The Sensitivity of Contrail Radiative Forcing to the Radiative Transfer Parameterization

Contrail cirrus, ice clouds produced by aircraft passing through cold, supersaturated air, may account for over half of radiative forcing caused by aviation. Because shortwave and longwave radiative effects of upper-tropospheric ice clouds are opposite in sign, the net contrail radiative forcing remains poorly constrained with an uncertainty of 70 %. Previously, an uncertainty of 35 % was attributed to the choice of radiative transfer code based on a model intercomparison. In this work, we reassess the diversity in contrail instantaneous radiative forcing in a modern set of radiative transfer codes using both idealized calculations and a realistic contrail outbreak. We find that forcing estimates are most sensitive to the ice optical property parameterization, and homogenizing the ice optics halves disagreement. Among codes with identical ice optical properties, diversity is largest when forcing is small because codes may disagree on the sign; conversely, diversity is smallest when the magnitude of the forcing is large. When averaged over a realistic simulation of contrail properties, larger daytime diversity contrasts good nighttime (longwave-only) agreement, with space- and time-averaged contrail radiative forcing agreement within 22 %, and 11 % when ice optics are homogenized. These results shows that while accounting for microphysical uncertainties remains important, the identification of the most radiatively-impactful contrails as targets for mitigation strategies is robust to the choice of radiative transfer code. Finally, in a calculation analogous to a previous benchmark, the uncertainty due to radiative transfer code is assessed at 12 %, leading to an overall uncertainty of 60 % in global contrail radiative forcing.