Tracking the Jets as Lagrangian Objects

Accurately determining the position of the upper-tropospheric jets on synoptic scales is key to understanding climate variability and regional weather patterns. Various methods have been developed to track jets and their evolution in time, based on the traditional view that jets are fast and narrow air currents. While such methods have been useful in diagnosing variability and trends, there remains little consensus on long-term trends associated with global warming, and jet mean characteristics vary across methods.

A shared characteristic of many existing algorithms is their reliance on instantaneous or time-mean meteorological fields. However, in this Eulerian framework, the continuity of jets as they meander over time is easily overlooked. As a result, jets often appear as collections of discontinuous streaks that can move within the phase space at nonphysical speeds, while weaker but continuous features drive transport on synoptic scales. In addition to this issue, many algorithms employ climatological parameters that introduce complex sensitivities and may affect long-term trend analyses.

We propose to address these issues by viewing jets as fundamentally Lagrangian objects, so as to use both their distinct morphology _and_ their temporal coherence as sources of information for their detection. We define jets as maxima of synoptic transport and develop a detection algorithm called 'JetLag' (LAGrangian JET). JetLag relies on two primary parameters whose choice is informed by Rossby wave dispersion theory. We apply JetLag to the historical record (1941-2024) as captured by the ERA5 reanalysis, and find that mean jet characteristics are virtually insensitive to parameter choice. We discuss JetLag's improved ability to detect jets at all latitudes, in all seasons, and in all states of the climate. We further compare JetLag to Eulerian detection methods to show changes in the representation of jets, both in the average and in the variability. Notably, Lagrangian jets occur at slightly more poleward latitudes and present larger variability on decadal to multidecadal scales--with important implications for long-term trend detection.