Contrail forecast and nowcast evaluation using satellite-based LIDAR data

Abstract

Contrail avoidance promises to be a near-term solution for mitigating part of aviation’s climate impact [1]. Atmospheric regions that allow for contrails to form and persist have been shown to be horizontally wide but vertically thin [2], motivating the idea that small vertical deviations are sufficient for avoiding the most impactful contrails [1]. Nonetheless, the concept of contrail avoidance relies on skillful forecasts of the regions where contrails will form and persist. Recent comparisons of NWP data and humidity measurements and contrail observations show that the prediction of contrail persistence is problematic [3,4]. Since simulation studies that have previously investigated contrail avoidance have assumed the prediction of these regions to be correct [1], real-world contrail avoidance strategies may be less effective than thought previously [4]. There is thus a need to both understand and improve the performance of prediction methods that could be utilized for contrail avoidance.

Previous work has [5] has resulted in a dataset of over 3000 contrail cross-sections found in CALIOP LIDAR data, obtained by collocating contrails detected using GOES-16 imagery [6]. We have now developed an algorithm that finds the location where an aircraft’s exhaust plume intersects CALIOP data. This allows us to estimate which contrail cross-section corresponds to which flight, as well as estimate which flights did not form a persistent contrail. The resulting dataset is used for the evaluation of existing forecast methods that rely on numerical weather prediction data, as well as a nowcasting algorithm that relies on contrail detections and altitude estimates from GOES-16 data [5,6].

This new forecast evaluation dataset and method can be used to better understand the limitations of existing approaches and enable the development of improved techniques for persistent contrail prediction.