 |
 |
|
Figure 1. Zonal means of the temperature [K] (a) and ozone mixing ration [ppmv] calculated from data from August 14, 1997. |
W. E. Ward1,2, J. Oberheide3, M. Riese3, P. Preusse3, and D. Offermann3
1CRESS/CRESTech, York University, North York,
Ontario, Canada.
2now at University of New Brunswick, Canada.
3Physics Department, University of Wuppertal,
Wuppertal, Germany.
The Cryogenic Infrared Spectrometers and Telescopes for the Atmosphere (CRISTA) instrument has flown on two shuttle missions providing observations of constituents and temperature throughout the stratosphere and mesosphere at high vertical and horizontal resolution. During the second mission (CRISTA 2: August 7-17, 1997) a wave two signature was observed in the southern midlatitudes in the temperature data between 17 and 80 km and in the ozone data between 17 and 63 km (the limits of ozone data availability for the current data version). This signature has a period of ~12.5 days and a vertical wavelength of ~45 km. It is one of the few observational instances showing vertical and equatorward propagation of a planetary wave well into the mesosphere. The relative phase of the signature in the temperature and ozone varies with height with the two being in phase below 30 km and 180 degrees out of phase above 40 km. These signatures are consistent with those expected from a migrating planetary wave with the ozone signature being dynamically driven at lower altitudes and photochemically driven at higher altitudes. The CRISTA results above the middle stratosphere show a greater agreement with photochemical equilibrium calculations than previous analyses. These results clearly illustrate the penetration of a dynamical feature from the tropopause into the mesosphere and its effect on constituent distributions throughout this height range.
The larger scale, long term systematic variation of ozone in the middle atmosphere has been a topic of major interest over the past 20 years [see Solomon, 1999; Jucks and Salawitch, this issue; Brasseur et al., this issue]. This activity has been fueled by evidence indicating that the total column amount of ozone is decreasing due to human use of chlorofluorocarbons thereby putting life on the planet at risk [Farman et al., 1985]. Much of this research has been concentrated on resolving questions associated with the chemistry of ozone, determining the morphology and time history of the large scale variations in its distribution and understanding the causes for these variations.
Ozone also varies over smaller spatial and shorter time scales. This variability is part of the natural response of ozone to dynamical and photochemical processes and is effectively a source of noise to those studying the long term systematic variations [Randel and Cobb, 1994]. An ability to model these smaller and shorter scale variations, however, provides an opportunity to confirm that the basic physics and chemistry associated with ozone is understood [Froidevaux et al., 1989; Randel et al., 1990; Smith, 1995; Allen et al., 1997].
In this paper, we present a simple analysis of temperature and ozone measurements from the 8 days during the CRISTA 2 mission (August 8-16, 1997) when CRISTA was actively taking measurements. At this time a large planetary wave 2 signature was present in the southern hemisphere. This wave extended from the tropopause to about 80 km in the temperature data and to at least 63 km in the ozone data (the upper boundary for ozone data in Version B01 data). Above 40 km the ozone and temperature perturbations are anti-correlated and below 30 km they are correlated. Similar correlations have been reported [Barnett et al., 1975; Douglass et al., 1985; Froidevaux et al., 1989; Randel, 1993; Fishbein et al., 1993, Sabutis et al., 1997] and modelled [Hartmann and Garcia, 1979; Rood and Douglass, 1985; Rose and Brasseur, 1989; Smith 1995; Reddmann et al., 1999] and the CRISTA result conforms to the accepted understanding of this feature. Compared to previous results, the CRISTA data are of finer vertical resolution and greater vertical extent. The data presented here show that the planetary wave and its influence on the ozone concentration extend across the stratopause into the mesosphere.
The plan of this paper is as follows. First the CRISTA results are presented and the observed correlations described. These correlations are then analysed from a photochemical perspective [Froidevaux et al., 1989]. The paper concludes with some comments on the relevance of these results to the interpretation of high resolution satellite observations.
The CRISTA instrument is designed to measure temperature and constituent distributions at high resolution by monitoring their thermal emissions (4-71 um) using several IR grating spectometers [see Offermann et al., 1999; Riese et al., 1999; Grossmann (this volume) for details of the instrumentation and data analysis procedure]. The basic configuration for the CRISTA 2 flight was the same as the CRISTA 1 flight. CRISTA collected spectra at the limb behind the spacecraft using three telescopes oriented so that the two adjacent telescopes viewed at azimuths of 18° relative to the central telescope. The temperature and ozone data reported on in this paper were retrieved from the CO2 12.6 um (for the stratosphere) and 15 um (for the mesosphere) emissions and the ozone 12.7 um emissions respectively. Temperature errors were of the order of 2 K and the ozone uncertainty was 10-15%. The data presented in this paper are a combination of observations from all three telescopes and ascending and descending nodes of the satellite orbit.
In Figure 1 zonal means of the temperature and ozone mixing ratio for this mission are presented. The mission date of mid-August corresponds to northern hemisphere (NH) late summer/southern hemisphere (SH) late winter. The temperature field is similar to summer solstice conditions with a warm stratopause and cold upper mesosphere in the summer (southern) hemisphere. The ozone field is roughly symmetric about the equator with a gradient of ~1 ppm/10° of latitude at 50° S.
|
|
|
Figure 1. Zonal means of the temperature [K] (a) and ozone mixing ration [ppmv] calculated from data from August 14, 1997. |
A planetary wave feature is seen in the CRISTA temperature and ozone fields. Figures 2 and 3 are SH polar views of these fields at heights of 20, 30, 40, 45, 50, and 55 km for August 15, 1997 (representative for all days of the mission). In each subplot, the range of ozone mixing ratio or temperature is listed in square brackets to the lower left. A wavenumber 2 signature is prominent in the temperature figures. A similar feature is seen clearly in the ozone fields at heights above 40 km and less distinctly at heights of 20 and 30 km. In each view, contours are used to isolate the regions of maximum and minimum temperature and ozone mixing ratio (dark contours at 70% of the full range and light contours at 30% of the full range). At 40 km and above, the regions of enhanced ozone correspond closely to regions of low temperature. Below 40 km the correlation is reversed.
 |
|
Figure 2. Polar views of the temperature field in the SH at 6 different altitudes. Dark contours are plotted at 70% and light contours are plotted at 30% of the full contour interval. |
 |
|
Figure 3. As in Figure 2 except for the ozone mixing ratio [ppmv]. |
Figure 4 shows the amplitude and phase of this wave 2 disturbance as a function of latitude and height for the temperature perturbation field (upper subplot) and the relative ozone variation (lower subplot) calculated by fitting in a least-mean-squares (LMS) sense for the wave parameters using latitude/height bins of 5° by 1.5 km. The amplitude appears as the grey scale map and the phase as the superimposed black contouring. The phase corresponds to the longitude at which the maximum amplitude occurs on August 9.
 |
 |
|
Figure 4. Height verses latitude cuts of the amplitude and phase of the wave2 temperature (uppera panel; in Kelvin) and ozone perturbations (lower panel; % change in ozone mixing ratio relative to the zonal mean value) in the Southern Hemisphere. The amplitudes are in grey scale and the phases are the labeled contours in degrees of longitude (see text for details). |
The temperature perturbation field shows amplitude maxima near 50°S of ~11, 13, and 9 K occuring at 25, 45 and 70 km respectively with equatorward propagation in the meosphere. Although expected from modelling studies (see Holton and Alexander, this issue, for a general description of planetary wave propagation characteristics) this is one of the few instances where this behaviour has been observed in atmospheric data. The minimum in the temperature at 35 km corresponds to the height of the maximum geopotential wave amplitude. The best fit for the wavenumber 2 signature was for an eastward propagating disturbance with a period of ~12.5 days. Using the 90° and -90° phase contours, the wavelength is seen to be ~45 km. LMS calculations also indicate the existence of a wavenumber 1 disturbance of approximately the same period and amplitude below 30 km poleward of the wave 2 disturbance. These features are consistent with previous descriptions of SH wave 2 planetary wave events in the late winter/early spring [Manney et al., 1991].
The ozone perturbation fields (Figure 4b) have significant amplitude above 25 km. The phase remains constant throughout this region at ~90° of longitude. Because the phase of the temperature perturbation varies with height, the ozone variations are roughly in phase below 30 km and out of phase above 40 km as expected from the plan views described above.
These data show that during the CRISTA 2 mission, the dynamics of the southern hemisphere mid to high latitudes are dominated by a wave 2 planetary wave extending from 20 km to 80 km with a maximum temperature amplitude of the order of 13 K at 45 km. Both the temperature and ozone fields show the planetary wave signature with the phase of the ozone signatures changing sign relative to the temperature at about 40 km.
The conceptual framework for temperature/ozone correlations such as those noted above was first developed by Hartmann and Garcia [1979]. They included both ozone chemistry, parameterized linearly in terms of its response to temperature perturbations and relaxation of odd oxygen perturbations, and transport effects. The behaviour of ozone was shown to vary with height as a result of the variation of its photochemical lifetime relative to dynamical timescale (taken to be ~1 day). Where its photochemical lifetime is short relative to dynamical time scales (the upper stratosphere and mesosphere, above ~35 km), the photochemistry would be expected to dominate and where its lifetime is slow relative to the dynamics (lower stratosphere, below ~35 km) transport effects would be expected to dominate. Because the ozone production rate increases with decreasing temperature, in regions dominated by photochemistry a negative correlation between temperature and ozone mixing ratio would occur. In regions dominated by transport, the effect was shown to vary. For large scale waves which penetrate to a sufficient height into the upper stratosphere (i.e. their maximum amplitude is above the transition region so that phase relations between the various velocity components associated with wave growth hold) they showed that the correlation would be positive. Work by Rood and Douglass [1985] demonstrated that care must be taken in applying these concepts. In circumstances where there are strong dynamical processes (such as stratospheric warmings), the advective terms can dominate the correlations even in regions where photochemistry is fast.
An alternate treatment of the temperature dependence of the ozone photochemistry under photochemical equilibrium conditions excluding consideration of dynamical effects was introduced by Barnett et al. [1975] and developed further in a number of studies [Haigh and Pyle, 1982; Froidevaux et al., 1989; Smith, 1995]. They pointed out that the temperature dependence could be represented by
| [O3] = B exp(Thetap/T) |
(1) |
where [O3] is the ozone concentration, and B and Thetap are constants whose value depended on the concentrations of the reactants involved in the ozone chemistry. Values vary with Thetap ~1405 K for pure oxygen chemistry, ~1200 K for NOx chemistry, ~400 K for HOx chemistry and ~200 K for Clx chemistry [Haigh and Pyle, 1982]. For the analysis of data at a given level, the derivative of the log of this expression provides the relationship
|
d[O3]/[O3] = - ThetaE/T2 * dT |
(2) |
where ThetaE is an empirical constant to be determined. The advantage of this formulation is that it removes the dependence on the absolute value of the ozone mixing ratio and allows the temperature dependence of the chemistry to be investigated directly [Froidevaux et al., 1989]. Inclusion of additional feedback effects between temperature and ozone concentration arising from UV heating variations due to variations in ozone and radiative damping of temperature perturbations results in the relationship [Brasseur and Solomon, 1986]:
|
d[O3]/[O3] = (alpha/PH - ThetaE/T2) dT |
(3) |
where alpha is the radiative relaxation rate and PH is the gross UV heating rate. Approximate values for PH~10 K day-1 from Brasseur et al. [1990] and alpha~0.24 day-1 [Brasseur et al., 1987] are used in the calculations which follow.
Scatterplots of the ozone and temperature perturbations are provided in Figure 5 using SH data at 30 (upper panel) and 50 km (lower panel) from August 14, 1997. For this figure and the following analysis, temperature and ozone perturbations are calculated relative to the zonal mean values calculated for the corresponding latitude/height bin (see Figure 1 for the mean fields). At 30 km where the correlation is expected to be dynamically driven d[O3]/[O3] verses dT/T is plotted (Figure 5a) and as expected, the correlation is positive. Further interpretation of this correlation is possible but requires a detailed dynamical analysis which we will not undertake here.
 |
 |
|
Figure 5. Scatter plots of normalized ozone perturbation vs normalized temperature perturbation at 30 km (upper panel) and -[1490 K] dT/T2 at 50 km (lower panel). See text for details. |
At 50 km the correlation is expected to be photochemically driven. In accordance with (3), values of d[O3]/[O3] and -[ThetaE/T2]dT are plotted (Figure 5b). The ThetaE value of 1490 K which is used is the value for which a maximum correlation is obtained (feedback effects are included). The resulting relationship is close to linear, suggesting that the photochemical description used here is appropriate and the dynamics play a second order role. Such an interpretation can also be supported on other grounds. As demonstrated by Froidevaux et al. [1989], since the dominant dynamical feature has a period of 12.5 days parcels are able to adjust to local photochemical equilibrium. In addition, since the zonal mean gradient in O3 is gradual above 40 km (~ 1 ppm per 10° of latitude at 50° as noted earlier), meridional transport is unlikely to contribute significantly to the observed correlations.
Optimal values of ThetaE are calculated as a function of height between 37 and 55 km for SH data for August 14 with and without feedback effects included (Figure 6, left subplot). The right hand subplot shows the correlation coefficient corresponding to this value. The correlation coefficient is greater than 0.8 above about 40 km but drops rapidly below this height. This is as expected since below about 40 km the photochemical time scale becomes of the same order as the dynamical time scale and the correlation would be expected to break down. The variation of ThetaE (feedback effects included) with height is similar to that calculated by Froidevaux et al. [1989] for photochemical equilibrium conditions (filled dots in Figure 4, left subplot) except that the peak at ~40 km is higher (1900 K instead of 1500 K). The decrease with height of Theta seen here also appears in their calculations and is attributed to the increased influence of the HOx cycle with its lower temperature dependence on the observed Theta. This is in contrast with the ThetaE calculated from the Limb Infrared Monitor of the Stratosphere (LIMS) observations which increases with height [Froidevaux et al., 1989]. The no feedback case also compares favourably with the corresponding Froidevaux et al. [1989] calculations. In terms of the quality of fit there appears to be little difference whether feedback is included or not. However, for data interpretation use of the ThetaE values calculated with feedback is physically more realistic.
 |
|
Figure 6. Plots of the value of Theta derived from SH data from August 14 with (solid line), without (dashed line) feedback effects and from the photochemical equilibrium calculations of Froidevaux et al., [1989] (solid circles, left panel) and the associated correlation coefficient (right panel). |
Although CRISTA ozone data only extend to ~60 km at the present time, the observed correlation would be expected to be present at greater heights as the HOx remains important for ozone photochemistry thoughout the mesosphere [Allen et al., 1984; Summers and Conway, this issue]. At greater heights (above ~65 km) however day/night differences and tidal dynamics become important and need to be taken into account [see Zhu et al., this issue]. As a result, the data interpretation at these heights will become more complex .
Correlations between ozone and temperature have been reported in the literature for a number of years. Such correlations are reported in long term statistical studies [Randal and Cobb, 1994; Sabutis et al., 1997], studies of zonal means [Froidevaux et al., 1989], Fourier decomposition of large scale disturbances [Elson et al., 1994] and large scale wave breaking events [Fishbein et al., 1993]. In general ozone and temperature were observed to be positively correlated below ~30 km where the dynamical timescale is less than the photochemical timescale and negatively correlated above about 40 km where the photochemical timescale is less than the dynamical timescale. Modelling studies [Hartmann and Garcia, 1979; Rood and Douglass, 1985] confirmed these basic results but also indicated that a careful analysis of the dynamical contributions to the ozone transport was necessary before the observed correlations could be attributed to any specific process.
CRISTA observations provide another opportunity to examine this relationship between ozone and temperature. Variations in ozone and temperature data in the SH midlatitudes are shown to be associated with a wave 2 planetary wave. This wave extends throughout the data set from the lower stratosphere at 20 km to the mid-mesosphere at 80 km. The data are of sufficient accuracy that correlations may be calculated on a measurement-by-measurement basis. Correlations between these two parameters conform to previously published results and show ozone perturbations to be anti-correlated with temperature perturbations across the stratopause and correlated with temperature perturbations in the lower stratosphere. The CRISTA observations above 40 km agree closely with the photochemical equilibrium calculations of Froidevaux et al. [1989] in spite of the presence of strong planetary wave signatures.
Further work with this data set is certainly possible. The CRISTA observations are of higher vertical and horizontal resolution than previously published work. Velocities may be derived from the temperature fields and the potential vorticity and Eliassen-Palm flux calculated and used to diagnose the dynamical conditions at this time [as in Manney et al., 1991]. A more extensive correlation analysis than that provided here could be undertaken and related more directly to the dynamical conditions. The development of a linear advective-photochemical model [Hartmann and Garcia, 1979; Douglass et al., 1985; Randel, 1990] would allow the contribution of the advective terms to be evaluated. Further evaluation of the contribution of the various photochemical cycles to the ozone variability could be undertaken by explicitly including other species measured by CRISTA into the photochemical equilibrium analysis undertaken here. It would be of interest to determine whether the close agreement with photochemical calculations noted here is a matter of superior data quality or the particular dynamical conditions present at the time of these measurements.
The Centre for Research in Earth and Space Technology is supported by the Technology Fund of the Province of Ontario. The CRISTA project was supported by grants 50 OE 8503 and 9501. of Deutsche Agentur für Weltraumangelegenheiten (DARA), Bonn, Germany. The CRISTA 2 instrument was flown as part of the Space Shuttle mission STS-85 of the National Aeronautics and Space Administration (NASA), USA.
Allen, D.R., J.L. Stanford, L.S. Elson, E.F. Fishbein, L. Froidevaux and J.W. Waters Allen, D.R., et al., The 4-day wave as observed from the Upper Atmosphere Research Satellite Microwave Limb Sounder, J. Atmos. Sci., 54, 420--434, 1997.
Allen, M., et al., The vertical distibution of ozone in the mesosphere and lower thermosphere, J. Geophys. Res., 89, 4841-4872, 1984.
Barnett, J.J., J.T. Houghton, and J.A. Pyle, The temperature dependence of the ozone concentration near the stratopause, Q. J. R. Meteorol. Soc., 101, 245-257, 1975.Brasseur, G. and S. Solomon, Aeronomy of the Middle Atmosphere, 441 pp., D. Reidel, Norwell, Mass., 1984.
Brasseur, G., A. De Rudder, G.M. Keating, and M.C. Pitts, Brasseur, G., et al., Response of middle atmosphere to short term solar ultra-violet variations: 2 Theory, J. Geophys. Res., 92, 903--914, 1987.
Brasseur, G., M.H. Hitchman, S. Walters, M. Dymek, E. Falise, and M. Pirre Brasseur, G.,et al., An interactive chemical dynamical radiative two-dimensional model of the middle atmosphere, J. Geophys. Res., 95, 5639--5655, 1990.
Brasseur et al., Natural and human-induced perturbations in the middle atmosphere, this volume, 2000.
Douglass, A.R., R.B. Rood, and R.S. Stolarski, Douglass, A.R., et al., Interpretation of ozone temperature correlations 2. Analysis of SBUV ozone data, J. Geophys. Res., 90, 10693--10708, 1985.
Elson, L.S., G.L. Manney, L. Froidevaux, and J.W. Waters, Elson, L.S., et al., Large-scale variations in ozone from the first two years of UARS MLS data, J. Atmos. Sci., 51, 2867--2876, 1994.
Farman, J.C., B.G. Gardner, and J.D. Shanklin, Large losses of total ozone in Antarctica reveal seasonal ClOx/NOx interaction, Nature, 315, 207--210, 1985.
Fishbein, E.F., et al., MLS observations of stratospheric waves in temperature and O3during the 1992 southern winter, Geophys. Res. Lett., 20, 1255--1258, 1993.
Froidevaux, L., et al., The mean ozone profile and its temperature sensitivity in the upper stratosphere and lower mesosphere: An analysis of LIMS observations, J. Geophys. Res., 94, 6389--6417, 1989.
Grossmann, K.U., Recent Improvements in middle atmosphere remote sounding techniques: The CRISTA-SPAS Experiment, this volume, 2000.
Haigh, J.D., and J.A. Pyle, Ozone perturbations experiments in a two-dimensional model, Q. J. R. Meteorol. Soc., 108, 551--574, 1982.
Hartmann, D.L., and R.R. Garcia, A mechanistic model of ozone transport by planetary waves in the stratosphere, J. Atmos. Sci., 36, 350--364, 1979.
Holton, J.R., and M.J. Alexander, The role of waves in the transport circulation of the middle atmosphere, this volume, 2000.
Jucks, K.W., and R.J. Salawitch, Future changes in upper stratospheric ozone, this volume, 2000.
Manney, G.L. et al., The behavior of wave 2 in the southern hemisphere stratosphere during late winter and early spring, J. Atmos. Sci., 48, 976--998, 1991.
Offermann, D., et al., The Cryogenic Infrared Spectrometers and Telescopes for the Atmosphere (CRISTA) experiment and middle atmosphere variability, J. Geophys. Res., 104, 16311--16325, 1999.
Randel, W.J., Kelvin wave-induced trace constituent oscillations in the equatorial stratosphere, J. Geophys. Res., 95, 18641--18652, 1990.
Randel, W.J., Global normal-mode Rossby waves observed in stratospheric ozone data, J. Atmos. Sci, 50, 406--420, 1993.
Randel, W.J. and J.B. Cobb, Coherent variations of monthly mean total ozone and lower statospheric temperature, J. Geophys. Res., 99, 5433--5477, 1994.
Reddmann, T., R. Ruhnke, and W. Kouker, Use of coupled ozone fields in a 3-D circulation model of the middle atmosphere, Ann. Geophysicae, 17, 415--429, 1999.
Riese, M., et al., Cryogenic Infrared Spectrometers and Telescopes for the Atmosphere (CRISTA) limb scan measurements, data processing, and atmospheric temperature and trace gas retrieval, J. Geophys. Res., 104, 16349--16367, 1999.
Rood, R.B. and A.R. Douglass, Interpretation of ozone temperature correlations 1. Theory, J. Geophys. Res., 90, 5733--5743, 1985.
Rose, K. and G. Brasseur, A three-dimensional model of chemically active trace species in the middle atmosphere during disturbed winter conditions, J. Geophys. Res., 94, 16387--16403, 1989.
Sabutis, J.L. et al.,Wintertime planetary wave propagation in the lower stratosphere and its observed effect on northern hemisphere temperature-ozone correlations, J. Geophys. Res., 102, 21709--21717, 1997.
Smith, A.K., Numerical simulation of global variations of temperature, ozone, and trace species in the stratosphere, J. Geophys. Res., 100, 1253--1269, 1995.
Solomon, S., Stratospheric ozone depletion: A review of concepts and history, Rev. Geophys., 37, 275--316, 1999.
Summers, M.E., and R.R. Conway, Insights into middle atmospheric hydrogen chemistry from analysis of MAHRSI observations of hydroxyl (OH), this volume, 2000.
Zhu, X., J.-H. Yee, and D.F. Strobel, Coupled models of photochemistry and dynamics in the mesosphere and lower thermosphere, this volume, 2000.