Uncertainties in contrail modelling of high impact regions

Aviation impacts Earth’s climate by the formation of contrails, where the largest fraction of the total atmospheric impact is caused by only a small fraction of flights. Forecasting of high impact regions which are likely to produce strongly warming contrails can be achieved with the Lagrangian plume model CoCiP (Contrail cirrus Prediction tool). The model calculates the instantaneous energy forcing per meter flight distance (EFm) during a contrails’ lifetime. It can be operated in two modes: Either along specific flight trajectories or on a 4D grid over an extended area. Simulations depend on meteorological input data and information on aircraft type, heading and contrail overlap. In this study we analyse the impact of these input parameters on the EFm and assess the spatial uncertainty of high-impact regions.

We use the pycontrails open-source implementation of CoCiP to simulate the potential contrail occurrence on a grid over Europe multiple times. We successively vary i) the aircraft type across an ensemble of the 10 most common types, ii) the heading from northward to eastward and iii) the meteorology across the 10-member ensemble of ECMWF ERA5 data for 2019. Additionally, we simulate the contrail formation from historical flight trajectories in 2019 provided by the Global Aviation emissions Inventory based on ADS-B (GAIA), both with and without accounting for contrail radiative interactions.

The standard deviation in EFm is largest for the aircraft ensemble, followed by the meteorology ensemble. This reveals that the aircraft type has significant impact on the climate effect from contrails. Within the aircraft ensemble, 26% of high-impact grid points are agreed upon by all ensemble members, whereas 8% are considered as high impact by only one member. The radiative interaction between contrails leads to a reduction in EFm by 2%, with a stronger effect during night than during daytime. In areas of high air traffic density, the reduction increases up to 5%.

Our results help to assess uncertainties in the prediction of contrail-sensitive airspaces and of individual flights necessary for operational contrail avoidance.