Investigation of the diabatic Rossby-wave dynamics
Maxi Böttcher (1), Heini Wernli (2)
(1) Institute for Atmospheric Physics, University of Mainz
(2) Institute for Atmospheric and Climate Science, ETH Zurich
1. Introduction
Diabatic Rossby-waves (DRWs) are low-tropospheric positive potential vorticity (PV) anomalies that are continuously regenerated through diabatic processes leading to a rapid propagation along a typically intense baroclinic zone. It has been hypothesized that DRWs can be important precursors for rapid cyclone development. Previously, the mechanism of DRWs has been studied mainly in idealized channel flows. Here a detailed case study is performed of a DRW that is involved in explosive cyclone development over the North Atlantic. Operational ECWMF analyses and forecasts as well as the meso-scale COSMO model are used to investigate the dynamics of the event. To generalize the findings of the case study, a 10-year statistics of DRW occurrence has been calculated with the aid of a sophisticated tracking algorithm that is adaped to the special DRW mechanism.
2. Forcing of vertical motion during the DRW life-cycle
After generation within a mesoscale convective system the low-level positive PV anomaly propagates as a DRW along the intense baroclinic zone that extends eastwards over the North Atlantic. Moist warm air at the leading edge of the PV vortex ascends and produces positive low-level PV steadily downstream of the system by latent heat release.
The DRW appears in Fig. 1a as the elongated high PV structure to the east of Florida at 00 UTC 19 Dec 2005. It is clearly recognisable that the blue contour indicating the tropopause at 250 hPa is far to the north during this propagation phase of the DRW life cycle. The DRW propagates rapidly towards the east without significant forcing from upper levels.
A quasi-geostrophic omega diagnostic that provides a height-attributable solution of the omega equation has been applied to corroborate this finding and to distinguish between ascent forced by the upper-level waves and by the DRW system itself at lower levels. In the vicinity of the DRW, Fig. 1b shows ascent in the downstream region and weaker descent to the rear of the DRW, both forced from lower levels. Other structures of ascent forced by the upper levels are far away from the DRW.
(a) (b)
Figure 1. Snapshot during DRW propagation at 00 UTC 19 Dec 2005: (a) PV (vertically averaged between 975 and 800 hPa, in pvu, see color bar), SLP (grey lines, interval 4 hPa) and the PV = 2.5 pvu contour at 250 hPa (blue line), and (b) vertical velocity at 700 hPa forced from lower levels (1000-750 hPa, red lines, solid and dashed lines for ascent and descent, respectively) and from upper levels (650-100 hPa, blue lines, style as before).
After more than a day of propagation, on 19 Dec 2005 the DRW comes into the influence of an upper-level trough approaching from the west. The coupling between upper-level features and the DRW induced circulation introduces the phase of the DRW intensification. The upper-level system leads to vertical ascent that couples with the ascending air of the DRW, leading to rapid SLP deepening (Fig. 2a). The overlap of the vertical motion forced from different tropospheric levels during intensification is depicted in Fig. 2b. The interaction of the upper and lower-level positive PV anomalies results in an explosive decrease of the surface pressure of 34 hPa from 00 UTC 20 Dec to 00 UTC 21 Dec. The cyclone attains a SLP minimum of 971 hPa at 00 UTC 21 Dec 2005.
(a) (b)
Figure 2. Snapshot during DRW intensification at 12 UTC 20 Dec 2005. Shown are the same fields as in Fig. 1.
3. Investigation of operational ECMWF forecasts
The predictability of the event is investigated by four operational forecasts started 00 and 12 UTC at 17 and 18 Dec. Despite the fact that the DRW was represented well in each of the forecasts two of them failed the explosive intensification. Figure 3a shows the low-level PV evolution that features slightly decreasing low-level PV values during the DRW propagates and a strong increase in case the intensification took place. The intensifying DRWs reached a somewhat deeper sea level pressure (SLP) minimum at nearly the same time as in the analysis (Fig. 3b).
(a) (b)
Figure 3. Time development of the 4 forecasts and the analysis: (a) low-level PV maximum [pvu] and (b) SLP [hPa]. The vertical dashed line separates the propagation from the intensification phase.
By application of a tracking tool the environmental factors that influence the DRW during propagation and intensification can be calculated. It is shown that the baroclinicity downstream and the moisture supply to the south of the system influence the DRW growth mechanism during the propagation (Fig. 4).
(a) (b)
Figure 4. Time development of the 4 forecasts and the analysis during the propagation phase: (a) low-level specific humidity [g/kg] and (b) baroclinicity at 950hPa [K].
The intensification went wrong in two of the forecasts due to the missing interaction of DRW and upper-level trough. The jet is crossed by the DRW in the two forecasts with the intensifying cyclones (Fig. 5 a).
(a) (b)
Figure 5. Same as Fig. 3, here (a) distance to the jet [km] and (b) upstream ascending vertical motion at 500hPa [Pa/s]. The vertical dashed line separates the propagation from the intensification.
The ascending vertical motion at 500hPa upstream of the DRW is assumed to be mostly affected by the upper-level dynamics. It shows very weak values for the poor forecasts whereas the intensifying DRWs in the forecasts close to the analysis are influenced by strong vertical ascent at the 20 Dec 2005 12UTC (Fig. 5 b). The wrong phasing of the upper-level trough to the low-level DRW led to the missed interaction and intensification in the poor forecasts.
4. Simulations with the meso-scale COSMO model
The non-hydrostatic regional COSMO model is a suitable instrument to simulate the DRW case in a higher spacial and temporal resolution of 14km and an hourly data output, and in an artificially modified environment. The purpose of the experiments was to test the mechanism of the DRW and its robustness. In the control run of the COSMO model the DRW life cycle is reproduced similar to the development in ECMWF analyses.
Artificial modifications were performed in limited areas in the vicinity of the DRW to avoid the influence to other weather systems. It could be shown that a DRW (i) not develops in a completely dry model setup, (ii) decays with moisture denial in a limited box over the track and (iii) decays by moisture denial in the moisture source region to the south of the track. This confims the essential role of moisture for the DRW mechanism.
Another experiment where the latent heat release in a box over the DRW track was suppressed led also to the stop of the propagation and the decay of the system. Figure 6 shows the comparison between
(a) (b)
Figure 6. DRW in the COSMO model: low-level averaged PV (pvu, colours), SLP [hPa] and PV at 250hPa (blue contours, 1.5 and 2pvu) at 19 Dec 2005 06 UTC for (a) control run and (b) run with suppressed latent heat release in the green box.
control run (ctl) and the run with suppressed latent heat release (nolh) in the green box at 06 UTC 19 Dec after 36 hours of model run. The DRW in nolh weakened rapidly after 10 UTC 18 Dec when the region of ascent and condensation at the downstream side of the DRW approached the modified box where no latent heating took place. The run nolh confirms the dependence of the PV generation on latent heat release that continuously regenerate the DRW.
The importance of DRWs is to act as possible precursor to explosive1) cyclone deepening. In the investigated case the DRW contributed to the strong pressure deepening of 34hPa/24h according to ECMWF analyses and 24 hPa/24 h in the COSMO control simulation (black lines in Fig. 3b and 7b).
After the decay of the DRW in nolh the approaching upper-level trough generated a new cyclone. The new created vortex was positioned about 400km further west than the DRW in ctl and therefore closer to the edge of the upper-level trough. The circles in Fig. 7a mark the position of the surface cyclones at 00 UTC 20 Dec at the begin of the interaction of the low- with the upper-level PV anomaly.
(a) (b)
Figure 7. (a) Track and (b) time development of the SLP of the COSMO simulations ctl (black) and nolh (green). The circle in (a) marks the position of the cyclones at 00UTC 20 Dec.
Surprisingly, the new generated cyclone in nolh underwent a more intense pressure deepening of 37hPa/24h and reached a stronger minimum SLP of 968.5hPa than the cyclone in ctl that was supported by the DRW as precursor (Fig. 7b). The position of the new vortex in nolh seemed to be in a more favourable position to a strong intensification than the DRW, or the unfavourable position of the DRW to the upper-level trough prevented the maximum pressure deepening, respectively.
5. DRW statistics 2001-2010
Ten years of ECMWF analyses for the northern hemisphere were analyzed with the aid of a DRW tracking algorithm to assess the occurrence of DRWs and their intensification.
In the northern Atlantic region 431 DRWs occurred and almost as twice as many in the Pacific. More frequent are DRWs in the summer months (Fig. 8), but explosive pressure deepening of the DRWs was more frequent in the winter half-year. In the summer months no or small intensifications occurred. It means that 15% of the Atlantic and 20% of the Pacific DRWs got to the interaction with an upper-level trough for an explosive deepening into a strong extratropical cyclone.
(a) North Atlantic (b) North Pacific
Figure 8. Frequency distribution of the monthly DRW occurrence 2001-2010. The red color marks the amount of DRWs that underwent an explosive intensification.
Figure 9 shows the portion of 'Bombs'1)from DRWs among all cyclone 'Bombs' from the ERA-Interim data set. In the investigated 10 years 39 'Bombs' occurred in the Atlantic and 51 in the Pacific in total. Explosive intensifying DRWs constituted between 5 and 100% of all cyclone 'Bombs' of the month. Even in the summer months where few 'Bombs' occur in general, the DRW contribution to the total number is noticeable. With this result one can expect 5 to 12 'Bombs' from DRWs per year.
(a) North Atlantic (b) North Pacific
Figure 9. Amount of 'Bombs' from DRWs (red) among all cyclone 'Bombs' per month.
1) explosive pressure deepening is fulfilled with at least 1 Bergeron in 24h; p[in Bergeron]=p[in hPa] · sin(60)/sin(latitude) per 24h, these cyclones are also called 'Bombs'