. 7
( 13)


tropical forests, the jungle teems with life of enormous variety. In drier
regions sparse vegetation exists, of a kind which can survive for long
periods with the minimum of water; animals there are also well adapted
to dry conditions.
Water is also a key substance for humankind; we need to drink it, we
need it for the production of food, for health and hygiene, for industry and
transport. Humans have learnt that the ways of providing for livelihood
can be adapted to a wide variety of circumstances regarding water supply
except, perhaps, for the completely dry desert. Water availability for
domestic, industrial and agricultural use averaged per capita in different
countries varies from less than 100 m3 (22 000 imperial gallons) per year
to over 100 000 m3 (22 million imperial gallons)21 “ although quoting
average numbers of that kind hides the enormous disparity between those
in very poor areas who may walk many hours each day to fetch a few
gallons and many in the developed world who have access to virtually
unlimited supplies at the turn of a tap.
The demands of increased populations and the desire for higher stan-
dards of living have brought with them much greater requirements for
fresh water. During the last ¬fty years water use worldwide has grown
over threefold (Figure 7.6); it now amounts to about ten per cent of the
estimated global total of the river and groundwater ¬‚ow from land to sea
The impact of climate change on fresh water resources 157

(Figure 7.5). Two-thirds of human water use is currently for agriculture,
much of it for irrigation; about a quarter is used by industry; only ten
per cent or so is used domestically. Increasingly, water stored over hun-
dreds or thousands of years in underground aquifers is being tapped for
current use. With this rapid growth of demand comes greatly increased
vulnerability regarding water supplies.
The extent to which a country is water stressed is related to the
proportion of the available freshwater supply that is withdrawn for use.
Withdrawal exceeding twenty per cent of renewable water supply has
been used as a threshold of water stress. Under this de¬nition approx-
imately 1.7 billion people, one-third of the world™s population, live in
water-stressed countries. This number is projected to rise to around ¬ve
billion by 2025, dependent on the rate of population growth, this without
taking into account any effect on water supplies due to climate change.
For instance, in India about seventy-¬ve per cent of available water is cur-
rently used for irrigation. There is therefore very little surplus for future
growth, the only river in north India with a surplus is the Brahmaputra.
Similar situations where nearly all the available water is currently used
prevail in much of central and western Asia.22 Many other developing
countries especially in Africa face similar situations.
A further vulnerability arises because many of the world™s major
sources of water are shared. About half the land area of the world is
within water basins which fall between two or more countries. There are
forty-four countries for which at least eighty per cent of their land areas
falls within such international basins. The Danube, for instance, passes
through twelve countries which use its water, the Nile water through
nine, the Ganges-Bramaputra through ¬ve. Other countries where water
is scarce are critically dependent on sharing the resources of rivers such as
the Euphrates and the Jordan. The achievements of agreements to share
water often bring with them demands for more effective use of the water
and better management. Failure to agree brings increased possibility
of tension and con¬‚ict. The former United Nations Secretary General,
Boutros Boutros-Ghali, has said that ˜The next war in the Middle East
will be fought over water, not politics™.23

The impact of climate change on fresh water
The availability of fresh water will be substantially changed in a world
affected by global warming. We saw in Chapter 6 (Figure 6.5(b)) that,
although there remains substantial uncertainty in model predictions of
precipitation change, it is possible to identify some areas where it is
158 The impacts of climate change

Figure 7.7 Simulations of
average monthly runoff in the
Sacramento basin of California
comparing (a) current climate
with (b) changed climates
with a 4 —¦ C temperature
increase and a twenty per cent
increase in rainfall and (c) with
the same temperature
increase but with a twenty per
cent decrease of rainfall.

likely that there will be signi¬cant increases or decreases. For instance,
precipitation is expected to increase in northern high latitudes in winter
and the monsoon regions of south-east Asia in summer, while other
regions (e.g. southern Europe, Central America, southern Africa and
Australia) can expect signi¬cantly drier summers. Further, increase of
temperature will mean that a higher proportion of the water falling on the
Earth™s surface will evaporate. In regions with increased precipitation,
some or all of the loss due to evaporation may be made up. However, in
regions with unchanged or less precipitation, there will be substantially
less water available at the surface. The combined effect of less rainfall
and more evaporation means less soil moisture available for crop growth
and also less runoff “ in regions with marginal rainfall this loss of soil
moisture can be critical.
The runoff in rivers and streams is what is left from the precipitation
that falls on the land after some has been taken by evaporation and by
transpiration from plants; it is the major part of what is available for hu-
man use. The amount of runoff is highly sensitive to changes in climate;
even small changes in the amount of precipitation or in the tempera-
ture (affecting the amount of evaporation) can have a big in¬‚uence on
it. To illustrate this, Figure 7.7 shows simulations, carried out for the
Sacramento Basin in California, USA (a region where water is stored for
some of the year in mountain snow), of changes in runoff with changed
climate conditions. With a 4 —¦ C regional temperature rise and twenty
per cent decreased rainfall, the runoff in the summer months falls to
between twenty per cent and ¬fty per cent of its normal value. Even with
twenty per cent increased rainfall and the same temperature increase,
The impact of climate change on fresh water resources 159

summer runoff still remains well below normal. Watersheds in arid or
semi-arid regions are especially sensitive because the annual runoff is in
any case highly variable.
Some watersheds in mid latitudes in the Northern Hemisphere, where
snowmelt is an important source of runoff, can also be severely affected.
For these places, as temperatures rise, winter runoff will increase sub-
stantially and spring high water will be much reduced. Further, as we
saw earlier in the chapter, up to one-half of the mass of mountain glaciers
and small ice caps may melt away over the next hundred years which
could substantially change the seasonal distribution of river ¬‚ow and
water supply for hydroelectric generation and agriculture.
A detailed hydrological study of Asia under climate change illus-
trates the very different impacts on different parts of the continent.24 Pro-
jections from a climate model for 2050 under a scenario similar to A1B
were introduced into a hydrological model of different river catchments
and the changes in runoff in the river basins estimated. In arid or semi-
arid regions of Asia, surface runoff is expected to decrease drastically
so affecting the volume of water available for irrigation and other pur-
poses. Average annual runoff in the basins of the Tigris, Euphrates, Indus
and Brahmaputra rivers was estimated to decline by twenty-two, twenty-
¬ve, twenty-seven and fourteen per cent respectively. Other areas will
experience substantial increases in runoff, for instance by thirty-seven
and twenty-six per cent, respectively, in the Yangtze (Changjiang) and
Huang He rivers. Substantial increases were also projected for Siberian
Watersheds that are particularly vulnerable to climate change can be
identi¬ed by asking certain questions about them.25

r How much water storage is there in the watershed relative to the annual
¬‚ow? In Colorado in the United States, for instance, the storage is four
times the annual ¬‚ow, whereas in the Atlantic States it is only one-tenth
of the annual ¬‚ow.
r How large is the demand as a percentage of the potential supply? This
varies a great deal. For instance, in North America, for the Rio Grande
and for the lower Colorado demand approximately equals supply and
very little of either of these rivers reaches the sea. Therefore, though
the Colorado has substantial storage and is therefore not very sensitive
to annual variations, the amount of use in its lower reaches means that,
over a number of years, any reduction of its ¬‚ow is bound to imply
lower water availability.
r How much groundwater is being used? There are many places in the
world where groundwater is being used faster than it is being replen-
ished. To give two examples, for more than half the land area of the
160 The impacts of climate change

United States over a quarter of the groundwater withdrawn is not being
replenished, so every year the water has to be extracted from deeper
levels; and in Beijing in China the water table is falling by 2 m a year
as its groundwater is pumped out.
r How variable are the stream and river ¬‚ows? This question is particu-
larly relevant to arid and semi-arid areas. Detailed studies taking these
criteria into account have been carried out for a number of areas; one
example for the MINK (Missouri, Iowa, Nebraska and Kansas) region
of the United States is shown in the box.

Study of the ˜MINK™ region in the United States
The United States Department of Energy has carried out a detailed
study26 of the likely effects of climate change on a region (known as the
MINK region) in the centre of the United States comprising the states
of Missouri, Iowa, Nebraska and Kansas. Included within the region are
parts of four major river basins “ the Missouri, the Arkansas, the Upper
and the Lower Mississippi. Water is already a scarce resource within
the MINK states; much of the area™s irrigation relies on non-renewable
groundwater supplies. These will diminish with time, so that even in the
absence of climate change less water will be available, especially for
To provide an analogue of the climate which might be expected with
increased carbon dioxide, the period of the 1930s was chosen, when the
average temperature in the region was about 1 —¦ C warmer than in the
period 1950“1980 (the ˜control™ period) and the average precipitation
about ten per cent lower than in the control period.
Water would become scarcer under the analogue climate compared
with the control.27 The hotter and drier conditions would increase evapo-
ration and reduce runoff. Stream¬‚ow would drop by about thirty per cent
in the Missouri and the Upper Mississippi basins and by about ten per cent
in the Arkansas. Most streams would fall well short of supplying both
the desired instream ¬‚ows and the current levels of consumption use.
Under the analogue climate, irrigated agriculture would be bound
to decline substantially because of the increased constraints on ground-
water use coupled with less water availability from other sources. This
would also result in a drive to increased ef¬ciency, albeit at greater cost.
Maintaining the high priority currently given to navigation on the main
stem of the Missouri would become very costly.

So far when mentioning changes in temperature or rainfall it is
changes in the average with which we have been concerned; for in-
stance, the simulations in Figure 7.7 are for average conditions. But,
as has been constantly emphasised, the severity of climate impacts
The impact of climate change on fresh water resources 161

depends to a great degree on extreme conditions. This is well illus-
trated by looking at the scale of natural disasters involving water “ either
too much water in ¬‚oods or too little in droughts. Some of the most
damaging ¬‚oods of recent years were mentioned in Chapter 1 (page 5)
“ see also Table 7.3 (page 183). Droughts do not appear high up on the
table of natural disasters, not because they are unimportant, but because,
unlike most other disasters, their effects tend to be felt over a long per-
iod of time. The ˜dust bowl™ years in the 1930s in the United States
are still within living memory, as are the droughts and famines in India
in 1965“7 which, it is estimated, claimed one and a half million lives.
Recent decades have seen a series of damaging droughts in the Sahel
region and in other parts of Africa28 “ which are still recurring only too
frequently on that continent.
Any temperature or rainfall record shows a large variability. The
inevitable result of variability added to higher average temperatures
(meaning higher evaporation) and higher average rainfall will be a greater
number and greater intensity of both droughts and ¬‚oods.29 For instance,
associated with the substantial changes in average runoff expected by
2050 in parts of Asia mentioned above will be increases in the num-
ber and intensity of ¬‚oods and droughts. Some of the areas likely to be
affected are just those areas that are particularly vulnerable at the mo-
ment “ although, as was also implied in Chapter 6, droughts and ¬‚oods
are increasingly likely to occur in some locations where, at present, such
disasters are rare. Very few quantitative estimates have been made of the
likely increase in ¬‚oods or droughts as a result of the increase of green-
house gases. One estimate quoted in Chapter 6 (page 131) projected an
increase of a factor of ¬ve in intense precipitation events in parts of
Europe under a doubled atmospheric carbon dioxide concentration.
The monsoon regions of southeast Asia are an example of an area that
may be particularly vulnerable to both ¬‚oods and droughts. Figure 7.8
shows the predicted change in summer precipitation over the Indian sub-
continent as simulated by a regional climate model (RCM) for 2050
under a scenario similar to SRES A1B. Note the improvement in detail
of the precipitation pattern that results from the use of the increased
resolution of the regional model compared with the global model
(GCM), for instance over the Western Ghats (the mountains that rise
steeply from India™s west coast) there are large increases not present
in the global model simulation. The most serious reductions in water
availability simulated by the regional model are in the arid regions of
northwest India and Pakistan where average precipitation is reduced to
less than 1 mm day’1 “ that coupled with higher temperatures leads to a
sixty per cent decline in soil moisture. Substantial increases in average
precipitation are projected for areas in eastern India and in ¬‚ood-prone
162 The impacts of climate change

Figure 7.8 Predicted GCM RCM
changes in monsoon rainfall
(mm/day) over India between
the present day and the
middle of the twenty-¬rst
century from a 300-km
resolution GCM and from a
50-km resolution RCM. The
RCM pattern is very different
in some respects from the
coarser resolution pattern of
the GCM.

0.2 1
“3 “1 “0.2 0 3
Precipitation (mm day’1)

Bangladesh where the projected increase is about twenty per cent. What
is urgently required for this part of the world and elsewhere is much
better information linking changes in average parameters with likely
changes in frequency, intensity and location of extreme events.
There is another reason, not unconnected with global warming, for
the vulnerability of water supplies: the link between rainfall and changes
in land use. Extensive deforestation can lead to large changes in rainfall
(see box on page 173). A similar tendency to reduced rainfall can be
expected if there is a reduction in vegetation over large areas of semi-
arid regions. Such changes can have a devastating and widespread effect
and assist in the process of deserti¬cation. This is a potential threat to
the drylands covering about one-quarter of the land area of the world
(see box on page 163).
What sort of action can be taken to lessen the vulnerability of hu-
man communities to changes in water availability or supply? Irrigation
accounts for about two-thirds of world water use, and is of great im-
portance to world agriculture. Irrigation is applied to about one-sixth
of the world™s farmland, which produces about one-third of the world™s
crops. In some areas the ratio is much higher; for instance, over eighty
per cent of the agricultural land in California is irrigated. Most irriga-
tion is through open ditches, which is very wasteful of water; over sixty
per cent is lost through evaporation and seepage. Microirrigation tech-
niques, in which perforated pipes deliver water directly to the plants,
provide large opportunities for water conservation, making it possible
to expand irrigated ¬elds without building new dams.30 Management of
the existing infrastructure can be improved, for instance by arranging for
the integration of different supplies, and conservation in the domestic
and industrial sectors can be encouraged. Most of these actions will cost
The impact of climate change on fresh water resources 163

Drylands (de¬ned as those areas where precipitation is low and where
rainfall typically consists of small, erratic, short, high-intensity storms)
cover about forty per cent of the total land area of the world and support
over one-¬fth of the world™s population. Figure 7.9 shows how these arid
areas are distributed over the continents.
Deserti¬cation in these drylands is the degradation of land brought
about by climate variations or human activities that have led to decreased
vegetation, reduction of available water, reduction of crop yields and ero-
sion of soil. The United Nations Convention to Combat Deserti¬cation
(UNCCD) set up in 1996 estimates that over seventy per cent of these
drylands, covering over twenty-¬ve per cent of the world™s land area, are
degraded33 and therefore affected by deserti¬cation. The degradation can
be exacerbated by excessive land use or increased human needs (gener-
ally because of increased population), or political or economic pressures
(for instance, the need to grow cash crops to raise foreign currency). It
is often triggered or intensi¬ed by a naturally occurring drought.
The progress of deserti¬cation in some of the drylands will be in-
creased by the more frequent or more intense droughts that are likely to
result from climate change during the twenty-¬rst century.

Figure 7.9 The world™s drylands, by continent. The total area of
drylands is about sixty million square kilometres (about forty per cent of
the total land area), of which ten million are hyper-arid deserts.
164 The impacts of climate change

money, although they may be much more cost-effective ways of coping
with future change in water resources than attempting to develop major
new facilities.31
In summary, what are the likely effects of global warming on water
supplies? Firstly, the current vulnerability of many communities to water
shortage should be noted. This is especially true of arid and semi-arid
regions where the increasing demands of human communities mean
that droughts, even for short periods, are more disastrous than before.
Vulnerability is well demonstrated in many areas of the world where the
amounts of groundwater extraction greatly exceed its replenishment “ a
situation that cannot continue for very long into the future. Because of
population growth these vulnerabilities will increase and will exacerbate
the negative effects of global warming.
Secondly, climate change because of global warming will result in
large changes in water supplies in many places. Although the present state
of knowledge regarding regional and local climate change does not allow
scientists to identify precisely the most vulnerable areas, they are able
to indicate the sort of area which will be most affected. Such areas are
those arid and semi-arid areas with reduced rainfall leading to greater
aridity and even deserti¬cation; continental areas where decreased sum-
mer rainfall and increased temperature result in a substantial loss in soil
moisture and much increased vulnerability to drought; and areas where
increased rainfall could lead to a greater incidence of ¬‚oods. The chang-
ing pattern of climate extremes, especially droughts and ¬‚oods, will be
the cause of most of the problems. It is also the case that regions such
as southeast Asia that are dependent on unregulated river systems are
more sensitive to change than regions such as western Russia and the
western United States that have large, regulated water resource systems.
Thirdly, some of the adverse impact of climate change on water supplies
can be reduced by taking appropriate alleviating action, by introducing
more careful and integrated water management32 and by introducing
more effective disaster preparedness in the most vulnerable areas.

Impact on agriculture and food supply
Every farmer understands the need to grow crops or rear animals that are
suited to the local climate. The distribution of temperature and rainfall
during the year are key factors in making decisions regarding what crops
to grow. These will change in the world in¬‚uenced by global warm-
ing. The patterns of what crops are grown where will therefore also
change. But these changes will be complex; economic and other factors
will take their place alongside climate change in the decision-making
Impact on agriculture and food supply 165

There is enormous capacity for adaptation in the growth of crops for
food “ as is illustrated by what was called the Green Revolution of the
1960s, when the development of new strains of many species of crops
resulted in large increases in productivity. Between the mid 1960s and
the mid 1980s global food production rose by an average annual rate
of 2.4% “ faster than global population “ more than doubling over that
thirty-year period. Grain production grew even faster, at an annual rate
of 2.9%.34 There are concerns that factors such as the degradation of
many of the world™s soils largely through erosion and the slowed rate of
expansion of irrigation because less fresh water is available will tend to
reduce the potential for increased agricultural production in the future.
However, with declining rates of population growth, there remains opti-
mism that, in the absence of major climate change, the growth in world
food supply is likely to continue to match the growth in demand at least
during the early decades of the twenty-¬rst century.35
What will be the effect of climate change on agriculture and food sup-
ply? With the detailed knowledge of the conditions required by different
species and the expertise in breeding techniques and genetic manipula-
tion available today, there should be little dif¬culty in matching crops to
new climatic conditions over large parts of the world. At least, that is the
case for crops that mature over a year or two. Forests reach maturity over
much longer periods, from decades up to a century or even more. The
projected rate of climate change is such that, during this time, trees may
¬nd themselves in a climate to which they are far from suited. The tem-
perature regime or the rainfall may be substantially changed, resulting in
stunted growth or a greater susceptibility to disease and pests. The im-
pact of climate change on forests is considered in more detail in the next
An example of adaptation to changing climate is the way in which
farmers in Peru adjust the crops they grow depending on the climate
forecast for the year.36 Peru is a country whose climate is strongly in¬‚u-
enced by the cycle of El Ni˜ o events described in Chapters 1 and 5. Two
of the primary crops grown in Peru, rice and cotton, are very sensitive
to the amount and the timing of rainfall. Rice requires large amounts
of water; cotton has deeper roots and is capable of yielding greater pro-
duction during years of low rainfall. In 1983, following the 1982“3 El
Ni˜ o event, agricultural production dropped by fourteen per cent. By
1987 forecasts of the onset of El Ni˜ o events had become suf¬ciently
good for Peruvian farmers to take them into account in their planning.
In 1987, following the 1986“7 El Ni˜ o, production actually increased
by three per cent, thanks to a useful forecast.
Three factors are particularly important in considering the effect
of climate change on agriculture and food production. The availability
166 The impacts of climate change

of water is the most important of the factors. The vulnerability of water
supplies to climate change carries over into a vulnerability in the growing
of crops and the production of food. Thus the arid or semi-arid areas,
mostly in developing countries, are most at risk. A second factor, which
tends to lead to increased production as a result of climate change, is the
boost to growth that is given, particularly to some crops, by increased
atmospheric carbon dioxide (see box below). A third factor is the effect of
temperature changes; in particular, under very high temperatures, yields
of some crops are substantially reduced.

The carbon dioxide ˜fertilisation™ effect
An important positive effect of increased carbon dioxide (CO2 ) con-
centrations in the atmosphere is the boost to growth in plants given by
the additional CO2 . Higher CO2 concentrations stimulate photosynthe-
sis, enabling the plants to ¬x carbon at a higher rate. This is why in
glasshouses additional CO2 may be introduced arti¬cially to increase
productivity. The effect is particularly applicable to what are called C3
plants (such as wheat, rice and soya bean), but less so to C4 plants (for
example, maize, sorghum, sugar-cane, millet and many pasture and for-
age grasses). Under ideal conditions it can be a large effect; for C3 crops
under doubled CO2 , an average of +30%.37 However, under real con-
ditions on the large scale where water and nutrient availability are also
important factors in¬‚uencing plant growth, experiments show that the
increases, although dif¬cult to measure accurately, tend to be substan-
tially less than the ideal.38 In experimental work, grain and forage quality
declines with CO2 enrichment and higher temperatures. More research
is required especially for many tropical crop species and for crops grown
under suboptimal conditions (low nutrients, weeds, pests and diseases).

Detailed studies have been carried out of the sensitivity to climate
change during the twenty-¬rst century of the major crops which make up
a large proportion of the world™s food supply (see box below). They have
used the results of climate models to estimate changes in temperature
and precipitation. Many of them study the effect of CO2 fertilisation and
some also model the effects of climate variability as well as changes in
the means. Some also include the possible effects of economic factors
and of modest levels of adaptation. These studies in general indicate that
the bene¬t of increased CO2 concentration on crop growth and yield does
not always overcome the effects of excessive heat and drought. For cereal
crops in mid latitudes, potential yields are projected to increase for small
Impact on agriculture and food supply 167

increases in temperature (2“3 —¦ C) but decrease for larger temperature
rises.39 In most tropical and subtropical regions, potential yields are pro-
jected to decrease for most increases in temperature; this is because
such crops are near their maximum temperature tolerance. Where there
is a large decrease in rainfall, tropical crop yields would be even
more adversely affected.
Taking the supply of food for the world as a whole, studies tend to
show that, with appropriate adaptation, the effect of climate change on
total global food supply is not likely to be large. However, none of them
have adequately taken into account the likely effect on food production of
climate extremes (especially of the incidence of drought), of increasingly
limited water availability or of other factors such as the integrity of the
world™s soils, which are currently being degraded at an alarming rate.40
A serious issue exposed by the studies is that climate change is likely
to affect countries very differently. Production in developed countries
with relatively stable populations may increase, whereas that in many
developing countries (where large increases in population are occurring)
is likely to decline as a result of climate change. The disparity between
developed and developing nations will tend to become much larger,
as will the number of those at risk from hunger. The surplus of food in
developed countries is likely to increase, while developing countries will
face increasing deprivation as their declining food availability becomes
much less able to provide for the needs of their increasing populations.
Such a situation will raise enormous problems, one of which will be
that of employment. Agriculture is the main source of employment in
developing countries; people need employment to be able to buy food.
With changing climate, as some agricultural regions shift, people will
tend to attempt to migrate to places where they might be employed in
agriculture. With the pressures of rising populations, such movement is
likely to be increasingly dif¬cult and we can expect large numbers of
environmental refugees.
In looking to future needs, two activities that can be pursued now
are particularly important. Firstly, there is large need for technical ad-
vances in agriculture in developing countries requiring investment and
widespread local training. In particular, there needs to be continued de-
velopment of programmes for crop breeding and management, especially
in conditions of heat and drought. These can be immediately useful in the
improvement of productivity in marginal environments today. Secondly,
as was seen earlier when considering fresh water supplies, improvements
need to be made in the availability and management of water for irriga-
tion, especially in arid or semi-arid areas of the world.
168 The impacts of climate change

Modelling the impact of climate change on world
food supply
An example illustrating the key elements of a detailed study of the impact
of climate change on world food supply is shown in Figure 7.10.41
A climate change scenario is ¬rst set up with a climate model of
the kind described in Chapter 5. Models of different crops that include
the effects of temperature, precipitation and CO2 are applied to 124
different locations in 18 countries to produce projected crop yields that
can be compared with projected yields in the absence of climate change.
Included also are farm-level adaptations, e.g. planting date shifts, more
climatically adapted varieties, irrigation and fertiliser application. These
estimates of yield are then aggregated to provide yield-change estimates
by crop and country or region.
These yield changes are then employed as inputs to a world food
trade model that includes assumptions about global parameters such
as population growth and economic change and links together national
and regional economic models for the agricultural sector through trade,
world market prices and ¬nancial ¬‚ows. The world food trade model
can explore the effects of adjustments such as increased agricultural in-
vestment, reallocation of agricultural resources according to economic
returns (including crop switching) and reclamation of additional arable
land as a response to higher cereal prices. The outputs from the total pro-
cess provide information projected up to the 2080s on food production,
food prices and the number of people at risk of hunger (de¬ned as the
population with an income insuf¬cient either to produce or to procure
their food requirements).
The main results with this model for the 2080s regarding the impact
of climate change following the IS 92a scenario (Figure 6.1) are that
yields at mid to high latitudes are expected to increase, and at low lati-
tudes (especially the arid and sub-humid tropics) to decrease. This pattern
becomes more pronounced as time progresses. The African continent is
particularly likely to experience marked reductions in yield, decreases
in production and an estimated sixty million or more additional people
at risk of hunger as a result of climate change.
The authors emphasise that, although the models and the methods
they have employed are comparatively complex, there are many fac-
tors that have not been taken into account. For instance, they have not
adequately considered the impact of changes in climate extremes, the
availability of water supplies for irrigation or the effects of future tech-
nological change on agricultural productivity. Further (see Chapter 6),
scientists are not yet very con¬dent in the regional detail of climate
change. The results, therefore, although giving a general indication of
the changes that could occur, should not be treated as a detailed predic-
tion. They highlight the importance of studies of this kind as a guide to
future action.
Impact on agriculture and food supply 169

Climate models

Observed change
climate scenarios

Crop models
CO2 effects adaptations
wheat, rice, maize, soya bean

Crop yield by site
and scenario
evapotranspiration, irrigation,
season length

Aggregation of site results
Agroecological zone analysis

Yield functions by region
Yield = function of temperature,
precipitation and CO2

Yield change estimates
Commodity group and
projections Population
Economic World food trade
growth rates model
Economic consequences
Shifts in trade
Incidence of food poverty

Figure 7.10 Illustrating key elements of a study of crop yield and food trade under a changed climate.
170 The impacts of climate change

The impact on ecosystems
A little over ten per cent of the world™s land area is under cultivation “
that was the area addressed in the last section. The rest is to a greater
or lesser extent unmanaged by humans. Of this about thirty per cent
is natural forest and between one and two per cent plantation forest.
The variety of plants and animals that constitute a local ecosystem is
sensitive to the climate, the type of soil and the availability of water.
Ecologists divide the world into biomes “ regions characterised by their
distinctive vegetation. This is well illustrated by information about the
distribution of vegetation over the world during past climates (e.g. for
the part of North America shown in Figure 7.11), which indicates what
species and what ecosystems are most likely to ¬‚ourish under different
climatic regimes.
Changes in climate alter the suitability of a region for different
species, and change their competitiveness within an ecosystem, so that
even relatively small changes in climate will lead, over time, to large
changes in the composition of an ecosystem. Since climate is the domi-
nant factor determining the distribution of biomes (Figure 7.12), infor-
mation gleaned from paleo sources could be used to produce maps of the
optimum distribution of natural vegetation under the climate scenarios
expected to occur with global warming.
However, changes of the kind illustrated in Figure 7.11 took place
over thousands of years. With global warming similar changes in climate
occur over a few decades. Most ecosystems cannot respond or migrate
that fast. Fossil records indicate that the maximum rate at which most
plant species have migrated in the past is about 1 km per year. Known
constraints imposed by the dispersal process (e.g. the mean period be-
tween germination and the production of seeds and the mean distance that
an individual seed can travel) suggest that, without human intervention,
many species would not be able to keep up with the rate of movement of
their preferred climate niche projected for the twenty-¬rst century, even
if there were no barriers to their movement imposed by land use.42 Nat-
ural ecosystems will therefore become increasingly unmatched to their
environment. How much this matters will vary from species to species:
some are more vulnerable to changes in average climate or climate ex-
tremes than others. But all will become more prone to disease and attack
by pests. Any positive effect from added ˜fertilisation™ due to increased
carbon dioxide is likely to be more than outweighed by negative effects
from other factors.
Trees are long-lived and take a long time to reproduce, so they cannot
respond quickly to climate change. Further, many trees are surprisingly
sensitive to the average climate in which they develop. The environmental
The impact on ecosystems 171

Figure 7.11 Vegetation maps of the south-eastern United States during past
climate regimes: (a) for 18 000 years ago at the maximum extent of the last ice
age, (b) for 10 000 years ago, (c) for 5000 years ago when conditions were
similar to present. A vegetation map for 200 years ago is similar to that in (c).
172 The impacts of climate change

Figure 7.12 The pattern of world biome types related to mean annual
temperature and precipitation. Other factors, especially the seasonal variations
of these quantities, affect the detailed distribution patterns (after Gates).

conditions (e.g. temperature and precipitation) under which a species can
exist and reproduce are known as its niche. Climate niches for some typ-
ical tree species are illustrated in Figure 7.13; under some conditions
a change as small as 1 —¦ C in annual average temperature can make a
substantial difference to a tree™s productivity. For the likely changes in
climate in the twenty-¬rst century, a substantial proportion of existing
trees will be subject to unsuitable climate conditions. This will be partic-
ularly the case in the boreal forests of the Northern Hemisphere where,
as trees become less healthy, they will be more prone to pests, die-back
and forest ¬res. One estimate projects that, under a doubled CO2 scen-
ario, up to sixty-¬ve per cent of the current boreal forested area could
be affected.43
A decline in the health of many forests in recent years has received
considerable attention, especially in Europe and North America where
much of it has been attributed to acid rain and other pollution originating
The impact on ecosystems 173

Forests, deforestation and climate change

in rainfall of over thirty per cent.45 A much more
Extensive changes in the area of forests due to de-
forestation can seriously affect the climate in the drastic experiment in which the Amazonian forest
region of change. Also, changes in temperature or was removed and replaced with a desert surface
rainfall that occur because of long-term changes in showed a reduction in rainfall by seventy per cent
climate can also have a major impact on forests. We to levels similar to those of the semi-arid regions
of the Sahel part of Africa.46 Such a model exper-
look at these effects in turn.
Changes in land use such as those brought about iment does not represent a realistic situation, but
by deforestation can affect the amount of rainfall, it illustrates the signi¬cant impact that widespread
for three main reasons. Over a forest there is a lot deforestation could have on the local climate.
more evaporation of water (through the leaves of More recent work has been with interactive
the trees) than there is over grassland or bare soil, models that include not just the effect of changes
hence the air will contain more water vapour. Also, in land use or forestation on the climate but also,
a forest re¬‚ects twelve to ¬fteen per cent of the sun- in a dynamic way, the effect on forests and other
light that falls on it, whereas grassland will re¬‚ect vegetation of changes in climate. In an experiment
about twenty per cent and desert sand up to forty with such a model that assumes carbon dioxide
emissions following the IS 92a scenario,47 substan-
per cent. A third reason arises from the roughness
of the surface where vegetation is present. tial reductions in precipitation are projected for
An American meteorologist, Professor Jules areas of Amazonia, that lead to die-back of the
Charney, suggested in 1975 that, in the context of Amazonian forest and signi¬cant release of carbon
the drought in the Sahel, there could be an impor- to the atmosphere (one of the positive climate feed-
tant link between changes of vegetation (and hence backs mentioned in the box on page 40). As the
changes of re¬‚ectivity) and rainfall. The increased forest dies back, the rainfall is further reduced be-
energy absorbed at the surface when vegetation is cause of the change in properties of the land sur-
present and the increased surface roughness both face, leading, by the end of the twenty-¬rst century,
tend to stimulate convection and other dynamic ac- to the replacement of much of the forest cover by
tivity in the atmosphere so leading to the production semi-arid conditions. Such results are still subject
of rainfall. to considerable uncertainties (for instance, those
Early experiments with numerical models that associated with the model simulations of El Ni˜ o n
included these physical processes demonstrated the events under climate change conditions and the con-
effect and indicated, for instance, a reduction of nections between these events and the climate over
about ¬fteen per cent in rainfall if the forest north of Amazonia), but they illustrate the type of impacts
30 —¦ S in South America were removed and replaced that might occur and emphasise the importance of
by grassland.44 Similar model experiments for Zaire understanding the interactions between climate and
over a smaller region showed an average reduction vegetation.

from heavy industry, power stations and motor cars. Not all damage to
trees, however, is thought to have this origin. Studies in several regions
of Canada, for instance, indicate that the die-back of trees there is related
to changes in climatic conditions, especially to successions of warmer
winters and drier summers.48 In some cases it may be the double effect of
pollution and climate stress causing the problem; trees already weakened
174 The impacts of climate change

Figure 7.13 Simulated environmental realised niches (the realised niche
describes the conditions under which the species is actually found) for three tree
species, Arolla Pine, Norway Spruce and Common Beech. Plots are of biomass
generated per year against annual means of temperature (T) and precipitation
(P). Arolla Pine is a species with a particularly narrow niche. The narrower the
niche, the greater the potential sensitivity to climate change.

by the effects of pollution fail to cope with climate stress when it comes.
The assessment of the impact of climate change carried out for the MINK
region of the United States (see box on page 160) concluded that, under
the warmer, drier conditions of the analogue climate they studied, decline
and die-back of the forested part of the region would reduce the mass of
timber in the forest by ten per cent over twenty years.49 The results of
The impact on ecosystems 175

these studies are indicative of the more serious levels of forest die-back
that are likely to occur with the rapid rate of climate change expected
with global warming (see box on page 173).These stresses on the world™s
forests due to climate change will be concurrent with other problems
associated with forests, in particular those of continuing tropical defor-
estation and of increasing demand for wood and wood products resulting
from rapidly increasing populations especially in developing countries.
If a stable climate is eventually re-established, given adequate time
(which could be centuries), different trees will be able to ¬nd again at
some location their particular climatic niche. It is during the period of
rapid change that most trees will ¬nd themselves unsatisfactorily located
from the climate point of view.
It was mentioned in Chapter 3 that forests represent a large store of
carbon; eighty per cent of above-ground and forty per cent of below-
ground terrestrial carbon is in forests. We also saw in Chapter 3 that
tropical deforestation due to human activities is probably releasing be-
tween 1 and 2 Gt of carbon into the atmosphere each year. If, because
of the rate of climate change, substantial stress and die-back occurs in
boreal and tropical forests (see box on page 173) a release of carbon will
occur. This positive feedback was mentioned in Chapter 3 (see the box
on page 40). Just how large this will be is uncertain but estimates as high
as 240 Gt over the twenty-¬rst century for the above-ground component
alone have been quoted.50
The above discussion has largely related to the impact of climate
change on natural forests where the likely impacts are largely negative.
Studies of the impacts on managed forests are more positive.51 They sug-
gest that with appropriate adaptation and land and product management,
even without forestry projects that increase the capture and storage of
carbon (see Chapter 10), a small amount of climate change could in-
crease global timber supply and enhance existing market trends towards
rising market share in developing countries.
A further concern about natural ecosystems relates to the diversity of
species that they contain and the loss of species and hence of biodiversity
due to the impact of climate change. Signi¬cant disruptions of ecosys-
tems from disturbances such as ¬re, drought, pest infestation, invasion of
species, storms and coral bleaching events are expected to increase. The
stresses caused by climate change, added to other stresses on ecological
systems (e.g. land conversion, land degradation, deforestation, harvest-
ing and pollution) threaten substantial damage to or complete loss of
some unique ecosystems, and the extinction of some endangered species.
Coral reefs and atolls, mangroves, boreal and tropical forests, polar and
alpine ecosystems, prairie wetlands and remnant native grasslands are
examples of systems threatened by climate change. In some cases the
176 The impacts of climate change

threatened ecosystems are those that could mitigate against some climate
change impacts (e.g. coastal systems that buffer the impact of storms).
Possible adaptation methods to reduce the loss of biodiversity include
the establishment of refuges, parks and reserves with corridors to allow
migration of species, and the use of captive breeding and translocation
of species.52
So far we have been considering ecosystems on land. What about
those in the oceans; how will they be affected by climate change? Al-
though we know much less about ocean ecosystems, there is considerable
evidence that biological activity in the oceans has varied during the cycle
of ice ages. Chapter 3 noted (see box on page 35) the likelihood that it
was these variations in marine biological activity which provided the
main control on atmospheric carbon dioxide concentrations during the
past million years (see Figure 4.4). The changes in ocean water tempera-
ture and the possible changes in some of the patterns of ocean circulation
are likely to result in changes in the regions where upwelling occurs and
where ¬sh congregate. Some ¬sheries could collapse and others expand.
At the moment the ¬shing industry is not well adapted to address major
Some of the most important marine ecosystems are found within
coral reefs that occur in many locations throughout the tropical and
subtropical world. They are especially rich in biodiversity and are par-
ticularly threatened by global warming. Within them the species diver-
sity contains more phyla than rainforests and they harbour more than
twenty-¬ve per cent of all known marine ¬sh.54 They represent a signif-
icant source of food for many coastal communities. Corals are partic-
ularly sensitive to sea surface temperature and even one degree Celsius
of persistent warming can cause bleaching (paling in colour) and ex-
tensive mortality accompanies persistent temperature anomalies of 3 —¦ C
or more. Much recent bleaching, for instance that in 1998, have been
associated with El Ni˜ o events.55

The impact on human health
Human health is dependent on a good environment. Many of the factors
that lead to a deteriorated environment also lead to poor health. Pollution
of the atmosphere, polluted or inadequate water supplies, and poor soil
(leading to poor crops and inadequate nutrition) all present dangers to
human health and wellbeing and assist the spread of disease. As has
been seen so far in considering the impacts of global warming, many
of these factors will tend to be exacerbated through the climate change
that will occur in the warmer world. The greater likelihood of extremes
of climate, such as droughts and ¬‚oods, will also bring greater risks to
The impact on human health 177

health from increased malnutrition and from a prevalence of conditions
more likely to lead to the spread of diseases from a variety of causes.
How about direct effects of the climate change itself on human
health?56 Humans can adapt themselves and their buildings so as to live
satisfactorily in very varying conditions and have great ability to adapt to
a wide range of climates. The main dif¬culty in assessing the impact of
climate change on health is that of unravelling the in¬‚uences of climate
from the large number of other factors (including other environmental
factors) that affect health.
The main direct effect on humans will be that of heat stress in the
extreme high temperatures that will become more frequent and more
widespread especially in urban populations (see box and Figure 6.6). In
large cities where heat waves commonly occur death rates can be doubled
or tripled during days of unusually high temperatures.57 Although such
episodes may be followed by periods with fewer deaths showing that
some of the deaths would in any case have occurred about that time,
most of the increased mortality seems to be directly associated with the
excessive temperatures with which old people in particular ¬nd it hard
to cope. On the positive side, mortality due to periods of severe cold in
winter will be reduced. The results of studies are equivocal regarding

Heat waves in Europe, 2003
Record extreme temperatures were experienced in Europe during June,
July and August 2003. At many locations temperature rose over 40 —¦ C.
In France, Italy, the Netherlands, Portugal and Spain, over 21 000 ad-
ditional deaths were attributed to the unrelenting heat. Spain, Portugal,
France and countries in Central and Eastern Europe suffered from in-
tense forest ¬res.58 Figure 7.14 illustrates the extreme rareness of this
Figure 7.14 Distribution
event. of average summer
temperatures (June, July,
August) in Switzerland
from 1864“2003 showing
a ¬tted gaussian
probability distribution “
standard deviation 0.94 —¦ C.
Average temperatures with
10,100 and 1000 year
return periods are also
indicated. The 2003 value
is 5.4 standard deviations
from the mean showing it
as an extremely rare event
(from Schar et al. 2004,
Nature 427, 332“6).
178 The impacts of climate change

whether the reduction in winter mortality will be greater or less than
the increase in summer mortality. These studies have largely been con-
¬ned to populations in developed countries, precluding a more general
comparison between changes in summer and winter mortality.
A further likely impact of climate change on health is the increased
spreading of diseases in a warmer world. Many insect carriers of dis-
ease thrive better in warmer and wetter conditions. For instance, epi-
demics of diseases such as viral encephalitides carried by mosquitoes
are known to be associated with the unusually wet conditions that oc-
cur in the Australian, American and African continents associated with
different phases of the El Ni˜ o cycle.59 Some diseases, currently largely
con¬ned to tropical regions, with warmer conditions could spread into
mid latitudes. Malaria is an example of such a disease that is spread
by mosquitoes under conditions which are optimum in the tempera-
ture range of 15“32 —¦ C with humidities of ¬fty to sixty per cent. It
currently represents a huge global public health problem, causing an-
nually around 300 million new infections and over one million deaths.
Under climate change scenarios, most predictive model studies indi-
cate a net increase in the geographic range (and in the populations at
risk) of potential transmission of malaria and dengue infections, each
of which currently impinge on forty to ¬fty per cent of the world™s
population. Other diseases that are likely to spread for the same rea-
son are yellow fever and some viral encephalitis. In all cases, how-
ever, actual disease occurrence will be strongly in¬‚uenced by local envi-
ronmental conditions, socio-economic circumstances and public health
The potential impact of climate change on human health could be
large. However, the factors involved are highly complex; any quanti-
tative conclusions will require careful study of the direct effects of
climate on humans and of the epidemiology of the diseases that are
likely to be particularly affected. Some remarks about how the health
impacts of extremes and disasters might be reduced are given in the next

Adaptation to climate change
As we have seen, some of the impacts of climate change are already
becoming apparent. A degree of adaptation therefore has already be-
come a necessity. Numerous possible adaptation options for responding
to climate change have already been identi¬ed. These can reduce ad-
verse impacts and enhance bene¬cial effects of climate change and can
also produce immediate ancillary bene¬ts, but they cannot prevent all
Costing the impacts: extreme events 179

damages. Examples of adaptation options are listed in Table 7.1. Many
of the options listed are presently employed to cope with current climate
variability and extremes; their expanded use can enhance both current
and future capacity to cope. But such actions may not be as effective in
the future as the amount and rate of climate change increase. To make
a list of possible adaptation options is relatively easy. If they are to be
applied effectively, much more information is urgently required regard-
ing the detail and the cost of their application over the wide range of
circumstances where they will be required.
Of particular importance is the requirement for adaptation to ex-
treme events and disasters such as ¬‚oods, droughts and severe storms.60
Vulnerability to such events can be substantially reduced by much more
adequate preparation.61 For instance, following hurricanes George and
Mitch, the Pan American Health Organisation (PAHO) identi¬ed a range
of policies to reduce the impact of such events62 :
r Undertaking vulnerability studies of existing water supply and sanita-
tion systems and ensuring that new systems are built to reduce vulner-
r Developing and improving training programmes and information sys-
tems for national programmes and international cooperation on emer-
gency management.
r Developing and testing early warning systems that should be coord-
inated by a single national agency and involve vulnerable communities.
Provision is also required for providing and evaluating mental care,
particularly for those who may be especially vulnerable to the adverse
psychosocial effects of disasters (e.g. children, the elderly and the

Costing the impacts: extreme events
In the previous paragraphs the impacts of climate change have been
described in terms of a variety of measures; for instance, the number of
people affected (e.g. by mortality, disease or by being displaced), the gain
or loss of agricultural or forest productivity, the loss of biodiversity, the
increase in deserti¬cation, etc. However, the most widespread measure,
looked for by many policymakers, is monetary cost or bene¬t. But before
describing what has been done so far to estimate the overall costs of
impacts, we need to consider what is known about the cost of damage
due to extreme events (such as ¬‚oods, droughts or windstorms). As has
been constantly emphasised in this chapter these probably constitute the
most important element in climate change impacts.
180 The impacts of climate change

See Table 3.6 on p. 76 of Watson, R. et al. (eds.) 2001, Climate Change 2001: Synthesis Report.

Because the incidence of such extreme events has increased signi¬-
cantly in recent decades, information about the cost of the damage due
to them has been tracked by insurance companies. They have catalogued
both the insured losses and, so far as they have been able to estimate, the
total economic losses “ these latter have shown an approximately tenfold
increase from the 1950s to the 1990s (see Figure 1.2 and box below).
Although factors other than climate change have contributed to this in-
crease, climate change is probably the factor of most signi¬cance. The
estimates for the 1990s of annual economic losses from weather-related
Costing the impacts: extreme events 181

disasters amount to approximately 0.2% of global world product (GWP)
and vary from about 0.3% of aggregate GDPs for the North and Central
American and the Asian regions to less than 0.1% for Africa (Table 7.2).
These average ¬gures hide big regional and temporal variations. For in-
stance, the annual loss in China from natural disasters from 1989 to 1996
is estimated to range from three to six per cent of GDP, averaging nearly
four per cent63 “ over ten times the world average. The reason why the
percentage for Africa is so low is not because there are no disasters there “
Africa on the whole has more than its fair share “ but because most of the
damage in African disasters is not realised in economic terms, nor does
it appear in economic statistics. Further such averaged numbers hide the
severe impact of disasters on individual countries or regions which, as
we mention below with the example of Hurricane Mitch, can prove to
be very large indeed.
The percentages we have quoted are conservative in that they do not
represent all relevant costs. They relate to direct economic costs only and
do not include associated or knock-on costs of disasters. This means, for
instance, that the damage due to droughts is seriously underestimated.
Droughts tend to happen slowly and many of the losses may not be
recorded or borne by those not directly affected. Another reason for
treating the information in the box with caution is because of the large
disparities between different parts of the world and countries regarding
per capita wealth, standard of living and degree of insurance cover. For
instance, probably the most damaging hurricane ever, Hurricane Mitch,
that hit Central America in 1998 does not appear in Table 7.3 as the total
insured losses were less than one billion dollars. In that storm, 600 mm
of rainfall fell in forty-eight hours, there were 9000 deaths and economic
losses estimated at over six billion dollars. The losses in Honduras and
Nicaragua amounted to about seventy and forty-¬ve per cent respectively
of their annual gross national product (GNP). Another example that does
not appear in Table 7.3 for the same reason is the ¬‚oods in central Europe
in 1997 that caused the evacuation of 162 000 people and over ¬ve billion
dollars of economic damage.
How about the likely costs of extreme events in the future? To es-
timate those we need much more quantitative information about their
likely future incidence and intensity. Very few such estimates exist. One
was mentioned in Chapter 6 (Figure 6.8) “ a possible factor of ¬ve
in extreme precipitation events in Europe under doubled pre-industrial
carbon dioxide concentration. A speculative but probably conservative
calculation of a global average ¬gure for the future might be obtained as
follows. Beginning with the 0.2% or 0.3% of GDP from the insurance
companies™ estimate of the current average costs due to weather-related
Table 7.2 Fatalities, economic losses and insured losses (both in 1999 US dollars) for disasters in different regions as
estimated by the insurance industry for the period 1985“99. The percentage from weather-related disasters (including
windstorms, ¬‚oods, droughts, wild¬re, landslides, land subsidence, avalanches, extreme temperature events, lightning,
frost and ice/snow damages) is indicated in each case. Total losses are higher than those summarised in Figure 1.2
because of the restriction of Figure 1.2 to losses from large catastrophic events

America: North, Central,
Africa South Caribbean Asia Australia Europe World

Number of events 810 610 2260 2730 600 1810 8820
Weather-related 91% 79% 87% 78% 87% 90% 85%
Fatalities 22 990 56 080 37 910 429 920 4400 8210 559 510
Weather-related 88% 50% 72% 70% 95% 96% 70%
Economic losses
(current $US billion) 7 16 345 433 16 130 947
Weather-related 81% 73% 84% 63% 84% 89% 75%
Insured losses
(current US$ billion) 0.8 0.8 119 22 5 40 187
Weather-related 100% 69% 86% 78% 74% 98% 87%

Data from Munich Re, presented in Figure 8.6 in Vellinga, P., Mills, E. et al. 2001. In McCarthy, Climate Change 2001: Impacts,
Chapter 8.
The insurance industry and climate change

The impact of climate on the insurance industry is climatic factors such as changes in precipitation,
mainly through extreme weather events. In devel- ¬‚ooding and drought events. There are differences
oping countries there may be very high mortality in balance between the causes by region and type
of event. Because of the complexities involved in
from extreme weather but relatively small costs
delineating both the socio-economic and the cli-
to the industry because of low insurance penetra-
matic factors, the proportion of the contribution
tion. In developed countries the loss of life may
from human-induced climate change cannot be de-
be much less but the costs to the insurance in-
¬ned with any certainty “ although it is interest-
dustry can be very large. Figure 1.2 illustrates the
large growth in weather-related disasters and the ing to note that the growth rate in damage cost of
weather-related events was three times that of non-
associated economic and insured losses since the
1950s and Table 7.2 the distribution of the disasters, weather-related events for the period 1960“99.
fatalities and economic losses from 1985 to 1999 Recent history has shown that weather-related
around the continents. Some idea of the types of losses can stress insurers to the point of bankruptcy.
Hurricane Andrew in 1992 broke the twenty bil-
disaster that cause the largest economic loss can be
lion dollar barrier for insured loss and served as a
gleaned from Table 7.3.
wake-up call to the industry. The insurance indus-
Part of the observed upward trend in historical
try, therefore, is very concerned to make estimates,
disaster losses is linked to socio-economic factors
such as population growth, increased wealth and as accurately as possible, for future trends as the
urbanisation in vulnerable areas; part is linked to rate of climate change increases.

Table 7.3 Individual events included in the aggregates in Table 7.2 that incurred
over ¬ve billion dollars of economic loss and over one billion dollars of insured loss

Economic losses Ratio: insured/
Year Event Area (bn $US) economic losses

1995 Earthquake Japan 112.1 0.03
1994 Northridge Earthquake USA 50.6 0.35
1992 Hurricane Andrew USA 36.6 0.57
1998 Floods China 30.9 0.03
1993 Floods USA 18.6 0.06
1991 Typhoon Mireille Japan 12.7 0.54
1989 Hurricane Hugo Caribbean, USA 12.7 0.50
1999 Winterstorm Lothar Europe 11.1 0.53
1998 Hurricane Georges Caribbean, USA 10.3 0.34
1990 Winterstorm Daria Europe 9.1 0.75
1993 Blizzard USA 5.8 0.34
1996 Hurricane Fran USA 5.7 0.32
1987 Winterstorm W. Europe 5.6 0.84
1999 Typhoon Bart Japan 5.0 0.60

Data from Munich Re, presented in Table 8.3 in Vellinga, P., Mills, E. et al. 2001. In
McCarthy, Climate Change 2001: Impacts, Chapter 8.
184 The impacts of climate change

extreme events, then multiplying by two to allow for the factors men-
tioned above (e.g. associated or knock-on costs) and further multiplying
by three to allow for the possible increase in extreme events, say by
the middle of the twenty-¬rst century, we end up with a ¬gure of be-
tween one and two per cent of GDP. Further, this again is a ˜money™
estimate. The real total costs of extreme events taking into account all
damages (including those that cannot be expressed in money terms)
are likely to be very signi¬cantly larger especially in many developing

Costing the total impacts
We now turn to consider all the impacts of anthropogenic climate
change, attempts that have been made to express their cost in monetary
terms and the validity of the methods employed. The IPCC 1995 Report
contained a review of four cost studies64 of the impacts of climate change
in a world where the atmospheric carbon dioxide concentration had
doubled from its pre-industrial level. The most detailed studies had been
carried out for the United States. For those impacts against which some
value of damage can be placed, estimates fell in the range of ¬fty-¬ve to
seventy-¬ve thousand million dollars per annum or between 1.0% and
1.5% of the US GDP in 1990. For other countries in the developed world,
estimates of the cost of impacts in terms of percentage of GDP were
similar. For the developing world, estimates of annual cost were typically
around ¬ve per cent of GDP (with a range of from two to nine per cent
of GDP). Aggregated over the world the estimates are between 1.5% and
2% of globally aggregated GDP (sometimes called global world product
or GWP). These studies provided the ¬rst indication of the scale of the
problem in economic terms. However, as the authors of these economic
studies explain, their estimates were crude, were based on very broad
assumptions, were mostly calculated in terms of the impact on today™s
economies rather than future ones and should not be considered as precise
More recent studies have given more consideration to the possibili-
ties for adaptation (not forgetting the cost of adaptation) through which
there is large potential to reduce the damage cost of climate change.
This is especially the case in the agricultural sector.65 In that sector,
again for doubling of atmospheric carbon dioxide concentration, stud-
ies of global aggregate economic impact vary from the slightly negative
to the moderately positive depending on underlying assumptions (see
also Impact on agriculture and food supply, pages 164ff.). But the ag-
gregate hides large regional differences. Bene¬cial effects are expected
Costing the total impacts 185

predominantly in the developed world; strongly negative effects are ex-
pected for populations that are poorly connected to regional and global
trading systems. Regions that will get drier or are already quite hot for
agriculture also will suffer, as will countries that are less well prepared
to adapt (e.g. because of lack of infrastructure, capital or education).
Overall, climate change is likely to tip agriculture production in favour
of well-to-do and well-fed regions at the expense of less well-to-do and
less well-fed regions. However, these studies have largely ignored the
increasing in¬‚uence of climate extremes and as yet inadequately con-
sidered important factors such as water availability “ largely because of
the lack of detailed information regarding these.
A further factor to which more consideration has recently been given
is the possibility of what are often called ˜singular events™ or irreversible
events of large or unknown impact. Some of these have been mentioned
earlier in this chapter or in previous chapters. Some examples are given
in Table 7.4. It is clearly extremely dif¬cult to provide quantitative es-
timates of the probability of such events. Nevertheless it is important
that they are not ignored. One recent study66 has allocated a potential
damage cost to these of about one per cent of GWP for a warming of
2.5 —¦ C and about seven per cent of GWP for a warming of 6 —¦ C. Such
calculations are necessarily based on highly speculative assumptions,
but in that particular study these singular events represent the largest
single contributor to the total overall cost.
These further factors, such as more adaptation and singular events,
have worked in both directions in the estimates of cost, some reduc-
ing and some increasing them. There has tended therefore to be a
greater spread in the overall results, thus emphasising the large un-
certainties hidden in the calculations.67 Further, little or no allowance
has been made in recent studies for the important in¬‚uence of extreme
If some allowance is made for the impact of extreme events (as
described in the last section), the studies so far suggest, in very gen-
eral terms, that the cost of damage due to climate change induced by a
doubling of the pre-industrial carbon dioxide concentration that can be
expressed in monetary terms is typically around one or two per cent of
GDP for developed countries and perhaps ¬ve per cent or more of GDP
in developing countries. However, what is also clear is that any estimates
at the moment must be considered as preliminary and uncertain, because
of inadequacies in both the assumptions that have to be made and the
available data necessary for the calculations.
An initial inspection of the sort of cost ¬gures we have been pre-
senting in the last few paragraphs might suggest that the costs of global
186 The impacts of climate change

Table 7.4 Examples of singular non-linear events and their impacts a

Singularity Causal process Impacts

Non-linear response of Changes in thermal and freshwater Consequences for marine ecosystems
thermohaline circulation forcing could result in complete and ¬sheries could be severe.
(THC) shutdown of North Atlantic THC or Complete shutdown would lead to a
regional shutdown in the Labrador and stagnant deep ocean, with reducing
Greenland Seas. In the Southern deepwater oxygen levels and
Ocean, formation of Antarctic bottom carbon uptake, affecting marine
water could shut down. Such events ecosystems. It would also represent
are simulated by models and also a major change in heat budget and
found in the paleoclimatic record. climate of northwest Europe.

Disintegration of West WAIS may be vulnerable to climate Considerable and rapid sea level rise
Antarctic Ice Sheet change because it is grounded below would widely exceed adaptive
(WAIS) sea level. Its disintegration could raise capacity for most coastal structures
global sea level by four to six metres. and ecosystems.
Signi¬cant sea level rise from this
cause is very unlikely during the
twenty-¬rst century, but a contribution
of up to three metres over the next
1000 years is considered possible.
Positive feedbacks in the Climate change could reduce the Rapid, largely uncontrollable
carbon cycle ef¬ciency of current oceanic and increases in atmospheric carbon
biospheric carbon sinks. Under some concentrations and subsequent
conditions the biosphere could climate change would increase all
become a source.b impact levels and strongly limit
Gas hydrate reservoirs also may be adaptation possibilities.
destabilised, releasing large amounts
of methane to the atmosphere.
Destabilisation of Climate change “ alone or in This could have severe social effects,
international order by combination with other environmental which, in turn, may cause several
environmental refugees pressures “ may exacerbate resource types of con¬‚ict, including scarcity
and emergence of scarcities in developing countries. disputes between countries, clashes
con¬‚icts as a result of These effects are thought to be highly between ethnic groups and civil
multiple climate change non-linear, with potential to exceed strife and insurgency, each with
impacts critical thresholds along each branch potentially serious repercussions
of the causal chain. for the security interests of the
developed world.

Table 19.6 from Smith, J. B. et al. 2001. Vulnerability to climate change and reasons for concern: a synthesis.
In McCarthy, Climate Change 2001: Impacts, Chapter 19.
See box on climate carbon cycle feedbacks in Chapter 3, page 40.
The overall impact of global warming 187

warming, although large, at a few per cent of GWP are not imposs-
ibly large, and that we might be able to buy our way out of the problems
of its impact. There are, however, two factors which have been omitted
from the studies cited above. The ¬rst is that the studies have only been
concerned with the impacts of global warming up to about the middle
of the next century when, for a business-as-usual scenario, the equi-
valent carbon dioxide concentration will have doubled from its pre-
industrial value. The longer-term impacts, if the growth of greenhouse
gases continues under business-as-usual, are likely to be much greater.68
The second factor is that impacts cannot be quanti¬ed in terms of mon-
etary cost alone. For instance, the loss of life (see question 7), human
amenity, natural amenity or the loss of species cannot be easily expressed
in money terms. This can be illustrated by focusing on those who are
likely to be particularly disadvantaged by global warming. Most of them
will be in the developing world at around the subsistence level. They
will ¬nd their land is no longer able to sustain them because it has been
lost either to sea level rise or to extended drought. They will therefore
wish to migrate and will become environmental refugees.
It has been estimated that, under a business-as-usual scenario, the to-
tal number of persons displaced by the impacts of global warming could
total of the order of 150 million by the year 2050 (or about three million
per year on average) “ about 100 million due to sea level rise and coastal
¬‚ooding and about ¬fty million due to the dislocation of agricultural pro-
duction mainly due to the incidence and location of areas of drought.69
The cost of resettling three million displaced persons per year (assuming
that is possible) has been estimated at between $US1000 and $US5000
per person, giving a total of about ten thousand million US dollars per
year.70 What the estimated cost for resettlement does not include, how-
ever (as the authors of the study themselves emphasise), is the human
cost associated with displacement. Nor does it include the social and po-
litical instabilities that ensue when substantial populations are seriously
disrupted because their means of livelihood has disappeared. The effects
of these could be very large.

The overall impact of global warming
The incidence of various impacts of global warming is complex and far
from uniform over the world. The following is a brief summary of what
we have learnt in this chapter.
r There are many ways in which our current environment is being de-
graded due to human activities; global warming will tend to exacerbate
these degradations. Sea level rise will make the situation worse for
188 The impacts of climate change

low-lying land that is subsiding because of the withdrawal of ground-
water and because the amount of sediment required to maintain the
level of the land has been reduced. The loss of soil due to overuse of
land or deforestation will be accelerated, with increasing droughts or
¬‚oods in some areas. In other places, extensive deforestation will lead
to drier climates and less sustainable agriculture.
r To respond to the impacts from the changes brought about by global
warming, it will be necessary to adapt. In many cases this will involve
changes in infrastructure, for instance new sea defences or water sup-
plies. Many of the impacts of climate change will be adverse, but even
when the impacts in the long term turn out to be bene¬cial, in the short
term the process of adaptation will mostly have a negative impact and
involve cost.
r Many of the important impacts of climate change are likely to arise
because of changes in the frequency and/or intensity of extreme events
(see Table 7.5 for a summary). For example, some parts of the world
are expected to become warmer and drier, especially in summer, with
a greater likelihood of droughts and heat waves; in other parts a greater
incidence of ¬‚oods is expected.
r Through adaptation to different crops and practices, ¬rst indications
are that the total of world food production will not be seriously affected
by climate change “ although studies have not yet taken into account
the likely occurrence of climate extremes. However, the disparity in
per capita food supplies between the developed and the developing
world will almost certainly become larger.
r Because of the likely rate of climate change, there will also be a se-
rious impact on natural ecosystems, especially at mid to high lati-
tudes. Forests especially will be affected by increased climate stress
causing substantial die-back and loss of production, associated with


. 7
( 13)