The Polar Stratosphere in a Changing Climate (POLSTRACC) mission employed the German High Altitude and Long Range Research Aircraft (HALO). The payload comprised an innovative combination of remote sensing and in situ instruments. The in situ instruments provided high-resolution observations of cirrus and polar stratospheric clouds (PSCs), a large number of reactive and long-lived trace gases, and temperature at the aircraft level. Information above and underneath the aircraft level was achieved by remote sensing instruments as well as dropsondes. The mission took place from 8 December 2015 to 18 March 2016, covering the extremely cold late December to early February period and the time around the major warming in the beginning of March. In 18 scientific deployments, 156 flight hours were conducted, covering latitudes from 25° to 87°N and maximum altitudes of almost 15 km, and reaching potential temperature levels of up to 410 K. Highlights of results include 1) new aspects of transport and mixing in the Arctic upper troposphere–lower stratosphere (UTLS), 2) detailed analyses of special dynamical features such as tropopause folds, 3) observations of extended PSCs reaching sometimes down to HALO flight levels at 13–14 km, 4) observations of particulate NOy and vertical redistribution of gas-phase NOy in the lowermost stratosphere (LMS), 5) significant chlorine activation and deactivation in the LMS along with halogen source gas observations, and 6) the partitioning and budgets of reactive chlorine and bromine along with a detailed study of the efficiency of ClOx/BrOx ozone loss cycle. Finally, we quantify—based on our results—the ozone loss in the 2015/16 winter and address the question of how extraordinary this Arctic winter was.
The POLSTRACC mission provided a wealth of data on the structure and composition of the Arctic lowermost stratosphere and upper troposphere during the extraordinarily cold winter 2015/16.
The Polar Stratosphere in a Changing Climate (POLSTRACC) campaign was the first comprehensive aircraft mission focusing on the Arctic lowermost stratosphere (LMS) over a whole winter–spring season. Changes of radiatively active trace gases such as ozone and water vapor in the lowermost stratosphere have a relatively large impact on climate forcing (e.g., Riese et al. 2012, and references therein). The coupling of the stratosphere with surface weather and climate has received considerable attention in recent years (e.g., Kidston et al. 2015, and references therein). The Arctic is the region with the largest observed climate change in the last decades (known as Arctic amplification; IPCC 2013). Yet, little is known of how climate change will affect chemical composition and processes in the Arctic stratosphere and how the feedback on surface climate will be.
After the discovery of the Antarctic ozone hole by Farman et al. (1985), the question was raised if comparable ozone depletion could also take place over the Arctic. Indeed, during cold stratospheric Arctic winters, substantial ozone depletion was observed also over the Arctic (e.g., Sinnhuber et al. 2000; Rex et al. 2004; Sinnhuber et al. 2011; WMO 2018), approaching in individual years losses that are comparable to the losses observed in the Antarctic (Manney et al. 2011). Weber et al. (2003) have shown that the interannual variability of high-latitude springtime total ozone columns over both hemispheres are quantitatively controlled by the activity of planetary-scale waves entering the stratosphere. Exceptional ozone columns such as in the disturbed Antarctic winter–spring of 2002 (Sinnhuber et al. 2003) or in the unusually cold Arctic winter–spring of 2010/11 (Sinnhuber et al. 2011) can be understood by exceptional wave activity in these winters.
Cold conditions in the winter polar stratosphere enable the formation of polar stratospheric clouds (PSCs) that serve as surfaces for heterogeneous chemical reactions, which are activating chlorine from the chemical less reactive reservoir compounds hydrochloric acid (HCl) and chlorine nitrate (ClONO2) into reactive ClOx (=Cl + ClO + Cl2O2). ClOx can very efficiently destroy ozone in catalytic cycles once sunlight is available. Polar stratospheric clouds also heterogeneously convert nitrogen oxides (NOx) into nitric acid (HNO3), inhibiting the deactivation of ClOx into chlorine nitrate (sometimes called denoxification). If polar stratospheric cloud particles can grow large enough to sediment efficiently, the nitric acid taken up in or on the PSC particles will be removed irreversibly from parts of the lower stratosphere (so-called denitrification), so that the halogen catalyzed ozone destruction can proceed even if temperatures are not cold enough for ongoing chlorine activation (Waibel et al. 1999; Fahey et al. 2001).
Heterogeneous chlorine activation can also take place on cold liquid aerosols, becoming efficient at temperatures below about 195 K (e.g., Drdla and Müller 2012). At temperatures below about 192 K sulfuric acid aerosol particles can efficiently take up HNO3 so that they can grow considerably in size, becoming supercooled ternary (H2O/H2SO4/HNO3) solution (STS) polar stratospheric cloud particles (Carslaw et al. 1997; Schreiner et al. 1999). They are supercooled, because they coexist at temperatures below about 196 K with thermodynamically favorable nitric acid trihydrate (NAT) solid PSC particles. At temperatures below about 188 K, ice particles can form (Voigt et al. 2000a). At temperatures a few kelvins above the ice frost point Tice (Murphy and Koop 2005), PSCs consisting of solid nitric acid trihydrate particles (NAT), STS droplets, and their mixtures can form, as shown by mass spectrometric composition measurements (Schreiner et al. 1999; Voigt et al. 2000a,b, 2005). While the role of the different PSC particles for chlorine activation is of secondary importance (Kirner et al. 2015), sedimentation of the larger NAT and ice particles becomes important for the irreversible redistribution of nitrogen compounds (denitrification and renitrification; Fahey et al. 2001).
Despite many similarities between ozone losses in cold Arctic winters and the Antarctic ozone hole, there are also fundamental differences (Solomon et al. 2014). In particular, although the degree of the ozone loss may be comparable in certain years, the actual amount of ozone in the Arctic is always larger than in the Antarctic ozone hole.
The connection between polar stratospheric ozone depletion and climate change is twofold, concerning the following questions: 1) How will climate change due to the anthropogenic emission of greenhouse gases affect the polar winter stratosphere? 2) What is the impact of changes in the polar stratosphere on surface climate?
Overall, the stratosphere is cooling in response to the increase in greenhouse gases (Goessling and Bathiany 2016). This is clearly seen in the middle and upper stratosphere. Climate models project a further cooling. However, climate models also project an increase in the strength of the large-scale meridional Brewer–Dobson circulation that would lead to a warming of the polar lower stratosphere (WMO 2018, and references therein). This has led to the notion, that the cold, undisturbed, Arctic stratospheric winters may be getting colder while overall there may be no clear trend or even a warming (Rex et al. 2004; Sinnhuber et al. 2011; Bohlinger et al. 2014). In spite of these factors, the Arctic winter lower stratosphere experiences new record low temperatures roughly every 5 years since the early 1990s. However, whether there is robust statistical evidence remains an open topic (e.g., Rieder and Polvani 2013).
There is now clear evidence that the Antarctic ozone hole has significantly affected surface climate change on the Southern Hemisphere (e.g., Thompson et al. 2011; WMO 2018). Ozone depletion in the Antarctic ozone hole has resulted in a cooling of the stratosphere between late winter and early summer, resulting in a poleward shift of the westerly winds that propagate down to the surface. This in turn leads to changes in rainfall and weather patterns. For the Arctic, however, the situation is much less clear. Although in individual Arctic winters the amount of ozone chemically depleted approaches that in the Antarctic ozone hole (e.g., Manney et al. 2011), the loss is generally smaller than in the Antarctic, the polar vortex does not persist as long into spring when solar radiation becomes important, and the large interannual variability in the Arctic makes detection and attribution more difficult. More recently, a number of studies have provided evidence that Arctic ozone changes can impact surface climate (Smith and Polvani 2014; Calvo et al. 2015; Ivy et al. 2017).
Little focus has so far been placed on the lowermost Arctic stratosphere, say below 100-hPa pressure altitude. The majority of the column integrated chemical ozone depletion in the Arctic takes place between approximately 17- and about 25-km altitude (e.g., WMO 2018), but much less is known about the upper troposphere–lowermost stratosphere (UTLMS) region where trace gas changes have a relatively large radiative effect for surface climate forcing (e.g., Riese et al. 2012). There is still large uncertainty of the expected ozone recovery in the extratropical and high-latitude UTLMS and the associated radiative forcing (Bekki et al. 2013). The polar vortex is less well defined in the LMS below about 100-hPa pressure altitude, and chemical ozone changes are more difficult to quantify at these altitudes. Moreover, satellite observations of trace gases often have limited precision and accuracy in the UTLMS region, so that our understanding of the chemical and transport processes in this altitude region is incomplete (Hegglin et al. 2013). Nevertheless, there is evidence for chlorine activation and resulting chemical ozone depletion also in the lowermost Arctic (and Antarctic) stratosphere from satellite observations (e.g., Santee et al. 2011). Short-lived halogen compounds, predominantly from natural sources, can have a relatively large impact on ozone depletion in the polar lowermost stratosphere (Hossaini et al. 2015; Sinnhuber and Meul 2015; Fernandez et al. 2017), but these processes are still not well represented in current chemistry climate models.
THE POLSTRACC MISSION: OBJECTIVES AND MAIN ACTIVITIES.
To address remaining questions on chemistry and transport processes in the Arctic upper troposphere–lowermost stratosphere under the present load of ozone-depleting substances and state of climate variables, the POLSTRACC campaign was conducted in the Arctic winter 2015/16, several years after past intensive Arctic campaigns [e.g., SOLVE/THESEO-2000 in 1999/2000 (Newman et al. 2002), EUPLEX/SOLVE-II in 2002/03, RECONCILE in 2009/10 (von Hobe et al. 2013)]. By deploying the German High Altitude and Long Range Research Aircraft (HALO), POLSTRACC was designed to study the Arctic UTLMS along with its coupling with lower latitudes throughout an entire winter–spring cycle, with the following specific scientific objectives (see also Fig. 1):
investigate the structure, composition, and dynamics of the bottom of the polar vortex and its coupling with the upper troposphere and extravortex air masses, to quantify transport and mixing, and to identify the role of special events such as tropopause folds and polar lows;
study polar stratospheric clouds and high-altitude cirrus in the UTLMS in Arctic winter and spring;
examine the vertical redistribution of HNO3 by PSCs and underlying processes;
investigate chlorine activation/deactivation in the Arctic LMS and assess the role of bromine compounds; and
quantify ozone loss in the LMS.
Selected scientific highlights achieved from the mission data addressing these objectives are discussed in the section “Highlighted results.”
Overall mission strategy.
POLSTRACC was designed to cover the whole Arctic winter from the onset of the polar vortex in early winter until its breakdown in spring, utilizing the capabilities of HALO, requiring flight operations between December and March over a wide geographical area. HALO (www.halo.dlr.de/; www.halo-research.de) is a Gulfstream 550 twin-engine jet aircraft, with about 10,000-km range, up to 15-km ceiling altitude, and a payload capacity of about 3 tons. The POLSTRACC (www.polstracc.kit.edu/polstracc/img_auth.php/POLSTRACC_Whitebook_v2.4.2.pdf) payload consisted of a dedicated combination of in situ and remote sensing instrumentation providing highly resolved and accurate observations of the polar UTLMS region and making use of the synergy from parallel in situ and remote sensing measurements. A large suite of meteorological diagnostics and forecast tools along with Lagrangian chemistry-transport models driven by ECMWF forecasts was used for detailed flight planning.
The POLSTRACC campaign was arranged in three phases: 1) early winter: 17–21 December 2015, 2) midwinter: 12 January–2 February 2016, and 3) late winter–spring: 26 February–18 March 2016. During phases 2 and 3, HALO was operated mainly from Kiruna, Sweden (68°N, 20°E), while the early winter survey flights during phase 1 and the late winter surveys during phase 3 were conducted from HALO’s home base at Oberpfaffenhofen, Germany (48°N, 11°E).
In total, 18 scientific flights adding up to 156 flight hours covered the time span from 17 December 2015 until 18 March 2016. The maximum ceiling level achieved was 14.6 km (FL480), the minimum air temperature was –76°C, and the maximum ranges with and without fuel stops were 9,332 and 8,060 km, respectively.
Synergies from combining the projects POLSTRACC, GW-LCYCLE, and SALSA to one joint campaign.
The POLSTRACC mission was operated in close connection with two other projects dedicated to complementary science issues by sharing the same instrumentation and joining forces in several aspects: GW-LCYCLE (Gravity Wave Life Cycle Experiment) and SALSA (Seasonality of Air mass transport and origin in the Lowermost Stratosphere using HALO Aircraft). The usage of a shared payload, the same campaign base, and the same infrastructure considerably reduced the efforts of payload/aircraft certification, campaign logistics, and flight operations. The combination of the projects into one joint and focused campaign has extended the available datasets of the individual projects and therefore broadened their scientific scopes, enabling the study of the links between the processes that have been in the focus of the individual projects.
GW-LCYCLE (Wagner et al. 2017) addressed dynamical coupling processes by gravity waves from the troposphere into the mesosphere, with the goal to improve gravity wave parameterizations in climate models. GW-LCYCLE employed—in addition to HALO measurements—observations on board the Falcon 20-E research aircraft operated by Deutsches Zentrum für Luft- und Raumfahrt (DLR), ground-based observations, correlative satellite observations, and numerical modeling. Science issues dedicated to GW-LCYCLE were embedded in the overall flight planning operated from Kiruna leading to six flights focusing on gravity wave science aspects (Dörnbrack et al. 2018), with a special focus on the tomographic capabilities of the Gimballed Limb Observer for Radiance Imaging of the Atmosphere (GLORIA) aboard HALO (Friedl-Vallon et al. 2014; Riese et al. 2014; Krisch et al. 2017). During the parallel deployment of HALO and Falcon at Kiruna during phase 1, coordinated flights of Falcon and HALO were realized.
The overall scientific rationale of SALSA (http://gepris.dfg.de/gepris/projekt/109259784) was to improve the understanding of UTLS transport and chemistry in the transition region between the subtropics and the Northern midlatitudes through a better knowledge of the seasonality of trace gas distributions. In conjunction with the POLSTRACC mission, a special focus was laid on the winter–spring season. For this purpose, several HALO flights were conducted to study intrusions of subtropical and midlatitude air to the Arctic. Along with two flights, dedicated to SALSA science, heading toward the subtropics at the begin and end of the whole joint mission, HALO was covering a much more extended latitudinal area (25°–87°N) than in previous Arctic aircraft campaigns.
HALO flights are coded here as “PGS-nn,” reflecting the combination of POLSTRACC, GW-LCYCLE, and SALSA project goals, while the discussion of this paper only focuses on POLSTRACC science issues.
POLSTRACC utilizes a combination of 10 in situ and three remote sensing instruments on HALO (Table 1; Fig. 2). This dedicated UTLS payload weighs about 3 tons. A similar payload had already been used before during the TACTS/ESMVAL mission in 2012 (Müller et al. 2016) and was later on used during the WISE mission in 2017 (www.wise2017.de/; Kunkel et al. 2019). Further deployment of this payload with minor modifications is used for studying the Southern Hemisphere mid- to high-latitude UTLS during the SOUTHTRAC mission recently operated in September and November 2019.
This payload exploits the synergy of local but highly specific and precise in situ observations at the aircraft location with 2D and 3D distributions of atmospheric parameters achieved from remote sensing instruments capable of adding information on the atmosphere above and/or below the aircraft. Dropsondes were released for specific case studies as well as validation, and for providing temperature, tropospheric humidity, and wind velocity. The overall payload is capable of measuring a large suite of parameters with respect to physical conditions (pressure, temperature, wind velocity, turbulence), chemical composition (trace gas composition), clouds (PSCs, cirrus clouds, and aerosol), and dynamic processes (using combinations of these parameters, in particular tracers of different lifetimes).
The three remote sensing instruments provided vertical distributions of selected parameters (particularly PSCs, aerosol, temperature, O3, H2O, NO2, HNO3, ClONO2, OClO, BrO), below the aircraft by applying passive infrared limb imaging (GLORIA; Friedl-Vallon et al. 2014; Riese et al. 2014), and ultraviolet (UV)–visible (VIS)–near-infrared (NIR) differential optical absorption spectroscopy (mini-DOAS; Hüneke et al. 2017), and above the aircraft using the new multiwavelength lidar system WALES (Wirth et al. 2009).
The 10 in situ instruments apply various techniques, including mass spectrometry (MS), gas chromatography (GC), chemiluminescence, and different spectroscopic techniques, allowing highly precise and time-resolved observations of a large number of avionic and meteorological data, O3, gas phase and total H2O, tracers of different lifetimes (CO2, N2O, CH4, CO, SF6, CFCs, NMHCs, halogens, bromine-containing VSLS, and others), chemically active species such as HNO3, HCl, ClONO2, NO, NOy, and particulate NOy and H2O.
A detailed description of the payload can be found in the online supplemental material (https://doi.org/10.1175/BAMS-D-18-0181.2).
Related ground-based and satellite observations.
Cooperation with ground-based observation stations, balloon sounding activities (ozone soundings and radiosondes), and satellite-borne projects [i.e., Cloud–Aerosol Lidar with Orthogonal Polarization (CALIOP; Pitts et al. 2018), Microwave Limb Sounder (MLS; Santee et al. 2011), Atmospheric Chemistry Experiment–Fourier Transform Spectrometer (ACE-FTS; Bernath et al. 2005)] were initiated and pursued to extend the scientific scope of POLSTRACC and for flight planning. Tight cooperation was realized with the CALIOP and MLS teams, both during the campaign and for data analysis. CALIOP science team members were present in Kiruna during the midwinter phase of the campaign and CALIOP data quick looks were regularly used for flight planning during the PSC season. One flight was realized for matching spaceborne CALIOP and airborne WALES PSC measurements. Both CALIOP and MLS data were later on used in a number of joint publications by POLSTRACC consortium members.
Data from ground-based stations in the European Arctic, particularly of the ground-based lidars in Kiruna (68°N, 20°E) and Andenes (69°N, 16°E) on the windward and leeward sides of the mountains, were embedded in the mission planning of the intensive observation periods and in scientific data exploitation. During midwinter, frequent radiosonde and balloon-borne lightweight hygrometers were launched from Kiruna Airport and Andenes. Ozone-lidar operations and ozone radiosondes launches from Ny-Ålesund/Spitsbergen (79°N, 12°E) were coordinated with HALO flight routes covering the Spitsbergen area.
Modeling activities were used for flight planning of the POLSTRACC campaign and for postflight data analysis. In particular, the Chemical Lagrangian Model of the Stratosphere (CLaMS) and the chemistry climate model ECHAM–MESSy Atmospheric Chemistry (EMAC) were closely involved in POLSTRACC.
The chemistry transport model CLaMS is based on a Lagrangian formulation of tracer transport (McKenna et al. 2002; Konopka et al. 2004) and driven by ECMWF forecasts and reanalyses. CLaMS considers an ensemble of air parcels on a time-dependent irregular grid that is transported and mixed employing 3D trajectories. The model provides global and high resolution (up to 20-km horizontal resolution and a-few-hundred-meters vertical resolution) and covers both the troposphere and the stratosphere. CLaMS also includes a denitrification scheme including an implementation of nucleation, growth, and sedimentation of NAT particles that are transported on Lagrangian trajectories (Grooß et al. 2014, and references therein). The formation of cirrus clouds was simulated by the model system CLaMS-ice, in which a two-moment bulk cirrus scheme (Spichtinger and Gierens 2009) was coupled to CLaMS trajectories; in forecast mode this model system was used to predict areas of cirrus occurrence. CLaMS was used for flight planning and forecasting; for the analysis of photochemistry, transport, and mixing processes; for studies of PSCs and denitrification; and for analysis of the radiative budget. For postflight analysis of the results, a CLaMS simulation was performed using ERA-Interim (Dee et al. 2011), and satellite observations by MLS for initialization and boundary conditions (Grooß et al. 2018).
Simulations with the chemistry climate model EMAC complemented the CLaMS activities by providing a large-scale and long-term context and for investigating the climate impact of the investigated processes (Khosrawi et al. 2017). The chemistry climate model EMAC is based on the climate model ECHAM5 (fifth-generation European Centre Hamburg general circulation model) and the Modular Earth Submodel System (MESSy; Jöckel et al. 2010). EMAC is a global model, covering the atmosphere from the surface up to about 80-km altitude. EMAC was used in a version nudged with meteorological analyses or reanalyses (ECMWF operational and ERA-Interim) at T106 spectral truncation to support point-by-point comparisons with observations.
Further modeling activities in the POLSTRACC framework include high-resolution simulations of the UTLS on regional scale by the seamless chemistry–climate model ICON-ART (Schröter et al. 2018).
Campaign and flight management.
The campaign was run from Oberpfaffenhofen at the beginning and end of the mission, and from Kiruna during midwinter–early spring. An interactive website (www.polstracc.kit.edu), based on the MediaWiki software, was set up as a living document platform, including up to date information of important aspects to keep all campaign participants on site and at remote locations informed.
A small core group consisting of the mission coordinators, meteorologists, modelers, and, on a case by case basis, other interested scientists regularly discussed and prepared possible flight plans on the basis of available data (MLS and CALIOP quick looks, ground-based observations, meteorological analyses) and CTM predictions of a series of forecast products. Detailed scientific flight planning made use of a dedicated Mission Support System (MSS) tool (Rautenhaus et al. 2012) that was extended and adapted to the POLSTRACC needs. Forecast products were based on ECMWF deterministic forecasts and chemical transport model forecasts with CLaMS (Grooß et al. 2014). More details and an example of the process of planning, implementation, execution, and reporting of science flights are given in the supplemental material.
The Arctic winter 2015/16 was extraordinary in the sense that in November–December 2015 the coldest and strongest Arctic polar vortex for almost 70 years formed (Matthias et al. 2016), setting the stage for the greatest potential yet seen for record Arctic ozone loss (Manney and Lawrence 2016). Synoptic PSCs formed over a long period of time in early winter, vertically extending down to altitudes lower than 15 km (Pitts et al. 2018; Voigt et al. 2018). Temperatures below the ice frost point were detected throughout January 2016 at 30 hPa (Dörnbrack et al. 2016). In February, three minor stratospheric warmings led to slightly warmer conditions in the polar vortex, with temperatures still below the NAT stability temperature TNAT (Hanson and Mauersberger 1988). The final stratospheric warming started around 5 March 2016 (Manney and Lawrence 2016), which resulted in a split of the vortex, and ended the cold conditions. The exceptional coldness of the Arctic winter 2015/16 and its evolution with respect to the long-term climatology is illustrated in Fig. 3.
The cold conditions during POLSTRACC mission phases 1 and 2 provided ideal conditions to investigate PSC formation, redistribution of HNO3, and chlorine activation. The major stratospheric warming during phase 3 allowed us to address in detail air mass exchange and mixing under these distorted conditions.
Overview of scientific flights.
In total, 18 scientific flights were conducted in the time span from 17 December 2015 to 18 March 2016. Tables 2–4 present for every single flight of the three mission phases summaries of primary flight objectives and addressed research topics along with essential operational flight parameters.
A detailed description of the individual flights including maps of flight paths on maps of potential temperature is provided as a supplement to this paper.
Mission phase 1: Early winter.
Two survey flights before the Christmas break were conducted from the home base at Oberpfaffenhofen to provide an overview of the state and composition of the UTLMS on long transects covering a wide range of latitudes (Table 2, Fig. 4). In PGS-04 the focus was put on the subtropical to middle latitudes dedicated to SALSA science objectives. PGS-05 was the first flight of POLSTRACC to the Arctic, probing the transition region between middle latitudes and the Arctic (48°–81°N), in a situation when the vortex had already cooled down to temperatures below PSC existence temperatures over large areas.
Mission phase 2: Midwinter.
Eight flights between 12 January and 2 February 2016 (Table 3, Fig. 4) covered the midwinter period, that was characterized by a stable polar vortex and record-low stratospheric temperatures. Apart from the transfer flight PGS-06 on 12 January, all flights were performed from Kiruna Airport (ENSQ). These flights were conducted in the coldest period of the winter with synoptic PSCs often occurring in the lowermost stratosphere down to the local tropopause (Fig. 5). A tropopause fold in connection with mountain wave activity sampled during PGS-06 was studied in detail by Woiwode et al. (2018) (see the “Highlighted results” section, Fig. 10). Enhanced HNO3 at the bottom of the vortex was already detected by GLORIA on this flight which indicated nitrification of the LMS (Johansson et al. 2018; Braun et al. 2019). Further clear signs of renitrified and chlorine-activated air were found in filamented vortex air over Scandinavia in PGS-07. In situ instruments detected several layers of enhanced nitric acid at flight level (FL) 450 (45,000 ft; 1 ft ≈ 0.305 m). At the uppermost FL 470 sharp enhancements of particulate NOy were measured by the NOy channel of the forward facing inlet. PGS-08 was planned to probe air underneath the cold pool of the vortex and to study the extent and character of synoptic PSCs above the Norwegian Sea between Scandinavia, Iceland, and Spitsbergen with a turning point as far as 87°N. Evidence of ongoing nitrification of the Arctic LMS was observed by GLORIA, reaching up to 7 ppbv of gas-phase HNO3, while particulate NOy was detected from in situ measurements at flight altitude. PGS-09 was particularly devoted to the study of PSCs matching air masses of two overpasses of CALIPSO satellite tracks (Voigt et al. 2018; cf. Fig. 11 and Fig. 12 in the “Highlighted results” section). An exceptional finding during the POLSTRACC flights was the in situ detection of large HNO3-containing particles at low stratospheric altitudes around 14 km by the Atmospheric Nitrogen Oxides Measuring System (AENEAS) instrument, as illustrated for PGS-10 in Fig. 13 in the “Highlighted results” section. PGS-12 went south to the northern tip of Ireland where a warm conveyor belt (WCB) pushed the tropopause height up to FL 450. A very strong jet stream was crossed twice by the flight pattern, from the Arctic stratosphere to subtropical air and on the way back to Kiruna (see supplemental material for more detail). On a section of the same flight observations at constant solar zenith angle (SZA) were arranged to investigate chlorine/bromine chemistry by DOAS UV–VIS–NIR observations (see Fig. 17 in the “Highlighted results” section).
Coordinated flights of Falcon and HALO aircrafts were conducted over Scandinavia during a mountain wave event (PGS-11). Particularly PGS-10 and PGS-13 were tailored to explore in detail gravity waves with HALO flights above Iceland and Spitsbergen, respectively, by making use of hexagon-like flight patterns in order to enable tomographic retrievals from GLORIA measurements (Ungermann et al. 2011; Krisch et al. 2017).
Mission phase 3: Late winter–spring.
Eight flights between 26 February and 18 March covered the late winter period (Table 4, Fig. 6). The vortex started to shrink with the sudden major warming around 5–6 March (Manney and Lawrence 2016) followed by a vortex split and a displacement of the vortex far off the pole.
Three flights (PGS-14, PGS-17, PGS 19) were primarily devoted to access strongly subsided air masses between Greenland and Canada in order to characterize low N2O, high DO3, ClOx, DNOy, and ClONO2 at potential temperatures up to 410 K at various stages of the vortex breakdown. Detailed observations of chlorine species during PGS-14 are presented in the “Highlighted results” section (Fig. 15). During the southward leg of PGS-14 over Baffin Island, strongly subsided air masses were measured with volume mixing ratios of over 1,200 ppb for ozone [Fast and Accurate In Situ Ozone Instrument (FAIRO)], 203 ppb for N2O (TRIHOP), and 7.05 ppt for SF6 [Gas Chromatograph for the Observation of Stratospheric Tracers (GhOST)]. During PGS-14 PSCs were still present across the North Atlantic sector.
The flight PGS-15 was devoted to study a polar low approaching the Norwegian coast associated with gravity waves, and flights PGS-16 and PGS-18 were focusing on gravity wave situations over Greenland.
The transfer flight (PGS-19) back to Oberpfaffenhofen took place on 13 March, with a detour over Greenland to probe again patterns of aged vortex air at high θ levels. Finally, two flights were performed from Oberpfaffenhofen. PGS-20 went toward the Canary Islands (25°N) to cover the transition region between middle and low latitudes in spring with respect to SALSA science objectives. PGS-21 studied once again and for the last time aged vortex air situated over northern Scandinavia when patches of enhanced ClONO2 were still measured by GLORIA (Fig. 16).
The following section presents briefly selected scientific highlights achieved from the POLSTRACC mission for the Arctic winter 2015/16 for the following topics: 1) synergy between in situ and remote sensing instruments and cross validation, 2) state of polar vortex: transport and mixing, 3) case study of a tropopause fold at the polar-front jet stream with vertically propagating mountain waves, 4) large-scale ice PSCs, 5) denitrification of the stratosphere and extensive nitrification of the LMS, and 6) halogen chemistry and ozone depletion in the Arctic lowermost stratosphere.
Synergy between in situ and remote sensing instruments and cross validation.
The combination of in situ and remote sensing instruments in the POLSTRACC payload enabled us to extend the findings sampled in situ at flight altitude throughout the lowermost stratosphere and sometimes even across the tropopause into the Arctic upper troposphere. Comparisons between in situ instruments and the GLORIA limb sounder are shown in Figs. 7a–f. These comparisons show an overall agreement and also reflect the different data precisions and temporal resolutions of the measurements. Differences between these datasets mostly result from atmospheric variability, as measurements of the in situ instruments are recorded directly at the aircraft, while remote sensing instruments sample air masses along the whole line of sight through the atmosphere although with a strong weight around the tangent altitude. Due to these different measurement geometries, atmospheric variabilities with horizontal gradients lead to differences in the observed trace gases between in situ and remote sensing measurements. Interestingly, particularly strong discrepancies are manifested in comparisons of HNO3, ClONO2, and O3 at similar parts of the flight which indicates strongly filamented structures along the GLORIA line of sight (Johansson et al. 2018).
In addition, GLORIA and WALES measurements of O3 have been compared to ozone sondes, launched from Ny-Ålesund and these comparisons are presented in Figs. 7g and 7h. GLORIA measurements during flight PGS-08 on 20 January 2016, spatially close to Ny-Ålesund show on average an agreement with the profile measured by the ozone sonde within 25 ppbv, even in the upper troposphere. This comparison extends the finding of Johansson et al. (2018), who found GLORIA O3 retrievals to agree with FAIRO in situ O3 measurements within the estimated errors of the instruments. Ozone measurements above the aircraft were possible with the WALES instrument during flight PGS-05 on 21 December 2015. The comparison with a coordinated ozone sonde, launched from Ny-Ålesund, shows agreement between the averaged WALES profiles, which have been recorded close to the ozone sonde.
State of polar vortex: Transport and mixing.
An overview of the trace gas measurements during POLSTRACC is presented in Fig. 8 and Fig. 9 as a function of equivalent latitude and potential temperature (e.g., Hoor et al. 2004; Engel et al. 2006). Equivalent latitude is calculated from potential vorticity and thus together with potential temperature is conserved for adiabatic and reversible transport. Changes in the mean tracer distributions in potential temperature–equivalent latitude space correspond to diabatic and irreversible transport and mixing. Due to the prevailing very cold temperatures in the LMS during December 2015 and January 2016, the potential temperatures reached during POLSTRACC phase 1 at high (equivalent) latitudes were limited to about 370 K. In contrast, during March 2016 we could probe high-latitude air masses up to 410-K potential temperatures. Mean age of air is calculated from SF6 measurements of the GhOST instrument reaching up to 5 years in late February and March in high-latitude air in the LMS. These old mean ages correspond to air masses with inorganic chlorine (Cly) of about 2 ppbv. Diabatic subsidence of air masses from the polar vortex into the lower and lowermost high-latitude stratosphere lead to an ageing of air, together with a decrease in N2O. Interestingly, Krause et al. (2018) find that the polar lower stratosphere showed at the same time an increase in mean age, as well as an increase in tropospheric carbon monoxide (CO). This is strong evidence for mixing of younger (or tropospheric) air masses into the polar stratosphere. Simulation with the CLaMS model reproduce this change in the age of air spectrum, with a general aging and a simultaneous increase in the fraction of young air (Krause et al. 2018). Nevertheless, N2O observations in comparison to previous Arctic winters indicate that the 2015/16 Arctic polar vortex was rather unperturbed until early March with little in-mixing of outside vortex air.
Tropopause fold at the polar-front jet stream with vertically propagating mountain waves.
On 12 January 2016, the polar-front jet stream was located above northern Italy. At the same time, mountain waves were excited above the French Alps, Corsica–Sardinia, and the Apennines and resulted in an unexpected stall event of HALO during PGS-06, requiring intervention by the pilots (Bramberger et al. 2018). Mountain waves carried large vertical energy fluxes of 8 W m–2 and propagated vertically from the troposphere to the stratosphere without significant dissipation. Below the same flight section, GLORIA observations resolved the mesoscale fine structure of a tropopause fold at the polar-front jet stream in interaction with a mountain wave (Woiwode et al. 2018). The GLORIA observations show dry intrusions into the troposphere at the cyclonic shear side of the jet stream, and moist air entraining into the stratosphere at the anticyclonic shear side (Fig. 10). Tracer–tracer correlations of water vapor and ozone as a function of potential temperature constructed from the GLORIA observations resolve an active mixing region, which is perturbed by vertically propagating mountain waves. The study confirms the conceptual model of tropopause folds (Shapiro et al. 1980), validates the high quality of short-term deterministic forecasts by the ECMWF Integrated Forecast System (IFS), and suggests that mountain waves modulate exchange processes associated with tropopause folds.
Large-scale ice PSCs in the extremely cold 2015/16 Arctic stratospheric winter.
The cold winter conditions during winter 2015/16 (Manney and Lawrence 2016; Dörnbrack et al. 2016) set the stage for excessive large-scale PSC formation. Thus, in January 2016, the CALIOP instrument on board the CALIPSO satellite measured the largest PSC coverage throughout its 11-yr climatology in the Arctic (Pitts et al. 2018). A three-dimensional visualization of the Arctic PSC observed on 22 January 2016, the day of PGS-09 is shown in Fig. 11. The version 2.0 CALIOP PSC detection algorithm (Pitts et al. 2018) is used to calculate total backscatter ratios at 532 nm, that is the ratio of total backscatter to molecular backscatter (Pitts et al. 2013). Single orbits are merged to the composite plot of backscatter ratios observed by CALIPSO on 22 January 2016. The vertical and horizontal extension of the PSC with a volume larger than 7 × 107 km3 exceeds PSC coverage in any other year of the climatology (Pitts et al. 2018).
PSCs above HALO flight altitudes were detected with the WALES lidar instrument on board HALO (Wirth et al. 2009; Grooß et al. 2014). On 22 January 2016, a widespread PSC was measured on flight PGS-09 from Kiruna to Greenland over a horizontal distance larger than 2,200 km (Fig. 12). The phase of the PSC was derived from backscatter and depolarization information in the cloud as described in Voigt et al. (2018). High backscatter and depolarization ratios, indicative for polar stratospheric ice clouds, were detected in a cloud layer with a vertical thickness of 2–4 km over distances of several hundred kilometers. The synoptic-scale ice PSC particles were surrounded by cloud layers mainly consisting of NAT, nondepolarizing liquid STS droplets as well as their mixtures (Voigt et al. 2018). Large-scale Arctic ice PSCs were only observed by CALIPSO in the 2009/10 Arctic winter but with lower ice area coverage than in 2015/16 (Pitts et al. 2018). The ice PSC has two branches with high or low particle depolarization ratio, indicative for different ice nucleation pathways. Ice nucleation on NAT (Voigt et al. 2018) may account for the higher particle depolarization ratios while ice nucleation in STS with meteoric inclusions (Engel et al. 2013) may contribute to the lower branch. The observation of widespread Arctic ice PSCs in the winter 2015/16 advances the knowledge on ice PSC nucleation. The new observational database can be used for model evaluation in polar latitudes. In addition to dehydration, the ice PSCs can contain traces of nitric acid, which may substantially contribute to renitrification as observed in the Arctic winter 2015/16.
Denitrification of the stratosphere and extensive nitrification of the LMS.
While denitrification of the polar stratosphere by HNO3-containing PSC particles is a well-known phenomenon (Fahey et al. 1990, 2001; Waibel et al. 1999), less attention has been paid to associated nitrification of lower layers due to sublimation of the particles. As a consequence of cold temperatures in the Arctic winter 2015/16, denitrification of layers above ∼100 hPa and associated nitrification of lower layers were found in satellite-borne observations (Khosrawi et al. 2017). An exceptional finding during the POLSTRACC flights was the detection of large HNO3-containing particles at around 14 km by the in situ instrument AENEAS. An example is shown for PGS-10 in Fig. 13. The observations show that these particles sediment down into the LMS under sufficiently cold conditions and make up a substantial proportion of total available NOy.
Frequently, extensive nitrification of the Arctic LMS was diagnosed using two-dimensional vertical cross sections of gas-phase HNO3 measured by GLORIA across wide ranges and down to potential temperatures well below 340 K or 12 km, respectively, as shown exemplary for PGS-08 in Fig. 14 (Braun et al. 2019). High maximum HNO3 mixing ratios of locally up to 11 ppbv were observed, and HNO3 enhancements due to nitrification between 5 and 7 ppbv were derived from the GLORIA measurements in January based on tracer–tracer correlations (Braun et al. 2019). Fine structures in the gas-phase HNO3 distribution suggest recently nitrified regions. Nitrification of the LMS was also derived during the same flight from in situ observations using the well-established correlation between total reactive nitrogen (NOy) and nitrous oxide (N2O), yielding excess NOy values up to 6 ppbv, in good agreement with the GLORIA observations. Comparisons of mesoscale structures in gas-phase HNO3 observed by GLORIA with CLaMS simulations also agree. However, significant differences remain in the fine structures, and nitrification of the LMS is underestimated by several parts per billion by volume by the model. Sensitivity experiments using CLaMS suggest that details of the particle microphysics are needed to explain the data until the end of January, while air subsidence, transport, and mixing increasingly affect nitrification patterns in the LMS rather toward the end of the winter (Braun et al. 2019). Until mid-March 2016, air masses depleted in NOy descended down and became accessible to the in situ instruments onboard HALO. At that time, the correlation with N2O reveals deviations from the expected NOy volume mixing ratio by up to ∼5 ppbv at a potential temperature of ∼400 K (Fig. ES11 in supplemental material).
Chlorine–bromine chemistry in the Arctic lowermost stratosphere.
As noted above, the cold temperatures in the Arctic lower stratosphere in winter 2015/16 led to widespread occurrence of polar stratospheric clouds over a large altitude range, reaching down to HALO flight altitudes. The POLSTRACC observations allowed us to study in detail the halogen partitioning and ozone loss in the lower and lowermost Arctic stratosphere. The amount of available inorganic chlorine in the lowermost stratosphere is controlled by subsidence within the polar vortex and mixing across the vortex edge. Available inorganic chlorine (Cly) was derived from GhOST CFC-12 observations using a correlation based on whole air sampler measurements from balloon-borne observations in the Arctic stratosphere (Wetzel et al. 2015), updated for current tropospheric chlorine source gas mixing ratios. The sum of in situ measured HCl and ClONO2 from AIMS agreed very well with the estimated Cly for many flights, but sometimes falls considerably short, indicating the presence of active chlorine compounds. An interesting case study is provided by PGS-14 on 26 February 2016. High values of Cly were observed, reaching 1.7 ppbv at potential temperatures of 380 K and above (Figs. 15a,b). For this situation, the sum of observed HCl and ClONO2, however, is substantially smaller than inferred Cly, suggesting a significant contribution from active chlorine, ClOx (defined here as the difference between Cly and the sum of HCl and ClONO2) of up to 1.2 ppbv (Fig. 15; see also Marsing et al. 2019). In fact, the Cly and ClOx concentrations derived from the in situ observations agree very well with calculations from the CLaMS model. GLORIA observations (Johansson et al. 2018, 2019) provided simultaneous observations on the vertical extent of the chlorine processing in the lowermost stratosphere (Figs. 15c,d). Cly levels of about 1 ppbv and more extended down to potential temperatures of about 350 K. ClONO2 was reduced at flight altitudes, consistent with a large contribution of active chlorine at potential temperatures above 380 K. Higher concentrations of ClONO2 with peak mixing ratios of about 0.5 ppbv were observed lower down between 350 and 380 K.
Measurements by GLORIA of two-dimensional distributions of ClONO2 over the course of the winter were used to evaluate the chlorine partitioning in the CTM CLaMS and the CCM EMAC. A comparison of flights PGS-06 (12 January 2016), PGS-14 (26 February 2016), and PGS-21 (18 March 2016) is shown in Fig. 16. These comparisons reveal that CLaMS succeeds to reproduce even small-scale structures in the trace gas distributions, while EMAC is able to reflect the overall structure of measured ClONO2 enhancements. Remaining differences can be explained by the lower spatial resolution of the EMAC model. Interestingly, both models underestimated maximum mixing ratios of ClONO2 for the late winter flight PGS-21, which indicates that the diabatic descent over the course of the winter is underestimated in the meteorological analyses, which define dynamics in both simulations. In addition, it was found that CLaMS tends to overestimate ClONO2 at potential temperature altitudes of 325–350 K, which is suspected to result from the lower boundary condition of the model (Johansson et al. 2019).
Information on bromine comes from observed source gases, including very short-lived bromine containing source gases (VSLS) and measurements of bromine monoxide (BrO) from the mini-DOAS instrument. Figure 17 shows observed BrO and estimated Bry from source gas observations in comparison with CLaMS model calculations. The modeled BrO/Bry ratio was used to infer Bry from BrO observations. The particular flight profile in the later part of PGS-12 on 31 January 2016 under quasi-constant solar zenith angles provides ideal stable conditions for the interpretation and analysis of the BrO and OClO observation and the estimates of reactive ClOx from inferred Cly and measured HCl and ClONO2. The measurements of OClO and BrO support a detailed study of the efficiency of ClOx/BrOx ozone loss cycle for which the branching ratio of the ClO + BrO reaction into BrCl + O2 is still somewhat uncertain (Canty et al. 2005). Further, the combined measurements of organic and inorganic bromine can be tightly monitored as a function of air mass age.
Ozone depletion in the Arctic lowermost stratosphere.
Substantial ozone depletion in the lower and lowermost stratosphere, consistent with the chlorine activation in this altitude region is diagnosed from in situ and remote sensing observations of ozone during POLSTRACC. Chemical ozone depletion can be estimated by reference to a passive model tracer or through correlation with a long-lived tracer gas such as N2O. Figure 18 shows observed ozone for 20 January and 13 March 2016, in comparison with CLaMS model simulations. The red solid line indicates the modeled passive ozone in the absence of chemical loss. On 20 January there is still little difference between the model simulations with and without photochemistry due to the limited amount of sunlight up to this time. The situation is very different in the later part of the campaign: on 13 March 2016 a large difference between the model simulations with and without chemistry emerges (red shaded area in Fig. 18, bottom), indicating chemical ozone depletion reaching more than 1 ppm or about 50% at 380 K and above by mid-March. The generally good agreement between observed and modeled ozone based on CLaMS simulations including chemistry provides confidence for the modeled passive ozone tracer as a reference to diagnose chemical ozone loss.
The agreement between the CLaMS results and observations of long-lived chemical species like ozone or N2O indicates that the transport in the lower and lowermost stratosphere as well as the photochemistry (and microphysics) are represented well in the model. In early winter, the discrepancies between observed and modeled HCl mixing ratios are slightly larger. This is potentially caused by the same unknown processes that cause a larger discrepancy at higher altitudes in early winter (Grooß et al. 2018). The overall excellent agreement between the CLaMS model and observations of ozone and reactive chlorine and bromine species indicates that we have an adequate understanding of the processes of chlorine activation and ozone depletion in the lower and lowermost stratosphere. A realistic treatment of subsidence and mixing of air masses in this altitude range is crucial for models to properly reproduce the observed chemical ozone depletion in the Arctic UTLS region, as this controls the amount of available inorganic chlorine.
CONCLUSIONS AND OUTLOOK.
The POLSTRACC Arctic aircraft campaign took place in winter 2015/16, several years after previous intensive Arctic campaigns under declining levels of ozone depleting substances (WMO 2018) and accelerated Arctic climate change (Overland et al. 2019). POLSTRACC focused particularly on the Arctic lowermost stratosphere and upper troposphere and its coupling with midlatitudes. With mission phases in December 2015, throughout January, and from late February until mid-March 2016, the campaign covered the important development of chemical and transport processes from early winter until the breakdown of the polar vortex by mid-March during the 2016 major stratospheric warming. The capabilities of the HALO research aircraft with its long-range flights and high ceiling altitude were critical for the success of the mission. From the base in Kiruna, Sweden, flights deep into the Arctic almost to the North Pole (maximum latitude reached 87.1°N during flight PGS-13) and probing high altitude air masses up to potential temperatures of 410 K were possible. Refueling stops not only allowed to extend the range, but also enabled observations at maximum flight levels in the region of interest.
The POLSTRACC payload combined precise in situ instrumentation with remote sensing instruments for measurements of temperature, trace gases, and cloud properties in the UTLS region. High-resolution ECMWF forecasts, together with chemistry transport modeling from the CLaMS model and the Mission Support System (MSS) flight planning tool (Rautenhaus et al. 2012) and satellite data from MLS and CALIPSO, were of great value in flight planning. The excellent agreement between the GLORIA infrared remote sensing data with in situ observations (Johansson et al. 2018) enabled us to extend the findings sampled in situ at flight altitude throughout the lowermost stratosphere and sometimes even across the tropopause into the Arctic upper troposphere. An impressive demonstration of these capabilities has been provided for a case study of a tropopause folding event and assessment of the quality of ECMWF meteorological forecasts (Woiwode et al. 2018).
The Arctic winter 2015/16 was characterized by record cold temperatures in the lower stratosphere in early winter, followed by a major warming beginning in early March. Probably one of the largest surprises during the POLSTRACC campaign was the occurrence of extensive PSCs, even down to HALO flight altitude, accompanied by observation of nitrogen redistribution, chlorine activation and ongoing ozone depletion at flight altitudes. The PSC observations provided new evidence for ice nucleation on nitric acid trihydrate (NAT) particles (Voigt et al. 2018). The coordinated flight with orbits of the CALIPSO satellite provided an unprecedented opportunity to validate the spaceborne CALIOP lidar with independent high spectral resolution observations by the airborne WALES lidar for an improved PSC classification. Nitric acid containing particles were observed at flight altitude—however, a quantification of the particle sizes and the amount of HNO3 contained is challenging because of uncertainties due to the enrichments in the inlets (Afchine et al. 2018). Extensive denitrification in the lower stratosphere and a nitrification at lower altitudes were observed throughout the winter (Khosrawi et al. 2017; Braun et al. 2019). Heterogenous chlorine activation with a shift in the partitioning of inorganic chlorine from HCl toward chlorine nitrate (ClONO2) was observed at different stages throughout the winter (Johansson et al. 2019), not observed in detail before at these altitudes in the Arctic. The flight under constant solar zenith angle conditions on 31 January 2016 provided a unique opportunity to study the chemistry of the BrO/OClO system and thus the ozone loss from the interhalogen reaction of ClO with BrO in the lower stratosphere from DOAS UV–visible observations.
Observations of different trace gases give a detailed picture of the impact of transport and mixing on the composition of the Arctic lowermost stratosphere throughout the winter. Subsidence of air masses from the polar vortex throughout the winter led to observed air masses with mean age of air of up to 5 years and inorganic chlorine levels of up to 2 ppm at potential temperatures between 380 and 410 K. Simultaneously with the aging of the air in the Arctic winter lower stratosphere throughout the winter, an increase in the contribution of tropospheric carbon monoxide (CO) was observed between January and March at levels up to 370 K, showing the importance of in-mixing of younger air masses into the Arctic lower stratosphere (Krause et al. 2018). The investigation of the importance of this in-mixing for the distribution of water in the lowermost stratosphere and the resulting radiative effects is currently under way, based on the POLSTRACC in situ and remote sensing observations. Substantial chemical ozone depletion was observed in the Arctic lower and lowermost stratosphere between January and late February and early March 2016, reaching local losses of up to 50%. Further studies should investigate the consequences of these large observed ozone losses in the lowermost Arctic stratosphere for radiative and dynamical coupling with the surface and troposphere and how well this is represented in current chemistry climate models. Comparison with satellite data will help to put the airborne observations in winter 2015/16 into a longer-term perspective. The winter of 2015/16 with record low temperatures in early and midwinter in the lower stratosphere was another winter in a row of record cold winters about every 5 years observed since the early 1990s (e.g., Rex et al. 2004; Sinnhuber et al. 2011; Bohlinger et al. 2014). Whether or not this is a manifestation of climate change in the Arctic stratosphere remains unclear at present as well as possible consequences of a changing Arctic. The POLSTRACC data will help to test and ultimately improve relevant chemical and transport processes in chemistry climate models.
POLSTRACC data are available through the HALO database at https://halo-db.pa.op.dlr.de/.
We thank the pilots, engineers, and scientists from DLR Flight Department–Flight Experiments for their excellent support during the POLSTRACC campaign. Andreas Minikin provided vigilant overall project management and Frank Probst responsively supported the operational activities. Andreas Giez, Volker Dreiling, and Martin Zöger supervised the BAHAMAS observations. We acknowledge the cooperative spirit of the airport manager Kurt Mäki and the team of the tower of the Kiruna Airport. We thank Andreas Dörnbrack, Sonja Gisinger, Martina Bramberger, Maxi Böttcher, and Johannes Wagner for the excellent meteorological support and Reinhold Spang for support with CLaMS forecasts. Reimar Bauer contributed to extending and adapting the MSS tools to the POLSTRACC needs. Ole Kirner and Farah Khosrawi provided near-real-time EMAC simulations; Jennifer Schröter provided the ICON-ART simulations. We acknowledge the use of meteorological forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) and the MLS satellite data. We thank Peter von der Gathen and his team for coordinated ozone sonde launches. We thank the instrument and data analysis teams of AENEAS (Michael Lichtenstern, Paul Stock, Greta Stratmann), FISH (Armin Afchine), GhOST (Ronja Heinemann, Timo Keber), GLORIA (Marleen Braun, Andreas Ebersoldt, Thomas Gulde, Michael Höpfner, Anne Kleinert, Thomas Latzko, Guido Maucher, Hans Nordmeyer, and Christof Piesch at KIT and Tom Neubert, Georg Schardt, Heinz Rongen, Markus Dick, Axel Schönfeld, and Isabell Krisch at FZJ), HAGAR (Valentin Lauther, Emil Gerhardt, Johannes Wintel), mini-DOAS (Tilman Hüneke, Oliver-Alex Aderhold, Bodo Werner), and WALES (Axel Amediek, Silke Gross, Mathieu Quatrevalet, Patrick Vranken) for persistent commitment. We gratefully acknowledge the funding by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) within the Priority Program “Atmospheric and Earth System Research with the High Altitude and Long Range Research Aircraft (HALO)” SPP 1294 through the following grants: Andreas Engel and Peter Hoor (EN 367/13-1, HO 4225/7-1); Andreas Marsing, Christiane Voigt, and Tina Jurkat-Witschas (VO1504/4-1 and JU3095-1/1); Klaus Pfeilsticker (PF-384/7-1, PF-384/7-2, PF384/9-1, PF384/9-2, and PF-384/16-1); C. Michael Volk (VO 1530/4-1); and Wolfgang Woiwode (WO 2160/1-1). The authors gratefully acknowledge the computing time for the CLaMS simulations granted on the JURECA supercomputer at Jülich Supercomputing Centre (JSC) under the VSR project ID JICG11 and for the EMAC and ICON-ART model simulations at the Steinbuch Center for Computing (SCC). The authors acknowledge support by the state of Baden-Württemberg through bwHPC.
CURRENT AFFILIATION: Excelitas Technologies GmbH and Co. KG, Wiesbaden, Germany
A supplement to this article is available online (10.1175/BAMS-D-18-0181.2)