Page tree
Skip to end of metadata
Go to start of metadata

Extratropical Cyclone Diagrams

Overview

A comprehensive set of post-processed ENS products use a feature-based approach to represent objectively the location and behaviour of near-surface, synoptic-scale features typically associated with adverse weather (eg fronts, frontal waves, cyclonic features).  Co-location masking, a feature-type hierarchy and a minimum separation threshold, are all used together to help keep all cyclonic features 300km or more apart.  Mean sea level pressure, estimated from 1000hPa geopotential height and temperature, is shown as a reference point on many charts.  A tracking algorithm is used to follow the cyclonic features as they evolve in each ensemble member.  As a severe weather event approaches, the products can:

  • indicate an increasing risk of a major storm in the area of interest,
  • highlight the track that the storm is likely to take,
  • suggest the degree of confidence that can be placed in that track (see Fig8.1.9.8).

 

Fig8.1.9.1: To view extratropical cyclone forecasts:

  1. On charts page, click on extratropical cyclones.
  2. On extratropical cyclones page select extratropical cyclones.
  3. Select nominal data time of interest.
  4. Select feature of interest - in this case a frontal wave (orange).
  5. Display of product (in this case ENS member tracks, 1km wind speeds, 300hPa wind speeds, central pressures and vorticity).
  6. Step to next or previous nominal data time of interest.
  7. Select alternative presentation of forecast features (e.g.fronts, dalmation charts etc) from drop-down menu.

Fig8.1.9.2: Product options from drop-down menu.

Example Charts

The extratropical cyclone diagrams provide a comprehensive display of the variation between the forecasts of each member of the ENS regarding positions of fronts, depth of depressions, and strength of winds at 1km altitude. In the examples the features near Brittany relate to an extreme windstorm which in terms of European losses was the major windstorm of thr 2016-17 winter.  For interpretation of the charts see a guide to using cyclone database products.


Fig8.1.9.3: An example of a chart showing positions of fronts diagnosed from ENS members and HRES (see legend below chart for details) illustrating the variation in positions.  Nominal data time of forecast: 00UTC 03 March 2017.

 

Fig8.1.9.4: An example of a "Dalmation Plot" showing the centres of cyclonic features coloured to show an analysis of the cyclone class as derived from ENS members and HRES (see legend below chart for details) showing the variation in forecast positions. 
Nominal data time of forecast: 00UTC 03 March 2017.  Note that not all the spots denote genuine low pressure centres; it is only the barotropic lows (black spots) that are guaranteed to be.


Fig8.1.9.5: An example of a "Dalmation Plot" showing the centres of cyclonic features, coloured to show an analysis of the forecast maximum wind strength, at 1km altitude, within 300km of each centre derived from ENS members and HRES (see legend below chart for details).  Chart highlights show the variation in positions and intensity.  Nominal data time of forecast: 00UTC 03 March 2017.  Note several members suggest a maximum wind of 65-85kn in the vicinity of northwest France.

Fig8.1.9.6:   An example of a chart showing the percentage of ENS members predicting a cyclonic feature point will track within 300km in a 24-hour period T+72 to T+96 (i.e. 00UTC 06 March to 00UTC 07 March 2017).  Nominal data time of forecast: 00UTC 03 March 2017.  Only cyclonic features with a maximum wind speed exceeding 60kn at 1km altitude within 300km of the centre at some point in the 24h period are included.  A probability greater than 60% (darker orange) is shown over the western English Channel and NW France.  For a cyclonic feature moving west-to-east in this part of the world the strongest winds will ordinarily be found to the south of the low track.  This needs to be taken into account - indeed it is important for the user to not misinterpret the shading on these strike probability charts as being like a simple wind gust probability chart.


Fig8.1.9.7: An example of the clickable T+0 ENS CNTL chart nominal data time 12UTC 04 March 2017.  Automatically diagnosed positions of fronts are indicated by lines and important synoptic scale cyclonic features are shown by filled circles.  Black: Barotropic low, Orange: Frontal Wave, Green: Diminutive frontal wave.

Clicking on the frontal wave depression (orange) at the point shown by the red arrow in Fig8.1.9.7 will display the subsequent forecast movement and development as shown in Fig8.1.9.8 (albeit that plumes for some parameters have been omitted on this figure).


Fig8.1.9.8A(top): Forecast tracks of frontal wave (arrowed in Fig8.1.9.7) from ENS members.  

Fig8.1.9.8B(left): Forecast central pressure (hPa) of each depression developing from the frontal wave as identified by the ENS members.

Fig8.1.9.8C(right): Forecast wind strengths (kn) at 1km altitude within 300km of each depression developing from the frontal wave as identified by the ENS members.

The ENS CNTL and HRES are shown respectively by thin and thick green lines (though in some locations they overlap).  Most ENS members forecast the track of the selected cyclonic feature to curve towards Britain before moving SE into northwest France.   Almost all ENS members deepen the low, some to below 995hPa with winds at 1km altitude reaching more than 60kn and a few greater than 70kn.  The threat of severe weather is clearly shown but it is necessary to inspect the ENS members, meteograms, EFI charts etc. to identify the associated risk.  Nominal data time of forecast: 12UTC 04 March 2017.

 Fig8.1.9.9A(top): Ensemble EFI and SOT charts for mean daily 10m wind speed (left) and M-climate for this (right) at 99th quantile (typically 1 in 100 occasions in the ENS realises more than the values shown).  Nominal data time of forecast: 00UTC 03 March 2017.

 Fig8.1.9.9B(bottom): Ensemble EFI and SOT charts for maximum 10m wind gusts (left) and M-climate for this (right) at 99th quantile (typically 1 in 100 occasions in the ENS realises more than the values shown).  Nominal data time of forecast: 00UTC 03 March 2017.

 EFI exceeding 0.7 in much of France suggests unusual winds, and 0.8 in some places suggesting very unusual winds for those locations.  Meanwhile areas of non-zero SOT suggest a genuinely extreme event is possible. 

Considerations when dealing with small cyclones

The spatial resolution utilised in generating the extra-tropical cyclone charts (~50km) is rather larger than the resolution intrinsic to HRES or ENS, and is primarily applicable to monitoring mid-latitude depressions and their associated features. This is partly by design, in that we are trying to capture "synoptic scale features", and not every minor nuance in the model fields. Mid latitude depressions typically have a length scale of order 1000km and the program can extrapolate realistic central pressures from the surface pressure pattern.  As the feature is followed through the forecast period, feature specific plumes of central pressure, upper and lower altitude winds, and vorticity are plotted.  These are presented in the subsidiary diagrams.  However, resolution of the input data means a reduced capacity to correctly represent certain aspects (such as depth or maximum 1km winds) of small, deep vigorous cyclones, where the length scale is < ~200 km, say.  Nevertheless, the forecast positions of small vigorous centres are normally well captured.



Fig8.1.9.10: Extra-tropical cyclone chart for a small depression in the Mediterranean Sea.  Forecast data time 12UTC 27 Sep 2018.  The initial position of the low centre near Benghazi is identified by almost all ENS members, but just one or two and HRES locate the centre a little further east (identification of these may be "corrupted" because part of the depression in model fields was over the high ground in northeast Libya).  The forecasts generally fall into two groups: the larger group tracks the cyclone initially northwest and then more persistently northeast, the other consisting of HRES (thick green line) and a few  ENS members track a depression towards Cyprus.  The plumes show the very different evolution of these two groups. Closer inspection of HRES fields shows a particularly complex surface pressure evolution, whereby the main feature actually develops west of the tracked low, thereby showing more consistency with the main plume. This illustrates an additional difficulty associated with developing small-scale cyclones - as far as this HRES run was concerned the main feature wasn't actually present at analysis time.

  

Fig8.1.9.11: Extra-tropical cyclone chart considering a small depression in the Meditteranean Sea with HRES MSLP charts. Forecast data time 00UTC 28 Sep 2018.  The initial position of the low centre is identified by ENS members and HRES and the subsequent forecast tracks and plumes are fairly similar. The central pressure of the depression is analysed by HRES as 991hPa and by T+48 HRES and ENS forecasts show the centre to have a similar depth.  However, the plume indicates pressures no lower than 1000hPa which is a consequence of the lower resolution. While confidence may be placed in extra-tropical cyclone charts associated with larger mid-latitude depressions, users should not take as precise the values given for small vigorous features although the trend of vorticity and other parameters will give a fair indication of trends.

Considerations regarding Identification of Fronts

Level at which fronts are identified.

Fronts are identified using a vertically-interpolated level that is everywhere 1km above the model orography.  This terrain-following approach, and the choice of 1km, have many advantages:

In tests it has been found that:

  • choosing higher levels (e.g. 850hPa, 800hPa, or equally 1.5km, 2.0km) there is increased likelihood that the identified front is displaced from the surface discontinuity due to frontal slope.  It is not possible to systematically account for any displacement from the surface discontinuity since frontal slopes vary, even perhaps with the possibility of overrunning.
  • choosing lower levels (e.g. 950hPa, 1000hPa, or equally 0.5km, 0.0km), the impact of the surface becomes ever more significant and the connection with any deeper tropospheric features is lost (e.g. coastlines become semi-permanent fronts, which is of course undesirable).

Choosing the level 1km above orography offers the best of both worlds - it is close enough to the surface to be mostly co-located with surface frontal signatures (e.g. a frontal pressure trough) but far enough away from the surface to be representative of the lower troposphere while not being over-influenced by discontinuities in the orography.  In general the 1km level is lower than the 850hPa level and actually represents the real model airmass over mountains and not a less meaningful underground extrapolation (see Fig8.1.9.12).  

Identification of type of front

A thermal variable (wet bulb potential temperature, θw) is used in order to incorporate a moisture component.  It is quite common practice to use a similar variable (equivalent potential temperature, θe) around Europe.  In tests it was found that using a pure thermal variable like temperature regularly generated spurious dry fronts downwind of topographic barriers (from the Foehn effect).

Fronts are located on the warm sides of bands of stronger gradients in the wet bulb potential field situated at a level 1km above the model orography. They are classified as cold or warm fronts by the sign of the geostrophic advection of the wet bulb potential temperature field at 1km. Thus:

  • a cold front is identified where the geostrophic wind blows from cold θtowards warm θw.
  • a warm front is identified where the geostrophic wind blows from warm θtowards cold θw.

The use of geostrophic wind, which relates directly to isobaric crossing, accords with practice in much of Europe.  Tests showed that if we had used the full wind instead, there would have been occasions where the type identified disagreed with the type identified using the geostrophic wind.

Quasi-stationary fronts are not indicated on Extratropical Cyclone Diagrams.  Applying thresholds to the magnitude of the cross-front geostrophic wind - ensuring that that is small - could in principal facilitate this. However, it was deemed more ihelpful for users to find and show incipient frontal waves (as indicated by coloured dots) and follow their subsequent track and development, and for that we need the meeting points of cold and warm fronts.

Of course within the above generally robust framework there are inevitably going to be occasions when the 1km level doesn't work so well (e.g. for very shallow sloping fronts), or when θmight give a somewhat different type of front or front position to that given by temperature on the same level (due to odd humidity structures, or when the geostrophic advection might give a different sign to full advection). Proximity of shallow meso-scale troughs or lows (e.g. heat lows) can also induce differential flow across an identified front, but in the medium range such detail may not be relied upon. These cases appear to be rare, but it is possible that some geographic regions will experience such issues more than others because of the effects of peculiarities in one or more of the following:

  • complex topography 
  • complexities and even uncertainties in land/sea/lake/ice boundaries
  • atypical humidity structures (for example the warm airmass may be dry and the cold airmass moist)

Correspondence with surface isobaric troughs

Where frontal activity is weak there will not necessarily be an associated surface isobaric trough. Around anticyclones, relatively inert cold fronts can give a change in weather type (e.g. from cloudy and mild to clear and cold) whilst exhibiting a very weak isobaric trough or indeed no pressure trough at all.  However, active fronts normally are associated with marked surface pressure troughs on charts.

Users should inspect other available IFS output to understand the structure of the atmosphere as forecast, and in particular should not rely on one model solution alone (e.g. HRES or ENS control) but instead assess uncertainty using the ensemble of frontal positions ("spaghetti fronts") and other products (e.g. meteograms). Users can then the interpret (and potentially adjust) IFS objective fronts in light of their knowledge and experience of the local geography and its influence.


Fig8.1.9.12: Selection of altitude used in the identification of fronts.


Additional Sources of Information

(Note: In older material there may be references to issues that have subsequently been addressed)

Read the guide to using cyclone database products for interpretation of the charts.

Watch a comprehensive lecture on extratropical cyclone diagrams (10sec delay before start).


Updated 29/08/19 - Identification of Fronts


  • No labels