An Evaluation of Cirrus Cloud Thinning through Improved Integration of Satellite Retrievals and Climate Modeling

Of the three most recognized aerosol approaches to climate engineering, cirrus cloud thinning or CCT is arguably the most difficult to evaluate using a climate model due in part to the sub-grid scale processes and mechanisms relevant to ice nucleation. To avoid this complication, a new satellite retrieval for cirrus clouds has been developed (Mitchell et al., 2018, ACP) that retrieves the cirrus cloud ice particle number concentration N, effective diameter D_e and ice water content IWC for optical depths between 0.3 and 3.0. Broadly speaking, cirrus clouds form through either homo- or heterogeneous ice nucleation (henceforth hom and het), where hom (het) is characterized by relatively high (low) N. Retrieval results suggest het is dominant during summer over oceans at mid-latitudes. This “het” condition was applied globally to represent cirrus formed via het using the Whole Atmosphere Community Climate Model version 6 (WACCM6), replacing the cirrus D_e supplied by the Morrison-Gettelman cloud microphysics scheme (MG2) with the D_e-temperature relationship derived from this het condition. The observed global cirrus cloud condition (based on the CALIPSO satellite), which was related to temperature (T), latitude, season and surface type (land vs. ocean), was represented in WACCM6 by 48 D_e-T relationships. Two 40-year WACCM6 simulations are being conducted, with one based on the het condition (henceforth the Het simulation) and one based on the observed condition (henceforth the CALIPSO simulation).

The MG2 code was revised so that predicted ice particle size distributions (PSD) in cirrus clouds conform with CALIPSO D_e values as a function of latitude, season and surface type. New relationships were developed between effective diameter D_e and PSD mean size D_mean, along with new ice fall speed relationships that are number- and mass-weighted (functions of D_mean and D_e, respectively). These are used to insure that the MG2 scheme is internally consistent with our CALIPSO D_e retrievals that partly determine the cirrus PSD in our modified MG2 scheme. This modified MG2 code is freely available to other investigators that have interest in interfacing satellite D_e retrievals with the cloud components of climate models.

Differences between the cloud radiative effect (CRE) in the CALIPSO and Het simulations will provide an estimate of the CRE contribution from hom-produced cirrus clouds. Conversely, this will also provide an estimate of the CRE from CCT, assuming that all hom cirrus are converted to het cirrus clouds. Since the CALIPSO retrievals of N and D_e indicate that hom cirrus clouds are most common at high latitudes and during winter at mid-latitudes, the CCT CRE will be strongest at high latitudes and during winter at mid-latitudes.