An historic cold spell

Between the 26th February and 2nd March 2018 a significant cold spell affected much of NW Europe, and in particular the UK & Ireland with large lowland areas affected by disruptive snowfall at some point - likely the most widespread and heavy snowfall so late in the season at low level since 2001 across the north and 1986 in the southeast - and indeed at the height of the cold spell (1st March) a maxima of −4.7C set a new record at Tredegar, Blaenau Gwent, Wales (previously −4.6C at Cassley, Sutherland on 2 March 2001) as the coldest ever March day in the UK.

This cold spell was well forecast in advance with good agreement and consistency from all of the available model output even as far as 7 days in advance - this led to the team at Wattisham being able to pre-warn the airfield manager that snow was likely to disrupt operations on the Wednesday previous to this event occurring, with the now infamous moment that Harry told Andy (airfield manager) that snow was coming to which he immediately then marched into flight planning and declared loudly "Harry says it's going to snow next week, so make sure you get whatever you need to get done this week because I'm not clearing the snow!".

Whilst further south and west the most significant impacts came from Storm Emma, this blog aims to summarise the most remarkable part of the cold spell further east across East Anglia and Lincolnshire, which occurred the day before on 28th February 2018. The ASXX for 12 Z on the day (see Figure 1, below) clearly shows the main situation with a large intense anticyclone of 1052 hPa centred over northern Sweden, with an associated strong, severely cold and deeply unstable easterly flow established across NW Europe. As shown in Figure xx; also below, by the dashed brown line, the thickness (depth of atmosphere between 1000 and 500 hPa) is unusually low at sub-510 Decametres (DM) over all of the UK and Ireland. The radiosonde sounding from Herstmonceux (see Figure 2) from 00 Z gave a thickness of 508.4 DM and the Albemarle radiosonde ascent from 12 Z (not shown here) recorded the lowest thickness in this spell of 502.2 DM - rivalling the UK record low of 497.4 DM recorded at Hemsby, Norfolk on 12th January 1987.

Lake-effect snowfall

This well documented process is perhaps best recognised and most frequently observed across North America, downwind of the extremely cold Arctic outbreaks that push south-eastwards across the 'Great Lakes' affecting mainly Michigan, northwesternmost Pennsylvania and western New York. The effect is well marked and in the most severe and prolonged outbreaks can deliver in excess of several metres of snowfall in a relatively short period - whilst at the same time on the western shore of the same lake almost no snowfall occurs - this manifests itself even in the climatology with as much as 3 to 4 times more snow falling annually in favoured downwind locations.

It occurs elsewhere in the world but works most efficiently where there is a large and relatively warm body of water closely situated to areas that are likely to be affected by deep and severely cold Arctic airmasses; other such regions include parts of northern Europe bordering the Baltic Sea, occasionally central and southern Europe bordering the Black Sea and even the areas downwind of cold outbreaks over the Sea of Japan. In very rare circumstances, such as with this historic cold event, it can even occur over the North Sea and bring heavy and disruptive snowfalls to eastern counties of England and Scotland, most notably in the aforementioned January 1987 event when snowfalls were widely 30cm or greater with deep drifts effectively cutting off remote communities for several days.

Deep instability

Whilst initially dry and very cold with therefore steep lapse rates, an Arctic or Polar Continental airmass over central Europe will be almost entirely cloud free when exiting the coast of the Netherlands, Belgium and north-western Germany, but can very rapidly pick-up and accumulate moisture and blossom with deep convection as it crosses the North Sea, heated intensely from the relatively warm surface. This process is akin to the 'lake-effect snows' of North America and as shown in Figure 3; below right, which is a snapshot of conditions over the east of England at 09 Z on 28th February 2018, demonstrates well this process in action.

The steering flow at this time becoming just south of due easterly (100-110 deg), but even in spite of a relatively short sea track of just 80 to 100 nautical miles the dry and clear air over the Netherlands and Belgium with dew points at the surface of minus 11 to minus 13 degrees Celsius becomes deeply unstable and convects to give glaciated CBs within 30 to 50 nautical miles of the east coast.

It is worth mentioning that given the depth of the cold air the tops of these CBs were only around 670 hPa (~10,500 feet) where temperatures were around minus 28 degrees Celsius. As can be seen from a rough parcel modification made on the 00 Z Herstmonceux ascent (Figure 2; right) a surface temperature of minus 1 would be enough to support relatively large levels of CAPE (Convective available potential energy) for wintertime convection and release deep and moist convection supportive in excess of CB level; though it is also worth highlighting that in reality the temperature was slightly lower than this over the southern North Sea - even though the sea surface temperature was around 6 degrees Celsius!

As we know, the Bergeron–Findeisen process reaches a maximum level of efficiency at around minus 12 degrees Celsius (balancing between the level of supersaturation with respect to ice and the available amounts of water in the atmosphere), and therefore with much of the convective cloud close to, or around this level, this allowed maximum efficiency in snow crystal growth which can occur within 10 to 20 minutes and therefore well within the time taken to cross the southern North Sea between the Netherlands and East Anglia, even with the (relatively) short sea track. So what started out as shallow Cu just off the coast quickly towers into towering cumulus and then CB all within a relatively short distance and constantly re-developing in the same location before being brought towards the coast.

Snowfall on 28th February

Whilst the previous two days had brought snow showers across parts of eastern and south-eastern England, the heaviest and most disruptive falls were relatively localised to (but not exclusively) parts of northern Kent, southeast Essex, eastern Lincolnshire, Norfolk and Suffolk with large variations across small areas, with shower distributions generally dependent on a subtle balance of the surface flow and coastal geography, where frictionally-backed convergence helped to maintain and enhance snow showers downstream.

On the 28th the deepest and coldest air tracked westwards across the UK, and as discussed earlier the lowest thickness was measured at Albemarle in Northumberland. By inference, and the greater sea track, this was where the deepest instability was predicted to be, and indeed near to Newcastle during the morning there were several SFERICs observed by the Met Office ATD network with thundersnow reported from Newcastle airport at 12 Z and 15 Z - this was highlighted and predicted well by Jonathan Wilkinson's recently developed lightning diagnostic - available from the GPP output suite (see Figure 4).

The depth and intensity of the cold air meant that the type of snow observed was different to that typically seen in the UK, and this had some important consequences with regard to the interpretation of model output (as discussed well in the guidance produced by the Guidance Unit at Exeter). Snowflakes can contain varying amounts of water, with the 'stickier' and wetter snow being the most commonly observed in the UK and generally formed at temperatures close to, or slightly below zero degrees Celsius. This snow can itself be quite disruptive, especially when heavy falls are accompanied by a strong enough flow (typically 15 to 20 knots) allowing it to aggregate and accumulate on power lines, and then potentially down both power lines and power poles to cause widespread disruption to the electricity grid, however these snowfalls typically give snow depth-water equivalent ratios of between 5:1 and 8:1. During this cold spell because mean temperatures within the lowest few thousand feet were at or below minus 10 degrees Celsius, the snow was drier and of a more typically powder snow consistency. Therefore in this instance snowflakes are commonly the "classic" dendritic type, do not aggregate together and give a powdery dry snow that has higher snow depth-water equivalent ratios of between 12:1 and 20:1 (generally the colder and drier the source region, the higher the ratio).

All of this meant that if interpreting the UKV forecast snow depths literally, which use a ratio much nearer to that of the typically observed wetter snow in the UK (e.g. 10:1), this would give an underestimate of between 1.5 and 2 times. This can be seen when comparing Figure 5; right (UKV forecast snow depths on 28 February 2018) with Figure 6; below (observed snow depths over East Anglia on 28 February 2018) with for example 5-10 cm lying snow forecast at Wattisham by 15 Z and 16 cm observed at 12 Z.

During the early hours of 28th February 2018 snow showers became heavy and widespread along the coast of Norfolk, Suffolk and Essex and then proceeded to move inland under the influence of the veering steering flow, allowing snow showers to push in across East Anglia and make significant progress inland despite temperatures here being widely minus 4 to minus 6 degrees Celsius and unable to of itself sustain or support convection. Initially at Wattisham there were just snow remains left from the previous day, but by 06 Z already 7 cm of snow had fallen with a further 8 cm then falling in just the next 4 hours to 10 Z; the heaviest fall being between 08 and 09 Z with 3 cm fresh snow in the hour and almost continuous moderate to heavy snow showers feeding in constantly from the North Sea with very little (if any) break between showers.

As you can see from Figure 6 (with thanks to Dan Holley) snow depths continued to vary widely across the region even across relatively small distances, however in general across the east of the region depths ranged from around 10 to 15 cm quite widely with locally up to 20 cm in places with more generally the larger totals tending to be along the coast from north Norfolk all the way down in to Essex, with 14 cm at Clacton-on-Sea.

This heavy and almost relentless onslaught of snow showers was replicated across most of East Anglia through much of the morning bringing significant disruption to the road and rail network, and then even as instability became limited by less cold air moving west by the afternoon, the strengthening east to north-easterly surface flow quickly helped to pick up, blow and drift the powder snow around giving some atrocious visibility at times. Even when there were no snow showers in the vicinity the wind picking up was enough to pick up and blow enough snow to reduce the visibility to below fog limits and occasionally down to 200 M (see Figure 7, right; for the observations from Wattisham airfield).

Due to the coldest air arriving around the time of maximum heating, there were some unusually low maxima recorded during the day on the 28th February, below are listed some of the lowest across East Anglia - all of which were the coldest maxima recorded in February since 1991 (see above in Table 1), although a number of new records were set - e.g. at Marham.

Summary

This was a historic cold spell for the UK and Ireland as a whole, with few precedents in terms of its severity within recent history especially for the lateness in the season in which it occurred. For East Anglia it will primarily be well remembered for the disruptive snowfalls that occurred on the morning of 28th February 2018 which widely brought 10-15 cm across eastern parts of Norfolk and Suffolk in what was already a memorably cold and snowy spell for some localities within a few hours. In terms of the forecast and warnings given for this spell, the Sudden Stratospheric Warming was well predicted far in advance at the start of February, and once underway saw the longer range forecasts that the Met Office puts out from mid-February correctly predicting a significantly cold and snowy end to the winter and start of spring.