Method and Observations

Three flights made early in the POLARIS campaign are examined to ascertain if there is any evidence of the polar sunrise ozone depletion phenomena. This is done through an examination of the BrO columns below the aircraft. The flights used are 26 April, 2 May, and 6 May 1997. Again, all flight tracks and altitude profiles are given in Appendix D. The albedo for these flights was fairly constant between 0.5 and 0.7 and is consistent with snow and ice as the reflecting surface. If substantial clouds were present, the albedo would experience a larger variability. For all ensuing forward model calculations, a thin aerosol layer representative of snow and ice was used between 0 and 3 km. The method is straightforward: BrO inside the PBL is estimated to determine if the levels are enhanced over the expected background amounts of 2 pptv (actually an upper limit) (Berg et al., 1983). The height of the PBL throughout this period has been measured to be in the 0.3-0.8 km range (S. Gaines, personal communication, 1998; J. Davies, personal communication, 1998), but a conservative height of 1 km is adopted here. Due to the large VCDs of O₃ below the aircraft during the spring at these latitudes ($\sim 200$ DU), it is not possible to detect the small depletion signature, at most 4 DU, over the background.

There is no straightforward method of converting nadir ACDs into PBL mixing ratios. However, by making conservative estimates (or upper limits) concerning the amount of BrO in the free troposphere and stratosphere, a lower limit to the BrO PBL mixing ratio can be deduced. The measured ACD is broken up into components from three layers,

$$\rm F_{nadir} = \rm\delta_{nadir} (C_1 + C_2 + C_3)$$ (11.47)

such that C₁ is the BrO VCD in the PBL (0-1 km), C₂ the BrO VCD in the free troposphere (1-8 km) and C₃ is the VCD between the tropopause and height of the ER-2 (8- $z_{\rm ER2}$ km). To improve the accuracy of these calculations, air mass factors are calculated separately for each layer so that $\delta_{\rm nadir}\rightarrow\delta_{\rm nadir,i}$. Solving for C₁ gives,

$$\rm C_1 = \rm\frac{F_{nadir} - \delta_{nadir,2} C_2 - \delta_{nadir,3} C_3} {\delta_{nadir,1}}.$$ (11.48)

An estimate of ${\rm C_3}$ can be obtained from Section 6.5.4 in which a BrO VCD of 2.3 $\times 10^{13}$ cm^-2 between 12 and 19.5 km was retrieved from limb measurements made on 6 May 1997 (see Table 6.10). Scaling this for 8-20 km results in a VCD of 3.7 $\times 10^{13}$ cm^-2. To ensure that the stratospheric component has not been underestimated, ${\rm C_3}$ is taken as twice this. The free tropospheric column is estimated by using a constant mixing ratio of 3 pptv which gives ${\rm C_2}=2 \times 10^{13}$ cm^-2 from 1-8 km.

In addition to assuming the enhanced BrO resides totally within the PBL, and for reasons which will be made clear below, three other scenarios are investigated in which some of the enhanced BrO is present in the free troposphere. These scenarios are: enhanced BrO between 0 and 5 km (C₁: 0-5 km, C₂: 5-8 km); enhanced BrO between 1 and 5 km (C₁: 1-5 km, C₂: 5-8 km); and 50 pptv between 0-1 km with the remainder placed between 1-5 km (C₁: 1-5 km, C₂: 5-8 km, C $_{1a}=1.4\times10^{14}$ cm²: 0-1 km).

Flights made on 2 May and 6 May 1997 are examined first with summaries of the analyses given in Figures 6.28 and 6.29, respectively. The first panel in each shows a time series of both the total ACD and the tropospheric (0-8 km) ACD. The total ACD is roughly $3\times10^{14}$ cm^-2 throughout much of both flights. Subtracting the estimated stratospheric component, roughly a constant over both flights, amounts to a reduction of about 25%, although this is quite variable. In fact, the tropospheric ACD becomes very close to zero in places, which could indicate that there is little tropospheric BrO but this could also mean the stratospheric component has been overestimated.

Panels (b) through (d) in Figures 6.28-6.29 are scatter plots of the BrO ACD as a function of latitude, altitude, and solar zenith angle, respectively. There appears to be no significant correlation with any of these parameters. This lack of correlation suggests that the bulk of the BrO must be in the troposphere as the tropospheric air mass factors are only weak functions of solar zenith angle and altitude of the ER-2. Air mass factors for stratospheric absorbing layers change much more rapidly with both SZA and ER-2 altitude, as was observed from Figure 6.11 and Table 6.4.

Panels (e) and (f) of these same Figures show the vertical column density and mixing ratios which result from the three of the four adopted scenarios (the fourth is omitted in these plots). A summary of these scenarios are given in Table 6.5 including

 

  

Figure 6.29: As Figure 6.28 but for the 6 May 1997 flight.

 

Table 6.11: Summary of nadir DOAS analysis of BrO vertical columns and mixing ratios for four different scenarios (see text for further details).

BrO Layer	Flight Date	Typical Mixing	Line Coloura
(km)		Ratio (pptv)
0-1	26 April 1997	95	black
	2 May 1997	12
	5 May 1997	20
1-5	26 April 1997	30	red
	2 May 1997	7
	5 May 1997	5
0-5	26 April 1997	25	green
	2 May 1997	5
	5 May 1997	4
1-5 (after placing	26 April 1997	15	blue
50 pptv from 0-1)

^a See Figures 6.28-6.30.

`typical' mixing ratio values for each flight and scenario. With all the enhanced BrO confined to the PBL, mixing ratios vary from 0-35 pptv on 2 May and 0-30 pptv on 6 May with averages of about 12 and 20 pptv, respectively. These values are similar to previous BrO mixing ratios reported (Hausmann and Platt, 1995). Examining results from the other scenarios, mixing ratios are observed to vary between 0 and 13 pptv for the 0-5 km case and about 2 pptv higher for the 1-5 km case. If 2 pptv is used as an upper limit for background BrO levels, significant portions of these flights still indicate BrO enhancement. There is no vertical information available to discriminate between the different scenarios and so on the basis that strong temperature inversions usually restrict mixing with free tropospheric air, the 0-1 km scenario is most likely representative.

 

Results from the third flight, made on 26 April 1997, are shown in Figure 6.30. The measured ACDs were 5 or 6 times larger than observed from the other two flights. Throughout the 26 April flight, the ACD hovers near 10¹⁵ cm^-2 and with a smaller (relative) variation. In this case, subtraction of the stratospheric ACD amounted to less than a 10% decrease. There is also no apparent relationship with latitude, SZA, and altitude of the ER-2, again suggesting that the large majority of the BrO is tropospheric. The VCDs and mixing ratios are shown for all four scenarios. The 0-1 km scenario produced mixing ratios between 70-120 pptv. However, upon allowing some of the BrO to reside in the free troposphere, well mixed up to 5 km, mixing ratios are reduced to 15-30 pptv.