Quantification of filaments penetrating the subtropical barrier

A method based on contour advection is introduced that aims to quantify the formation of filaments both equatorward and poleward of the subtropical barrier. It is applied to diagnosed potential vorticity fields for every day of January and February in 1997 and 1998. An isosurface of modified potential vorticity is found to represent the region of largest isentropic gradients of potential vorticity and is hence assumed as the tropopause. Tropopause-penetrating structures ("filaments") forming in contour-advected potential vorticity fields are identified, and statistics of the abundance of such structures are derived. Filamentation as measured by this method exhibits a large temporal variability on the scale of days to weeks. For example, tropospheric cutoff systems developing in the wake of Rossby wave breaking events cause strong filamentation to occur in the lowermost stratosphere. The timescales governing filamentation are a function of altitude, reflecting the differing types and amplitudes of waves inducing filamentation at different isentropic levels. Zonal asymmetries arise as Rossby waves favorably break near the end of the North Atlantic storm track. A difference in the intensities of filamentation in January and February of 1997 and 1998 suggests that the relatively low values of ozone in the lowermost stratosphere observed in 1997 are related to increased filamentation-induced stratosphere-troposphere exchange, compared to 1998. For example, around 30% more filaments are found in the hemispheric and bimonthly mean at the 330 K isentropic surface in January and February of 1997 than during the same months of 1998. Single events can represent most of the filamentation occurring in a month, and hence interannual variability in the frequency of filamentation can be considerable.