- Email: [email protected]
Tracking potential sources of peak ozone concentrations in the upper troposphere over the Arabian Gulf region
Atmospheric Environment 101 (2015) 257e269
Contents lists available at ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Tracking potential sources of peak ozone concentrations in the upper troposphere over the Arabian Gulf region Teresa K. Spohn a, b, 1, Bernhard Rappenglück a, b, * a b
Department of Earth and Atmospheric Sciences, University of Houston, 4800 Calhoun Rd, Houston, TX, 77204-5007, USA Qatar Environment and Energy Research Institute, P.O. Box 5825, Doha, Qatar
h i g h l i g h t s For the first time temporally highly resolved O3 sondes in the Middle East region. Upper tropospheric ozone in the Arabian Gulf region is on average 80 ± 13 ppbv. O3 in the 6e12 km range is higher when air masses came from the Mediterranean. High pressure may cause subsidence of O3 from upper troposphere/lower stratosphere. Convective activity and associated lightning can substantially increase O3.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 August 2014 Received in revised form 12 November 2014 Accepted 12 November 2014 Available online 13 November 2014
In August 2013, the Qatar Environment and Energy Research Institute (QEERI), was the first to launch temporally highly resolved ozonesondes in the Middle East region. The data from 20 launches consistently show changes in meteorological parameters at about 5.5 km above the surface, which are more pronounced following a change in synoptic conditions on 15 August 2013, including temperature inversions, corresponding change in potential temperatures, relative humidity, and significant wind shear. These changes are typically associated with a large scale subtropical subsidence layer in accordance with previous aircraft studies in this region. Below the inversion layer, the ozone follows typical patterns for lower tropospheric measurements, starting in the surface layer up to 0.5 km above the ground level around noon at about 66 ± 15 ppbv. However, above the subsidence inversion, ozone mixing ratios begin to increase to 80 ± 13 ppbv between 6 and 12 km with maximum values ~100 ppbv around 8 km, then decreasing again before reaching the stratosphere. Three-day HYSPLIT back trajectories indicate that ozone levels are typically about 17% lower in the 6 e12 km range under wind flow conditions from the East than in cases when trajectories came from the Mediterranean. High pressure may lead to subsidence of ozone from the upper troposphere/lower stratosphere and eventually cause an increase of ozone mixing ratios by ~18% above average between 6 and 7 km, i.e. slightly above subtropical subsidence layer. Under the impact of regional convective activity and associated lightning, ozone mixing ratios can increase by more than 35% averaged over the 9 e12 km altitude range. In both cases maximum ozone in the mid to upper troposphere reached more than 100 ppbv. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Ozone Middle East Qatar Tropospheric transport Lightning
1. Introduction Ozone (O3) in the troposphere is of interest because of its negative effects on human health and global warming (US
* Corresponding author. E-mail address: [email protected] (B. Rappenglück). 1 now with the National University of Ireland, Galway. http://dx.doi.org/10.1016/j.atmosenv.2014.11.026 1352-2310/© 2014 Elsevier Ltd. All rights reserved.
Environmental Protection Agency, 2012). Tropospheric ozone is primarily formed photo-chemically through the break-down of nitrogen oxides (NOx) by sunlight in the presence of volatile organic compounds (VOCs) (e.g. Lelieveld et al., 2009). Contrary to the tropospheric ozone, ozone in the stratosphere is formed through photolysis of oxygen and forms a barrier protecting the Earth from the sun's ultra-violet radiation; however, it is sometimes mixed into the troposphere through StratosphericeTropospheric Exchange (STE) (e.g. Lelieveld et al., 2009).
258
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
There are very few studies of ozone in the Arabian Gulf region (Li et al., 2001; Lelieveld et al., 2009; Zanis et al., 2014). While they all concur that there is unusually high ozone in the mid to upper troposphere in the Middle East region, they do not agree on the reasons for it. Li et al. (2001) focus on a study in July 1997. They conducted sensitivity simulations to determine contributions to ozone at various levels within the troposphere, as well as distinguish between the influence of anthropogenic and lightning sources. Li et al. claim that there are unusually high summertime levels of ozone, over 80 ppbv in the mid to upper troposphere in the Middle East region. The reason for this is the anti-cyclonic circulation over the Arabian Peninsula and the Indian sub-continent “funneling” in pollution from Europe and nitrogen oxides (NOx) generated by lightning outflow from the Indian Monsoon. Pollution from eastern Asia transported in the Easterly Tropical Jet Stream is said to contribute as well. Results indicate that the largest source of ozone (35e50%) is due to production in the upper troposphere and largescale subsidence in the region. Losses in these areas due to midlevel outflow are only little according to the Li et al. study. They estimate that 20e30% of the tropospheric ozone column has been caused by anthropogenic sources and 10e15% is from lightning. According to the paper, stratospheric ozone is not thought to be a major contributor, although it seems to have a greater impact than in other regions of the world. Li et al.'s study used the GEOS-CHEM (Goddard Earth Observing System) global 3-D model of atmospheric chemistry in conjunction with vertical ozone profiles collected by the MOZAIC (Measurement of Ozone and Water Vapour on Airbus in-service Aircraft; MOZAIC, 2014) program over the regions of Tel-Aviv, Dubai, and Tehran. Data from the NASA (National Aeronautics and Space Administration) Data Assimilation Office for the years 1993e1997 were used as input for the model along with an anthropogenic base emissions inventory from 1985, adjusted for the study period, which included NOx, Non-Methane Hydrocarbons (NMHC), and Carbon Monoxide (CO). Lightning-NO production associated with deep convection was parameterized following Price and Rind (1992). Lelieveld et al. (2009) postulate that the Middle East is an ozone “hotspot” due to long-range transport of pollutants as indicated by the tracer CO, unusually strong STE, substantial natural upwind sources of NOx such as lightning, the lack of precipitation, and contribution from local emissions such as those from oil and gas refineries combining to create ideal conditions for ozone production and entrainment. They find that there is a distinct ozone maximum between the surface and 750 hPa, which is even more pronounced in the summertime when conditions favor photochemistry. The average ozone mixing ratio in the mid to upper troposphere in the summer is around 80 ppbv. A comparison with other subtropical areas showed that the diel variation of midtropospheric ozone in Bahrain was related more to the longrange transport of pollutants than to local production, and that after removing anthropogenic sources from the model, the region still had higher ozone than the other areas. Unlike Li et al. (2001), this study suggests that stratospheric ozone does have a major impact on the regional tropospheric ozone column, making up about two thirds of it in winter and one quarter in the summer. The Lelieveld study uses the EMAC (European CentereHamburg 5th generation model MESSy [Modular Earth Submodel System] Applied to Atmospheric Chemistry) model in conjunction with satellite-retrieved ozone imagery from the Tropospheric Emission Spectrometer (TES, 2014) and Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY, 2014) to identify where the highest ozone concentrations are. The model is nudged toward the meteorological conditions in 2006 based on re-
analysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) and uses EDGAR 3.2 (Emissions Database for Global Atmospheric Research 2000) for inputs of anthropogenic emissions. The model includes a stratospheric ozone tracer as well and follows its transport and destruction in the troposphere. However, it does not include recycling processes of that stratospheric ozone tracer in the troposphere. The most recent study by Zanis et al. (2014) indicates that the enhanced tropospheric summer ozone levels in the Eastern Mediterranean and Middle East are the result of stratospheric ozone being transported to the troposphere through subsidence induced by high pressure systems in the region. While local photochemical formation is still the dominant contributor to ozone concentration, stratosphere to troposphere transport (STT), a type of STE, plays a critical role in places with favorable conditions. The Zanis et al. (2014) research was conducted in a similar way to the previous Lelieveld et al. (2009) study, using the EMAC model with ECMWF inputs and a stratospheric ozone tracer, and comparing the results to TES satellite data, but this time for a 12 year climatological study from 1998 to 2009. Both the model and satellite show pools of higher ozone concentration in the upper and middle troposphere over the Eastern Mediterranean, with the stratospheric ozone tracer indicating a 40e45% contribution from the stratosphere to the middle troposphere. This is attributed to the large-scale subsidence and limited outflow resulting from the anticyclonic motion of the high pressure systems prevalent in the area during the summer months. Reid et al. (2008) report about dust profiles obtained through aircraft measurements, mostly over the area of the United Arab Emirates and vicinity, and also give a detailed overview of meteorological patterns in the region, including the mention of a subtropical subsidence inversion. While the focus of the paper was on aerosols and dust in the region, their flights encountered polluted air masses from Europe, India, and possibly Africa as well. All of these publications agree that the Middle East area is of specific interest due to some unique meteorological conditions and presumably elevated ozone levels which may have a large-scale impact and contribute to global warming as the ozone is exported from the region. Due to the scarcity of data sets in that area, they all call for more in situ measurements. In this paper we present data analysis from ozonesondes launched from Doha, Qatar, in August 2013 and explore potential source areas for ozone in the upper troposphere. 2. Methods 2.1. Data collection Electro Chemical Concentration (ECC) ozonesondes (Droplet Measurement Technologies, 2013) in conjunction with iMet-1 radiosondes (InterMet Systems, 2008) attached to a 1.2 kg Kaymont brand weather balloon were used. Sondes were equipped with GPS (Global Positioning System). The iMet-1 radiosonde system (InterMet Systems, 2013) provides data for pressure (accuracy: 0.5 hPa), temperature (accuracy: 0.2 C), relative humidity RH (accuracy: 5% RH), wind speed (accuracy: 0.1 m s1), wind direction (accuracy: 5 for wind speeds <14 m s1, and 2 for wind speeds >14 m s1), altitude (accuracy: 15 m), and position (accuracy: 5 m). The principle of ECC sondes are described thoroughly in Russell III et al. (1998). ECC sondes were used because they routinely outperform other types of ozonesondes such as the Brewer-Mast and KC96 Carbon Iodine Cell, as demonstrated in the 1996e2000 Jülich OzoneSonde Intercomparison Experiment (JOSIE) (Smit and Kley, 1996; Smit et al., 2007) and are currently the most commonly used world-wide (Global Atmospheric Watch, 2013).
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Depending on the launch conditions and amount of helium, the balloon rose at a rate of 2e5 m s1 and reached around 30 km above sea level (asl) before burst. For the altitude ranges considered in this study the accuracy of the ECC sondes varies from ±5% (1000 hPa) to ±12% (200 hPa) and the precision from ±4% (1000 hPa) to ±12% (200 hPa) (Droplet Measurement Technologies, 2013). Smit et al. (2007) report even slightly lower values. Comparison studies with routine lidar measurements made at the Observatoire de Haute Provence during 1990e1995 show that for fifteen simultaneous ECC versus lidar profiles, the mean of the differences observed between 4 and 7 km was 2.5 ± 1.8 ppb (4 ± 3%) (Ancellet and Beekmann, 1997). In the period from 4 to 21 August 2013, 20 ozonesondes were launched from the College of the North AtlanticeQatar campus in Doha (25.36 N, 51.48 E) at various times of day, including early morning (7:00 LT (local time)), mid-day (between 12:00e13:00 LT), and night (23:00e24:00 LT). Local time is UTC (Universal Time Code) þ 2 h. Data validation and verification were done in accordance with standardized methods listed in the U.S. Environmental Protection Agency's Manual EPA-QA/G-8. (US Environmental Protection Agency, 2002)
3. Observations The radiosonde data consistently show a marked change in temperature, relative humidity, wind direction, and potential temperature at an altitude of about 5.5 km asl e factors indicative of a strong inversion. Previous observational and modeling studies (Li et al., 2001; Reid et al., 2008; Lelieveld et al., 2009) have shown that this phenomenon at about 5 km is in fact a subtropical subsidence inversion, which seems to occur persistently during the summer months. This is also seen in regular radiosonde launches from other locations in the Middle East area, e.g. Al-Qaisumah and Riyadh (Saudi-Arabia), Abu Dhabi (United Arab Emirates) and Kuwait City (Kuwait) (not shown, but retrieved from the University of Wyoming global radiosonde depository at http://www.weather. uwyo.edu/upperair/mideast.html). Fig. 1 displays the average ozone profile, including the 25% and 75% percentiles, for all 20 Doha ozonesondes launched from 4 to 21 August 2013, segregated into 100 m altitude bins. This figure shows that above the subsidence inversion, ozone begins to increase significantly within the free middle troposphere, peaking around 8 km altitude range and then decreasing again before entering the stratosphere between 14 and 16 km (see Table 1 for more information). This is fairly in agreement
259
with modeled height of tropopause crossings for the Middle East region by Tyrlis et al. (2014). Ozone in the troposphere is mainly formed locally near the surface in complex reactions involving NOx and VOC precursors, however it has been found that pollution in the surface layer can reach upper tropospheric levels, due to strong convective processes (Baehr et al., 2003; Borbon et al., 2012). The persistent subsidence inversion over the Middle East area during the summer acts largely to suppress convective processes and the upper troposphere is therefore decoupled from the lower troposphere (Li et al., 2001; Reid et al., 2008; Lelieveld et al., 2009). As in-situ formation of ozone is limited in the upper troposphere, ozone aloft is likely due to transport processes. Apart from horizontal transport, vertical transport (e.g. lifting or subsidence) can play an important role. Through STE processes, stratospheric ozone can also contribute. Previous research has postulated that ozone in the mid to upper troposphere is being transported from Europe and the Mediterranean (Li et al., 2001). Henceforth we focus the discussions on the ozone distribution in the upper troposphere to trace the potential source areas of the elevated ozone found in the mid to upper troposphere between 6 and 12 km over the Arabian Gulf area. For this purpose we consider mid-day launches only and do not include early morning launches done in addition after 15 August. This way there is consistency of only one launch per day at approximately the same time. From the remaining 14 mid-day launches we needed to remove one launch (09 August) due to problems in the GPS data acquisition which led to unreliable wind data. In the following data analysis we therefore focus on an overall dataset of 13 mid-day launches. For data analysis we applied a series of 72 h HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory Model) (Draxler and Rolph, 2013) back trajectories at altitudes of 6, 9, and 12 km, based on the GPS coordinates of each ozonesonde when it was at these altitudes. In order to identify meteorological patterns for the month of August, the World Wide Lightning Location Network (WWLLN) lightning data (WWLLN Management Team, 2013), as well as National Oceanic and Atmospheric Administration (NOAA) surface, 200 and 500 hPa weather analysis charts of the region (NOAA National Climatic Data Center, 2013) were used to determine what factors might have influenced the long-range transport of ozone. 3.1. HYSPLIT back trajectories Fig. 2 shows the results of the HYSPLIT back trajectory analyses. Of the 13 launches used in this study, 6 days between 5 and 13 August showed trajectories coming from the East, from northern India and the Himalayas. On two days, 14 and 15 August, the trajectories curled around locally, indicating a shift in large-scale wind flows. After 15 August, the remaining 5 set of trajectories came from the northwest, the Mediterranean and North Africa. Uncertainties of the HYSPLIT trajectories are assumed to be in the range of 15e30% of the travel distance (Draxler and Rolph, 2013). Although, there might be some limitations associated with the accuracy of backward trajectories, they at least may provide information about major regimes. The HYSPLIT back trajectories showed generally that each air packet ended up within approx. 20e25% of its initial altitude. 3.2. Weather charts
Fig. 1. Average ozonesonde profiles in the time period 4e21 August 2013. Data are all 20 ozonesonde launches segregated into 100 m altitude bins.
Fig. 3 displays 500 hPa and 200 hPa weather charts for the Middle East area. The surface level analysis and lifted index charts are characterized by a constant “heat low” in the region due to high surface temperatures. This is capped by high pressure, indicated by
260
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Table 1 Median, mean, standard deviation, and range of measured ozone for all 13 launches at various altitudes. Launch times close to local noon. Layer
Upper troposphere (6 / 12 km) Entire troposphere (0 / 14 km) Surface layer (0 / 0.5 km) a
Numbera
Ozone [ppbv] Median
Mean
Standard deviation
Maximum
Minimum
79.4 72.4 66.4
80.0 71.6 64.3
12.8 14.4 14.8
109 109 95
46 27 29
2878 ± 183 5767 ± 355 165 ± 25
Number of data points per launch.
elevated geopotential heights on the 500 hPa chart, preventing convective lifting, and resulting in thin altostratus clouds at about 6 km altitude most days. On 12 and 13 August, a stronger ridge moved in over Qatar, indicated by the yellow arrow on the 500 hPa chart. This ridge began to move out of the Qatar region on 14 August and had receded completely by 15 August. At the 200 hPa level, there is a persistent powerful jet streak north of Qatar between 40 and 45 N from the beginning of August until 15 August, when it suddenly disappears, after which an upper level trough forms, pulling in air from the Mediterranean. 3.3. Lightning data Lightning data for the northern hemisphere for the month of August was obtained from the World Wide Lightning Location Network (WWLLN; http://wwlln.net/new/network/) and plotted in Google Earth. These data are gathered by a global network of sensing stations using VLF (Very Low Frequency between 3 and 30 KHz) to detect “sferics”delectromagnetic discharges caused by lightning. With a current network of 40 stations, the WWLLN is capable of detecting only about 30% of all lightning strikes. This may be considered as an upper limit; it may be likely less in the in the area of this study. Due to this limitation, any lightning activity which is detected can be considered a lower limit and the actual lightning activity may be largely underestimated, and therefore the potential lightning-influenced ozone production in these areas as well. The Indian Summer Monsoon took place during August, and there was in fact heavy lightning activity over India, often along the HYSPLIT trajectories. There was however also considerable lightning in the Middle East, even in Qatar itself on some days. In particular, on 6 August the country experienced some thunderstorms (Fig. 4). 3.4. Balloon data As illustrated in Table 1, the average measured ozone for all 13 launches between 6 and 12 km was 80 ± 13 ppbv, with a median of ~79 ppbv. This is higher than the average total tropospheric ozone between the surface and 14 km of 72 ± 14 ppbv. The average ozone below 500 m was ~64 ± 15 ppbv. Note that for consistency, and to ensure none of the data were above the boundary layer, 500 m was used as a representation of surface ozone, rather than the entire planetary boundary layer which varies between 1 and 2 km. All launches were characterized by pronounced changes in relative humidity above 5.5 km as shown exemplary in Fig. 5. Unlike the launches before 15 August which showed only weak temperature inversions, the launches after 15 August all showed a marked temperature inversion of 2e3 C over a distance of less than 0.5 km at around 5.5 km altitude. On 7 of the 13 days, the ozonesondes measured maximum ozone over 100 ppbv between 6 and 12 km. Three of these launches took place before 15 August, at which point the large scale synoptic condition changed and thus did the wind direction throughout
most of the troposphere. On 6 August a peak ozone mixing ratio over 100 ppbv occurred at an altitude above 12 km. On 12 August the peak ozone mixing ratio was 105 ppbv between 8 and 10 km, and on 13 August a nearly identical “plume” between 6 and 8 km (Fig. 6) can be visually discerned. 4. Discussion In this section we will focus on three examples to explain the causes of high ozone concentrations in the Arabian Gulf region. Each case has specific characteristics which can indicate the likely source of the ozone peaks. 4.1. Case 1: ozonesonde launch on 6 August and the effect of lightning In this study we do not verify a distinct impact on ozone due to lightning related to the Indian Monsoon as this would require a comparison with transport processes from the Indian subcontinent under non-Monsoon conditions. However, this study shows that lightning from the Indian Monsoon is not associated with as high of ozone values at upper-tropospheric levels in the Arabian Gulf region as are observed when back trajectories point to the Eastern Mediterranean region. Air masses originating from the Eastern Mediterranean region contain about 30e40 ppb higher ozone values than air masses impacted by the Indian Monsoon and associated lightning activities as shown in Fig. 8. Fig. 8 also demonstrates that these different ozone regimes are clearly accompanied by different relative humidity regimes. While high relative humidity values can be expected for air masses impacted by the Indian Monsoon, it is remarkable that air masses originating from the Eastern Mediterranean contain extremely low relative humidity for the entire preceding 72 h. This distinct feature is visible in all the other profiles of the corresponding specific synoptic conditions. As mentioned before, the days 14 and 15 August were transitional days between these large-scale flows, which also showed transition of the relative humidity profiles. The Eastern Mediterranean is a well-known hot-spot for enhanced tropospheric ozone levels (e.g. Kouvarakis et al., 2002; Kourtidis et al., 2002; Lelieveld et al., 2002, 2009). This is largely due to large-scale subsidence in the summer (Zanis et al., 2014; Tyrlis et al., 2014), low tropopause height, and enhanced tropopause folding activity over the Levantine region during the summertime, in connection with the Asian monsoon (Tyrlis et al., 2014). The numerical simulations by Zanis et al. (2014) indicate that there is a large contribution of stratospheric ozone to the pool of high ozone values over the Eastern Mediterranean/ Middle East in the middle and lower free troposphere and Sprenger et al. (2007) showed that stratospheric air mixed into the upper troposphere can remain there up to a few days. The extremely low relative humidity values in air masses coming from the Eastern Mediterranean area are in support of these previous modeling studies. Apart from these ozone transports in the upper troposphere, however, also local thunderstorm activity in the Arabian Gulf could
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
261
Fig. 2. HYSPLIT back trajectories in Google Earth (Google Inc., 2013). Altitude color coding: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
262
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 3. NOAA 12:00UTC weather charts sequence for 11 to 14 August depicting a ridge over Arabian Peninsula on 12 and 13 August 2013 on the 500 hPa charts as indicated by the yellow arrow (left panel, top to bottom) and jet streak movement (shown in green) at 40e45 N around 15 August 2013 on the 200 hPa charts from 13 to 16 August (right panel, top to bottom). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
be contributing to higher ozone. On 6 August there was considerable lightning activity in the entire region, about 50% more than on other days before 15 August, and the HYSPLIT back trajectory for the ozonesonde, which was launched that day, either passed directly
through (trajectory at 9 km altitude in Fig. 9) or close to a nearby thunderstorm in the Arabian Gulf with correspondingly increased ozone at 12 km (Fig. 9). It should be noted that the location, altitude, spatial extension of the lightning activity as well as the
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
263
Fig. 4. Lightning on 6 August 2013 (WWLLN Management Team, 2013).
Fig. 5. Profiles of ozone, relative humidity, air temperature, potential temperature, wind speed, and wind direction as obtained from the ozonesonde on 19 August 2013 at 10:10 UTC (12:10 LT).
264
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 6. Above: Comparison of vertical ozone profiles on 12 and 13 August 2013. Note the plume height at 9e10 km on 12 August and 8 km on 13 August. Below: Vertical profiles of ozone and relative humidity in the altitude range 6e12 km.
number of strikes is indicative for the overall lightning activity. As mentioned above the WWLLN data may only detect about 30% of all lightning strikes. In addition, uncertainties associated with the back trajectories are in the range of 15e30% of the travel distance (Draxler and Rolph, 2013). Thus it is likely that both trajectories, at 9 km and at 12 km altitude, might have been impacted by lightning. The peak ozone mixing ratio that day was 105 ppbv at 12 km, with a
smaller 101 ppbv peak at 10 km (Fig. 9). This is significantly higher than on the preceding day when under the same flow conditions about 50e60 ppbv ozone were observed from top of the boundary layer height around 2 km altitude up to 10e14 km altitude. Lightning is an important source of NOx in the upper troposphere worldwide (Schumann and Huntrieser, 2007; Huntrieser et al., 2012), including the Middle East area (Li et al., 2001) and the balloon's
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
265
Fig. 7. Above: Altitude plot of the 72-h backward trajectory ending at the location of the ozonesonde at about 9 km altitude. Below: relative humidity along this trajectory.
Fig. 8. Comparison of high lightning and low ozone (5 August 2013, 10:07 UTC (12:07 LT); upper left plot and blue line in graph) with low lightning and high ozone days (19 August 2013, 10:10 UTC (12:10 LT); upper right plot and red line in graph). Lightning data for 5 Aug (blue), 4 Aug (pink), 3 Aug (green), and 2 Aug (yellow); Lightning data for 19 Aug (blue), 18 Aug (pink), 17 Aug (green), and 16 Aug (yellow). Altitude color coding: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. In addition, vertical profiles of relative humidity on the corresponding days as well as along 72-h backward trajectories ending at the ozonesonde location at about 9 km altitude are shown. Upper RH trajectory is for 5 August, lower RH trajectory for 19 August. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
266
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 9. Above: lightning and HYSPLIT back trajectories on 6 August 2013, 10:00 UTC (12:00 LT), highlighting the 9 km trajectory passing through the lightning storm. The start point of the back trajectory indicates the position of the ozonesonde at that given time. Altitude color coding for back trajectories: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. Below: vertical ozone profile on 6 August 2013.” (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
actual trajectory took it directly downwind from this source. Huntrieser et al. (2007) report up to 60e70 nmol mol1 enhancement of ozone in the anvil outflow over a horizontal scale of about 400 km. This also due to the fact that the atmospheric lifetime of NOx in the upper troposphere is significantly longer than in the boundary layer and that photochemical processes are largely NOx limited (Huntrieser et al., 2007 and references therein). The amount of NOx produced also depends on the lightning stroke length, peak current, release height, and vertical wind shear (Huntrieser et al., 2008). Fig. 9 shows that ozone exceeds 60 ppbv, the preceding day's average tropospheric value observed under similar Easterly flow conditions, around 7 km altitude and then continuously increases and reaches maximum values. This coincides with strongest vertical wind shears between 9 and 12 km, which encompasses the top of the cloud deck, as reflected in the relative humidity data (not
shown). Ozone mixing ratios averaged over 9e12 km altitude were 66.8 ± 2.2 ppbv on 5 August versus 91.2 ± 2.4 ppbv on 6 August, which corresponds to an increase of about 36.5% over this altitude range. These values rank among higher values found in other studies on lightning impact on ozone in the upper troposphere (Cooper et al., 2007; Morris et al., 2010 and references therein). In conclusion, this suggests that ozone produced as a result of this lightning activity may have contributed appreciably to the concentrations measured by the ozonesonde over Qatar. 4.2. Case 2: ozonesondes launched on 12 and 13 August The 12 and 13 of August were the only two days of the month during which a strong ridge was centered over the Arabian Peninsula, as shown on the 500 hPa chart in Fig. 3. Zanis et al.
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
267
Table 2 Mean and standard deviation of measured ozone at various altitudes separated into different source regions according to the 72 h back trajectories as shown in Figs. 2 and 7. The transition days August 14-15 (see Fig. 2) are not included. Launch times close to local noon. Ozone in ppbv. Layer
Upper troposphere (6 / 12 km) Entire troposphere (0 / 14 km) Surface layer (0 / 0.5 km) a b
Launches August 5e13a (“East”)
Launches August 16e21b (“West”)
Mean
Standard deviation
Mean
Standard deviation
74.0 66.1 61.3
6.9 5.0 2.1
86.2 77.6 66.7
8.9 3.0 4.5
Based on 6 ozonesondes. Based on 4 ozonesondes.
(2014) discuss how the anti-cyclonic activity associated with high pressure systems causes ozone-rich air to be pulled downward isentropically out of the stratosphere into the upper troposphere through subsidence in the Eastern Mediterranean and Middle East region. Their study used a stratospheric ozone tracer in the model which showed the air mass descending over time. The observations on 12 and 13 August show an example for subsidence, as the ozone plume is between 8 and 10 km on 12 August, and has fallen to between 6 and 8 km only 24 h later, yet maintains its distinct shape (see section highlighted by the rectangle in Fig. 6). In Fig. 6 we included the nighttime ozonesonde launch on 12 August so that this figure displays a sequence of ozonesondes almost every 12 h apart. Note that the ozone values between 6 and 12 km altitude are significantly higher than the average ozone profiles at these altitude ranges (see Fig. 1). Both days show subsidence inversions (not shown) at similar altitudes (5.5 km on 12 August and 6 km on 13 August, i.e. close to the 500 hPa level). These inversions are associated with wind shear, mostly northerly below, whereas aloft winds come from the east throughout the upper troposphere coincident with the large scale synoptic situation as shown in Fig. 3. Both inversions are also associated with generally decreasing relative humidity aloft, typical for subsidences. Also, starting with these inversions ozone mixing ratios increase with increasing altitude reaching maximum values between 8 and 11 km altitude on 12 August and between 6 and slightly above 8 km altitude on 13 August. The levels with enriched ozone are confined by another thermal inversion at about 12 km altitude on 12 August and slightly above 8 km altitude on 13 August. On 12 August the layer of maximum ozone is accompanied by low relative humidity around 10e15%, similar to the relative humidity values observed in the lower stratosphere on that day. Due to these overall low relative humidity values the thermal inversion about 12 km altitude does not show a pronounced change in relative humidity. This is in contrast to 13 August, when the lower thermal inversion can be clearly identified as a subsidence inversion with decreasing relative humidity values aloft. This is also reflected in the HYSPLIT analysis for that day (Fig. 7), which shows that air masses arriving at the ozonesonde location at about 9 km altitude have originated above 11.5 km altitude and experienced a drastic decrease in relative humidity along the trajectory. While on 13 August the relative humidity profile shows some variation within the layer between 6 km and slightly above 8 km altitude ozone mixing ratios are largely anticorrelated with relative humidity in that layer indicative for sustaining similar atmospheric properties from the previous day. The anticorrelation of ozone with relative humidity can be best seen in the nighttime ozone profile on 12 August between ~8 km and ~9 km altitude. The main difference between the vertical meteorological profiles of those two days is the change in the upper tropospheric subsidence inversions from around 12 km (12 August) to slightly above 8 km altitude (13 August) within 24 h, which would translate into an entrainment of enriched ozone layers from upper tropospheric areas on the order of 4e5 cm s1 estimated as an upper boundary. Based on the
HYSPLIT analysis an entrainment velocity on the order of 1 cm s1 can be estimated. While on 12 August ozone mixing ratios were about 93.0 ± 11.8 ppbv averaged over the altitude range of 8e11 km, they decreased to 74.3 ± 9.4 ppbv on the subsequent day for the same altitude range, which is a decrease by ~20%. On the other hand, ozone mixing ratios averaged over 6e7 km were 66.1 ± 4.2 ppbv on 12 August and increased to 77.9 ± 12.5 ppbv on 13 August (increase by ~18%). The altitude range 7e8 km represents a transitional region with limited change as the plume of enriched ozone passes this region during subsidence. This subsidence process will be accompanied by horizontal dispersion, which is reflected by the fact that the ozone mixing ratios in the ozone lamina slightly decrease and are confined to a shrinking altitude range.
4.3. Case 3: ozonesonde launched on 19 August The majority of cases of high ozone concentrations in the region during the measurement period occur when the HYSPLIT back trajectories come from the Mediterranean, as exemplified by the profile of the ozonesonde launched on 19 August (Fig. 8), and corroborated by all launches on and after 15 August. Here the ozone begins to increase at a lower altitude, around 7 km, and remains high up to 12 km. In these instances the ozone peak is about 100 ppbv, and most likely the result of long-range transport of pollutants (Li et al., 2001) in the 6e12 km range (72 h back trajectories in this altitude range point towards the Eastern Mediterranean). The strong temperature inversion at 6 km acts as a boundary layer, preventing mixing from below or transfer from aloft. Between 11.5 and 12 km altitude wind direction changes by about 180 (not shown) which coincides with a sharp drop in ozone mixing ratios and which caps the high ozone levels observed beneath, i.e. in the 6e12 km range. Table 2 shows general results of the two large-scale flow regimes observed during the measurement period as evident in Figs. 2 and 8. The flow regime denoted as “East” refers to the time period impacted by Monsoon activity over India, whereas the flow regime denotes as “West” refers to the source region over the Eastern Mediterranean. From Table 2 it is evident that Westerly flow regimes are associated with higher ozone mixing ratios than Easterly flow regimes during the measurement period of August 2013. Compared to Easterly flow conditions, ozone values in Westerly flow conditions are enhanced by about 17% through the entire troposphere and in particular in the 6e12 km altitude range, which is double the enhancement in the surface layer. It is noteworthy to keep in mind that ozone values under Easterly flow conditions are impacted by thunderstorm activities related to the Monsoon season. As a consequence an even higher ozone enhancement in Westerly versus Easterly flow regimes in the 6e12 km altitude range may have been masked by Monsoon lightning activities. The observed ozone enhancement associated with back trajectories point towards source regions over the
268
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Eastern Mediterranean and support recent model results (Lelieveld et al., 2009; Zanis et al., 2014). 5. Conclusions During August 2013, 20 ozonesondes were launched from Doha, Qatar. A subset of 13 mid-day launches were further analyzed. Although limited to one month in one summer season some interesting observations were obtained. Upper tropospheric ozone in the Arabian Gulf region in August 2013 was on average 80 ± 13 ppbv, as predicted by models used in previous studies (Li et al., 2001; Lelieveld et al., 2009); however the source of high ozone is more in agreement with Lelieveld et al. (2009) and Zanis et al. (2014) than postulated by Li et al. (2001). The extremely low relative humidity of a few percent observed in air masses originating from the Eastern Mediterranean in our study is in support of the modeling studies by Lelieveld et al. (2009) and Zanis et al. (2014) who suggested high ozone levels in the middle and upper troposphere as a result of stratosphereetroposphere transport over the Eastern Mediterranean and subsequent transport from this area to the Arabian Gulf under continued subsidence processes. Elevated ozone levels in the upper troposphere occur because of the meteorological conditions allowing precursors to be transported into the area, and as a result of subsidence. When the wind flow in the upper troposphere is from the East, the ozone levels are about 17% lower in the 6e12 km range than when the wind is from the Mediterranean. High pressure may lead to subsidence of ozone from the upper troposphere/lower stratosphere and eventually cause an increase of ozone mixing ratios by ~18% between 6 and 7 km, i.e. slightly above the subtropical subsidence layer. Under the impact of regional convective activity and associated lightning, ozone mixing ratios can increase by more than 35% averaged over the 9e12 km altitude range. In either case maximum ozone in the mid to upper troposphere reached more than 100 ppbv. This study was limited to 13 launches in the month of August. There are no seasonal comparisons. More work needs to be done in order to learn more about the temporal variability in regional ozone patterns and long-term trends. Acknowledgments We are grateful for the support provided by Qatar Foundation and the Qatar Civil Aviation Authority. Logistical and technical support provided by Messrs. M.A. Ayoub, L. Ackermann, and Mrs. M. Shalaev from QEERI is greatly appreciated. The research results have not been subject to QEERI scientific and policy review and therefore do not necessarily reflect the views of QEERI and no official endorsement should be inferred. Many thanks also to Dr. B. Holzworth for providing the WWLLN data for the time period of the study. References Ancellet, G., Beekmann, M., 1997. Evidence for changes in the ozone concentration in the free troposphere over Southern France from 1976e1995. Atmos. Environ. 31, 2835e2851. Baehr, J., Schlager, H., Ziereis, H., Stock, P., van Velthoven, P., Busen, R., Stroem, J., Schumann, U., 2003. Aircraft observations of NO, NOy, CO, and O3 in the upper troposphere from 60N to 60S e interhemispheric differences at mitlatitudes. Geophys. Res. Lett. 30 (11), 1598. Borbon, A., Ruiz, M., Bechara, J., Aumont, B., Chong, M., Huntrieser, H., Mari, C., Reeves, C.E., Scialom, G., Hamburger, T., Stark, H., Afif, C., Jambert, C., Mills, G., Schlager, H., Perros, P., 2012. Transport and chemistry of formaldehyde by mesoscale convective systems in West Africa during AMMA 2006. J. Geophys. Res. 117, D12301. Cooper, O.R., Trainer, M., Thompson, A.M., Oltmans, S.J., Tarasick, D.W., Witte, J.C., Stohl, A., Eckhardt, S., Lelieveld, J., Newchurch, M.J., Johnson, B.J.,
Portmann, R.W., Kalnajs, L., Dubey, M.K., Leblanc, T., McDermid, I.S., Forbes, G., Wolfe, D., Carey-Smith, T., Morris, G.A., Lefer, B., Rappenglück, B., Joseph, E., Schmidlin, F., Meagher, J., Fehsenfeld, F.C., Keating, T.J., Van Curen, R.A., Minschwaner, K., 2007. Evidence for a recurring eastern North America upper tropospheric ozone maximum during summer. J. Geophys. Res. 112, D23304. http://dx.doi.org/10.1029/2007JD008710. Draxler, R., Rolph, G., 2013. HYSPLIT (HYbrid Single-particle Lagrangian Integrated Trajectory) Model Access. Retrieved 08 27, 2013, from NOAA ARL READY. http:// www.arl.noaa.gov/HYSPLIT.php. Droplet Measurement Technologies, 2013. ECC Ozonesonde. Retrieved August 27, 2013, from Droplet Measurement Technologies. http://www.dropletmeasurement.com/ products/airborne/ECC_Ozonesonde#Specifications. Global Atmospheric Watch, 2013. Quality Assurance and Quality Control for Ozonesonde Measurements in GAW. Report #201. World Meteorological Organization, Geneva. Google Inc, 2013, July 12. Google Earth 7.1.1.1888 [Software]. Huntrieser, H., Schlager, H., Roiger, A., Lichtenstern, M., Schumann, U., Kurz, C., Brunner, D., Schwierz, C., Richter, A., Stohl, A., 2007. Lightning-produced NOx over Brazil during TROCCINOX: airborne measurements in tropical and subtropical thunderstorms and the importance of mesoscale convective systems. Atmos. Chem. Phys. 7, 2987e3013. €ller, H., Giez, A., Betz, H.-D., Forster, C., Huntrieser, H., Schumann, U., Schlager, H., Ho Pinto Jr., O., Calheiros, R., 2008. Lightning activity in Brazilian thunderstorms during TROCCINOX: implications for NOx production. Atmos. Chem. Phys. 8, 921e953. € ller, H., Grewe, V., 2012. Thunderstorms: trace species generators. Huntrieser, H., Ho In: Schumann, U. (Ed.), Atmospheric Physics: Background e Methods e Trends Research Topics in Aerospace. Springer-Verlag, ISBN 978-3-642-30182-7, pp. 115e133. InterMet Systems, Inc, 2008. iMet-1 Product. Retrieved August 27, 2013, from International Met Systems. http://www.intermetsystems.com/InterMet_1.html. InterMet Systems, Inc, 2013. Technical Description for the iMet-3050 and 3150 403 MHz GPS Upper-air Sounding Systems and the iMet-1 Radiosonde. Document No. 200,731 Revision-D, 37 pages. Kourtidis, K., Zerefos, C., Rapsomanikis, S., Simeonov, V., Balis, D., Perros, P.E., Thomson, A.M., Witte, J., Calpini, B., Sharobiem, W.M., Papayiannis, A., Mihalopoulos, N., Drakou, R., 2002. Regional levels of ozone in the troposphere over eastern Mediterranean. J. Geophys. Res. 107, 8140. http://dx.doi.org/ 10.1029/2000JD000140. Kouvarakis, G., Vrekoussis, M., Mihalopoulos, N., Kourtidis, K., Rappenglueck, B., Gerasopoulos, E., Zerefos, C., 2002. Spatial and temporal variability of tropospheric ozone in the boundary layer above the Aegean Sea (eastern Mediterranean). J. Geophys. Res. 107, 8137. http://dx.doi.org/10.1029/ 2000JD000081. Lelieveld, J., Berresheim, H., Borrmann, S., Crutzen, P.J., Dentener, F.J., Fischer, H., de Gouw, J., Feichter, J., Flatau, P., Heland, J., Holzinger, R., Korrmann, R., Lawrence, M., Levin, Z., Markowicz, K., Mihalopoulos, N., Minikin, A., Ramanathan, V., de Reus, M., Roelofs, G.-J., Scheeren, H.A., Sciare, J., Schlager, H., Schultz, M., Siegmund, P., Steil, B., Stephanou, E., Stier, P., Traub, M., Williams, J., Ziereis, H., 2002. Global air pollution crossroads over the Mediterranean. Science 298, 794e799. € ckel, P., Pozzer, A., Hadjinicolaou, P., Cammas, J., Beirle, S., Lelieveld, J., Hoor, P., Jo 2009. Severe ozone air pollution in the Persian Gulf region. Atmos. Chem. Phys. 9, 1393e1406. Li, Q., Jacob, D.J., Logan, J.A., Bey, I., Yantosca, R.M., Hongyu, L., Martin, R.V., Fiore, A.M., Field, B.D., Duncan, B.N., Thouret, V., 2001. A tropospheric ozone maximum over the middle east. Geophys. Res. Lett. 28 (17), 3235e3238. MOZAIC (Measurements of OZone, water vapour, carbon monoxide and nitrogen oxides by in-service AIrbus airCraft), 2014. Retrieved from http://www.iagos.fr/ web/rubrique2.html. Morris, G.A., Thompson, A.M., Pickering, K.E., Chen, S., Bucsela, E.J., Kucera, P.A., 2010. Observations of ozone production in a dissipating tropical convective cell during TC4. Atmos. Chem. Phys. 10, 11189e11208. http://dx.doi.org/10.5194/acp10-11189-2010. NOAA National Climatic Data Center, 2013. Service Records Retention System (SRRS). Retrieved from Analysis and Forecast Charts. http://nomads.ncdc.noaa. gov/ncep/NCEP. Price, C., Rind, D., 1992. A simple lightning parameterization for calculating global lightning distributions. J. Geophys. Res. 97, 9919e9933. Reid, J.S., Piketh, S.J., Walker, A.L., Burger, R.P., Ross, K.E., Westphal, D.L., Bruintjes, R.T., Holben, B.N., Hsu, C., Jensen, T.L., Kahn, R.A., Kuciauskas, Arunas P., Mandoos, A.A., Mangoosh, A.A., 2008. An overview of UAE2 flight operations: observations of summertime atmospheric thermodynamic and aerosol profiles of the southern Arabian Gulf. J. Geophys. Res. 113, D14213. Russell III, J., Smit, H., Hofmann, D., Rodgers, C., 1998. Stratospheric Processes and their Role in Climate (SPARC): Assessment of Trends in the Vertical Distribution of Ozone. WMO Ozone Res. And Monit. Proj. Rep. 43, (World Clim. Res. Prog., Geneva). Schumann, U., Huntrieser, H., 2007. The global lightning-induced nitrogen oxides source. Atmos. Chem. Phys. 7, 3823e3907. http://dx.doi.org/10.5194/acp-73823-2007. SCIAMACHY e SCanning Imaging Absorption spectroMeter for Atmospheric ChartographY. (2014). Retrieved from http://www.sciamachy.org/. Smit, H.G.J., Kley, D., 1996. Jülich ozone sonde intercomparison experiment (JOSIE). World Meteorol. Organ. Glob. Atmos. Watch 130.
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269 Smit, H., Straeter, W., Johnson, B., Oltmans, S., Davies, J., Tarasick, D.W., Hoegger, B., Stubi, R., Schmidlin, F., Northam, T., Thompson, A., Witte, J., Boyd, I., Posny, F., 2007. Assessment of the performance of ECC-ozonesondes under quasi-flight conditions in the environmental simulation chamber: insights from the Juelich Ozone Sonde Intercomparison Experiment (JOSIE). J. Geophys. Res. 112, D19306. Sprenger, M., Wernli, H., Bourqui, M., 2007. Stratosphereetroposphere exchange and its relation to potential vorticity streamers and cutoffs near the extratropical tropopause. J. Atmos. Sci. 64, 1587e1602. http://dx.doi.org/10.1175/ JAS3911.1. TES (Tropospheric Emission Spectrometer). (2014). Retrieved from http://tes.jpl. nasa.gov/. Tyrlis, E., Skerlak, B., Sprenger, M., Wernli, H., Zittis, G., Lelieveld, J., 2014. On the linkage between the Asian summer monsoon and tropopause fold activity over
269
the eastern Mediterranean and the Middle East. J. Geophys. Res. 119, 3202e3221. http://dx.doi.org/10.1002/2013JD02113. US Environmental Protection Agency (2002). EPA-QA/G-8. Retrieved from http:// www.epa.gov/QUALITY/qs-docs/g8-final.pdf. US Environmental Protection Agency, 2012, November 1. Ground-level Ozone. Retrieved August 27, 2013, from US Environmental Protection Agency. http:// www.epa.gov/glo/health.html. WWLLN Management Team, 2013. World Wide Lightning Location Network (Retrieved from wwlln.net). Zanis, P., Hadjinicolaou, P., Pozzer, A., Tyrlis, E., Dafka, S., Mihalopoulos, N., Lelieveld, J., 2014. Summertime free tropospheric ozone pool over the Eastern Mediterranean/Middle East. Atmos. Chem. Phys. 14, 115e132. http://dx.doi.org/ 10.5194/acp-14-115-2014.
Contents lists available at ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Tracking potential sources of peak ozone concentrations in the upper troposphere over the Arabian Gulf region Teresa K. Spohn a, b, 1, Bernhard Rappenglück a, b, * a b
Department of Earth and Atmospheric Sciences, University of Houston, 4800 Calhoun Rd, Houston, TX, 77204-5007, USA Qatar Environment and Energy Research Institute, P.O. Box 5825, Doha, Qatar
h i g h l i g h t s For the first time temporally highly resolved O3 sondes in the Middle East region. Upper tropospheric ozone in the Arabian Gulf region is on average 80 ± 13 ppbv. O3 in the 6e12 km range is higher when air masses came from the Mediterranean. High pressure may cause subsidence of O3 from upper troposphere/lower stratosphere. Convective activity and associated lightning can substantially increase O3.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 August 2014 Received in revised form 12 November 2014 Accepted 12 November 2014 Available online 13 November 2014
In August 2013, the Qatar Environment and Energy Research Institute (QEERI), was the first to launch temporally highly resolved ozonesondes in the Middle East region. The data from 20 launches consistently show changes in meteorological parameters at about 5.5 km above the surface, which are more pronounced following a change in synoptic conditions on 15 August 2013, including temperature inversions, corresponding change in potential temperatures, relative humidity, and significant wind shear. These changes are typically associated with a large scale subtropical subsidence layer in accordance with previous aircraft studies in this region. Below the inversion layer, the ozone follows typical patterns for lower tropospheric measurements, starting in the surface layer up to 0.5 km above the ground level around noon at about 66 ± 15 ppbv. However, above the subsidence inversion, ozone mixing ratios begin to increase to 80 ± 13 ppbv between 6 and 12 km with maximum values ~100 ppbv around 8 km, then decreasing again before reaching the stratosphere. Three-day HYSPLIT back trajectories indicate that ozone levels are typically about 17% lower in the 6 e12 km range under wind flow conditions from the East than in cases when trajectories came from the Mediterranean. High pressure may lead to subsidence of ozone from the upper troposphere/lower stratosphere and eventually cause an increase of ozone mixing ratios by ~18% above average between 6 and 7 km, i.e. slightly above subtropical subsidence layer. Under the impact of regional convective activity and associated lightning, ozone mixing ratios can increase by more than 35% averaged over the 9 e12 km altitude range. In both cases maximum ozone in the mid to upper troposphere reached more than 100 ppbv. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Ozone Middle East Qatar Tropospheric transport Lightning
1. Introduction Ozone (O3) in the troposphere is of interest because of its negative effects on human health and global warming (US
* Corresponding author. E-mail address: [email protected] (B. Rappenglück). 1 now with the National University of Ireland, Galway. http://dx.doi.org/10.1016/j.atmosenv.2014.11.026 1352-2310/© 2014 Elsevier Ltd. All rights reserved.
Environmental Protection Agency, 2012). Tropospheric ozone is primarily formed photo-chemically through the break-down of nitrogen oxides (NOx) by sunlight in the presence of volatile organic compounds (VOCs) (e.g. Lelieveld et al., 2009). Contrary to the tropospheric ozone, ozone in the stratosphere is formed through photolysis of oxygen and forms a barrier protecting the Earth from the sun's ultra-violet radiation; however, it is sometimes mixed into the troposphere through StratosphericeTropospheric Exchange (STE) (e.g. Lelieveld et al., 2009).
258
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
There are very few studies of ozone in the Arabian Gulf region (Li et al., 2001; Lelieveld et al., 2009; Zanis et al., 2014). While they all concur that there is unusually high ozone in the mid to upper troposphere in the Middle East region, they do not agree on the reasons for it. Li et al. (2001) focus on a study in July 1997. They conducted sensitivity simulations to determine contributions to ozone at various levels within the troposphere, as well as distinguish between the influence of anthropogenic and lightning sources. Li et al. claim that there are unusually high summertime levels of ozone, over 80 ppbv in the mid to upper troposphere in the Middle East region. The reason for this is the anti-cyclonic circulation over the Arabian Peninsula and the Indian sub-continent “funneling” in pollution from Europe and nitrogen oxides (NOx) generated by lightning outflow from the Indian Monsoon. Pollution from eastern Asia transported in the Easterly Tropical Jet Stream is said to contribute as well. Results indicate that the largest source of ozone (35e50%) is due to production in the upper troposphere and largescale subsidence in the region. Losses in these areas due to midlevel outflow are only little according to the Li et al. study. They estimate that 20e30% of the tropospheric ozone column has been caused by anthropogenic sources and 10e15% is from lightning. According to the paper, stratospheric ozone is not thought to be a major contributor, although it seems to have a greater impact than in other regions of the world. Li et al.'s study used the GEOS-CHEM (Goddard Earth Observing System) global 3-D model of atmospheric chemistry in conjunction with vertical ozone profiles collected by the MOZAIC (Measurement of Ozone and Water Vapour on Airbus in-service Aircraft; MOZAIC, 2014) program over the regions of Tel-Aviv, Dubai, and Tehran. Data from the NASA (National Aeronautics and Space Administration) Data Assimilation Office for the years 1993e1997 were used as input for the model along with an anthropogenic base emissions inventory from 1985, adjusted for the study period, which included NOx, Non-Methane Hydrocarbons (NMHC), and Carbon Monoxide (CO). Lightning-NO production associated with deep convection was parameterized following Price and Rind (1992). Lelieveld et al. (2009) postulate that the Middle East is an ozone “hotspot” due to long-range transport of pollutants as indicated by the tracer CO, unusually strong STE, substantial natural upwind sources of NOx such as lightning, the lack of precipitation, and contribution from local emissions such as those from oil and gas refineries combining to create ideal conditions for ozone production and entrainment. They find that there is a distinct ozone maximum between the surface and 750 hPa, which is even more pronounced in the summertime when conditions favor photochemistry. The average ozone mixing ratio in the mid to upper troposphere in the summer is around 80 ppbv. A comparison with other subtropical areas showed that the diel variation of midtropospheric ozone in Bahrain was related more to the longrange transport of pollutants than to local production, and that after removing anthropogenic sources from the model, the region still had higher ozone than the other areas. Unlike Li et al. (2001), this study suggests that stratospheric ozone does have a major impact on the regional tropospheric ozone column, making up about two thirds of it in winter and one quarter in the summer. The Lelieveld study uses the EMAC (European CentereHamburg 5th generation model MESSy [Modular Earth Submodel System] Applied to Atmospheric Chemistry) model in conjunction with satellite-retrieved ozone imagery from the Tropospheric Emission Spectrometer (TES, 2014) and Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY, 2014) to identify where the highest ozone concentrations are. The model is nudged toward the meteorological conditions in 2006 based on re-
analysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) and uses EDGAR 3.2 (Emissions Database for Global Atmospheric Research 2000) for inputs of anthropogenic emissions. The model includes a stratospheric ozone tracer as well and follows its transport and destruction in the troposphere. However, it does not include recycling processes of that stratospheric ozone tracer in the troposphere. The most recent study by Zanis et al. (2014) indicates that the enhanced tropospheric summer ozone levels in the Eastern Mediterranean and Middle East are the result of stratospheric ozone being transported to the troposphere through subsidence induced by high pressure systems in the region. While local photochemical formation is still the dominant contributor to ozone concentration, stratosphere to troposphere transport (STT), a type of STE, plays a critical role in places with favorable conditions. The Zanis et al. (2014) research was conducted in a similar way to the previous Lelieveld et al. (2009) study, using the EMAC model with ECMWF inputs and a stratospheric ozone tracer, and comparing the results to TES satellite data, but this time for a 12 year climatological study from 1998 to 2009. Both the model and satellite show pools of higher ozone concentration in the upper and middle troposphere over the Eastern Mediterranean, with the stratospheric ozone tracer indicating a 40e45% contribution from the stratosphere to the middle troposphere. This is attributed to the large-scale subsidence and limited outflow resulting from the anticyclonic motion of the high pressure systems prevalent in the area during the summer months. Reid et al. (2008) report about dust profiles obtained through aircraft measurements, mostly over the area of the United Arab Emirates and vicinity, and also give a detailed overview of meteorological patterns in the region, including the mention of a subtropical subsidence inversion. While the focus of the paper was on aerosols and dust in the region, their flights encountered polluted air masses from Europe, India, and possibly Africa as well. All of these publications agree that the Middle East area is of specific interest due to some unique meteorological conditions and presumably elevated ozone levels which may have a large-scale impact and contribute to global warming as the ozone is exported from the region. Due to the scarcity of data sets in that area, they all call for more in situ measurements. In this paper we present data analysis from ozonesondes launched from Doha, Qatar, in August 2013 and explore potential source areas for ozone in the upper troposphere. 2. Methods 2.1. Data collection Electro Chemical Concentration (ECC) ozonesondes (Droplet Measurement Technologies, 2013) in conjunction with iMet-1 radiosondes (InterMet Systems, 2008) attached to a 1.2 kg Kaymont brand weather balloon were used. Sondes were equipped with GPS (Global Positioning System). The iMet-1 radiosonde system (InterMet Systems, 2013) provides data for pressure (accuracy: 0.5 hPa), temperature (accuracy: 0.2 C), relative humidity RH (accuracy: 5% RH), wind speed (accuracy: 0.1 m s1), wind direction (accuracy: 5 for wind speeds <14 m s1, and 2 for wind speeds >14 m s1), altitude (accuracy: 15 m), and position (accuracy: 5 m). The principle of ECC sondes are described thoroughly in Russell III et al. (1998). ECC sondes were used because they routinely outperform other types of ozonesondes such as the Brewer-Mast and KC96 Carbon Iodine Cell, as demonstrated in the 1996e2000 Jülich OzoneSonde Intercomparison Experiment (JOSIE) (Smit and Kley, 1996; Smit et al., 2007) and are currently the most commonly used world-wide (Global Atmospheric Watch, 2013).
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Depending on the launch conditions and amount of helium, the balloon rose at a rate of 2e5 m s1 and reached around 30 km above sea level (asl) before burst. For the altitude ranges considered in this study the accuracy of the ECC sondes varies from ±5% (1000 hPa) to ±12% (200 hPa) and the precision from ±4% (1000 hPa) to ±12% (200 hPa) (Droplet Measurement Technologies, 2013). Smit et al. (2007) report even slightly lower values. Comparison studies with routine lidar measurements made at the Observatoire de Haute Provence during 1990e1995 show that for fifteen simultaneous ECC versus lidar profiles, the mean of the differences observed between 4 and 7 km was 2.5 ± 1.8 ppb (4 ± 3%) (Ancellet and Beekmann, 1997). In the period from 4 to 21 August 2013, 20 ozonesondes were launched from the College of the North AtlanticeQatar campus in Doha (25.36 N, 51.48 E) at various times of day, including early morning (7:00 LT (local time)), mid-day (between 12:00e13:00 LT), and night (23:00e24:00 LT). Local time is UTC (Universal Time Code) þ 2 h. Data validation and verification were done in accordance with standardized methods listed in the U.S. Environmental Protection Agency's Manual EPA-QA/G-8. (US Environmental Protection Agency, 2002)
3. Observations The radiosonde data consistently show a marked change in temperature, relative humidity, wind direction, and potential temperature at an altitude of about 5.5 km asl e factors indicative of a strong inversion. Previous observational and modeling studies (Li et al., 2001; Reid et al., 2008; Lelieveld et al., 2009) have shown that this phenomenon at about 5 km is in fact a subtropical subsidence inversion, which seems to occur persistently during the summer months. This is also seen in regular radiosonde launches from other locations in the Middle East area, e.g. Al-Qaisumah and Riyadh (Saudi-Arabia), Abu Dhabi (United Arab Emirates) and Kuwait City (Kuwait) (not shown, but retrieved from the University of Wyoming global radiosonde depository at http://www.weather. uwyo.edu/upperair/mideast.html). Fig. 1 displays the average ozone profile, including the 25% and 75% percentiles, for all 20 Doha ozonesondes launched from 4 to 21 August 2013, segregated into 100 m altitude bins. This figure shows that above the subsidence inversion, ozone begins to increase significantly within the free middle troposphere, peaking around 8 km altitude range and then decreasing again before entering the stratosphere between 14 and 16 km (see Table 1 for more information). This is fairly in agreement
259
with modeled height of tropopause crossings for the Middle East region by Tyrlis et al. (2014). Ozone in the troposphere is mainly formed locally near the surface in complex reactions involving NOx and VOC precursors, however it has been found that pollution in the surface layer can reach upper tropospheric levels, due to strong convective processes (Baehr et al., 2003; Borbon et al., 2012). The persistent subsidence inversion over the Middle East area during the summer acts largely to suppress convective processes and the upper troposphere is therefore decoupled from the lower troposphere (Li et al., 2001; Reid et al., 2008; Lelieveld et al., 2009). As in-situ formation of ozone is limited in the upper troposphere, ozone aloft is likely due to transport processes. Apart from horizontal transport, vertical transport (e.g. lifting or subsidence) can play an important role. Through STE processes, stratospheric ozone can also contribute. Previous research has postulated that ozone in the mid to upper troposphere is being transported from Europe and the Mediterranean (Li et al., 2001). Henceforth we focus the discussions on the ozone distribution in the upper troposphere to trace the potential source areas of the elevated ozone found in the mid to upper troposphere between 6 and 12 km over the Arabian Gulf area. For this purpose we consider mid-day launches only and do not include early morning launches done in addition after 15 August. This way there is consistency of only one launch per day at approximately the same time. From the remaining 14 mid-day launches we needed to remove one launch (09 August) due to problems in the GPS data acquisition which led to unreliable wind data. In the following data analysis we therefore focus on an overall dataset of 13 mid-day launches. For data analysis we applied a series of 72 h HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory Model) (Draxler and Rolph, 2013) back trajectories at altitudes of 6, 9, and 12 km, based on the GPS coordinates of each ozonesonde when it was at these altitudes. In order to identify meteorological patterns for the month of August, the World Wide Lightning Location Network (WWLLN) lightning data (WWLLN Management Team, 2013), as well as National Oceanic and Atmospheric Administration (NOAA) surface, 200 and 500 hPa weather analysis charts of the region (NOAA National Climatic Data Center, 2013) were used to determine what factors might have influenced the long-range transport of ozone. 3.1. HYSPLIT back trajectories Fig. 2 shows the results of the HYSPLIT back trajectory analyses. Of the 13 launches used in this study, 6 days between 5 and 13 August showed trajectories coming from the East, from northern India and the Himalayas. On two days, 14 and 15 August, the trajectories curled around locally, indicating a shift in large-scale wind flows. After 15 August, the remaining 5 set of trajectories came from the northwest, the Mediterranean and North Africa. Uncertainties of the HYSPLIT trajectories are assumed to be in the range of 15e30% of the travel distance (Draxler and Rolph, 2013). Although, there might be some limitations associated with the accuracy of backward trajectories, they at least may provide information about major regimes. The HYSPLIT back trajectories showed generally that each air packet ended up within approx. 20e25% of its initial altitude. 3.2. Weather charts
Fig. 1. Average ozonesonde profiles in the time period 4e21 August 2013. Data are all 20 ozonesonde launches segregated into 100 m altitude bins.
Fig. 3 displays 500 hPa and 200 hPa weather charts for the Middle East area. The surface level analysis and lifted index charts are characterized by a constant “heat low” in the region due to high surface temperatures. This is capped by high pressure, indicated by
260
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Table 1 Median, mean, standard deviation, and range of measured ozone for all 13 launches at various altitudes. Launch times close to local noon. Layer
Upper troposphere (6 / 12 km) Entire troposphere (0 / 14 km) Surface layer (0 / 0.5 km) a
Numbera
Ozone [ppbv] Median
Mean
Standard deviation
Maximum
Minimum
79.4 72.4 66.4
80.0 71.6 64.3
12.8 14.4 14.8
109 109 95
46 27 29
2878 ± 183 5767 ± 355 165 ± 25
Number of data points per launch.
elevated geopotential heights on the 500 hPa chart, preventing convective lifting, and resulting in thin altostratus clouds at about 6 km altitude most days. On 12 and 13 August, a stronger ridge moved in over Qatar, indicated by the yellow arrow on the 500 hPa chart. This ridge began to move out of the Qatar region on 14 August and had receded completely by 15 August. At the 200 hPa level, there is a persistent powerful jet streak north of Qatar between 40 and 45 N from the beginning of August until 15 August, when it suddenly disappears, after which an upper level trough forms, pulling in air from the Mediterranean. 3.3. Lightning data Lightning data for the northern hemisphere for the month of August was obtained from the World Wide Lightning Location Network (WWLLN; http://wwlln.net/new/network/) and plotted in Google Earth. These data are gathered by a global network of sensing stations using VLF (Very Low Frequency between 3 and 30 KHz) to detect “sferics”delectromagnetic discharges caused by lightning. With a current network of 40 stations, the WWLLN is capable of detecting only about 30% of all lightning strikes. This may be considered as an upper limit; it may be likely less in the in the area of this study. Due to this limitation, any lightning activity which is detected can be considered a lower limit and the actual lightning activity may be largely underestimated, and therefore the potential lightning-influenced ozone production in these areas as well. The Indian Summer Monsoon took place during August, and there was in fact heavy lightning activity over India, often along the HYSPLIT trajectories. There was however also considerable lightning in the Middle East, even in Qatar itself on some days. In particular, on 6 August the country experienced some thunderstorms (Fig. 4). 3.4. Balloon data As illustrated in Table 1, the average measured ozone for all 13 launches between 6 and 12 km was 80 ± 13 ppbv, with a median of ~79 ppbv. This is higher than the average total tropospheric ozone between the surface and 14 km of 72 ± 14 ppbv. The average ozone below 500 m was ~64 ± 15 ppbv. Note that for consistency, and to ensure none of the data were above the boundary layer, 500 m was used as a representation of surface ozone, rather than the entire planetary boundary layer which varies between 1 and 2 km. All launches were characterized by pronounced changes in relative humidity above 5.5 km as shown exemplary in Fig. 5. Unlike the launches before 15 August which showed only weak temperature inversions, the launches after 15 August all showed a marked temperature inversion of 2e3 C over a distance of less than 0.5 km at around 5.5 km altitude. On 7 of the 13 days, the ozonesondes measured maximum ozone over 100 ppbv between 6 and 12 km. Three of these launches took place before 15 August, at which point the large scale synoptic condition changed and thus did the wind direction throughout
most of the troposphere. On 6 August a peak ozone mixing ratio over 100 ppbv occurred at an altitude above 12 km. On 12 August the peak ozone mixing ratio was 105 ppbv between 8 and 10 km, and on 13 August a nearly identical “plume” between 6 and 8 km (Fig. 6) can be visually discerned. 4. Discussion In this section we will focus on three examples to explain the causes of high ozone concentrations in the Arabian Gulf region. Each case has specific characteristics which can indicate the likely source of the ozone peaks. 4.1. Case 1: ozonesonde launch on 6 August and the effect of lightning In this study we do not verify a distinct impact on ozone due to lightning related to the Indian Monsoon as this would require a comparison with transport processes from the Indian subcontinent under non-Monsoon conditions. However, this study shows that lightning from the Indian Monsoon is not associated with as high of ozone values at upper-tropospheric levels in the Arabian Gulf region as are observed when back trajectories point to the Eastern Mediterranean region. Air masses originating from the Eastern Mediterranean region contain about 30e40 ppb higher ozone values than air masses impacted by the Indian Monsoon and associated lightning activities as shown in Fig. 8. Fig. 8 also demonstrates that these different ozone regimes are clearly accompanied by different relative humidity regimes. While high relative humidity values can be expected for air masses impacted by the Indian Monsoon, it is remarkable that air masses originating from the Eastern Mediterranean contain extremely low relative humidity for the entire preceding 72 h. This distinct feature is visible in all the other profiles of the corresponding specific synoptic conditions. As mentioned before, the days 14 and 15 August were transitional days between these large-scale flows, which also showed transition of the relative humidity profiles. The Eastern Mediterranean is a well-known hot-spot for enhanced tropospheric ozone levels (e.g. Kouvarakis et al., 2002; Kourtidis et al., 2002; Lelieveld et al., 2002, 2009). This is largely due to large-scale subsidence in the summer (Zanis et al., 2014; Tyrlis et al., 2014), low tropopause height, and enhanced tropopause folding activity over the Levantine region during the summertime, in connection with the Asian monsoon (Tyrlis et al., 2014). The numerical simulations by Zanis et al. (2014) indicate that there is a large contribution of stratospheric ozone to the pool of high ozone values over the Eastern Mediterranean/ Middle East in the middle and lower free troposphere and Sprenger et al. (2007) showed that stratospheric air mixed into the upper troposphere can remain there up to a few days. The extremely low relative humidity values in air masses coming from the Eastern Mediterranean area are in support of these previous modeling studies. Apart from these ozone transports in the upper troposphere, however, also local thunderstorm activity in the Arabian Gulf could
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
261
Fig. 2. HYSPLIT back trajectories in Google Earth (Google Inc., 2013). Altitude color coding: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
262
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 3. NOAA 12:00UTC weather charts sequence for 11 to 14 August depicting a ridge over Arabian Peninsula on 12 and 13 August 2013 on the 500 hPa charts as indicated by the yellow arrow (left panel, top to bottom) and jet streak movement (shown in green) at 40e45 N around 15 August 2013 on the 200 hPa charts from 13 to 16 August (right panel, top to bottom). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
be contributing to higher ozone. On 6 August there was considerable lightning activity in the entire region, about 50% more than on other days before 15 August, and the HYSPLIT back trajectory for the ozonesonde, which was launched that day, either passed directly
through (trajectory at 9 km altitude in Fig. 9) or close to a nearby thunderstorm in the Arabian Gulf with correspondingly increased ozone at 12 km (Fig. 9). It should be noted that the location, altitude, spatial extension of the lightning activity as well as the
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
263
Fig. 4. Lightning on 6 August 2013 (WWLLN Management Team, 2013).
Fig. 5. Profiles of ozone, relative humidity, air temperature, potential temperature, wind speed, and wind direction as obtained from the ozonesonde on 19 August 2013 at 10:10 UTC (12:10 LT).
264
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 6. Above: Comparison of vertical ozone profiles on 12 and 13 August 2013. Note the plume height at 9e10 km on 12 August and 8 km on 13 August. Below: Vertical profiles of ozone and relative humidity in the altitude range 6e12 km.
number of strikes is indicative for the overall lightning activity. As mentioned above the WWLLN data may only detect about 30% of all lightning strikes. In addition, uncertainties associated with the back trajectories are in the range of 15e30% of the travel distance (Draxler and Rolph, 2013). Thus it is likely that both trajectories, at 9 km and at 12 km altitude, might have been impacted by lightning. The peak ozone mixing ratio that day was 105 ppbv at 12 km, with a
smaller 101 ppbv peak at 10 km (Fig. 9). This is significantly higher than on the preceding day when under the same flow conditions about 50e60 ppbv ozone were observed from top of the boundary layer height around 2 km altitude up to 10e14 km altitude. Lightning is an important source of NOx in the upper troposphere worldwide (Schumann and Huntrieser, 2007; Huntrieser et al., 2012), including the Middle East area (Li et al., 2001) and the balloon's
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
265
Fig. 7. Above: Altitude plot of the 72-h backward trajectory ending at the location of the ozonesonde at about 9 km altitude. Below: relative humidity along this trajectory.
Fig. 8. Comparison of high lightning and low ozone (5 August 2013, 10:07 UTC (12:07 LT); upper left plot and blue line in graph) with low lightning and high ozone days (19 August 2013, 10:10 UTC (12:10 LT); upper right plot and red line in graph). Lightning data for 5 Aug (blue), 4 Aug (pink), 3 Aug (green), and 2 Aug (yellow); Lightning data for 19 Aug (blue), 18 Aug (pink), 17 Aug (green), and 16 Aug (yellow). Altitude color coding: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. In addition, vertical profiles of relative humidity on the corresponding days as well as along 72-h backward trajectories ending at the ozonesonde location at about 9 km altitude are shown. Upper RH trajectory is for 5 August, lower RH trajectory for 19 August. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
266
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Fig. 9. Above: lightning and HYSPLIT back trajectories on 6 August 2013, 10:00 UTC (12:00 LT), highlighting the 9 km trajectory passing through the lightning storm. The start point of the back trajectory indicates the position of the ozonesonde at that given time. Altitude color coding for back trajectories: red 6 km, blue 9 km, green 12 km altitude of the ozonesonde upon arrival time of the air parcel. Below: vertical ozone profile on 6 August 2013.” (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
actual trajectory took it directly downwind from this source. Huntrieser et al. (2007) report up to 60e70 nmol mol1 enhancement of ozone in the anvil outflow over a horizontal scale of about 400 km. This also due to the fact that the atmospheric lifetime of NOx in the upper troposphere is significantly longer than in the boundary layer and that photochemical processes are largely NOx limited (Huntrieser et al., 2007 and references therein). The amount of NOx produced also depends on the lightning stroke length, peak current, release height, and vertical wind shear (Huntrieser et al., 2008). Fig. 9 shows that ozone exceeds 60 ppbv, the preceding day's average tropospheric value observed under similar Easterly flow conditions, around 7 km altitude and then continuously increases and reaches maximum values. This coincides with strongest vertical wind shears between 9 and 12 km, which encompasses the top of the cloud deck, as reflected in the relative humidity data (not
shown). Ozone mixing ratios averaged over 9e12 km altitude were 66.8 ± 2.2 ppbv on 5 August versus 91.2 ± 2.4 ppbv on 6 August, which corresponds to an increase of about 36.5% over this altitude range. These values rank among higher values found in other studies on lightning impact on ozone in the upper troposphere (Cooper et al., 2007; Morris et al., 2010 and references therein). In conclusion, this suggests that ozone produced as a result of this lightning activity may have contributed appreciably to the concentrations measured by the ozonesonde over Qatar. 4.2. Case 2: ozonesondes launched on 12 and 13 August The 12 and 13 of August were the only two days of the month during which a strong ridge was centered over the Arabian Peninsula, as shown on the 500 hPa chart in Fig. 3. Zanis et al.
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
267
Table 2 Mean and standard deviation of measured ozone at various altitudes separated into different source regions according to the 72 h back trajectories as shown in Figs. 2 and 7. The transition days August 14-15 (see Fig. 2) are not included. Launch times close to local noon. Ozone in ppbv. Layer
Upper troposphere (6 / 12 km) Entire troposphere (0 / 14 km) Surface layer (0 / 0.5 km) a b
Launches August 5e13a (“East”)
Launches August 16e21b (“West”)
Mean
Standard deviation
Mean
Standard deviation
74.0 66.1 61.3
6.9 5.0 2.1
86.2 77.6 66.7
8.9 3.0 4.5
Based on 6 ozonesondes. Based on 4 ozonesondes.
(2014) discuss how the anti-cyclonic activity associated with high pressure systems causes ozone-rich air to be pulled downward isentropically out of the stratosphere into the upper troposphere through subsidence in the Eastern Mediterranean and Middle East region. Their study used a stratospheric ozone tracer in the model which showed the air mass descending over time. The observations on 12 and 13 August show an example for subsidence, as the ozone plume is between 8 and 10 km on 12 August, and has fallen to between 6 and 8 km only 24 h later, yet maintains its distinct shape (see section highlighted by the rectangle in Fig. 6). In Fig. 6 we included the nighttime ozonesonde launch on 12 August so that this figure displays a sequence of ozonesondes almost every 12 h apart. Note that the ozone values between 6 and 12 km altitude are significantly higher than the average ozone profiles at these altitude ranges (see Fig. 1). Both days show subsidence inversions (not shown) at similar altitudes (5.5 km on 12 August and 6 km on 13 August, i.e. close to the 500 hPa level). These inversions are associated with wind shear, mostly northerly below, whereas aloft winds come from the east throughout the upper troposphere coincident with the large scale synoptic situation as shown in Fig. 3. Both inversions are also associated with generally decreasing relative humidity aloft, typical for subsidences. Also, starting with these inversions ozone mixing ratios increase with increasing altitude reaching maximum values between 8 and 11 km altitude on 12 August and between 6 and slightly above 8 km altitude on 13 August. The levels with enriched ozone are confined by another thermal inversion at about 12 km altitude on 12 August and slightly above 8 km altitude on 13 August. On 12 August the layer of maximum ozone is accompanied by low relative humidity around 10e15%, similar to the relative humidity values observed in the lower stratosphere on that day. Due to these overall low relative humidity values the thermal inversion about 12 km altitude does not show a pronounced change in relative humidity. This is in contrast to 13 August, when the lower thermal inversion can be clearly identified as a subsidence inversion with decreasing relative humidity values aloft. This is also reflected in the HYSPLIT analysis for that day (Fig. 7), which shows that air masses arriving at the ozonesonde location at about 9 km altitude have originated above 11.5 km altitude and experienced a drastic decrease in relative humidity along the trajectory. While on 13 August the relative humidity profile shows some variation within the layer between 6 km and slightly above 8 km altitude ozone mixing ratios are largely anticorrelated with relative humidity in that layer indicative for sustaining similar atmospheric properties from the previous day. The anticorrelation of ozone with relative humidity can be best seen in the nighttime ozone profile on 12 August between ~8 km and ~9 km altitude. The main difference between the vertical meteorological profiles of those two days is the change in the upper tropospheric subsidence inversions from around 12 km (12 August) to slightly above 8 km altitude (13 August) within 24 h, which would translate into an entrainment of enriched ozone layers from upper tropospheric areas on the order of 4e5 cm s1 estimated as an upper boundary. Based on the
HYSPLIT analysis an entrainment velocity on the order of 1 cm s1 can be estimated. While on 12 August ozone mixing ratios were about 93.0 ± 11.8 ppbv averaged over the altitude range of 8e11 km, they decreased to 74.3 ± 9.4 ppbv on the subsequent day for the same altitude range, which is a decrease by ~20%. On the other hand, ozone mixing ratios averaged over 6e7 km were 66.1 ± 4.2 ppbv on 12 August and increased to 77.9 ± 12.5 ppbv on 13 August (increase by ~18%). The altitude range 7e8 km represents a transitional region with limited change as the plume of enriched ozone passes this region during subsidence. This subsidence process will be accompanied by horizontal dispersion, which is reflected by the fact that the ozone mixing ratios in the ozone lamina slightly decrease and are confined to a shrinking altitude range.
4.3. Case 3: ozonesonde launched on 19 August The majority of cases of high ozone concentrations in the region during the measurement period occur when the HYSPLIT back trajectories come from the Mediterranean, as exemplified by the profile of the ozonesonde launched on 19 August (Fig. 8), and corroborated by all launches on and after 15 August. Here the ozone begins to increase at a lower altitude, around 7 km, and remains high up to 12 km. In these instances the ozone peak is about 100 ppbv, and most likely the result of long-range transport of pollutants (Li et al., 2001) in the 6e12 km range (72 h back trajectories in this altitude range point towards the Eastern Mediterranean). The strong temperature inversion at 6 km acts as a boundary layer, preventing mixing from below or transfer from aloft. Between 11.5 and 12 km altitude wind direction changes by about 180 (not shown) which coincides with a sharp drop in ozone mixing ratios and which caps the high ozone levels observed beneath, i.e. in the 6e12 km range. Table 2 shows general results of the two large-scale flow regimes observed during the measurement period as evident in Figs. 2 and 8. The flow regime denoted as “East” refers to the time period impacted by Monsoon activity over India, whereas the flow regime denotes as “West” refers to the source region over the Eastern Mediterranean. From Table 2 it is evident that Westerly flow regimes are associated with higher ozone mixing ratios than Easterly flow regimes during the measurement period of August 2013. Compared to Easterly flow conditions, ozone values in Westerly flow conditions are enhanced by about 17% through the entire troposphere and in particular in the 6e12 km altitude range, which is double the enhancement in the surface layer. It is noteworthy to keep in mind that ozone values under Easterly flow conditions are impacted by thunderstorm activities related to the Monsoon season. As a consequence an even higher ozone enhancement in Westerly versus Easterly flow regimes in the 6e12 km altitude range may have been masked by Monsoon lightning activities. The observed ozone enhancement associated with back trajectories point towards source regions over the
268
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269
Eastern Mediterranean and support recent model results (Lelieveld et al., 2009; Zanis et al., 2014). 5. Conclusions During August 2013, 20 ozonesondes were launched from Doha, Qatar. A subset of 13 mid-day launches were further analyzed. Although limited to one month in one summer season some interesting observations were obtained. Upper tropospheric ozone in the Arabian Gulf region in August 2013 was on average 80 ± 13 ppbv, as predicted by models used in previous studies (Li et al., 2001; Lelieveld et al., 2009); however the source of high ozone is more in agreement with Lelieveld et al. (2009) and Zanis et al. (2014) than postulated by Li et al. (2001). The extremely low relative humidity of a few percent observed in air masses originating from the Eastern Mediterranean in our study is in support of the modeling studies by Lelieveld et al. (2009) and Zanis et al. (2014) who suggested high ozone levels in the middle and upper troposphere as a result of stratosphereetroposphere transport over the Eastern Mediterranean and subsequent transport from this area to the Arabian Gulf under continued subsidence processes. Elevated ozone levels in the upper troposphere occur because of the meteorological conditions allowing precursors to be transported into the area, and as a result of subsidence. When the wind flow in the upper troposphere is from the East, the ozone levels are about 17% lower in the 6e12 km range than when the wind is from the Mediterranean. High pressure may lead to subsidence of ozone from the upper troposphere/lower stratosphere and eventually cause an increase of ozone mixing ratios by ~18% between 6 and 7 km, i.e. slightly above the subtropical subsidence layer. Under the impact of regional convective activity and associated lightning, ozone mixing ratios can increase by more than 35% averaged over the 9e12 km altitude range. In either case maximum ozone in the mid to upper troposphere reached more than 100 ppbv. This study was limited to 13 launches in the month of August. There are no seasonal comparisons. More work needs to be done in order to learn more about the temporal variability in regional ozone patterns and long-term trends. Acknowledgments We are grateful for the support provided by Qatar Foundation and the Qatar Civil Aviation Authority. Logistical and technical support provided by Messrs. M.A. Ayoub, L. Ackermann, and Mrs. M. Shalaev from QEERI is greatly appreciated. The research results have not been subject to QEERI scientific and policy review and therefore do not necessarily reflect the views of QEERI and no official endorsement should be inferred. Many thanks also to Dr. B. Holzworth for providing the WWLLN data for the time period of the study. References Ancellet, G., Beekmann, M., 1997. Evidence for changes in the ozone concentration in the free troposphere over Southern France from 1976e1995. Atmos. Environ. 31, 2835e2851. Baehr, J., Schlager, H., Ziereis, H., Stock, P., van Velthoven, P., Busen, R., Stroem, J., Schumann, U., 2003. Aircraft observations of NO, NOy, CO, and O3 in the upper troposphere from 60N to 60S e interhemispheric differences at mitlatitudes. Geophys. Res. Lett. 30 (11), 1598. Borbon, A., Ruiz, M., Bechara, J., Aumont, B., Chong, M., Huntrieser, H., Mari, C., Reeves, C.E., Scialom, G., Hamburger, T., Stark, H., Afif, C., Jambert, C., Mills, G., Schlager, H., Perros, P., 2012. Transport and chemistry of formaldehyde by mesoscale convective systems in West Africa during AMMA 2006. J. Geophys. Res. 117, D12301. Cooper, O.R., Trainer, M., Thompson, A.M., Oltmans, S.J., Tarasick, D.W., Witte, J.C., Stohl, A., Eckhardt, S., Lelieveld, J., Newchurch, M.J., Johnson, B.J.,
Portmann, R.W., Kalnajs, L., Dubey, M.K., Leblanc, T., McDermid, I.S., Forbes, G., Wolfe, D., Carey-Smith, T., Morris, G.A., Lefer, B., Rappenglück, B., Joseph, E., Schmidlin, F., Meagher, J., Fehsenfeld, F.C., Keating, T.J., Van Curen, R.A., Minschwaner, K., 2007. Evidence for a recurring eastern North America upper tropospheric ozone maximum during summer. J. Geophys. Res. 112, D23304. http://dx.doi.org/10.1029/2007JD008710. Draxler, R., Rolph, G., 2013. HYSPLIT (HYbrid Single-particle Lagrangian Integrated Trajectory) Model Access. Retrieved 08 27, 2013, from NOAA ARL READY. http:// www.arl.noaa.gov/HYSPLIT.php. Droplet Measurement Technologies, 2013. ECC Ozonesonde. Retrieved August 27, 2013, from Droplet Measurement Technologies. http://www.dropletmeasurement.com/ products/airborne/ECC_Ozonesonde#Specifications. Global Atmospheric Watch, 2013. Quality Assurance and Quality Control for Ozonesonde Measurements in GAW. Report #201. World Meteorological Organization, Geneva. Google Inc, 2013, July 12. Google Earth 7.1.1.1888 [Software]. Huntrieser, H., Schlager, H., Roiger, A., Lichtenstern, M., Schumann, U., Kurz, C., Brunner, D., Schwierz, C., Richter, A., Stohl, A., 2007. Lightning-produced NOx over Brazil during TROCCINOX: airborne measurements in tropical and subtropical thunderstorms and the importance of mesoscale convective systems. Atmos. Chem. Phys. 7, 2987e3013. €ller, H., Giez, A., Betz, H.-D., Forster, C., Huntrieser, H., Schumann, U., Schlager, H., Ho Pinto Jr., O., Calheiros, R., 2008. Lightning activity in Brazilian thunderstorms during TROCCINOX: implications for NOx production. Atmos. Chem. Phys. 8, 921e953. € ller, H., Grewe, V., 2012. Thunderstorms: trace species generators. Huntrieser, H., Ho In: Schumann, U. (Ed.), Atmospheric Physics: Background e Methods e Trends Research Topics in Aerospace. Springer-Verlag, ISBN 978-3-642-30182-7, pp. 115e133. InterMet Systems, Inc, 2008. iMet-1 Product. Retrieved August 27, 2013, from International Met Systems. http://www.intermetsystems.com/InterMet_1.html. InterMet Systems, Inc, 2013. Technical Description for the iMet-3050 and 3150 403 MHz GPS Upper-air Sounding Systems and the iMet-1 Radiosonde. Document No. 200,731 Revision-D, 37 pages. Kourtidis, K., Zerefos, C., Rapsomanikis, S., Simeonov, V., Balis, D., Perros, P.E., Thomson, A.M., Witte, J., Calpini, B., Sharobiem, W.M., Papayiannis, A., Mihalopoulos, N., Drakou, R., 2002. Regional levels of ozone in the troposphere over eastern Mediterranean. J. Geophys. Res. 107, 8140. http://dx.doi.org/ 10.1029/2000JD000140. Kouvarakis, G., Vrekoussis, M., Mihalopoulos, N., Kourtidis, K., Rappenglueck, B., Gerasopoulos, E., Zerefos, C., 2002. Spatial and temporal variability of tropospheric ozone in the boundary layer above the Aegean Sea (eastern Mediterranean). J. Geophys. Res. 107, 8137. http://dx.doi.org/10.1029/ 2000JD000081. Lelieveld, J., Berresheim, H., Borrmann, S., Crutzen, P.J., Dentener, F.J., Fischer, H., de Gouw, J., Feichter, J., Flatau, P., Heland, J., Holzinger, R., Korrmann, R., Lawrence, M., Levin, Z., Markowicz, K., Mihalopoulos, N., Minikin, A., Ramanathan, V., de Reus, M., Roelofs, G.-J., Scheeren, H.A., Sciare, J., Schlager, H., Schultz, M., Siegmund, P., Steil, B., Stephanou, E., Stier, P., Traub, M., Williams, J., Ziereis, H., 2002. Global air pollution crossroads over the Mediterranean. Science 298, 794e799. € ckel, P., Pozzer, A., Hadjinicolaou, P., Cammas, J., Beirle, S., Lelieveld, J., Hoor, P., Jo 2009. Severe ozone air pollution in the Persian Gulf region. Atmos. Chem. Phys. 9, 1393e1406. Li, Q., Jacob, D.J., Logan, J.A., Bey, I., Yantosca, R.M., Hongyu, L., Martin, R.V., Fiore, A.M., Field, B.D., Duncan, B.N., Thouret, V., 2001. A tropospheric ozone maximum over the middle east. Geophys. Res. Lett. 28 (17), 3235e3238. MOZAIC (Measurements of OZone, water vapour, carbon monoxide and nitrogen oxides by in-service AIrbus airCraft), 2014. Retrieved from http://www.iagos.fr/ web/rubrique2.html. Morris, G.A., Thompson, A.M., Pickering, K.E., Chen, S., Bucsela, E.J., Kucera, P.A., 2010. Observations of ozone production in a dissipating tropical convective cell during TC4. Atmos. Chem. Phys. 10, 11189e11208. http://dx.doi.org/10.5194/acp10-11189-2010. NOAA National Climatic Data Center, 2013. Service Records Retention System (SRRS). Retrieved from Analysis and Forecast Charts. http://nomads.ncdc.noaa. gov/ncep/NCEP. Price, C., Rind, D., 1992. A simple lightning parameterization for calculating global lightning distributions. J. Geophys. Res. 97, 9919e9933. Reid, J.S., Piketh, S.J., Walker, A.L., Burger, R.P., Ross, K.E., Westphal, D.L., Bruintjes, R.T., Holben, B.N., Hsu, C., Jensen, T.L., Kahn, R.A., Kuciauskas, Arunas P., Mandoos, A.A., Mangoosh, A.A., 2008. An overview of UAE2 flight operations: observations of summertime atmospheric thermodynamic and aerosol profiles of the southern Arabian Gulf. J. Geophys. Res. 113, D14213. Russell III, J., Smit, H., Hofmann, D., Rodgers, C., 1998. Stratospheric Processes and their Role in Climate (SPARC): Assessment of Trends in the Vertical Distribution of Ozone. WMO Ozone Res. And Monit. Proj. Rep. 43, (World Clim. Res. Prog., Geneva). Schumann, U., Huntrieser, H., 2007. The global lightning-induced nitrogen oxides source. Atmos. Chem. Phys. 7, 3823e3907. http://dx.doi.org/10.5194/acp-73823-2007. SCIAMACHY e SCanning Imaging Absorption spectroMeter for Atmospheric ChartographY. (2014). Retrieved from http://www.sciamachy.org/. Smit, H.G.J., Kley, D., 1996. Jülich ozone sonde intercomparison experiment (JOSIE). World Meteorol. Organ. Glob. Atmos. Watch 130.
T.K. Spohn, B. Rappenglück / Atmospheric Environment 101 (2015) 257e269 Smit, H., Straeter, W., Johnson, B., Oltmans, S., Davies, J., Tarasick, D.W., Hoegger, B., Stubi, R., Schmidlin, F., Northam, T., Thompson, A., Witte, J., Boyd, I., Posny, F., 2007. Assessment of the performance of ECC-ozonesondes under quasi-flight conditions in the environmental simulation chamber: insights from the Juelich Ozone Sonde Intercomparison Experiment (JOSIE). J. Geophys. Res. 112, D19306. Sprenger, M., Wernli, H., Bourqui, M., 2007. Stratosphereetroposphere exchange and its relation to potential vorticity streamers and cutoffs near the extratropical tropopause. J. Atmos. Sci. 64, 1587e1602. http://dx.doi.org/10.1175/ JAS3911.1. TES (Tropospheric Emission Spectrometer). (2014). Retrieved from http://tes.jpl. nasa.gov/. Tyrlis, E., Skerlak, B., Sprenger, M., Wernli, H., Zittis, G., Lelieveld, J., 2014. On the linkage between the Asian summer monsoon and tropopause fold activity over
269
the eastern Mediterranean and the Middle East. J. Geophys. Res. 119, 3202e3221. http://dx.doi.org/10.1002/2013JD02113. US Environmental Protection Agency (2002). EPA-QA/G-8. Retrieved from http:// www.epa.gov/QUALITY/qs-docs/g8-final.pdf. US Environmental Protection Agency, 2012, November 1. Ground-level Ozone. Retrieved August 27, 2013, from US Environmental Protection Agency. http:// www.epa.gov/glo/health.html. WWLLN Management Team, 2013. World Wide Lightning Location Network (Retrieved from wwlln.net). Zanis, P., Hadjinicolaou, P., Pozzer, A., Tyrlis, E., Dafka, S., Mihalopoulos, N., Lelieveld, J., 2014. Summertime free tropospheric ozone pool over the Eastern Mediterranean/Middle East. Atmos. Chem. Phys. 14, 115e132. http://dx.doi.org/ 10.5194/acp-14-115-2014.