View all publications

The state of the atmosphere

Two trends are undisputed among atmospheric scientists: ozone levels have declined and greenhouse gas levels have increased. In both cases, the trends have been marked departures from the natural state of the atmosphere that has prevailed for hundreds of thousands, and probably millions, of years. Where uncertainty and speculation are more prevalent is in predicting the impacts of these changes.

A general consensus exists that solar radiation in temperate zones has intensified as the ozone layer has thinned, and that global temperatures have increased as greenhouse gases have accumulated.

The magnitude of these effects and their significance for humans and the environment is still uncertain. If there are effects, however, they will increasingly manifest themselves in sunburn-related problems in plants, animals and humans, and in climate changes that will alter growing seasons, snowlines, sea levels, rainfall patterns and species distributions.

The state of the ozone layer

It is now well established that, in the past two decades, the ozone layer has been significantly depleted (World Meteorological Organisation, 1995). Ground-based measurements show an overall drop of about 5 percent through the 1980s, very similar to that observed from space by an orbiting satellite instrument, the Total Ozone Mapping Spectrometer (TOMS) (Reinsel et al., 1994). In mid-latitudes, an annually averaged decrease in total ozone of 4 to 5 percent per decade has taken place since 1979 (UK Stratospheric Ozone Review Group, 1996). A drop of 10 percent in total ozone concentrations increases UV-B radiation on the Earth's surface by some 20 percent (World Meteorological Organisation, 1995).

New Zealand's ozone monitoring shows a decline in ozone concentrations during the 1980s, with a more abrupt decline from the mid-1980s (see Figure 5.17). In only two of the 120 months between 1985 and 1995 did average ozone levels exceed the average 1970s level. In Figure 5.17, the seasonal fluctuations have been removed to show a natural phenomenon known as the global Quasi-Biennial Oscillation (QBO), which is a natural two-yearly fluctuation in the amount of ozone caused by the global circulation pattern.

Figure 5.17: Deviations in monthly mean ozone, Invercargil/Lauder, 1970-94

Deviations from the monthly mean level of ozone (based on 1970 to 1979 levels) monitored at Invercargill/Lauder fluctuate biannually. There has been an abrupt reduction in monthly deviations since 1985.

Sources: Nichol and Coulmann (1990); NIWA, unpublished data

The good news for the ozone layer is that world governments and the chemicals industry have acted just in the nick of time to stave off a much larger catastrophe (Prather et al., 1996; Andersen and Miller, 1996). Provided countries strictly comply with the provisions of the Montreal Protocol, the international agreement to phase out the use of ozone-depleting substances, ozone depletion is expected to peak between 1997 and 1999 (Montzka et al., 1996; Prather et al., 1996; Kerr, 1996a; Pearce, 1996a; World Meteorological Organisation, 1995). Within that period chlorine concentrations will top out at around 4 ppb (see Figure 5.18).

Figure 5.18: Atmospheric chlorine levels with and without the Montreal Protocol and its 1992 amendments.

Predictions for changes in atmospheric chlorine levels as a result of the Montreal Protocol and its 1992 amendments. Without the protocol, levels are predicted to reach 9 parts per billion in 2010. The protocol is predicted to reduce levels to a peak of 4 parts per billion in between 1995 and 2000, and then reduce the concentration to below 2 parts per billion in 2050 - below this level the Antarctic Ozone Hole should stop forming.

Adapted from: United Kingdom Stratospheric Ozone Review Group (1993); Pruther et al. (1996)

The bad news is that it will be some time before the damage is undone, because CFCs have such a long lifetime. By the year 2050 chlorine concentrations could be down to about 2 ppb, about half their present level and the ozone layer will have largely recovered. But even with the Montreal Protocol's constraints on the use of ozone-depleting substances, it will be another century before the amount of chlorine in the atmosphere returns to anywhere near its natural level.

The Antarctic ozone hole

When the first alarm bells were sounded about ozone depletion in the early 1970s, no-one predicted that CFCs would actually tear a hole in Earth's 'bullet-proof vest'. When the Antarctic ozone 'hole' was finally announced in 1985, it came as a surprise (Farman et al., 1985; Gribbin, 1988). But evidence of the hole's development had been accumulating for some time. Data gathered at the British Antarctic Survey station in Halley Bay showed that total ozone had been declining each spring (September October) since the observations began in 1957.

These early observations were made with a ground-based spectrophotometer. With the advent of satellite-based TOMS measurements in 1979 it became possible to cover a much wider area. The satellite data confirmed the ground-based results and showed that the springtime ozone depletion covers most of the Antarctic continent (Newman et al., 1991).

At first, however, the congruence between ground and satellite data was not noticed. The scientists operating the TOMS equipment had dismissed the low Antarctic ozone readings as errors because they were so far below ozone readings from other parts of the world (Gribbin, 1988).

Compared to declines of 1-2 percent elsewhere in the world, the springtime ozone over a large area of Antarctica was dropping by 10-20 percent (see Figure 5.20). The decline accelerated around 1980 and by 1983 ozone levels in some parts of the Antarctic were falling from their normal springtime low of 300 Dobson Units (DU) to levels of 225 DU and lower. Below 225 DU, the thinning ozone layer became known as the 'ozone hole'. At sea level, this would represent a thinning of the layer from 3 millimetres to just over 2 mm. The hole continued to widen and deepen through the 1980s, with ozone levels in some parts declining to less than 150 DU.

Using data from 1979 to 1990, the best computer models successfully predicted the global ozone characteristics for 1991, including a small central area where total ozone fell to less than half its normal level (i.e. less than 150 DU). The models allowed for seasonal effects, the global circulation effect known as the global Quasi-Biennial Oscillation (QBO), and the solar cycle. However, the record low global ozone levels of 1992 and 1993 were 1-2 percent lower than predicted (Gleason et al., 1993; Herman and Larko, 1994).

Figure 5.19: The Antarctic Ozone Hole 1980-1996.

The size of the Antarctic Ozone Hole has increased from 1980 to 1996.

In 1993, the ozone hole extended over almost 24 million km2 (an area larger than North America). At its October peak, in the atmospheric zone between 14 and 19 km, ozone levels fell to just 1 percent of their usual level (World Meteorological Organisation, 1995). The most likely explanation for this abrupt drop was the eruption of Mount Pinatubo in June 1991, although the mechanism by which ozone might be affected by the volcanic eruption is still being debated by atmospheric chemists and climate modellers. In 1994, things got back to where they had been with ozone levels recovering as the volcanic particles in the stratosphere declined (World Meteorological Organisation, 1995). Despite the return to recent ozone levels, 1994 still produced a sharp reminder of the critical state of the ozone hole: Wednesday 28 September made the record books as the day with the lowest ozone value ever recorded over Antarctica: 88 DU.

Although the 1995 ozone hole did not quite match previous ones in extent or ozone loss, it did set a different record: it lasted longer than any other. The 1995 hole covered an area greater than 15 million km2 for a total of 71 days compared with the previous high of 63 days in 1993 (GECR, 1995).

The latest hole, which formed in September and October of 1996, covered more than 22 million km2. The ozone between altitudes 16 and 22 km was completely annihilated. The area with ozone depletion greater than 50 percent had grown from the small hole of 1991 to an area covering almost two thirds of the Antarctic (World Meteorological Organisation, 1996a, 1996b).

The ozone 'hole' forms over Antarctica and nowhere else because of the peculiar combination of extremely low temperatures (below -80°C) and strong air currents which circle Antarctica during the winter (the Antarctic Vortex). Of these, temperature is the more important (Gribbin, 1993). The Vortex effectively seals off Antarctica from other global air currents, trapping ozone and halogens above the continent.

Meanwhile the intense cold produces ice particles in the polar stratospheric clouds. Chlorinated chemicals settle on these particles and react chemically with each other. These reactions turn the chlorine part of the CFC molecules into unstable compounds which begin attacking ozone when they are activated by the returning spring sunlight (see Box 5.4).

After the Antarctic Ozone Hole was discovered, some scientists took the view that it might be a natural event caused by volcanic chlorine emissions from Mount Erebus rather than manufactured chlorinated chemicals. Eventually, however, Mount Erebus was exonerated (Zredna-Gostynska et al., 1993). Most of the chlorine Mount Erebus throws up takes the form of hydrogen chloride (HCl), which (like other chlorine from natural sources) readily dissolves in the water vapour of the lower atmosphere well before it can reach the stratosphere.

For Mount Erebus to affect the ozone layer, the volcano would have to inject a large proportion of its hydrogen chloride directly into the stratosphere, above a height of about 10 km. Mount Erebus has been active since it was first observed by James Ross in 1840, but appears never to have erupted with the force necessary to send chlorine directly into the stratosphere. The mountain itself is almost 4,000 m high (3,794 m), but the volcanic plume seldom rises above 5,000 m. The amount of gas Mount Erebus emits also bears no relation to the size of the ozone hole. In the summer of 1983, chlorine emissions from Mount Erebus were about 170 tonnes a day. In the following seven summers, when ozone depletion was even more severe, the chlorine emissions ranged from one-tenth to one-quarter of the 1983 figure (Zreda-Gostynska et al., 1993).

Figure 5.20: Average October total ozone, Halley Bay, Antartica, 1957-94.

The average total ozone over Halley Bay, Antarctica, has decreased each October from 1957 to 1994. Levels were around 330 Dobson units in 1957 and in 1994 were around 120 Dobson units.

Source: WMO (1995).

Another popular misconception is that New Zealand is directly affected by the Antarctic Ozone Hole. At no time since its discovery has the Antarctic Ozone Hole itself ever extended as far north as New Zealand. In springtime, when the ozone hole covers the Antarctic, New Zealand usually lies under an area rich with ozone (see Figure 5.19). However, for about a day in September 1993, an unusual distortion in the polar vortex pushed an area of low ozone over southern New Zealand. This resulted in ozone values of 270 Dobson Units (DU) compared with the month's long-term average of about 380 DU.

Until then the record September low had been 300 DU.

The Antarctic Ozone Hole is expected to begin getting smaller sometime around 2010 and should cease to form by 2050 if current controls on ozone-depleting substances are maintained indefinitely. The period of most intense ultraviolet-B radiation is likely to be over the next dozen years or so.

The state of solar ultraviolet radiation

Ozone depletion has led to a significant increase in ultraviolet radiation in New Zealand since the 1970s. Between the mid 1970s and the 1990s, ozone declined by approximately 5-7 percent. Over the same period, it is estimated that skin-damaging (UV-B) solar radiation probably increased by about 6-9 percent (McKenzie, 1996). Ultraviolet radiation has been implicated in the development of malignant melanoma, as well as sunburn, skin ageing, and cataract formation, and several studies are underway to gather hard evidence of the links between ultraviolet radiation in New Zealand and its relationship to skin disorders.

Zheng and Basher (1993) analysed data taken from a Robertson-Berger (R-B) meter at Invercargill from 1981 to 1990. They concluded that in clear sky conditions there is a significant trend to higher UV-B and that this corresponds with the downward trend in ozone. They estimated that UV-B radiation had increased during that decade by about 6 percent in clear-sky summertime at Invercargill. The increase was particularly steep between 1982 and 1985.

Since 1990, NIWA scientists have been monitoring spectrally-resolved ultraviolet radiation with narrow band spectroradiometers at the Lauder atmospheric research facility in central Otago. Although these instruments cannot log data continuously, they are much more accurate than R-B meters. Spectrally-resolved measurements clearly show that, when ozone levels go down, ultraviolet radiation levels go up. In New Zealand, however, this trend has not shown up with the spectroradiometer monitoring because the ozone levels have been relatively stable since the instruments were introduced (McKenzie, 1996).

To date the most conclusive evidence that ozone depletion would logically result in more harmful ultraviolet radiation reaching the Earth's surface has come from data collected by the satellite-borne Total Ozone Mapping Spectrometer (TOMS). Recent analysis of the TOMS data shows that ultraviolet radiation capable of damaging DNA must have been increasing at a rate of about 8 percent per decade in the spring, early summer, and autumn around latitude 40°, a zone which includes the lower North Island of New Zealand. Further south, the low temperatures that foster ozone loss appear to have boosted the increase in solar radiation to 10 percent to 12 percent per decade. Only the tropics and subtropics have been spared any ultraviolet increases (Kerr, 1995c).

White et al. (1992) measured ultraviolet radiation at Auckland, Wellington, and Christchurch. Although their relatively short data series prevented them from making statements about trends, they determined the actual levels of erythemal (sunburn) and carcinogenic (melanoma) ultraviolet radiation for the three sites, including its variation with latitude, season, and cloud cover.

Latitudinal differences in ultraviolet have been measured by cross-calibrated sensors in the north and south of New Zealand. These measurements show that averaged over a year, the north receives 25 percent more ultraviolet than the south. The north of the country receives approximately 10 percent more ultraviolet than the south in summer, and twice the amount in winter. The dominant factor in the amount of ultraviolet reaching New Zealand is the overhead position of the Sun, which causes large seasonal variations, especially in the south. At both sites, clouds reduced the amount of ultraviolet received by 25-30 percent compared with clear sky conditions. Seasonal changes in ozone also cause significant changes in ultraviolet, and day-to-day variations cause ultraviolet fluctuations of up to 10 percent (McKenzie et al ., 1996).

Although ozone levels over New Zealand have been fairly stable since the early 1990s, further depletion can be expected as the ozone-depleting substances released over the past 50 years continue to float into the stratosphere. As a result, levels of UV-B radiation are not likely to decline until well into next century when the ozone layer starts to recover (see Figure 5.18).

Seckmeyer and McKenzie (1992) made the particularly interesting comparison between New Zealand (Lauder in central Otago) and a site of comparable latitude in Germany. They showed that in the New Zealand summer of 1990-1991 the amount of ultraviolet radiation harmful to plants and animals was 30 to 60 percent more than the amount occurring in summer in northern Germany. Two further surveys have confirmed that result (McKenzie et al., 1993; Seckmeyer et al., 1995). The large difference has been attributed mainly to the lower amounts of stratospheric ozone above New Zealand, as well as to the greater amount of tropospheric ozone (caused by pollution and aerosols) over northern Europe.

Box 5.6: The potential impacts of excessive ultraviolet-b (UV-B) radiation

Ultraviolet-B (UV-B) radiation can damage property, causing paint to fade, window glazing to yellow, and car roofs to become chalky. At high levels it can also damage many of the plastics used in the building industry and elsewhere. Costly though this sort of damage is, most of the concern about ozone depletion has centred on the potential impacts on health and the environment (UNEP, 1994). UV-B radiation has increased measurably in temperate latitudes as a result of ozone depletion (Kerr and McElroy, 1993; Appenzeller, 1993b; Kerr, 1995c; Madronich et al., 1995; Slaper et al., 1996). In New Zealand, the increase of about 6-9 percent since the 1970s is compounded by the fact that our summer coincides with the period of lowest ozone and with the phase of the Earth's orbit which takes us slightly closer to the Sun (McKenzie, 1996). As a result, peak summertime levels of sunburning radiation are 40 percent more intense here than at comparable northern latitudes. Because the majority of people living in New Zealand are fair-skinned, and the majority of our farm animals spend the entire year outdoors, the potential health risks from UV-B are higher here than in many other countries.

UV-B rays are known to damage the proteins that make up living tissue (such as skin or plant tissue). They can also cause mutations in the underlying genes that manufacture these proteins. Such damage can cause the exposed tissue to die (e.g. the peeling skin which follows sunburn), or form scar tissue (e.g. cataracts and the yellow lumps on the eyeball known as pterigia), or grow aberrantly (e.g. skin cancers). Excessive exposure to UV-B can also suppress the efficiency of the body's immune system, reducing resistance to disease, such as cancers, and latent viral infections (e.g. Herpes simplex and H. zoster). Three forms of skin cancer are associated with UV-B damage: the non-melanocytic skin cancers (which include both squamous and basal cell carcinomas) and melanoma. Non-melanocytic skin cancer does not spread and usually responds to treatment. Melanoma, however, (the type most people refer to as 'skin cancer' or simply 'melanoma') can spread and cause death if not treated. Early detection of melanoma results in successful treatment in 95-100 percent of cases, but a delay of 5 years in detection can lower the chance of survival to 20 percent.

Over the past few decades, skin cancer has become a serious problem. The number of melanoma cases in virtually all fair-skinned Caucasian populations around the world has risen over the past 40 years. Incidence rates of melanoma in New Zealand's non-Māori population have continuously been among the highest registered worldwide, and are only exceeded by rates in some Australian States, and among white Hawaiian men. New Zealand men and women have the highest melanoma death rates in the world (Bulliard and Cox, 1996). The main cause of rising skin cancer rates in New Zealand, Australia, and North America appears to be human behaviour, that is the fad for fair-skinned people to tan their skins by 'sunbathing'. About 9 out of 10 melanomas occurring among populations of European descent in Australia and New Zealand are attributable to sun exposure. Potentially, this makes melanoma one of the most preventable cancers (Bulliard and Cox, 1996). Worldwide, around 10,000 people die from melanoma each year. Roughly 2 percent of these are New Zealanders-180 in 1991, rising to about 260 today, and projected to reach around 380 by 2005 (Cox, 1995; Ansley, 1996). During the 1980s, the number of recorded new melanoma cases increased by about 2.5 percent per year and had reached 1,000 by 1991 (Public Health Commission, 1994). The rate appeared to have levelled off from the late 1980s (Bulliard and Cox, 1996) but the latest figures indicate that this was probably because of under-reporting. Melanoma reporting became compulsory for medical practitioners from 1 July 1994 under the Cancer Registry Act 1993. In the first full year of compulsory reporting (1 July 1994-30 June 1995) the New Zealand Health Information Service's provisional figure for the number of melanoma cases was 1,618. The provisional figure for the 1993 calendar year (January-December) was 1,037.

Although ozone depletion increases the biologically harmful solar ultraviolet radiation reaching the surface of the Earth, the actual contribution of ozone depletion to the skin cancer trend is not known. Estimated increases in skin cancer rates due to ozone depletion generally refer to non-melanocytic skin cancer. This is because melanoma risk is related to intermittent ultraviolet exposure whereas squamous cell carcinoma, and to a lesser extent basal cell carcinoma, are associated with chronic, and thus cumulative, ultraviolet exposure. Any direct links between ozone depletion, increased ultraviolet radiation, and rising rates of skin cancer are also blurred by the time lag between skin damage and the onset of cancer. However, researchers have estimated that under a 'no restrictions' scenario, unlimited production of ozone-depleting substances would have led to a quadrupling of skin cancer incidence by the year 2100-assuming no change in human behaviour to limit exposure to the sun (Slaper et al., 1996). Limited restrictions under the original Montreal Protocol would have seen a doubling of the incidence by the same date. However, a scenario based on the much tighter 1992 Copenhagen Amendments to the Montreal Protocol, shows a peak increase in the incidence of skin cancer of only 10 percent occurring 60 years later (see Figure 5.18).

UV-B related cancers are not confined to humans. Where pigs and goats, cattle and horses have inadequate access to shade they also fall victim to skin and eye cancers, and their exposure is less subject to seasonal fashions. It is not known whether their melanoma rates have increased in recent years. Plants are also affected by excessive UV-B light, and, because they rely on direct sunlight for energy, many species have developed protective mechanisms against high levels of UV-B. Despite this, experiments have shown that cotton, peas, beans, melons, and cabbage grow more slowly under intense UV-B, and pollen fails to germinate in some plants. Plant hormones and chlorophyll, the chemical mainly responsible for photosynthesis, can also be damaged. New Zealand scientists are currently studying ways of making crop plants more ultraviolet-resistant through selective breeding programmes and gene modification (Markham and Ryan, 1996).

Aquatic ecosystems are also vulnerable. UV-B can penetrate clear water to a depth of many metres, posing a threat to single-celled algae which are known from laboratory studies to be sensitive to UV-B. These organisms are at the very base of the aquatic food chain, and are also major consumers of carbon dioxide and emitters of anti-greenhouse sulphate aerosols. A serious reduction in algal biomass could therefore reduce fish populations and enhance the build-up of greenhouse gases in the atmosphere.

A study of Antarctic algae which were shielded from UV-B radiation by ice found no change in species composition over two decades of ozone depletion (McMinn et al., 1996). However, a survey of algal blooms in open water found that biomass (total weight) fell by 6-10 percent in the waters where UV-B was most intense, beneath the springtime 'ozone hole' (Smith et al., 1992). This represents about 7 million tonnes of lost photosynthesis per year, or a 2 percent depletion of the Southern Ocean's phytoplankton. The researchers noted that this reduction, though significant, was small compared to the seasonal advance and retreat of the ice pack which causes biomass reductions as high as 50 percent.

The state of the climate

Over the years, instruments not very different from those being used today have measured the major climate variables of pressure, temperature, rainfall, and wind. Measurements at many New Zealand sites go back to the 1860s. Together, those indicators of the climate and its rate of change, give a guide to the state of our present atmospheric environment, and signal some trends.

Figure 5.21: New Zealand temperatures

New Zealand average temperatures increased from 1853 until around 1940, were reasonably constant from 1940 to 1970 and have been increasing since then. The warmest decade is the 1980s.

Source: NIWA, unpublished data


Temperature is one of the most useful climate indicators available. Although it is difficult to identify rapid global temperature changes for the Earth in prehistoric times, global temperatures can be deduced for most of the planet's history. Good instrumental measurements have been taken from an increasing number of regions since the mid-nineteenth century.

Climatologists attempt to track climate change by using global average temperature as an indicator. Despite the difficulties in establishing a true average for the whole globe through all seasons, the efforts are sufficiently accurate to show global temperature trends. Enormous amounts of meteorological data have been collected and archived over the past century. These high-quality records are described in the IPCC Report (Houghton et al., 1992).

The climate monitoring records begin about 1860, and show a general temperature increase through to about 1940, little change to 1970, then an increase through to the 1990s. Globally, the four warmest years from the whole period are all in the 1990s: 1995 was the Earth's warmest year, while 1990, 1991, and 1994 were the second, third, and fourth warmest in that order. The 1995 global average surface temperature was 0.40°C above the 19611990 baseline, just eclipsing the previous record of 0.36°C of 1990. The 199195 period also tied for the warmest ever half-decade, despite the effects of Mount Pinatubo which erupted in the Philippines in June 1991 and cooled global temperatures in 1992 and 1993.

The New Zealand land surface temperature record has been constructed from data collected at 22 reference climatological stations (Salinger et al., 1992a). Although each data series represents a single site, it is possible to derive from them a single figure representative of an annual temperature for the whole of New Zealand (see Figure 5.22). The pattern of temperature change in New Zealand has been similar to the global pattern over the past 130 years, but recent evidence suggests that New Zealand's temperature is increasing at a rate higher than the global average. Global average surface temperatures for the land and sea have increased by 0.45°C over the past 100 years, but New Zealand's temperature since 1900 has risen by 0.7°C (Salinger, 1995).

The warmest decade in New Zealand was the 1980s, and just as 1995 was the Earth's warmest year on record, it was also New Zealand's warmest year since 1990. Global temperatures have remained high in the 1990s, but recurring El Niño episodes since 1990 have had a cooling effect on New Zealand's climate. New Zealand's very much cooler year of 1992 was directly attributable to the combination of Mount Pinatubo's eruption and the persistent El Niño event. The national average temperature began increasing again from 1992 as the cooling effect of Mount Pinatubo faded away (see Figure 5.22).

Figure 5.22: New Zealand average temperatures 1980-94.

New Zealand's average temperature between 1980 and 1994 has been lower during El Nino episodes (1983, 1987, 1992). 1992 was particularly cold with the El Nino being combined with the cooling effects of the Mt Pinatubo eruption in June 1991.

Source: NIWA, unpublished data


Short-term fluctuations in local or regional rainfall from year to year, or decade to decade, reflect airflow changes and their interaction with the mountains, as well as changes in airflow induced by El Niño/La Niña events. Salinger et al. (1992) conclude that in the last decade, most North Island sites were drier than normal by up to 20 percent. Most western South Island stations recorded wetter than normal conditions in the 1980s, although there were some exceptions, such as Blenheim, Lincoln, and Dunedin, which were all drier than normal. These trends were caused by an unusually high number of El Niño episodes in that period (see Figures 5.23 and 5.24).

Figure 5.23: The Southern Oscillation Index (SOI), 1970-95.

The Southern Oscillation Index has fluctuated over 1970 to 1995. It has tended to be lower since 1981, and an intense El Nino in 1983 dropped the Southern Oscillation Index to around minus 3. During the persistent El Nino from 1992 to 1994 it was around minus 2.

Source: NIWA, unpublished data

Box 5.7: The potential impacts of global warming and climate change

Nobody knows for sure what the effects of climate warming will be in any particular region, but there is broad agreement on the general changes that might occur (Tegart et al., 1990; Tegart and Sheldon, 1993; Watson et al., 1996). Models predict that the world's desert regions will become more arid, a third to a half of the mountain glaciers will melt, snowlines will rise by 150550 metres, sea levels will rise by 2080 centimetres, and the distributions of tropical and sub-tropical organisms will spread north and south by 150-550 kilometres (Watson et al., 1996). The rising sea may inundate low-lying coastal areas, including a number of Pacific Islands. It may also erode coastlines and beaches and make groundwater supplies salty. Rainfall patterns will change, in some cases bringing more extreme conditions of flood or drought. One recent study found that, although the slight global warming of the past century has not changed the overall amount of rainfall, there has been an increase in rainfall variability with more extremes (Tsonis, 1996). Tropical cyclones may extend their range north and south. River and groundwater flows may change, reducing water supplies in some areas and increasing floods in others.

Natural ecosystems, and wild species of plants and animals, will undergo changes in their distributions and ranges. Warm-adapted species will increase and cold-adapted species will decline. Some alpine, sub-polar and cool temperate species may become extinct. Warmer seas will also favour algal growth, including increased blooms of toxic plankton. The abundance and distribution of some fish, marine mammal and seabird species may decline, though global marine fisheries production is expected to remain the same. Temperate forests may experience climatic stress in some areas, and although changed climatic conditions may favour the expansion of tropical forests, this is unlikely to offset the decline caused by human activities (Watson et al., 1996).

Agricultural production in temperate regions will increase because of significantly longer growing seasons and fewer frosts, but will also be subject to greater weed and pest invasions from warmer regions. In the tropics, agricultural production will decline, causing increased risk of hunger and famine in areas of high population and low income.

Other health impacts are expected to be wide-ranging and mostly adverse (Watson et al., 1996; WHO, 1996). While cold-related deaths in temperate cities will decrease, heat wave deaths in warmer urban areas will increase, especially for the elderly and the poorly-housed. More significantly, infectious tropical and sub-tropical diseases, especially mosquito-borne pathogens (e.g. malaria, filariasis, dengue and yellow fever), will become more widespread (Stone, 1995; Weinstein, 1996). The proportion of the world's population at risk from them may expand from 45 percent to 60 percent. Rats and insect pests will also expand their ranges (Watson et al., 1996; Saul, 1996). Temperate countries will become more vulnerable to a wider range of pests, weeds and diseases.

The impacts on New Zealand are uncertain, but two scenarios are commonly suggested (Royal Society, 1988, 1990 and 1992; Ministry for the Environment, 1990; Wratt et al., 1991). The most likely scenario is based on New Zealand's previous period of maximum warmth 810,000 years ago and assumes a 1.5°C increase in temperature by 2050. The frost-free season would be some 40 days longer, the snowline would be 100-300 metres higher and the sea would be 2040 cm higher. Westerly winds would decline by 10 percent, bringing fewer rainy days to western regions. However, rain would be heavier than at present so that the west and north would be about 10-15 percent wetter and the east and south would be 510 percent drier (Salinger and Hicks, 1990).

Westerlies would also decline in the second scenario, which assumes a temperature increase of 3°C. This would be accompanied by more frequent rain in the north associated with moist northerlies from the cyclone belt, and less rain in the south, particularly in Southland and Otago. The frost-free season would be at least 60 days longer than at present, the snowline 300-400 metres higher and the sea level about 30-60 cm above present levels (Salinger and Hicks, 1990). Under both scenarios, rainfall extremes are expected to be more frequent and widespread, causing flooding and erosion (Campbell and Ericksen, 1990; Griffiths, 1990). Droughts would increase in the east, with a 40 percent reduction in surface water in parts of Otago, Canterbury and Hawke's Bay. Sea level rise and increased rainfall would cause drainage problems in low-lying areas. Stormwater and sewerage systems could become overloaded more frequently, spilling contaminated water into surface waters. Smaller South Pacific nations may be more vulnerable to many health effects of climate change, including the direct effects of sea level rise and storms (Woodward, 1995).

Many of the most important medical conditions in New Zealand vary geographically and may be climate sensitive. Examples include asthma, heart disease, cot death, melanoma, and communicable diseases such as Hepatitis B. Little is known about how long the frequency and severity of these conditions may change under global warming. (There may be gains as well as losses.) There is potential for introduced disease, particularly arboviruses that cause Ross River fever and dengue, to be introduced and become established. Indirect health effects resulting from changes in food production, migration, and the extension of infectious diseases from tropical zones may also be important (Woodward, 1995). Researchers at the Wellington School of Medicine already say there is evidence that outbreaks of dengue fever (a viral disease spread by mosquitoes) in the South Pacific are closely linked to El Niño events (Hales, et al., 1996).

It is generally assumed that New Zealand's pastoral agriculture would be able to adapt to the new conditions, but crops, horticulture and viticulture may undergo some disruption if moisture levels and other micro-climatic conditions change significantly, and if invasive sub-tropical pests and weeds (such as the grass, paspalum) spread here. A complicating factor is the role of the El Niño-Southern Oscillation climate pattern and whether this is likely to aggravate the impacts of climate change (see Box 5.3).

A project called CLIMPACTS is making integrated assessments of the climatic effects on New Zealand agriculture, horticulture, grasslands, and soil nationally and regionally as well as at specific sites. CLIMPACTS is a collaborative programme involving two universities and five Crown Research Institutes (CRIs). It will increase the knowledge of the environmental effects of climate variability and change in New Zealand by pooling information and data from various sector-oriented science programmes into a mathematical model. This will be used as a basis for improved decision making in New Zealand (Royal Society of New Zealand, 1996).

Figure 5.24: Variations in average yearly rainfall, 1920-90 (expressed in percentage terms, with -0.2 corresponding to 20% less rainfall, and 0.5 corresponding to 50% more rainfall).

Variations in average yearly rainfall from 1920 to 1990 in Kaitaia, Gisborne, Auckland, Napier (Whatkatu), New Plymouth, Wellington, Hokitika, Blenheim, Milford Sound and Lincoln. Over the last decade, stations in the North Island, Blenheim and Lincoln have tended to have less rainfall than usual. Hokitika and Milford Sound stations have had more rainfall than usual.

Source: Salinger et al. (1992b)