The North Atlantic Waveguide and Downstream Impact Experiment (NAWDEX) explored the impact of diabatic processes on disturbances of the jet stream and their influence on downstream high-impact weather through the deployment of four research aircraft, each with a sophisticated set of remote sensing and in situ instruments, and coordinated with a suite of ground-based measurements. A total of 49 research flights were performed, including, for the first time, coordinated flights of the four aircraft: the German High Altitude and Long Range Research Aircraft (HALO), the Deutsches Zentrum für Luft- und Raumfahrt (DLR) Dassault Falcon 20, the French Service des Avions Français Instrumentés pour la Recherche en Environnement (SAFIRE) Falcon 20, and the British Facility for Airborne Atmospheric Measurements (FAAM) BAe 146. The observation period from 17 September to 22 October 2016 with frequently occurring extratropical and tropical cyclones was ideal for investigating midlatitude weather over the North Atlantic. NAWDEX featured three sequences of upstream triggers of waveguide disturbances, as well as their dynamic interaction with the jet stream, subsequent development, and eventual downstream weather impact on Europe. Examples are presented to highlight the wealth of phenomena that were sampled, the comprehensive coverage, and the multifaceted nature of the measurements. This unique dataset forms the basis for future case studies and detailed evaluations of weather and climate predictions to improve our understanding of diabatic influences on Rossby waves and the downstream impacts of weather systems affecting Europe.
Multiaircraft and ground-based observations were made over the North Atlantic in the fall of 2016 to investigate the importance of diabatic processes for midlatitude weather.
Progress in understanding the processes controlling midlatitude weather is one of the factors that have contributed to a continuous improvement in the skill of medium-range weather forecasts in recent decades (Thorpe 2004; Richardson et al. 2012; Bauer et al. 2015). Additionally, numerical weather prediction (NWP) has undergone a revolution in recent years, with the development and widespread use of ensemble prediction systems (EPSs) to represent forecast uncertainty (Bauer et al. 2015). However, the short-term prediction of high-impact weather (HIW) events (e.g., strong winds and heavy precipitation), and the medium-range prediction of extratropical cyclones, including their tracks and intensity, are still major challenges (e.g., Frame et al. 2015). Recent research into midlatitude weather has focused on quantifying model errors and predictability, and in particular on investigating the role of diabatic processes such as those related to clouds and radiation, whose interaction with the dynamics of the flow must be understood and represented more accurately in models in order to further improve forecast quality.
Detailed observations are needed to characterize the weather systems and embedded physical processes across a range of spatial and temporal scales that encompass cloud microphysical variability and Rossby waves. In September and October 2016, the North Atlantic Waveguide and Downstream Impact Experiment (NAWDEX) made new multiscale observations in the North Atlantic basin from eastern Canada to western Europe. Weather features expected to be associated with forecast errors were extensively probed, providing a high-quality set of observations that are not assimilated routinely and thus can be used for validation of the NWP systems.
The fall season was chosen for the experiment because diabatic processes are particularly active as a result of relatively high sea surface temperatures and the intensification of the jet stream as the high latitudes cool. Many of the weather phenomena central to the growth of disturbances on the jet stream and midlatitude predictability are active in fall, such as extratropical cyclones with intense fronts and warm conveyor belts (WCBs), carrying air from the oceanic boundary layer into ridges at the tropopause level. There is also the possibility of North Atlantic tropical cyclones (TCs) recurving poleward into midlatitudes and undergoing extratropical transition (ET), which is a process known to be associated with low predictability (Harr et al. 2000). Coherent mesoscale depressions of the tropopause, known as tropopause polar vortices (TPVs; Cavallo and Hakim 2010; Kew et al. 2010), can disturb the jet stream if they move equatorward from the Arctic.
NAWDEX contributes to the World Weather Research Programme (WWRP) and its High Impact Weather project (Jones and Golding 2015) and aims to provide the observational foundation to further investigate cloud diabatic processes and radiative transfer in North Atlantic weather systems, which will form the basis for future improvements in the prediction of HIW over Europe.
THE ROLE OF DIABATIC PROCESSES.
Weather in Europe strongly depends on the life cycles of Rossby waves that propagate along the slowly varying part of the North Atlantic jet stream (Martius et al. 2010). The strong meridional potential vorticity (PV) gradient associated with the jet stream serves as a waveguide for propagating Rossby waves. Frequently, small disturbances in the jet entrance region over eastern North America grow in baroclinic weather systems and evolve into large-amplitude features in the European sector (Schwierz et al. 2004). Figure 1 portrays an idealized North Atlantic flow situation that could result in HIW in the form of high winds and heavy precipitation over northern Europe. In addition to Rossby waves amplifying through baroclinic instability, diabatic processes are able to modify upper-tropospheric PV at the level of the midlatitude jet stream, which impacts the wavelength and amplitude of the downstream Rossby wave development (e.g., Massacand et al. 2001; Knippertz and Martin 2005; Grams et al. 2011; Teubler and Riemer 2016).
The majority of the precipitation and cloud diabatic processes in extratropical cyclones occur within a coherent airstream known as the WCB. It carries warm, moist air from the low-level warm sector of a cyclone to the ridge at tropopause level within 1–2 days (Browning et al. 1973; Carlson 1980; Wernli and Davies 1997). The boundary layer humidity in the inflow of WCBs (region 1 in Fig. 1) can impact the outflow height of WCBs (Schäfler and Harnisch 2015). For some WCBs, the inflow region coincides with a filament of strong horizontal water vapor transport, a so-called atmospheric river, which can contribute to intense rain in the midlatitudes (Lavers and Villarini 2013). During the ascent of WCBs (region 2 in Fig. 1), embedded convection, and turbulent fluxes influence the level of the outflow layer, the direction taken by outflow air masses, and the shape of the upper-level ridge (Martínez-Alvarado et al. 2014; Joos and Forbes 2016). The latent heating in WCBs is strong both in the early phase of the ascent when condensation dominates and later when mixed-phase clouds are formed and vapor deposition on ice crystals and snow becomes important (Joos and Wernli 2012).
The effect of the heating on the PV structure is to produce a positive PV anomaly in the lower troposphere (Wernli and Davies 1997), which influences the structure and evolution of midlatitude surface cyclones (e.g., Kuo et al. 1991; Davis et al. 1993; Binder et al. 2016). Above the level of maximum latent heating, PV is reduced by cloud diabatic processes, leading to negative PV anomalies in the upper-tropospheric WCB outflow region (Wernli 1997; Pomroy and Thorpe 2000; Madonna et al. 2014; Methven 2015). The divergent outflow winds (region 3 in Fig. 1) tend to amplify the upper-level downstream ridge and to intensify the jet stream by strengthening the PV gradient (Archambault et al. 2013). If the outflow layer is higher, the negative PV anomaly is stronger and more of the air mass enters the anticyclonic branch of the WCB flowing into the downstream ridge (Grams and Archambault 2016). In addition, a sharp peak in longwave radiative cooling near the tropopause, associated with a step change in water vapor, creates a reinforcement of the positive PV anomaly in upper-level troughs (Chagnon et al. 2013) and plays a key role in maintaining and strengthening TPVs (Cavallo and Hakim 2012).
Diabatic processes also play a key role in weather systems that act as triggers to disturb the midlatitude waveguide. Recurving TCs undergoing ET (Jones et al. 2003) can enhance the anticyclonic and divergent flow at upper levels, excite and amplify Rossby waves, and cause downstream forecast errors, as well as HIW events (e.g., Agusti-Panareda et al. 2004; Harr et al. 2000; Riemer and Jones 2010). Radiatively maintained TPVs, which are positive PV anomalies above the tropopause, can disturb the Rossby waveguide from the polar latitudes.
Rossby wave breaking leads to PV filamentation, forming smaller-scale PV anomalies such as PV streamers and cutoff vortices. They form frequently over the eastern North Atlantic and Europe (e.g., Wernli and Sprenger 2007), and several studies have reported their relevance for triggering HIW, in particular heavy precipitation (e.g., Martius et al. 2006; Chaboureau and Claud 2006; Grams and Blumer 2015). Synoptic wave breaking events are also important for the large-scale flow itself as they reinforce weather regimes such as blocking ridges (Michel and Rivière 2011; Spensberger and Spengler 2014). Blocks are also strongly influenced by diabatic processes in air masses ascending from the lower troposphere (Pfahl et al. 2015).
Disturbances of the waveguide and associated errors can amplify and propagate downstream, and may cause significant forecast errors over Europe (Madonna et al. 2015; Martínez-Alvarado et al. 2016) (region 4 in Fig. 1). In NWP models, diabatic processes such as those associated with convection, cloud microphysics, and radiation are represented by parameterizations of varying degrees of fidelity and may contain both systematic and random errors that influence forecast skill. A distinct Rossby wave pattern associated with the poleward transport of warm and moist air over the eastern United States and strong diabatic activity has been identified as a common precursor 6 days before the worst forecast busts over Europe (Rodwell et al. 2013). Upscale error growth experiments in numerical models show that the growth of small-scale perturbations is initially confined to regions where condensation is occurring, with the regions of large error amplitude gradually expanding to affect the synoptic-scale weather pattern (Zhang et al. 2007; Selz and Craig 2015). Doyle et al. (2014) found forecasts of an extratropical cyclone with severe impacts in western Europe to be very sensitive to the initial low-level moisture, which influenced the moisture supply in a WCB. At upper levels, global NWP models fail to maintain a sufficiently sharp tropopause, showing a decrease in sharpness with forecast lead time (Gray et al. 2014). This influence on the waveguide can have major implications for the representation of the downstream propagation and amplification of Rossby waves in NWP (Harvey et al. 2016) and the associated prediction of HIW.
Previous studies using measurements to study the influence of diabatic processes on the Rossby waveguide have been primarily based on routinely collected observations by operational meteorological services. These observations rely largely on satellite data, which models predominantly assimilate in cloud-free areas, and on sparse in situ measurements, all of which are combined in the data assimilation system using model forecasts as a background estimate. This approach to studying diabatic processes has significant limitations since these processes tend to be strongest in cloudy and precipitating regions, which are particularly challenging for both observation and modeling systems. The processes associated with diabatic heating are characterized by a high degree of small-scale variability, particularly in the vertical (e.g., sharp vertical gradients of cloud microphysical processes and their interactions with radiative forcing), which are typically poorly resolved by satellite and conventional in situ observations. Furthermore, rapid error growth and systematic model errors lead to large errors in the background forecast in precisely these regions, which are poorly characterized by error covariance matrices based on climatology and/or sampling using an ensemble of limited size. A field campaign has the potential to address some of these difficulties by deploying specialized observing systems with high resolution and the ability to measure both in and around clouds.
NAWDEX was proposed with the overarching hypothesis that diabatic processes have a major influence on the jet-stream structure, the downstream development of Rossby waves, and eventually HIW. Specific science goals were formulated (Table 1), which require observations of moisture advection in the boundary layer and of the vertical distribution of stability, water vapor, liquid droplets, and ice crystals. These observations will be used to investigate spatial variability within clouds and the implications for diabatic processes. Detailed wind measurements in the layer of the divergent outflow of the WCB are needed to investigate the interaction of diabatically modified air masses with the upper-level jet. This includes observations of horizontal and vertical gradients of wind, temperature, and humidity, as well as hydrometeors in clouds. Accordingly, high-resolution cross sections of wind, temperature, and humidity from the lower stratosphere down to the surface, inside and outside of clouds, are the central observational requirements, which are not available from conventional observations.
EXPERIMENTAL DESIGN AND OBSERVATIONS.
The need for a new field experiment emerged from a series of campaigns coordinated by the World Meteorological Organization’s The Observing System Research and Predictability Experiment (THORPEX; Parsons et al. 2017). This series includes the Atlantic THORPEX Regional Campaign (ATReC; Rabier et al. 2008), Winter Storm Reconnaissance (WSR, e.g., Szunyogh et al. 2000), the THORPEX Pacific Asian Regional Campaign (T-PARC, Weissmann et al. 2011), and the Convective and Orographically Induced Precipitation Study/European THORPEX Regional Campaign (COPS/ETReC 2007; Wulfmeyer et al. 2011), which all focused on the impact of additional observations on improving forecast accuracy. This idea was pioneered by the Fronts and Atlantic Storm Track Experiment (FASTEX) in 1997 (Joly et al. 1999), where the concept of targeting observations using sensitive area calculations was introduced. The synthesis of these campaigns and data assimilation experiments denying observations in data-rich areas showed that the impact of targeted observations on global forecast systems is weaker than originally anticipated, although they improve forecasts on average [see review by Majumdar (2016)]. At the same time, as discussed above, evidence was growing that forecast errors often originate in regions where diabatic processes are strong and observation and modeling systems are least reliable. This provided the motivation for a new campaign, NAWDEX, that rather than targeting regions of forecast sensitivity, instead focused on observing the processes that are thought to be most uncertain in NWP models.
Diabatic processes are difficult to measure directly but can be constrained via their observable effects on the structure and evolution of weather systems. In the decade before THORPEX, detailed diagnostic case studies using aircraft measurements [e.g., Experiment on Rapidly Intensifying Cyclones over the Atlantic (ERICA; Hadlock and Kreitzberg 1988) and FASTEX] had already shown that diabatic processes, in particular diabatic heating and cooling, can impact the large-scale dynamics via PV modification (Neiman et al. 1993; Pomroy and Thorpe 2000). However, the processes are difficult to accurately quantify since they depend on finescale structures (e.g., sharp gradients) in the water vapor and cloud fields and are influenced by transport and mixing over a wide range of spatial and temporal scales throughout the lifetime of the cyclone (recall Fig. 1). Two single-aircraft field campaigns organized within THORPEX explored how aircraft observations could be used to accurately constrain the impact of diabatic heating in midlatitude cyclones. The Diabatic Influence on Mesoscale Structures in Extratropical Storms (DIAMET) experiment (Vaughan et al. 2015) made airborne in situ measurements of liquid droplets and ice crystals and used them to infer the microphysical processes acting, their positions relative to mesoscale structures (such as fronts and PV anomalies), and their role in the weather system dynamics. Although the observations were limited to the one-dimensional (1D) aircraft flight path, they provided a basis for the modeling studies of Dearden et al. (2014), who obtained estimates of heating rates from various microphysical processes represented by a Lagrangian model initialized with in situ observations of size distributions for cloud droplets and ice particles. However, using a model to extrapolate the measurement information in space and time represents an additional source of uncertainty in the quantification of the processes. The THORPEX-NAWDEX-Falcon project (Schäfler et al. 2014) attempted to constrain this uncertainty by carrying out in situ observations of clouds, humidity, and wind in ascending WCBs, and trying to resample the same air masses at a later time to obtain a Lagrangian estimate of the integrated diabatic effects. NAWDEX was conceived to expand upon the design of these previous experiments by combining high-resolution remote sensing and in situ instrumentation to provide accurate measurements of atmospheric structures including strong gradients, using multiple aircraft to sample air masses at different stages of the WCB ascent and advection along the tropopause.
To allow these observations to be related to the development of weather forecast errors, NAWDEX employed four research aircraft and ground-based stations spanning the northern part of the North Atlantic with the aim of observing the processes influencing the development of disturbances to the North Atlantic waveguide across the Atlantic. This includes upstream triggering of disturbances on the waveguide by phenomena with strong latent heat release, the continuous effects of clouds and radiation near the tropopause, the dynamical interactions between large-scale disturbances, and the potential impact on weather over Europe from the Mediterranean to Scandinavia.
Airborne platforms and payload.
NAWDEX employed four research aircraft: the German High Altitude and Long Range Research Aircraft (HALO) and the Deutsches Zentrum für Luft- und Raumfahrt (DLR) Dassault Falcon 20, the French Service des Avions Français Instrumentés pour la Recherche en Environnement (SAFIRE) Falcon 20, and the British Facility for Airborne Atmospheric Measurements (FAAM) BAe 146. FAAM operated from the United Kingdom and HALO and the two Falcon aircraft from Keflavik, Iceland, in an area covering the North Atlantic, north of 45°N, and northern and central Europe. The payloads were chosen to observe the required profiles of wind, temperature, moisture, and cloud properties, and in the case of FAAM, in situ cloud microphysics.
The strategy was to deploy HALO with its extended range to observe moisture transport and diabatic processes in weather systems upstream of Iceland that impact the midlatitude waveguide. HALO is a modified Gulfstream G-550 ultra-long-range business jet with a maximum flight range of about 10,000 km and a maximum endurance of 10 h (Krautstrunk and Giez 2012; Wendisch et al. 2016), which allows for access to remote regions over the central North Atlantic that are not accessible by other European research aircraft. The high ceiling of almost 15 km in combination with a sophisticated remote sensing payload (see the “Active remote sensing observations for future satellite missions Aeolus and EarthCARE” sidebar and Table 2) allow HALO to fly above the main commercial aircraft routes and to probe features of interest from above. The two Falcon aircraft, with a maximum range of 3,000 km, a maximum endurance of about 4 h, and a ceiling up to 12 km, aimed to observe the approaching cyclones and evolving jet streams close to Iceland. The DLR Falcon was equipped with two wind lidar systems and the SAFIRE Falcon with a remote sensing payload for clouds and winds (see the “Active remote sensing observations for future satellite missions Aeolus and EarthCARE” sidebar and Table 2). The FAAM BAe 146, with a maximum endurance of 5 h and a ceiling of 10 km, was equipped with a range of in situ instrumentation for meteorological, cloud, and chemical measurements together with a downward-pointing aerosol lidar and passive spectral radiometers. Its flights from East Midlands, United Kingdom, were aimed at observing the microphysics and turbulence in WCBs and the structure of the jet stream.
HALO and the SAFIRE and DLR Falcon aircraft were equipped with remote sensing instruments that are specifically relevant for the future EarthCARE satellite missions (Illingworth et al. 2015) and Aeolus (ESA 2008) of the European Space Agency. NAWDEX observations, through coordinated flights of multiple aircraft and of aircraft with satellite overpasses, provide data from comparable airborne instruments for the preparation and future validation of these satellite instruments.
HALO was equipped with the HSRL (532 nm) and the water vapor DIAL WALES, HAMP with a 35.6-GHz cloud radar and microwave radiometers, the cloud spectrometer (specMACS), and the visible to near-infrared SMART instrument (Table 2). The French Falcon was equipped with the radar–lidar (RALI; Protat et al. 2004) payload consisting of the 94-GHz RASTA cloud radar and the UV High Spectral Resolution LNG lidar (Table 2). These aircraft provide the most complete instrumentation package available at the European level to mimic upcoming EarthCARE measurements and thus provide valuable data for preparing the EarthCARE mission and for future validation. Coordinated flights with both aircraft as well as Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO)/CloudSat underpasses during NAWDEX delivered independent measurements for testing EarthCARE level 2 (L2) algorithms at different wavelengths and for performing a first rehearsal of the validation/calibration for EarthCARE.
Figure SB1a illustrates the complementary character of lidar and radar measurements taken during the HALO research flight (RF) on 1 October 2016. Optically thin ice clouds at cloud top are only visible in the lidar measurements (green-marked curtain), while optically thicker cloud regions are only visible in cloud radar measurements (red-marked curtain).
The DLR Falcon was equipped with a DWL payload consisting of the A2D direct-detection DWL and a 2-µm scanning coherent/heterodyne detection DWL. The A2D is the prototype of the satellite-borne wind lidar instrument on Aeolus and provides range-resolved line-of-sight wind speeds with high data coverage by exploiting both molecular and particulate backscatter return. With a view to the prelaunch activities for the upcoming Aeolus mission, NAWDEX offered the opportunity to extend the A2D dataset and to perform wind measurements in dynamically complex scenes, including strong wind shear and varying cloud conditions, as well as multiple instrument calibrations, which are a prerequisite for accurate wind retrieval. RALI on board the SAFIRE Falcon complemented the A2D instrument with wind measurements in clouds and aerosol-rich layers.
Figure SB1b shows collocated wind observations from the A2D and the 2-µm DWL from a flight of the DLR Falcon east of Iceland on 4 October 2016. The good vertical coverage, limited only by a dense cloud layer, is achieved by combining complementary information from both aerosol backscatter (A2D Mie channel and 2-µm DWL) and molecular backscatter (A2D Rayleigh channel).
HALO, SAFIRE, and the FAAM aircraft were equipped with dropsonde dispensers to measure air temperature, wind, and humidity profiles. Global NWP centers could access the dropsonde data from HALO and SAFIRE via the Global Telecommunication System in near–real time. The potential for coordinated application of the various instruments on board multiple aircraft was realized through specific instrument-driven science goals (Table 1 and see the “Active remote sensing observations for future satellite missions Aeolus and EarthCARE” sidebar). Table 2 indicates which of the research aims listed in Table 1 are addressed by each instrument.
In parallel with NAWDEX, the National Oceanic and Atmospheric Administration (NOAA) Sensing Hazards with Operational Unmanned Technology (SHOUT) campaign took place in the tropical and subtropical western North Atlantic. SHOUT utilized the unmanned National Aeronautics and Space Administration (NASA) Global Hawk aircraft with a suite of remote sensing platforms and dropsondes to study the impact of the observations on TC forecasts. During the campaign, a tropical storm (TS) moved into the midlatitudes and underwent ET, providing an unprecedented scientific opportunity to observe the interaction of such a system with the jet stream using a combination of upstream flights with the SHOUT Global Hawk and downstream flights with NAWDEX aircraft.
NAWDEX observations took place in the North Atlantic basin between 17 September and 22 October 2016. Figure 2 shows the tracks of the 47 research flights of the four aircraft, together amounting to 205 flight hours. Performing research flights over the North Atlantic is complicated because of the dense transatlantic air traffic. Commercial airliners are tightly staggered along predefined flight routes, the so-called North Atlantic tracks (NATs), between altitudes of 9 and 12 km. Operating research aircraft beneath the NATs offers high flexibility for the flight planning; however, the base height of the NATs is often too low to observe the tropopause and jet-related maximum wind speeds. Furthermore, the location of the NATs changes from day to day, depending on the forecast wind situation. Height changes and the release of dropsondes from high altitudes are not possible in the NAT area. The requirement of air traffic control (ATC) authorities to have detailed flight plans 2–3 days in advance created challenging circumstances in weather situations with reduced predictability (i.e., in situations with large changes between subsequent forecasts). Therefore, NAWDEX combined modern forecasting tools, including ensemble and adjoint-based diagnostics, and new visualization techniques to incorporate forecast uncertainty in the planning process (see the “Forecast products for investigating forecast uncertainty” sidebar).
NAWDEX focused on weather phenomena that are poorly represented in NWP, so a strong effort to estimate forecast uncertainty was essential for the planning of the IOPs. Deterministic forecasts from the ECMWF, the Met Office, the Naval Research Laboratory (NRL), Météo-France, the IMO, and the Danish Meteorological Institute (DMI) were available. Additionally, ensemble forecasts from the ECMWF, Met Office [Met Office Global and Regional Ensemble Prediction System (MOGREPS-G)], and Météo-France [Prévision d’Ensemble Action de Recherche Petite Echelle Grande Echelle (ARPEGE), known as the PEARP short-range ensemble] played an essential role.
Each day a standard set of synoptic charts and tailored weather products (e.g., PV on isentropic surfaces and WCB trajectories) were produced using a common map projection and predefined cross sections. Ensemble diagnostics of the mean and spread of several variables, as well as tailored ensemble forecast products for NAWDEX-relevant features (e.g., WCB and cyclone frequencies, and tropopause height), were created. These forecast products were provided via websites. In addition, an interactive web interface allowed the flight-planning team to compute backward and forward trajectories from planned flight tracks, facilitating the planning of flights to attempt Lagrangian resampling of air masses.
Flight planning typically requires cross-section information, for example, to obtain an accurate picture of tropopause height, winds speeds, and cloud layers and to assess forecast uncertainties along hypothetical flight routes. The NAWDEX community had access to special flight-planning tools that allowed an interactive visualization of forecast products. Central to forecasting and flight-planning operations was the Mission Support System (MSS; Rautenhaus et al. 2012). In addition, the interactive 3D forecast tool Met.3D (Rautenhaus et al. 2015a) provided specialized forecast products. Two workstations were set up at the operation center in Keflavik to run Met.3D and enable the novel ensemble forecasting workflow described in Rautenhaus et al. (2015b). Ensemble forecasts by ECMWF could be interactively analyzed in combined 2D–3D depictions. WCB trajectories and derived probabilities of WCB occurrence could be combined with additional forecast information. The ability of Met.3D to interactively navigate the ensemble data proved particularly useful, facilitating analysis of the uncertainty for features such as the predicted tropopause position.
Figure SB2 shows an example of forecast products used for planning the IOP 4 flight. The +60-h deterministic IFS forecast shows ex-TS Karl as a deep surface cyclone south of Greenland (Fig. SB2a) with cyclonically wrapped PV contours resulting from an advection of low-PV air to upper levels in the outflow of a WCB (not shown). High WCB probabilities with two distinct maxima north and east of Karl indicate that the location of the tropopause and WCB outflow is predicted with high certainty (Fig. SB2b). Images from Met.3D (Figs. SB2c,d) show the relation between the jet stream, WCB, and the tropopause in the ECMWF ensemble mean along cross sections intersecting the waveguide and the WCB east of the surface cyclone. A cross section with ensemble-mean PV (Fig. SB2c) shows a low tropopause north of the jet (depicted by an isosurface of wind speed), whereas a high tropopause appears to the south. This coincides with high probabilities of WCB (Fig. SB2d). WCB trajectories of a selected ensemble member show two distinct branches (Fig. SB2d). One branch wraps cyclonically around the cyclone and features a lower outflow compared to the second branch, which follows anticyclonic pathways at higher elevations, contributing to the elevated WCB probability maximum there. Real-time adjoint products from Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS) (Doyle et al. 2014) were used to identify regions of initial-condition sensitivity. At 1200 UTC 24 September, the maximum moisture sensitivity is located in the low to midlevels and is positioned along the eastern portion of TS Karl (Fig. SB2e). The adjoint sensitivity is computed using a kinetic energy response function located in a box (450 km × 600 km in the horizontal and extending from the surface to 700 hPa) centered on the ascending WCB at the 48-h forecast time at 1200 UTC 26 September when the IOP 4 flights were planned. Optimal perturbations derived from the adjoint sensitivity show an increase in wind speeds from 30 to over 45 m s–1 in the WCB, highlighting the importance of the midlevel moisture associated with Karl (48 h prior) for the intensification of the WCB.
HALO covered large parts of the central and eastern North Atlantic and reached flight distances up to 7150 km (∼9 h). The flights were performed either at altitudes between 11.5 and 14.2 km above the NATs for remote sensing observations or at ∼8 km to release dropsondes beneath the NATs. The two Falcon aircraft remained in radar-controlled airspaces near Greenland, Iceland, and the United Kingdom. The FAAM BAe 146 flights were north and west of the United Kingdom. A total of 289 dropsondes were released (Fig. 3a).
The research flights occurred within 13 intensive observation periods (IOPs), which were consecutively numbered and had durations of 1–6 days. Each IOP was associated with a particular weather system development and addressed one or more NAWDEX science objectives (Table 3). For easier communication, the IOPs were given names, which either corresponded to the cyclone naming of the Free University of Berlin or the National Hurricane Center (NHC), or were invented by the NAWDEX team (Table 3). Some IOPs overlap in time when different weather systems were observed simultaneously by the different aircraft.
To exploit instrument synergies and enable direct instrument comparisons, coordinated flights were performed; that is, the same air mass was nearly simultaneously probed by different aircraft on common flight legs. In total, 16 coordinated legs, with a total flight time of 14.5 h and a distance of about 10,000 km, were achieved. The longest coordinated leg with the SAFIRE Falcon and HALO on 14 October 2016 had a distance of 1,365 km (1.8 h). On two occasions the coordination involved three aircraft: HALO and the two Falcons flew together for ∼30 min (∼300 km) on 9 October between the United Kingdom and Iceland, and on 14 October, FAAM, HALO, and the SAFIRE Falcon had a common leg between the Faroe Islands and Scotland (55 min, 570 km).
Ground-based facilities and observations.
During several IOPs additional ground-based observations were taken to complement the aircraft operations and to enhance the temporal and spatial coverage of routine observations. In total 589 additional radiosondes from 40 stations in 14 countries were launched (Fig. 3b and Table 4). Of these launches, 253 were achieved through the cooperation of national meteorological agencies in the European Meteorological Services Network (EUMETNET), complemented by additional radiosondes from Iceland, the United Kingdom, France, and Norway. Launches from land stations or commercial ships were requested daily depending on the predicted evolution of weather systems. Furthermore, two additional radiosondes were launched daily during the campaign from six stations in eastern Canada, upstream of the main NAWDEX area (336 in total).
Special ground-based observations were conducted in Iceland, the United Kingdom, and France (Fig. 3b). At Keflavik International Airport, a radiosonde facility was set up by DLR to increase the frequency of the operational soundings. In cases of orographically induced gravity waves (GWs), large balloons were launched to reach altitudes up to 42 km. Also in Keflavik, a Doppler cloud radar [the Bistatic Radar System for Atmospheric Studies (BASTA); Delanoë et al. (2016)] allowed several comparisons with its airborne counterpart on board the SAFIRE Falcon during overflights. In the United Kingdom, a mesosphere–stratosphere–troposphere (MST) radar, Raman lidar, and radiosondes were operated at Capel Dewi in Wales, together with another MST radar wind profiler at South Uist in Scotland. Additionally, the MST radar at Andøya, Norway, measured tropospheric winds upon request. Two observational sites were active in France during the campaign. The site in Lannion (Brittany) operated a wind profiler, the BASTA Doppler cloud radar, and a GPS station. The Site Instrumental de Recherche par Télédétection Atmosphérique (SIRTA) near Paris, France (Haeffelin et al. 2005), operated radar and lidars, and launched radiosondes.
The fall of 2016 was a favorable period for observing midlatitude weather over the North Atlantic. The average synoptic situation for the campaign period was characterized by an increased frequency of relevant weather systems compared to climatology (Fig. 4). One of the most prominent features was a long-lasting blocking high and surface anticyclone covering large parts of Scandinavia (Fig. 4a). Extratropical cyclones occurred more frequently than normal south of Iceland and Greenland (Fig. 4b), in the core area of airborne NAWDEX observations. Consistent with the increased frequency of cyclones relative to climatology, the WCB frequency (Fig. 4c) shows increased activity over large parts of the North Atlantic. During the campaign, a succession of events with poleward transport of warm air and ascent of low-PV air into the upper troposphere was observed that appeared to strengthen the downstream anticyclonic anomaly. Most midlatitude cyclones (Fig. 4d) approached Iceland from the southwest, which was favorable for reaching them with Falcon flights from Keflavik. Only a small fraction of the extratropical cyclones moved into central and northern Europe. Six TSs occurred during NAWDEX. Ian (12–16 September), Julia (13–16 September), Karl (14–25 September), and Lisa (19–25 September) did not exceed TS strength, while Matthew (29 September–9 October) and Nicole (4–18 October) were classified as major hurricanes. Ian, Karl, and Nicole underwent ET and moved far into the midlatitudes. TPVs originating over the Canadian polar region were observed twice when they moved southward over the Davis Strait and interacted with the midlatitude waveguide.
North Atlantic weather regimes during NAWDEX show Scandinavian blocking to be the dominant regime (blue line in Fig. 5a), corresponding to the anomalous anticyclone activity over northern Scandinavia (Fig. 4a). In late September the block decayed and a short period with a positive North Atlantic Oscillation (NAO) prevailed before the Scandinavian blocking pattern was again established.
A broad measure of forecast quality during NAWDEX is provided by the anomaly correlation coefficient (ACC) of the midtropospheric geopotential height pattern over the eastern North Atlantic, as predicted by the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecast System (IFS) in fall 2016 (Fig. 5b). Periods of increased 120-h forecast errors and high spread in the ensemble forecasts are evident, and four of these periods were directly relevant to NAWDEX. Three periods of reduced forecast skill (23–27 September, 29 September–3 October, and 5–10 October) occurred during NAWDEX and two periods among the four were accompanied by a weather regime transition (Fig. 5a). Forecast uncertainty was high on 26 September during the onset of a positive NAO phase, and on 1 October during the return to the Scandinavian blocking regime. High uncertainty also occurred prior to the campaign, for forecasts initialized between 10 and 14 September, again covering a regime transition to Scandinavian blocking. This period affected NAWDEX as it complicated the planning of the transfer flight to Keflavik (IOP 1) 5 days later.
The progression of weather systems across the North Atlantic during NAWDEX can be conveniently described as a storyline characterized by upstream triggers, their dynamic interaction with the jet stream, subsequent development of disturbances, and downstream weather impacts over Europe. Three such sequences occurred completely within the NAWDEX period, and their timespan is indicated by dark gray shading in Fig. 5. In each case, low predictability was found in 5-day forecasts for the eastern North Atlantic initialized within the trigger stage (marked by a drop in forecast skill in Fig. 5b), while the final impact stage was associated with significant changes in the weather over Europe at the verification time 5 days later. The snapshots for each sequence in Fig. 6 show that the interaction of the trigger disturbance with the waveguide featured intensification of a surface cyclone, with a diabatic contribution consistent with the first three regions identified in the conceptual model presented in Fig. 1. However, the subsequent development and impact stages differed markedly, with the pattern of low PV in the downstream ridge affecting weather even farther downstream than suggested by region 4 in Fig. 1. The temporal continuity between the snapshots in Fig. 6 is shown by labeling several coherent long-lived features (identified in the caption). Prominent ridges (R1–R9) along the North Atlantic waveguide are identified as northward excursions of the jet stream (and the PV gradient). Since each ridge is characterized by low-PV air, the associated flow tends to be anticyclonic.
Sequence A is triggered by TS Karl leaving the subtropics and moving northward into the midlatitudes (Fig. 6, sequence A1). Large ensemble spread and changes between consecutive forecast runs showed that the subsequent evolution was very sensitive to uncertainties in the location and timing of the interaction of Karl with ridge R2, the trough upstream, and the associated weak surface cyclone (not shown). The interaction that took place was a merging of Karl with a low-level cyclone, leading to rapid reintensification and the formation of a cyclonic hook at tropopause level separating ridge R2 from the new ridge R3 (Fig. 6, sequence A2). The ridge building is intensified by diabatically produced low PV in the WCB outflow. Hence, in the subsequent development, the jet stream is unusually strong on its southern flank, forming a jet streak that propagates ahead from Karl, reaching Scotland the following day (Fig. 6, sequence A3). The impact on European weather occurs through the formation of a new cyclone, Walpurga (W in Fig. 6, sequence A4), which develops to the west of ridge R3, helping to amplify it. Moisture-laden air on the western flank of ridge R3 is drawn around the subtropical high. During IOP 5, HALO observed the moist boundary layer in this atmospheric river–type flow that extends to Norway, where it causes heavy, persistent rainfall, similar to the case studied in Sodemann and Stohl (2013).
Sequence B begins as sequence A ends, in a southwesterly flow situation with a long PV streamer that formed through the merger of the trough west of R3 and the large cutoff feature C (Fig. 6, sequences A3 and A4). The trigger for this sequence appears to follow from the vortex rollup of the streamer through shear instability, resulting in a new cutoff over Newfoundland (V in Fig. 6, sequence B1), which then interacts and merges with a large-scale trough west of R5 advancing rapidly from the northwest. Note that ridge R5 and its upstream trough wrap up cyclonically during the development so that the trough catches up with the cutoff to the south of R5. The tropopause was very low just in the very center of this system, which therefore has been named the stalactite cyclone (St in Fig. 6, sequence B2). In the development stage, a second cyclone (F in Fig. 6, sequence B3) intensified rapidly between ridge R6 and the trough to its west. The poleward-moving air in R6 crossed Iceland and reinforced the anticyclonic anomaly formed by ridge R5 of the stalactite cyclone. The impact of the sequence comes not as a classical severe weather event, but through the establishment of a strong blocking anticyclone over northern Europe, which persisted for the next 2 weeks.
Sequence C begins with two upstream triggers. A TPV originating in the Canadian Arctic is carried rapidly southeastward on the poleward flank of the jet stream (T in Fig. 6, sequences B4 and C1). It is hypothesized that the TPV locally enhanced the cyclonic circulation about the tip of the large-scale trough (T in Fig. 6, sequence C1), which eventually wrapped cyclonically over Iceland (Fig. 6, sequence C2). At the same time, the remnants of cutoff C appear to be associated with the emergence of a small surface cyclone, which has been named Sanchez (S in Fig. 6, sequence C1). The European dipole block (cf. Rex 1950) is well established at this time so that the ridge R8 and the cyclonic PV anomaly over Iceland are held stationary and a PV filament forms in the deformation region on their western side. The filament is unstable and experiences vortex rollup, forming three tropopause-level cyclonic vortices. The key interaction in this sequence occurs as the low-level Cyclone Sanchez passes the southernmost cutoff, but then phase locks with the central cutoff, resulting in baroclinic intensification (S in Fig. 6, sequence C2). As the sequence develops, the resulting cutoff cyclone progresses slowly eastward (Fig. 6, sequence C3) and is responsible for some of the most dramatic high-impact weather during NAWDEX, with heavy precipitation and flooding across southern France and northwestern Italy in the southerly flow ahead of it (Fig. 6, sequence C4). But this HIW is not the only significant outcome of sequence C. Returning to stage C2, ridges R8 and R9 are similar in their horizontal extent, but the tropopause is much higher above R9 than R8 (not shown) with the result that the anticyclonic circulation induced by R9 is stronger and R8 is stretched out meridionally between R6 and R9 (Fig. 6, sequence C3). As NAWDEX draws to a close, the ridge R9 extends rapidly into the Arctic, reinforcing the block and forming a PV anomaly in the shape of the Icelandic character Þ (the first letter of Þor, pronounced Thor, the ancient Norse god of HIW).
It is important to note that the large-amplitude ridge building leading to the Thor block is not R8, which developed as part of the interaction phase of sequence C, but rather R9, which is associated with a second cyclone that develops to the west (and which may play a role in the cutoff of Sanchez). Indeed for all three sequences, the development stage leading to weather impacts over Europe appears to be associated with a second cyclone that forms in an environment modified by the interaction of the trigger disturbance with the midlatitude flow, and whose development is difficult to project because of the low predictability of the environment.
HIGHLIGHTS OF NAWDEX.
Observations in NAWDEX were organized into IOPs that focused on key weather systems involved in the longer sequences (Table 5). Several of these IOPs are unprecedented in terms of the phenomena that were sampled or the comprehensive coverage and multifaceted nature of the measurements. While the analysis of the data is just beginning, a first impression of the results can be obtained from four highlights that illustrate the unique sets of multiplatform and multi-instrument observations that were obtained.
Extratropical transition of TS Karl.
The evolution of TS Karl is the central feature of sequence A discussed above and IOP 4. It was the first extratropical transition sequence that has been observed with research aircraft through all stages of its development, including TS status, ET, reintensification with impacts on jet-stream strength, moisture transport, and downstream HIW (Table 5). By flying over the TS and its northwestern flank twice, the SHOUT Global Hawk observed the development stage that occurred far south of the midlatitude jet stream on 22–23 September (Fig. 7a) and the ET phase on 24–25 September (Fig. 7b). On 26 September, HALO observed the interaction with the waveguide and reintensification phase of the storm by flying over the cyclone center (Fig. 7c), WCB ascent, the low-valued PV air in the WCB outflow, and the dry intrusion (not shown). When Karl moved rapidly toward Scotland, decaying in strong horizontal shear on 27 September, IOP 5 focused on the intense jet streak at the tropopause level and the strong moisture transport along the equatorward side of the jet with a combination of HALO, FAAM, and DLR Falcon flights (Fig. 7d).
IOP 4 will contribute to answering several of the posed research questions (see Tables 1 and 3). The large number of dropsonde and special radiosonde measurements that were assimilated into operational forecasts in real time will provide a basis for observational impact and predictability studies. Detailed airborne remote sensing observations will allow examination of the role of diabatic processes and their representation in numerical models. Both the synergies of the instruments and the storm-following observational strategy give unprecedented information about this intense and long-lived cyclone and a unique opportunity to analyze forecast error growth due to in situ processes versus downstream propagation.
Cloud physics in a WCB.
IOP 3 focused on observing the vertical cloud structure and cloud microphysical processes in a WCB that was related to the Midlatitude Cyclone Vladiana south of Iceland and west of Scotland on 23 September 2016 (Fig. 8a and Table 5). The WCB transported moist air northeastward just west of the United Kingdom, as indicated by the low-valued PV air in the upper troposphere (Fig. 8a).
HALO first stayed beneath the NATs at altitudes of ∼8 km on the way to the southwesternmost point of the flight (white circle) to begin the first of three sections across the WCB. On this leg to Ireland, 12 dropsondes were released before HALO climbed to ∼13 km in Irish airspace. Over northern Ireland, HALO and FAAM joined to perform coordinated remote sensing and in situ observations of the WCB. HALO measured the WCB by remote sensing from above while FAAM performed four in situ legs at different altitudes to measure cloud microphysical parameters inside the WCB. After the coordinated leg, HALO crossed the WCB a third time and observed the outflow of the WCB between Scotland and Iceland.
Figure 8 focuses on the first and second crossings of the WCB. The water vapor lidar measured water vapor profiles throughout the troposphere and lower stratosphere in the absence of clouds (Fig. 8b). On the western side of the cross sections, where HALO was located along the stratospheric side of the waveguide, the postfrontal troposphere was cloud free except for boundary layer clouds reaching up to 2 km. The water vapor shows high variability, which portrays the dynamically modulated transport of moisture related to Cyclone Vladiana. On both crossings of the waveguide, one west to east and one east to west, a tilted dry layer is visible at altitudes between 5 and 9 km (1110–1125 and 1305–1325 UTC), related to a dry intrusion west of the low-level cold front. The wedge-shaped moist layer on top is associated with high moisture values in the WCB outflow. The second crossing at high altitudes depicts a strong vertical moisture gradient, on top of the elevated moist layer, that marks the tropopause and extends farther east into the area where WCB clouds reach high altitudes. A decrease of the tropopause height is detected toward the west on the second leg. The radar shows two vertically (∼11.5 km) and horizontally (∼400 km) extended and coherent clouds (Fig. 8c) representing the double crossing of the WCB. In between (i.e., on the eastern side of the WCB), cloud tops are lower and the clouds are intermittent. The sharp vertical gradient in radar reflectivity at about 3-km altitude marks the melting layer.
On the second transect the FAAM aircraft performed in situ measurements on flight legs beneath HALO (Figs. 8b,c). HALO met FAAM at the beginning of its second WCB leg (purple diamond marker in Fig. 8c), where FAAM started its lowest leg at about 3-km altitude, just above the melting layer, with subsequent legs at 4, 6, and 7.5 km. The in situ observations show that both mixed-phase and ice-only clouds were encountered during the low-level run, but during the high-level runs only ice was observed. The ice water content (IWC) in the WCB shows maximum values of 0.4 g m−3 on the lowest two legs (Fig. 8d). Ice images show large differences in the form of the particles at different altitudes. On the lowest leg, large aggregates (∼6 mm) dominate close to the freezing level, while at the highest level higher concentrations of small irregularly shaped crystals (<1 mm) prevail.
HALO also observed the interaction of Vladiana’s WCB outflow with the jet stream in coordination with the DLR Falcon (not shown). IOP 3 contributes to all research aims (Table 3) and future work on the cloud microphysics observations will investigate, for example, the correlation of increased IWC with particularly high radar reflectivity. Data from liquid and ice particle size distributions will be used to improve the retrieval of cloud properties from the HALO remote sensing instruments. Overall, this is a unique set of comprehensive and complementary airborne observations of a WCB, its embedded microphysical processes, and its outflow interaction with the jet stream.
Wind observations in the jet stream and outflow of a WCB.
Figure 6 (sequence B2) shows the stalactite cyclone that formed previously via merging of two near-surface vorticity maxima with a very intense, small-scale, upper-level PV anomaly south of Newfoundland (not shown). The rapid development of the cyclone occurred in the mid-Atlantic between 30 September and 2 October. On 2 October (IOP 6), a coordinated flight of the DLR and SAFIRE Falcons observed WCB ascent and outflow when the stalactite cyclone was most intense (Fig. 9a). The aircraft flew together to intersect the jet stream on the northwestern edge of ridge R5, wrapping cyclonically around the stalactite cyclone. On a common leg between Iceland and Greenland both aircraft crossed the jet stream (Fig. 9b) and made complementary wind observations (Figs. 9c,d). The Doppler wind lidar (DWL) on the DLR Falcon observed two wind maxima up to 50 m s−1 in cloud-free regions and in optically thin cirrus in the WCB outflow. Complementarily, the SAFIRE radar observed in-cloud winds in the region of WCB ascent in the mid- and lower troposphere. Only in dry and aerosol-poor air masses over Greenland (i.e., on the stratospheric side of the waveguide) does the combination of both instruments provide poor data coverage. The SAFIRE Falcon released nine dropsondes when crossing the jet stream, yielding further profiles of winds, temperature, and moisture.
Future research on IOP 6 will be mainly dedicated to predictability issues associated with the blocking formation downstream of the cyclone. The block formed at a time when a loss of predictability in the ECMWF forecasts occurred (Fig. 5b). Winds measured by the two aircraft will help to characterize the role of the WCB outflow in the ridge building. The observed high winds and strong vertical gradients were repeatedly observed on flights across the jet stream with observed maxima up to 80 m s−1 and were often related to strong vertical wind speed gradients up to 30 m s−1 km−1. A unique aspect of this example is the benefit of coordinated flights with complementary instruments to address one of the key objectives of NAWDEX (Tables 1 and 3): observing the strong wind shear and PV gradients near a WCB outflow.
HIW related to cutoff Cyclone Sanchez.
Cutoff Sanchez was initiated in the central North Atlantic and reached southern Europe between 12 and 14 October 2016 (Fig. 6, sequence C). On its leading edge moisture was advected northward (Fig. 10a) on 13 October when it triggered heavy precipitation and strong winds over France and Italy. The 24-h accumulated precipitation in the Herault region reached ∼250 mm (Fig. 10b), and wind gusts exceeding 100 km h−1 were observed along the French Mediterranean coast (Fig. 10c). As in typical Cévenol episodes, strong southerlies brought warm and moist air from the Mediterranean Sea toward the Massif Central and caused heavy orographic precipitation over the mountain ranges of the Cevennes. Upper-level cutoffs like Sanchez are known to be favorable synoptic conditions for triggering convective mesoscale events (Nuissier et al. 2008), which were intensively studied during the recent Hydrological Cycle in the Mediterranean Experiment (HyMex) field campaign (Ducrocq et al. 2016). Part of the air masses responsible for the HIW subsequently reached as far north as the SIRTA site near Paris, causing precipitation during the afternoon of 13 October (Fig. 10d). This episode illustrates one of the key NAWDEX aims (Table 1), which is to investigate how HIW events over Europe are associated with complex waveguide dynamics (in this case the formation of a PV cutoff) over the upstream North Atlantic. The combination of the ground-based data with NAWDEX observations both from aircraft and from the many additional radiosondes taken during IOP 9 will enable detailed studies of the forecast sensitivity of HIW to upstream initial condition errors.
NAWDEX was the first field experiment with synergistic airborne and ground-based observations from the entrance region to the exit region of the climatological storm track, and was undertaken to investigate the role of diabatic processes in altering jet-stream disturbances, their development, and their effects on HIW downstream.
Because of the operational limitations on research aircraft flights over the North Atlantic and Europe and the need for high-resolution profile measurements of thermal and dynamic properties, NAWDEX focused on airborne remote sensing observations and the deployment of multiple aircraft. Advanced instruments for remote sensing of wind, water vapor, and cloud properties provide an integrated picture of the atmospheric structure in regions where diabatic processes were active, from the synoptic to subkilometer scale. The unique combination of the four aircraft and the first deployment of HALO in a campaign focusing on midlatitude dynamics allowed observations over large parts of the North Atlantic. Often, the same weather system could be sampled at different stages of its development, and the interactions of successive weather systems have been observed following the NAWDEX storyline. Additional ground-based observations and an enhanced density of operational radiosonde releases yielded very high coverage with high-resolution vertical profiles from the ground to the lower stratosphere. The region with enhanced atmospheric profiling extended from eastern Canada to most parts of Europe. The coverage and fidelity of the resulting observations will enable future studies to estimate diabatic heating through the use of models and diagnostics constrained by the NAWDEX observations, particularly in situations when the atmospheric flow is especially sensitive to small changes in diabatic heating. Over the 13 IOPs it was possible to address all of the original campaign objectives (Tables 1 and 3). Table 5 lists a number of particular highlights and “firsts” that have drawn the attention of the NAWDEX scientists.
The success of the observational campaign was possible because of the favorable meteorological conditions, with many cyclones and WCBs in the vicinity of Iceland. Importantly, the NAWDEX period contained episodes of reduced predictability, indicating that uncertainties originating in the estimated atmospheric state and model formulation grew rapidly. The suggestion that these uncertainties spread via their impact on the life cycle of a “second cyclone” forming to the west, rather than through a process of downstream development, shows that NAWDEX has the potential to make an important contribution to the study of predictability of midlatitude weather and the representation of uncertainty in EPSs. Since there were also episodes of HIW in Europe connected to disturbances of the North Atlantic waveguide, NAWDEX also offered a unique opportunity to explore HIW predictability.
To the best of our knowledge, the NAWDEX period provides the most complete set of combined wind, humidity, temperature, and cloud profile observations of the North Atlantic jet stream yet assembled. This dataset will form the basis of detailed case studies and evaluations of weather and climate prediction models for many years. The widespread coverage of high-resolution multivariate cross sections across the jet stream and weather systems developing from one side of the North Atlantic to the other enables examination of the whole chain of processes from the triggering of disturbances on the waveguide to the ultimate impact on weather systems affecting Europe.
Many international institutions contributed to the implementation of NAWDEX and the results presented in this overview article. NAWDEX was only possibly due to the close cooperation of many colleagues including the planning team, the instrument groups, and the flight operation crews ( appendix). The HALO and DLR Falcon campaign received funding from DLR [project “Klimarelevanz von atmosphärischen Spurengasen, Aerosolen und Wolken” (KliSAW)], ETH Zürich, NRL Monterey, the German Science Foundation (DFG; within SPP1294 HALO and SFB/TRR165 Waves to Weather), the European Facility for Airborne Research (EUFAR; project NAWDEX-Influence), and the European Space Agency (ESA), providing funds related to the preparation of Aeolus (WindVal II, Contract 4000114053/15/NL/FF/gp) and EarthCARE (EPATAN, Contract 4000119015/16/NL/CT/gp). Special thanks are due to the Max Planck Institute Hamburg for sharing the HALO payload during the NARVAL-II and NAWDEX campaigns in summer and fall 2016 and for their general support of NAWDEX. We are grateful to the DLR flight experiments team, in particular, the colleagues of flight operations for the careful preparation and the outstanding support in Keflavik. The authors are grateful to the HALO and Falcon pilots and the technical and sensor team from DLR flight operations for excellent support prior to and during NAWDEX. The SAFIRE Falcon contribution to NAWDEX received direct funding from L’Institut Pierre-Simon Laplace (IPSL), Météo-France, Institut National des Sciences de l’Univers (INSU) via the LEFE program, EUFAR Norwegian Mesoscale Ensemble and Atmospheric River Experiment (NEAREX), and ESA (EPATAN, Contract 4000119015/16/NL/CT/gp). The U.K. funding for the FAAM aircraft flights, dropsondes, and additional radiosondes was provided by the Met Office. Special thanks are due to the SHOUT mission for their open collaboration, which enabled coordinated observations of Tropical Storm Karl. Observations at Capel Dewi were funded by the National Centre for Atmospheric Science (NCAS). We thank the European Meteorological Service Network (EUMETNET) for funding additional radiosondes and for providing access to the data. Environment and Climate Change Canada (ECCC) provided funding for enhanced radiosonde frequency during the campaign. The Icelandic Meteorological Office (IMO) and DLR are thanked for providing additional radiosondes from Iceland. We thank ECMWF for providing access to data within the framework of the Support Tool for HALO Missions (SPDEHALO) special project. We gratefully acknowledge many discussions with U.S. colleagues, in particular Chris Davis, Pat Harr, and Heather Archambault, during the scientific preparation of NAWDEX. We thank Michael Sprenger (ETH Zürich) for preparing the data used in Fig. 4. MB and CMG acknowledge funding from the Swiss National Science Foundation (Projects 200020_165941 and PZ00P2_148177, respectively). JDD and CAR acknowledge the support of the U.S. Navy Chief of Naval Research through the NRL Base Program PE 0601153N, and the ONR PE 0602435N. We are very grateful to three anonymous reviewers for their constructive comments that helped to improve the manuscript.
APPENDIX: THE NAWDEX TEAMS.
Table A1 shows the NAWDEX participants, along with their affiliation and role(s) in the campaign.