Northern Hemisphere Contrail Properties Derived from MODIS Imagery

One of the recommendations of the Federal Aviation Administration's (FAA) Aviation-Climate Change Research Initiative (ACCRI) is the development of a global climatology of line-shaped contrails and contrail cirrus detectable via satellite remote sensing methods. Such a database is necessary for the validation of a new generation of atmospheric models that represent contrail formation explicitly. In addition to contrail coverage, information on the associated optical properties of the contrails is also needed to reduce our uncertainty of the regional and global contributions of contrails and contrail-generated cirrus to climate change.

We use an automated contrail detection algorithm (CDA) to determine the coverage of linear persistent contrails over the Northern Hemisphere. The contrail detection algorithm is an extension of the Mannstein et al. (1999) method, and uses several channels from Terra and Aqua MODIS data to reduce the occurrence of false positive detections. A set of three contrail masks of varying sensitivity is produced to define the potential range of uncertainty in contrail coverage estimated by the CDA. Global aircraft emissions waypoint data provided by FAA allow comparison of detected contrails with commercial aircraft flight tracks. A pixel-level product based on the advected flight tracks defined by the waypoint data and U-V wind component profiles from the NASA GMAO GEOS-4 reanalysis has been developed to assign a confidence of contrail detection for the contrail mask. To account for possible contrail cirrus missed by the CDA, a post-processing method based on the assumption that pixels adjacent to detected linear contrails will have radiative signatures similar to those of the detected contrails is applied to the Northern Hemisphere data.

Results from a year of MODIS observations (2006) will be presented, representing a near-global climatology of contrail coverage. In addition, a summary of the contrail properties (including optical depth, particle size, and radiative forcing) derived from the multi-spectral CERES cloud property retrieval method will be presented.