Analysis of Halogen Occultation Experiment HF versus CH4 correlation plots: Chemistry and transport implications

,


Introduction
Several chemical species in the stratosphere are regarded as long-lived tracers because the timescales for their chemical sinks or sources are usually longer than the timescales for atmospheric dynamical processes.The mixing ratios of those trace gases, for example, can therefore be used as labels of moving air parcels for a time period shorter than the chemical timescale or diffusive mixing timescale.In fact, satellite observations [Jones and Pyle, 1984;Kumer et al., 1993;Russell et al., 1993a, b]  The interrelationship between long-lived tracers has been a very powerful tool used in studies of polar ozonehole related transport and chemical processes.Their correlation curves have been used to infer unavailable measurements or to identify chemical perturbations, such as denitrification and ozone loss [Fahey et al., 1990;Profrill et al., 1990].A commonly used assumption in such studies is that the correlation between tracers is universal.By this it is meant, for example, that it is reasonable to assume that observed tracer correlation curves in the midlatitude are also applicable in the polar vortex and any departure from the curve would be due to perturbed chemistry.Although some tracer fields in the lower stratosphere are indeed very close to "slope equilibrium" and their VMR correlations should merge to a universal nearly compact curve, it should be borne in mind that air in the polar region and near the ozone hole has been isolated from the rest of the atmosphere for a long season and it originated from a higher altitude [Russell et al., 1993a] where chemical processes may not be negligible.Hall and Prather [1995] studied N20-O3 correlations for a modeled vortex with only normal gasphase chemistry.Their results showed that the correlation curves of the two species in the lower stratosphere inside the vortex depart from that of outside, which suggests less ozone would be found on a N20 surface across the vortex boundary.In this pa-per we will examine the correlation curves for HALOE simultaneously measured CH4 and HF in the polar vortex and compare them with those outside the vortex.HALOE data clearly show that CH4-HF correlations follow different curves for measurements made inside and outside the vortex.It is therefore considered to be a better tracer than the others for use in studies of dynamical processes in the upper stratosphere.As pointed out by Mahlman et al.

HALOE Observations of CH4 and HF
[1986] and Holton [1986], chemical effects would tend to fiatten the mixing ratio surfaces of long-lived tracers.We expect that the isopleth (or isosurfaces) for CH4 VMR graphed as pressure versus latitude would be steeper compared to other tracer fields, such as HALOE HF.
Plate i shows HALOE CH4 pressure versus latitude cross sections for the same four time periods in 1993 used by Luo et al. [1994] for the HF fields.These four periods are basically representative times for northern winter, spring, summer, and fall.We will not repeat the detailed discussions on implications of the seasonal characterized structures of the CH4 field as we did for HF.The HALOE-observed CH4 and HF global patterns are very similar.For example, both CH4 and HF fields show a pronounced "double-peak" structure in the April-May midstratosphere and a less pronounced "double-peak" in October-November; the two fields show similar tiltings of their tropical minimum and maximum regions in the solstice seasons; as the result of strong mixing by waves, both tracer VMR surfaces show a relatively fiat area (the "surf zone") in the midlatitude October-November period; and distinct vortex-descent features are shown in both CH4 and HF in the spring Antarctic region.The similarities in HALOE-measured CH4 and HF global patterns indicate that HALOE measurements are internally consistent and that our current understanding of fluorine chemistry in the stratosphere is roughly correct.HF is indeed a long-lived trace gas and HALO E global observations of HF provide us with useful data for studies of stratospheric transport and its coupling with the halogen chemistry.
Although the similarity in meridional structures of tracers can be qualitatively explained by seasonally dependent dynamical processes, detailed observationmodel comparisons reveal problems in the understanding and in the model treatment of realistic dynamical processes and their coupling with chemical processes.The correlations between chemical tracers, in-cluding model-simulated hypothetical tracers, are powerful tools in diagnosing the effect of interactive processes.As a step toward quantitatively using HALOEmeasured CH4 and HF fields, we will examine the interrelationship between these two simultaneously measured species, which we believe will reveal the differences between the two very similar fields and the role of stratospheric chemistry and transport in determining their distributions and relationship.We examine here one day of HALOE observations in which profiles for both inside and outside the vortex are available.Figure 6  Figure 7 shows HALOE-observed CH4 versus HF in September-November 1992 for the southern hemisphere on the 650 K potential temperature surface.We found that the data seem to follow a compact curve.However, comparing to our collections of global CH4-HF curves,

Global
and model simulations [e.g., and Remsberg, 1993].While observations have very limited latitudinal and seasonal coverage, model results seem to suggest that a "universal" compact, nearly linear correlation between very longlived tracers does exist, such as N20 versus CH4 and N20 versus CFC-12 for example.Some models also show a slight latitudinal departure of the correlation curves.HALOE observations provide long-term nearly global measurements of CH4 and HF and the data are of excellent quality.In this paper we will study the correlations between HALOE-observed CH4 and HF and compare them with NCAR two-dimensional model resuits [Brassear et al., 1990] which provide global distributions of stratospheric tracers.The model however, is not able to simulate polar vortex conditions.
Since launch in September 1991, the HALOE instrument on U ARS has been operating essentially without flaw.This solar occultation instrument consists of eight optical channels measuring atmospheric absorption features by various stratospheric gases and provides retrieved volume mixing ratio profiles of a number of key chemical species at every spacecraft sunrise and sunset event.A detailed description of the instrument and its measurement coverage can be found in the work of Russell et al., [1993b].The measurement line-of-sight tangent point of HALOE moves gradually between the south and the north and scans nearly the whole globe in about a month.The accumulated global data set then can be used to create a zonal mean pressure versus latitude cross section for the volume mixing ratio of the species measured.The HALOE HF and CH4 channel validation papers [Russell et a/.,1995; Park et a/.,1995] provide further information on data quality and comparisons with correlative measurements.In general, HALOE HF observations agree with correlative balloon underflight observations to within 7% or less throughout the stratosphere above the 70 mbar level.The precision (repeatability) of the measurements is _•0.04 part per billion by volume (ppbv) over the range from the tropopause to the stratopause.The estimated CH4 error (random plus systematic) is of the order of 7% over the altitude range from 12 to 40 km.The precision is _•0.05 part per million by volume (ppmv) from 25 to 75 km degrading below 25 km to 0.1 ppmv at 16 km.Luo et al. [1994] described HALOE-measured stratospheric hydrogen fluoride (HF) in detail, including its zonal mean meridional structures for different seasons, its global column amounts and comparisons with previous measurements, and results from the NCAR twodimensional model.HALOE has provided the first global observations of HF, the dominant reservoir species for fluorine released from man-made CFCs.HF is believed to be very inactive chemically with no known chemical (photochemical) removal processes in the stratosphere.The conversion of the fluorine in CFCs to HF in the middle-high stratosphere leads to a monotonically increasing HF field with altitude.Its seasonally dependent global distributions, as described by Luo et al. [1994], show a very similar pattern to the long-lived tracers, such as CH4 and N20 observed by the Stratospheric and Mesospheric Sounder (SAMS) and CH4 observed by HALOE itself.The production timescale for HF is equivalent to the timescale for CFC dissociation.This species is therefore a good dynamical tracer in the stratosphere, as discussed, for example, by Brasseur and Solomon [1986].The tropospheric source gas CFCs have roughly a few percent annual increase rate, and the observations of stratospheric HF and CF20 also show similar rates of increase with a few years delay time [Zander et al., 1994; Luo et al., 1994].The total fluorine amount in different regions in the stratosphere is predicted to be different depending on the average "age" of the air.Simultaneous measurements of fluorine-containing species globally should therefore provide useful data in studies of air exchanges in different regions of the atmosphere.The HF distribution itself, however, is hard to use as an indicator of the air age because the fluorine partitioning among F-containing species depends on altitude, events are usually located at two latitude circles and they gradually sweep between south and north.The measurement tangent point passes the equator about 15-20 times annually or at least once a month.HALOE observes each polar region (600 latitude and poleward) in three time periods every year.In addition to a summer month when observations are limited to below 700 latitude, there are two time periods in each hemisphere when the observations reach nearly 80 o latitude; i.e., mid-March to the end of April and mid-August to mid-September in the Arctic and mid-February to mid-March and the end of September to end of October in Antarctica.The global coverage of HALO E measurements allows us to survey the CH4 and HF relationship for different latitude bands and seasons.We found an interesting shift of CH4-HF correlations for data taken inside the polar vortex relative to those measured outside.The discussion on the vortex tracer correlation will be in the next subsection.In this subsection we focus on the HALOE global CH4-HF relationships without considering the polar vortex and compare with results from the NCAR two-dimensional model, which does not simulate the polar vortex.Examinations of HALOE simultaneously measured stratospheric CH4 and HF for different latitude bands and seasons show that in general, measurement pairs in CH4-HF scatterplots fall onto a compact curve for a given latitude band.The tropical correlation curve and the high-latitude correlation curve define the envelope of all the possible correlations.Figures la, lb and lc show the CH4-HF scatterplots for the tropical region for the entire year of 1992, mid-February to mid-March 1992 southern polar region and mid-August to mid-September 1992 northern polar region.The two available polar region measurements were taken during the late summer in each hemisphere, before the polar vortices formed.Methane and HF are indeed observed to have simple relationships.In the tropical and high southern latitude regions the correlation curves seem slightly bent compared to the curve in high northern latitudes which is nearly linear.Since most chemical activity (removal of CH4 and formation of HF) occurs in the tropicallow latitude region, the slope of the CH4-HF curve is mostly determined by the relative effects of CI-Iq re-HALOE CH4 MIXING RATIO SS JUL.I-AUG.4,1993 (d) HALOE CN4 MIXING RATIO SS OCT.11 -NOV.20,199•5 0.1 l"•-'•'•-l"rq f'"i'"'"•'• •---i -• ]'•'t q •-•-I-'• • ; I :.]';I •'ratio of CH4 removal to HF production changes at about 6 mbar (CH4 = ~ 1.0 ppm from Plate 1).The observed CH4-HF correlations globally are found to lie in between the tropical bent curve and high-latitude curves depending on the competitive effects of seasonal mixing processes and advection.For example, the CH4-HF correlation for CH4 <0.8 ppm in the February to March southern polar region (Figure lb; also see Fig-ure 2) is found to bend slightly toward and merge to the tropical curve (Figure la) compared to that in August to September northern polar region (Figure lc).During the 1992 southern late summer, relatively larger variability in HALO E-observed tracer profiles at southern high latitudes is found than that shown in data of northern late summer high latitudes (Figures lb versus lc), implying differences in wave mixing processes which would draw tracer correlations at high latitudes to near that of tropics.It is worth mentioning that at the top and bottom pressure levels defined in Figure1, measurements in the tropical region and at high latitudes show differ-CH4 (and HF).For instance, tropical CH4 at 100 mbar is found to be greater than 1.6 ppm, while 100 mbar CH4 at polar latitudes is observed to be less than 1.6 ppm.At 0.5 mbar, data in polar regions show smaller CH4 values than those in the tropics.Figure l c also indicates that when CH4 is <~0.2 ppm, HF becomes constant at ~ 1.2 ppb in the 1992 mesosphere, which was perhaps nearly the total inorganic fluorine amount at that level.Note that atmospheric concentrations of HF and CH4 are increasing with time [World Meteorological Organization (WMO), 1992;Gunson et  al., 1994], so that future measurements of these species will presumably reveal higher VMRs.Figure2shows the averaged CH4-HF correlation curves for data in Figure1.The high-latitude curves for the two hemispheres are shown in Figures la and lb with the tropical curve plotted for comparison.It is found that in the very low (near 100 mbar) and very high stratosphere (near 0.5 mbar) the scattered points of CH4-HF for tropical and polar regions nearly coincide.In the middle stratosphere the ranges of CH4-HF points for tropics and polar regions are clearly separated.For example, a constant CH4 mixing ratio surface (e.g., CH4 -1.0 ppm) would correspond to a larger value of HF in the tropics (e.g., HF -0.654-0.05ppb) but a lower value of HF in polar regions (e.g., HF = 0.454-0.05ppb); this is more than a 35% change in HF.

Figure 2
Figure 2 and the above statement indicate that in the stratosphere, CH4 constant VMR surfaces (its isopleths or contour lines) have steeper slopes than HF although this is not easily seen by visually comparing the pressure versus latitude cross sections for CH4 (Plate 1) and HF [Luo et al., 1994, Plate 1].HALOE results of global CH4-HF correlations agree with theoretical analyses which state that the vertically stratified tracer having the longer chemical lifetime will have steeper VMR isopleths than the tracer with the shorter chemical timescale [Mahlman et al., 1986].A model simulation of CH4-HF correlations is shown in Figure 3.The NCAR two-dimensional model treats stratospheric radiative, dynamical, and chemical processes interactively.More detailed descriptions about the model can be found in the work of Brasseur et al. [1990], and the fluorine chemistry employed is described by Luo et al. [1994].Qualitatively, model CH4-HF correlations agree very well with HALOE observations in Figure 2. Model CH4-HF correlations at high latitudes

Figure 5 .
Figure 5. Scatter diagram for HALOE CH4 versus HF points from HALOE simultaneously measured volume mixing ratios of CH4 and HF for latitudes 65øS and poleward in the time period September 26 to October 24 1992 (data version 17).The pressure range is between 100 mbar and 0.5 mbar.The two curves are from Figure 2a, the averaged tropics CH4-HF correlation, and the averaged polar region correlation observed in February- Figure 7. HALOE-measured CH4-HF scatter diagram for data sampled on the 650 K potential temperature surface in the southern hemisphere between September 8 and November 11, 1992.The solid curve is the averaged CH4-HF correlation for entire year 1992 between 10øS and 10øN latitudes (Figure 2).The dashed curve is CH4-HF for event 19 (green in Plate 2) inside the vortex on October 10, 1992 (pressure levels 100-15 mbar).