Large Uncertainty in Observed Meridional Stream Function Tropical Expansion

Abstract. Recent tropical expansion rate estimates vary substantially, as a multitude of methods and reanalysis datasets yield conflicting results. Among the many methods of estimating the tropical width, the meridional stream function 500 hPa zero-crossing is the most widely used, as it is directly related to the poleward edge of the Hadley Cell (HC). Other common metrics use atmospheric phenomena associated with the HC as a proxy, for instance the zonal surface wind zero-crossing. As each of these metrics require different data, each with varying error, the level of data-driven uncertainty differs between each metric. While previous work has analyzed the statistical and dynamical relationships between metrics, to date no study has quantified and compared the uncertainty in different HC metrics. In this study, we use ERA5 ensemble members, which include small perturbations in atmospheric variables based on the data error, to quantify the uncertainty associated with six commonly used HC metrics as well as the range of their trend estimates. In the Northern Hemisphere, the tropical expansion rate calculated by the stream function is roughly 0.05 degrees per decade, while the Southern Hemisphere rate is 0.2 degrees per decade. Of the six metrics, only the meridional stream function and precipitation minus evaporation have substantial uncertainties. The stream function errors are large due to uncertainty in the underlying meridional wind data and the presence of large regions of near-neutral circulation at the poleward edge of the tropics. These errors have decreased in recent decades because of improvements in the assimilated observations. Despite these improvements, we recommended using the zonal surface wind zero crossing to analyze tropical extent trends in reanalyses. This is particularly important in the Northern Hemisphere, before the year 2000, and when studying individual seasons other than winter.