edited by Mark Pascoe



There is a psychological barrier to predicting a rainfall expected only once every 150 years or more. It’s like backing the least fancied runner in a horse race. This is why meteorologists almost never forecast an extreme event. However, these are the ones that are the most important, and can have major effects on the community. Hence this publication.

A secondary purpose of this treatise is to have some sort of documentation of past events. Within it you will see some huge rainfalls being reported for events which to us occurred an enormously long time ago, such as the the record-breaking Hawkes Bay rainfall of 1924, and the cyclone of 1868. By 1870, just a few decades into the period of European settlement, New Zealand rivers had been responsible for 1115 recorded drownings. Drowning became known as “the New Zealand death”. But is the frequency of flooding increasing, and are the floods of today any worse? Our knowledge of river flow records is handicapped by lack of information on past events, as indeed it is for extreme rainfalls. A simple tabulation of high rainfall events will produce an increase over time, but one would expect that anyway as the rainfall network and population density increases. The consensus in New Zealand would appear to point to an increase in localised events, with little change in the frequency of widespread, major floods.

As global warming brings more extreme and more frequent floods, there is an increasing need for them to be accurately forecast. At present, MetService does an excellent job of forecasting heavy rainfalls (defined for verification purposes as exceeding 50 mm per six hours, or 100 mm per 24 hours). What we do not do so well with is the extreme events – the ones that inundate farmland and houses, cause massive landslips, drown stock, and take lives. So the main objective of this publication is to provide some insights into what makes an event extreme.

Although some of the content is technical, this publication is designed to appeal to a fairly broad range of meteorological knowledge. It is hoped that professional meteorologists, particularly heavy rain forecasters, will be assisted in doing their job. It is hoped some insights will be thrown up into what turns a mundane heavy rain event into something lethal.

This work gathers together 122 extreme events between 1858 and 2005. The events are divided into geographical regions, and are placed in the region that experienced the most severe weather. They have been selected based primarily (but not exclusively) on rainfall return periods, rather than severity of flooding. In no way can the list pretend to be exhaustive, and I am sure that a number of significant events, for various reasons, have not made it onto these pages.

Some caution needs to be exercised in interpreting return periods. An extreme rainfall that occurs over a small area is much less significant than one that occurs over a large area. A thunderstorm can deliver a high rainfall to an area as small as 1 square kilometre. The land area of New Zealand is 268,000 square kilometres. This means that thunderstorms are going to produce 268,000/100 = 2680 one-hundred-year rainfalls in various tiny parts of New Zealand in an average year. (Statisticians will tell me that the above is not entirely accurate, but the general idea is correct.)


MetService Expert Forecasters (Erick Brenstrum, John Crouch, Andy Downs, Allister Gorman, Bob Lake, Paul Mallinson, Ian Miller, Steve Ready): Severe Weather and Severe Convection Logs.
John Crouch: Much of the material in the Severe Convection Section.
Steve Ready: Data on Cyclone Bola.
Allan Penney, Errol Lewthwaite (NIWA ). Assistance with dredging the archives.
Marianne Watson of Horizons Regional Council, and Craig Thompson of NIWA : The rainfall recurrence interval map for the 2004 Manawatu Flood Event. (Fig. 112).


Cumberland, K., 1985: Rivers and Lakes. In: New Zealand. Whitcoulls Publishers, Wellington.
Mosley, M.P., and C.P. Pearson, 1997: Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society.


Northland and Auckland are the first places in New Zealand in the firing line for ex-tropical cyclones. Lows of sub-tropical origin can also produce heavy rains. In heavy rain situations, the wind flow over the area is usually (but not always) easterly or northeasterly. Auckland, however, does receive some sheltering by the Coromandel Peninsula in flows from the east, especially from the City southwards. The exact direction of the flow is important in determining which parts of the Auckland Region,
if any, will receive heavy rain. In Northland, the Hikurangi area usually receives the most substantial falls.

Convective activity plays a major role in heavy rain events in Northland and Auckland, particularly in summer. Because these convective rainfalls tend to affect only limited areas, there are a large number of events which produce a 100-year rainfall somewhere or other.

A critical element in the forecasting of heavy rain events in this area, particularly Auckland, is the presence of a slow-moving low level convergence zone in the presence of suitable upper-level support.

26th January–6th February

Taheke: 231mm/24hr (return period over 150 years)

A 10-day period of wet weather included the large 24-hour rainfall of 231mm in 24 hours at Taheke on the 2nd February. Most bridges in Northland were
swept away.

A blocking high near the Chatham Islands maintained a long period of moist northeasterlies over the northern North Island, culminating in the passage of a front.

1935 15th February

Auckland City: 114mm/24hrs (return period 10 years)
incl 89mm/2hrs (return period over 150 years)
and 57mm/30 mins (return period over 150 years)

Considerable damage was caused to shops and warehouses in Auckland City, and to low-lying areas in the suburbs, after two hours of torrential rain. Workmen went swimming in Fort St to locate blocked stormwater gratings.

The flooding was caused by convective activity within a slack, low pressure system.

1936 1st-2nd February

Whangarei 290mm/24hrs (return period over 150 years)
Russell 301mm/24hrs (return period over 150 years)
Warkworth 275mm/24hrs (return period 120 years)

Auckland City 162mm/24hrs (return period 45 years)

Stock losses: Heavy
Damage: £115,000
2012 Dollars: 12.7 million

Hawkes Bay:
Maraetotora 254mm/24hrs (return period 125 years)
Stock Losses: Serious
Road Damage: £13,400
2012 Dollars: 1.5 million

Total Deaths: Several
Total Nationwide Damage: £10 million
In 2012 Dollars: 1 billion.

An ex-tropical cyclone caused widespread devastation across a large part of the country. The centre passed just west of Northland and Auckland, then veered
southeast to cross the central North Island, emerging near Napier.

In parts of Auckland there were floods of record proportions, and Whangarei was flooded to depths of up to a metre. Napier was isolated by floodwaters. There was also extensive flooding in Whanganui, Manawatu and Wairarapa. The Ashley River in Canterbury rose to unprecedented heights. The wind blew in windows from Kaitaia to Picton, the worst hit being the Manawatu, where the grandstands of the A&P Association, the Awapuni Racecourse and the Sportsground were demolished.

The rainfall distribution for the event is shown in Fig 1.

Fig 2 shows the MSL chart for 9am February 2nd. The central pressure is drawn as below 970 hPa. Palmerston North is reporting a wind of 10 on the Beaufort Scale, implying winds of storm force (averaging between 48 and 55 knots). This figure also shows isallobars (rate of pressure change), in hectopascals per three hours. The biggest pressure falls are seen to be occurring off the southern Hawkes Bay coast.

thirtysix thirtysix2
Fig 1. Rainfalls 9am January 31st to 9am February 2nd, 1936. Fig 2. MSL analysis for 9am 2nd February, 1936.

1966 16th February

Whenuapai: 260mm/24hrs (return period over 150 years)
incl 107mm/1 hr (return period well over 150 years)

A downpour over Whenuapai broke the official record for a one-hour fall, a record which stood for 35 years. Heavy rain was quite widespread through Northland and Auckland. The 24-hour rainfall at Whenuapai was 260mm, and about 400mm fell at Kokopu. The MSL analysis for 6pm 16th (Fig 3) did not suggest that anything extraordinary was about to occur.

The flow remained very anticyclonic-looking through to 6am on the 17th (Fig 4), when a north-south front suddenly appeared on the analysed charts.

A look at the upper air (Fig 5) reveals a trough in the Tasman Sea, exhibiting a slight negative tilt. The flow over the North Island at 200 hPa at midday on the 16th was anticyclonic.

whenuapai Fig 3. MSL analysis for 6pm 16th February 1966.
whenuapai2 misc
Fig 4. MSL analysis for 6am 17th February, 1966. Fig.5. 500 hPa analysis for noon 16th February, 1966.
Convection obviously played a major role in this event. However, the appearance of the north-south front on the surface charts the next morning suggests that things were also occurring on the synoptic scale. There could well have been a small-scale upper-level perturbation which would not have been picked up by the sparse upper-air network.

1971 18-19th April

Stock Losses: 74

Major slipping occurred on hillsides, with roads damaged. Buildings were damaged and paddocks flooded.

The rainfall chart (not shown) showed a double centre. One centre with rainfalls over 6.5 inches (165mm) lay over the Settlement Road, Taipuha area. The maximum rainfall within this area was estimated at 250mm, and this fell in the two and a half hours to 10pm 18th. This is 2.59 times the 150- year return period rainfall, and would place this storm at number three in the hierarchy (see Appendix 1). The other centre lay over Whangarei Heads. The rainfall here was estimated at 230mm in the five and a half hours to 2.30am on the 19th. This is also a well over 150-year return period rainfall.

MSL charts (Fig 6) showed a frontal band lying slow-moving over Auckland, while Northland lay within a humid north to northeast airstream. The rain was accompanied by intense electrical activity. Upper level charts (not shown) revealed a negative-tilt trough moving onto Northland. The flow at this level appears quite difluent, suggesting upper-level divergence was a significant forcing mechanism.

taipuha taipuha2
Fig 6. MSL analyses for midnight 17th (left) and midnight 18th (right) April 1971.

1974 22-23rd February

Paitu: 258mm/6hrs to noon 23rd (return period well over 150 years)
incl 134mm/2hrs to 11am 23rd (return period over 150 years)
One bridge was lost, rivers overflowed their banks, and many fences were flattened. The two-hour rainfall recorded at Paitu (near Kaeo) between
9am and 11am was a huge 134.3mm. Total rainfall for the two-day duration of the event exceeded 350mm. The distribution of rainfall is shown in Fig 7.
kaeo Fig 7. Rainfalls for 48 hours to 9am, 23rd February, 1974.

The MSL chart (Fig 8) shows Northland lying within a moist and very humid northeast flow. Dewpoints were up around 21C. Satellite imagery (not shown) revealed a broad cloud band covering the whole of the North Island. Convection was a major contributor.

It’s a little hard to tell from the figure (Fig 9), but upper level divergence may also have been a significant factor. At 500 hPa, a ridge lay over Northland.

kaeo2 kaeo3
Fig 8. MSL analysis for noon, 23rd February, 1974. Fig 9. 300 hPa analysis for noon, 23rd February, 1974.



1975 30th May

145.5mm/8hrs to 3pm 30th. (return period 30 years)
Matapouri: 277mm/24hrs (return period over 150 years)
Damage: $150,000.
2012 Dollars: 1.4 million
The most severe damage was restricted to an area of only 60 sq km, although considerable damage was caused over an area of 230 sq km. Two houses were demolished by slips, and six houses were flooded. There were also some stock losses. Most of the rain fell in two separate events, of 2.5 and 4.5 hours. It is estimated that at the storm centre, 335mm fell in just eight hours, with a return period of well over 150 years. The MSL chart (Fig 10) featured a low west of Northland, with a slow moving front over the area.
Strong low level convergence was a major factor in the extreme rainfall. There was also significant upper level divergence, hinted at in the 300 hPa chart shown in Fig 11.
ngun ngung2
Fig 10. MSL analysis for noon, 30th May, 1975. Fig 11. 300 hPa analysis for midday, 30th May, 1975.

1981 19-20th March

A massive flood stuck the Kerikeri area, and New Zealand’s oldest house, Kemp House, was inundated to ground floor window-sill level. The Waipapa Flats were completely covered with water, and in the Waipapa Landing area a house was completely washed away. Fifty boats were torn from their moorings. Most of the river valleys at the northern end of Kerikeri were scoured, and the
vegetation along their banks totally cleaned out. The isohyetal chart (Fig 12) shows a considerable area recording over 300mm in the seven-hour duration
of the event. 174mm in 2.5 hours was recorded, but intensities during shorter periods were not available.

The extreme rainfalls recorded put this event well up
in the hierarchy (see Appendix 1). The MSL chart (not shown) looked rather innocuous, with a large high centred
east of the South Island, and a weakly cyclonic easterly flow over Northland. The air was moist and unstable however, and the heavy rain was the result of an outbreak of severe convection.


F. Hunt: 448mm/9.5hrs
(return period well over 150 years)
incl 174mm/2.5 hrs
(return period well over 150 years)
Deaths: One.

Fig 12. Rainfall overnight 19/20th March, 1981 (less than 10 hours).



See Landslide Sweeps Through Te Aroha


See Bola Devastates Gisborne District

1997 30th June


Brynderwyn Hills: 310mm/24 hrs (return period over 150 years)
incl 96mm/1hr (return period over 150 years)
and 165mm/2hrs (return period well over 150 years)
Maungaturoto: 268mm/24hrs (return period over 150 years)
Purerua: 131mm/7hrs to 4pm (return period 80 years)

Convection flared up on a front as it moved southwards over Northland. As well as the flooding in Kerikeri and Kaeo, many roads, including State Highway 1, were closed by slips. A depression was moving slowly eastwards across the north Tasman Sea (Fig 13).
kk The front appeared to activate in response to an upper level trough, guided by a 100-knot mid-level northwest jet near Norfolk Island creating strong divergence across the far north of the North Island (Fig 14).
A strong easterly gradient added
to the rainfall intensities.
The warning was issued late, and significantly more rain fell than
was forecast. The convection mostly bypassed Auckland, and the warning issued for that region was a false alarm. The flow was too easterly, and Auckland was therefore sheltered by the Coromandel Peninsula.
Fig 13. MSL analysis for midnight
29th June, 1997.
Solid lines – isobars.
Dashed lines – 850 hPa temperature.
Shaded areas – 700 hPa humidity (%)
kk2 Fig 14. 250 hPa windbarbs for
midnight 29th June, 1997.
Thin lines – isotachs (knots).
Shaded areas – divergence.

1999 21st January

Homes were washed away in the small towns of Pawarenga, Panguru and Omapere after walls of water and logs swept down the valleys. In Pukekohe homes were flooded, and two rest homes had to be evacuated.
Intense convective activity was the main forcing mechanism in this event.

A broad slow-moving trough lay over northern New Zealand, with a weak easterly flow over the area (see Fig 15). Shears were low. Low-level convergence was also present, as the westerly sea breezes met the large-scale easterly flow. Diurnal afternoon heating, and warm, moist tropical air at low levels, all contributed to the development of large cumulonimbus clouds.
Fig 203, in Appendix 3, shows the sounding for Auckland for the time of the storm. It features tall ‘skinny’ CAPE, with a lifted index of –6.0C and a SLI of –2.8C.
The CHAMP model provided excellent guidance to forecasters, who were able to issue warnings for this event. The global models did not indicate heavy rain over the area, being unable to resolve the small-scale convective nature of the event.


Panguru: 100mm/3hrs
(return period over 150 years)
Bayly’s Beach: 90mm/2hrs
(return period over 150 years)
Omapere: 120mm/5hrs
(return period over 150years)
Awakino: 160mm/4hrs
(return period well over150 years)
Pukekohe: 160mm/3hrs
(in fact just over 2 hrs)
(return period well over 150 years)

Deaths: Two

Evacuations: Hundreds.



Fig 15. MSL analysis for 6pm 21st January 1999.

2000 21st April

Matauri Bay: 124mm/1 hour ending 2pm 21st. (return period well over 150 years)
Near Warkworth: 75mm/45 mins (return period over 150 years)
An unoffical New Zealand record for a one-hour rainfall was set when torrential rain hit parts of Northland and Auckland. The heaviest rain hit a few small areas, causing floods and disruption. A tornado reported in Pukanui (40km north of Kaitaia) also caused considerable damage.
The extreme convective rainfall occurred within a shallow low-pressure system, containing moist sub-tropical air (Fig 16).
Low-level convergence was present, along with weak CVA. The severe convection was only isolated in Northland, but in Auckland was widespread enough to cause the Regional Council some concern. No warnings were issued for either area.
mat Fig 16. MSL analysis for noon 21st April 2000.

2000 28-29th June

Awanui: 205mm/24hrs (return period 150 years)
Birkenhead: 110mm/24hrs to 9am 29th (return period 8 years)
Convection was again the main culprit when a downpour struck Awanui, just north of Kaitaia. Homes were evacuated, schools and roads closed, and a bridge washed away. Heavy rain also struck the North Shore, with a rest home evacuated, and the motorway flooded.
A low moved from the north Tasman Sea to the area west of Northland.
A strengthening northeast flow lay over the northern North Island, with a front moving slowly southwards onto Northland. A small low moved southwestwards along the front, accompanied by CVA and upper divergence, thus activating the front. (See Fig 17.) Shears were low.
While a heavy rain forecast was issued in good time for both Northland and Auckland, the blow-up of convection was not anticipated, and rainfall amounts in a few areas were much greater than forecast.
aw Fig 17. MSL analysis for midnight 28th June 2000.

2001 30th May

Leigh: 109mm/1 hour to 2.40am
(return period well over 150 years)
A severe convective storm washed a car into the family lounge and took the official record for a one-hour rainfall. The high rainfall was confined to a very localised area around Leigh, Mangawhai and Great Barrier Island. This area briefly lay under a “triple point” – the intersection of warm, cold and occluded fronts. The nearest sounding – that for Whenuapai at midnight – was unspectacular. Figs 18 and 19 show the MSL analyses for midnight 29th and 6am 30th. The triple point can be seen west of the area at midnight, and to the east at 6am. The 250 hPa analysis for this time (Fig 20) shows the area to be under the polewards exit region to a jet, and therefore in an area of strong upper level divergence. Above. Fig 18. MSL analysis for midnight 29th May 2001.
leigh leigh2
Fig 19. MSL analysis for 6am 30th May 2001. Fig 20. 250 hPa analysis for midnight 29th May, 2001.


Barnett, M.A.F., 1938: The Cyclonic Storms in Northern New Zealand on the 2nd February and the 26th March, 1936. New Zealand Meteorological Service Office Note 22.
Brenstrum, E.M., 2000: The Cyclone of 1936: The Most Destructive Storm of the 20th Century? Weather and Climate Vol. 20. p. 23-27.
Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
McGavin, T. : Severe Weather Events Analysis. New Zealand Meteorological Service
Mosley, M.P., and C.P. Pearson, 1997: Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society.
New Zealand Herald 18th February, 1966.
________ 21st March, 1981.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
New Zealand Gazette, 1936, Vol. I. New Zealand Government.
Pascoe, R.M., 2001: Forecasting Heavy Rain in Auckland. New Zealand Meteorological Service Internal Report 6.
Severe Weather Log. New Zealand Meteorological Service.

(including Coromandel Peninsula)

Northwesterly airstreams are the main bringers of heavy rain to the Waikato area. The area is dominated by the Waikato, New Zealand’s longest river. Because of its length, and also its relative flatness, it takes several days of heavy rain beforethe river will flood. Unfortunately, it also takes a long timefor the river level to go down again. As with other parts of thecountry, stationary fronts can cause flooding, and convectionoften plays a part.
Moist northeasterlies ahead of a frontal band or associated with an ex-tropical cyclone are required to produce heavy rain in the Thames/Coromandel area. Orography plays a major part in these events, and in the most severe, convection is also a factor.

1907 10-23 January

Ngaruawahia: about125mm/24 hrs (return period 20 years)
Hamilton: 95mm/24hrs (return period 15 years)
Stock losses: Heavy.

A long spell of wet weather led to the biggest flood of the 20th century for the Waikato River. The main railway south from Auckland was flooded south of Pokeno to depths up to 4.5 metres. Train passengers were transferred to boats there and taken to Taupiri to catch another train. Refreshments were served on the way at Mercer station, although the station floor was under 30cm of water.

Vast areas of land were flooded, and crop and stock losses were heavy. The wettest period was from the 14th to the 16th, when a northerly flow between a high to the east and a slow-moving trough in the Tasman Sea brought a prolonged spell of rain. MSL analyses for the 14th and 15th (not shown) reveal a slow moving depression with central pressure about 990 hPa lying just west of the North Island.


See Ormond Valley a Sea of Water

1953 7 July
SINCE 1907

152mm/4 days (return period 15 years)
Taupo: 139mm/4 days (return period 15 years)
incl. 89mm/18 hours. (return period 15 years)

Although rainfall amounts were similar to the flood of April 1924, the flooding in 1953 was much more serious. About 10,000 hectares of developed and partly developed farmland along the banks of the Waikato River was inundated. About 2000 hectares of first-class dairying land was under water for two to three weeks, and considerable areas were affected for longer periods. The main highway was covered at many points, and in one place was impassable for several days.
However, river levels were well below those recorded in 1907.

Interestingly, it was not a northwest flow that brought the heavy rain. A rain band remained slow moving over the Waikato area for some time. Fig 21 shows the sequence of MSL charts for the period of heaviest rainfall. During this period, the low-level flow over most of the North Island was from the northeast.
Fig 22 shows that, as indeed might be expected, heavier falls occurred in the Bay of Plenty.
Fig 23 shows the system was cut off to a high level in the atmosphere, and thus can be expected to have been very slow moving.

wko wko2 wko3
Fig 21. Sequence of MSL charts for the
period of heaviest rainfall.
Fig 22. Four-day rainfalls (inches)
9am July 3rd to 9am July 7th, 1953.
Fig 23. 500 hPa analysis for
4pm 4th July 1953.
(Heights in hundreds of feet)

1954 18-20th May

Waihi: 232mm/48hrs (return period 8 years)
Te Aroha: 467mm/72 hrs (return period over 150
Paeroa: 297mm/48 hrs (return period 45 years)

Parts of Paeroa were evacuated during this flood event and the Ohinemuri River rose seven metres. Low-lying areas of Thames were flooded to a depth of two metres, with the Waihou River rising nine metres. Although significant flood events occurred in June 1920 and April 1923, this was reported to be the worst flood in 60 years. In actuality, it was probably not as bad as that of 1910 (only 44 years before).
Fig 24 shows the classic set-up for heavy rain on the Coromandel Peninsula. The low remained virtually stationary, while its associated frontal band drifted slowly southwards over the affected area.

cop Fig 24. MSL analysis for 6am 18th May 1954.

1958 15-25th February

Rangipo Prison Farm: 305mm/2 days (return period over 150 years)
The second worst flood of the 20th century covered thousands of hectares of farmland, drowned large numbers of livestock, and flooded hundreds of houses in Otorohanga, Te Kuiti and Huntly. The floods were the culmination of a very wet month, with a small area around Hamilton having the wettest ever month in over 50 years of observations.
On the 14th, a northeast flow spread onto the area, while a slow moving depression lay in the northeast Tasman Sea. A frontal band crossed Waikato on the 18th (Fig 25), and the low crossed the North Island on the 19th. A temporary clearance was brought by a weak ridge on the 20th.
wkob Fig 25. MSL analysis for 6am 18th February, 1958.
However, with the approach of a deep and very active trough from the Tasman Sea (Fig 26), general rain soon developed again.
wkob2 Fig 26. MSL analysis for midday 22nd February, 1958.
A frontal band became slow moving over Waikato for more than a day, before being pushed through by the following front (Figs 27 and 28).
wkob3 Fig 27. MSL analysis for 6am 23rd February, 1958.
wkob4 Fig 28. MSL analysis for 6am 24th February, 1958.
The clearance occurred on the afternoon of the 24th. An intense high lay east of the Chathams throughout the period. There was some evidence of blocking in the 500 hPa charts, (Fig 29), but the latter part of the event was dominated by a sharp 500hPa trough (Fig 30).
wkob5 wkob6
Fig 29. 500 hPa analysis for midday, 15th February, 1958. Fig 30. 500 hPa analysis for midday, 24th February, 1958.

1981 12-14 April

Waihi: 521mm/72 hrs (return period over 150 years)
Paeroa: 346.5mm/48 hrs (return period 90 years)
Kauaranga Catchment: 870mm/72 hours (return period over 150 years)
Evacuations: About 1000.
Damage: $30 million
2012 Dollars: 119 million.
After three days of heavy rain, many homes in Paeroa were flooded to street level when the Ohinemuri River broke its banks. In the small town of Waikino five buildings were swept away. 750 people were evacuated in Paeroa, and 200 in Thames. Gale force winds affected many coastal areas, with houses losing roofs, and power lines blown down. This was another three-day event, and although rainfalls in some areas were lower than in 1954, this flood exceeded that of 1954 in severity and areal distribution of heavy rainfalls. This therefore made it the worst flood on record for the area to date.
The weather situation for the period of the storm was rather typical for heavy rainfall events for this area. However, the event was more severe and longer lasting than normal.
Blocking was a major contributing factor. As can be seen from Fig 31, a very large 500 hPa low developed.
waihi Fig 31. 500 hPa analysis for midnight 12th April 1981
This low proved very difficult to shift. The surface low, while not particularly deep at any stage, was prevented from moving southwards by an anticyclone which moved slowly across the southern South Island. There was another high to the east of the country, entraining tropical air onto the Coromandel region. Fig 32 shows the MSL situation at 6am 12th April, 1981.
waihi2 Fig 32. MSL analysis for 6am 12th April 1981.
Fig 33, a day later, shows that while the high centre has moved from the Tasman Sea to the area southeast of the South Island, the low has not moved much at all.
waihi3 Fig 33. MSL analysis for 6am 13th April 1981.
A ship eastnortheast of Coromandel at this time gave an indication of the high dew point air (21C) that was being entrained onto the affected area.
Cooler air lay behind the frontal band, typefied by a dew-point of only 15C at Auckland.
The rain band associated with the depression was particularly broad and active, as can be seen from Fig 34.
waihi4 Fig 34. NOAA 6 infrared image for 6am 12th April 1981.

1985 16-17 February

Coromandel 331mm/24hrs (return period 85 years)
incl 266mm/8hrs (return period over 150 years)
Thames 251mm/24 hrs (return period 150 years)
Te Aroha 127mm/6hrs. (return period 100 years)

Pukekohe 166mm/24hrs (return period 50 years)

Deaths: 4
Damage: $7 million
2012 Dollars: 19 million
Evacuations: Hundreds

In the early hours of 17th February an avalanche of boulders, logs and sludge swept through the main streets of Te Aroha. One home was completely destroyed, killing a woman and two of her children. Nearly every shop and about 50 houses were damaged. In Thames, 200 homes and more than 30 businesses were flooded, and parts of the road lay under one metre of water. Further north in Waioumu, landslides resulted in another death.
MSL charts in Fig 35 and 36, shown for midday 15th and midday 17th February, suggest that the flooding resulted from the northeast flow and frontal passage
tearoha tearoha2
Fig 35. MSL analysis for midday 15th February, 1985. Fig 36. MSL analysis for midday 17th February 1985.
characteristic of such events in this area. However, the fact that gradients were not unusually strong indicates that orographic effects were no stronger than in many more ordinary heavy rain events.
Thus there must have been other factors involved.
It was in fact convection that made this such an extreme event.
This is well illustrated in Fig 37, which clearly shows that the heaviest rainfalls were on the western side of the Coromandel Peninsula, and further downstream, in the flat South Auckland area.
tearoha3 tearoha4
Fig 37. Rainfalls and isohyets (mm) for the 24 hours up to 9am 17th February 1985. Fig 38. NOAA 9 infrared image at 3am 17th February 1985.
This is a characteristic of an unstable spillover situation, where, while orography plays its part, maximum rainfalls occur on the lee side of the ranges.
The infrared photograph in Fig 38 shows that cloud top temperatures over the area were colder than minus 60C, indicating tops reaching 12,000 metres.
The highest tops, while extending westwards into the Tasman Sea, covered quite a small portion of the frontal band, and the sounding at Auckland Airport at midday on the 16th (Fig 39) was unlikely to have been representative of the areas most affected.
This is typical of convective events in that they often affect small areas.
There is some evidence to suggest that low shear was also a factor in this event.
Fig 40 shows that the depression was cut off to above 250 hPa, with – unusually – two closed contours appearing at this level.
Southerly winds of over 100 knots lay to the west of the upper low. This pattern also indicates the slow moving nature of the system, although the duration of the storm was less than either the 1954 or 1981 events.
tearokap tearoha5
Fig 39. Sounding from Auckland Airport for midday 16th February 1985. Fig 40. 250 hPa analysis for midday 16th February, 1985.
This flood affected different areas than in 1981. Apart from Te Aroha, which was severely affected, the worst damage extended from Kopu to Tapu, on either side of Thames. The fact that the rain fell in a shorter time period than in 1981 makes a comparison based solely on flood damage difficult. What can be said is that from Coromandel to Thames rainfalls were the highest on record, and 24-hour totals in this area were in excess of the three-day accumulations in the 1981 event. In some areas there were also extreme rainfalls for durations less than 24 hours, particularly at Te Aroha, which received 127mm in six hours. The 266mm in 8 hours at Coromandel where, amazingly, no major flooding occurred, has a return period of over 150 years.

1996 29-31st December

242mm/24 hrs (return period 30 years)
Pinnacles: 426mm/24 hrs (return period 40 years)
Golden Cross: 372mm/24 hrs (return period 55 years)
Tamahunga (Auckland Region): 160mm/24hrs (return period 10 years)
Damage to roads: $2 million.
2012 Dollars: 2.9 million
The remains of Cyclone Fergus struck during the holiday period. Hundreds of holiday makers were trapped in the area by floodwaters, and forced to seek shelter
in Civil Defence centres. Water supplies in Pauanui and Whitianga were cut off, and a state of emergency was declared for the area. Heavy rain with extensive
flooding also affected Northland, with up to 409mm recorded for the event there.
The centre of the low passed across the Bay of Plenty overnight 30-31st, but it was the moist east to northeast flow on its southern flank that brought the torrential rain. Pinnacles station in the Kauaeranga catchment began continuous records in 1991, and the 426mm recorded there in a 24-hour period between the 29th and 30th is the highest recorded 24-hour rainfall to date for any place in the Coromandel. There were some extremely high intensities in the last six hours of the event.
Golden Cross (established 1990) also recorded its highest ever 24-hour rainfall in this event – 372mm.
Fig 41 shows the MSL situation at midday on the 30th. Wet bulb potential temperatures were up to 23C over northern New Zealand.
Rainfall forecasts were slightly underestimated for the Coromandel/Kaimai ranges. Northland reported an isolated rainfall far above that forecast, but many areas had falls far below the forecast. There was no significant influx of cold air into the cyclone, and there was no convection within the low when it moved onto New Zealand. However, places near where the centre made landfall (between Whakatane and Hicks Bay) experienced some wind damage, suggesting that there was a remnant of the core present.
fergus Fig 41. MSL analysis for midday 30th December 1996.

1998 1-20 July


Matangitangi 203mm/72 hrs to 7.40pm 11th (return period 42 years)
Ngaroma 271mm/72 hrs to 4.55pm 11th. (return period 22 years)
incl 188mm/24 hrs (return period 7 years)

Bay of Plenty:

Ranger Stn 391mm/72hrs (return period 120 years)
incl 326mm/48hrs (return period 80 years)

Evacuations: 30 houses
Damage: $17.85 million
2010 Dollars: 26.2 million

Episode 1

A low deepened rapidly in the Tasman Sea. The low was triggered by strong CVA, also very strong upper level divergence in the poleward exit region to a 180-knot
southwest jet and the entrance to an anticyclonically-curved northwest jet (not shown).
Very moist air, with wet bulb potential temperature (WBPT) around 17C was brought out of the subtropics, and 850 hPa winds strengthened to 50-60 knots.
The MSL chart for midnight 1st July is shown in Fig 42.
The high slowed the movement of the cold front as it crossed the North Island.
Rainfall rates reached 60 mm per hour in the Bay of Plenty.
Forecast amounts were good for the area around Taupo, but underestimated elsewhere.

weeks Fig 42. MSL analysis for midnight 1st July, 1998.

Episode 2

Strong north to northwest winds spread over the North Island ahead of an active cold front. A complex low lay in the Tasman Sea.
Fig 43 shows the mean sea level (MSL) analysis for midday on the 9th.
The development of a small low, seen west of Taranaki in Fig 43, occurred beneath strongly difluent flow (Fig 44), and the resulting increased gradient
enhanced the orographic contribution to the heavy rain.
The front subsequently weakened over the eastern Bay of Plenty.

weeks2 weeks3
Fig 43. MSL analysis for midday 9th July, 1998. Fig 44. 250 hPa windbarb analysis for midnight 8th July, 1998.
Thin lines are isotachs (knots), and divergence is shaded.

Episode 3

A wave can be seen in the midday 10th July analysis over northern New Zealand (Fig 45).
The imagery revealed a large baroclinic leaf (not shown), which rapidly developed, then spread rain back over the whole of the northern half of the North Island. Although rain from this episode may not have been quite up to the criteria required for a warning, it was certainly of significance, coming as it did on the top of the previous heavy rain.
Computer guidance had led forecasters to expect a further burst of rain in the eastern Bay of Plenty only, and not for the wave to come back as far west as it did.
There was in fact a fourth burst of heavy rain on the 14/15th July. Another 50-60mm of rain fell over the affected areas in an easterly airstream.

weeks4 Fig 45. MSL analysis for midday 10th July, 1998.

2002 21 June

125mm/25mins (return period well over 150 years)
Thorndon Bay: 150mm/2hrs (return period well over 150 years)
Tapu: 83mm/1 hour (return period over 150 years)
Waiomu: 31mm/15 mins (return period 140 years)
Te Aroha: 97mm/1 hour (return period over 150
Putaruru: 120mm/2 hrs (return period well over 150 years)

Deaths: One.
Evacuations: Many
Damage: $13 million.
2012 dollars: 17 million.

A camper was swept out to sea and drowned, and 356 houses were inundated, when an explosively deepening depression passed to the west of the Coromandel Peninsula. States of civil emergency were declared in Thames and Putaruru. An unofficial report of 125mm in 25 minutes was
received from Coromandel Township. The world record for a 20-minute rainfall is around 200mm, recorded at a location in Romania.
Fig 46 shows the rapid development of the low, which was well predicted by computer models. However, what was not predicted so well was the strong convection affecting a small part of the Coromandel Peninsula, causing extreme damage.
A sharp upper trough moved across the north Tasman Sea with negative tilt near northern New Zealand. Subtropical air moved down the moist conveyor into the low, overriding cool air from around the high to the south.
bomb bomb2
Fig 46. MSL analyses for midnight 19th June 2002 (top left); noon 20th June 2002 (top right) midnight 20th June 2002 (right) bomb3

The 250 hPa analysis (Fig 47) shows a strongly difluent flow over
the Coromandel Peninsula.

Fig 47. 250 hPa analysis for midnight 20th June 2002.

2003 25-28th February

See Heaviest Rain Since Bola in Some Places

2003 26th November

Hamilton: 55.6mm/1 hr to 6pm 26th
(return period 110 years)
A downpour over Hamilton caused flooding. The cause was strong convection within a very warm, moist airmass behind a frontal band. Models had forecast the instability, indicating CAPE in the 1100-1200 j/kg range. In fact, CAPE reached around 1400 j/kg. Storm motion was high at about 20kt, but a line of thunderstorms developed, with multi-cell storms tracking over the same area.
Because the heavy rain affected only a small area, no warning was required.
However, the severe convection forecasters, working on a trial basis, had been expecting a high risk of thunderstorms, with heavy rain in the range of 10-25mm/hr.

2004 28-29th February

Ngaroma: 218mm/36hrs (return period 80 years)
incl 183.5mm/24hr (return period 55 years)
Mangatoetoe: 289mm/36hr (return period over 150 years)

Evacuations: 90.
Damage: $4 million.
2012 Dollars: 5 million.

The Tongariro River burst its banks and 100 people in Turangi were evacuated. The return period for the flow in the Tongariro and Upper Waipa Rivers was about 100 years, although only 20 years for the Lower Waikato. Flooding occurred in many other parts of the Waikato Region, particularly Otorohanga, and low-lying areas adjacent to the lower Waipa and Waikato Rivers, with roads closed and farms inundated. Elsewhere in the country, rivers in Manawatu and Whanganui came close to bursting their banks. SH1 just south of Kaitaia was closed by flooding. There was surface flooding around Auckland (where there were also trees blown down), in Hastings, on the Kapiti Coast, and in the Hutt Valley. Flooding was exacerbated in many areas by the already wet catchments, after a stormy month.
The area of heaviest rain was in the upper reaches of the Tongariro and Waipa Rivers, where intensities reached 27mm per hour.

A deep low over the Tasman Sea moved slowly southeastwards, crossing the South Island on the night of the 28th. Its associated front was preceded by a very moist, strong northerly flow. The analysis for midday 28th February is shown in Fig 48. The front moved slowly across central and northern New Zealand, and linked up with the remains of Tropical Cyclone Ivy, which moved rapidly southsoutheast passing just east of East Cape early on the afternoon of the 29th. The presence of Ivy ensured that very humid tropical air was brought down onto the country – this is what turned the event into an extreme one in a small area. Orographics played a major role in the event, with upper level divergence also a factor.

turangi Fig 48. MSL analysis for midday, 28th February, 2004

2005 18-19th May

See Matata and Tauranga Swamped

2005 16-17th July

Pauanui: 313mm/36hrs (return period 50 years)
incl 215mm/12hrs (return period 65 years)
There was extensive flooding on the Coromandel Peninsula around Tairua and Pauanui with land under water and roads cut. Thousands of people were trapped in the area by a culvert blow-out. 309mm was reported to have fallen overnight at Tairua – a rainfall of this magnitude has a return period of over 150 years.
A depression lay west of the North Island. An associated frontal band extending east-west from the low moved slowly south, with its progress slowed by a blocking high between the South Island and the Chathams (Fig 49).
Low-level convergence within a strong easterly flow affected the Coromandel, and rainfall was enhanced by upper divergence (Fig 50) and CVA (Fig 51)
second second2
Fig 49. MSL analysis for midnight, 16th July, 2005.
Shaded area: 700 hPa humidity.
Dotted lines: 850 hPa temperature.
Fig 50. 250 hPa windbarbs for midnight, 16th July, 2005.
Shaded area: Divergence.
Solid lines: Isotachs.

within a short wave trough moving southeast over the area. Convection was a major contributor to the rainfall. No useful hourly falls are available for the Coromandel area, but parts of Auckland received up to 45mm in an hour, and several properties were flooded. There was also flooding and road closures in Northland.

The convection was the key to this event. The active portion of the front was quite small, but tracked right down over Northland, Auckland and Coromandel. Forecast soundings did not produce much CAPE, and actual soundings were unspectacular. The only clue to impending heavy rain was given by CHAMP12. This often
produces spurious heavy rain from low-level convergence zones, but can be useful as a “first alarm” level for convection forecasters. This needs to be done judiciously though (for example, “blobs” of high rainfall in inland areas on the CHAMP fields can usually be ignored in winter, where afternoon convection is not an issue and there is no warm sea close by to act as a moisture source).


fl3 Fig 51. 500 hPa analysis for midnight, 16th July, 2005.
Shaded area: Upward motion.
Dashed lines: Geostrophic vorticity.


Collen, B and J.W.D. Hessell, 1981: The Thames-Coromandel Floods of April 1981. New Zealand Meteorological Service Technical Information Circular 183.
Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
Finkelstein, J., 1954: The Waikato Flood of July 1953. New Zealand Meteorological Service Tech. Note 108.
Holmes, V, and S.M. Burgess, 1985: The Thames-Coromandel Floods of February 1985. New Zealand Meteorological Service Scientific Report 17.
McGavin, T. : Severe Weather Events Analysis. New Zealand Meteorological Service
Munro, A.J., and M.G. Chapman, 1998: Waikato Regional Flood Event of 9-20 July 1998. Environment Waikato Technical Report 1998/15.
New Zealand Gazette, 1958, Vol. I. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Severe Weather Log. New Zealand Meteorological Service.
White, Ron, 2003: Weather Bomb – The Coromandel Experience. Tephra June, 2003, Vol. 20.

Bay of Plenty

Floods are quite common in the Bay of Plenty, and can be very severe. This is because of the region’s exposure to the north – air arriving from the north tends to be warmer than air coming from other directions, and warm air can carry more moisture than cold air. Heavy rain events in this area generally occur
in strong pre-frontal northerly airstreams. If the front stalls over the Bay of Plenty, very large accumulations can ensue.
Convection is often a factor in heavy rain events. Such events do not need a northerly fetch to produce flooding, but in these cases the flooding is localised.

1944 22 February


Rotorua: 146mm/24 hrs (return period 10 years)
(mostly in three hours) (return period over 150 years)
East of Rotorua: about 380mm/3 days
Edgecumbe-Urewera strip: Up to 240mm/24 hrs

Stock Losses: Considerable.

The cause of this major event was described by Kerr (1944) as a “cyclone of tropical origin”.
This does not necessarily mean it was ever a “true” tropical cyclone – many ordinary lows develop in the tropics and then develop as they move to higher
Fig 52, for 6am 22nd February 1944, illustrates a classical heavy rain situation for the Bay of Plenty.
The blocking high to the east of the country prevented the eastward movement of the front across the Bay of Plenty. Heaviest rain occurred within the front, and in the strong northnortheast flow on its eastern side.
disas Fig 52. MSL analysis for 6am 22nd February, 1944.

1948 17-18th April

Tauranga: 230mm/24hrs (return period 95 years)
incl 221mm/12 hrs (return period over 150 years)
and 212mm/6 hours (return period over 150 years)
and 34mm/10 minutes (return period over 150 years)

Opotiki: 182mm/12hrs (return period over 150 years)
incl 145mm/6 hours (return period over 150 years)

Official Records
Period Amount(mm) Site Date
10 minutes 34 Tauranga 17 April 1948
1 hour 134 Cropp Waterfall 8 January 2004
12 hours 566 Hokitika at Prices Flat 11-12 May 1978
24 hours 758 Cropp Waterfall 27-28 December 1989
48 hours 1049 Cropp Waterfall 11-13 December 1995
1 calendar month 2927 Cropp Waterfall December 1995

1964 9-11th March

Mt. Pirongia Makeokeo:
277mm/48hrs to 9am 11th (return period 150 years)

This flood covered most of the Opotiki floodplain and filled nearly all Opotiki’s 650 homes with muddy water. The Whakatane River was also affected. Up to 300mm of rain fell in three days.
A depression which had formed in the north Tasman Sea moved onto Taranaki and combined with another centre near Southland. A frontal band associated with the
first low hesitated as it crossed the North Island. It was finally pushed through by the front associated with the second low. Fig 53 shows the MSL situation at midday 10th March.

The 500 hPa chart for this time (Fig 54) shows a sharp trough in the Tasman Sea. The rain finally eased after both fronts passed across the Bay of
Plenty during the afternoon of the 11th.

muddy muddy2
Fig 53. MSL analysis for midday 10th March, 1964. Fig 54. 500 hPa analysis for midday 10th March, 1964.

1970 11-14th August

Te Teko: 295mm/48hrs to 9am 15th (return period 120 years)
incl 262mm/24hrs to 9am 14th (return period over 150 years)
Lake Tarawera: 200mm/24hrs to 9am 15th (return period 95 years)

Evacuations: 100

whak whak2
Fig 55. MSL analysis for midday, 12th August, 1970. Fig 56. 500 hPa analysis for midday, 12th August, 1970.

Roads were closed, and the Rangitaiki River burst its banks at Te Teko and Edgecumbe.
An intense anticyclone lay east of the country, while a broad, complex trough covered the Tasman Sea (Fig 55).

The 500 hPa flow at this time showed a blocking signature (Fig 56).
The trough moved slowly onto New Zealand, and a strong, moist northerly airstream affected Bay of Plenty for a considerable time. The low seen near Northland at midnight on the 13th (Fig 57) subsequently moved southeastwards into the Bay of Plenty, bringing a burst of enhanced rainfall.

whak3 Fig 57. MSL analysis for midnight,13th August, 1970.

1979 21st March

Te Puke: 375mm/48hrs (return period over 150 years)
Evacuations: Dozens
A train was derailed between Tauranga and Mount Maunganui. Dozens of people were evacuated, with one home destroyed and others damaged. Worst hit was Te Puke, which lost its water supply. SH2 between Te Puke and Tauranga was closed to all but heavy traffic.
Fig 58 shows a low in the Tasman Sea at midday 20th March, with a northeast flow onto the Bay of Plenty. Twenty-four hours later (Fig 59), not much had changed.
derail derail2
Fig 58. MSL analysis for midday 20th March, 1979. Fig 59. MSL analysis for midday, 21st March, 1979.
Factors contributing towards this being an extreme event were:
1. The slow moving nature of the low. Fig. 60 shows the classical omega block pattern east of New Zealand.
2. Upper-level CVA and divergence over Bay of Plenty. The area lay in the poleward exit region to a weak jet (not shown).
3. There is also a suggestion in Fig. 59 of some low level convergence.
4. Reports of thunderstorms at Hamilton at the same time suggest that convection may have played a part as well.
fig60 Fig 60. 500 hPa analysis for midday, 21st March, 1979.


See Flooding Lasts For Weeks

1999 30th April-2nd May

Rotorua: 214mm/24hrs (return period 80 years)
incl 169mm/5hrs (return period over 150 years)
and 47mm/1 hour (return period 35 years)
The first ever cancellation of the Fletcher Marathon resulted from rainfalls exceeding 150-year return periods. Homes and streets were flooded. In Northland and Auckland homes were flooded also, and roads and schools closed. Some places in Northland recorded 120mm in three and a half hours.
The culprit was a frontal band associated with a large low in the Tasman Sea. Fig 61 shows the MSL analysis for midday 1st May.
The front was preceded by a very moist, unstable northeasterly airstream. The heavy rain developed in a convergence zone between this northeasterly and the northnorthwest flow behind the front.
Fairly strong upper-level divergence lay over the area in a diffluent region ahead of a short wave trough and south of the subtropical jet. The air was potentially unstable, and shears in the middle layer of the atmosphere were low. Rear-feeder cumulonimbus clouds were most active over Rotorua as new cells formed over Tauranga. The slow-moving nature of the front was a factor in the large rainfalls.
marathon Fig 61. MSL analysis for noon 1st May 1999.

2003 6th April

Ranger Stn: 175.5mm/6hrs (return period over 150 years)
Te Puke: 156mm/6hrs (return period over 150 years)

The Harbour Board had their dredge in the Whakatane River washed out to sea. Frontal activity was generated primarily by strong upper level divergence
associated with a strong northwest jet. (Fig. 62)
However, surface gradients were weak, as can be seen from Fig 63. Also, the front was expected to move quickly across the area.
dredge dredge2
Fig 62. 250 hPa analysis for midday, 6th April, 2003. Fig. 63. MSL analysis for midday, 6th April, 2003.
Indeed, Fig 64 suggests that at midday the event was just about all over for the Bay of Plenty, but at this point the front stalled. Heavy rain lasted until about 9pm, with rates reaching 35mm per hour.
A watch was issued for this event, but was never upgraded to a full warning.
dredge3 Fig 64. GMS satellite image for midday, 6th April, 2003.

2004 17-19th July

249mm/2 days (return period 100 years)
257mm/3 days (return period 40 years)
Much of the Rangitaiki Plains and parts of Whakatane township were submerged in this event. A stopbank breach flooded parts of Edgecumbe and surrounding farmland. The Whakatane River exceeded its hundred-year flow by a considerable margin, while it was almost exactly a hundred-year flood in the Rangitaiki.
The heaviest falls were in fact on the coast around Whakatane. As with the year 2000 event in Tauranga, the cause was a low level convergence zone stalling over the area.
The event started out typically enough, with a frontal band approaching from the west, preceded by a moist northerly flow. But perhaps it was a sign of things to come that heavy rain started in the eastern Bay of Plenty well before the front arrived.The front then became slow-moving over the Bay of Plenty, blocked by a high to the east (Fig 65.) Meanwhile, to the north, a subtropical low was developing. A broad trough developed, extending from the central North Island northwards, with a second shallow low centre developing in the Waikato area. (Fig 66)
100 1001
Fig 65. MSL analysis for midnight 15th July 2004. Fig 66. MSL analysis for midnight 16th July 2004.


The sub-tropical low moving southsoutheastwards gradually became the dominant feature, and the flow over the Bay of Plenty gradually turned to a light easterly
(Fig 67). During this time a narrow band of heavy rain persisted over the eastern Bay of Plenty. Even after the northerlies had gone, a zone of low-level convergence persisted in the eastern Bay of Plenty, between northeasterlies and southeasterlies.
Aloft, a major cut-off low lay slowmoving to the west of central New Zealand. A forward-tilting trough, with its associated CVA, moved around the low onto the North Island during the 17th (Fig 68), with the cut-off low then crossing the North Island on the 18th.

1003 1004
Fig 67. MSL analysis for midnight 17th July 2004. Fig 68. 500 hPa analysis for midday 17th July, 2004
Initial heavy rain warnings concentrated on the western Bay of Plenty, and underestimated the amount of rain in the east.
Computer models did not handle the situation well at all. They expected the development of the subtropical low to the north to push the front back westwards, where it was expected to stall and bring prolonged rain between Northland and the western Bay of Plenty. A study of this situation revealed some deficiencies in the way the model chooses to incorporate or reject incoming data (Nabi, 2005). Heavy easterly rain was later expected to spread down the North Island east coast and into Wellington. No significantly heavy rain occurred in any of these other places.

2005 3rd May

Tauranga AWS: 67mm/1 hour 8-9pm (return period 130 years)
Rotorua AWS: 40.6mm/1 hr (return period 80 years)
A deluge struck Tauranga in the evening, causing extensive flooding in the city. The rain was accompanied by an impressive lightning display. Active convection occurred on a convergence line moving eastwards within a large low-pressure area (see Fig 69). Subtropical air was being advected onto the Bay of Plenty ahead of the line of convergence.
Aloft, a cut-off low lay over northern New Zealand (not shown). Earlier in the day, an active frontal band had given up to 50mm in an hour to parts of Waikato, including Hamilton. Around mid-afternoon, Rotorua AWS had recorded 40.6mm in one hour.
Forecast soundings were unspectacular.
deluge Fig 69. MSL analysis for 6pm 3rd May, 2005.

2005 18th May

Bay of Plenty:
Tauranga AWS 347mm/24hrs (return period over 150 years)
incl 217mm/12hrs (return period about 150 years)
and 133mm/3hrs (return period over 150 years)
Awakaponga 95mm/1hr (return period over 150 years)
Hikuai about 900mm/72 hours (return period over 150 years)

Evacuations: About 500.

Several rain storms hit the region, destroying houses, swamping farms, and forcing hundreds from their homes. Worst hit were Tauranga and the small town of Matata. At Matata a torrential downpour in the hills (close to Awakaponga) turned small streams into raging torrents which swept down carrying logs, huge boulders and river silt. Houses and vehicles were buried under up to 5 metres of silt, and cars were left bobbing in the lagoon. Residents ran for their lives.
On the Coromandel Peninsula eastern areas were cut off, with roads rendered impassable, and about five homes in Whangamata were inundated. Rainfall was mostly coastal. It is thought that hourly intensities exceeded 100mm per hour at Hikuai, Pauanui, Whangamata and Opoutere, as well as at the high-level Pinnacles gauge, with return periods in excess of 150 years.
A shallow low brought warm, moist, unstable air down onto the Bay of Plenty (Fig 70).
Convergence zones within this airstream produced high hourly intensities. One of these stalled over Tauranga for a time. The sun highlights the cumulonimbus tops over Tauranga in this image (Fig 71).
It is likely that another convergence zone produced the downpour in the hills above Matata that night.
matata matata2
Fig 70. MSL analysis for 9am 18th May, 2005.
The heavy rain occurred in the broad convergence area between the northerlies over Northland, and the northeasterlies over East Cape.
Fig 71. Visible satellite image taken midday 18th May, 2005.
Model guidance failed to predict the southward movement of the warmest air, and each run differed in its positioning of the moisture convergence zones.
Fig 72 shows that the air in the vicinity did in fact become very unstable. The CAPE in this sounding is 1728, with LI –5.8, and storm motion 14kt.
matata3 matata4
Fig 72. Whenuapai sounding for midday 18th May 2005. Matata after the 18th May 2005 flood. (Whakatane Beacon)


Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Kerr, I.S., 1944: Synoptic Situations Associated with Flood Producing Rains in New Zealand. New Zealand Meteorological Service Circular Note 33.
McGavin, T. : Severe Weather Events Analysis. New Zealand Meteorological Service
Nabi, M., Using the MM5/3DVAR Scheme to Investigate the Impact of Model Forecasts: A Case Study. New Zealand Meteorological Service Student Projects Series.
New Zealand Gazette, 1964 Vol. I. New Zealand Government.
New Zealand Herald 20-22nd March, 1979
Severe Weather Log. New Zealand Meteorological Service.

Gisborne/Hawkes Bay

Easterly airstreams ahead of active fronts, or accompanied by slow-moving fronts, are the most common situations producing heavy rain in the east of the North Island. The most severe events include some convective element. Strong moist easterlies associated with ex-tropical cyclones can also produce heavy rain, the most notable example in recent years being Cyclone Bola (1988).
This area takes the top slot in the country in the extreme events hierarchy at the end of this publication (Appendix 1). The hierarchy is based on the multiplication factor by which the maximum rainfall for the event exceeds its 150-year return period. Using this criterion, the events of 1924, February 1938, 1953, April 1938, 1910, and 1941, all of which occurred in the Gisborne/Hawkes Bay area, take six of the slots in the top ten of heavy rain events in recorded history in New Zealand.
Also of interest is that not one of the above-mentioned events occurred in the last 50 years.

1867 25th May-4th June

Napier: 381mm/4 days (return period over 150 years)

Stock Losses: Heavy

Several feet of floodwater flowed through Napier, and silt to a depth of 30-50cm was deposited. Land that had not been flooded within the memory of the oldest inhabitant was under water. After initial heavy rain fine weather followed, but on 3rd June rain recommenced, and continued on the following day.

1893 4th December

The Waipawa River reached its highest known levels. The Tukituki River broke its banks and swept over Clive. Papakura, Meeanee and Taradale were also inundated.

1897 14-17th April

Hawkes Bay:
Tikowhai 507mm/96hrs
Raetihi 160mm/96hrs
Ruanui 257mm/96hrs
incl 178mm/24hrs (return period over 150 years)

Deaths: 12
Stock losses: Hundreds of thousands
Damage: 150,000 pounds
2012 Dollars: 27 million

Hundreds of families homeless.

“A terrible calamity” was how The Hawkes Bay Herald described this event. Flooding extended from Wairoa to Woodville. It was estimated that three-fifths of the Heretaunga Plains were under water, with roads and railway washed away.
Clive, about 10km south of Napier, was worst affected. Only the tops of the houses were visible in some parts of the town, and in some instances houses were washed out to sea.
Of the people drowned, ten were in rescue boats that had also been swept out to sea.

It was reported that “Eleven and a half inches of rain (290mm) fell in 24 hours”, presumably in the Napier area. Another report gave the Napier rainfall as 355mm in four days. Inland, it was reported that 21 inches (533mm) fell in 30 hours in parts of the higher reaches of the Tukituki, Ngaruroro and Tutaekuri Rivers.
Heavy rain began in inland areas at first, soon spreading to the coast.
Extreme flooding also occurred on the Rangitikei River. In fact, this remains the largest flood in the Rangitikei in European history, with a return period for the flooding of maybe as much as 500 years. All six bridges across the river were destroyed, and settlers in the Lower Rangitikei had several feet of water through their houses. Near the river 100 acres of green forest, including a stand of three-hundred-year-old totara trees, was swept away.
The Whangaehu River also had an extremely high flood, as did the Manawatu River.
The Whanganui River also received a big flood, and southwesterly gales were reported by the Whanganui Herald.
The severe weather was caused by a depression with central pressure around 990 hPa, moving slowly eastwards across the North Island. There is evidence to suggest strong similarities between this event and that which caused the disastrous floods in Manawatu in February 2004. Heavy rain in Hawkes Bay and Wairarapa suggests a flow from the easterly quarter, yet Whanganui reported southwest gales. This is strongly suggestive of convergence between southeast and southwest flows, as occurred in the 2004 event.

sea Photograph of the flood damaged Waitangi railway
bridge at Clive, 1897.
Permission of the Alexander Turnbull Library, Wellington, must be obtained before any re-use of this image.

1906 14-17th July

Deaths: Two.
This flood was one of the heaviest in the history of the district. Two people were drowned at Makauri, and Julius Caesar had the irritation of three metres of water in his paddocks. However, snow melt was a major contributing factor, and no remarkable rainfalls were recorded.

1910 30th March-1st April

Gisborne City 365mm/3 days (return period over 150 years)
incl 195mm/24 hrs (return period 70 years)
Te Rata, Whatatutu 584mm/60 hrs (return period well over 150 years)
Coromandel Peninsula: Waihi 309mm/12hrs 30min (return period over 150 years)
The whole of Makauri, on the outskirts of Gisborne City, was flooded, and the Ormond Valley was practically a sea of water, with many houses having 1-2 metres of water in them. At Whatatutu, 584mm fell in 60 hours – a phenomenal rainfall for this area.
Even so, flooding was nowhere near as serious as in 1906.
In the Coromandel area, Paeroa and surrounding land was flooded to an extent previously unknown. Large volumes of mining tailings and slimes were deposited on fertile lands.
The culprit was an ex-tropical cyclone, with central pressure just below 980 hPa, moving southeastwards to pass near Cape Reinga and just northeast of East Cape.

1917 13th June

Napier: 254mm/3days (return period 70 years)
incl 187mm/36hrs (return period 45 years)
Morere: 522mm/4days (return period 60 years)
incl 319mm/24hrs (return period 80 years)
Tutira: 511mm/4days (return period 110 years)
incl 427mm/48hrs (return period 130 years)
This flood was said to have been bigger than that in 1897, and nearly as big as the 1867 event. The Tutaekuri River rose 15cm higher than in 1897. Several small
towns were flooded.
The main synoptic feature was a low, with central pressure around 990 hPa, which moved very slowly eastwards across the far north of the North Island.

1924 11th March

Riverbank, Rissington: 512mm/10hrs to 5.30pm 11th (return period well over 150 years)
incl 229mm/2.75hrs to 11.45am 11th (return period well over 150 years)
Tutira: 203mm/3hrs to 10am 11th. (return period well over 150 years)

Deaths: One.
Stock losses: Heavy.

This phenomenal downpour, in Hawkes Bay’s Esk Valley, resulted from a clash of airmasses, with warm air from a depression meeting a surge of cold air to the south. (See Fig 73.) Widespread thunderstorms were occurring within the warm airmass, suggesting that the air was already unusually unstable. This warm air became trapped in Hawkes Bay, then becoming undercut by a cold southerly, leading to violent convection.

Convective rainfalls can on occasion be confined to a single thunderstorm, but in this case the area of 500 mm-plus (20 inches approx) rainfall was quite large (Fig 74).

rissing rissing2
Fig 73. MSL analysis for 9am 11th March 1924. Fig 74. 48-hour rainfalls to 9am 12th March 1924.
The area of 250mm-plus rainfall (10 inches approx) rainfall extended from Napier in the south to the Waikare River in the north. The bulk of this rain fell in only about 14 hours.
The official New Zealand all-time record for a 12-hour fall is 566mm at Prices Flat near Hokitika, on 11012th May 1978. This is rivalled by the Rissington report of 512mm in ten hours.
When one is confronted with claims of such extreme falls, it is necessary to look for some corroborating evidence. Plenty is available:
• A farmer at Eskdale reported the following:
- 37mm in the four hours from 5am to 9am,
- a further 127mm in the two and a quarter hours to 11.15am,
- a further 102mm in the two hours to 1.15pm,
- then a further 191mm in the four and a half hours to 5.45pm, but with the ‘bottle’ having overflowed.
- His total of (at least) 489mm for the whole event (14 hours or so), included at least 419mm in under nine hours.
• This compared with that of a neighbour about a kilometre away, who reported 514mm for the event.
• A pluviograph at Waipunga recorded 356mm in 18 hours, this included 102mm in 1hr 20mins.
• At one locality, where Territorials were in camp, five inches (127mm) was reported to have fallen in one hour. This value is not much below another official all-time New Zealand record, the 134 mm which fell at Cropp Waterfall, Westland on 8th January, 2004.
The Esk River rose five metres, including two metres in the space of 15 minutes. The Moeangiangi River rose seven and a half metres.
* To the north of the Esk, at Tutira, the fall included 203mm in 3 hours.
Not surprisingly, the level of some Hawkes Bay rivers exceeded those of the 1897 flood. Stock losses were heavy, but a flood warning issued by the Meteorological Office enabled many to save their flocks. Damage was not as great as in 1897, because the heaviest rain was more localised, and did not affect a heavily populated area.


See the '36 Cyclone

1938 19th February

Tarewa: 200mm/24hrs (return period 10 years)
Puninga: 204mm/24 hrs (return period 10 years)
Tolaga Bay: 194mm/24hrs (return period 20 years)
Signal Station Gisborne: 263mm/3days (return period 65 years)

Deaths: 22
Road Damage: 43,800 pounds.
2012 Dollars: 4.5 million.

A camp for construction workers building the Wairoa-Gisborne railway had been located about four metres above normal water level on the banks of the Kopuawhara Stream, just north of Mahia Peninsula. It was made up of houses for married men up on higher ground, and a cookhouse and huts for 47 single men close to the river bank. In the early morning hours of 19th February, a five-metre wall of water hit the camp. Most of the huts and part of the cookhouse were swept away, and 21 lives
were lost.
Another life was lost at Boyd’s Camp. Many bridges were overtopped, and road damage was considerable.
The heaviest rain fell away from rain gauges. The Maungakotukutuku River recorded a freak discharge, which has been subjected to considerable review and scrutiny. Testimony to the extremity of the event was the considerable scour, and large boulders, logs, stumps and other debris deposited onto the downstream grass flats. The measured discharge was equivalent to a rainfall rate of 132mm per hour, over an area of 19 square km.

Pressures were high to the south of New Zealand and low to the north, with a northeast flow over the North Island. A mesoscale low developed off Gisborne within this flow, then moved southwards. Torrential rain occurred from East Cape to Hawkes Bay, and southwards to Cook Strait. Electrical activity was a major contributor to the heavy rain. The charts for 3pm 18th and 9am 19th February (Figs 75 and 76) show the dramatic difference from one day to the next, and highlight the extreme difficulty in predicting heavy rain at times.

kopu kopu2
Fig 75. MSL chart for 3pm 18th Feb 1938. Fig 76. MSL chart for 9am 19th Feb 1938.
kopu3 kopu4
Kopuawhara No. 4 public works camp before the flood,
Mahia Peninsula. 17 Feb 1938

Flood debris from Kopuawhara No. 4 public works camp,
Mahia Peninsula. 20 Feb 1938

Permission of the Alexander Turnbull Library, Wellington, must be obtained before any re-use of these images.


1938 23-25th April

Puketitiri: 1001mm/3 days (return period well over 150 years)
Putorino: 815mm/3 days (return period well over 150 years)
Napier: 274mm/3days (return period 100 years)
incl 169mm/24hrs (24th) (return period 50 years)
Hastings: 271mm/3 days (return period over 150 years)
incl 194mm/24 hrs (24th) (return period 130 years)

Deaths: One.
Evacuations: Many
Damage to Roads and
200,000 pounds
2012 Dollars: 21 million

This was a longer-lived storm than that of 1924, and consequently flooding was more severe.
Slipping of hillsides occurred on a spectacular scale, and Napier and Hastings were inundated to depths of up to a metre.
The main feature of the synoptic situation was a slow-moving low to the north of the North Island (see Fig 77). The extreme rainfalls occurred in a fairly small area. This area lay in a broadconvergence zone between a humid eastnortheast flow and cooler eastsoutheasterlies. The distribution of rainfall was similar throughout the three days of the storm. Rainfalls for April 24th are shown in Fig 78. On the 25th, rainfalls were only slightly less.
esk2 esk
Fig 77. MSL analysis for 9am 24th April, 1938. Fig 78. Rainfall (inches) for 24th April, 1938.
esk3 Hillside erosion scars after the Esk Valley floods.


1941 May 4th

406mm/24hrs (return period well over 150 years)
171mm/24hrs (return period 125 years)

The Porangahau River rose 14 metres above normal and burst its banks, and the township was inundated to depths up to 1.2 metres. The Akitio River rose
15 metres in a few hours, with two bridges washed away. Manawatu was also severely affected by flooding.
A depression moved eastwards across the North Island. A cold front, followed by a strong southeast flow, stalled over southern parts of Hawkes Bay for a
considerable time, as can be seen from Figs 79 and 80.

por por2
Fig 79. MSL analysis for midday, 3rd May, 1941. Fig 80. MSL analysis for midday, 4th May, 1941.

1948 May 14th

Whatatutu 344mm/72hrs (return period over 150 years)
Kanakanaia 339mm/72hrs (return period 45 years)
Hawkes Bay:
Onepoto 307mm/72hrs (return period 20 years)
Tuai 260mm/72 hrs (return period 15 years)

Stock Losses: over 16,000.
Total Damage: 336,356 pounds
2012 Dollars: 23 million

This flood is still regarded as the most destructive in Poverty Bay history. The flood spread over practically all the flats along the main and tributary streams of the Waipaoa River, including the extensive Poverty Bay flats. 8500 hectares were inundated, and parts of Gisborne City were flooded. Further south, the Wairoa River rose to a record height and flooded buildings in Wairoa township. The flood flow of 11,450 cumecs has been exceeded only by floods on the Buller River.

A low developed on a cold front just east of Gisborne on the 12th. The low drifted slowly east, then northwest. A moist east to southeast flow was maintained over the area right through to the 15th, when the gradient over Gisborne started to weaken, and the low began to drift away northeastwards. Fig 81 shows the MSL analysis for 6pm on the 14th.

The rainfall distribution is shown in Fig 82.

pov Fig 81. MSL analysis for 6pm 14th May 1948.
pov2 Fig 82. Rainfall (inches) for 12th and 13th May, 1948.

1953 January 27-28th

Mangarouhi Valley:
349mm/12hrs (return period well over 150 years)
mostly in 6hrs!
Kauranaki: 224mm/9hrs
(return period well over 150 years)
233mm/14.5 hrs (return period over 150 years)
Wilder Settlement:
208mm/9hrs (return period well over 150 years)

Stock Losses: Substantial.

A very active depression developed east of Banks Peninsula and took an unusual northward track, producing flooding from north Canterbury to southern Hawkes Bay.
Exceptionally heavy rain occurred along an axis through Wanstead, Elsthorpe and Maraetotora during the night of the 27th, when the low was centred over Northland. Porangahau was inundated, and bridges in the Dannevirke area were under water. Stock losses included 6500 sheep drowned. The Manawatu River experienced its sixth biggest flood on record.
The MSL chart for midnight 27th (Fig 83) is suggestive of low-level convergence in the area.
The 500 hPa chart for 4pm 27th January (Fig 84) reveals an omega block in the Tasman Sea.
excep excep2
Fig 83. MSL chart for midnight, 27th January, 1953. Fig 84. 500 hPa analysis for 4pm, 27th January, 1953.

1963 June 2-4th

Tareha: 310mm/48 hrs (return period approx 100 years)
incl 195mm/24hrs (return period approx 100 years)

Stock Losses: Heavy

About 40 houses were flooded to window-sill level. Thousands of acres of farmland were inundated. The valley was filled with a swirling torrent of water almost half a mile wide, carrying trees and other debris. After the flood, the Tongoio Valley was declared unsafe for habitation.
An estimated 365mm fell at Flatrock Station in the 15 hours to 6pm June 4th. This is 1.59 times the 150-year return period rainfall.
A deep, complex low moved eastwards over northern New Zealand (Fig 85). The 500hPa analysis for midday 4th (Fig 86) is rather unusual. While it does not fit any of the classic blocked patterns, the slack gradients and long fetch of easterly quarter winds is indicative of strong blocking.
tongoio tongoio2
Fig. 85. MSL analysis for 6am 4th June 1963. Fig 86. 500 hPa analysis for midday 4th June, 1963.

1988 March 6-8th

Gisborne/Hawkes Bay:
Te Puia Springs 419mm/24 hrs (return period over 150 years)
Aniwaniwa 771mm/3 days (return period over 150 years)
Glenross/Waimata 917mm/4days (return period well over 150 years)
Tutira 649mm/2 days (return period over 150 years)
Whangarei 234mm/2 days (return period 30 years)

Deaths: Three
Evacuations: Hundreds.
Damage: $111 million.
2012 dollars: 205 million.

It took the Gisborne District years to recover from the massive landslides wrought by ex-tropical cyclone Bola. The rain also cut the water supply to Gisborne City, and three people were killed when their car was swept away by floodwaters. States of emergency were declared in Wairoa, Gisborne and the East Cape. Intense rainfall also occurred in Northland. 593mm was recorded for the event at one location there, and a state of emergency was declared in Dargaville.
As can be seen from Fig 87, the centre of the low did not come anywhere near Gisborne. This was a classic case of extratropical transition.
bola Fig 87. Track of Cyclone Bola.
Fig 88 shows that the system had lost its tropical cyclone appearance by the time the heavy rain started in Gisborne.
The heavy rain occurred in an area where tropical air brought down by the low converged with cooler air, brought up from the southeast on the northern flank of an anticyclone.
Gisborne also lay within a “double jet” structure – that is, it lay in the region between the poleward exit region to one jet, and the equatorwards entrance region to another (Fig 89).
The orography of Gisborne did the rest.
bola2 bola3
Fig 88. GMS 3 low-resolution infrared imagery for midday 6th March 1988.
The arrow indicates the surface centre, also its direction of movement.
Fig 89. 200 hPa analysis for midday 7th March 1988.
Fig 90 shows rainfalls for the event for the five-day period starting 1am NZDT 5th March.
A heavy rain warning was issued for the event, but the amounts that eventuated were far in excess of the forecast.
bola4 Fig 90. Five-day observed rainfalls (mm) for Cyclone Bola.

2001 9th December


Surface flooding caused millions of dollars worth of damage in Napier and Hastings when 50mm of thundery rain fell in the hour before noon – a return-period rainfall of over 150 years. Thundery downpours occurred in Hawkes Bay, Wairarapa, Wellington and Marlborough. Fig 91 shows the area lying in a convergence zone between warm moist northerly air and a cooler southerly.

Fig 92 suggests that overrunning was a significant factor.

hvy hvy2
Fig 91. MSL analysis for midday 9th December, 2001. Fig 92. 500 hPa analysis for midday 9th December, 2001.


See Central Wellington Flooded

2003 25-28th February

Waikura 438mm/48hrs to end 27th (return period 130 years)
Hikuwai 423mm/60hrs (return period 120 years)
incl 381mm/48hrs (return period 100 years)

Coromandel Peninsula:
Pinnacles about 650mm/2days (return period 100 years(?))
Golden Cross about 400mm/2days (return period 35 years)

North of Tolaga Bay, up to 600mm was reported in two days. At Maungatuna, near Tolaga Bay, six people were forced from their homes. The Hikuwai and Uawa Rivers threatened to burst their banks, and residents were on standby to evacuate. Roads were closed on the Coromandel Peninsula, and four houses were evacuated at Paeroa. The HIRDS2 return period for the Pinnacles rainfalls is probably too high.
A frontal band moved southwards onto Gisborne, embedded in a strong easterly flow (see Fig. 93). Upper-level divergence was a significant factor in the rainfall,
particularly on the 26th, and strong gradients advecting very warm moist air onto the mountains were another major cause. This was demonstrated by the fact that heavy rain continued over Coromandel Peninsula after the front had moved south of the area.
sss Fig 93. MSL analysis for midday, 27th February, 2003.

2004 15-16th February

See Parts of Manawatu Devastated

2004 18th October

Tamatea: 179mm/3 hrs to 4.30am 18th (return period well over 150 years)
Ben Nevis: 165.6mm/6 hrs (return period over 150 years)
incl 70.5mm/1 hr (return period over 150 years)

Thunderstorms caused localised flooding in Napier around 2am, and around 8am northwest of Porangahau.
The thunderstorms formed on and south of a convergence line between northwesterly and easterly flows on the southeastern side of a deepening low.

Napier reported a wind change from southeast to southwest during the thunderstorm episode. Thunderstorms were clustered along the Hawkes Bay coast.

loc Fig 94. Analysis fields for midnight 17th October, 2004.
Top left: MSL pressure, 850hPa temp and 700hPa relative humidity.
Top right: 850hPa wind barbs and 850hPa wet bulb potential temp.
Bottom left: Wind barbs, isotachs and divergence at 250 hPa.
Bottom right: Height and vorticity at 500hPa, and 700 hPa upward motion.

Fig 94 shows several analysis fields for midnight 17th October. It can be seen that while WBPTs were very low, strong CVA and upper divergence were present over Hawkes Bay. Satellite imagery (Fig 95) attests to the fact that a very unstable atmosphere prevailed over the area. Indeed, the sounding from Paraparaumu, admittedly some way from the affected area, produced around 800j/kg of CAPE when edited with the actual Napier temperature and dew point (Fig 96).

loc2 loc3
Fig 95. GMS IR imagery for 4am 18th October, 2004. Fig 96. Paraparaumu tephigram for midnight 17th modified using actual temperature and dewpoint at Napier.

2005 18th March

Pukeorapa: 126mm/2hrs to 4am 18th (return period over 150 years)
incl 99mm/1hr to 3.30am (return period well over 150 years)
and 56mm/30mins to 3am (return period well over 150 years)
Heavy rain occurred throughout Hawkes Bay, but a small area in the north received a short-period deluge.
A slow-moving 500 hPa low lay over and east of the North Island, with a surface low trapped beneath it (Fig 97). The southeast airstream affecting northern Hawkes Bay was very warm and moist, and the imagery (Fig 98) suggests that the rain was being generated by low-level processes. However, why such an extreme rainfall occurred remains a mystery. Some other parts of Hawkes Bay did receive 24-hour rainfalls exceeding 100mm.
locdel locdel2
Fig 97. MSL analysis for midnight 17th March 2005. Fig 98. GMS IR imagery for 4am 18th March 2005.

2005 21st October

Hikuwai: 371.5mm/36hrs to 11am 22nd (return period 120 years)
incl 273mm/15hrs to 10pm 21st (return period about 150 years)
Te Puia: 365.5mm/36hrs to 10am 22nd (return period 60 years)

Evacuations: 50
Crop Damage: $8.4 million.
2012 Dollars: 10 million.

Gisborne City was isolated by flooding and slips. Both Hikuwai and Waipaoa Rivers peaked at within half a metre of Bola levels. The flooding hit at a critical part of the season, and severely affected the squash, sweet corn, maize and broccoli crops. The worst affected area was from Te Puia to Tolaga Bay. Rainfall intensities reached 44mm per hour, with the highest rainfall of 385mm in 36 hours recorded just north of Tolaga Bay.
A low formed to the northeast of the North Island on the 20th. The low deepened as it moved southeastwards, passing close to East Cape on the night of the 21st. The low was deep and slow-moving and cut off to a high level (Figs 99, 100). The heavy rain occurred within the wrap-around to the low.
crops crops2
Fig 99. MSL analysis for 6pm 21st October, 2005. Fig 100. 500 hPa analysis for midnt 21st October, 2005.


Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
Hill, H., 1897: On the Hawkes Bay Plain: Past and Present. Transactions of the Royal Society of New Zealand 1868-1961. National Library of New Zealand website.
Kidson, E., 1930: The Flood Rains of 11th March, 1924, in Hawkes Bay. New Zealand Journal of Science and Technology, Vol. XII, No. 1, pp 53-60.
McLintock, A.H., 1966: The Encyclopaedia of New Zealand. Published on the Internet.
New Zealand Gazette, 1910, January-June. New Zealand Government.
________, 1938, Vol. I. New Zealand Government.
________, 1941, Vol. II. New Zealand Government.
________, 1953, Vol I. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Salinger, J. in Awesome Forces, 1998: Edited by Geoff Hicks and Hamish Campbell. Te Papa Press.
Sinclair, M.R., 1989: Preprints, 12th Conf. Wx. Analysis and Forecasting, Monterey, California, Oct. 2-6, 1989.
___________, 1993: A Diagnostic Study of the Extratropical Precipitation Resulting from Tropical Cyclone Bola. Mon. Wea. Rev., 121, 2690-2707.
The Hawkes Bay Herald, April 19, 1897.
The Manawatu Herald April 17, 1897, and April 22, 1897.
The Taranaki Herald, April 20, 1897.
The Whanganui Herald, April 17, 1897.


The predominant heavy-rain bearing flow in this area is the northwesterly. A straight northerly flow can bring heavy rain to Taranaki, while the Manawatu River, rising as it does east of the main divide, can receive flood rains in easterly flows. Also in easterly flows, significant spillover can occur on the western side of the Ruahines. Flows from between west and southwest generally do not contain enough moisture to bring heavy rain. Post-frontal northwesterly flows can occasionally cause problems. If the flow is strong and moist, heavy rain can continue, particularly at high altitudes.
That cyclonic southerly flows could bring heavy rain did not become obvious until a devastating flood in the Manawatu in February 2004. Research for this publication revealed that the 1897 flood, which still holds the record for the highest known flow in the Rangitikei River, also occurred in a cyclonic southerly flow. What happens is that broad-scale southeasterly winds converge over the Manawatu/Central Plateau area with south or southwesterly winds that have come through Cook Strait.


See Biggest Flood on Record for Wairarapa


See Houses Washed Out To Sea

1902 13-15th June

Palmerston North:
91mm/3days (return period 5 years)
281mm/3 days (return period 150 years)
This was an easterly rainfall, and the Manawatu River flooded land to a depth of up to six metres. Floodwaters were up to a metre higher than in 1880.

1904 22-25th May

102mm/3 days (return period 15 years)

Stock Losses:

Opinions differ as to which flood was bigger for Whanganui: 1904, or 1940. However, the 1904 event was the highest known at that time for the town, and was probably the bigger of the two. East Town in Whanganui township was extensively flooded.

A 992hPa low moved southeastwards onto the South Island, and another (985hPa) followed it. A prolonged period of northwesterlies affected the Whanganui River headwaters. Figs 101, 102 and 103 are the original charts for 9am on the 24th, 25th and 26th respectively. Fronts were not analysed in those days, but a possible frontal analysis is superimposed in red. This suggests that a front remained stationary over or near the area for three days.

whang whang2
Fig 101. MSL analysis for 9am 24th May, 1904. Fig 102. MSL analysis for 9am 25th May, 1904.
whang3 whang4
Fig. 103. MSL analysis for 9am 26th May, 1904. Flooding at Whanganui, showing the railway station, 1904.
Permission of the Alexander Turnbull Library, Wellington,
must be obtained before any re-use of this image.


See Worst Ever For Hokitika
npa Flooding on Devon Street, New Plymouth, late 1930s.
Permission of the Alexander Turnbull Library, Wellington, must be obtained before any re-use of this image.

1940 24-25th February

Central Plateau:
Taumarunui 147mm/24hrs (return period 125 years)

Whangamomona 229mm/24hrs (return period 130 years)
New Plymouth 150mm/24hrs (return period 30 years)
Tangarakau 218mm/24hrs (return period over 150 years)

This flood occurred during the year which celebrated 100 years since the signing of the Treaty of Waitangi. For Whanganui this was the third major flood in half
a century, equalling, in the opinion of many, the great flood of 25th May 1904. A period of southeasterly rain may have produced comparable flooding at the mouth, even if flooding in the upper and mid reaches was higher in 1904. At the height of the flood, one of the Whanganui River steamers sailed over the top of a bridge and into a garden, where peaches were picked from the tree tops.

In Taranaki, parts of New Plymouth were under water, and roads everywhere were blocked by slips and washouts. The RNZAF dropped supplies to remote settlements in north Taranaki which were isolated by floodwaters.

On the 20th a high centre moved to the northeast of the country, where it remained for several days. On the 23rd rain commenced in Taranaki. A disturbance approached from the Tasman Sea, and one of its centres developed into a small but active depression.
The southeastward passage of this low brought the heavy rain to the central North Island (Fig 104).
The rain occurred firstly within a long fetch of northerlies, and later in a southeast flow as the low moved away.


Fig 104. MSL analysis for 3pm, 24th February, 1940

(central pressure in the low is about 985 hPa).


See Porangahau Inundated


See Exceptional Rainfalls

1971 23-24th February

New Plymouth:
290mm/24hrs (return period over 150 years)
Bell Block: 329mm/24hrs (return period over 150 years)
Tarata: 461mm/24hrs (return period well over 150 years)

New Plymouth’s city centre was flooded, with shops ruined and the main routes in and out of the city closed. Wide areas of Waitara were covered with metres of water, and in rural Taranaki roads were closed by slips and flooded rivers, and schools were closed.

A low developed just south of Norfolk Island on the 21st, to the south of a weak tropical depression. The low then moved southwards while deepening, with a northeast flow developing over the North Island. Frontogenesis then occurred west of the North Island, downstream from a strong upper trough in the Tasman Sea. The MSL chart for 6am on the 24th (Fig 105) might suggest that the flow over Taranaki was anticyclonic. However, the tropical depression had played a role in bringing very
warm and moist air down into New Zealand latitudes, and strong ascent was occurring within a warm conveyor belt over Taranaki.
Fig 106 shows that the rain was associated with a sharp upper-level trough.

worst worst2
Fig 105. MSL analysis for 6am 24th February 1971. Fig 106. 500 hPa analysis for noon 24th February 1971.
The frontal band gradually weakened during the 24th, but it remained quite slow-moving, and rain continued in Taranaki until it finally moved off on the morning of the 25th.
Although the long northerly fetch bringing very moist air from the tropics was a major factor in this event, the blocked nature of the situation was also very important.

1990 8-11th March


Pukeiti: 751mm/6days incl 445mm/24hrs to 9am 9th (return period 140 years)
Lake Mangamahoe: 498mm/6 days incl 294mm/24hrs to 9am 9th (return period over 150 years)
North Egmont Visitor Centre: 970mm/6 days
incl 419mm/24hrs to 9am 9th (return period 45 years)

Stock losses: Major.

Although it did not come anywhere near New Zealand, ex-tropical cyclone Hilda was blamed for this event, with moisture from the system finding its way onto
Waitara was evacuated owing to fears of a breach which did not eventuate. The Whanganui River broke its banks though, leading to the evacuation of several houses. Farmland was inundated, and there was extensive damage to roads, with many remote parts isolated.
Hilda moved into the north Tasman Sea, then transitioned on the 8th into a large low-pressure area.
The heaviest rainfalls occurred in the 24 hours to 9am 9th March.

The MSL analysis for midnight 8th March (Fig 107) shows a very anticyclonic-looking flow over Taranaki, and the low-level flow over the area is quite northeasterly.
However, the 500 hPa analysis (Fig 108) shows air at this altitude being brought from very low latitudes.
This air was rising as it moved through the eastern Tasman Sea and onto Taranaki – this is the warm conveyor belt (WCB). Satellite imagery (not shown) revealed the frontal band extending right up into tropical latitudes. Very humid air was being advected onto Taranaki – dew points reached as high as 21C.

hilda hilda2
Fig 107. MSL analysis for midnight 8th March, 1990. Fig 108. 500 hPa analysis for midnight 8th March, 1990.
Twenty-four hours later (Fig 109) the flow has turned more northerly, while the 500 hPa analysis (Fig 110) still shows air being fetched from low latitudes. The
WCB, in a weakened state, finally moved off the area on the afternoon of the 10th.
hilda3 hilda4
Fig 109. MSL analysis for midnight 9th March, 1990. Fig 110. 500 hPa analysis for midnight, 9th March, 1990.

The similarity to the 1971 event was quite marked, and the main causes of the heavy rain were the same, viz.:

1. a tropical cyclone bringing tropical air into the north Tasman Sea, which was subsequently dragged further south by troughing,

2. a strong WCB, and

3. a blocking high to the east.

1995 20-21st April


New Plymouth AWS: 134mm/5hrs (return period over 150 years)

New Plymouth was cut off by floodwaters, with flooding in some central city streets. There were landslides in the city, drain covers popped, and raw sewage spilled across the streets. However, the heavy rain was quite localised, with no major rivers or streams breaking their banks.

On the 19th and 20th a high to the east extended a ridge across the North Island, while a front approached from the west. The front moved onto Taranaki and then stalled.

Fig 111 shows the MSL chart for midday on the 21st. It was just before this time that the New Plymouth AWS received its downpour of 134mm in 5 hours. The flow over Taranaki looks anticyclonic, as it did for much of the period of the 1971 and 1990 Taranaki floods. The rain eased off shortly after this time, but it did continue to fall, with some heavy bursts, right through to the 23rd. Flooding in Taranaki was not as severe as in the events of 1971 or 1990. Blocking was again a major factor, and the WCB affected the area for a considerable time. The difference was probably the absence of a long fetch from the tropics.

wet Fig 111. MSL chart for midday 21st April 1995.

2004 15-16th February


Apiti, Table Flat 200mm/24 hours (return period 130 years)
Kopua 176mm/24 hours (return period 100 years)
Dannevirke, Pine Gr. 225mm/24 hours (return period over 150 years)

Birch Lane (Lower Hutt) 234mm/48hrs (return period 50 years)
incl 199.5mm/24hrs (return period 50 years)
and 161.5mm/12hrs (return period 50 years)

Hawkes Bay:
Shagrock 228mm/24hrs (return period 150 years)
incl 48mm/1hr (return period 150 years)
Wallingford 197mm/24 hrs (return period 125 years)

Evacuations: Many
Stock Losses: High.
Damage: $112 million.
2010 dollars: 140 million.

This was the most devastating flood in New Zealand for many a year. Insurance payouts of 112 million 2004 dollars made it the fourth most expensive natural disaster since 1968 – after the 1987 Bay of Plenty earthquakes, the loss of the Wahine in 1968, and the Otago/ Southland floods of 1984.

Farmland was inundated, and stock losses were high. The region lost 100 million cubic metres of soil (one-sixth of Mount Ngauruhoe), which was either deposited on farmland or washed out to sea. It was the third largest flood on record for the Rangitikei River, after 1897 and 1926, while for the Whangaehu River it was the biggest, much bigger than the 1897 event.

In the Wellington Region slips and flooding were widespread, with 50-year flows in the Ruamahanga River in the Wairarapa, also the Waiwhetu Stream in the Hutt Valley.

In Hawkes Bay there were evacuations in Porangahau Village. Fences and gates were demolished, and there were numerous slips, with roads blocked.

Rainfalls exceeded the 150-year return period over a wide area, coloured purple in Fig 112.


Fig 112. 24-hour rainfall return periods for the storm of 15-16th February, 2004.
Over 150 years – Purple.
Over 100 years – Dark blue.
Over 75 years – Light blue.
Over 50 years –Teal.
Over 25 years – Dark green.
Over 10 years – Light green.
Over 5 years – Yellow.
Over 2 years – Orange.
Less than 2 years - White.

See Acknowledgements

The synoptic situation was markedly similar to that of the 1897 event, with a deep low moving slowly eastwards over the North Island. Fig 113 shows a strong southeast airstream affecting the southern half of the North Island.
mnu2 Fig 113. MSL analysis for midday,15th February, 2004.
An interesting feature of this event was that rainfalls on the lee (western) side of the Ruahine Range were almost as high as those in the east. This is not all that uncommon, and the explanation can be seen in Fig. 114.
mnu3 Fig 114. Diagram showing low level air (light blue) being overrun
by air of higher potential temperature. (Taken from a case study of
the 15th August 2001 snowstorm on the central plateau.)

This figure is taken from a case study of a winter storm that dumped up to 40cm of snow on the central plateau. Low-level air blowing through Cook Strait is being overrun by air of higher potential temperature coming from the southeast (the warm conveyor belt), with warm advection adding to the upward motion.

In the February 2004 event, an intrusion from the north of dry air with high Isentropic Potential Vorticity added deep instability to the mix. The 2004 storm in fact resulted from a polar outbreak more typical of winter – but with temperatures much higher, precipitation amounts were correspondingly much greater.

Another interesting aspect of the storm was the presence of a striated cloud head over the southern North Island as well as a striated delta further east. The striations over the southern North Island can be seen clearly in Fig 115.

On the right of the picture can be seen part of the cirrus shield of a larger striated delta – this pattern has been associated with major cyclogenesis. The striations over the North Island were moving westwards, with both wavelength and amplitude increasing with time and distance downstream. This suggests that energy was being fed into the system. The striations showed both in radar returns, and as peaks in rainfall intensity (not shown).

mnu4 mnu5
Fig 115. GMS visible satellite image for 11am, 15th February, 2004. Manawatu flooding, February, 2004.
(Manawatu District Council)


Dixon, R.S., K.A. Browning, and G.J. Shutts, 2000: The mystery of striated cloud heads in satellite imagery. Atmospheric Sciences Letters.
Feren, G., 1995: The “striated delta” cloud system – a satellite image precursor to major cyclogenesis in the eastern Australian – western Tasman Sea region. Wea. and Forecasting, 10, 286-306.
Horizons (Manawatu-Whanganui Regional Council) website: www.horizons.govt. nz
Miller I.D., 2005: Some 3D aspects of a striated cloud head over New Zealand including radar and raingauge data. pdf
New Zealand Gazette, 1940, Vol. I. New Zealand Government.
______, 1926, Vol. III. New Zealand Government.
New Zealand Herald 8th November 1965
______, 12th March 1990
______, 21st April 1995
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Pascoe, R.M., and P.E. Bruce, 2003: Cyclogenesis East of the North Island and the Central North Island Snowstorm of 15th August, 2001. Unpublished. (Available
from New Zealand Meteorological Service.)


The weather in this area is greatly influenced by Cook Strait. In Wellington there can be only two rain-bearing wind directions because there are essentially only two wind directions – north and south. In Wairarapa, the main flood-bearing wind is the northwesterly. The Tararua Ranges can receive huge rainfalls, which often spill over well to the east of the divide, and because the Tararuas are so steep the Wairarapa rivers rise very quickly. Wairarapa can also receive heavy falls in easterly flows, the brunt of which is borne by the eastern hills.

Wellington City and the Hutt Valley get their heaviest rainfall in southerlies. Forecasters particularly look out for where a shallow southerly is overlain by moist northerlies – the so-called “overrunning situation”. However, for overrunning to produce extreme rainfalls there needs to be other forcings at work, such as upper-level divergence.

1858 17th January

Deaths: 13.

An early settler wrote: In the year 1858 there was a tremendous flood in the Hutt River; the water reached from hill to hill; great trees came floating down carrying away houses with them; thirteen people were drowned.

As a result, many settlers left the Hutt Valley.

1880 24th March

Flooding affected over 20,000 hectares, and the return period of the flow into Lake Wairarapa was over 100 years. This is thought to have been the biggest flood in the Wairarapa since European settlement.
In the Manawatu River, this flood was of a comparable size to that of 1902.


See Houses Washed Out to Sea

1898 17th June


This flood was of a similar magnitude to that of 1858. These two floods have been estimated to have reached 2000 cumecs in the Taita Gorge, and as such are thought to have been the biggest floods in the Hutt River in Pakeha times. Water was knee deep in Lower Hutt township.

Surprisingly, this was a northerly rainfall. A period of northerly gales brought minor wind damage, and the railway line was washed out near Paraparaumu. Snow melt was thought to have been a factor.

hutt Flooding, Lower Hutt, June, 1898. Railway Ave in middle of picture.
(Local History Collections at Hutt City Library)

1924 17-19th December

264mm/3 days (return period over 150 years)
incl 179mm/24hrs to 9am 19th (return period over 150 years)
Low-lying parts of Masterton were flooded. Three days of rain included a severe electrical storm. The synoptic situation involved a depression moving
eastwards across the North Island.

1932 28-30th August

Putara: 370mm/3 days (return period 70 years)
including 203mm/24 hrs (return period 15 years)

Stock losses: Heavy.

Widespread heavy rain affected the lower North Island over a three-day period. Several Wairarapa rivers broke their banks, and the town of Masterton was flooded. Particularly badly affected was southern Wairarapa, where the flooding reached record proportions.

Individual rainfalls were not unusually high, but heavy rainfalls were sustained over three days, and affected all parts of the Wairarapa catchments. In this case a period of northwesterly rain was followed by a period of southeasterly rain. Snow melt in the northwesterly stages of the event added to the flooding.

Fig 116 shows a strong northwest flow onto the area.

In Wellington there were several gusts exceeding 100km/hr, while at Eketahuna the worst gale for 20 years uprooted trees and damaged buildings.
Fig 116 shows a wave developing west of central New Zealand. This wave had the effect of slowing the front.

Subsequently a cut-off low developed from this wave (Fig 117).

swrp swrp2
Fig 116. MSL analysis for 9am 28th August, 1932. Fig 117. MSL analysis for 9am 29th August, 1932.

A weak southerly reached the southern North Island, and there was a considerable period where this weak southerly was overlain by the moist humid northerly
airstream. The heaviest rainfalls in Wairarapa occurred in the 24 hours to 9am on the 29th, the flooding reaching its climax on
the night of the 28th.

In winter, this is a classic snow-producing situation, and in fact snow fell to amazingly low levels, with 20cm on the ground near Masterton and 15cm at Featherston.

1939 11th December

Kelburn: 127mm/13 hrs (return period 130 years)

Stock Losses: Large

This flood, described at the time (and still today) as the greatest in living memory, covered hundreds of hectares, isolated houses, flooded roads, destroyed and damaged bridges, and drowned large numbers of stock. There were also slips and flooding in Wellington City. At one point at Silverstream the entire flat between the hills forming the valley was one sheet of turbulent water, fences and roads being invisible.

The storm was brought by a deep low moving eastwards over northern New Zealand. Northeast gales blew over the Auckland Peninsula, while strong to gale southerlies were experienced around Wellington.

1947 27-29th June

129mm/24hrs. (return period 50 years)
Greytown: 106mm/24hrs (return period 20 years)
Martinborough: 182mm/24hrs (return period over 150 years)

Deaths: One.

Stock Losses: Large.

A week of southerly rain brought the worst floods to the Wairarapa since 1897. 9000 sheep and 300 cattle were lost, and 60,000 acres flooded. Masterton and many other Wairarapa towns were flooded, and southern Wairarapa was cut off. It was a hundred year event for the Ruamahanga River.

The main episode occurred between the 26th and the 28th, with a long period of strong, cyclonic southerlies affecting central New Zealand.

1976 20th December

Taita: 230mm/8hrs (return period well over 150 years)

Damage: $30 million
In 2012 dollars: 218 million.
Evacuations: Many.
Deaths: One.

When one mentions “The Hutt Valley Floods” this is the event which comes to the mind of many people, although February 2004 has doubtless supplanted that in the minds of some. The photograph of the Petone overpass (below) shows that this was an event of unusual severity. Several houses in the Hutt Valley were demolished, and Lambton Quay in Wellington lay under 30cm of water.

The heavy rain was produced by mesoscale processes, the synoptic situation, shown in Fig 118, being nothing out of the ordinary.

huttv huttv2
Fig 118. MSL analysis for 6am 20th December 1976. Petone, 20th December, 1976: Flooding from Korokoro Stream
(The Evening Post/courtesy of Dominion Post)

The cause was a convergence line,extending from the hills west of Wellington City and up the Hutt Valley, between weak southeasterlies and weak northwesterlies. Fig 119 shows the position of this convergence line at 9am 20th December. Contributing factors were high instability and considerable upper-level moisture.

During the period of the intense rainfall, the winds through a deep layer of the atmosphere (2km to 10km) above the convergence line appear to have been northeasterly in direction, and quite light in strength. The dashed arrow in Fig 119 indicates the direction of this upper-level flow. This northeasterly, containing a high
level of moisture, overran the cooler and denser southerly air below.

huttv3 Fig 119. Local wind velocities at 9am 20th December 1976. At each observation point (black dot) the arrow indicates the direction towards which the wind was blowing and the number is the Beaufort wind force (where C indicates calm conditions – for rest see below). The dashed arrow indicates the direction of the steering flow and the broad line is the line of convergence.
Beaufort Scale:
1: 1-3 knots
2: 4-6 knots
3: 7-10 knots
4: 11-16 knots
5: 17-21 knots
6: 22-27 knots

Fig 120 shows the 24-hour rainfall during 19 and 20 December (most of the rain fell in 12 hours or less). The rain eased when the southerly winds developed sufficient strength to destroy the deep vertical structure of the convergence zone, and began to move it northwards.

Fig 120. 24 hour rainfall during 19 and 20 December 1976. All points within the dashed line received 24 hour rainfalls of return period greater than 50 years.

1998 26th June

Kelburn: 69.5mm/1hr 40mins to 9.10pm (return period over 150 years)

Evacuations: Several.

Vehicles were submerged and several properties damaged when a short period of intense thunderstorm activity struck parts of Wellington. People were evacuated from homes in Thorndon and a hotel in Oriental Bay which were struck by slips. Lightning also caused power cuts. In Karori, 80mm was reported to have fallen in an hour
and a quarter.

Fig 121 shows the MSL situation for midnight on the 26th.

A cold front had moved onto the southern North Island, followed by a southerly change. Nothing appeared to be particularly unusual, but cooling aloft, relatively humid air at low levels ahead of the front, and strong CVA ahead of an upper trough caused high instability in a strongly convergent low-level flow. Multi-cellular
cumulonimbus clouds developed over the hills west of Wellington.
Heavy rain had been forecast for the Tararuas only. However, the heavy rain around Wellington affected only a small area, from Karori to the CBD.

flash Fig 121. MSL analysis for midnight 26th June 1998.

2002 10th January

Kelburn over 50mm in less than 1 hour (return period over 150 years)

Hawkes Bay:
Hastings 77mm in 90 minutes. (return period well over 150 years)
Napier 70mm in 90 minutes. (return period over 150 years)

Water was knee-deep in Wakefield and other city streets after a violent thunderstorm struck Wellington City. Slips caused the evacuation of homes. The trigger was the sea-breeze convergence zone.

Thunderstorm activity also occurred in Wairarapa, within a shallow trough of low pressure containing humid, unstable air and light winds. Further north, streets and properties were flooded in Napier and Hastings, in what was described as the second hundred-year event in a month.

Fig. 122 shows the likely sounding over Wellington that day, adjusted from the actual Paraparaumu sounding. It shows the deep instability and high, ‘skinny’ CAPE characteristic of soundings within high rainfall storms. The low storm motion also contributed to the high rainfall. The “straight-up-and down” structure of the storm, as seen in the radar cross-section (Fig. 123) points to the low shears that prevailed over the area, leading to intense rainfall over a small area.

tep rad
Fig. 122. Likely sounding for noon 10th January 2002
(modified from actual Paraparaumu sounding).
Fig 123. Radar cross-section of thunderstorm over Wellington 10th January 2002. Strongest radar returns (and thus heaviest rain) can be seen in the core of the cloud.

2003 3-4th October


A deluge struck the Kapiti Coast, caused by a slow-moving convergence zone between northeast winds to the north and northwest winds to the south. A Convair freight aircraft became caught in severe icing and crashed into the sea, killing both pilots. Paekakariki was worst hit. A huge landslide crashed into a motel, and the Paekakariki Hill Road was rendered impassable by slips. State Highway 1 was flooded, with motorists trapped, and the Kapiti Coast was declared a disaster area.

The severe flooding affected only a small area. A private weather station 3 to 4km north of Paekakariki recorded 92mm in 24 hours, including about 60mm in 6 hours.
However, rainfalls were likely to have been considerably higher in the hills immediately above Paekakariki. Fig 124 shows the radar rain accumulation for the period 7am to 10pm 3rd October 2003. Based on this, the return period of the rainfall was estimated at greater than or equal to 120 years.




Fig 124. Radar rain accumulation for 7am-10pm 3rd October 2003.

A maximum of 93mm lies just south of Kapiti Island, and another of 124mm in the Marlborough Sounds.




The convergence zone developed within a frontal conveyor belt in a very strong low level north to northwest flow. See Fig. 125. 850 hPa WBPTs were 15-16C. Fig. 126 shows the analysis for midday 3rd October.
paekak3 paekak4
Fig. 125. Schematic diagram of convergence 3rd October, 2003. Fig. 126. MSL analysis for midday 3rd October, 2003.


See Parts of Manawatu Devastated


See Massive Rainfall Floods Waikawa

2005 30th March

Castlepoint: 115mm/3hrs to 9pm 30th (return period well over 150 years)
incl 92mm/2hrs (return period over 150 years)
and 57mm/1hr (return period over 150 years)
Masterton: 148mm/24hrs (return period 90 years)
Porangahau: 115mm/3.5hrs (return period over 150 years)

Residents of the coastal settlements of Mangatoetoe and Mataikona were isolated for several days. Bridges, fences and roads were washed away, and dozens
of houses were extensively damaged. Thousands of tonnes of logs – debris from forestry harvesting – were catapulted over farmland and hurled through the sides of farm buildings, wiping out equipment and killing farm stock. There was intense electrical activity out to sea, which encroached onto the coast in some places. Southern parts of the Wainuiomata Coast Road were hit by the worst flooding in living memory, with the road washed out in several places. One place (Homeburn) recorded 444mm in 36 hours.

Waipukurau and Waipawa, in central Hawkes Bay, experienced 25 to 30mm of rain in 15 minutes (return period about 150 years), and water poured
into four businesses. There was also flooding in Picton, in the Marlborough Sounds.

The extreme rainfalls were the combined result of a frontal band, with strong gradients and high dew-point air, which remained slow moving over the area for a considerable period, and convective activity embedded within the front (see Fig. 127).

cow Fig. 127. MSL analysis for midnight, 29th March, 2005


Kidson, E., The Wairarapa Floods of August 1932. New Zealand Journal of Science mand Technology. Vol. 14, No. 4. p. 220-227.
Kilmister Reminiscences, Alexander Turnbull Library, MS 1117, dated 12 May 1932.
McGavin, T. : Severe Weather Events Analysis. New Zealand Meteorological Service
Mosley, M.P., and C.P. Pearson, 1997: Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society.
New Zealand Gazette, 1940, Vol. I. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Severe Weather Log. New Zealand Meteorological Service.
Tomlinson, A.I., 1977: The Wellington and Hutt Valley flood of 20 December 1976. New Zealand Meteorological Service Technical Information Circular 154.
Wellington Regional Council Website.


A strong northerly ahead of an active front is the main bearer of heavy rain for the top of the South Island.

Marlborough is most vulnerable in flows from 330 to 360 degrees. As the flow tends more northerly, spillover onto the south side of the Wairau River tends to increase. The event of July 1983, one of the biggest known floods, occurred with isobars from almost due north. As the flow tends northeasterly, the proportion of total rainfall spilling south of the Wairau continues to increase. However, total rainfall tends to decrease, with the area gaining increasing shelter from the North Island. As the flow tends more easterly, Marlborough again becomes exposed, and the catchment of the Taylor River, which runs through Blenheim, receives its heaviest falls in flows from between northeast and southeast.

If a dominant direction for an impending event can be identified for Nelson, this is very useful information to hydrologists in determining which river catchments will be affected. The wettest part of the Nelson area is the Tasman Mountains, which receive rain in the more westerly flows. The northeasterly can be extremely wet in Golden Bay. Heavy rain can occur in this area even if there is a southeasterly blowing out of Cook Strait, as convergence develops between this relatively localised
southeasterly and the broadscale northeast flow.

Purely convective major flooding events are rare, but the summer months can see big thunderstorms developing in the high dry mountains south and southeast of Blenheim.

1866 1-3rd February

An extended period of rain led to severe flooding. All three bridges over the Brook Street Stream disappeared, and the bridges over the Matai sustained severe damage. This was said to have been an easterly rainfall. The broadscale flow was probably from the northnortheast.

1867 27-28th January

Nelson: 239mm/24hrs (return period over 150 years)
Flooding with this event was not as great as might have been expected from the amount recorded. The flood was said to have been not as severe as that a year previously. Thunder and lightning were a major feature, so the extreme rainfall recorded by the Meteorological Registrar was probably quite localised. It appears to have been a typical north or northwesterly event. However, most of the thunder and lightning activity was reported to have been emanating from the western ranges, which leads one to wonder what devastation was occurring there. . . .


See Ex-Tropical Cyclone Sweeps over New Zealand

7th February

Known as the “Great Earth Flood”, this event was said to have raised river-bed levels by more than three metres in the lower Motueka River. The intense rainfall generated a huge number of slips.

1923 5-8th May

Spring Creek, 165mm/2 days (return period 50 years)
incl 102mm/24hrs (return period 10 years)
Duntroon, 234mm/2 days (return period over 150 years)
incl 161mm/24hrs (return period 65 years)

Emscote, Stag and Spey 934mm/4 days (return period well over 150 years)
incl 775mm/2 days (return period well over 150 years)
and 424mm/24hrs to 9am 7th (return period over 150 years)
Keinton Combe, 366mm/24hrs (return period well over 150 years)
incl 330mm/12hrs (return period well over 150 years)
Christchurch: 86mm/24hrs (return period 15 years)

Deaths: Two.
Stock Losses: Huge
Road and Public Works Damage: 41,800 pounds.
2012 dollars: 3.9 million.

This was the worst of several events to have flooded Blenheim in its history. The town was inundated to a depth of 1.2 metres, and rowing boats plied Market Place. There were huge stock and property losses. The town is now protected by a detention dam. It is estimated that 125-150mm fell in Blenheim overnight (return period over 150 years) but the gauge overflowed. This event and that of 1868 are the only two known occasions when the weather pattern changed from eastsoutheast to
northnorthwest (or vice versa) and brought major floods to both Wairau floodplain river systems.

There was also severe flooding along the Kaikoura coast, and in Canterbury. Phenomenal rainfall was recorded at Keinton-Combe, and at Emscote, Stag and Spey, putting this event into the top five in the hierarchy (see Appendix 1). Three spans of the Waiau River bridge were washed away, while the Conway and Hurunui River bridges were completely destroyed. There was also extensive flooding in Christchurch.

On the 3rd, there was evidence of the development of a low to the west of the country, and by the 5th the low was centred west of Waikato. At the same time pressures were rising to the south, leading to a strong easterly gradient over Canterbury and Marlborough. The low was very slow-moving, and did not start to move away to the northeast of the North Island until the 9th.


Flooding, Market Street North, Blenheim ca 1930s. 1/4-016885-G.

Permission of the Alexander Turnbull Library, Wellington, must be obtained before any re-use of this image.

1924 1-2nd November

Picton: 431mm/45 hours (return period over 150 years)
Ward: 152mm/1 day (return period 70 years)
Flooding followed three days of rain. The rain was produced by an ex-tropical cyclone, and occurred within an easterly airstream.
Heavy falls were not widespread throughout the country.


See National River Flow Record

1931 3-4th April

Rai Valley: 563mm/40 hrs (return period well over 150 years)
incl 290mm/14hrs (return period over 150 years)
Yncya Bay: 343mm/48hrs (return period well over 150 years)
Bainham: 611mm/48hrs (return period well over 150 years)

West Coast:
Karamea 428mm/24hrs (return period well over 150 years)

Stock Losses: Large.
Road Damage: 5400 pounds
2012 Dollars: 540,000.

The Rai River rose by over six metres. Considerable areas in the Rai Valley were flooded, and the township was cut off. Much of the countryside was scarred by slips, and roads were blocked.

A low developed off the Queensland coast. It deepened as it moved southeast, and crossed the central South Island. Northerly gales affected a wide area, with Cook
Strait experiencing its most severe northerly gale for many years. There were large floods in the Hutt and Otaki Rivers in the lower North Island.

1933 31st January

Rai Valley: 512mm/36 hours (return period well over 150 years)

Deaths: One.

A disastrous flood affected the Rai Valley, and one person was drowned trying to ford the Rai River.
The heavy rain was caused by an active front which crossed the area only very slowly, preceded by a north to northwest flow.

1939 25th November-2nd December


Takaka was isolated, with parts of the town under one and a half metres of water. Old settlers stated that the flooding was the heaviest in the history of the district.

Closer to Nelson City, this is the biggest flood on record for the Waimea River.

A blocking high near the Chathams held up a front. In the early stages of the event, the geostrophic flow over the area was northeasterly, and winds in Cook Strait were southeasterly. The low level flow over the Nelson Province was from the eastnortheast. With time, the flow gradually shifted around to the north, as rain continued to fall. The rain had started on the 25th November, and did not cease until the front cleared the area on the 2nd December.

1954 16-17th June

Cobb Dam: 618mm/48hrs (return period well over 150 years)
incl 510mm/24hrs (return period well over 150 years)
This flood saw huge rainfalls in the Cobb Dam area. However, the extreme intensities were not widespread. It also produced the greatest flow in the Wairau River since recordings began in 1936. A wave in the north Tasman Sea was caught up by a trough movingfrom southeast Australia. Rapid cyclogenesis followed, the low moving
southeast towards Fiordland. A strong pre-frontal northwest flow affected the northern South Island for 12 to 14 hours. The synoptic situation (Fig128) was very similar to that of the 1975 event.
idk Fig 128. MSL analysis for midnight, 16th June, 1954.

1967 10th August


Flooding caused in Takaka by this event was variously described as “the worst and most damaging floods in (the region’s) history”, and “the worst since the 1920s”. Many homes in Takaka were invaded by floodwaters.

Blocking was a major factor. An intense anticyclone lay well east of New Zealand. An approaching cold front was slowed in its eastwards progress, before finally pushing through the Golden Bay area on the morning of the 11th. Fig 129 shows the analysis for noon on the 10th.

The 500 hPa flow (Fig 130) shows the typical blocked situation that leads to so many extreme rainfalls. It features a medium wave trough over eastern Australia, with short wave troughs moving through it; and a medium wave ridge east of New Zealand.

worser worser2
Fig 129. MSL analysis for noon 10th August 1967. Fig 130. 500 hPa analysis for noon 10th August, 1967.

1970 30-31st August

Nelson Airport:
137mm/24hr to 9am 30th (return period 50 years)
incl 102mm/6hr to midnt 29th (return period 100 years)
Roding: 233mm/48hrs (return period 65 years)

Deaths: Two.

A long spell of wet weather culminated in two days of incessant rain. All the streams flowing through Nelson City overflowed. One of them, Brook Stream,
increasing its usual flow by more than 100 times, took off in new directions. Property, bridges and servicing pipes were extensively damaged, and one person was drowned. Another died when her house was hit by a slip. A blocking high to the east led to a frontal band parking itself over theNelson area (Fig 131).

The 500 hPa pattern (Fig 132) shows the typical blocked pattern.

brook brook2
Fig 131. MSL analysis for midnight, 29th August, 1970. Fig 132. 500 hPa analysis for midday, 30th August, 1970.

1975 April 1st

Cobb Dam 121mm/17hrs (return period 3 years)
Uruwhenua 198mm/24hrs (return period 8 years)

Onamalutu 157mm/24hrs (return period 25 years)
Waihopai 75mm/24hrs (return period 3 years)
The Branch 127mm/24hrs (return period 20 years)

Damage: Minor

Effects from this flood were not severe, but it did produce the largest ever peak flow into the Cobb reservoir. In fact, the weather situation that produced the high
rainfall was very similar to that of the 1954 flood – also a major event for the Cobb. In the Wairau River, the flow was similar to that of 1954. The cause was a“bomb” that developed from a wave near Lord Howe Island.

Development occurred in the equatorwards entrance to a 150 kt jet, and in an area of strong vorticity advection. (Fig 133).

Figs 134 and 135 show a deepening of the low from 998 hPa to 968 hPa in an 18-hour period.

This is almost two bergerons. (A bergeron is considered the threshold to be called a “bomb”). Deepening rates of this magnitude are rarely seen in the New Zealand area.

A period of very strong northwest winds affected the northern South Island ahead of the front, with the wind strength supplemented by the rapid movement of the low, and by outflow from thunderstorms accompanying the front.

No heavy rain warning was issued for this event.

cobb Fig 133. 300 hPa analysis for midnight 31st March 1975.
cobb2 cobb3
Fig 134. MSL analysis for midnight 31st March 1975. Fig 135. MSL analysis for 6pm 1st April 1975.

1976 9th April

Moss Bush: 151mm/6hrs (return period 111 years)
incl 139mm/4hrs (return period 171 years)
and 92mm/2hrs (return period about 110 years)

Deaths: One.

A small but fairly intense low moved southeastwards over the northern South Island. The rain included a short period of localised very high-intensity rain in the Riwaka area, generated within a convergence zone between northeasterlies and easterlies. A lady standing on
a bridge in the Graham Valley died when it was swept away.

The MSL chart for 9am 9th April is shown in Fig 136. At this time, the wind at Farewell Spit was 070/25kt, while a 20kt easterly prevailed at Nelson City, and there were southerly gales in Cook Strait.

Upward motion was assisted by strong upper divergence, as illustrated in the strongly difluent flow that can be seen over the northern South Island in Fig 137.

There was also CVA over the area.

riwaka riwaka2
Fig 136. MSL analysis for 9am 9th April, 1976. Fig 137. 300 hPa analysis for midday, 9th April, 1976.

1983 15th April

Cobb: 165mm/2hrs (return period well over 150 years)
incl 106mm/1hr (return period well over 150 years)

Successive 30-minute rainfalls at Cobb were 31mm, 38mm, 35mm and 61mm, with a maximum hourly fall of 106mm.

A low moved southeastwards from the northwest Tasman Sea, and an associated front became stationary over the northern South Island for a time. The low
weakened away as it crossed the northern South Island.

However, the extreme rainfall occurred while the low and cold front were still well out in the Tasman Sea.

Fig. 138 shows the mean sea level analysis at the time the heavy rain started. This was obviously a localised convective event.

Fig. 139 shows the area lying in the equatorwards entrance zone to a jet, but apart from that there is little clue to the cause of the extreme rainfall.

cobbd cobbd2
Fig. 138. MSL analysis for midnight, 14th April, 1983. Fig. 139. 300hPa analysis for midnight, 14th April, 1983.

1983 July 8-9th

Uruwhenua 462mm/48hr (return period 140 years)
Patriarch 338mm/48hr (return period over 150 years)
Wairau Valley 225mm/48hrs (return period 130 years)
The Leatham 256.5mm/48hrs (return period over 150 years)

West Coast:
Reefton 214mm/48hrs (return period 110 years)
Upper Mai Mai 216mm/24hrs (return period 100 years)

Number evacuated: 500 plus.
Stock losses: About 26,000.
Damage: $6-7 million
2012 dollars: 18-21 million.

The Wairau flood control scheme failed to cope with the July 1983 event. Floodwaters spilled out over much of the Wairau Plain, causing hundreds of evacuations. The flow had a return period of about 150 years, and was probably comparable to that of the 1868 event.

Residents of Takaka were also evacuated. Here floodwaters were reported to be 30cm deeper than in the 1967 event.

On the West Coast, extreme rainfalls were recorded in the Reefton area and in the Grey Valley. However, floodwaters did not enter Greymouth.

The synoptic situation over the period of this storm featured a complex, slow-moving depression covering the whole of the Tasman Sea, with an anticyclone east of the
country also slow-moving. A narrow band of moist tropical air flowed onto the country for a period of 48 hours during the 8th, 9th and 10th of July.

The mean sea level analysis for midday 9th July, 1983 is shown in Fig 140. A broad frontal cloud sheet covered the whole country (Fig 141).

A large upper level low covered much of the Tasman Sea (Fig 142).

over over2
Fig 140. MSL analysis for midday 9th July 1983. Fig 141. NOAA 7 infrared satellite image taken 4pm 9th July 1983.
over3 Fig 142. 500 hPa analysis for midday 8th July, 1983.

The warmth of the air at low levels made the air unstable, thus assisting the upward motion process.

The direction of the flow was also a critical factor in the magnitude of the flooding. The flow was very northerly, which meant that spillover to the south side of the Wairau River was very high, making it a “whole catchment” event. Strong winds also contributed to high spillover.

The rainfall pattern (Fig 143) shows the typical distribution of rainfall in this area in northerly flows.

A heavy rain forecast was issued for this event, but the amount forecast was only “120mm or more in the next 24 to 36 hours”. The acquisition of radar coverage has been a great help in forecasting for this area.

over4 Fig 143. Isohyets of 48hr rainfall 9am 8th July to 9am 10th July, 1983.

1985 10th January

Roding: 198mm/24hrs (return period 65 years)
incl 141mm/4hrs (return period well over 150 years)
Takaka Hill: 144mm/6hrs (return period 70 years)
incl 133mm/4hrs (return period 130 years)

Homes Evacuated: Two

Nelson residents had to resort to boiling their drinking water after flooding washed away the inlet to the Roding River reservoir. At Pelorus Bridge fifteen campers were evacuated by front-end loader. In Golden Bay Takaka was isolated, two homes at Rockville were evacuated, and hundreds of cattle were washed down the Aorere River.

This was a major event for the Roding area (south of Richmond). The main front, as seen in Fig 144, consisted of an active line of convection. (The second front behind it was weaker.)

A difluent 300 hPa flow (not shown) affected the northern South Island, with a trough lying just west of the South Island.

elec Fig 144. MSL analysis for midnight, 10th January, 1985.

1990 12th August

Moss Bush: 349mm/72hrs (return period 90 years)
incl 314m/48hrs (return period 116 years)
and 224mm/24hrs (return period about 50 years)

Deaths: One.
Evacuations: 45.

Flooding cut roads and damaged homes. A civil emergency was declared for Motueka, and twelve people were rescued from Riwaka by coastguard boat. Access to Golden Bay was cut.

A large low moved from the central Tasman Sea southeastwards to cross the central South Island. Central pressures within the low were down to about 980hPa. A moist northeast flow affected the Nelson area for a considerable period. A warm front crossed the area at 6am 11th. The MSL chart for noon 11th August is shown in Fig 145.

At this time the wind at Stephens Island was 100/35kt, while that at Farewell Spit was 060/30kt. This indicates low level convergence within the Nelsonregion. Nelson City was reporting continuous heavy rain at this time. The small low seen off the West Coast was moving southsouthwestwards, and its presence helped prolong the
period of heavy rain. During the latter stages of the event, several convective bands crossed the area, as the flow turned slowly around to the northwest.

The 500 hPa chart for midnight on the 11th (Fig 146) shows the characteristic blocked pattern. A negative-tilt trough can be seen extending onto northern and central New Zealand from the Tasman Sea.

golden golden2
Fig 145. MSL analysis for noon, 11th August, 1990. Fig 146. 500 hPa analysis for midnight, 11th August, 1990.

2004 17th February

Waikawa 160mm/24hrs (return period 25 years)
incl 85mm/1hr to 8.15am (return period well over 150 years)
and 55mm/30 mins to 7.35am (return period well over 150 years)

Miramar 30mm/1 hour (return period 35 years)
incl 14mm/10mins (return period 60 years)

A convergence zone spread into the Cook Strait area between a cool moist southerly flow and a milder northeast flow. The northeasterly overran the denser southerlies, leading to rapid ascent of moist air. The slow-moving nature of the convergence zone led to some extremely heavy rainfalls. The synoptic situation is seen in Fig 147.

In the Marlborough Sounds, 1000 people were evacuated due to fears that a dam above Picton would burst. Floodwaters swept away camper vans and flooded houses.

In the Wellington Region severe surface flooding occurred in Miramar, and a rock slide smashed into a house in Karaka Bay.

waikawa Fig 147. MSL analysis for 6am 17th February, 2004.


Challands, N., Pointer, M.W., and Quayle, A.M., 1983: The Marlborough- Nelson Bays floods of 8-10 July 1983. New Zealand Meteorological Service Technical
Information Circular 194.
Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
Fenemor, A.D., 1989: Motueka and Riwaka Catchments Water Management Plan. Richmond: Nelson Catchment Board/Tasman District Council.
Nelson Evening Mail, 1867, January 28, National Library of New Zealand, Digital Collections.
New Zealand Gazette, 1923, Vol. II. New Zealand Government.
_________, 1925, Vol. I. New Zealand Government.
_________, 1931, Vol. II. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Pascoe, R.M., 1982: The flood of 1 April 1975 in the Wairau River, Marlborough, New Zealand. New Zealand Meteorological Service Office Note 105.
______A scheme for forecasting spillover rainfall in Marlborough and Wairarapa. Unpublished. Available from New Zealand Meteorological Service.
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Severe Weather Log. New Zealand Meteorological Service.
The Dominion, 11th August 1967.
______, 18th February, 2004.
Thomson P. A. In: One-Eyed and Blinkered, 2001, by Cynthia Brooks. Marlborough Civil Defence.
Williman, E.B., 1993: Wairau River Flood Frequency Analysis. Marlborough District Council

West Coast

The West Coast is the wettest part of the country, lying well within the roaring forties and prone to heavy rain from fronts that pass through at regular intervals. The rainfall is greatly augmented by orographic uplift from the Southern Alps, which rise to 3754 metres. Heaviest rainfall occurs in the northwest flows ahead of these fronts, and in the largest falls the maximum rainfall occurs 10-20km upwind of the highest ground. Extreme rainfalls can occur when the front becomes slow-moving, or when it contains active thunderstorms. Sometimes flooding can result from a number of active fronts passing across the area in the space of several days. The West
Coast holds the official rainfall records for all durations from 12 hours upwards. However, rain is distributed relatively evenly throughout the year, and extreme events are less common than elsewhere in the country.

1872 8-9th February

Deaths: One
This may not have been the biggest flood to inundate the Greymouth CBD, but because it was the first, it no doubt came as a great shock. Seventy buildings were washed out to sea; one, so it was said, with the lights still burning. The river flowed where the business district once stood. In the Buller this flood is thought to have been the highest in European times. In the Taramakau catchment, the whole township of Greenstone was washed away. Spells of heavy rain had been occurring for upwards of a week, and culminated in a 36-hour period of continuous rain commencing on the night of the 7th.
grey Flooding on Mawhero Quay, Greymouth, 1872. PA-o-415-04.
(Permission of the Alexander Turnbull Library, Wellington,
must be obtained before any re-use of this image.)

1926 3-5th November

Damage in Westport to Public Services: 50,000 pounds.
2012 Dollars: 4.5 million.

This flood holds the record for being the worst ever for Westport, the peak being higher than the 1950 flood by some feet. Practically the whole of the town was flooded. This event is likely to have seen a flow in excess of 12,000 cumecs in the Buller River – with the possible exception of the 1872 event, the highest anywhere in New Zealand in the last 150 years. The river broke its banks and flooded practically the whole of Westport.

This was also a major event for Marlborough. Large parts of Spring Creek and Tuamarina were inundated, and the Ferry Hotel at Tuamarina was inundated to a depth of over a metre. In Canterbury, the Waimakariri rose to an abnormal height and inundated much of the surrounding countryside.

The flood followed a wet ten days, culminating in the passage of an active front. Snow melt was also a factor.

1935 20-22nd February

West Coast:
Hokitika 233mm/24 hrs (return period 80 years)

New Plymouth 184mm/24hrs (return period 65 years)
Stratford 243mm/24hrs (return period 70 years)

Damage (West Coast): 3,380 pounds
2012 dollars: 370,000.
Road Damage (Taranaki): 25,517 pounds
2012 dollars: 2.75 million.

It takes a lot of rain to flood Hokitika, the river being usually well contained in its riverbed. However, they got a flood in February 1935. The lower portions of the town were flooded to a depth of 1.2 metres, and many houses and businesses were invaded. The southern span of the Kaniere Bridge over the Hokitika River was swept out to sea. The Ahaura Valley was devastated, with Humphreys estimated to have received 432mm/24 hours (2.22 times the 150-year return period).

This event also affected Taranaki, where flooding was serious and damage widespread. Water swept through business premises in the central part of New Plymouth; 20,000 pounds worth of damage (2.1 million 2012 dollars) was done to stock alone.

Unstable air seems to have prevailed over the country throughout the month, with frequent thundery outbeaks. These continued until the last week, but the later stages of the event were associated with more general rain. A north to northwest flow prevailed over the country during the last two days, and a blocking high to the east caused a succession of fronts to slow considerably as they crossed New Zealand. Fig 148 shows the MSL analysis for the latter stages of the event.

hok Fig 148. MSL analysis for 9am 22nd February, 1935.
Isobars in inches of mercury.
(One inch of mercury = 33.86 hPa.
Central pressure of low is approx. 970 hPa.)

1953 26th March


Described by the inhabitants of Whataroa as the heaviest rain they had ever experienced, 17 inches (432mm) was estimated to have fallen in ten hours (1.57 times the 150-year return period). The Waitangi-taona River changed its course.

Fig 149 shows that a front stalled in its northeastwards progress over the South Island. The low near the Kermadecs (off the far northeast of the map in Fig 149) is
suggestive of blocking, as is the 500 hPa chart for 4pm 26th (Fig 150).

what what2
Fig 149. MSL analysis for midnight 26th March 1953. Fig 150. 500 hPa analysis for 4pm 26th March 1953.
(Heights in hundreds of feet)

1978 26-29th March

Haast: 610mm/24 hours to 9am 27th (return period well over 150 years)

After a whopping 610mm fell in the 24 hours to 9am on the 27th, a state of emergency was declared in Haast. The town was isolated by slips, and SH 73 closed. There were a lot of trampers in the area for Easter, and many people were evacuated by helicopter.

Two active frontal systems passed over the West Coast. The first can be seen over Fiordland in Fig 151. The second front lies attached to a wave depression in the mid Tasman Sea. Close to this second front a ship was reporting thunderstorms. The first front stalled briefly over northern Fiordland, and was caught up by the second. Fig 152 shows the wave depression close to Haast at 6am on the 27th..

haast haast2
Fig 151. MSL analysis for midnight, 26th March, 1978. Fig 152. MSL analysis for 6am, 27th March, 1978.

In the latitudes of Northland the same ship, moving eastwards with the front, was still reporting thunderstorms – active convection was a feature of this event.

Upper air charts showed a northwest jet extending from the central Tasman Sea to the area south of the South Island (not shown), and a very weak shortwave trough in the eastern Tasman Sea. Upper level divergence assisted the frontal uplift. The 500 hPa flow (Fig 153) looks like the classic blocked set-up – however, systems
were mobile. CVA associated with the short wave trough seen in the eastern Tasman in Fig 153 was also a factor. The 500 hPa trough over eastern Australia at midnight on the 26th weakened as it crossed the Tasman Sea, and moved over Westland early on the 29th.

haast3 Fig 153. 500 hPa analysis for midnight, 26th March, 1978.
The flow turned to the northwest again quite quickly, another pair of fronts crossing the area on the 28th, adding more moisture to the already very wet catchments.

1982 10-12th March

Alex Knob 1810mm/3days (return period over 150 years)

The three-day rainfall at Alex Knob up the Waiho Valley is a New Zealand record.

There were hundred-year floods in the Waiho, Whataroa, Paringa and Whanganui Rivers, and a 50-year flood in the Waitangi-taona River. Damage throughout South Westland was extreme and widespread. Franz Josef reported 350mm/13 hours, and 650mm for three days, but with the rain gauge having overflowed. Another gauge, 4km from Alex Knob, overflowed after 1310mm rain.

The MSL charts for midday on the 10th, 11th and 12th are shown in Fig 154. These charts illustrate a rather atypical West Coast heavy rain situation, which usually involves a strong northwest flow. A clue to the causes of the heavy rain can be seen from a closer examination, particularly of the midday 12th analysis. Firstly, the front became slow moving over the area for a time. Blocking was therefore a significant factor. A more significant clue though is the connection of the front to the low between Queensland and New Caledonia. A nephanalysis done at this time (not shown) depicted a broad frontal zone extending from this low and down over the South Island. Humid air from the subtropics was rising through the atmosphere as it moved towards New Zealand, the so-called “warm conveyor belt” – thus creating this broad and active band of cloud. With a trajectory from such low latitudes, this activity was unusually intense.

alex alex2
alex3 Fig 154. MSL analyses for midday 10th (top left), 11th (top right) and 12th (bottom)
March, 1982.

The 500 hPa chart, seen in Fig 155, confirms that air was being dragged from very low latitudes.

A similar pattern – this connection of the front to the tropics or subtropics – is also seen in several Taranaki extreme rainfall events, particularly that of
8-11th March, 1990.

alex4 Fig 155. 500 hPa analysis for midday, 10th March, 1982.

1988 19-20th May

Westport Aero: 272mm/13 days
Hokitika Aero: 221mm/13 days
Greymouth Harbour: 251mm/13 days

Evacuations: 400
Damage: $4 million.
2012 Dollars: 7.5 million.

Most of Greymouth was flooded, with the water over 1 metre deep in places. Both the Grey and Inangahua Rivers had record floods. There was extensive
damage in the Grey Valley.

The flooding followed 12 days of solid rain. Towards the end of the period, northwesterly rain was produced by a low moving southeastwards in the Tasman Sea. A front crossed the area at about noon on the 19th, but with a low remaining in the central Tasman Sea, the flow remained northwesterly and rain continued, before easing around midnight.

1988 13-14th September

Inchbonnie: 462mm/4 days (return period 15 years)
Hokitika Aero: 204mm/4 days (return period 3 years)

Deaths: One.
Damage: $8-10 million.
In 2012 dollars: 14-18 million.

The Greymouth CBD has been flooded several times in its history, often to depths of up to 1.5 metres, including three times in the 1980s. However, this was the worst flooding in Greymouth’s history, with floodwaters reaching 2.5 metres deep. 200 houses were flooded, and in the Grey Valley 500-600 sheep were lost. The flow in the Grey River was estimated to have a return period of greater than 120 years.

Snow melt was a major factor in this event, and rainfalls were not out of the ordinary. A series of troughs affected the area, followed by a particularly active one on the 12th and 13th.


See Alexandra Inundated


See Otago Flooded Again

1997 14-16th December


The lower Grey River reached to near the top of its banks, but recent flood protection was equal to the task, and Greymouth was (just) saved from flooding. 300-350mm fell in the ranges of north Westland and Buller.

A strong moist northwesterly flow ahead of a front dragged a lane of subtropical air onto the northern South Island (Fig 156).

saved Fig 156. MSL analysis for midnight 15th December, 1997.
850 hPa temperature – dashed lines.
700 hPa humidity - shaded.
The air was very unstable, with wet bulb potential temperatures around 18C. Conditional instability from layer lifting was assisted by very low shears – less than
15kt between 850 and 400 hPa – and intense convection resulted. Strong divergence associated with the equatorwards entrance to a jet (Fig 157) enhanced the rainfall.
saved2 Fig 157. 250 hPa windbarbs for midnight 15th December, 1997.
Solid lines – isotachs. Divergence – shaded.
A warning was issued, but rainfall amounts were greater than initially forecast, partly due to the northwesterly flow lasting longer than expected.
The period was extended in subsequent bulletins.


See Alexandra Election-Year Floods

2004 8th January

Cropp Waterfall: 134mm/1hr to midnight 8th (return period over 150 years (?))
Cropp Hut: 105mm/1hr to midnight 8th (return period over 150 years (?))

A major weather system crossed the country, bringing flooding and slips to the West Coast and closing roads. Further afield, a woman was killed in Queen Charlotte Sound when the tent she was sleeping in was blown away, and three Cook Strait power pylons near Molesworth were destroyed. There was also wind damage on the West Coast, with trees brought down. Northwesterly wind damage on the West Coast is generally caused by strong gusts (or tornados) associated with active
convective cells.

The extreme rainfall, which was probably quite isolated, occurred within an instability line on the leading edge of an active cold front. Unfortunately, the satellite image for the time of onset of the extreme rainfall was not available. However, Fig. 158 clearly shows the instability line at 10pm.

The two Cropp stations are located very close together, inland from Hokitika. The return period calculated by HIRDS2 is probably too high.

Fig 159 shows the MSL analysis at the time of the extreme rainfall. The weather system was accompanied by a strong upper trough, with significant upper divergence, and 850 hPa wet bulb potential temperature up to 20C (not shown). No sounding near the affected area is available.

onehour onehour2
Fig. 158. GMS IR image for 10pm, 8th January, 2004. Fig 159. MSL analysis for midnight, 8th January, 2004.


Benn, J.L., 1990: A Chronology of Flooding on The West Coast, South Island, New Zealand, 1846-1990. The West Coast Regional Council.
Henderson, R.D., 1993: Extreme storm rainfalls in the Southern Alps, New Zealand. In: Extreme hydrological events – precipitation, floods and droughts
(Proceedings of Yokokama Symposium, July, 1993) I.A.H.S. Publ. No. 213:113-120.
New Zealand Herald 14th and 15th September, 1988.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Severe Weather Log. New Zealand Meteorological Service.


Flood events in Canterbury occur in two categories. The more common are the northwesterly events, where heavy rain occurs on the West Coast and in the Southern Alps. Significant rainfall can spill over into the headwaters of the large Canterbury rivers, particularly the Rangitata, Rakaia and Waimakariri. The main factors favouring high spillover are unstable air and strong winds normal to the Alps. In these cases, flooding is confined to the immediate environs of the main rivers.

More significant but less common are the easterly events, often associated with a large, slow-moving depression over or near the North Island. These situations can produce widespread rain throughout the region. Surface flooding occurs, as well as flooding in the smaller rivers and water courses. In these cases, a slow-moving front is usually found over the area, and convection is a factor in the most severe.

1868 2-4th February

Nelson City 292mm/48hrs (return period over 150 years)
incl 179mm/24hrs (return period 150 years)

Mt Peel 205mm/24 hrs (return period 150 years)

Deaths: At least 15
Number Evacuated: Hundreds

The storm of 1868 brought widespread flooding from Marlborough to Otago, also to Nelson, and sank several ships. It appears to have been an ex-tropical cyclone. Reports from the Marlborough area were of a period of easterly gales followed by strong northerlies, and atmospheric pressure reaching a minimum of 968 hPa. At New Plymouth, the glass bottomed out at a similar pressure, with southeast gales giving way to a dead calm, then southwest gales, soon afterwards turning to northwest gales. Heavy rain and devastating flooding extended as far south as Otago, where a number of lives were lost, including nine at Totara Station, near the
Waioreka River. The Taieri Plain, from a point about 11 miles from Dunedin, was covered by an almost unbroken sheet of water. Canterbury was the worst affected, with the flooding there still being the highest on record. Christchurch was flooded, with 60cm of water flowing through the Magistrates’ Court, and one and a half metres in Fendalton Road. In Kaiapoi there was nearly two metres of water in some streets. At Kaikoura houses and farm land were washed out to sea. The whole of the Wairau Plain from above Blenheim to the sea was reported to have become “one vast sheet of water”. The Blenheim rainfall recorder had an accident occur to his gauge on one of the days involved, but estimated about 90mm to have fallen there over 48 hours. In Nelson, this was the third year in a row in which major flooding had occurred at the end of January or beginning of February.

Easterly gales led to the sinking of four ships off the Otago coast, with further loss of life. Another ship was lost off the coast of Hawkes Bay in a northeasterly gale. Gales were also reported from the West Coast, with the barometer falling to about 980 hPa. An easterly gale also caused problems for shipping around Auckland.

The approximate track of the low, as inferred from newspaper reports around the country, is shown in Fig 160.

extc Fig 160. Approximate track of the 1868 cyclone.


See Blenheim Under Water


See National River Flow Record
waimak Waimakariri River floods the Main North Road, November, 1926.
(Christchurch City Libraries Image 0090)

1945 20-21st February

Strathmore: 354mm/48 hrs (return period over 150 years)
incl. 234mm/24 hrs (return period 150 years)

Damage: 372,000 pounds
2012 Dollars: 30 million.

One and a half metres of flood water flowed through Temuka, and 23 bridges were either destroyed or rendered impassable after two days of heavy rain.

The charts for noon 18th and noon 19th (Figs 161 and 162) show the complexity of the build-up to the heavy rain event. In particular, note the situation on 19th February. Two depressions lie in the Tasman Sea, while warm air from the subtropics is about to collide with cold sub-antarctic air.

By 20th February (Fig 163) the two lows have combined into one deeper one, while one can imagine a very strong baroclinic zone lying over South Canterbury.

The analyses for noon 20th February (Fig 163) and noon 21st February (Fig 164) show the classic isobaric pattern for heavy rain in the eastern South Island.

temuka temuka2
Fig 161. MSL analysis for noon, 18th February, 1945. Fig 162. MSL analysis for noon, 19th February, 1945.

The intensification of an anticyclone to the south of New Zealand, bringing the cold air northwards and halting any southeastward movement of the low, was an important factor in producing the extreme rainfalls.The chart of rain distribution for this (Fig 165) shows how little orography influenced rainfall intensities. Convection
must have been a major contributor.

Flooding was widespread, with many towns isolated. The flooding was made worse by the fact that the ground was already well saturated.

temuka3 Fig 165. Total rainfall, (inches) 20th and 21st February, 1945.

1957 26-27th December

A month of wet weather culminated in a depression forming off Westland and deepening rapidly as it moved southwards. The northwesterly rains of the associated active front were consequently enough to trigger large floods in the Waimakariri and Rakaia Rivers. The author had shifted house from the settlement of Kainga on the banks of the Waimakariri only six months previously – his ex-residence was flooded to a depth of a metre. The river has not broken its banks since then. In the Southern Alps, December 1957 was the wettest month in over 50 years of records. Arthurs Pass recorded 1194mm of rain during the month, and The Hermitage at Mount Cook 1295mm, including 491mm on the 26th.

1975 13th March


Kaikoura Coast:
Blue Duck about 400mm/24 hrs.
Kaikoura Met 205mm/24 hrs (return period 65 years)
incl 193mm/12 hrs (return period over 200 years)
and 155 mm/6hrs (return period over 200 years)

Christchurch Airport 102mm/24hrs (return period 35 years)
including 98mm/12hrs (return period about 100 years)
and 80mm/6hrs (return period 150 years)

Ruahine Ranges:
Pohangina Saddle 612mm/72 hrs
(return period over 150 years)
incl 389mm/24hrs (return period over 150 years)

Evacuations: Many.

Kaikoura township was isolated and its business area flooded, State Highway 1 and the main trunk railway were closed by 85 slips and washouts, and motorists and goods trains were marooned after Alison passed west of the country during 11-14th March. Kaikoura was subsequently declared a disaster area.
There was also flooding in Christchurch, with families marooned or evacuated.

The ex-tropical cyclone in fact caused damage in many parts of the country, with trees uprooted, house roofs lifted, boats sunk and farms and towns flooded all the way from Northland down to Otago. Greymouth received the greatest wind damage in living memory.

Fig 166 shows the MSL charts at midnight (NZST) on the 11th and 12th March. A strong northeast airstream can be seen flowing onto the South Island. Anecdotal reports suggest that, at a time when the wind was tearing roofs off buildings and demolishing Victoria Park in Greymouth, it was flat calm in Kaikoura, but coming down “in buckets”.

alison Fig 166. MSL analyses for midnight 11th (left) and midnight 12th (right) March, 1975.

Orographic enhancement of the very moist humid tropical air was obviously a major contributor to the heavy rain. As with all extropical cyclones, it is the interaction of the low with other weather systems that leads to an extreme event. In this case, the anticyclone passing south of the South Island helped maintain the very strong pressure gradient on the southeastern flank of the low, and also ingested some cold air into the system. The low weakened as the high moved away.

Fig 167 (A) shows the rainfall for the 24-hour period up to 9am 13th March. Within the 250mm isohyet, the Marlborough Catchment Board gauge at Blue Duck reported a rainfall of over 400mm during this 24-hour period. Fig 167 (B) shows the return period of the 24-hour rainfall. Six and 12 hour return periods were even higher, as can be seen from Figs 167(C) and 167(D).


Fig 167. Rainfall pattern resulting from Cyclone Alison.

A: the total rainfall (mm) over the 24 hours to 9am 13th March on the
Kaikoura Coast.

B: the return period (years) of the maximum 24 hour
rainfall during Cyclone Alison.

C: as in B but for 12 hour rainfall.

D: as in B but for 6 hour rainfall.

1986 12-13th March

Kakahu Bush: 145mm/12 hrs (return period over 150 years)
incl 106mm/6hrs (return period over 150 years)
Rangiora: 101mm/6hrs (return period over 150 years)
Heatherleigh: 167mm/24hrs (return period well over 100 years)

Deaths: One.
Damage: $66 million.
2012 Dollars: 157 million.
Stock losses: Nearly $15 million.
2012 Dollars: 35 million.
Evacuations: 2000.

After surviving several years of drought, all the farmers in the Hakataramea Valley were forced to evacuate their farms when a very intense rainstorm brought flooding to large parts of South Canterbury and North Otago. Many homes in Temuka were evacuated, as was the entire town of Pleasant Point.

Return periods of the 24-hour rainfalls were of the order of 75 years. However, since most of the rain fell in only 12 hours, this was an event with a return period of well over 100 years. In some places total rainfalls were up to 250mm, including a burst of up to 50mm in two hours at the end.

Three features collaborated to make this an extreme event. Firstly, there was a tropical cyclone near the Kermadecs. Warm air from this system spread down onto the South Island in a weak anticyclonic flow. There the air became caught up in the second system – a large, slow-moving depression centred in the Tasman Sea. The third system, a high-latitude anticyclone, then brought cool southeasterly air up the South Island east coast. Convergence between this cool air and the warm tropical air then led to intense convection. The thermal gradient in the lowest 3-500 metres across the convergence zone was 6-7C.

The sequence of MSL charts from 10th to 15th March can be seen in Fig 168.

haka Fig 168. Sequence of daily weather maps at midnight during 10-15 March 1986.

The 300 hPa analysis for 6am 13th March (Fig 169) shows the South Island lying within a strongly difluent flow, in the poleward exit region to a jet stream. Strong upward motion is therefore indicated.

The satellite photograph shown in Fig 170 shows intense convection occurring over the Banks Peninsula/North Canterbury area. Cloud top temperatures are -60 to -70C, indicating cloud top heights of about 13,000 metres. At around this time, Rangiora in North Canterbury was raining at up to 37mm per hour. Timaru can be seen to be lying beneath warmer tops, yet even there, it was raining at up to 19mm per hour. The clouds look less convective there, but synoptic-scale forcing was
obviously very strong. Further south, in South Canterbury and North Otago, the rain was enhanced by the ranges near the coast. The fact that the Hakataramea
Valley (in the lee of the Hunter Hills) was badly affected, was another indication of intense convection. Compare this with the 1985 Coromandel Peninsula event, where high rainfalls also occurred well downstream of the ranges in the South Auckland area.

haka2 haka3
Fig 169. 300 hPa analysis for 6am 13th March 1986. Fig 170. Enhanced infra-red satellite photograph taken at about 3am 13th March 1986.

Rainfalls for the 24-hour period to 9am 13th March are shown in Fig 171, while Fig 172 shows that most of this rain fell in less than 12 hours.

Also of interest in Fig. 172 is how abruptly the rain intensified, and later ceased. As the high to the south moved northeastwards, the flow over Canterbury turned northeast, and the convergence zone broke up.


haka4 haka5
Fig 171. Rainfall distribution 9am 12th to 9am 13th March 1986. Fig 172. Rainfall accumulations from recording raingauges
during 11-13 March 1986.
If weather surveillance radar had been available in March 1986, a better indication of the potential severity of this event could have been given, and a significant proportion of the $15 million worth of stock losses could have been saved. A heavy rain warning was issued, but was too late to be of any real use to catchment authorities. It takes an unusual situation to produce an easterly heavy rain event in Canterbury. Forecasters therefore tend to suffer from lack of familiarity with such situations, although the advent of computer models has eased this problem.

2000 17-19th August

Lees Valley: about 240mm/2 days (return period 125 years)

Stock losses: Large.
Evacuations: Many.
Road damage: $1 million.
2012 dollars: 1.4 million.

Numerous roads (including State Highway 1) were closed, homes were flooded and evacuated, and there were large stock losses over extensive areas of Canterbury, in a flood that reached a 50-year high in the Selwyn River. The main trunk railway line was closed for nearly a week.

The situation was very typical for easterly heavy rain events in Canterbury (Fig 173). Blocking was strong, and a large low developed over the North Island.

ashley Fig 173. MSL analysis for midnight 18th August, 2000.

A narrow cold front moved up onto Canterbury as the low deepened. A strong southeast flow developed over Canterbury, and this was overrun at slightly higher levels by milder and moister east to northeasterlies. Rainfall rates reached 10 mm per hour or so, and with this type of set-up being long lasting, large accumulations resulted. Two-day falls reached 200 mm in many places, with 240mm in the Lees Valley.

Forecasters had actually been expecting colder temperatures, with heavy snow forecast initially for South Canterbury. The heavy rain warning for the north was extended southwards in subsequent warnings. Rain amounts were mostly underestimated, although amounts in the initial warning for North Canterbury turned out to be close to the mark.


Benn, J.L., 1990: A Chronology of Flooding on the West Coast, South Island, New Zealand, 1846-1990. The West Coast Regional Council.
Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
Edie, E.G., C.J. Seelye and J.D. Raeside, 1946: Notes on the Canterbury Floods of February, 1945. New Zealand Meteorological Service Office Note 29.
McGavin, T. : Severe Weather Events Analysis. New Zealand Meteorological Service
Marlborough Express, 1868 February 8 and February 15. National Library of New Zealand, Digital Collections
Mosley, M.P., and C.P. Pearson, 1997: Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society.
Nelson Evening Mail, 1868, February 4 and February 5. National Library of New Zealand, Digital Collections
New Zealand Gazette, 1958, Vol. I. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
North Otago Times, 1868, February 4 and February 7, National Library of New Zealand, Digital Collections.
Otago Witness, 1868, February 8th. National Library of New Zealand, Digital Collections.
Poole, A.L., 1983: Catchment Control in New Zealand. Water and Soil Misc. Pub. 48.
Severe Weather Log. New Zealand Meteorological Service.
South Canterbury Catchment Board and Regional Water Board, 1987: Report on Flood, 13 March, 1986. South Canterbury Catchment Board and Regional Water
Board Publication No. 47.
Taranaki Herald, 1868, February 8. National Library of New Zealand, Digital Collections.
Thompson, C.S., and H.A.L. Osborn, 1986: The South Canterbury Flood of March 1986. New Zealand Meteorological Service Scientific Report 23.
Timaru Herald, 1868, February 5, National Library of New Zealand, Digital Collections.
Tomlinson, A.I., 1975. Cyclone Alison. New Zealand Meteorological Service Technical Information Circular 148.
Wellington Independent, 1868, February 8, National Library of New Zealand, Digital Collections.




Flooding in Otago and Southland is often caused by northwesterly rain spilling over from the Southern Alps and Fiordland mountains. High spillover occurs when the
component of the windflow at right angles to the Alps is strong, and when the airstream is unstable. In Otago, easterly airstreams can produce flood rains, often occurring when a large, slow-moving low lies to the north. One of the events in this chapter occurred in a southerly airstream. This illustrates a
problem in forecasting heavy rain in the south and east of the South Island – events often occur in an unprecedented way, so forecasters can’t draw on past experience.

The Southland Region takes in Fiordland, one of the wettest parts of the country. Heavy rain is common, and many heavy rain events go by unremarked. As with the West Coast the incidence of extreme return-period events in Fiordland is the lowest of anywhere in New Zealand.

1863 25-27 July

A spell of heavy snow in the ranges was followed by unusually warm, torrential, northwesterly rain. The Clutha shot up six metres in a night, while the Shotover rose 11 metres above normal. Arrowtown was inundated by its River Arrow. At least 33 people were drowned, most of whom were gold miners, living in tents or huts along along the river banks. The same winter saw a series of heavy snowstorms, with the death toll from the season unofficially estimated at between one and two hundred.


See Ex-Tropical Cyclone Sweeps Over New Zealand

25th September- 10th October


Deaths: Two
There were two separate episodes of northwesterly rain during this event. It is considered the biggest on record for the Clutha River as far as flow
is concerned. In 1999 the Clutha rose to a higher level, but this was only because of silting caused by the Roxburgh Dam. The Clutha carried houses, bridges and livestock down to the sea. Nearly every town along its banks was inundated. Parts of Queenstown were under two metres of water. Balclutha became part of the Clutha River.

1923 23rd April

Musselburgh: 229mm/24hrs (return period over 150 years)
Sullivan’s Dam: 238mm/24hrs to 6am 23rd (return period over 150 years)
incl 146mm/12hrs to 6am 23rd. (return period over 150 years)

Many residents fled their homes. The South Dunedin/St Kilda/Musselburgh/Tainui area was one big lake. A sawmill in the Leith Valley was demolished and
swept away.

A period of easterlies affected Otago, as a low moved southeastwards across the North Island.

dunedin Flood water in Harrow St, near Dunedin Railway Station, 1923.
(Dunedin City Council Archives Ref. 334/50)

1929 19-20th March

Ross Creek: 279mm/24 hrs. (return period over 150 years)
Whare Flat: 239mm/24hrs. (return period over 150 years)
Sullivans Dam: 236mm/24hrs. (return period over 150 years)
Musselburgh: 104mm/24hrs (return period 35 years)
Oamaru: 156mm/24hrs (return period 150 years)

This was estimated at a greater-than- 200-year event for the Water of Leith, with the flow estimated at 200 cumecs. There was serious flood damage in
Dunedin City. In keeping with the ‘200’ theme, silt, debris and huge boulders deposited on the streets provided work for 200 unemployed for a fortnight. Bridges were swept away.

Surface analyses at the time showed a 1000 hPa low moving eastwards across the South Island.

leith Flooding in the Water of Leith – date unknown.
(Otago Regional Council)

1949 19th March

Queenstown: 98.5mm/15.5 hours to 9.30am 19th (return period 80 years)
124.5mm/24 hours on 18th (return period 90 years)

The flood in the Shotover River was the highest since 1863. There was also flooding in the Clutha, but without serious damage.

A frontal band became slow-moving over the central South Island for a time, before being pushed southwards by a northwest flow. A deepening depression then moved southeastwards over Otago. Extreme rainfalls affected only a small area.

1978 13-14th October

Roxburgh: 116.5mm/24hrs (return period over 150 years)
Balclutha: 115.6mm/24hrs (return period 110 years)
Makarora: 295mm/48hrs (return period 25 years)
incl 240mm/24hrs (return period 30 years)

Stock losses: Thousands.
Evacuations: Many.

Widespread flooding occurred in much of central and south Otago and in eastern Southland. Several towns were affected. Invercargill Airport was flooded by up
to one metre of water, and at Balclutha thousands of animals were drowned. The flow in the Clutha River in this event was claimed to be similar to that of 28 October 1878, with current wisdom putting the 1878 event as the larger of the two.

The synoptic situations for the two events seem to have been similar also, with warm rain and snow melt dominant factors. The rain was brought by a front becoming stationary over the area.

The MSL analysis for midday 13th October is shown in Fig 174, and there was little change to this situation until 9am 14th October, when pressures started to rise
and the front commenced a northeast movement over the country. Upward motion within the front was assisted by a northwest jet, with the axis over Southland containing winds of 100 kt at 500 hPa (Fig 175).

big big2
Fig 174. MSL analysis for midday 13th October, 1978. Fig 175. 500 hPa analysis for midday 13th October 1978.

A blocking high centred just north of the country also contributed by allowing very warm moist air to flow onto the areas just ahead of the front, making for strong baroclinicity within it. Wind shear between the 1000 and 500 hPa levels was 90kt. Small wave depressions crossed the affected area during the event, and strong
convection in some areas gave almost continuous thunderstorms.

The Invercargill sounding for midday on the 13th (Fig. 176) shows the wedge of cold low-level air behind the surface front, with a deep warm and very moist layer overriding it.

24-hour rainfalls in Fiordland exceeded 300 mm, with rain in the headwaters contributing significantly to the subsequent flooding.

The 1878 event may have been produced by a similar synoptic situation, with warm rain and snow melt described as dominant factors.

big3 Fig 176. Invercargill sounding at noon 13th October, 1978.

1980 16-17th January

Gore: 106mm/24 hours (return period 115 years)
Invercargill: 74mm/24hrs (return period 30 years)

Stock losses: Some.
Evacuations: Many.

Although the area was visited by its second “100-year flood” in 15 months, the synoptic situations causing the two floods was very different. Fig. 183 (see the 1984 event) shows the areas where the maximum 24-hour falls exceeded the 50-year return period for the 1978, 1980 and 1984 events. The area for the 1980 event is slightly smaller and displaced a little towards the southwest.

The cause of the rain was the rapid development of an unusually deep depression just east of Otago (see Fig 177). The deepening rate of the low was 1.6 bergeron,
putting it well within the definition of a “bomb”. A period of northwesterlies had brought moist tropical air onto New Zealand. Subsequent strong baroclinicity, along with increasing shear vorticity, were major contributors to the development of the low.

The satellite image (Fig. 178) illustrates the vigour of the system. With rain on this occasion occurring within a strong southerly flow, rainfall
in Fiordland was minimal.

seconds seconds2
Fig 177. MSL analysis for 6am 17th January, 1980. Fig 178. TIROS -N infrared photo taken 3.10pm 16th January 1980.

1980 5th June

Over 1400
Stock Losses: Heavy.
Damage: $30 million
2012 Dollars:136 million.

The biggest flood in the last 150 years in the Taieri River was probably the one in 1868. However, rainfall during this 1980 event was described as the heaviest for over 50 years. The threat of inundation by the Taieri River caused the evacuation of more than 1400 people. Fortunately major flooding did not eventuate. However, there were heavy property and stock losses, and a state of civil emergency was declared for a large part of the Taieri Plain.

The weather pattern was a very typical one for heavy easterly rain in Otago, featuring a large, slowmoving low (Figs 179 and 180).

taieri taieri2
Fig. 179. MSL analysis for midday, 4th June, 1980. Fig 180. MSL analysis for midday, 5th June, 1980.
The 500 hPa analysis (Fig. 181) featured a blocked setup, with a cut-off 500 hPa low over New Zealand.
taieri3 taieri4
Fig. 181. 500 hPa analysis for midday, 5th June, 1980. Farm buildings at Riverside, Outram, 5th June, 1980
(from Otago Regional Council website)

1984 26-27th January

West Arm: 279mm/24hrs (return period 90 years)
Winton: 116.5mm/24 hrs (return period 130 years)
Riverton: 132mm/24hrs (return period over 150 years)

Evacuations: 3720.
Damage: $70-80 million
2012 Dollars: 210-240 million.

Although it affected a different area to the previous two events, this third “100-year flood” in six years was more serious than the previous two, with the true return period well over the 100-year mark. Insurance payouts were higher than for any flood in New Zealand’s history. Thousands of people were evacuated, and floodwaters at Invercargill Airport were up to three metres deep. Flooding affected country areas also – the town of Otautau was devastated – and the entire Southland province was declared a disaster area.

The synoptic situation at the time of heaviest rain, as seen in Fig 182, is, at first sight, very similar to that of the 1978 event.

southland Fig 182. MSL analysis for noon 26th January, 1984.

Strong northwesterlies prevailed aloft and, although not up to 100kt at 500 hPa, as they were in 1978, winds were 150kt at 250 hPa.

As with the 1978 event, the heavy rain was caused by the frontal band stalling over the southern South Island, with major contributions from a very strong, warm, moist pre-frontal flow, and orographic contribution from the Fiordland mountains. Values of precipitable water reached the high 30s for a time, comparable to values which had been seen the day before associated with the same front over Australia. Thus the air reaching the southern South Island probably had its origin in the tropical areas northwest and north of Australia.

A major difference from the 1978 event was the area affected. The very heavy falls (return period exceeding 50 years) occurred in a fairly narrow band in the two cases,
and this band was further west in the 1984 event. Fig. 183 shows the area in which maximum 24-hour rainfalls exceeded the 50-year return period values for all
three storms of October 1978, January 1980 and January 1984.

As with the 1978 event, 24-hour rainfalls in Fiordland exceeded 300mm (peaking at 575mm at Dumpling Hut). Three-day rainfalls exceeded 800mm in some places (Fig. 184).

southland2 southland3
Fig. 183. Areas in which maximum 24-hour rainfalls exceeded the 50-year return period values for the storms of October 1978, January 1980 and January 1984. Fig. 184. Isohyets of 3-day rainfalls, 25-27 January, 1984.

1994 21-22nd January

Makarora 257mm/48hrs (return period 10 years)
Routeburn Stn 255mm/48hrs (return period 10 years)

West Coast:
Colliers Creek 682mm/24hrs (return period well over 150 years – HIRDS1 gives about 100 years)
incl 473mm/12hrs (return period well over 150 years – HIRDS1 gives about 100 years)

Inflows into the Waitaki headwater lakes in this event were larger than in 1995, and flooding could therefore have been worse than it actually was, if lake levels had not been so low. However, Alexandra was still inundated. Lost in all the media attention on Alexandra was that on the West Coast, Colliers Creek was reaching near-record 12- and 24-hour rainfalls. However, the return periods as calculated by HIRDS2 for this rainfall appear to be far too high – even the HIRDS1 calculation seems too high.

The cause was a front that moved only very slowly over Fiordland, blocked by a high east of the country. A southerly spread up the east coast later in the event, bringing widespread rain to eastern areas. Figs. 185 and 186 show the MSL analyses at 6am on the 21st and 22nd.

colliers colliers2
Fig. 185. MSL analysis for 6am, 21st January, 1994. Fig 186. MSL analysis for 6am, 22nd January, 1994.

1995 12-13th December

Southern Alps:
Mt Cook 445mm/48hrs (return period12 years)
Panorama Ridge 840mm/48hrs (return period 80 years)

Makarora 296mm/48hrs (return period 20 years)
Hawea Flat 169mm/48 hrs (return period over 150 years)
Alexandra 103mm/48 hrs (return period over 150 years)

West Coast:
Cropp Waterfall 1049mm/48 hrs (return period over 150 years – HIRDS1 gives 40
Franz Josef 592mm/48hrs to 9am 13th (return period 30 years)

The rain brought abnormally high lake levels which flooded parts of Queenstown and Wanaka. Major rivers flowing out of these and other mountain lakes then flooded many properties downstream, especially in Alexandra, which had the worst flood in its 131-year history. In Westland roads and bridges were damaged and power and phone lines were cut. In Hakataramea a state of civil emergency was declared on the evening of the 14th, just before a stopbank protecting the town was breached.

With this event, Cropp Waterfall station on the West Coast, inland from Hokitika, took the New Zealand record for a 48-hour rainfall of 1049mm. It also took out the record for the wettest calendar month, amassing 2927mm for the month of December. Franz Josef also received its highest ever recorded 48- hour rainfall.

A principal cause was the stalling of a front as it moved over the southern South Island. On the evening of 11th December (Fig. 187) an active cold front crossed
Fiordland and Southland, preceded by a long fetch of northerlies from the subtropics. During the 12th this remained stationary, blocked by a large high to the east, while a weak southerly spread into eastern districts. Overrunning of moist warm air produced high rainfalls in eastern areas on this day. Convection also played a major role, and helped the heavy rain to spill over well east of the divide.

Fig. 188 shows the analysis for midnight on the 12th. On the 13th, the front moved slowly for much of the day, bringing further heavy rain. In the evening, however, a small wave depression moved southeast along the front ending up south of Canterbury, and the southwest flow around the northern side of this low gave the front the push it needed to move northeast.

Snow melt helped turn this into a major flood event.

cropp cropp2
Fig. 187. MSL analysis for midnight, 11th December, 1995. Fig. 188. MSL analysis for midnight 12th December 1995.

1999 15-18th November

Milford Sound 651mm/3 days (return period 40 years)
incl 326mm/24 hrs (return period 7 years)

Southern Alps:
Mt Cook 411mm/3 days (return period over 150 years)

Rees Valley Station 275mm/4 days (return period 100 years)
Makarora 290mm/3days (return period 8 years)
Shotover at Peats 129m/42 hrs (return period 15 years)

Evacuations: Over 1000.

A new record flood struck Alexandra, with about 200 businesses and homes evacuated. Nine homes in the town were destroyed, with water reaching “up to their eaves”. The flood became an election issue when it was revealed that flooding in Alexandra over recent years had been exacerbated by sediment settling in Lake Roxburgh after construction of the hydroelectric dam in 1956. Businesses and homes were also inundated in Queenstown (for a whole week), Wanaka, Balclutha, Kaitangata and Mataura. It was described as a 150- year flood in the Clutha. The level of Lake Wakatipu exceeded its 1878 record by half a metre. The Oreti and Mataura Rivers also reached record levels.

The heavy rain was caused by a front that stalled over the area as a broad active trough approached from the Tasman Sea. A blocking ridge lay over and east of the North Island. The equatorwards entrance zone to a jet developed over the southern South Island, and this, along with a short wave trough, led to an invigoration of the front. Release of potential instability over the Alps produced severe thunderstorms. On the 17th, a wave developed on the trough and delayed the eastward progress of the system over the southern South Island. The original frontal cloud band remained almost stationary over Fiordland for two and a half days. Late on the 16th and early on the 17th, a southerly change pushed northwards over the south of the South Island. The air became progressively colder in that southerly flow, but the upper-level flow remained northwesterly. This resulted in significant overrunning rainfall. The sequence of charts can be seen in Fig. 189.

rox rox2
rox3 Fig. 189. MSL charts for noon 15th (top left), 16th (top
right), and 17th (bottom left) November, 1999.

Fig. 190 shows the strong upper-level divergence over the southern South Island. The equatorward entrance region lay over the southern South Island until late on the
17th before starting to pull away. At about this time the front finally started to make more definite northward progress.

Initial heavy rain forecasts warned of substantial amounts of rain (up to 250mm in 24 hours). However, hydrologists were not prepared for the high intensities that occurred, nor for the rain band to become slow-moving over the area. It is estimated that 600mm of rain fell on the main divide during one 24-hour period. Computer models did not latch onto the slow-moving nature of the system, and, over several successive runs, persisted in predicting a steady eastward movement.

rox4 Fig. 190. 250 hPa analysis for noon 16th November, 1999.
Solid lines – isotachs;
shaded – divergence.

2002 12th January

The Dasher: 251.5mm/36hrs (return period over 150 years)
incl 217mm/24hrs to 6.30pm 12th (return period over 150 years)

There was widespread flooding from Kaikoura to North Otago, with SH1 cut in a couple of places. The Waikouiti River kept the highway closed for 24 hours. Over four days, the rainfall in the Kakanui mountains ranged from 100 to nearly 500mm. However, extreme rainfall affected only a very small area.

A large, slow-moving depression was moving eastwards towards the North Island, with an east-to-northeast flow covering central and southern New Zealand. A frontal rain band moved southwards onto the South Island. However, the extreme rainfall at The Dasher occurred well to the south of the frontal band, and well before any warnings were issued. Fig. 191 shows the flow at various levels at midnight 11th January, along with several superimposed fields. The panel at top left, the MSL analysis, shows an area of high 700 hPa moisture over the southeast of the South Island. The panel at top right shows WBPTs in the Otago area to be around 14C. The bottom left panel shows weak upper-level divergence in the general area of the southern South Island (the 12-hour forecast of upper divergence had an area into the
second level of shading over the north Otago/south Canterbury area). The panel at bottom right shows 700 hPa upward motion over north Otago. In the 500hPa field, an omega block can be seen to the south of the South Island.

dasher Fig. 191. Computer analyses from UKMO model for
1am NZDT, 12 January, 2002.
Top left: MSL, solid lines; 850hPa temperature, dashed lines;
700 hPa humidity, shaded.
Top right: 850 hPa windbarbs; 850 hPa WBPT, shaded.
Bottom left: 250 hPa windbarbs; 250 hPa divergence, shaded.
Bottom right: 500 hPa heights, solid lines; 500 hPa upward motion, shaded.

2005 7th January

Moa Flat: 34mm/30mins to 5.30pm. (return period 130 years)

Local streams flooded, washing over roads and paddocks. Farmers reported over 80mm falling in six hours, with stock losses. There was surface flooding at Ettrick. The heaviest rainfall reported was the 34mm in 30 minutes at Moa Flat, but this site was not at the centre of the rain event.

A weak northwest flow over the area produced high temperatures. Winds were light, leading to slow storm motions.

The Invercargill sounding at midnight, modified for the actual temperatures (Fig. 192) produces over 2000 j/kg of CAPE. Storm motion from the sounding is 8kt, with
surface-6km shear 27kt, and SSI 65. The satellite images (Fig. 193) show development of the cumulonimbus cloud.

herriot Fig. 192. Invercargill sounding at midnight 7th February, 2005, modified for actual mid afternoon temperature/ dewpoint (24/14) in affected area.
herriot2 herriot3
Fig. 193. GMS visible images for 7.00pm (left) and 7.52pm (right) NZDT 7th February 2005.

2005 7th February

Dunedin City: 17mm/15 mins (return period 120 years)

City shops, streets and properties were flooded. In the Dunedin hills, 34 mm fell in 20 minutes (near 6pm).

This was associated with a cold front crossing the area (see Fig. 194).

The forecast sounding produced that morning gave 1400 j/kg of CAPE, with a storm motion of 15 knots. Based on this, forecasters predicted a high risk of localised falls of 10 to 20mm per hour.


Cowie, C.A., 1957: Floods in New Zealand, 1920-53: with notes on some earlier floods. Soil Conservation and Rivers Control Council.
Hessell, J.W.D., and J.A. Renwick, 1980: The Otago/Southland floods of January 1980. New Zealand Meteorological Service Technical Information Circular 178.
________, and J.H.A. Lopdell, 1979: The Southland/Otago floods of October 1978. New Zealand Meteorological Service Technical Information Circular 171.
Hill, H.W., and A.M. Quayle, 1984: The Southland Flood of January 1984. New Zealand Meteorological Service Technical Information Circular 198.
Hutchins, Graeme, 2006: High Water – Floods in New Zealand. Grantham House.
McGavin: Severe Weather Events Analyses. NZ Meteorological Service.
Mosley, M.P., and C.P. Pearson, 1997: Floods and Droughts: the New Zealand Experience. New Zealand Hydrological Society.
New Zealand Gazette, 1929, Vol. II. New Zealand Government.
NIWA, 2002: HIRDS v2.0 – High Intensity Rainfall Design System
Otago Daily Times 23rd and 24th April 1923.
________ 14th January, 2002.
Severe Weather Log. New Zealand Meteorological Service.
Thompson, C.S., and H.A.L. Osborn, 1986: The South Canterbury Flood of March 1986. New Zealand Meteorological Service Scientific Report 23.



Temporal and Spatial Distribution

Fig. 195 shows the distribution of events throughout the year. Of the 122 events, 34 were considered predominantly convective. February is usually thought of as a dry month in New Zealand, and for most places is in fact the driest month of the year. However, as can be seen from the figure, it is also the month for extreme rainfalls. This is due to a combination of the high sea and land temperatures in the late summer. September, with its combination of low sea and land temperatures, has the lowest frequency of extreme events. Also of note in Fig. 195 is that convective events occur well into winter, due to the lag between land and sea temperatures. Then
in November and December, although sea temperatures are still cold, convection starts again due to warming land temperatures.

extreme Fig. 195. Monthly distribution of extreme events.

Only seven ex-tropical cyclones have produced extreme rainfalls in the last 150 or so years. A notable absentee from this list is Giselle, which, while producing hurricane force winds in Cook Strait and sinking the inter-island ferry “Wahine”, did not produce extreme rainfalls.

Events were also plotted on a map of New Zealand. This is not shown, since for most events it was not known precisely where the maximum rainfall occurred. However, the plot did highlight the following:
There were no extreme rainfall events in Fiordland. Rainfall there is the most reliable of anywhere in the country, but variability is low.
There were not many events for Manawatu, although there were a few memorable ones.
Gisborne, Hawkes Bay and Wairarapa showed up strongly, also Marlborough and Nelson.

Something also needs to be said about the decadal distribution. This is shown in Fig. 196, but this tells us more about the New Zealand population over
the years and the growth of the rainfall network than the actual incidence of extreme rainfalls. However, there is an interesting dip in the number of events from the 1940s to the 1960s.

decadal Fig. 196. Decadal distribution of extreme rainfalls.
A look at the New Zealand temperature record for the last 150 years (Fig. 197) suggests that no explanation for this dip lies therein. However, the warming trend from about 1900 does suggest that an increase in the number of events over the years might be expected.
temp Fig. 197. Historical temperature record for New Zealand.

Event Classification

Following Hand et al (2004), the events were classified according to the main forcing mechanism, and plotted on a chart of rainfall depth versus duration. This is shown in Fig 198.

The classifications were as follows:

(a) Convective Events [abbreviated to Conv. in Appendix 3 and marked with a diamond in Fig. 198]: Forcing was either from insolation or a meso-scale feature such as a convergence line or sea breeze.

(b) Convective with Frontal Forcing [abbreviated to C*** in Appendix 3 and marked with a cross in Fig. 198]: Forcing was from a synoptic scale feature such as a front. Of the ten most extreme events (see hierarchy Appendix 1), six were of this category.

(c ) Frontal with Convective Forcing [abbreviated to F*** in Appendix 3 and marked with a triangle in Fig. 198]: Rainfall was widespread and continuous over a large area and clearly associated with a synoptic scale frontal system. There was significant embedded convection characterised by pulses of very heavy rain or

(d) Frontal [marked with a square in Fig. 198]: Rainfall was associated with a synoptic scale frontal system.

(e) Orographic [abbreviated to Orog. in Appendix 3 and marked with an asterick in Fig. 198]: Rainfall was not directly associated with a frontal system or deep convection. The main rainfall mechanism was the prolonged ascent of moist cloudy air over high ground. Such a situation would characteristically give continuous
drizzle near sea level, but orographic enhancement would lead to many hours or even days of steady rain in the hills.

(f) Ex-tropical cyclone [abbreviated to Ex-TC in Appendix 3 and marked with a star in Fig 201].

A comparison of Fig. 198 with Fig. 16 in Hand et al shows that in the New Zealand cases also, each storm type tends to occupy a unique portion of the graph. However, this does not occur in quite such a neat fashion as for the British events. This is not surprising, since there are very few parts of New Zealand for which orography does not play a role, and in some places it plays a very significant role. For example, the 610mm in 24 hours on the 29th March 1987 was a West Coast event, so that while convection on the front played a major role, so did orography. Another notable example of a storm type being out of place was the Kerikeri storm of 19-20th March 1981, where a huge amount of rain fell in a brief convective event. Other anomalies were the 512mm in 10 hours in Hawkes Bay in 1924 (the most
extreme rainfall in the data-base) and the 349mm in 12 hours in 1953. The latter three anomalies all fall in the top five in the Extreme Rainfalls
Hierarchy (Appendix 1).

These four exceptions aside, event totals of over 300mm were all associated with frontal (non-convective) or orographic forcing (plus two ex-tropical cyclones).

hand Fig 198. Plot of rainfall amount versus duration for different event classifications (see text for abbreviations.)
(Off the scale: 1810mm in 72 hours – Frontal))
Note the tendency for storm types to cluster together on the graph.


Many heavy rainfall events occur when a front becomes slow-moving. Often the eastward progress of a front will be stalled by a ‘blocking high’. On many occasions, the front will just weaken away as it moves into the high. However, a certain 500 hPa pattern became a familiar sight to the author during this study. This pattern featured a broad trough over Australia and the western Tasman Sea, and a broad ridge over New Zealand and the Chathams. Examples can be seen in Figs. 56, 106, 108, 130, 132, 146, 150, 153 and 155. Such a large-scale pattern tends to be slow-moving, and is typical of ‘blocking’.

Late January 2005, the 500 hPa pattern shown in Fig. 199 was noted.

blocking Fig. 199. 500hPa anal for midnight, 31st January, 2005.

This kind of pattern brought two weeks of northeasterlies, with fine, warm weather in most places, although in the northern North Island it was very humid with outbreaks of rain. However, early in February, the trough in the Australian Bight amplified as it approached Victoria, leading to the development of a deep, slow-moving low. In the 24 hours to 9am 3rd February, 120mm was recorded in Melbourne, Australia, the highest ever in its 149-year record. The lesson from this situation is that the synoptic-scale systems within the block need to be in the “right” place for extreme rainfalls to occur

Another well-documented blocking pattern is the so-called “omega block.” This pattern is associated with easterly or northeasterly rainfalls and is seen in Figs 31, 60, 84, 181 and 191.

The Southern Oscillation

An attempt was made to find a relationship between extreme rainfall and ENSO (El Nino Southern Oscillation) based on the following
(a). A suggestion in some studies that neutral values of the Southern Oscillation Index (SOI) favour heavy rainfalls.

(b). A study by Environment Bay of Plenty, (Ellery and Blackwood, 2005) that showed a strong correlation between Bay of Plenty floods and negative values of the Interdecadal Pacific Oscillation (IPO). The IPO is a climatic fluctuation with a period of one to three decades–during its positive (negative) phase, El Nino (La Nina) predominates. La Nina years tend to bring an increased frequency of northeasterlies to the North Island, so intuitively, areas such as Bay of Plenty, Coromandel Peninsula, Auckland, and Northland should have more heavy rain events during this phase of the SOI. On the other hand, El Nino events bring an increased west to southwest flow to the South Island, so that during El Ninos, the West Coast and the Southern Alps should have a greater number of heavy rain events.

A comparision of the historical record of extreme rainfalls with values of the SOI and IPO is handicapped by the following facts:

(a) The steady but unrealistic increase (apparent from the data) in the incidence of events over the historical record.
(b) The fact that the incidence of extreme rainfall events may not bear much relationship to the incidence of mere “heavy” rain events. It certainly bears no relationship to average rainfalls. This problem is highlighted by the fact, as already mentioned, that the driest month of the year (February) is also the month with the greatest number of extreme events.


It was found that 75% of extreme rainfall events occurred with neutral values of the SOI (ie between -10 and +10), 14% with positive values, and 11% with negative values. Bearing in mind that neutral values of the SOI occur 69% of the time anyway, and also that the data shows an apparent but probably unrealistic maximum since the year 2000, when the SOI has been neutral, this would suggest that at least on a nationwide basis, the state of the SOI has little bearing on the occurrence of extreme events.

The regional distribution of extreme rainfall events among the three phases of the SOI is shown in Fig. 200. One would expect, from the average flows known to occur with the various Southern Oscillation regimes, that in the first group of four regions, more than 15% of events should occur in association with positive (greater than 10) values of the SOI. In fact, this was true only for Northland/Auckland.

soi Fig. 200. Percentage of events occurring under the different SOI regimes for the regions of New Zealand.

In the second to last group, one would expect that more than 16% of events would occur in association with negative (less than -10) values of the SOI. The 50% in Canterbury represents only one event, and in fact only 10% of events in this area occurred in association with negative values, while 20% occurred in association with positive values.

However, the last group, mainly easterly and southerly rainfalls, show a strong preference for neutral values of the SOI (92%). This appears consistent with the nature of El Ninos and La Ninas, since El Ninos bring increased westerlies to these areas (ie dry conditions) while La Ninas are associated with anticyclonic regimes (also dry conditions).

A relationship with the IPO (also known as the Pacific Decadal Oscillation) was not much easier to find. Fig 201 (from Rodionov) shows monthly values of the IPO since 1900.

ipo Fig. 201. Monthly values of the IPO: 1900-Nov 2005

Using a method described in Rodionov 2004, regime shifts are calculated to have occurred in 1948, 1976, and 1999.

The two areas that are the most interesting to compare with extreme rainfall incidence are the Bay of Plenty and the Southern Alps. The former (latter) should receive a greater number of extreme rainfalls during the negative (positive) phase of the IPO ie (more La Ninas (El Ninos)).

The table shows the number of extreme events for the two areas in each phase of the IPO.

IPO Years Bay of Plenty Events Southern Alps Events
Positive 1900-1948 2 2
Negative 1949-1976 2 2
Positive 1977-1999 3 10
Negative 2000-2006 5 1

Before about 1950, data on rainfall events is probably deficient. That aside, the table bears out the theory reasonably well, particularly since the latest negative phase has only just started, and the Bay of Plenty could accumulate many more events before it ends. However, Northland/Auckland and the Coromandel Peninsula might be expected to show similar trends to Bay of Plenty, and this did not occur (not shown).

Also, according to Rodionov (2006), it is still not completely certain whether or not there really was a regime change in 1999.

Summary of Insights

(A) February is the month with the highest incidence of extreme rain events, while September has the least.

(B) Convection can play a role in events well into winter.

(C) No extreme events show in the record for Fiordland.

(D) Events can be classified according to the main forcing mechanism, and similar types tend to cluster together when plotted on a graph of rainfall amount versus duration.

(E) Blocking is a recurrent theme in extreme rainfall situations.

(F) On a nationwide basis, there is no predisposion of extreme rainfalls towards positive, negative or neutral values of the SOI, apart from in Canterbury and Otago, where easterly and southerly events strongly favoured neutral values.

(G) Bay of Plenty and the Southern Alps show some correlation with phases of the IPO, with the negative (positive) phase favouring Bay of Plenty (Southern Alps) events, while the positive (negative) phases correspond to a paucity of events in the Bay of Plenty (Southern Alps). This correlation with the IPO did not show up anywhere else.

(H) The most extreme events were the convective ones where forcing was from a synoptic-scale feature such as a front.

(I) Of the non-convective extreme events, a strong warm conveyor belt, with air at all levels coming from low latitudes, was noticed on a number of occasions. This tended to occur in combination with blocking.


Ellery, G., and P. Blackwood, 2005: Impacts of the Bay of Plenty Floods of May, 2005. Presented to “Enhanced Weather Services for Decision Making” Seminar,
MetService, Wellington, 3rd Nov. 2005.
NIWA Science, 2005: Past Climate Variations over New Zealand.
Rodionov, S.N., 2004: A sequential algorithm for testing climate regime shifts. Geophys. Res. Lett., 31, doi:10.1029/2004GL019448.
Rodionov, S., J. Overland, and N. Bond, 2006: Pacific Climate Overview – 2005.

Appendix 1

A Hierarchy of Extreme Rainfalls in New Zealand

Included in the following table are all rainfalls which reached or exceeded 1.50 times the 150-year return period value.

Most return periods have been calculated using NIWA’s HIRDS2 program. This does appear to have a few deficiencies, particularly in the very wet areas such as the Coromandel Peninsula and the Southern Alps. In these areas, return periods calculated using HIRDS2 appear to be far too high. In fact HIRDS1 gives much more realistic values for the Southern Alps.

Rissington GSB/HBY 11th March 1924 10 512 2.91  
Coromandel WKO 21st June 2002 0.42 125 2.84 "Weather bomb"
Rissington GSB/HBY 11th March 1924 3 229 2.49  
Maungakotukutuku GSB/HBY 19th February 1938 1 132 2.24

Kopuawhara Disaster (est. from river level)

F. Hunt NLD/ALD 19-20th March 1981 9.5 448 2.22 Kerikeri floods
Keinton-Combe CNY 5-8th May 1923 12 330 2.10  
Mangarouhi Valley GSB/HBY 27-28th January 1953 12 349 2.06  
Tutira GSB/HBY 11th March 1924 3 203 2.03  
Puketitiri GSB/HBY 23-25th April 1938 72 1001 2.00 Esk Valley floods
Putorino GSB/HBY 23-25th April 1938 72 815 1.99 Esk Valley floods
Emscote Stag & Spey CNY 5-8th May 1923 48 775 1.95  
Whatatutu GSB/HBY 31st March 1910 60 584 1.95  
Porangahau GSB/HBY 4th May 1941 24 406 1.93  
Cobb NSN/MRB 15th April 1983 2 165 1.87  
Emscote Stag & Spey CNY 5-8th May 1923 96 934 1.83  
Brynderwyn Hills NLD/ALD 30th June 1997 2 165 1.81  
Pukeorapa GSB/HBY 18th March 2005 1 99 1.79  
Taita WGN/WRP 20th December 1976 8 230 1.78  
Tarata TKI/WGI/MNU 23-24th February 1971 24 461 1.77  
Kauranaki GSB/HBY 27-28th January 1953 9 224 1.75  
Colliers Creek WEST COAST 21-22 January 1994 12 473 1.73  
Cobb NSN/MRB 15th April 1983 1 106 1.73  
Waikawa NSN/MRB 17th February 2004 1 85 1.72  
Whenuapai NLD/ALD 16th February 1966 1 107 1.69  
Colliers Creek WEST COAST 21-22 January 1994 24 682 1.69  
Pukekohe NLD/ALD 21st January 1999 3 160 1.68  
Haast WEST COAST 26-29th March 1978 24 610 1.66  
F. Hunt NLD/ALD 19-20th March 1981 2 174 1.62  
Glenross/Waimata GSB/HBY 6-8th March 1988 96 917 1.60 Bola
Matauri Bay NLD/ALD 21st April 2000 1 124 1.60  
Waikawa NSN/MRB 17th February 2004 0.5 55 1.59  
Cobb Dam NSN/MRB 16-17th June 1954 24 510 1.57  
Awakino NLD/ALD 21st January 1999 4 160 1.57  
Paitu NLD/ALD 23rd February 1974 6 258 1.56  
Thorndon Bay WKO 21st June 2002 2 150 1.55 "Weather Bomb"
Putaruru WKO 21st June 2002 2 120 1.55 "Weather Bomb"
Keinton-Combe CNY 5-7th May 1923 24 366 1.54  
Leigh NLD/ALD 30th May 2001 1 109 1.53  
Pukeorapa GSB/HBY 18th March 2005 0.5 56 1.53  
Cobb Dam NSN/MRB 16-17th June 1954 48 618 1.51  
Hastings GSB/HBY 10th January 2002 1 77 1.51  
Rai Valley NSN/MRB 4-5th November 1931 40 563 1.50  

Appendix 2

Chronology of Extreme Events

* well-founded estimate
**actual max intensity for event probably considerably more

17 Jan 1858 Hutt Valley ? ? ?
25-27 Jul 1863 Otago ? ? Orog
1-3 Feb 1866 Nelson City ? ? ?
27-28 Jan 1867 Nelson City 239 24 F***
25 May-4 Jun 1867 Napier 381** 96 Frontal
2-4 Feb 1868 Canterbury 205** 24 Ex-TC
8-9 Feb 1872 Greymouth ? ? ?
7 Feb 1877 Motueka ? ? ?
25 Sep-10 Oct 1878 Otago ? ? Orog
24 Mar 1880 Wairarapa ? ? Frontal
4 Dec 1893 Hawkes Bay ? ? ?
14-17 Apr 1897 Tikowhai, Hawkes Bay 507 96 Frontal
17 Jun 1898 Hutt Valley ? ? Frontal
13-15 Jun 1902 Ormondville, N Wairarapa 281 72 Frontal
22-25 May 1904 Whanganui 102** 72 Frontal
14-17 Jul 1906 Gisborne ? ? ?
10-23 Jan 1907 Waikato 125** 24 Frontal
30 Mar-1 Apr 1910 Gisborne City 365** 72 Ex-TC
26 Jan-6 Feb 1917 Taheke, Northland 231 24 Frontal
13 Jun 1917 Tutira, Hawkes Bay 427 48 Frontal
22 Apr 1923 Dunedin 230 24 Orog
5-8 May 1923 Emscote etc, N Canterbury 934 96 Orog
11 Mar 1924 Rissington, Hawkes Bay 512 10 C***
1-2 Nov 1924 Picton 431 45 Ex-TC
17-19 Dec 1924 Masterton 179 24 F***
3-5 Nov 1926 Buller/Westland ? ? ?
19-20 Mar 1929 Ross Creek, Dunedin 279 24 Orog
3-4 Apr 1931 Rai Valley, Marlborough 563 40 Orog
28-30 Aug 1932 Putara, Wairarapa 370** 72 Orog
31 Jan 1933 Rai Valley, Marlborough 512 36 Frontal
15 Feb 1935 Auckland City 89 2 Conv
20-22 Feb 1935 Hokitika 233** 24 Frontal
1-2 Feb 1936 Whangarei 290 24 Ec-TC
19 Feb 1938 Kopuawhara, Gisborne 132* 1 C***
23-25 Apr 1938 Puketitiri, Hawkes Bay 1001 72 Orog
25 Nov-2 Dec 1939 Nelson District ? ? Frontal
11 Dec 1939 Kelburn, Wellington 127** 13 Frontal
24-25 Feb 1940 Tangarakau, Taranaki 218 24 Frontal
4 May 1941 Porangohau, Hawkes Bay 406 24 Frontal
22 Feb 1944 Rotorua 146** 24 Frontal
20-21 Feb 1945 Strathmore, S Canterbury 354 48 Frontal
27-29 Jun 1947 Martinborough, Wairarapa 182 24 Orog
17-18 Apr 1948 Tauranga 212 6 F***
14 May 1948 Whatatutu, Gisborne 344 72 Frontal
19 Mar 1949 Queenstown 124 24 C***
27-28 Jan 1953 Mangarouhi Vly, Hawkes B 349** 12 C***
26 Mar 1953 Whataroa, Westland ? ? Frontal
7 Jul 1953 Hamilton 152** 96 Frontal
18-20 May 1954 Te Aroha 467 72 Orog
16-17 Jun 1954 Cobb Dam, Nelson 618 48 Frontal
26 Dec 1957 Southern Alps ? ? ?
15-25 Feb 1958 Rangipo Prison, Taupo 305 48 Frontal
2-4 Jun 1963 Tareha, Hawkes Bay 310** 48 F***
9-11 Mar 1964 Mt. Pirongia 277 48 Frontal
16 Feb 1966 Whenuapai 107 1 Conv
10 Aug 1967 Golden Bay, Nelson ? ? Frontal
11-14 Aug 1970 Te Teko, Bay of Plenty 262 24 Orog
30-31 Aug 1970 Nelson City 102 6 Frontal
23-24 Feb 1971 Tarata, Taranaki 461 24 Frontal
18-19 Apr 1971 Taipuha, Northland 250* 2.5 C***
22-23 Feb 1974 Paitu, Northland 134 2 C***
13 Mar 1975 Kaikoura 155 6 Ex-TC
1 Apr 1975 Onamalutu, Marlborough 157** 24 Frontal
30 May 1975 SE of Whangarei 200* 2.5 C***
9 Apr 1976 Moss Bush, Nelson 139 4 C***
20 Dec 1976 Taita, Hutt Valley 238 8 C***
26-29 Mar 1978 Haast 610 24 F***
13-14 Oct 1978 Roxburgh 116 24 Frontal
20-21 Mar 1979 Te Puke 375 48 F***
16-17 Jan 1980 Gore 106 24 Orog
5 Jun 1980 Eastern Otago ? ? Frontal
19-20 Mar 1981 F. Hunt, Kerikeri area 448 9.5 Conv
12-14 Apr 1981 Waihi 521 72 Orog
10-12 Mar 1982 Waiho Valley, Westland 1810* 72 Frontal
15 Apr 1983 Cobb Dam, Nelson 165 2 Conv
8-9 Jul 1983 The Leatham, Marlborough 256 48 Frontal
26-27 Jan 1984 Riverton, Southland 132 24 Frontal
10 Jan 1985 Roding, Nelson 198** 24 F***
16-17 Feb 1985 Coromandel Township 266 8 F***
12-13 Mar 1986 Kakahu Bush, Canterbury 106 6 F***
6-8 Mar 1988 Glenross/Waimata, Gisborne 917 96 Ex-TC
19-20 May 1988 Westland ? ? Orog
13-14 Sep 1988 Inchbonnie, Westland 462** 96 Orog
8-11 Mar 1990 L Mangamahoe, Taranaki 294 24 Frontal
12 Aug 1990 Moss Bush, Nelson 314 48 Orog
21-22 Jan 1994 Colliers Creek, Westland 682 24 Frontal
20-21 Apr 1995 New Plymouth 134 5 C***
12-13 Dec 1996 Cropp Waterfall, Westland 1049 48 Frontal
29-31 Dec 1996 Golden Cross, Coromandel 372 24 Ex-TC
30 Jun 1997 Brynderwyn, Northland 165 2 C***
14-16 Dec 1997 Buller/North Westland ? ? F***
26 Jun 1998 Kelburn, Wellington 50 1 C***
1-20 Jul 1998 Matangitangi, Waikato 203 72 Frontal
21 Jan 1999 Awakino, Northland 160 4 C***
30 Apr-2 May 1999 Rotorua 169 5 F***
15-18 Nov 1999 Mt Cook 411 72 Frontal
21 Apr 2000 Matauri Bay, Northland 124 1 Conv
28-29 Jun 2000 Awanui, Northland 205 24 F***
17-19 Aug 2000 Lees Valley, Canterbury 240 48 Orog
30 May 2001 Leigh, Auckland 109 1 C***
9 Dec 2001 Napier & Hastings 50** 1 C***
10 Jan 2002 Hastings 77 1.5 Conv
12 Jan 2002 The Dasher, Otago 217 24 Orog
21 Jun 2002 Coromandel Township 125 0.42 C***
25-28 Feb 2003 Waikura, Gisborne 438 48 Frontal
6 Apr 2003 Te Puke 156 6 Frontal
3-4 Oct 2003 Paekakariki, Kapiti Coast 120 3 Frontal
26 Nov 2003 Hamilton 56 1 Conv
8 Jan 2004 Cropp Waterfall, Westland 134 1 F***
15 Feb 2004 Dannevirke 225 24 Frontal
17 Feb 2004 Waikawa, Marlborough 55 0.5 Conv
28-29 Feb 2004 Mangatoetoe, Taupo 289 36 Orog
17-19 Jul 2004 Whakatane 249 48 Frontal
18 Oct 2004 Tamatea, Napier 179 3 C***
7 Jan 2005 Moa Flat, Otago 34 0.5 Conv
7 Feb 2005 Dunedin 17 0.25 C***
18 Mar 2005 Pukeorapa, Hawkes Bay 99 1 C***
30 Mar 2005 Castlepoint, Wairarapa 115 3 F***
3 May 2005 Tauranga 67 1 Conv
18 May 2005 Awakaponga, Bay of Plenty 95 1 Conv
16-17 Jul 2005 Pauanui, Coromandel 215 12 F***
21 Oct 2005 Hikuwai, Gisborne 273 15 Frontal

Appendix 3

Severe Convection

Severe convection plays a major role in many heavy rain events. Thunderstorms can occur in single cells, multi-cells, or so-called super cells. They can also occur in large clusters, and in fronts. Thunderstorms are generally responsible for the short-period extreme rainfalls, and play a major part in many of the longer-period extreme events. Up until recently, scattered thunderstorm activity was considered too difficult to predict, and heavy rain forecasts were attempted only when an area in excess of 1000 square kilometres was expected to be affected. However, in 2005, a dedicated severe convection shift was instituted at MetService. As well as heavy rain, the possibility of hail, severe straight line winds from convection, and tornados are now considered. Those parts of the country likely to be affected are assigned a probability (low, moderate or high) with the forecasts posted on the MetService website.

The fundamental tool for forecasting convection is the sounding of temperature, dew-point, and wind, as determined from weather balloons released at least twice per day from various parts of the country. There are only three sites in New Zealand where temperature soundings are made (Whenuapai, Paraparaumu and Invercargill), so we need to interpolate between these to get an idea of the atmospheric regimes at other places. Also, we need to know what the soundings are likely to look like at some future time. Computer models have become better and better at this. Although there is still a lot of room for improvement, forecast soundings are now good enough for useful convection forecasts to be made.

Beginners Guide to Using Tephigrams for Convection Forecasting

A parcel of air is lifted from the surface, possibly by flowing over a hill, or within a convergence zone. As it rises, its temperature falls off at the “dry adiabatic” lapse rate (sloping brown “potential temperature” lines on the tephigram – see Fig 202).

tephig Fig 202. The tephigram – temperature, dewpoint and wind sounding through the atmosphere

At the same time, its dew point falls, but not as rapidly as temperature, travelling up the mixing ratio (dashed green) lines. When the parcel’s temperature reaches its dewpoint, it becomes “saturated”. Whether it continues to rise will probably now be dependent on whether or not it is warmer than the surrounding air (or it may be forced up over a hill). If it continues to rise, it does so at the moist adiabatic lapse rate (solid green lines). The air will continue to rise until it is no longer warmer than the surrounding air. Because it will have some velocity at this point, it will “overshoot”.

Tephigram Parameters

The positive area between the parcel trajectory and the environmental temperature is the Convective Available Potential Energy (CAPE) (coloured red in Fig 202).

The green area in Fig 202 is equal to the red area, and represents the loss in kinetic energy of the air parcel as it overshoots.

The temperature difference between the lifted parcel and the environmental temperature at the 500 hPa level is the Lifted Index (LI).

The temperature difference between the lifted parcel and the environmental temperature at 700 hPa is the SLI700.

The altitude at which the environmental temperature reaches 0C is the freezing level.

The wet bulb temperature is the temperature a parcel of air would have if its moisture level was increased to saturation. This is roughly mid-way between its temperature and dewpoint. The altitude at which the wet bulb temperature reaches 0C is called the WBZ (wet Bulb Zero).

Heavy Rain Thunderstorm Sounding

A typical sounding for a thunderstorm producing heavy rain is shown in Fig 203.

This example is from an Auckland thunderstorm on 21st January, 1999. It features:

(1) “Tall skinny” CAPE. For heavy rain to occur, CAPE should be moderate to high, in the 1100-2000 j/kg range, indicating a low warm cloud base and a high equilibrium level.

(2) The sounding should show a deep moist environment. This is not so obvious in this example – however, locations close to water have a source of moisture which can readily be entrained into the atmosphere.

(3) A large lifted index (LI) in the range –4.5C to –6.0C.

(4) A high freezing level of greater than 3000 metres.

(5) A high WBZ of greater than 2500 metres.

(6) Slow storm motion of less than 10 knots (or a succession of cells moving over the same area – the so-called train effect).

(7) Weak environmental wind shear.

(8) High relative humidity near the ground.

A couple of personal observations are made at this point, based on the editor’s experience in forecasting convective rainfalls.

(A) A sounding or forecast sounding is not the last word when it comes to forecasting rainfall intensities. The role of any forecast dynamic processes must also be considered and added.

(B) Mesoscale models can be very good at picking up zones of moisture convergence (although not always in the right place). High intensity rainfall appearing in a mesoscale model can be treated as a “first alarm” level for severe convection. If the 100km or 60km models are indicating dynamic support in the form of strong upward motion, CVA, or upper divergence, then high intensity rainfall can be forecast with a fair amount of confidence. With the CHAMP12 model, for example, a forecast of 30-50mm per three hours, along with dynamic support from the coarser models, has been seen more than once to lead to actual rainfall intensities of 30-50mm per one hour. Sometimes these rainfall rates are accompanied by little or no thunderstorm activity.

(C) Conversely, if CHAMP12 is forecasting high intensity rainfalls without dynamic support, there is a good chance that it will be wrong. Many of the events described above were the result of severe thunderstorm activity. For the historically earlier cases, soundings are, of course, not available. Even for some of the later cases, (in particular the record Leigh rainfall of May 2001) the nearest soundings were not close enough to be representative of the airmass within the storm.

tephig2 Fig 203. Whenuapai sounding for midday, 21 January, 1999


© Mark Pascoe 2012