A comparison of contrail detectability from LEO and GEO satellite images

Abstract

Despite large uncertainties, persistent contrails are estimated to be responsible for more than half of the additional climate forcing attributed to aviation (Lee et al., 2021). Persistent contrails form only under ice-supersaturated conditions, present in vertically thin regions (Gierens et al., 2020), resulting in a potential to reduce the formation of aircraft-induced cirrus clouds by making altitude deviations of approximately 1km (Sausen et al., 2023). This requires knowledge of the location of the regions to be avoided. Currently, research groups are using images from geostationary (GEO) satellites to observe and detect persistent contrails and to use these detections to estimate regions of persistent contrail formation (Meijer et al., 2022; Ng et al., 2021; Vazquez-Navarro et al., 2010). Observations from GEO satellites provide frequently updated images (every five minutes for GOES-16 ABI), enabling “nowcasting” of contrail forming regions (McCloskey et al., 2021). We have conducted a preliminary assessment, which shows that data from GEO satellites does not resolve many contrails visible on images from satellites in the low-Earth orbit (LEO). This was particularly observed under conditions of high background cloudiness and shows a noticeable underestimation of the extent of regions of persistent contrail formation. Since LEO satellites orbit closer to the Earth’s surface (typically at an altitude around 800 km) compared to GEO satellites (35,786 km), instruments onboard LEO satellites often have a higher spatial resolution. However, most LEO satellites overfly most points on Earth only every 12 hours. We therefore find that, while high-resolution LEO images provide the potential for improved contrail detection at certain points in time, due to their more limited temporal resolution, they can only serve as an additional layer of information rather than as a sole source for contrail detection. Here we present a dataset of collocated observations from a variety of sources, including GEO and LEO infrared radiometers, LEO lidar, data from numerical weather prediction models, and contrail height estimates from correlations of satellite observations with flight data. With this data, we are building our “ground truth library” of contrail observations, which allowed us to systematically investigate the influence of different cloud and atmospheric parameters on contrail detection. Since attributing observed contrails to individual flights still is a highly manual process, we further investigated which instrument channels allow for the best human identification of contrails under different conditions. Considering weather data as an additional layer of information, especially when cloudiness impacts the visibility for all sources of visual data, allows us to assess the likelihood of contrail persistence in these regions and to identify the correlated uncertainty bounds for this process. Our investigation yields a multifaceted assessment of contrail detectability and the potential of different data sources to improve the identification of regions that allow for contrail persistence.