchmark large eddy simulation (LES) cumulus congestus case study from the aircraft-based NASA Cloud, Aerosol, and Monsoon Processes Philippines Experiment (CAMP2Ex) and demonstrate its observational constraints using novel polarimetric retrievals. The airborne Research Scanning Polarimeter (RSP) measures the polarized reflectance at cloud top, for which the rainbow structure in the polarized reflectance allows an accurate characterization of cloud droplet sizes (effective radius, Reff), geometric cloud boundaries, and derivation of cloud droplet number concentration (Nd) at cloud top. Herein, a congestus case study is constructed with input from observed aerosol distributions, large-scale thermodynamic and vertical motion conditions harvested from a mesoscale simulation, and evaluated against polarimetric retrievals and in situ cloud microphysics measurements in simulations using both bulk and bin microphysics schemes. Results show that the control bulk simulation reasonably reproduces characteristic profiles of cloud-top Nd that decrease with altitude, while the bin simulation realizes greater discrepancies due to weaker precipitation formation. Sensitivity tests indicate that the Nd profile is strongly sensitive to collision-coalescence and the vertically resolved aerosol particle size distribution. While Nd appears to be well-constrained, a persistent low-bias in cloud-top Reff is evident in both the bulk and bin schemes. Comparison of simulated and in situ droplet size distributions (DSDs) indicate that this low-bias in Reff originates from a cloud droplet mode that is too narrow relative to observations, whereby Reff is largely sensitive to the DSD structure within the 10-30 ?m size range. Finally, a thermal-tracking framework is employed to demonstrate that the dilution of Nd throughout a thermal's lifetime is heavily determined by an active collision-coalescence process and the height-varying aerosol distribution, and that in the absence of these two factors, the impact of entrainment on diluting Nd is largely offset by continuous aerosol activation within the thermal. Implications for developing warm-phase convective microphysics schemes for ESMs and expansion of the presented evaluation method to global, space-based polarimetry platforms are discussed. We propose that the presented case study offers a viable benchmark LES case study for cumulus congestus evaluation using single column model versions of parent ESMs.