Highlights from Research Area B: Moist processes and diabatic Rossby waves

Research area B of the PANDOWAE project investigated aspects of the role of diabatic processes in mid-latitude weather systems. Diabatic processes include latent heating and cooling due to i n-cloud and below-cloud microphysical processes, surface fluxes and radiation. One way of quantitatively analyzing the effect of diabatic processes on atmospheric processes is via the potential vorticity (PV) framework, since diabatic effects can alter the PV of an air parcel and thereby impact upon the larger-scale balanced flow (Hoskins et al. 1985). PANDOWAE research used this PV framework and combined Eulerian and Lagrangian diagnostics to analyze how diabatic processes, mainly associated with extratropical cyclones and extratropical transition systems modify (i) the upper-tropospheric jet and evolving Rossby wave structure downstream, and (ii) the structure and dynamics of the weather system itself. In addition trajectory-based techniques have been used to identify the evaporative moisture sources and transport pathways involved in the formation of deep cyclones and intense precipitation events.

a. Diabatic processes modify upper-tropospheric jet and downstream Rossby wave evolution (contributing to RA-B Questions b3 and b5)
Strongly ascending airstreams associated with mesoscale convective systems, tropical cyclones (TCs), the extratropical transition (ET) of TCs, and extratropical cyclones are associated with intense latent heating due to condensation and other cloud microphysical processes (Joos and Wernli 2012). On average, the PV in these ascending air parcels is first increasing, leading to low to mid-level cyclonic PV anomalies (see section b), and then decreasing to low tropospheric values of typi cally 0–0.5 pvu. The height of these low-PV out flows is often above the position of the climatological tropopause, thus leading to the formation of coherent and large-amplitude anticyclonic PV anomalies at the level and in the direct vicinity of the upper-tropospheric jet stream. In a detailed case study of the North Atlantic flow in August 2008, Grams et al. (2011) found that important diabatic PV modifications occurred at upper levels, associated with the cross-isentropic transport of low-PV air during the ET of Hanna and the development of an extratropical cyclone near Newfoundland. The diabatic outflow of so-called warm conveyor belts (WCBs) contributed to the amplification of ridges downstream of each cyclone (see Figure 1) and the downstream Rossby wave amplitude, in this case leading to the triggering of a Mediterranean cyclone and heavy precipitation event a few days later. An in-depth process understanding of these moist diabatic influences on the upper-tropospheric flow has been acquired in the context of ET research. Using the ex ample of Jangmi (2008), Grams et al. (2013a) showed that sustained diabatically driven out flow during ET is key to downstream ridgebuilding and jet acceleration.

Figure 1: The left panel shows PV (shading) on 320 K and sea level pressure (black contours every 5 hPa) at 12 UTC 09 Sep 2008. The black crosses mark the intersection of WCB trajectories with the isentropic layer between 317.5 and 322.5 K. The right panel shows a Meteosat IR image with 2, 5 and 10 pvu contours of PV on 320 K. The ‘L’ labels mark the surface position of different cyclones. Figure taken from Grams et al. (2011).

Thereby ascent transitions from being convective-upright to more WCB-like slantwise along the midlatitude baroclinic zone later during ET. Quinting and Jones (2016) corroborated this finding from a climatological perspective. Grams et al. (2013b) further showed that the phasing of the diabatic outflow and midlatitude flow features is key to the final Rossby wave amplitude. Grams and Archambault (2016) framed a generalized picture of the role of diabatic outflow in upper-tropospheric flow modification during ET and stress that diabatic outflow from different weather system categories act in concert to yield a highly amplified downstream Ross by wave pattern. Several other case studies and climatological investigations confirmed the important role of diabatic processes in WCBs and other weather systems with strong latent heat release in modifying the upper-level PV evolution (e.g., Madonna et al. 2014; Grams et al. 2015; Schäfler and Harnisch 2015; Schneidereit et al. 2016). The study by Schneidereit et al. (2016), performed jointly with RA-A, stressed that strongly enhanced diabatic WCB outflow was key to the formation and maintenance of a blocking anticyclone over Alaska that initiated a major sudden stratospheric warming event in January 2009. Schäfler and Harnisch (2015) emphasized that the boundary layer moisture in the WCB inflow affects the latent heat release and therefore the WCB out flo w height and eventually the amplitude of the negative PV anomaly induced by the WCB. Schäfler et al. (2011) showed that the boundary layer humidity in the inflow region of a WCB can differ by several g/kg when comparing state-of-the-art ECMWF analyses with lidar measurements.

As a consequence of this uncertainty of boundary layer humidity and shortcomings in the representation of cloud microphysical processes in model, the evolution of humidity, cloud water and cloud ice along warm conveyor belts, which typically span a temperature range from +20 to –50°C, is not well known. Therefore, PANDOWAE scientists, supported by DLR and ETH, organized the aircraft campaign T-NAWDEX-Falcon in October 2012 with the aim of collecting in-situ observations of humidity, ice particles and trace gases in WCBs. As a special challenge for the flight planning (Schäfler et al. 2014), one objective was to sample the same air parcel several times during consecutive flights as the air parcels rises along the WCB. Reaching this objective was difficult because of severe flight permission constraints over most parts of Europe and because flight plans needed to be submitted a few days ahead, i.e., at times when the forecasts used for the identification of the WCBs and the planning of the Lagrangian flight matches were still to a certain degree uncertain. Nevertheless, it is a clear success of T-NAWDE X-Falcon that for the first time a few Lagrangian flight matches in WCBs have been achieved (M. Boettcher, personal communication) and that a mainly theoretically working research group like PANDOWAE was able to define the objectives for a 2-week aircraft field experiment. Thanks to the excellent exchange with colleagues at the ECMWF, it was also possible to set up, within a short time period, a diagnostic tool on the ECMWF computer system, which routinely, twice a day, identifies WCBs in each of the 50 ensemble members and produces probability maps for the occurrence of WCBs (see example in Figure 2; Rautenhaus et al. 2015). Note that ensemble data is only archived on few pressure levels, which impedes the calculation of meaningful trajectories – and therefore such products can only be computed “on the fly ”, i.e., as long as the full resolution data is available on disk.

Figure 2: WCB ensemble probability (in % shaded), valid at 12 UTC 19 October 2012 (i.e., at the time of IOP3 of T-NAWDEX-Falcon), for forecasts initialized at (left) 12 UTC 17 October (+48h), and (right) 00 UTC 19 October (+12h). Figure taken from Schäfler et al. (2014).

b. Role of diabatic processes for the formation and intensification of intense cyclonic systems (contributing to RA-B Questions b1, b2, b4, b6, b7 and b8)
It has been well known from many case studies and “dry” sensitivity experiments with numerical models that latent heating and the associated low -level PV production contribute favorably to the intensification of extratropical cyclones (e.g., Hoskins and Berrisford 1988; Davis and Emanuel 1991; Kuo et al. 1991). At the time of maximum intensity, the typical vertical PV profile of a deep extratropical cyclone is characterized by a so -called “PV tower”, i.e., high PV values in t he upper troposphere due to adiabatic descent of stratospheric air towards the cyclone center, and – compared to climatology – also strongly enhanced PV values in the lower troposphere due to diabatic PV production. Campa and Wernli (2012) investigated composites of vertical PV profiles of about 10’000
winter cyclones in ERA-Interim and corroborated climatologically that for deeper cyclones, both the adiabatic upper-level and the diabatic lower-level PV anomalies are in general enhanced compared to weaker cyclones.

Diabatic Rossby waves (DRWs) constitute a particular category of cyclones, for whose existence diabatic effects and low-level PV production are essential (and not just an amplifying factor). DRWs are rapidly propagating, shallow, low-level mesoscale vortices in a strong baroclinic zone, which – if interacting with an approaching trough – can deepen explosively. Boettcher and Wernli (2011) investigated such a storm over the North Atlantic and confirmed the negligible forcing from upper levels during the rapid propagation phase. The study also revealed that operational ECMWF forecasts, even if they capture the DRW in the early stage, can have huge problems in predicting the correct track and intensity of t hese potentially high -impact cyclones. Based on an increased understanding of the dynamics of DRWs, Boettcher and Wernli (2013) then developed a specific DRW tracking method. The challenge in doing so was to distinguish DRWs on the one hand from “ordinary cyclones” (with a substantial upper-level forcing throughout the development) and on the other hand from low-level PV features that are related topography or fronts, but not to propagating cyclones. This tracking tool has been applied to 10 years of operational ECMWF analyses in order to produce a first Northern Hemisphere climatology of DRWs. DRWs are more frequent over the North Pacific than over the North Atlantic (see Figure 3) with on average 81 and 43 systems per year, respectively. Less than 15% of these DRWs intensify explosively, on average 12 per year over the Pacific and 5 over the Atlantic. For selected DRWs, sensitivity experiments with the COSMO model (B oettcher 2010; Campa 2012) – modifying for instance the low level humidity field in the region of the DRW propagation – revealed the complex interplay of ambient moisture and the evolution of the DRW.

Figure 3: Track density of DRW centers in the years 2001–10 from genesis till decay (number of events within 3° x 3° boxes) in the (left) North Atlantic and (right) North Pacific. The black lines denote values of 1 and 30. Black dots show the DRW genesis positions in the warm season May–October and open circles in the cold season Nov.–April, respectively. Figure from Boettcher and Wernli (2013).

Finally, RA-B investigated diabatic influences on the structure of TCs with focus on the ET stage. Lang et al. (2012a) showed that moist singular vectors and higher model resolution better represent uncertainty in the TC intensity and track and in particular perturb the TC outflow more strongly resulting in larger downstream impact compared to dry singular vectors or lower resolution ensemble forecasts (Figure 4). They concluded that this is mainly due to a better representation of moist diabatic processes. The combination of observational data from the THORPE X P acific -Asian Reginal Campaign (T-PARC) and numerical data allowed to document the role of diabatic processes and latent heat release on the structure of Typhoons Sinlaku (2008) and Jangmi (2008) during ET (Grams et al. 2013a; Förster et al. 2014; Quinting et al. 2014).

Figure 4: Total energy (colored surfaces) of the leading TL255 SV at 0000 UTC 17 Sep 2006 and PV (1 PVU; gray surface) of accompanying linearization trajectory: (e) evolved moist SV, and (f) evolved dry SV. Figure taken from Lang et al. (2012a).

c. The link ages of surface evaporation, moisture transport and severe weather (contributing to RA -B Questions b3 and b5)
A third category of research in area B studied aspects of the atmospheric moisture transport, in particular the evaporative sources of severe precipitation events and extratropical cyclones. The Lagrangian technique by Sodemann et al. (2008) is able to diagnose surface evaporation sources and transport pathways for the moisture in a target region of interest. The technique is based on the 6 - hourly changes of specific humidity q along trajectories; moisture sources due to surface evaporation are identified where q increases in air parcels located in the boundary layer. This approach has been applied to study the moisture transport leading to the severe June 2013 flood event in Central Europe (Grams et al. 2014) and a more local flood in the Swiss Alps in October 2011 (Piaget et al. 2015). For the event in 2013, characterized by a quasi-stationary PV cutoff at upper levels, an inverted baroclinicity (with warm air over the Baltic Sea and cold air over the Alps), and unusually westward tracking cyclones, almost all moisture sources have been diagnosed over continent al Europe, which, most likely was due to intense rainfall during the weeks preceding the flood. In contrast, for the Alpine
flood in 2011 a part of the moisture had a tropical origin (see Figure 5) and a subtropical cut-off low served as an important moisture collector, which enabled the very long-range transport from Western Africa, over the eastern North Atlantic, and eventually from a northwestern direction towards the Alps.

The same moisture source diagnostic was also used by Campa (2012) to investigate the origin of the moisture that leads to the diabatic PV production in the lower part of extratropical cyclones [cf. section (b)]. It was found that the main moisture uptake occurs 12-72 hours prior to the time of PV production. Cyclones in the eastern part of the North Atlantic and North Pacific collect their moisture over a much larger area than those in the western parts.

PANDOWAE research also contributed to the analysis of severe weather events that occurred during the special observation period of the Hydrological Cycle in t he Mediterranean Experiment (HyMeX) in autumn 2012. Röhner et al. (2016) investigated moisture uptake regions for severe and mainly convective heavy precipitation event over Spain and found that the early part of the event was fed with moisture that evaporated from the eastern North Atlantic and the later part with moisture from the Mediterranean Sea itself.

Figure 5: Evaporative moisture sources (shading) for the October 2011 Alpine flood event (i.e., the 12-h accumulated precipitation that fell in Switzerland (red box) from 1800 UTC 09 to 0600 UTC 10 October. The blue dots indicate the location of each trajectory at 10 days prior to arrival in the red box. Figure taken from Piaget et al. (2015).