Portal de Periódicos da CAPES

Comment on “A Lagrangian analysis of stratospheric ozone variability and long‐term trends above Payerne (Switzerland) during 1970–2001” by G. Koch et al.

2003; American Geophysical Union; Volume: 108; Issue: D21 Linguagem: Inglês

10.1029/2003jd003687

2156-2202

H. Teitelbaum, C. Basdevant, Albert Hertzog, N. Semane,

Atmospheric and Environmental Gas Dynamics

[1] In a recent paper devoted to the study of ozone variability and long-term trends above Payerne (Switzerland), Koch et al. [2002] (hereinafter referred to as K02) try, in their section 3, an explanation for an extreme ozone profile observed on 29 November 2000. By means of 10-day backward isentropic trajectories at various levels, they infer that this deep ozone minihole was due to the simultaneous arrival over Payerne of ozone-poor air masses coming, at low levels (340–440 K), from the subtropics and, at high levels (530–700 K), from polar origin. To further convince the reader, they reconstruct an ozone profile using ozone climatological values of the “origin” of the air masses; a synthetic profile that fits rather well the measured profile. [2] We think, however, that this interpretation of the minihole formation is erroneous and that the authors' arguments are misleading. Our criticism will be twofold. [3] First, although back-trajectories using European Centre for Medium-Range Weather Forecasts (ECMWF) or National Centers for Environmental Prediction (NCEP) data are reliable deterministically for a few days, the information they provide for longer periods can only be taken as statistically valid. Moreover, there is no justification for reconstructing an ozone profile with 10-day back-trajectories, or with 5-day or one-month back-trajectories. Further, the invariance of an air parcel composition along a trajectory is rather elusive as a given constituent can increase or decrease by diffusion or turbulent mixing. [4] Second, basically, the analysis given by K02 of the formation of this minihole fails because they focus uniquely on the vertical ozone profile measured at Payerne on 29 November 2000, failing to explain the minihole precursors that are clearly seen, since 27 November crossing western Europe, on TOMS ozone fields. Although no ozone profile is available for these days and locations, TOMS data are sufficiently reliable and coherent with the sounding in Payerne that any explanation must take into account the complete sequence from 27 November, when an ozone minihole appears above Spain, through 28 and 29 November when it is the deepest, until 30 November just before it disappears east of Italy. [5] Besides, this Payerne ozone minihole has already been studied by Semane et al. [2002]. In short, it is shown by Semane et al. [2002] that this very deep minihole is not the consequence of the advection of a low ozone bubble, but was first created by the local uplift of isentropic surfaces and then intensified in the encounter, at high levels, with the polar vortex. Since this type of explanation of minihole formation is still ignored by numerous people and while the explanation based on air mass advection seems to be still popular, we want to go further in the argument developed by Semane et al. [2002]. [6] Ozone minihole formation has been studied by several authors [e.g., McKenna et al., 1989; Teitelbaum and Sadourny, 1998; Teitelbaum et al., 2001; James and Peters, 2002]. Teitelbaum and Sadourny [1998], studying polar stratospheric clouds (PSCs) and associated miniholes in both hemispheres, showed that the main cause of both phenomena was a synoptic upwelling of isentropic surfaces, which decreases the ozone partial pressure and cools the air mass. Teitelbaum et al. [2001] identify and characterize the meteorological conditions that best explain the simultaneous occurrence of PSCs and miniholes in the Arctic. Following Hoskins et al. [1985], Rossby waves deform the zonally oriented jet at the tropopause, inducing in the upper troposphere and lower stratosphere an isentropic potential vorticity (IPV) anomaly and, where the jet is locally displaced toward the north forming a meander, an anticyclonic circulation. The anticyclonic circulation induces an upward deformation of isentropes above the tropopause and a downward deformation below. The integrated ozone partial pressure decreases in the lifted column of air, resulting in the localized ozone minimum. Thus the minihole is also associated with an elevated tropopause, a geopotential maximum on isentropic surfaces in the stratosphere, and a stratospheric temperature minimum. The intensity of the uplift and then that of the columnar total ozone decrease are linked to the shape of the jet meander. [7] To further demonstrate that the Payerne minihole has been formed by an uplift mechanism rather than by a transport process, we will develop in the following sections two types of arguments: First, we will show that the synoptic-scale situation from 27 to 29 November 2000 explains the minihole formation. Then we will demonstrate that air mass trajectories are not compatible with the minihole formation. [8] Full color maps of TOMS total ozone in Dobson units (DU) for 27, 28, 29, and 30 November 2000 are given by Semane et al. [2002, Figure 1]. The minihole is clearly visible over Spain on 27 November, the ozone column minimum reaching 216 DU. The minihole intensifies crossing western Europe and then vanishes; the TOMS ozone minimum is 196 DU on 30 November, 195 DU on 29 November, and 207 DU on 30 November, so very close to the 198 DU on 29 November 2000 reported by K02. Both the evolution of the synoptic-scale situation and the ozone minihole location during these 4 days are displayed in Figure 1. The horizontal velocity field on the 330 K isentrope, computed from ECMWF data, is displayed together with two IPV isolines. These two isolines indicate the jet position and the strong IPV gradient. The 220-DU and lower ozone levels are also displayed in grayscale. These maps clearly show that the minihole is located at the same place as, and moves with, the anticyclonic circulation that developed over Europe at that time. This anticyclonic circulation induces a quasi-adiabatic uplift of the air column above, and, the ozone mixing ratio being conserved in the uplift, the total column ozone decreases. The vertical structure of this phenomenon is described in Figure 6 of Semane et al. [2002], where, for 2 days, isentropes are drawn on a longitude/geopotential height cross section at the latitude of the ozone minimum on that day and the 2.5-IPV isolines that indicates the height of the tropopause. As expected, the tropopause is uplifted at the respective positions of the minihole. Isentropes are uplifted above the tropopause and lowered below. It is also shown that the minihole position corresponds to a geopotential maximum on isentropic surfaces in the lower stratosphere (an anticyclone). [9] To furthermore dismiss the ozone-poor air transport hypothesis, we present three sets of trajectories. In Figure 2 are drawn three-dimensional 3-day forward trajectories computed with the FLEXTRA trajectory model (A. Stohl, The FLEXTRA trajectory model, 1999, available at http://www.forst.uni-muenchen.de/EXT/LST/METEO/stohl/flextra.html) using ECMWF wind fields. Trajectories start from the minihole position on 27 November at six different altitudes ranging from 10 km to 30 km. It can be seen that none of these trajectories follow the minihole displacement, although their different altitudes span the ozone range associated with the maximum ozone deficit (see Figure 3 of K02). This further proves that a minihole is not associated with an air mass. [10] In Figure 3 are drawn 3 three-dimensional 10-day back-trajectories, computed with FLEXTRA and ECMWF data, arriving at the minihole location on 29 November at altitudes 12, 16, and 20 km. These back-trajectories are similar in their horizontal behavior to the isentropic back-trajectories of K02 at levels between 330 and 440 K. However, our interpretation is completely different in terms of the minihole formation. For instance, it can be seen in Figure 3 that the air mass displacement is far more rapid than the minihole displacement; air masses arriving on 29 November above Payerne were above the Atlantic Ocean, south of 30°N, on 27 November, i.e., at the time when the minihole has already appeared over Spain. [11] Moreover, 10-day isentropic back-trajectories are used by K02 to reconstruct the 29 November 2002 ozone profile. First, it should be noticed that there is no physical justification for considering 10-day trajectories or longer or shorter trajectories. The typical chemical lifetime of ozone in the lower stratosphere is, for instance, much longer than 10 days. Second, the methodology they used to reconstruct the ozone profile is particularly open to criticism. They actually take, at each level in the lower stratosphere, the climatological ozone value at the southernmost latitude reached during these 10 days. It can be seen in our Figure 3 (or Figure 5 of K02) that the trajectory latitudinal excursion is large and, once again, there is no justification for the ozone mixing ratio to be taken at the southernmost latitude. It can also be seen in Figure 3 that the air mass altitude varies noticeably along the trajectories which favor mixing and diffusion processes. Incidentally, the air mass uplift can be seen at the end of the trajectories. [12] For the upper part of the ozone profile, Figure 4 displays 3 three-dimensional 10-day back-trajectories arriving at the minihole location on 29 November at altitudes of 24, 26, and 28 km. These trajectories are similar to the isentropic trajectories at higher levels given by K02. Again they are not compatible, in direction and in time, with the minihole displacement. In contrast to the lower level, K02 decided here to take the ozone mixing ratio at the northernmost latitude along the trajectories, which looks like a rather educated guess! From our point of view, their argument is therefore not valid, and no transport phenomenon can explain the minihole formation. [13] We have showed that, as in most occurrences, the ozone minihole observed above Payerne (Switzerland) on 29 November 2000 was due primarily to an anticyclonic anomaly of the jet, which induced a quasi-adiabatic uplift of isentropic surfaces in the stratosphere and consequently a columnar ozone decrease. The hypothesis that it was due to the arrival of ozone-poor air bubbles as stated by K02 must therefore be rejected. [14] We are not convinced by the reply by Koch et al. [2003] (hereinafter referred as K03), principally for the following reasons: [15] 1. Figure 1 of K03 is supposed to demonstrate that 3-day or 10-day backward trajectories would have given similar results; however, our calculations (not shown) prove that 10-day trajectories reach far more southward latitudes at low levels and far more northward latitudes at high levels than 3-day trajectories. Thus, using their transport hypothesis, ozone profile would not be the same (note that the ozone profile is not reconstructed with the 3-day trajectories of K03). [16] 2. Reconstructing ozone profile from a two-dimensional climatology ignores the strong zonal and temporal variabilities of ozone and is therefore prone to very large uncertainties (this fact is admitted by K02); the demonstration of K02 and K03 is thus very dubious. [17] 3. If one would follow the argument of K03, as the circulation was southward in the middle stratosphere and northward in the lower stratosphere at that time over Europe, a large ozone hole should have appeared over western Europe and not a localized minihole. [18] 4. The mechanism developed by K02 and K03 cannot explain why a potential vorticity anomaly and a minihole develop at the same location and move together, while our mechanism reveals the link between both processes. [19] We are grateful to ECMWF for providing the meteorological fields. The air mass trajectory calculations have been performed with the FLEXTRA model of A. Stohl. The TOMS data were produced and provided by the NASA Ozone Processing Team at Goddard Space Flight Center.