Trajectory dispersion by unresolved wind variability in the UTLS
Tuesday, 16 December 2014
Lagrangian trajectory calculations, in which air-parcel paths are determined from wind fields from operational analysis and reanalysis data, are standard tools for diagnosing dynamical interactions and chemical concentrations in the upper troposphere and lower stratosphere. However, due to poor observational sampling, these calculations suffer from uncertainties that are difficult to account for. Potentially large contributors to these uncertainties are velocity fluctuations at space and time scales not resolved by the analysis data sets used to calculate trajectories. These fluctuations cause trajectories to disperse from paths determined by resolved winds. While filling the data voids at small scales is not a viable alternative for the foreseeable future, the small-scale variability has robust statistical relationships with variations at resolved scales. We examine these relationships and their uncertainties using wind fields from the high-resolution (12 km in the horizontal) ECMWF operational analysis and from aircraft observations from ATTREX. Estimates of the corresponding trajectory dispersion are then determined using a stochastically forced Lagrangian trajectory model.