Downstream impact of Extratropical Transition: idealised modelling

  • Contact: M. Riemer
  • Project Group: Modelling and hazard analysis of weather systems


1. Introduction and motivation

The influence a tropical cyclone (TC) can exert on the midlatitude weather patterns is still widely underestimated. A poleward moving TC usually decays as it approaches the midlatitudes and experiences structural changes, developing the features of an extratropical cyclone. This process is known as extratropical transition (ET). About half of the tropical cyclones in the North Atlantic undergo ET. The ET-system can then rapidly develop into an intense extratropical storm. Even before transforming into an extratropical system a TC can alter the midlatitude flow patterns by exciting a Rossby wave train (RWT) on the potential vorticity (PV) gradient associated with the midlatitude jet stream that (the jet stream?) disperses downstream. Through this interaction an ET-event in the western North Atlantic can initiate rapid cyclogenesis in Western Europe and enhance strong precipitation events in the Mediterranean. Case studies suggest that particularly in the latter case the moisture transport by the ET-system from the tropics in the midlatitudes plays an important role. Another important open question is, whether there are midlatitude flow patterns that are virtually stable or conducive for pronounced modification when interacting with a TC.
Numerical weather prediction systems show considerable difficulties in representing the development in the downstream region of an ET-event. Current investigations using ensemble prediction systems (Link Doris) show substantial reduction in predictability associated with ET in the whole ocean basin.

2. Method

We investigate the impact of TCs on the midlatitude flow pattern using numerical experiments with idealised initial conditions. The most simplified representation of the midlatitude circulation is a straight jet. A model tropical cyclone approaches from the South, interacting with the jet. For a more realistic representation of the midlatitudes we perturb the jet initially, thereby exciting a baroclinic wave with which the TC subsequently interacts. As a reference we compare this experiment with a run in which the baroclinic wave develops in the absence of a TC. Thus we have a solid basis for an objective diagnosis and quantification of the impact of a TC on the development of a baroclinic life-cycle.
The numerical model used is the PSU/NCAR mesoscale model MM5 employing parameterisations for cloud physics, convection and boundary layer processes. A higher-resolution nest tracks the tropical cyclone continuously during the integration. The initial state and the setup of the domains are illustrated in Fig. 1.

Fig. 1: left panel: horizontal cross section of the wind speed at 500 hPa, right panel: S-N vertical cross section through the centre of the TC and the jet axis, wind speed (coloured), potential temperature (contours) und location of the dynamic tropopause (thick contour). The TC is inserted initially approx. 2000 km south of the jet axis.

3. Numerical Experiments

As the tropical cyclone approaches the jet from the south the first clear sign of an interaction between the two systems can be seen in the upper troposphere where the outflow impinges on the jet. The subsequent evolution is depicted for selected time steps in Figure 2. At 120 h into the model run a jet streak has formed due to the interaction of the outflow with the jet stream and a ridge-trough couplet is emerging. 36 h later both features are now well pronounced. Beneath the left exit region of the jet streak a surface cyclone starts to develop. In the next 36 h the surface cyclone intensifies rapidly with a pressure drop of about 20 hPa. A further upper-level ridge-trough pattern develops downstream and initiates another surface cyclone. At the end of the simulation the upper-level wave pattern has extended over most of the domain and initiated the development of 3 surface cyclones.

Fig. 2:potential temperature on the dynamic tropopause (PVU=2, coloured), surface pressure (white contours, every 5 hPa) and wind speed on 200 hPa greater than 40 m/s (black contours, every 5 m/s) 120 h (a), 156 h (b), 192 h (c), und 240 h (d) into the integration of the straight jet experiment. Tick marks are labelled by number of grid points (horizontal resolution of 60 km).


Fig. 3: Hovmoeller of plot of meridional wind speed at 200 hPa of the experiment depicted in Fig. 1 (meridional average between grid point 60 and 120). Bold arrow depicts the translation of the ET-system, dashed arrow shows propagation of RWT. The scale of the x-axis is in grid points.

The development at upper levels can be seen as the excitation of a Rossby wave train (RWT) by the ET event. Its propagation can be conveniently depicted in a Hovmoeller plot of the 200 hPa meridional wind speed (Figure 3). The RWT can be identified after day 4. The ET system itself moves only slowly to the east while the Rossby wave energy propagates with a speed of approx. 1 500 km/day to the east.

Fig. 4: as Fig. 2, but for an experiment with an initial perturbation of the jet at near the tropopause (36 h (a) and 96 h (b)).


In another experiment the TC interacts with a developing ridge-trough pattern (Fig. 4). As the outflow impinges on the jet a pronounced jet streak forms on the eastern side of the ridge and the development of the downstream trough and the associated surface cyclone are significantly amplified compared to the run without the TC. In this case the impact of the ET-system disperses downstream as seen in an enhanced RWT (Fig. 5).

Fig. 5: as Fig. 3, but for the experiment depicted in Fig. 4, without (a) and with (b) a tropical cyclone.


4. PV-diagnosis

We use piecewise inversion of potential vorticity (PV) to investigate the importance of several physical processes for the excitation of the RWT. This diagnostic yields the wind- and temperature fields associated with the ET-system, the upper-level wave pattern and features of the lower and middle troposphere. We then diagnose the contribution of these wind fields to the excitation of the RWT. Fig. 6 shows advection of potential vorticity on the dynamic tropopause. The flow field attributable to the ET-system causes warm advection in the crest of the ridge and the base of the trough. This pattern of advection leads to an amplification of the wave. The wind field associated with the ridge-trough pattern leads to advection in the sides of the wave pattern only, contributing solely to the propagation of the wave. The feedback of the temperature and PV-anomalies of the lower troposphere are negligible.

By understanding the physical processes that lead to the excitation of the RWT and determine its propagation we expect to contribute to improving the predictability of the severe weather events alluded to in the introduction. In this respect this work is incorporated in ‚THORPEX, A World Weather Research Programme’. Link to THORPEX page

Fig. 6: advection of potential temperature (coloured) on the dynamic tropopause (PVU=2) associated with the wind field of the ET-system (left) as well as with the wave-like PV-anomaly in the upper troposphere (right). Contour lines denote the distribution of potential temperature and wind arrows associated with the respective PV-anomaly at the tropopause.The scale of the wind arrows differ from panel to panel. The scale of the axis is in grid points and the unit of advection of potential temperature is 10-5 K/s.