Transient contrails as an opportunity for upper tropospheric humidity estimation

When aircraft exhaust mixes with cold air, it forms an ice cloud: a contrail. If the ambient conditions are dry the contrail is transient, meaning that the ice crystals sublimate during mixing, in the first minutes after emission. Conversely, in humid air the contrail can persist and contribute a significant warming radiative forcing. Errors that are present in weather data therefore make contrails (and aviation's climate impact) hard to model. More humidity observations are needed to reduce model errors in this part of the atmosphere. The observation of persistent contrails implies the presence of ice-supersaturated regions. In this study, we establish the potential to extend these opportunistic observations using measurements of transient contrails, enabling direct measurement of relative humidity with respect to ice. A refined contrail jet phase bulk microphysics model is compared to ground camera detection and measurement of contrails immediately behind aircraft. Contrails are detected over an all-sky camera in Palaiseau, France using a cross-track peak detection methodcombining ADS-B aircraft positions and winds derived from aircraft-reported data. Around 5% of daytime overhead aircraft lead to a contrail detection (either transient or persistent). However, many contrails go unobserved. This observability limitation is evident when focusing on those aircraft that fly in air satisfying the Schmidt–Appleman temperature threshold condition, for contrail formation. The condition is satisfied in more than 99% of the observations where a contrail is detected, but fewer than 10% of observations behind aircraft where this condition is satisfied yield a detection. Invariably, this is due to natural cloud or the artefact being too faint or small to be detected using the current instrument and method. We demonstrate that temperature and relative humidity with respect to ice are the main controls on observed transient contrail lifetime. The model reproduces this dependence, though the introduction of a mixing model hybridising the core and bulk plume is critical to constrain this process. This result provides a foundation to infer the relative humidity directly, without requiring new sensors. Some limitations of the observing system remain to be overcome: higher resolution cameras, detection algorithms robust to advection error and understanding the conditions for observability would improve the method accuracy. On top of this, the clear-sky conditions required to detect transient contrails are relatively infrequent, which is an external limitation that must be understood. Nonetheless, these resultshighlight a new pathway to infer the humidity in a part of the atmosphere where accuracy is valuable, but models are insufficiently constrained.