Quantifying Cirrus Cloud Radiative Effects in Numerical Models by Incorporating Satellite Observations

The complex mechanisms governing the formation of cirrus clouds pose significant challenges in the accurate simulation of cirrus clouds within regional and global models, leading to uncertainties in predicting cirrus cloud response to aerosols and the efficacy of cirrus cloud thinning (CCT). CCT is characterized by a transition from homogeneous (hom) to heterogeneous (het) cirrus clouds, with significant implications for the cloud radiative effect (CRE). Key sources of uncertainty regarding aerosol-cloud interaction (ACI) and CCT include the relative contributions of hom vs. het ice nucleations, the "pre-existing ice" assumption in models, and the treatment of orographic gravity waves (OGW) in ice microphysics parameterizations. Recent advancements conducted by our group, leveraging global-scale Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) measurements, revealed that hom process plays a more substantial role in cirrus formation than traditionally assumed. This finding suggests that the efficiency of CCT might surpass previous estimates, as hom nucleation tends to produce optically thicker clouds, thus leading to a more pronounced CRE change.

In the first phase of our study, we use libRadtran, a 1-dimensional radiative transfer model (RTM), to simulate cirrus CRE at the top of the atmosphere (TOA), Earth’s surface (Sfc.), and within the atmosphere (Atm.). The RTM incorporates vertical profiles of ice water content (IWC) and effective diameter (D_e) from CALIPSO retrievals for both hom and het cirrus clouds, focusing on winter conditions across various geographical regions (Arctic, Antarctic, midlatitudes) and surface types (ocean and land). Over the Arctic during boreal winter, the change in Sfc. CRE between hom and het (ΔCRE), which is purely from terrestrial radiation, is approximately -10 W m^-2. Based on CALIPSO results, the frequency of hom cirrus is 0.2 for this time and region. Assuming that the total cirrus cloud fraction is 0.3, the estimated ΔCRE is -0.6 W m^-2 which implies a CCT cooling potential at the surface. Similar cooling effects are estimated over the Antarctic during the austral winter, whereas midlatitude regions show negligible surface ΔCRE due to compensating solar and terrestrial radiation effects. However, the Atm. ΔCRE in midlatitudes is about -1 W m^-2. Sensitivity tests reveal the significant influence of atmospheric thermodynamics and low-level liquid clouds on CRE magnitude. The RTM modeling provides instantaneous CREs, but it does not account for climatic effects from interactions between components of the Earth system.

In the second phase of our work, we incorporate the new CALIPSO retrievals into a global model, known as the Community Atmosphere Model, Version 6 (CAM6), through a simple yet prognostic method to address the first source of uncertainty. In addition, we deactivate the “pre-existing ice” treatment, which often results in unrealistically low ice particle concentrations due to the use of mean-layer IWC rather than the much lower IWC near the cloud top where ice nucleation occurs. This deactivation not only addresses the second source of uncertainty but also mitigates the third source, e.g., the absence of OGW for cirrus formation in CAM6, which would otherwise enhance the vertical motions necessary for achieving the supersaturations required for hom. In addition, we parameterize the CALIPSO-retrieved D_e from retrieved IWC across various temperature ranges, seasons, latitude bands, and surface types to refine the simulation of terminal ice fall velocity, cloud lifetime, and cloud cover under two different scenarios: “het” and “natural” (combination of het and hom). This ongoing work aims to provide a comprehensive assessment of ACI and CCT impacts on the atmospheric system, including potential influences on climate dynamics.