Low‐Level Marine Tropical Clouds in Six CMIP6 Models Are Too Few, Too Bright but Also Too Compact and Too Homogeneous

Abstract

Several studies have shown that most climate models underestimate cloud cover and overestimate cloud reflectivity, particularly for the tropical low‐level clouds. Here, we analyze the characteristics of low‐level tropical marine clouds simulated by six climate models, which provided COSP output within the CMIP6 project. CALIPSO lidar observations and PARASOL mono‐directional reflectance are used for model evaluation. It is found that the “too few, too bright” bias is still present for these models. The reflectance is particularly overestimated when cloud cover is low. Models do not simulate any optically thin clouds. They fail to reproduce the increasing cloud optical depth with increasing lower tropospheric stability as observed. These results suggest that most models do not sufficiently account for the effect of the small‐scale spatial heterogeneity in cloud properties or the variety of cloud types at the grid scale that is observed.

The Role of Mesoscale Cloud Morphology in the Shortwave Cloud Feedback

“…Consideration of changes in morphology occurrence under climate change may be helpful in predicting shortwave cloud feedback. Current models appear to poorly capture cloud heterogeneity and associated radiative effect (Konsta et al, 2022). The geographical pattern of the morphology feedback (Figure 3b-d) contributes regions of positive and negative feedback that may be useful to consider in understanding patterns of radiative feedback.…”

“…State-of-the-art GCMs from phase 6 of the Coupled Model Intercomparison Project (CMIP6) do not capture the radiative properties of low clouds largely due to poorly representing cloud heterogeneity. GCMs' inability to simulate optically-thin cloud features at lower CF is thought to be a contributor to this issue (Konsta et al, 2022). Opticallythin features are observed across mesoscale cloud morphologies (Leahy et al, 2012;Wood et al, 2018;Mieslinger et al, 2021) and are likely associated with precipitation processes during cloud morphology development and transition (O, Wood, & Tseng, 2018).…”

“…Opticallythin features are observed across mesoscale cloud morphologies (Leahy et al, 2012;Wood et al, 2018;Mieslinger et al, 2021) and are likely associated with precipitation processes during cloud morphology development and transition (O, Wood, & Tseng, 2018). In addition to the so-called "too few, too bright" bias (Nam et al, 2012;Engstrom et al, 2015a;Bender et al, 2017;Konsta et al, 2022), representation of morphology and generation of optically-thin features may also effect GCM biases in cyclone cold sectors (Bodas-Salcedo et al, 2014;Williams & Bodas-Salcedo, 2017) and simulated mean-state SST (e.g., coastal gradients, regional seasonal cycles) (Farneti et al, 2022;Hyder et al, 2018;Wang et al, 2022). These diagnosed model biases suggest that consideration of mesoscale cloud morphology will assist in improving mean-state cloud radiative properties and their subsequent environmental impacts in GCMs.…”

Improving stratocumulus cloud amounts in a 200-m resolution multi-scale modeling framework through 2 tuning of its interior physics

“…This section focuses on the Peruvian Sc region to examine further details of the unsteady evolution of boundary layer vertical structure and the associated changes in low clouds. This region, lying off the west coast of South America over the ocean (gray area in Figure 3b), is one of the most persistent Sc decks (Bretherton & Wyant, 1997) and poorly simulated in models (Konsta et al, 2022). strong turbulent mixing occurring during local nighttime, as expected (Hignett, 1991).…”

“…Sc also has a strong influence on the heat and moisture exchange between the troposphere and boundary layer (Randall et al, 1984). Despite the climatic significance of Sc, climate models do not agree on their seasonal cycle, spatial extent, radiative properties, and cloud feedbacks (Bony et al, 2011;Lin et al, 2014;Gettelman & Sherwood, 2016;Brunke et al, 2019;Vignesh et al, 2020;Tselioudis et al, 2021;Konsta et al, 2022;Zelinka et al, 2022), and even high resolution models simulate widely varying cloud properties in idealized case studies (Ackerman et al, 2009;Stevens et al, 2005;Bretherton et al, 1999).…”