View all publications

2 Projections of future New Zealand climate change continued

2.2.3 Daily temperature extremes

Daily temperature extremes (overnight minimum and daily maximum) will also vary with regional warming, in addition to changes in mean temperature. Box 2.1 illustrates that small changes in the mean temperature value can have a potentially large effect on how often a specified high temperature is exceeded, or how often temperatures below a certain low value (such as 0°C) occur.

Figure 2.8 is an example of how the frequency of frosts, and of hot days above 25°C, could change for two of the SRES scenarios (B2 and A2) that have so far been run using the NIWA regional climate model. Maps were generated by examining the time series of simulated daily maximum and minimum temperatures, and counting the number of times extremes occur. Larger changes in extreme temperatures occur for the higher emission scenario (A2).

Figure 2.8: Change in the number of days per year with extreme temperatures, between a control run (1980–1999) and two future climate runs under the B2 (left) and A2 (right) scenarios for 2080–2099. Top panel: days below freezing; bottom panel: days above 25°C.

Top panel: This panel has 2 maps.

The first map shows change in the number of days per year with minimum temperatures below freezing under the B2 scenario for 2080-2099. Between 40 and 70 less days per year with a minimum temperature below freezing are projected for Central Otago, Westland, and the region surrounding the northern end of the Southern Alps. Between 20 and 40 less days per year are expected for Canterbury, Southland, the north-west of the South Island and the central North Island. Between 0 and 20 less days are projected for the rest of the North Island.

The second map shows change in the number of days per year with minimum temperatures below freezing under the A2 scenario for 2080-2099. This map is fairly similar to the first but shows even greater reduction in days per year with minimum temperatures below freezing particularly in the central North Island, much of South Westland and along the Southern Alps and the north-western tip of the South Island.

Bottom panel: This panel also has 2 maps.

The first map in this panel shows change in the number of days per year with maximum temperatures above 25 degrees Celsius under the B2 scenario for 2080-2099. An increase of up to 40 days per annum is shown for all of the South Island and for the lower half of the North Island. For the Bay of Plenty, Waikato, and Northland increases of up to 60 days per annum are projected.  

The second map in this panel shows change in the number of days per year with maximum temperatures above 25 degrees Celsius under the A2 scenario for 2080-2099. In this map, patches of the eastern South Island around Canterbury and Blenheim have increases of up to 50 in the number of days over 25 degrees Celsius. In the North Island, increases of up to 70 days are expected for Northland, Auckland, and parts of the Waikato and Bay of Plenty. For the rest of the North Island increases of between 10 and 50 days per year with maximum temperatures above 25 degrees Celsius are expected.

Note: Maps generated from NIWA regional climate model simulations using the UK Met Office Unified Model. The emission scenarios B2 and A2 straddle the A1B scenario (see Figure 2.1). All changes in the top panel are negative, indicating fewer days of air frost in the future. All changes in the bottom panel are positive, indicating more hot days.

There are large decreases in the number of frost days in the central North Island and in the South Island16 (upper panel). For example, in the central plateau of the North Island, the number of air frosts is projected to decrease by around 25 days or so per year for this particular simulation under the B2 emission scenario, and by a few more with the A2 scenario by the end of the 21st century. For comparison, there are typically 30–40 frost days per year currently in this part of the North Island (away from the actual alpine areas such as Ruapehu).

A substantial increase is projected for the number of days above 25°C, particularly at already warm northern sites. For example, for Auckland under the B2 scenario, an additional 40 days or more per year by the end of the century are projected with the maximum daily temperature exceeding 25°C. Under the A2 scenario, this becomes more than 60 extra hot days. For comparison, Auckland currently has approximately 21 days per year with maximum temperature exceeding 25°C. Current values for other locations include: 26 days for Hamilton, 3 days for Wellington, and 31 days for Christchurch.

2.2.4 Heavy rainfall

A warmer atmosphere can hold more moisture (about 8% more for every 1°C increase in temperature), so there is an obvious potential for heavier extreme rainfall under global warming. The IPCC, in its Fourth Assessment Report, declared that more intense rainfall events are “very likely over most areas”.17 The mountainous nature of New Zealand, with its starkly contrasting rainfall climates, makes it difficult to be sure if this situation is universally applicable across the country. Any change in the mix of circulation patterns will have a major impact on the spatial distribution of precipitation.

An early study18 on New Zealand changes in extreme rainfall suggested that by 2030 there would be “no change through to a halving of the return period of heavy rainfall events” and by 2070 “no change through to a fourfold reduction in the return period”. The return period19 is the probable number of years between events with rainfall exceeding some specified high value.

More recent climate model simulations confirm the likelihood that heavy rainfall events will become more frequent. Studies have suggested empirical adjustments to historical rainfall distributions20 that can be applied to estimate a range of possible changes in extreme rainfall under global warming for a particular site. For example, for Auckland, the worst case (most severe) end of the range for 2100 indicates that a rainfall amount currently with a return period of 50 years (AEP=0.02) would have a return period of less than 10 years (AEP>0.10) by 2100 (see Appendix 3 for details). The same approach could be applied to other New Zealand sites with long rainfall records.

Preliminary analyses have now been made of changes in extreme rainfall, based on runs of the NIWA regional model under the B2 and A2 emission scenarios. Extreme value theory was applied to 30 years of daily rainfall data for the model ‘control’ climate (30 years ending 1999), and comparisons were made with the changed climate (ending 2099). For extremes with return periods of 30 years and longer, the increase in rainfall depth was approximately 8% per 1°C of local warming, when averaged over the entire country. This figure (8% per 1°C) matches the increased moisture content of the atmosphere with warming, and is the value widely accepted as a reasonable upper limit for heavy rainfall changes, provided the circulation patterns remain essentially the same.21

However, changes in extreme rainfall were not geographically uniform: in some parts of New Zealand, increases well in excess of 8% per 1°C of warming were found; in other parts of the country, there were decreases (ie, the change per 1°C of warming was negative). The pattern of changes in extremes can be related to the simulated change in storm track frequency and in intensity of cyclones crossing the country. This variation in circulation is very likely to be model-dependent, at least if it follows a similar pattern to the change in mean rainfall. Moreover, even a 30-year run for a particular model is probably not long enough to get stable statistics. Thus, it is recommended that the same changes to rainfall return periods be applied everywhere across New Zealand. It may be possible in future to have projections for regionally varying changes, but this will require comprehensive modelling studies. The recommended adjustment factors are given in Table 5.2, with a worked example in Appendix 4.22

Increased rainfall intensity has obvious implications for increased flooding. The Fourth Assessment Report summarises a number of international studies23 that analyse the increased risk of floods in a future warmer climate. In a recent New Zealand study24, three storm events for the West Coast Buller catchment were modelled, both for the current climate as well as for three different scenarios of temperature increase. Rainfall increased on average (for the three storms) by 3%, 5% and 33% for temperature changes of 0.5°C, 1.0°C and 2.7°C, respectively. Averaged over the three storm events, peak river flow increased by 4%, 10% and 37%, respectively, for the three temperature scenarios. Using these factors to modify the 1-in-50-year design storm, flooding was estimated to increase from 4% of the township being inundated under the current climate, to 13%, 30% and 80% for each of the temperature respective scenarios.

2.2.5 Snowfall and snowline

It is generally expected that snow cover will decrease and snowlines rise as the climate warms. However, there are confounding issues. As stated in section 2.2.4, warmer air holds more moisture, and during winter this moisture could be precipitated as snow at high elevations. There could also be instances of increased winter snowfall to low elevations, for the same reason. However, with the expected increase in temperatures, any snow cover will melt more quickly, and thus the duration of seasonal snow lying on the ground is expected to be shortened.

Figure 2.9 shows an example projection of snow amount changes from the NIWA regional climate model, run under the A2 emissions scenario. There are decreases in seasonal snow almost everywhere. This is particularly evident in the South Island, but decreases also occur over the North Island central plateau (although changes here are too small to be visible with the contour interval used in Figure 2.9).

A decrease in winter snowfall and an earlier spring melt can cause marked changes in the annual cycle of river flow. Some analyses of seasonal river discharge and flood magnitude have been carried out for major rivers of the world,25 but not for New Zealand as yet.

Figure 2.9: Change in winter snow (in kg/m2) between a control run (1980–1999) and a climate simulation under the A2 scenario (2080–2099).

Note: The contour intervals are not equally spaced. Changes in snow amount smaller than 1 kg/m2 are not shown (ie, white space). The snow amount is that lying on the ground averaged over the season, which is not the same as the average snowfall over the season. The unit 1 kg/m2’corresponds to a water equivalent of 1 mm rainfall.

Figure 2.9 represents changes in snow amount between a control run and a future climate simulation under the A2 scenario and shows decreases in seasonal snow cover almost everywhere across the country. These changes are particularly marked in the South Island.

2.2.6 Sea level

The rise of sea level26 around New Zealand is likely to be similar to the global projections of sea-level rise by the IPCC Fourth Assessment, 2007. This statement is based on the similarities between the New Zealand average and the global average over last century of around 1.8 mm/year. Sea-level rise will continue for several centuries even if greenhouse gas emissions are reduced.

Using the same approach as for global temperature change, the IPCC projects that mean sea level will rise by at least 18–59 cm by the 2090s (2090–2099 average) from the 1990s (1980–1999 average), taking the full range of SRES scenarios into account. A further 10–20 cm rise above current levels would occur if melt rates of Greenland and Antarctica were to increase linearly with the future temperature increases. The IPCC notes that even larger sea-level rises cannot be excluded, but no consensus was possible because of limited understanding of the processes involved. They say “The projections do not include uncertainties in carbon-cycle feedbacks nor the full effects of changes in ice sheet flow, therefore the upper values of the ranges are not to be considered upper bounds for sea level rise.”27 More information about sea-level change, its likely impacts and possible response options, is provided in the companion Coastal Hazards & Climate Change. A Guidance Manual for Local Government in New Zealand (Ministry for the Environment 2008).

2.2.7 Wind patterns

It is expected that the annual mean westerly wind component across New Zealand will increase this century.28 As shown in Box 2.2, this ‘mean westerly’ is built up from conditions when the actual east–west wind component is sometimes westerly (positive) and sometimes easterly (negative). The average southerly component is built up similarly from the north–south wind component.

Box 2.2: Westerly component of the wind across New Zealand

The solid arrows represent individual wind speed and direction values. The dashed horizontal arrows are the ‘westerly components’ of these winds – positive when they point to the right and negative when they point to the left. The ‘mean westerly component’ is the average of these individual westerly components. It is 1 m/s for the three periods shown above (ie, average of +3, –2 and +2). If the middle period of easterlies is excluded, the mean westerly component increases substantially to 2.5 m/s, even though the mean wind speed (average length of solid arrows) is hardly changed.

Table 2.6 provides projected changes in the seasonal and annual average westerly and southerly components of the flow across New Zealand. These are broad-scale changes only, calculated from model changes in the Auckland–Christchurch pressure difference (related to west–east wind flow) and in the Hobart–Chathams pressure difference (north–south wind flow). Changes determined from the 12 General Circulation Models under the A1B scenario were re-scaled to span the six SRES emission scenarios, as for the temperature and precipitation tables. The average change is shown for each season, as well as the range over all models and emission scenarios.

A strong seasonality is apparent in the projected wind changes from the models used in the Fourth Assessment Report,29 with increased westerly flow in winter and spring and decreased westerly flow in summer and autumn. In spring, the mean westerly flow increases by about 10% by 2040 and 20% by 2090. Winter westerlies increase even more, but there are projected decreases of 5–20% in the summer and autumn westerlies, in the average over all models and scenarios. There is clearly still substantial uncertainty about the projected future wind changes, as evidenced by the wide range across the climate models. Only for winter do all the models project increasing westerly flow across New Zealand.

Table 2.6: Projected changes in seasonal and annual westerly and southerly wind components (in m/s).

  Summer Autumn Winter Spring Annual
Westerly wind speed component

1970–1999 climate






Change by 2040:




[–1.6, +1.3]


[–2.1, +1.3]


[+0.2, +2.4]


[–0.6, +1.4]


[–0.5, +1.2]

Change by 2090:




[–2.5, +1.4]


[–2.3, +1.0]


[0.0, +3.6]


[–0.7, +2.0]


[–0.6, +1.5]

Southerly wind speed component

1970–1999 climate






Change by 2040:




[–0.8, +0.5]


[–0.6, +0.3]


[–0.7, +0.5]


[–0.4, +0.6]


[–0.5, +0.4]

Change by 2090:




[–1.2, +0.6]


[–0.6, +0.2]


[–1.3, +0.5]


[–0.6, +0.6]


[–0.9, 0.0]

Note: The ‘westerly’ component is derived from the Z1 Index (Auckland minus Christchurch pressure difference), and the ‘southerly’ component from the M1 Index (Hobart minus Chatham Islands pressure difference). For comparison, the current climatology (1970–1999) based on NCEP reanalyses is also given. A positive value means more westerly or more southerly, as appropriate.

Projected changes in the north–south wind component are less clear. Note that the climatological values are positive in all seasons except summer, meaning that the prevailing winds over New Zealand are from slightly south of west. There is a tendency for more northerly flow in future (ie, southerly component changes are negative), but the changes are not large enough to alter the prevailing wind direction from the west-southwest.

As noted in the example of Box 2.2, an increase in the mean westerly component does not in itself imply an increase in total wind speed. Strong30 winds are associated with intense convection (expected to increase in a warmer climate) and with intense low-pressure systems, which might also become more common (see extra-tropical cyclones below). Thus an increase in severe wind risk could occur.31 Since strong winds can cause damage to structures, forests and crops in New Zealand, there is much interest in trying to understand how climate change might impact on winds.

Higher temporal resolution is required to quantify future changes in wind speed. Figures 2.10 and 2.11 show preliminary analyses from the NIWA regional climate model, where data on daily time scales allow the full wind distribution to be examined. The two panels of Figure 2.10 together help clarify the relation between the westerly wind component and the total wind speed.

Figure 2.10: Change (in %) in wind in the winter season between a control run (1980–1999) and a climate simulation under the A2 scenario (2080–2099) for a 10-m wind westerly component (left) and 10-m wind speed (right).

This figure contains two maps. The first map depicts the percentage change in the westerly component of the wind. The change is calculated as the difference between that estimated in a control model simulation of the climate for 1980-1999 and a model simulation of the climate for 2080-2099 assuming the A2 emissions scenario. It shows the westerly wind component increasing in the south with 20-30 percent increase over Southland and Otago, increases in the west with 40-60 percent over Westland 20-40 percent for the region from Wellington to Taranaki including Wairarapa, and around 60 percent increase over Waikato. Little change is shown for Canterbury, Kaikoura, Hawkes Bay and most of Northland. The area north of Hokianga shows a 10 to 20 percent decrease.

The second map depicts changes in the 10 m windspeed for the same model simulations. This show much smaller changes, with increases of 4 to 6 percent over Southland, Otago and Canterbury. Most other areas show decreases, with Westland showing around 2 to 4 percent decrease, and north of a line between Wanganui and Hawkes Bay showing around a 2 percent decrease. Little change is mapped for Marlborough, Kaikoura, Wellington, Wairarapa, and Manawatu.

Figure 2.11: Change (in %) in the 99th percentile wind speed in the winter season between a control run (1980–1999) and a climate simulation under the A2 scenario (2080–2099).

Figure 2.11 shows a map of the 99th percentile 10 m wind speed change for the winter season. The change is calculated as the difference between that estimated in a control model simulation of the climate for 1980-1999 and a model simulation of the climate for 2080-2099 assuming the A2 emissions scenario. For the South Island, the map shows a decrease in strong winds over Westland and Buller of around 3 to 6 percent, and an increase of 3 to 6 percent over Fiordland, Southland Canterbury, Kaikoura, Marlborough and Nelson. Otago shows increases of up to 9 percent. The North Island shows a general increase, with 3 to 6 percent increases over Wellington, Wairarapa, Manawatu, Hawkes Bay, Gisborne, Waikato, Auckland and Northland. Taranaki and Bay of Plenty show decreases of around 3 percent.

Figure 2.10 shows (left panel) an increase in the mean westerly wind component over New Zealand in the order of 20% by late in the century. This result from the NIWA simulation, shown for the winter season only, is consistent with the global model run by the UK Met Office and the ensemble of other models used in the Fourth Assessment Report. The corresponding change in wind speed is shown in the right-hand panel. Over most of the North Island, in this single example, average winter wind speed is projected to either change minimally or actually decrease by the end of the century. The implication is that there are fewer days of high-speed easterly wind days in the simulation. Over the lower South Island, and south of New Zealand, the projected wind speed increases by around 5%, a much smaller percentage change than the change in mean westerly.

Since Figure 2.10 presents an average over a season, there remains the possibility of stronger winds in individual storms.32 More research is needed on intensity and storm tracks, but Figure 2.11 shows a preliminary analysis of the upper tail of the daily wind distribution. The 99th percentile represents the daily wind speed level that is exceeded only 1% of the time, which is currently about 10 m/s over the land (and 15 m/s or more over the ocean) in this model. Figure 2.11 suggests an increase in these strongest winds over much of the country by 2100. The changes are fairly small for the most part (averaging out at a 2.3% increase over all land points in the model), but reach about 10% in some eastern locations.

2.2.8 Ex-tropical cyclones, and mid-latitude storms

The IPCC Fourth Assessment Report concludes that it is likely33 that future tropical cyclones will become more intense, with larger peak wind speeds and more heavy precipitation. There is less confidence in projected changes in numbers of tropical cyclones. Tropical cyclones have changed their characteristics by the time they reach New Zealand, by which stage they are generally called ‘ex-tropical’ cyclones. Such systems tend to affect mainly the northern and eastern regions of the North Island, although occasionally they track further south. During El Niño periods, tropical cyclones tend to track further east in the South Pacific, and are less likely to affect New Zealand directly. Many climate models show an El Niño-like change in the mean state of the tropical Pacific over the next 50 years, but just how this might affect the likelihood of future ex-tropical cyclones reaching central New Zealand is not yet clear (see also section 3.2.1).

Of particular interest to New Zealand is the future behaviour of mid-latitude storms and low-pressure systems, also known as ‘extra-tropical cyclones’. The IPCC34 says: ‘Extra-tropical storm tracks are projected to move poleward, with consequent changes in wind, precipitation and temperature patterns’.

Based on our understanding of weather dynamics, an increase in southern hemisphere storminess is certainly possible. The mid-latitude cyclones draw their energy from both the horizontal north–south temperature gradient (which is expected to increase for at least the next century, as the tropics will warm faster than will southern polar regions) and the release of latent heat (which will increase with the temperature-related moisture increases). Latent heat drives the tropical cyclones, but its importance varies enormously from storm to storm with mid-latitude circulation systems.35 Also, a change in storm tracks could overwhelm any local impact of change in storm intensity. Nevertheless, increases in the peak wind speeds are possible, and increases in extreme precipitation likely, as discussed earlier in this chapter.

2.2.9 Ocean currents and wave patterns

The coupled atmosphere–ocean global climate models do not include enough detail to show the narrow ocean currents that flow around New Zealand, and very little analysis has been done of future wave patterns. This means that no quantitative statements about these climatic features can be made at this stage.

Changes in the winds do hold clues to what might be expected in the ocean.36 The warm currents flowing down the eastern coast of the North Island are part of the subtropical gyre and are driven primarily by the ‘twisting’ action of the winds over the subtropical South Pacific. These wind patterns over the subtropics show little change in the model runs, but are projected to increase in higher latitudes, and this increase is expected to impact on the ocean current systems. Recent research37 has linked warm conditions around New Zealand in the late 1990s to an enhanced Southern Annular Mode with increased westerlies south of New Zealand. The model projections of further increases in the westerly winds over the southern oceans could, therefore, be expected to cause a ‘spin-up’ of the subtropical gyre, with slightly stronger flows and warmer conditions in the gyre centre (ie, around New Zealand). The projected increased westerlies in high latitudes would also be expected to accelerate the cold Antarctic Circumpolar Current.

In addition, increased westerly winds could have impacts on stratification and, therefore, the input of nutrients from the deep ocean into the euphotic zone. The first mechanism for this is simply the increase in mechanical stirring from stronger winds. The second is increased wind-driven upwelling along the New Zealand coast. For a straight westerly, the coasts affected are the northward-facing coasts (ie, the West Coast of the South Island, and the northeast coast of the North Island), but changes in the north–south wind component could substantially modify which regions are affected.

Increased westerlies would also influence the ocean wave climate that impacts on New Zealand. In particular, coastal regions exposed to the prevailing winds would be subject to an increase in the frequency of heavy swells that would add to effects of higher sea levels. (Refer to the CoastalHazards and Climate Change manual (Ministry for the Environment 2008) for further discussion.)

16  Because the far north of New Zealand already experiences very few frosts, the frost frequency there cannot decrease substantially.

17 Table SPM.2, Summary for Policy Makers, IPCC 2007a.

18 Whetton et al 1996.

19  Return periods can be translated into annual exceedance probabilities (AEP) that specify the likelihood of a rainfall amount being exceeded in any given year. For example, a 100-year return period event has an AEP of 0.01, a 20-year return period event has an AEP of 0.05, and so on.

20 Semenov and Bengtsson 2002.

21  Pall et al 2007.

22  For situations that could involve serious loss of life, there should be consultation to determine the latest scientific advice on projected changes in extreme rainfall amounts.

23  Meehl et al 2007.

24  Gray et al 2005.

25  Arora and Boer 2001.

26  This refers to what is often called ‘absolute’ sea-level rise. What is actually measured by tide gauges is the ‘relative’ sea-level rise, which includes any local vertical land movement (uplift or subsidence of land or coastal seabed) that may occur, for example, as a consequence of earthquake activity. See the Coastal Hazards and Climate Change manual (Ministry for the Environment 2008) for further discussion.

27  IPCC 2007d

28 Mullan et al. (2001b) suggested a 10% increase in westerly component wind speed over the next 50 years.

29  With the Third Assessment models reported in the previous Guidance Manual edition (Ministry for the Environment 2004), six models for the 2030s and six for the 2080s, there was a tendency for increased mean westerly flow over New Zealand in all seasons individually.

30 We use the term ‘strong’ here in a non-technical sense to cover wind speeds above about 10 m/s. A ‘strong’ wind is formally defined on the Beaufort Wind Scale as Level 6 (in the range 22–27 knots, or
11–14 m/s), one level above ‘fresh’ and one level below ‘near gale’.

31 Knippertz et al. (2000) identified an increasing number of strong wind events over the North Atlantic in their climate model simulation, which they relate to the increasing number of intense cyclones.

32  This particular regional model simulation shows a decrease in the number of low-pressure centres crossing the North Island by the end of the century; hence fewer storms, even if individually stronger, could still result in lower average wind speeds over each season.

33  In IPCC terminology, ‘likely’ has the technical meaning of greater than 66% probability of occurrence.

34  IPCC 2007a.

35  Revell 2002; 2003.

36 Mullan et al. 2001b. See also the MfE 2008 Coastal Guidance Manual, section 2.3 and Factsheet 10.

37  Roemmich et al 2007.