A Winter Storm

Since Wednesday 6 July, stormy westerly conditions have affected New Zealand. In this blog, we’ll look at why.

The “Long Waves”

Below is the mean sea level analysis – the weather map – for 6am Sunday 10 July. In between big highs over the mid South Pacific and south of western Australia is a really large trough; it’s the area shaded light blue. The weather map has looked like this, more or less, since Wednesday 6 July: that is, the big features on it aren’t moving much.

MetService mean sea level analysis for 6am Sunday 10 July 2011.

There’s good reasons why these big features aren’t moving much. They reflect the so-called “long waves” in the troposphere (the troposphere is that part of the Earth’s atmosphere in which the weather occurs), which are stationary at the moment. Below is an image made from a Fourier analysis of the wave pattern in the Southern Hemisphere at midnight Saturday 09 July. There’s a trough (blue) in this wave pattern more or less in the same place as the one shaded light blue on the weather map above. And either side of this trough, there are ridges (pink) in about the same place as the big highs on the weather map above. For more about this, see my blog post on Wave Three.

The "long" and "medium" waves about half-way up the troposphere at midnight Saturday 09 July 2011.

Polar Air

Below is a plot of where the air arriving on New Zealand’s west coast at midnight Saturday 09 July came from. The air is from the Antarctic.

"Backward trajectories" of the air at mean sea level arriving at Cape Reinga, Farewell Spit and Secretary Island at midnight Saturday 09 July 2011. Each red line traces the path of the air over the period of a week. That is, the air arriving on New Zealand's west coast at midnight Saturday 09 July left the Antarctic about a week previously. Data courtesy NOAA Air Resources Laboratory.

On its way to New Zealand, this air travelled over a long stretch of ocean. It will have been colder than the surface of the sea it passed over, and will therefore have taken up heat from the sea surface. As heat transfers from the sea surface to the air immediately above it, “blobs” of air become warmer than their surroundings and rise upwards. This process is known as convection (see Chris Webster’s blog post about predictability and popcorn). If convection continues for long enough, showers and/or thunderstorms are the result. In the satellite picture below, more or less all of the clouds over the Tasman Sea, New Zealand and the seas to the south of the country are “blobs” of air which is rising (or has recently risen) convectively.

MTSAT-1R infra-red satellite image for midnight Saturday 09 July 2011. Image courtesy Japan Meteorological Agency.

The Jet

A major flood of air out of the Antarctic region, like this, has other consequences. The northern boundary of the cold air pushes against the warmer air further north. Thus, the north-south temperature contrast increases and simultaneously the strength of the westerly winds increases – not just at the Earth’s surface, but throughout the depth of the troposphere. (How this works might be the subject of a future blog post). This is at the heart of why many places have been windy over the last few days.

The axis of the strongest winds, in the mid- to upper troposphere, is known as the jet stream. In the plot below (for midnight Saturday 09 July), the polar jet has two branches (arrows in black): one curving across the south Tasman Sea and over the South Island, and the other crossing the south of the North Island. Over central New Zealand, wind speeds above 10,000ft were generally 80 kt (150 km/hr) or more.

Wind speed at 500 hPa (approximately 18,000 ft) in the New Zealand region at midnight Saturday 09 July 2011. Warmer colours are stronger winds. Data courtesy European Centre for Medium-range Weather Forecasting.

Thunderstorms and Tornadoes

Air ascending into convective clouds (showers and thunderstorms) comes down again. When a convective cloud collapses, the resulting downdraft hits the Earth’s surface and spreads out, much like water does when tipped from a bucket onto the ground. When the showers and thunderstorms are themselves fast-moving (on the afternoon of Saturday 09 July, storm motions on the Kapiti Coast were 70 to 90 km/hr), the winds near the Earth’s surface can become very strong when downdrafts occur: this is very likely to be the cause of some of the wind damage on the Kapiti Coast on the afternoon of Saturday 09 July. And this is one of the reasons why the various Severe Thunderstorm Outlooks, Watches and Warnings issued over the last few days have included the mention of damaging wind gusts.

Imagery (see below) from the Wellington radar for 4:00pm Saturday 09 July shows a line of thunderstorms extending across Cook Strait onto the Kapiti Coast. The tornado is very likely to have been associated with the strong thunderstorm shown just east of Waikanae at 4:00pm; eye witness reports suggest that the tornado crossed State Highway 1 around 3:55pm. The tornado’s genesis remains unclear: it may be that a low-level vortex was “spun off” the northern end of Kapiti Island just at the time that this thunderstorm passed by, and the strong ascending motion in the thunderstorm developed it into a tornado between there and landfall on the Kapiti Coast. Once again, it is remarkable that there was no loss of life.

Reflectivity image from the Wellington radar, 4pm Saturday 09 July 2011. Colours represent how strongly precipitation bounces the radar signal back to the radar.


The Southern Alps are a significant barrier: in westerly airstreams, generally only a moderate amount of precipitation falls any distance east of the Divide. How much precipitation falls east of the Divide, and how far east of the Divide it reaches, depends on a number of factors. Not the least of these is the Foehn Effect. Overnight Saturday 09 July and on the morning of Sunday 10 July, a reasonable amount of snow fell east of the Divide, in parts of Otago, to below 500 metres, in a northwesterly airstream. This is an uncommon occurrence, and reflects how deeply cold and showery the air passing across Otago was at the time.

Large Sea Waves

Since about the middle of the first week of July, sea waves arriving on New Zealand’s western coasts have been notably large. In some places, they have probably attained heights observed only once every year or two – and will remain high until late in the week ending Fri-15-Jul. The weather map below, from about the time large waves began arriving, shows why:

  • The fetch – that is, the expanse of ocean over which waves arriving on New Zealand’s western coasts have been generated (pink arrow) – is very long
  • The waves are still growing as they reach New Zealand’s western coasts, because the winds across New Zealand are themselves strong.

MetService mean sea level analysis for midnight Thursday 07 July 2011. The red arrow shows the approximate path sea waves arriving on New Zealand's western coasts have travelled.

New Plymouth tornado, Sunday 19 June 2011

The tornado in New Plymouth early on the morning of Sunday 19 June 2011, like every tornado that passes through an area where people observe it, was certainly dramatic.

Like many tornadoes that affect Taranaki, this tornado formed out to sea – not very far to the northwest of New Plymouth – in a line of thundery showers.

Radar imagery

At 4:15am, a mesocyclone (to be explained later in this blog) is identifiable over New Plymouth in imagery from the New Plymouth radar. Radar imagery 7.5 minutes earlier, at around 4:07am, hints vaguely at the mesocyclone’s existence but is far from conclusive. Radar imagery at around 4:22am suggests the mesocyclone either has decayed or is rapidly decaying. In other words, conditions supporting the development of a tornado in the New Plymouth area were favourable only for a very short time.

Reflectivity image from the New Plymouth radar for 4:16am Sunday 19 June 2011. Colours represent how strongly precipitation bounces the radar signal back to the radar.

Colour key for the above image. The further to the right along the scale, the more strongly precipitation reflects the radar signal.

Severe Weather Forecaster John Crouch has extracted radar data from the bottom few “sweeps” of the radar beam. The mesocyclone was sufficiently close to the New Plymouth radar for these sweeps to provide a reaonable view of it. Here they are below, animated. Each time step is 33 seconds; the first sweep is at about 100 metres above the ground near downtown New Plymouth, while the last is about 1000 metres above the ground. So, we’re looking at successively higher layers of the mesocyclone as time goes forward; the reflectivity pattern can be seen wrapping around the circulation associated with the mesocyclone.

Radar reflectivity loop covering the (just over) three-minute period from 4:15am to 4:18am Sunday 19 June 2011. A hook shape can be seen moving from the northwest across New Plymouth.

It’s important to note that this doesn’t mean it’s possible to produce 33-second radar imagery routinely, or in “operational time”. It’s only when phenomena very close to the radar are moving and evolving sufficiently rapidly that data from subsequent sweeps of the radar beam might be coherent enough for animations of them to be meaningful.

Life cycle

Note, in the above animation, how quickly the hook-shaped pattern moves and changes. It suggests strongly that the whole tornado event was over in a couple of minutes, which appears to be consistent with reports in the media. In the animation, there’s evidence of only one hook-shaped pattern in the New Plymouth area: it may be that there was only one tornado, and it wasn’t always reaching down to the surface along its path.

Warm seas to the northwest of New Plymouth provided some of the “fuel” for the thundery showers which passed across Taranaki in the early hours of the morning of Sunday 19 June 2011. This is very likely why the tornado was short-lived: when the mesocyclone came ashore, its fuel supply was cut off.

A little bit about mesocyclones

In brief, a mesocyclone is a local rotation and ascent of air about a vertical axis.

Hook-shaped patterns in radar reflectivity imagery are not uncommon – but are nothing like conclusive evidence of the presence of tornadoes. There also needs to be strong, coincident, rotation: this is one of the reasons why weather radars measure the inbound / outbound speed of the echo using the Doppler Effect.

In the case of the tornado in New Plymouth early on the morning of Sunday 19 June 2011, there was a strong velocity couplet observable at 4:15am, but not either side of this time.

Further reading

For more on tornadoes in New Zealand, see the blog about the Albany tornado of Tuesday 3 May 2011.

Albany Tornado, Tuesday 3 May 2011

Few weather events are as dramatic, dangerous or challenging to predict as tornadoes.

About the tornado of Tuesday 3 May 2011

On the afternoon of Tuesday 3 May 2011, a line of showers moved southwards across Northland. Ahead of this line the winds were moderate northeasterlies; behind it, they were moderate northwesterlies. Along the line, the winds converged – that is, pushed against each other. Below is a portion of a working chart for 3:00pm Tuesday 3 May 2011, drawn by one of the Severe Weather Forecasters.

Fine-scale mean sea level analysis, valid 3:00pm Tuesday 03 May 2011, drawn by a member of MetService's Severe Weather Team. At this stage, the main line of convergence is just passing through the Auckland area but there are still areas of activity behind it.

Anyway, when air gets piled together like this it prefers to go somewhere if it can, and in this case it went upwards. More about why this is important in a minute.

MetService’s Severe Weather Forecasters had been watching the line of showers since the day before when it was out over the Tasman Sea, and expected it to be “active” as it crossed northern New Zealand on Tuesday. Below is the Severe Thunderstorm Outlook published on www.metservice.com on the morning of Monday 2 May 2011. (This was re-issued with similar content on Monday evening and again on Tuesday morning).

Severe Thunderstorm Outlook for Tuesday 3 May, published on www.metservice.com at 10:21am Monday 2 May 2011

Around midday on Tuesday, Severe Weather Forecasters received an important new set of data from the Whenuapai balloon flight, which describes the profiles of wind, temperature and moisture aloft. Analysis of the data revealed that even more thunderstorm activity was likely in the afternoon over the Auckland region than had been previously thought, and the Severe Thunderstorm Outlook was re-issued with an increased risk for Auckland; see below. In other words, measurements of the properties of the air around Auckland revealed it was even more likely to go upwards than had previously been thought. Upwards air movement, if unchecked, leads to thunderstorms.

Severe Thunderstorm Outlook for the afternoon and evening of Tuesday 3 May, published on www.metservice.com at 12:38pm Tuesday 3 May 2011

However, the Whenuapai upper-air data also showed that the air around Auckland possessed little of the tornadic tendency (for the technically minded, low-level shear and low-level helicity are important factors) that it exhibited so tragically and destructively just a few hours later. That is, using high-quality facts from very close by, together with weather model data for the Auckland area, the Severe Weather Forecasters investigated the possibility of tornadoes over the Auckland area and decided there weren’t enough factors favouring this.

So why was there a tornado? The low-level convergence mentioned a few paragraphs back was a major contributor. The strong tendency of the air around Auckland to ascend was another. Together, though, these spell thunderstorms – and only a small proportion of thunderstorms have tornadoes associated with them. The third factor is how the air near the Earth’s surface in the Auckland area managed to acquire so much corkscrew-like motion (helicity) so quickly. We’ve not yet looked into this in depth, but it seems certain that the local geography in the area north of Albany played a significant part. And a correspondent who witnessed the development of this tornado, David Haysom, thinks similarly. Excerpts from David’s email to us are below.

Eyewitness account

“One of my work colleagues saw the tornado form on a block of land next to our office that is exposed clay. (Corner of Corban Ave and Munroe Lane). At that stage it was just a flurry of leaves in a circular motion.
This is about 200 metres from a large escarpment to the North.
It then moved southwards and when I saw it, it was only a couple of metres high and not much energy.
100 metres further on it was about 5 metres high.
By the time it reached the third road parallel to our office it was about 12 metres high and had turned white. (Don McKinnon Drive).
Above the mall was intensely black cloud and it headed for this growing rapidly in height and becoming very opaque which I guess was the water from the lakes.
It went across the lakes to the mall flattening trees instantly as it went or ripping them out.
It hit the mall on the corner that has the cinemas and it looked like an explosion of roofing iron.
Given the similar event in Albany 20 years ago, I think it is the Oteha Valley Rd escarpment that causes it. As the wind comes down it would get a twist as it exits the motorway gradient which is lower than the escarpment to its west.”

So … could MetService have forecast this tornado?

No. The closest we can get is being able to identify an area where there is a significant risk of small-scale severe weather, or where small-scale severe weather is already occurring. In the case of the latter, 30 to 60 minutes’ warning of small-scale weather coming your way is likely to be as good as it gets anytime soon. You can sign up to be emailed Severe Thunderstorm Watches and/or Severe Thunderstorm Warnings here.

Tornadoes like this – which is the most destructive seen in New Zealand for a few years – are dramatic but very localised: they last only a few minutes, and they affect relatively small areas. From David Haysom’s account above, the distance from the corner of Corban Avenue and Munroe Lane to the shopping centre in Albany is about 800 metres. MetService is able to forecast the conditions favourable for the formation of (severe) thunderstorms, and sometimes we can anticipate the likelihood of tornadoes that have nothing to do with severe thunderstorms. It all comes down to being able to represent phenomena at the time and space scales upon which they occur – which in this case is a few hundred metres and a few minutes respectively. For more about this, see Chris Webster’s Predictability and Popcorn blog entry.

But … I could see a “hook echo” on the weather radar?

Yes, there was a hook echo. Here it is at 2:54pm Tuesday 3 May 2011.

Radar reflectivity, 2:54pm Tuesday 2 May 2011. There is a hook-shaped radar echo near Albany.

And there was rotation around the hook in the “right” place (3-6 metres per second away from the radar; 0-3 metres per second towards – not exactly mesocyclone territory), as the Doppler velocity image for the same time shows (below).

Radial velocity, 2.54pm Tuesday 2 May 2011. Warm colours (yellow through to red) denote wind speeds away from the radar; cool colours (green through to blue) denote wind speeds towards the radar.

Are you doing anything about getting better at this?

Norm Henry, MetService’s General Manager of National Weather Services, described MetService’s advances in forecasting in a blog post on 08 April 2011. About the forecasting of small-scale severe weather, he says “An example of a long-term research programme that has delivered direct benefit to New Zealanders is the development of the Severe Thunderstorm Warning service. This work began about 7 years ago with a detailed investigation of the climatology of New Zealand thunderstorms, with emphasis on the types of convective environments that lead to localised severe weather. As a result, New Zealand-specific techniques to diagnose and forecast severe convective storms have been developed, along with appropriate guidance material based on observation data and MetService’s high-resolution models. MetService has also implemented state-of-the-art thunderstorm tracking software as part of its weather radar systems, allowing it to rapidly identify, track and forecast the behaviour of severe convective storms.”

A little bit more about tornadoes in New Zealand

Tornadoes in New Zealand are quite different from those that occur in the midwest of the United States primarily in the warm part of the year. In New Zealand, tornado occurrence is primarily related to convection along strong cold fronts – and thus they are largely a “cold season” phenomenon. Some notable recent instances include the Waitara tornado of 15 August 2004, the Greymouth tornado of 10 March 2005 (very likely a cold season event, even though it occurred in March), the New Plymouth tornadoes of 3-4 July 2007 and the Cambridge tornado of 17 October 2008. The Albany tornado of Tuesday 3 May 2011 wasn’t of this type. Of course, for those it affected this is nothing more than academic. It is tragic that there has been a fatality, a number of injuries and considerable property damage. As many have already said, it was a miracle that there was only one death.

Tropical cyclones: extra-tropical transition

On average, about nine tropical cyclones form in the South Pacific tropics between November and April. Three or four of these leave the tropics and nearly all of them undergo a marked transformation to a mid-latitude depression – a completely different weather system – before they reach New Zealand.

For a while after this extra-tropical transition, the system may be referred to as “low formerly cyclone so-and-so”. At this stage of its life, the system may still have considerable potential for severe weather, despite its name change – for example, the low that was formerly tropical cyclone Bola, in March 1988.

So, how does this extra-tropical transition take place?
When a well-developed tropical cyclone reaches its peak in the heart of the tropics, it has an eye. The eye is often fairly cloud-free, nearly circular, and surrounded by a ring of very active thunderstorms. In the early and middle parts of their lives, tropical cyclones stand up quite vertically in the atmosphere, like large columns.

Infra-red satellite image for 10pm 02 February 2011 (New Zealand Daylight Time); this image displays colour-coded cloud top temperature. Tropical cyclone Yasi is just about to make landfall on the Cassowary Coast of northern Queensland.

Besides encountering cooler seas, tropical cyclones heading towards New Zealand eventually come under the influence of the westerlies. The westerlies of the mid-latitudes increase in strength with height, a phenomenon known as vertical wind shear. This shear almost literally chops off the upper part of the tropical cyclone and sweeps it away, not unlike a woodcutter chopping off the upper part of a coconut tree to leave a section just above the ground (except it’s a much more gradual and subtle process). Along with the lower sea temperatures of the mid-latitudes, this destroys the positive feedback processes within the cyclone. What remains is the former cyclone’s low-level circulation, which may get carried off in the westerlies or become the focus of further development if conditions are right. Either way, tropical cyclones approaching the New Zealand area undergo drastic changes of structure and appearance as they undergo this extra-tropical transition.

Let’s take a look at the extra-tropical transition of Wilma, which affected New Zealand in late January.

While in the tropics, Wilma had a similar look to Yasi, but was a less intense tropical cyclone. In the above infra-red satellite image (for 2pm Wednesday 26 January 2011 New Zealand Daylight Time), we can see a well-developed eye and the ring of thunderstorms surrounding it.

24 hours later (2pm Wednesday 26 January 2011 New Zealand Daylight Time), at the margin of the tropics, in response to increasing wind shear, the eye of Wilma became obscured amongst a broader mass of colder topped clouds.

And 24 hours further on, Wilma’s cloud structure bore little resemblance to that of 26 January. The eye and eye wall are no longer identifiable.

Finally, we take a look at the visible image for the same time as the last infra-red image above. This shows the more active, colder cloud tops (bright white) are now east of Wilma’s low-level centre. Therefore, Wilma is on the verge of losing its tropical cyclone status and becoming classified as a depression. Like all other cyclones before Wilma, Wilma lost its tropical cyclone status before reaching New Zealand.

The Thunderstorm in History

One of the pleasures of reading history is coming across stories about the weather. Thunderstorms often figure in these. One of the most dramatic examples was recorded in the sixth century AD, by Gregory, Bishop of Tours, in his Historia Francorum (The History of the Franks).

In AD 536 there were three rulers of Frankish kingdoms: Childebert, the king of Paris; his brother Lothar, the king of Soissons; and the brother’s nephew Theudebert, the king of Metz. Childebert and Theudebert joined forces and set out with a large army to attack Lothar, who retreated to a fortified position on a hilltop. Hearing of the imminent battle, Queen Clothild, mother of Childebert and Lothar, went to the tomb of Saint Martin and prayed through the night for divine intervention to prevent her sons fighting.

The next morning, before battle preparations had been completed, a terrific thunderstorm laid waste to the aggressor’s camp. Tents were blown down, gear was scattered and horses driven away by hail and lightning. The hailstones were so large and pelted down with such force that many soldiers, including the two kings, were cut by them, driven to the ground and forced to shelter beneath their shields. Meanwhile, Lothar and his army were untouched by the thunderstorm. Accepting the event as divine chastisement, Childebert and Thuedebert did penance to God begging forgiveness for attacking their own kith and kin, then sued for peace and concord, which Lothar granted. Lothar’s dynasty prospered, leading eventually to the unification of France and the rule of Charlemagne.

The role of weather is also given a prominent place in The Oxford History of the French Revolution by William Doyle. Repeated drought during the 1780s caused soaring grain prices leading to repeated civil disturbances in many parts of France. Then, in July 1788, on the eve of the harvest, widespread hail storms devastated hundreds of square kilometres of crops in the Paris Basin, which was one of the most productive agricultural areas in France. Hailstones were so large they killed men and animals. The inability to gather tax revenue on the destroyed harvest bankrupted the French Government and the price of grain rose to almost 90% of a workers salary.

In order to try to gather tax from the nobility, who were largely exempt, the French King was forced to call the Estates General for the first time in over a hundred years. Once assembled the Estates General moved beyond the King’s control, passing laws he neither wanted nor anticipated. Within the year, the struggle for power escalated into violence, the Bastille was stormed, and the French Revolution was underway.

More intriguing is the story of Martin Luther and the thunderstorm. Having completed a masters degree and a visit home to his parents, Martin Luther was returning to University in Erfurt to study law when, on July 2 1505, near Stotterheim, he was caught in a thunderstorm. Thrown to the ground by a lightning bolt striking near him, he called out to St Anne, promising to become a monk if his life was spared. Two weeks later he abandoned law studies and entered a monastary, starting down a path that eventually changed European history for ever, splitting the church and triggering decades of war.

Told this way, there is a hint of myth about the story. In fact, Martin Luther seems not to have been too keen on a law career and to have been thinking about joining the church anyway, but this was bitterly opposed by his father. Perhaps the weather provided Martin Luther with an alibi. “ Sorry Dad, a thunderstorm made me do it.”

Thunderstorms in Wellington

Cumulonimbus (CB, thunderstorm) cloud is not a common part of the Wellington cloudscape, and certainly is not in response to diurnal heating as happens in other regions like Waikato and Bay of Plenty. When Wellington gets thunder and lightning it is because of the development of cumulonimbus other than diurnal heating. CB will develop in a cold unstable air mass, and can be triggered by low level convergence usually caused by the wind flow interacting with the land forms. This is why they can occur at any time of the day in Wellington, and it is what happened on Wednesday morning (6 May 2009) when the southern suburbs were rudely awoken by crashing thunder and the sky lit up with spectacular lightning.

It started at 5:45am when a small trough arrived in Cook Strait. The convergence was between the westerlies ahead, and the southerlies behind the trough being squeezed into the harbour entrance. That was the trigger and away it went! Half an hour later it was over and the trough moved away to the northeast.


The map is from our Lightning Detection Network and shows the lightning strikes between 5:45 and 6:10am as red dots concentrated over the southern suburbs and south of Island Bay and the harbour entrance.


The radar scan at 6am shows the position of the trough line (by its rain echoes) near Wellington and extending out to the east.

I was in Upper Hutt at this time, just getting out of bed. I heard nothing. While I was waiting for the train at 6:40am I could see the top of a CB to the south. “Hmmm. Interesting. Palliser Bay perhaps?”. When I got into the office here, the place was buzzing with excitement, and tales of lightning, hail, and deafening thunder.

Here is a great video posted by Matt Kovesdi. Thanks, Matt.
There is a growing collection of New Zealand weather videos on the video section of our website.