Balloon-borne measurements of temperature, water vapor, ozone and aerosol backscatter on the southern slopes of the Himalayas during StratoClim 2016–2017

4.1 Seasonal variability

Figure 2 illustrates the seasonal variability of the meteorological systems above the southern slopes of the Himalayas during NT16_AUG (top row), NT16_NOV (center) and DK17 (bottom row). The left column (Fig. 2a, c, e) shows latitude–pressure cross sections of potential vorticity (PV), potential temperature and total wind speed for NT (longitude 80^∘ E, Fig. 2a, c) and DK (85^∘ E, Fig. 2e), from ECMWF analysis data averaged over the time of the respective measurement periods. The right column (Fig. 2d, b, f) shows the average geopotential height at 100 hPa for the three measurement periods from ECMWF analysis data, superimposed with 2-week backward air mass trajectories initialized at 100 hPa at the time of each sounding during the three campaign periods. Trajectories are calculated using the Lagrangian analysis tool (LAGRANTO) (Wernli and Davies, 1997) using ERA-Interim reanalysis wind fields that are color-coded by pressure.

Figure 2(a, c, e) Latitude–pressure cross sections of potential vorticity (color scale), potential temperature (black contours, in K) and total wind speed (blue contours, in m s⁻¹) from the ECMWF analysis data (horizontal resolution O1280 interpolated to a ${\mathrm{0.125}}^{\circ }\times {\mathrm{0.125}}^{\circ }$ grid, vertical resolution L137, 6-hourly) averaged over the NT16_AUG (a, longitude 80^∘ E), NT16_NOV (c, longitude 80^∘ E) and DK17 (e, longitude 85^∘ E) campaign time periods, as given in Table 2. Black dashed lines show the latitude of NT (a, c) and DK (e). (b, d, f) Geopotential height at 100 hPa from ECMWF analysis data averaged over the NT16_AUG (b), NT16_NOV (d) and DK17 (f) campaign time periods (color scale), and 2-week LAGRANTO backward trajectories calculated along ERA-Interim wind fields (horizontal resolution T255 interpolated to a 1^∘ × 1^∘ grid, vertical resolution L60), initialized at 100 hPa at the time of each balloon sounding in NT16_AUG (b), NT16_NOV (d) and DK17 (f), color-coded by pressure. Black dashed lines show the longitude of NT (a, c) and DK (e). The white dashed rectangle in (d, f) shows the approximated ASMA area used for the confined fraction calculation (10–50^∘ N, 0–140^∘ E; see Sect. 5.2). Note that, for NT16_NOV, trajectories started 6 h before and 6 h after each sounding are also displayed (d).

Download

Figure 3Time series of temperature (a, c, e, g) and the water vapor mixing ratio (b, d, f, h) as a function of pressure, from ECMWF operational analysis data (6-hourly, same horizontal and vertical resolution as given in the caption of Fig. 2) for NT (c, d, g, h) and DK (a, b, e, f) from 1 to 31 August 2016 (a–d) and from 20 July to 21 August 2017 (e–h).

Download

During the NT16_AUG and DK17 campaigns (Fig. 2a, b, e, f), our stations were located inside the ASMA vortex. The continental-scale anticyclonic motion is confined by the subtropical westerly jet stream to the north (40–45^∘ N) and the equatorial easterly jet to the south (10–15^∘ N). Both NT and DK were found in the average geopotential height region exceeding 16.8 km at 100 hPa during their respective measurement periods, which was the highest in the ASMA. Backward trajectories show that the UTLS flow on the southern slopes of the Himalayas is mainly easterly, following the southern branch of the ASMA, and transporting air masses which had already been confined inside the anticyclone for several days. During the 2 weeks prior to their respective measurement, the air masses sampled at 100 hPa during our campaigns had undergone net diabatic ascent at an average rate of 0.7 and 0.4 K day⁻¹ (in isentropic coordinates) for NT16_AUG and DK17, respectively. The equatorial easterly jet was slightly stronger and extended further north during the ASM season in 2017 (Fig. 2e) compared with 2016 (Fig. 2a); this is also reflected by the corresponding backward trajectories (Fig. 2b, f). North of the subtropical westerly jet, the dynamical tropopause (corresponding to PV = 3–4 PVU in this region and season: Kunz et al., 2011) decreases steeply with altitude.

After the end of the monsoon season, the subtropical westerly jet migrates southward to 30–35^∘ N and intensifies in strength. Therefore, during the post-monsoon campaign NT16_NOV (Fig. 2c, d), our station was located below the jet stream and the associated tropopause break, resulting in the large LRT–CPT separation discussed in Sect. 3. In contrast to the ASM season, in November the UTLS winds on the southern slopes of the Himalayas are mainly westerly and follow the subtropical jet stream. This is consistent with wind speed and wind direction measurements by RS41 shown in Fig. S3 in the Supplement. We also note that the large standard deviation of the UTLS temperature in NT16_NOV (Fig. 1d) is likely related to the varying meridional position of the jet stream during the measurement period.

The dynamical features discussed above are consistent with the seasonal variations of the ITCZ, the jet streams and the ASM system in general, which are extensively discussed in previous literature, e.g., Lawrence and Lieleveld (2010), Munchak and Pan (2014), Ploeger et al. (2015), Garny and Randel (2016), Pan et al. (2016).

4.2 Interannual and regional variability

To assess whether the differences between the observations in NT16_AUG and DK17 are mainly caused by geographic difference, and different associated mesoscale weather features, or by interannual variability between the ASM in the 2016 and 2017 seasons, Fig. 3 shows time series of the UTLS temperature (left column) and the H₂O mixing ratio (right column) from ECMWF analysis data for both stations and both campaign periods.

In August 2016 (Fig. 3a–d) the UTLS was relatively warm at both locations, with CPT temperatures rarely dropping below −80 ^∘C (DK was slightly colder than NT – 0.4 K on average at 100 hPa), and the H₂O mixing ratio also never dropped below 4.5 ppmv at either site. The same day-to-day variability occurred at both locations with a time shift of about 6–12 h, which is consistent with DK being upstream of NT along the southern branch of the ASMA, and with wind speeds of around 20 m s⁻¹ in the UTLS. In July–August 2017 (Fig. 3e–h), temperatures and H₂O values and features were similar to 2016 until 3 August. Then, a period characterized by an extremely cold and dry tropopause began in both NT and DK, peaking between 7 and 10 August with a CPT colder than −83 ^∘C and H₂O mixing ratios lower than 3 ppmv. The minima were slightly more pronounced in DK than in NT but were correlated in time, suggesting that these features are related to a large-scale cooling and drying pattern occurring in the ASMA. We also note that a layer of high H₂O rises to high altitudes (70–85 hPa) after 3 August (Fig. 3f), forming the local maximum above the CPT which was also found in our DK17 measurements (Fig. 1e). This feature is remarkable and not in accordance with more typical climatological conditions observed during NT16_AUG.

Based on Fig. 3, we argue that the differences between the NT16_AUG and DK17 datasets are not due to local meteorological effects, which appear to have a negligible impact on the UTLS temperature and H₂O at the two measurement sites. Rather, these differences are attributed to interannual variability, and in particular to a period of an anomalously cold and dry UTLS on the southern slopes of the Himalayas, which occurred after 3 August 2017 and persisted on a large scale.

5.1 Asian tropopause transition layer (ATTL)

In the tropics, the thermodynamic transition between the troposphere and the stratosphere occurs over a layer of several kilometers in thickness, known as tropical tropopause layer (TTL). This layer is influenced by both upper tropospheric and lower stratospheric processes, and its properties control water vapor transport through the tropopause (e.g., Fueglistaler et al., 2009; Randel and Jensen, 2013). Amongst several different definitions of the TTL used in the literature, reviewed by Pan et al. (2014), Gettelman and de F. Forster (2002) identify its boundaries based on the temperature and potential temperature lapse rates only, which is particularly suited to balloon-borne measurements. In their definition, the upper boundary of the TTL is the CPT, and the lower boundary is the lapse-rate minimum (LRM), i.e., the point in altitude where the change of potential temperature (θ) with altitude (z) is minimum. This defines the TTL as the layer in which the temperature lapse rate switches from convectively dominated in the troposphere (small dθ∕dz, low stability), to radiatively controlled in the stratosphere (high dθ∕dz, high stability; Gettelman and de F. Forster, 2002). In addition, the LRM coincides with the mean convective outflow level (e.g., Gettelman and de F. Forster, 2002; Vömel et al., 2002; Paulik and Birner, 2012). Given the similarity between the UTLS region on the southern slopes of the Himalayas during the ASM season and that of the tropics, in this study we adopt the abovementioned definition of the TTL to study the thermal structure of our NT16_AUG and DK17 datasets. However, as our measurement sites are not tropical in a geographical sense, we refer to the TTL in this region and season as the Asian tropopause transition layer (ATTL).

Figure 4Mean profiles (solid lines) and standard deviations (colored shading) of dθ∕dz as a function of pressure for NT16_AUG (blue) and DK17 (red). Horizontal dashed lines show the mean LRM and CPT for NT16 (blue) and DK17 (red). The ATTL regions for the two datasets are highlighted by grey arrows. Note that the mean profiles and standard deviations of dθ∕dz were smoothed with a ±5 hPa (about 250 m) moving average to reduce the noise contributions from the geometric altitude measurement by RS41.

Download

Figure 4 shows mean profiles and standard deviations of dθ∕dz as a function of pressure for the two ASM season datasets. The average LRM is found at lower pressure in DK17 compared to NT16_AUG (169.5 vs. 180 hPa), corresponding to a roughly 400 m altitude difference, and the minimum in dθ∕dz of DK17 is more pronounced (1.5 vs. 2 K km⁻¹). This suggests that convection reached higher altitudes, on average, in the ASM in 2017 compared to 2016 on the southern slopes of the Himalayas (at least during our measurement periods). The average ATTL boundaries in terms of pressure (potential temperature) are 180–97.5 hPa (360–382 K) for NT16_AUG and 169.5–88 hPa (362.5–383.5 K) for DK17 (see Table 3; note that the data are binned with respect to pressure and the given potential temperature values are the average θ in the pressure levels where the LRM and CPT occur). We also observe that, due to the colder temperatures, the isentropic levels in DK17 are shifted to lower pressures compared to NT16_AUG (see Fig. S4 in the Supplement), meaning that the large altitude difference between the two LRMs and CPTs in NT16_AUG compared to DK17 (400–600 m) becomes small in isentropic coordinates (1.5–2 K).

5.2 Confined lower stratosphere (CLS)

Since the anticyclonic circulation extends up to above the CPT, it is important for interpreting the observed vertical gradients of chemical species and aerosols to quantify the vertical extent of the ASMA in the lower stratosphere. Here we estimate the top of the horizontal confinement effect of the ASMA during the NT16_AUG and DK17 campaign periods by means of air mass trajectories. For this, we consider 2-week LAGRANTO backward trajectories initialized at 5 hPa intervals between 40 and 150 hPa at the time of each balloon sounding in NT16_AUG and DK17 (i.e., same as in Fig. 2b, f, for 100 hPa), and 6 h before and 6 h after each sounding, using ERA-Interim wind fields. For each pressure level, we calculate the “confined fraction” of trajectories, defined as the fraction of trajectories which were already located inside the anticyclone 2 weeks before the measurements. For this purpose, based on the average geopotential height fields shown in Fig. 2, we approximate the ASMA area as the box of 10–50^∘ N latitude, 0–140^∘ E longitude (see the white dashed rectangle in Fig. 2b, f).

Figure 5Confined fraction of trajectories (see Sect. 5.2) as a function of pressure for the NT16_AUG (blue) and DK17 (red) campaign periods. Dashed lines mark the TOC level for NT16 (blue) and DK17 (red) and the 50 % confined fraction threshold (black).

Download

Figure 6Schematics of the vertical structure of the UTLS above the southern slopes of the Himalayas. The Asian summer monsoon anticyclone (ASMA) consists of two layers, the Asian tropopause transition layer (ATTL) and the confined lower stratosphere (CLS). These layers are confined by three levels: the lapse-rate minimum (LRM, green surface), the cold-point tropopause (CPT, red) and the top of confinement (TOC, blue). Dynamical features and relevant transport processes discussed in the paper are also sketched. Approximated pressure, altitude and potential temperature levels of the TOC, CPT and LRM derived from NT16_AUG and DK17 measurements are displayed on the respective surfaces.

Download

Figure 7Mean profiles (solid lines) and standard deviations (colored shading) of temperature (a), the H₂O mixing ratio (b) and the O₃ mixing ratio (c) as a function of altitude relative to the CPT, for NT16_AUG (blue) and DK17 (red). Dashed lines show the CPT (black) and the average LRM and TOC levels for NT16 (blue) and DK17 (red) (see the labels in a). The four layers defined in Sect. 5.1 (troposphere, ATTL, CLS and free stratosphere) are identified using grey labels.

Download

Figure 8Probability density functions (PDFs) of temperature (a, d, g, j), the H₂O mixing ratio (b, e, h, k) and the O₃ mixing ratio (c, f, i, l), calculated in the free stratosphere (a–c), CLS (d–f), ATTL (g, h) and troposphere (j–l) altitude regions as defined in Sect. 5.1, for NT16_AUG (blue) and DK17 (red).

Download

Figure 5 shows the resulting confined fractions for NT16_AUG and DK17 as a function of pressure. In both campaign periods, the confined fraction is high (above 60 %) up to 70–80 hPa, while above this pressure level it sharply decreases to zero. Confinement is higher for DK17 than for NT16_AUG throughout the entire UTLS, which is qualitatively consistent with the backward trajectories shown in Fig. 2. Based on these curves, we define the top of confinement (TOC) as the level of confined fraction equal to 50 %, corresponding to 73 hPa in NT16_AUG and 63.5 hPa in DK17. This level separates altitudes that are affected by horizontal confinement in the ASMA (below TOC) from the confinement-free stratosphere above the ASMA (above TOC). Note that the mean altitude and potential temperature levels of the TOC derived from the balloon measurements are given in Table 3.

Following the definition of the TOC, we further define the confined lower stratosphere (CLS) as the region of altitudes above the CPT and below the TOC. The CLS is the layer of lower stratosphere which is subject to confinement in the ASMA, in contrast to the free stratosphere above the anticyclonic vortex (i.e., above TOC). Figure 6 illustrates the vertical structure of the UTLS above the southern slopes of the Himalayas during the ASM season according to the abovementioned definitions of the ATTL and the CLS. In the following, we refer to this framework of significant levels and layers to discuss the vertical distributions and variability of water vapor, ozone, ice saturation and aerosols in the ASMA.

Figure 9(a) Mean profiles and standard deviation of ice saturation (S_ice) as a function of altitude relative to the CPT, for NT16_AUG (blue) and DK17 (red). (b) The supersaturated fraction (i.e., fraction of measurements with S_ice > 1) as a function of altitude relative to the CPT. (c, d, e) PDFs of ice saturation in the CLS, ATTL and troposphere altitude regions, respectively.

Download

5.3 Water vapor and ozone

To analyze our H₂O and O₃ measurements in relation to the thermodynamic structure of the UTLS, for each balloon sounding we define the altitude relative to the CPT as a new vertical coordinate. Figure 7 shows mean profiles and standard deviations of temperature, the H₂O mixing ratio and the O₃ mixing ratio in this coordinate system (note that, besides the CPT in black, the mean LRM and TOC levels are shown by blue dashed lines for NT16_AUG and red dashed lines for DK17). Figure 8 shows the probability density functions (PDFs) of temperature (left column), the H₂O mixing ratio (center) and the O₃ mixing ratio (right column) calculated in the free stratosphere (Fig. 7a–c), CLS (Fig. 7d–f), ATTL (Fig. 7g–i) and troposphere (Fig. 7j–l) regions, for NT16_AUG and DK17. The PDFs of the free stratosphere region are calculated for altitudes between the TOC and CPT + 5 km, while the troposphere PDFs are calculated for altitudes between the CPT and 6 km and LRM, i.e., covering the whole range of altitudes (with respect to CPT) as shown in Fig. 7.

Figure 10Examples of thin cirrus clouds measured during four individual soundings of the NT16_AUG campaign (a–d) and two soundings of the DK17 campaign (e, f). Solid lines show temperature (black), ice saturation (light blue), BSR at 455 nm (BSR₄₅₅, blue), BSR at 940 nm (BSR₉₄₀, red) and the color index (green). Vertical dashed lines mark the S_ice = 1 (light blue) and the color index = 7 (green) thresholds, used for cloud-filtering (see Sect. 6.1). Sounding identification numbers are noted in black in the top-right corner of each panel; for the date, time and payload of each sounding, see Table S1 in the Supplement.

Download

The water vapor mixing ratio in DK17 shows a minimum at the CPT and a local maximum in the CLS (Fig. 7b), centered about 1 km above the local CPT (i.e., not the average CPT, but evaluated for each profile individually). The H₂O minimum is conceivably due to the unusually high frequency of the occurrence of ice clouds near the CPT in DK17, which depletes water vapor from the gas phase in favor of the condensed phase and results in a strongly dehydrated CPT (see Sect. 5.4). The isolated H₂O maximum in the CLS is consistent with hydration by overshooting convective updrafts; this process injects ice crystals above the CPT, which then evaporate and release localized “pockets” of moist air. Convective updrafts overshooting the CPT were observed by Corti et al. (2008), and a similar hydration mechanism was hypothesized by Dauhut et al. (2015, 2016).

In both NT16_AUG and DK17, the PDFs of the H₂O mixing ratio show higher water vapor in the CLS (Fig. 8e) compared to the free stratosphere (Fig. 8b). In particular, the PDFs in the CLS are broad (3–7 ppmv) and skewed towards high values, while the PDFs in the free stratosphere are narrow (3–5 ppmv) and show the expected distribution of background stratospheric water vapor. The high H₂O mixing ratios in the CLS in DK17 are obviously related to the previously discussed isolated maximum, yet the enhanced frequency of occurrence of high H₂O mixing ratios is also observed in NT16_AUG, despite no local maximum being found in this dataset. This is consistent with the slow ascent of moist convective outflow air within the confined anticyclone, and may, in part, reflect the decreasing frequency of overshooting convective tops with altitude in NT16_AUG.

The ozone mixing ratio in DK17 shows a minimum slightly above the LRM (Fig. 7c), which is characteristic of deep convection, rapidly transporting ozone-poor air from the boundary layer to the convective outflow level (e.g., Gettelman and de F. Forster, 2002; Vömel et al., 2002; Paulik and Birner, 2012). The absence of this feature in NT16_AUG suggests that the average age of air, meant as the time elapsed since the last interaction with deep convection, was higher in NT16_AUG compared with DK17, such that the O₃ minimum is smeared out by mixing and additional photochemical production (which is enhanced in the ASMA due to the enrichment in ozone precursors; Gottschaldt et al., 2018). The higher dilution of the convective signature in NT16_AUG vs. DK17 is also consistent with the absence of an H₂O maximum above the CPT in NT16_AUG (Fig. 7b), and with the higher frequency of the occurrence of low O₃ mixing ratios in DK17 compared with NT16_AUG in the ATTL and CLS (Fig. 8f, i).

In summary, both the NT16_AUG and DK17 datasets show evidence of deep convection extending into the CLS, i.e., up to 1.5–2 km above the CPT. Convective features, such as low O₃ in the ATTL and high H₂O in the CLS, are more pronounced in DK17 than in NT16_AUG, indicating that DK17 likely sampled fresh convective outflow more frequently than NT16_AUG. This is also consistent with the higher altitude of the LRM in DK17 compared to NT16_AUG (Fig. 4) and suggests that convective activity on the southern slopes of the Himalayas was more frequent during the ASM season in 2017 than in 2016.

Transport to the CLS is likely due to a combination of different processes, including overshooting convection, slow diabatic ascent and adiabatic transport from regions with a higher CPT potential temperature in the ASMA (discussed in Sect. 7). Although we do not quantitatively evaluate these processes, we argue that the horizontal confinement effect of the ASMA plays an important role in shaping the vertical distributions of H₂O and O₃ above the CPT, by keeping the moist convective outflow horizontally confined (while it continues to rise slowly) and thereby increasing the frequency of occurrence of air parcels with high H₂O and low O₃ above the CPT. This is supported by the fact that differences in H₂O and O₃ between NT16_AUG and DK17 vanish in the free stratosphere (Fig. 8a–c), which is in agreement with our trajectory-based definition of TOC. Further analysis will be required to disentangle the relevance of the different abovementioned transport processes in moistening the CLS.

5.4 Ice saturation

Figure 9 shows the mean profiles and standard deviations of ice saturation (Fig. 9a) and histograms of supersaturated fraction (Fig. 9b) as a function of altitude relative to the CPT, and PDFs of ice saturation calculated for the troposphere, ATTL and CLS regions (Fig. 9c–e). As a result of colder temperatures (see Fig. 7a), much higher and more persistent ice saturations were measured in DK17 than in NT16_AUG throughout the entire ATTL. In both datasets, the average ice saturation is higher in the ATTL than in the troposphere, and in DK17 it shows a pronounced maximum with respect to average supersaturated conditions (i.e., more than 50 % of the measurements reach S_ice > 1) in the 1.5 km directly below the CPT. In contrast, the supersaturated fraction is 10–15 % in NT16_AUG over the same range of altitudes. The ATTL ice saturations and supersaturated fractions of NT16_AUG are comparable with previous measurements from Lhasa and Kunming, China from 2009 to 2010 (Bian et al., 2012), while the measurements in DK17 range significantly higher. This suggests that the frequency of occurrence of cirrus clouds in the ASM season in 2017 was unusually high, which is likely the reason for the H₂O minimum at the CPT observed in DK17 (Fig. 7b). Interestingly, we also note that ice supersaturations in DK17 frequently extend into the CLS, with about a 30 % supersaturated fraction in the first 500 m above the CPT (Fig. 9b, c). This implies that, in overshooting convective updrafts, ice crystals can regularly penetrate the CPT as condensed phase (e.g., see Fig. 10f) and hence efficiently hydrate the CLS.

6.1 Cirrus clouds

Figure 10 shows individual soundings as examples of thin cirrus clouds observed during the NT16_AUG (Fig. 10a–d) and DK17 (e, f) campaigns. Along with the temperature and ice saturation profiles, we show BSR at 455 nm (BSR₄₅₅), BSR at 940 nm (BSR₉₄₀), and the color index (CI). The color index is defined as the 940-to-455 nm ratio of the aerosol component of BSR, i.e., CI = (BSR₉₄₀−1) ∕ (BSR₄₅₅−1). CI is independent of the number density; therefore, it is a useful indicator of particle size (e.g., Cirisian et al., 2014) as long as particles are sufficiently small, so that Mie scattering oscillations can be avoided. Based on the size-dependence of the CI, considerations regarding the typical size range of ice crystals and aerosol droplets, and the evaluation of ice saturation measurements by CFH, a threshold of CI = 7 was empirically developed to discriminate in-cloud (CI > 7) from clear-sky measurements (CI < 7) (Vernier et al., 2015). This helps discern the BSR features in Fig. 10 as either ice cloud or aerosol signal, and is also used as a threshold for cloud-filtering. For example, in sounding NT004 (Fig. 10a), the sharp feature at 145 hPa with a CI ≈ 10 is likely an ice cloud (note the concomitant ice supersaturation above the thin cloud layer, suggesting sedimentation), while the broad enhancement in BSR between 95 and 140 hPa without CI enhancement is the signal of the ATAL. The main common characteristics of the cirrus clouds in Fig. 10 is their very small spatial and optical thickness, with BSR₉₄₀ < 20, while much larger values (BSR₉₄₀ ≫ 100) are expected in homogeneously nucleated cirrus clouds, as often observed in the midlatitudes (e.g., Brabec et al., 2012; Cirisian et al., 2014), and as also shown by the thick outflow cirrus below 120 hPa in DK002 (Fig. 10f). Low BSR indicates low ice crystal number densities, suggesting that these clouds are most likely formed by heterogeneous nucleation on solid ice nuclei, rather than by homogeneous freezing of sulfate aerosol liquid droplets. This hypothesis is currently being investigated by a dedicated microphysical modeling study. Similarly thin cirrus clouds were observed in more than half of the COBALD soundings in NT16_AUG (9 out of 17); therefore, they occur very frequently in the ASMA, and they were often found embedded within the ATAL (Fig. 10a, e).

Figure 11All clear-sky (i.e., aerosol only) data points (dots) and mean profiles (solid lines) of BSR₄₅₅ as a function of pressure, for NT16_AUG (blue) and NT16_NOV (green). Black lines show the mean LRM (dashed), CPT (solid) and TOC (dashed) levels for NT16_AUG. The troposphere, ATTL, CLS and free stratosphere regions are identified using grey labels, and the ATAL region is identified using a black arrow.

Download

Figure 12Mean profiles of temperature (red), dθ∕dz (purple), the H₂O mixing ratio (blue), the O₃ mixing ratio (orange), ice saturation (light blue) and clear-sky aerosol BSR₄₅₅ (black) as a function of pressure (left axis, light grey dashed lines) for NT16_AUG (a) and DK17 (b). Average potential temperature levels are shown on the right axis and marked by dark grey dashed lines. Note that the pressure scale is the same for the two panels, and the potential temperature levels vary according to the measurements. The average CPT, LRM and TOC levels in the two datasets are marked by black solid lines. The ATTL and CLS layers are highlighted using light green and orange shading, respectively. Note that the mean profiles of dθ∕dz, ice saturation and clear-sky aerosol BSR₄₅₅ are smoothed with a ±5 hPa (about 250 m) moving average (same as in Fig. 4).

Download

6.2 ATAL during the ASM season 2016

Figure 11 shows all clear-sky (i.e., aerosol only) BSR₄₅₅ data points and mean profiles from the NT16_AUG and NT16_NOV datasets. Similarly to Vernier et al. (2015), the cloud-filtering criterion that we applied consists of three thresholds from two independent measurements: BSR₉₄₀ < 2.5 and CI < 7 from COBALD, and S_ice < 0.7 from CFH. Only data points which simultaneously fulfill all of the three abovementioned specifications were classified as clear-sky and are shown in Fig. 11. The cloud-filtering method is illustrated using a scatter plot of BSR₉₄₀ vs. CI, which is given in Fig. S5 in the Supplement.

In NT16_AUG, a clear-sky BSR₄₅₅ enhancement (i.e., a mean value exceeding 1.04) starting approximately at the LRM (180 hPa) and extending up to the TOC (73 hPa) is the signature of the ATAL (Fig. 11), showing intensity and vertical extent comparable to those derived by satellite retrievals and previous COBALD measurements (Vernier et al., 2015, 2018). The enhanced aerosol BSR₄₅₅ covers both the ATTL and the CLS, with a maximum BSR₄₅₅ at the CPT. The fact that the onset of the ATAL coincides with the LRM suggests that the mean convective outflow level is also the onset of horizontal confinement in the ASMA. The maximum BSR₄₅₅ at the CPT is possibly an effect of colder temperatures, driving the partitioning of more condensable material (e.g., nitrates, see Vernier et al., 2018) to the aerosol phase in the ATAL. Above the CPT, the BSR₄₅₅ enhancement gradually fades with altitude until the TOC as the horizontal confinement effect of the ASMA vanishes, which is consistent with the decreasing confined fraction of the backward trajectories shown in Fig. 5. Above the TOC, the ATAL signal merges with the Junge layer of stratospheric aerosols, which extends into the free stratosphere up to about 10 hPa. The clear-sky BSR₄₅₅ enhancement is absent in NT16_NOV at all altitudes in the UTLS (except for the Junge layer in the free stratosphere), showing that the ATAL does not outlive the breakup of the anticyclonic vortex and the lack of a supply of precursor gases by deep convection after the end of the ASM season.