In this project, a novel computational method to model the evolution of aircraft contrail (ice particles forming in aircraft engine’s exhaust) using high-performance computing is being developed. The specific goals of the research are: (1) to carry out the first fully three-dimensional spatial large-eddy simulations of contrail formation that include the aircraft geometry and to develop accurate data-set of contrail evolution in the early phases of the aircraft wake; (2) to identify the 3D contrail features and fit parameters by journaling the simulation workflow using advanced visualization techniques; and (3): to apply reduced-order modeling, such as Polynomial Chaos reconstructions, and machine-learning techniques, such as Artificial Neural Networks, to estimate model parameters and to reduce the large dimensionality of the problem to ease integration into global atmospheric models.

EVL’ visualization work aims to create a multi-scale visual computing system that can identify the spatial features and parameters of the 3D contrails to analyze and explore multiple simulations of the ensemble data. These simulations can be performed for many different cases with specific parameter combinations. A guided summary and comparison of all of the simulations, aim to identify areas of interest across multiple outcomes. Additionally, the visual interface of the visualization system developed will provide users the ability to manipulate input or output parameters to find ways to minimize the formation of the contrails. An anticipated outcome, will be to enable climate researchers to select and examine specific simulations under given parameter values and gain more insight.