The role of the host climate model in quantifying the contrail cirrus climate impact

Aviation currently makes a 3.5% contribution to the anthropogenic effective radiative forcing of climate. The largest component of this forcing comes from contrail cirrus, estimated to be 2 times larger than the contribution from aviation CO2 emissions. However, there is still a large uncertainty (i.e. ~70%) in the contrail cirrus effective radiative forcing (ERF) estimates according to the latest aviation climate impact assessment. Here we implement the existing contrail parameterisation developed for the Community Atmosphere Model version 6 (CAM6) in the atmospheric component of the UK Earth System Model (UKESM), i.e. the Unified Model (UM). By analysing and comparing the results from both models, we are able to isolate and investigate for the first time how key features of the host climate model is affecting our ability to accurately quantify the contrail climate impacts. We show that differences in background humidity (particularly ice supersaturation) in the two climate models lead to substantial differences in the simulated contrail fractions, with UM values being 2-3 times larger than those from CAM6. We also find contrasting responses in overall global cloud cover due to air traffic, with contrails causing increases and decreases in total cloud fraction in the UM and in CAM6, respectively. The different complexity of the two models cloud schemes (i.e. single and double moment cloud scheme in UM and CAM6, respectively) results in substantial differences in the simulated contrail-driven changes in cloud ice water content. However, if we account for this difference in cloud scheme complexity by scaling the simulated UM contrail cirrus optical depth to match existing estimates, the contrail cirrus ERFs simulated by the two models are comparable. To conclude, the large dependence of the simulated contrail cirrus climate impact on the host climate model highlights the need for improved evaluations of the key model microphysical and radiative processes.