An Evaluation of the Representation of Tropical Tropopause Cirrus in the CESM/CARMA Model Using Satellite and Aircraft Observations

Observations from the third campaign of the National Aeronautics and Space Administration Airborne Tropical Tropopause Experiment (ATTREX 3) field mission and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations satellite mission are used to evaluate simulations of tropical tropopause layer (TTL) cirrus clouds in the Community Earth System Model's (CESM) Community Atmosphere Model, CAM5. In this study, CAM5 is coupled with a sectional ice cloud model, the Community Aerosol and Radiation Model for Atmospheres (CARMA). We find that both model variants underrepresent cloud frequency along the ATTREX 3 flight path and both poorly represent relative humidity in the TTL. Furthermore, simulated in-cloud ice size distributions contained erroneous amounts of ice crystals throughout the distribution. In response, we present a modified ice cloud fraction scheme that boosts the cloud fraction within the TTL. Due to coarse vertical model resolution in the TTL, we also prescribe a 2-K decrease in cold point tropopause temperatures to better align with observed temperatures. Our modifications improve both CAM5 and CAM5/CARMA's in-cloud ice size and mass distributions. However, only CAM5/CARMA has a significant improvement in cloud frequency and relative humidity. An investigation of cloud extinction in the ATTREX 3 region found that each model variant struggles to reproduce observed extinctions. As a first-order approximation, we introduce randomly generated temperature perturbations to simulate the effect of gravity waves into the CAM5/CARMA simulation. These gravity waves significantly increase the incidence of low extinction (<0.02 km(-1)) values, ice cloud fraction between 16 and 18 km, and ice crystal smaller than 100-mu m concentrations but provided only small changes to high extinction values. Plain Language Summary We use observations from aircraft and satellite missions to evaluate the representation of upper tropical tropospheric ice clouds in a widely used global climate model. The goal of this study is to produce a more realistic representation of ice clouds within a climatologically influential region. Such model improvements will help us better predict future climates. We find that the model used in this study fails to properly represent the ice clouds along the aircraft flight path and is too dry. In response, we introduce modifications that boost the cloud coverage in the upper troposphere. Our modifications improve the frequency of clouds along the aircraft flight path, as well as the simulated number and mass of ice crystals inside ice clouds. The modifications also moisten the model. We also find that simulations are unable to reproduce proper cloud extinctions. As a first step to address this problem, we introduce a set of randomly generated temperature fluctuations into the simulated upper troposphere. These temperature fluctuations improve instances of transparent clouds, but they fail to improve the amount of high extinction clouds. Future work should focus on better temperature fluctuation parameterizations and better cloud parameterizations to address these model shortcomings.