The Snow of ’39

Ranfurly Shield Rugby in the Snow
Ranfurly Shield Rugby in the Snow. Source: The Weekly News, August 1939

Among the benefits of giving public talks on the weather are the stories you get back from the audience. Once, after a lecture in Dunedin when I described the cyclone of 1936 as the worst storm to have hit New Zealand in the last century, a number of people mentioned the snow storm of 1939. I had never heard of it, but when I looked up the old newspapers, I was amazed to find that it affected pretty much the whole country.

On 31 July 1939, snow was reported by the lighthouse keeper at Cape Maria van Diemen at the top of the North Island. A few days earlier, it snowed in Dargaville and Ruapekapeka, and snow lasted on the hills behind Kaikohe for several hours. In Auckland, snow fell in many suburbs just before dawn on 27 July. Five cm of snow lay on the summit of Mt Eden, while the Bombay hills shone white for most of the morning.

In Gisborne, snow fell for nearly three hours, covering lawns and gardens. Snow fell to sea level at Castlepoint, too, and the road to Masterton was closed by drifts at Big Saddle on the Whakataki hill. Snow lay 15 cm deep in Masterton, where the town clock was stopped at 2.20am by the weight of snow on its hands.

In the east of the South Island, the falls were heavier. Snow cut off Banks Peninsula from Christchurch for a time, and lay 30 cm deep in Akaroa. Sixty men in a public works camp near Duvauchelle were without food for two days until a launch arrived with supplies. Stock sheltering in gullies were buried in snow drifts up to 10 m deep, and many perished. In places, snow covered the tops of fences, and sheep and cattle roamed at will. Although snow in Christchurch was only a few centimetres deep, a layer of frost on top caused numerous accidents and disrupted the tram service.

In Dunedin, conditions were much worse. Here snow and thunderstorms began during the evening rush hour on Monday, 24 July, and by morning snow lay 15 cm deep over the city. Only one bus made it out to the north over Mount Cargill, assisted by a gang of men with shovels. Electric trams did not commence running until the middle of the morning, and then only on certain lines on the flat.

Buses made some progress over the snow, but cars needed chains, and visibility was hampered for many whose windscreen wipers had frozen to the windscreens. The road to the south was closed, as was the airport. Some trains got through, running silently over the snow-covered rails.

The snow continued through Tuesday, and by Wednesday morning it was 35 cm deep at St Kilda and up to a metre deep in the more elevated suburbs of Roslyn and Maori Hill. Icicles 30 cm long were reported on some houses, and the weight of snow caused roofs and skylights to cave in. The only fatality occurred when a man clearing snow from a factory skylight fell 10 m to the floor.

The radio masts at Highcliffs, on Otago Peninsula, were struck by lightning, and the staff there isolated without adequate food. A rescue mission was launched under the command of the director of 4YA, H. Ninnis, who had been with Shackleton in the Antarctic. He was accompanied by four members of the Otago Ski Club. To transport the supplies they borrowed one of Captain Scott’s sleds from the Otago Museum. When they reached the station, one of the men coming out to receive the supplies sank up to his neck in snow.

Supplies were also running short in many parts of town, as no fresh meat, vegetables, milk or coal were coming into the city.

However, the abundant snow provided a rare opportunity for people to enjoy winter sports. Snowballing was rife, with battles between rival groups of businesses downtown. Any moving target was fair game, including cars and buses, some of which lost windows, and the police were obliged to intervene. Snowmen sprang up all over town, and one school even managed to stage a paper chase, using Condy’s crystals to stain the snow red and make a trail.

Skiers had a grand time eagerly assisting in rescue missions, including one farmer whose barn had collapsed. Tobogganing was also popular, with strips of linoleum and tea-trays called into service. Speeds of up to 70 km/h were reported and there were inevitable accidents. One concussion victim had to be carried by stretcher-bearers over 200 m of soft snow to reach the ambulance.

Inland, the train from Lumsden to Kingston was marooned at Eyre Creek when the engine charged into a drift and became stuck. Passengers and crew were forced to spend the night on the train, and had to raid the cargo for dinner.

Snow was not the only problem during the cold snap. Overnight frosts caused water pipes to burst in Palmerston North and Hastings, while at Paremata, just north of Wellington, eight hectares of the harbour froze over. Tidal waters also froze at Opotiki in the Bay of Plenty, where sheets of ice were left attached to wharf piles as the tide went out.

The cold continued for a while and Invercargill was covered in 15 cm of snow overnight from Friday to Saturday 12 August. The Ranfurly Shield game between Southland and Manawatu took place with only the lines cleared. Perhaps more used to the chilly conditions, Southland beat the North Islanders 17-3.

Up and down the country, it was regarded as the worst winter for snow in living memory. Considering how temperatures have warmed in the last seventy years, we may not see these conditions again.

Record-Breaking weather for Coastal Classic

The weather pattern was almost perfect for the 2009 HSBC Coastal Classic Yacht race from Auckland to Russell.

Part of the 2009 Coastal Classic fleet

A fleet of 205 boats set off from Devonport at 10am on the Friday before Labour weekend in prime conditions, with a 20 knot southwest breeze through Rangitoto channel and a steady 15 to 20 knot southwest wind in the Hauraki Gulf.

A new overall race record was set by Neville Crichton on keelboat Alpha Romeo, smashing the  seven hour barrier and completing the race in 6 hours and 43 minutes  (an average speed of 17.9 knots for this 119 nautical mile race), replacing the 7hr 20min multihull record set by Split Enz in 1998, and keelboat Zana’s 8hr 29min record set in 2005.

Two other race records were set: For a keelboat under 9.14m, Overload took 10hr 23 min, beating the old record of 11hr 31min by Extreme in 1996;  and for a multihull under 10.68m, Frantic Drift took 7hr 58min, beating the old record of 8hr 02 min set by Redken Cat in 1996.

Taeping, a 13.5 m catamaran, just missed the multihull 1996 record by seven minutes.

‘We had a fantastic ride all the way’, Crichton told Montgomery. ‘We hardly went under 18kts, it was a dream run.’

‘We cleared the field very early in the race. We carried spinnaker all the way to Kawau (26nm north if Auckland) and from there were two sail reaching all the way.’ ‘From Tutukaka onwards we had 25kts, and a beam reach which is just beautiful sailing. It probably suited multihulls better than us, but we were able to stretch our legs, at the start, and they never got back at us at us.’  ‘From Cape Brett to Russell it was right on the nose’

The weather pattern that allowed these records to tumble involved the combination of a cold front moving off and an anticyclone approaching from the Tasman Sea.   The map at 0000 UTC 23 Oct 2009 below (1pm Friday 23 Oct New Zealand daylight time) shows these features a few hours after the start of the race.

1pm Friday 23 October 2009
1pm Friday 23 October 2009

That black line drawn across northern Waikato and the Bay of Plenty region marks a trough line. This trough formed in the lee (or downwind) side of the North Island as the air pressure built on the windward side.  In this case the pressure rose faster over Hawke’s Bay than around Auckland, and consequently the isobars were twisted into a lee trough.  Sometimes, when conditions are conducive aloft, a lee trough deepens as it is displaced northwards,  and the result is a surge of easterly wind across the Bay of Plenty that is concentrated by the gap between the Coromandel Peninsula and Great Barrier Island, so that it moves across the Hauraki Gulf and arrives in Auckland with a wet flourish.

The weather map for 1200UTC on 23 Oct 2009 (1am Saturday New Zealand daylight time) shows that, in this event,  the southeast wind change remained mostly offshore.

1am Saturday 24 October
1am Saturday 24 October

Look closer at the 1020 hPa isobar and you can see that at 1am/1200UTC it was still over the Auckland area and moving eastwards.

At the briefing for this year’s Coastal Classic, MetService Weather Ambassador Bob McDavitt was able to interpret the weather information available and provide the helpful racing tip that if a vessel observed that their barometer was rising above 1020 hPa, then they should prepare for a southeast wind change.

In the animated map of wind observations below, add 13 hours to convert the timestamp to New Zealand daylight time.  The plotted winds flow from the feathered end to the pointed end; each barb is worth ten knots, and half a barb is five knots.  The observations highlighted in red are from Bean Rock in the Auckland harbour entrance and from Mokohinau Island, to the north of Little Barrier Island.  Mokohinau’s pressure is also plotted in code. The numbers given are the last three numbers of the pressure in tenths of hPa, with the leading  numbers deleted:  thus, 173 at the start of the loop with timestamp 20:30 translates to a pressure of 1017.3 hPa at 9:30am Friday (half an hour before race start).

click to play
click to play

Note how the pressure at Mokohinau rose above 1020 hPa briefly between 11pm Friday night, and 1am on Saturday.  Winds turned southeast to easterly along the Coromandel coast and through Colville Channel from 1am until dawn, showing that Bob’s race tip worked OK – where the pressure got above 1020 hPa there was a southeast change.

The southeasterly winds along the coast near noon on Saturday were more likely sea breezes rather than remnants of the easterly surge.

If this easterly surge had been strong enough to make it to the Bay of Islands, and if it did so at the right time, then Alpha Romeo would not have had those head winds in the final stretch and could have completed a quicker time.

This is a good example of how sailors can use a barometer to check that the weather is unfolding in the real world as forecast by the computers.

The Structure of Lows – part II

In my previous blog post I pointed out that tropical lows and cyclones don’t have fronts like the lows we’re used to around NZ, but rather, a core of warm air near the centre. I’d like to follow up by further contrasting tropical and mid-latitude lows, and looking a bit more closely at tropical cyclones and how they can affect our weather in New Zealand.

When tropical lows fully develop into cyclones they become the most damaging of tropical weather features. There are three main reasons for this:

  1. damaging winds at, or very close to the Earth’s surface
  2. intense rainfall leading to flooding
  3. near the coasts, the sea can be driven inland as a storm surge.

In an earlier blog post we saw that, on a weather map, the closeness of the isobars is related to the strength of the wind. There is a latitude effect too: for a given isobar spacing, if you were nearer the equator the wind would be stronger than if you were nearer the poles. This is an important point, because a tropical cyclone has very closely packed isobars near its centre and is relatively close to the equator.

Also, as you get nearer the equator the wind blows more across isobars, in contrast to the mid-latitudes where the wind blows approximately parallel with the isobars (like slot cars on a track). So in tropical cyclones there is a strong inflow into the centre as in the diagram below.

Inflow into a tropical cyclone (southern hemisphere). The innermost circle represents the "eye".
Inflow into a tropical cyclone (southern hemisphere). The innermost circle represents the "eye".

I don’t want to overcomplicate things, but there is a further effect based on the curvature of the isobars. If you’re interested send me a comment and I’ll explain it.


At the innermost circle in the diagram, the air is racing so fast that it doesn’t make it all the way into the centre. Inside this circle is the eye and a phenomenal contrast between the wind and rain at the outer wall and the relative calm of the interior.

The rainfall is usually at its most intense just outside the eye, and the cloud is very deep with extremely cold tops – see for example the infra-red satellite pictures described in the previous post. I remember being in Darwin many years ago when Cyclone Rachel was over the Timor Sea. We were some distance away from the centre of Rachel, but the cloud overhead was so dense that we needed lights on during the daytime if we wanted to read. We set a 24hr rainfall record for Darwin of 290mm, although the wind wasn’t particularly strong (compared with what I’m used to in Wellington).

Tropical cyclones sometimes affect New Zealand during our warmer months – a couple of examples were mentioned here. You may well recall some more recent examples, e.g. Drena and Fergus. After they form in the tropics they typically move quite slowly and erratically, then usually curve southeastwards into the mid-latitudes.

Once they move over colder water in mid-latitudes they undergo a transition in structure from the tropical to the extra-tropical (meaning outside the tropics, like “extraordinary” means “outside the ordinary”).

At this stage the cyclone tends to expand by drawing in colder contrasting air and evolving into a more typical mid-latitude storm complete with warm and cold fronts. On a few occasions (e.g. the Wahine Storm in 1968) they interact with a vigorous mid-latitude low to form a new system, complete with active fronts. Either way, the new storm isn’t a tropical cyclone anymore in terms of structure but can still be very damaging as a major mid-latitude depression.

Puysegur Point Storms

Years ago, I heard the wife of a lighthouse keeper talking on the radio about the weather at Puysegur Point, on the south coast of Fiordland, where she and her husband had been stationed for a time. Six months or so before they arrived there, a fishing boat had gone down in a storm with all hands. One day, as she walked along the black stony beach, something white caught her eye. Bending over, she picked up a single human tooth. She could not be certain it came from one of the lost fishermen but it symbolised for her the little that the sea gives back after it unleashes its fury on a small boat.

When I told this story to a retired lighthouse keeper I met a couple of years later, he countered with another about strong winds. Because Puysegur Point is so isolated, successive keepers farmed sheep in paddocks near the lighthouse in order to provide themselves and their families with fresh meat. As extreme winds were common, sheep were often blown over the cliffs and onto the rocks below. Eventually the keepers switched to cattle, as they were harder for the wind to dislodge.

The reason that the winds near Puysegur Point are frequently the strongest over the country is because it is located near the southern end of the Southern Alps. When westerly winds slam into the mountains, not all the air manages to rise over the ridgeline. Some pours around the end of the mountains.

Winds from both the northwest and southwest blow fiercely on this coast, but a northwest wind ahead of a front is the worst, for two reasons. First, the wind blows perpendicular to the Southern Alps. Second, as the air approaches New Zealand, the progressively colder sea surface cools it from below. This makes the air more stable, and so resistant to rising over the mountains, causing a greater proportion to flow around the Fiordland end of the Southern Alps.

These strong winds also show up clearly in measurements taken by a satellite borne instrument known as a scatterometer. This machine emits pulses of microwave energy and measures the returning signal reflected from the sea surface. The strength of the return signal depends on the size of the small capillary waves on the water, which in turn depend on the wind strength.

A good example of this effect occurred on 20 November 2003. The scatterometer winds in the map below show a stretch of 50 knot winds (solid triangles ) more than 100 kilometres long with a few readings of 60 knots. The associated weather map shows the isobars close together over southern Fiordland.

More examples of scatterometer winds can be found on NOAA’s QuikSCAT site.

Scatterometer winds 20 November 2003
QuikSCAT Scatterometer winds 20 November 2003
Weather Map 20 November 2003
Weather Map 20 November 2003

The Structure of Lows

You are probably familiar with seeing lows and highs on our weather maps around New Zealand. See, for example, previous blog posts on a mid-July northern lowHow Lows and Highs move and the satellite loop in Winds Aloft.

The lows or depressions that affect us in the mid-latitudes are accompanied by warm and cold fronts with marked contrasts between the warm/moist air and cold/dry air that follow behind these two types of front, respectively. See our Learning Centre for more.

Occasionally, at favourable times of the year, lows form in the tropics that, on the face of it, may seem to be quite similar to the lows around New Zealand. But they are actually very different. If you looked at only 2-D surface weather maps you might think that tropical lows and cyclones are like our mid-latitude lows, but their structures differ markedly.

In different parts of the world these tropical cyclones are given other names. Here is a satellite image of a typhoon affecting the Philippines (in middle-right of picture).

Infra-red satellite image, 10am NZ Daylight Time, 2 Oct 2010
Infra-red satellite image, 10am NZ Daylight Time, 2 Oct 2009

This typhoon, called “Parma”, was following hot on the heels of Typhoon Ketsana that caused serious flooding in the Philippines and then Vietnam in late September.

The image is specially enhanced to highlight the deepest and most active cloud, shaded white and blue in this case. The temperature of these coldest cloud tops is -90C. This is perhaps an example of a paradox: In the global troposphere the coldest air is at upper levels in the tropics, not in the mid-latitudes and not at the poles.

How colours relate to temperature in the previous infra-red satellite image

Also note that with this enhancement there is a double entry for black: cloud tops between -20 and -30C appear similar to warm tropical sea (high 20s) between the clouds.

As I implied above, these tropical lows don’t have fronts associated with them. They normally have a warm core with a much more symmetric structure than mid-latitude lows. This core of warm air near the centre also means that their strongest winds are near the ground.

Another difference is a region of light winds at the centre, called an eye. The eye can’t always be detected in satellite pictures, but does show in the recent example below. Typhoon Melor (upper right of picture, encircled by purple and blue) was curving from westwards to northwards over the North Pacific Ocean. Parma was just north of the Philippines at this time. 

As previous satellite image but exactly four days later

When the tropical lows fully develop they become the most damaging of tropical weather features. I will explain the reasons for this in my next blog post, with a focus on how they can affect our weather in New Zealand.

Early October Snow

The snow that closed the Desert Road and Napier-Taupo Road on Sun-Mon-Tue 4-5-6th  October 2009 was unseasonable.  It was caused by a low pressure system deepening over the area at the same time as a cold southerly flow arrived, resulting in moist air being cooled from below in a cauldron of lowering pressure.  This produced an unusually heavy amount of snow over a wide area.

The weather map for noon on Sunday shows the low pressure system forming over the Central North Island

Wetaher Map 1pm Sunday 4 October  2009 (0000UTC)
Wetaher Map 1pm Sunday 4 October 2009 (0000UTC)

And the satellite image for the same time shows the extent of the cloud —skies had already mostly cleared over the South Island

Satellite image (visible) 1pm Sunday 4 October 2009 (0000UTC)
Satellite image (visible) 1pm Sunday 4 October 2009 (0000UTC)

MetService issued a Severe Weather Watch for heavy snow on Saturday night and upgraded this to a Warning soon after 9am on Sunday, advising of 20-30 centimetres of snow over the higher ground, including the Hawke’s Bay hills.

MetService maintains a number of NZTA (New Zealand Transport Agency) weather stations on the Napier-Taupo and Desert Roads with their locations shown in this Google Map image:


The following stations line up along the Napier-Taupo road from west to east:

APA Taupo Airport

HLX: High Level Road

ERX:  Eastern Rangitaiki

TOX: Te Haroto

TEX: Te Pohue

NRA: Napier Airport

For these sites, the temperature (degrees Celsius) and rainfall (mm/hr) graphs from 1am NZDT Saturday 3 Oct (02 1200 UTC) until 1pm NZDT Tuesday (06 0000 UTC) are shown below.


Note the warm afternoon on Saturday (20 C in Napier), the rapid cooling to zero degrees air over the road around noon on Sunday – at around a third of the way through the rainfall event– and the steady warming on Tuesday morning.

The following stations line up from north to south along the Desert Road

TTX Tongariro

MRX Moturoa

DRT: Desert road

RUX Waiouru

For these sites, the temperature (degrees Celsius) and rainfall (mm/hr) from 1am NZDT Saturday 3 Oct (02 1200UTC) until 1pm NZDT Tuesday (06 0000 UTC) are shown below.


Comparing these traces with the Napier-Taupo trace we can see that the temperature drop and rise were more gradual on the Desert Road but the rainfall events were similar.  That peak of rain at Waiouru around noon on Tuesday is due to snowmelt – the snow that had accumulated in the rain gauge melting as conditions warmed..

It seems that around 30 centimetres of snow fell on the upper parts of these roads,  and the wind blew this around producing some drifts of around half to one metre. It took several days for the Napier-Taupo Road to reopen.

In most of the surrounding cities, Monday’s temperatures remained in single digits.  Only Gisborne managed to get above 10.  It was reported that snow lay on the roads around Rotorua:


Decode: AAA for Auckland Airport, GSA for Gisborne Airport, HNA for Hamilton airport, ROA for Rotorua Airport, and TGA for Tauranga Airport. Temperatre in degrees Celcius, and timestep from  1am NZDT Saturday 3 Oct (02 1200UTC) until 1pm NZDT Tuesday (06  0000UTC) are shown above.

This was indeed a freakish early October snow event, so late in the year, so low, so heavy and so widespread — these details deserve being kept as a bench mark for possible future comparisons.