Meridional distributions of NOx, NOy, and other species in the lower stratosphere and upper troposphere during AASE II

.


Introduction
In the upper troposphere and lower stratosphere, NO and NO2 (NOx) play important roles in the production and destruction of 03.The region near the tropopause is one where 03 has its greatest effect as a greenhouse gas and where the lifetime of NOx is relatively long, so a given emission can have a larger impact.The global distribution of NOx in this region is only poorly known, however.Ehhalt et al. [1992] show The measurement of NOy (NO, NO 2, HNO 3, C1ONO 2, 2N20 5, HO2NO2, etc.) shows a significant lag in air with large gradients in HNO3 (an inlet problem), as when crossing the tropopause.The resulting errors can be rather large (-35%), but underestimates on ascent tend to cancel overestimates on descent, so the effect is reduced for the averages shown here.Altitudes are computed relative to the altitude of the tropopause, the height of which is determined at 14-s intervals along the flight track using remotely sensed temperature profiles, as described by Gary [1989].Data from this instrument are not available for the January flights, so those flights are excluded from the present analysis.

The Meridional Distribution of NO•
Figure 2a shows the meridional distribution of NOx as a histogram, explicitly illustrating the binning used and also the number of 1-min NOx samples in each bin.In order to have an adequate number of samples in each bin, and at the same time preserve altitude resolution, the bin dimensions are 15 ø in latitude by 1 km in altitude.The midpoints of the bins are given on the horizontal axes of the histogram.For the 30ø-45øN latitude bin, most of the samples are closer to 45øN than 30øN.
As the numbers of samples on the individual columns indicate, the sampling is extensive above the tropopause (except for the lowest latitude bin), but limited below.Nonetheless there is consistency among the altitude profiles for the different latitude bins.In descending from the stratosphere to the troposphere, there is a decrease in NOx in all four latitude bins immediately below the tropopause.With further descent there is a marked increase in the middle two latitude bins, spanning 45ø-75øN, with a slight increase for the 30ø-45øN range, and little change with altitude for the 75 ø-90ON range.For the two latitude ranges with data in the lowest altitude range, there is a marked decrease in NOx at the lowest level.This suggests the presence of a layer of NOx peaking at 2-3 km below the tropopause (see Figure 1 for the longitudes sampled).However, an examination of the individual flight profiles reveals that this is not a consistent feature and is primarily due to two events (discussed later) during the 10 flights.Thus the appearance of a layer at 2-3 km below the tropopause should be viewed as reflecting the occasional occurrence of relatively high values there, but no significance is attached to the exact shape, position, and latitudinal dimension of the apparent layer.Although the elevated NOx may be intermittent, it occurs frequently enough and with high enough values (at least in this sample) to have an impact on   Unlike the maximum at 6-7 km, the transition in NOx at the tropopause (Figure 2) is "smeared out" and is not readily apparent when altitude relative to the ground is used (Figure 3).There is enough variation in the height of the tropopause that the low values just below the tropopause are distributed over several bins of altitude above the ground, causing the transition to be lost in that representation.In the stratosphere at higher latitudes (50ø-90øN), the NOx/NOy ratio increases from -3% to -8% as the tropopause is

Meridional Distributions of Carbon Species
The high levels of NO x and NOy in the troposphere are correlated with high levels of CO (Figure 6) and C2C14 (Figure 7).Similar layers appear for each of these species, as well as for a number of other reactive carbon species (not shown), including

Summary
Meridional distributions of NOx in the lower stratosphere and upper troposphere are presented within the context provided by distributions of other species.In the lower stratosphere, there is no significant altitude gradient but a substantial latitude gradient (Figure 2b).In the upper troposphere, variability complicates the picture.For background conditions, that is, in the absence of recent pollution-laden convection (though not necessarily without such impact in the longer term), the pattern is similar to that in the lower stratosphere with little systematic variation with height, but a latitude gradient from 30øN to 90øN (Figure 2c).One difference is that there is a modest gradient at the tropopause, leading to smaller mixing ratios in the troposphere at a given latitude.In frontal regions (two cases), convection leads to higher values within a few km of the tropopause.Although sporadic, these values occurred with sufficient frequency and magnitude to have a significant impact on the mean values in AASE II (Figure 2b vs. Figure 2c).
NOx and NOy, 5 s for CO).For the carbon species grab samples, the integration times are dependent on altitude, but typically in the range 20-120 s.The chemical measurement techniques are described elsewhere (in Walega et al. [1991] and Weinheimer et al. [1993] for NO x and NOy, in Anderson et al. [1993] and references therein for the reactive carbon species).At low mixing ratios of NO and NO2 the overall uncertainty in 1-min NOx values is 20 pptv (including bias and precision errors).The percentage error decreases with increasing NOx, and it is -35% for the highest values shown.

Fig. 1 .
Fig. 1.Combined flight track of the DC-8 for the 10 flights in February and March 1992.Points plotted with an 'x' are those at altitudes between 2 and 3 km below the tropopause, the level of the elevated mixing ratios.

Figure
Figure 2c is the same as Figure 2b, except it excludes data during the two events principally responsible for the apparent layer 2-3 km below the tropopause.Those high values were obtained during the final descent into Stavanger (Norway, 59øN, 6øE) on 920214 and during the final descent into Bangor (Maine, USA, 45øN, 69øW) on 920217.Winds from a mesoscale model (J.-F.Lamarque, personal communication) for 920214 demonstrate two potential sources of polluted surface air to account for the high values measured: (1) Back trajectories show transport of surface air from southern England to aircraft altitudes (-6 km) near Stavanger in approximately 48 hours.(2) There is also, at the time of measurement and slightly before, convergence in the vicinity of Stavanger, indicating the possibility of local convection.Thus the high values are possibly due to either advective/convective transport from England or convective transport over Norway.On 920217 there was a front oriented east-west near the U.S.-Canada border [NOAA, 1992] as the southbound DC-8 approached Bangor through this area at an altitude of 6 km.This, combined with the presence of precipitation in the region [NOAA, 1992], suggests that the source of the elevated species abundance was convection of polluted boundary layer air.Exclusion of these two periods gives a representation (Figure 2c) of winter background conditions (not strongly influenced by recent pollution-laden convection) at the latitudes and longitudes flown.After exclusion the standard deviations of the 1-min values in the grid boxes range from 3% to 92% of the respective means; the mean is 47%.Prior to exclusion of the high values, the maximum was greater than 200%.

Figure 3
Figure 3 shows the 1-min NOx measurements of Figures 2a,b (high values not excluded) binned and contoured in a similar manner, except using altitude relative to sea level.This facilitates comparison with the NO distributions in Ehhalt et al. [1992] and Wahner et al. [in preparation], and with the NOx profile in Carroll et al. [1990].The NOx inferred from the measured NO is expected to have a distribution similar to that for NO, as demonstrated for the summertime measurements [Ehhalt and Drummond, 1988], so it is meaningful to compare the NO from those studies with the NOx from AASE II.

Figure 4
Figure 4 shows the meridional distribution of NOy relative to the height of the tropopause.There is little variation with latitude, and from 1 km below the tropopause to 3 km above, there is a uniform altitude gradient of-600 pptv/km.In the troposphere, there is a layer 2-3 km below the tropopause, similar to the apparent layer in NOx.Interestingly, the NOy layer does not completely disappear when the two high-NOx periods are excluded (not shown).The greater persistence of this feature for NOy may reflect its longer lifetime.The region of high NOy extends to lower altitudes over the pole which is due to the polluted air seen in a profile measured at the pole on 920314.

Figure 5
Figure 5 shows the meridional distribution of the NOx/NOy ratio (high values not excluded).In the stratosphere, NOx/NOy increases in going from high to low latitudes, similar to NOx.Since there is little latitudinal gradient in NOy, the increase in NOx is due to a change in the NOy partitioning, one factor being the increased photolysis of HNO3 at lower latitudes.Over the pole at 2 km above the tropopause, NOx is 50 pptv and NOx/NOy is 3%.Toward 40øN, values approach 120 pptv and 6%.
4. The presence of C2C14 indicates an industrial origin, and CO a combustive origin, of the elevated NO x and N Oy in the troposphere.Moreover the relatively short lifetime of C2H4 (perhaps 2-4 days for these conditions) indicates the freshness of the emission.While CO and NOx are known to be emitted by aircraft, C2C14 is unlikely to be, so the earth's surface is the likely source of the emissions.This is corroborated by the meteorology discussed earlier for the two days that contribute most to the elevated NOx.