Memorandum by the Tyndall Centre for Climate
The Tyndall Centre is a national UK centre
for trans-disciplinary research on climate change. It is dedicated
to advancing the science of integration, to seeking, evaluating
and facilitating sustainable solutions to climate change and to
motivate society through promoting informed and effective dialogue.
The Tyndall Centre welcomes the opportunity to submit evidence
to this review and would like to be kept informed of subsequent
Emma L Tompkins, Emily Boyd, Sophie Nicholson-Cole,
Rachel Warren, Mike Hulme
Q. How are the current estimates of the scale
of climate change damage derived?
Damage functions relating some metric quantification
of a climate impact to temperature and /or temperature change,
are commonly used to compare the benefits of alternative mitigation
policies. This requires the use of some means to assess the global
damages accruing due to climate change across multiple sectors
and regions. Typically, the metric chosen is the monetised costs
of climate change. In the simplest case, a global aggregate damage
function is used to link the costs of climate impacts to temperature.
Such damage estimates are strongly affected by both the functional
forms and the parameters used to simulate impacts.
In the literature, few studies have estimated
climate impacts at a range of temperatures and thus the damage
functions are usually extrapolated from one or two benchmark estimatestypically
a "no climate change" case, and at "doubling of
CO2 concentrations on pre-industrial levels" (eg, Tol 2002b).
Pearce (1996) reviewed estimates of climate impacts on the US
economy and many functional forms are calibrated using these estimates.
Extrapolation implies that a functional form has to be assumed
from only two data points. The graph below illustrates the functional
forms of damage functions employed in the literature. The earlier
approaches (eg Manne et al 1995, Nordaus 1993a,b) use polynomials
(a type of mathematical formula) in the absolute level of temperature
change in a global aggregate approach. Dowlatabadi & Granger
Morgan (1993) use a different approach in which probabilistic
damage surfaces are estimated for market and non-market goods,
expressing GDP losses as a function of change in radiative forcing
and time. Later approaches include regional and sectoral disaggregation
(eg Agriculture, Industry) and use more complex functional forms
taking into account both the level and rate of change of temperature
(Tol 2002b). However, knowledge of impacts at the regional scale
is uneven and far too incomplete for a thorough regional analysis,
particularly in the developing world. This prevents a holistic
treatment of impacts in a rigorous fashion across regions. Thus
methods are frequently somewhat ad hoc. Another critical feature
of climate change is an expected increase in the variability of
the climate, and indeed this may be much more important that an
increase in mean temperatures and mean sea levels.
Of general concern to all these damage function
approaches are (i) the general lack of any, or an adequate, treatment
of damages due to extreme weather events, which are predicted
to increase in magnitude with climate change and are likely to
be responsible for a large component of climate change damage;
(ii) only one study (Azar & Lindgren 2003) takes any consideration
of the potential for large scale changes in the earth system such
as the melting of Greenland or West Antarctic Ice Sheets, the
breakdown of the thermohaline circulation, the collapse of the
Amazon rainforest or the release of methane clathrates; (iii)
the choice of discount rate dramatically affects the value of
damage that is incurred in 2100 as opposed to damage that occurs
in 2020, raising moral questions as to whether the quality of
life of future generations should be subject to a discount rate
at all; (iv) little or crude representation of adaptation, eg
by setting threshold on aggregate damage function of global temperature
range below which there is no modelled damage (Hope et al
1993) or with very optimistic assumptions of super-intelligent
farmer and full potential for CO2 fertilisation realised (Tol
2002b); (v) lack of treatment in damage functions of interaction
between climate damages in different sectors, eg water demand,
agriculture, and conservation of natural ecosystems; (vi) inadequate
treatment of future commitments to temperature and sea level rise
which are reached if emissions E are emitted in year Y; (vii)
in multi-sectoral approaches, how to aggregate across different
metrics, eg lives at risk/crop failure; (viii) if monetisation
is used to resolve this, which is often extremely controversial,
especially when considering how to value intangibles (ie, non-market
goods such as existence of ecosystems); (ix) if monetisation is
used, whether to use PPP or MER exchange rates to aggregate across
In summary, therefore, many different authors
have used different methods to arrive at very different estimations
for damage. The Figure above illustrates the drivers which, when
included or not in the calculations, tend to increase/decrease
the damage values arrived at. These damage values, normalised
to a single ton of carbon, are commonly referred to as the "social
cost of carbon". (this should not be confused with the costs
of removing, or preventing, that ton of carbon from entering the
atmosphere, ie, the mitigation cost). The same calculations of
damage are used in models which attempt to deduce the optimal
timing of mitigation.
In conclusion, damage valuation and cost-benefit
exercises for climate change policy are fraught with methodological
difficulties which are difficult to resolve. Hence, in many people's
opinion the damages due to climate change are best expressed in
terms of (for example): lives at risk due to hunger, flooding
of river/cost, exposure to disease, water stress, species extinctions,
ecosystem functioning problems, and large scale changes in the
earth system. The use of a database of different damages expressed
in terms of different metrics, not monetised except where uncontroversial,
and regionally specific, is more useful. Such databases are compiled
in the literature (for example, Hare 2003; there is process led
by Defra to gather an impacts database following the February
2005 Exeter conference).
Q. What are the uncertainties in these estimates?
Clearly, there are considerable uncertainties
in estimating climate damages. However, there is a more robust
understanding of the levels of climate impacts that could occur
as temperature rises, although even these are affected by the
future socio-economic development pathway, since this affects
both the stocks at risk and the adaptive capacity of the human
systems. Uncertainty increases as damage is monetised and then
aggregated across sectors, leading to very large discrepancies
(up to two orders of magnitude) between different estimates of
the social cost of carbon.
Q. In monetary terms, the impact of change
and the costs of control may be greater in rich countries than
in poor countries. But is this an adequate measure?
Those most exposed to the impacts of climate
change will experience the worst physical impacts, eg people in
low lying areas, people struggling with existing climate variability,
people with limited capacity to adapt to new conditions. However,
preparedness levels will determine how sensitive the different
groups are to that exposure. A group with high levels of preparedness
will experience much lower impacts and hence lower costs of impacts.
Preparedness refers to both minimising the causes of climate change
(mitigating), as well as being able to cope with the consequences
Wealth (or lack of it) is not the best indicator
of who will cope best with the impacts of climate change; a better
indicator is the level of preparedness. For example, recent research
on the impacts of hurricanes in the Caribbean reveals that two
countries with excellent preparedness, the Cayman Islands and
Cuba, which have very different levels of income, managed the
2004 hurricane season much more successfully, and with less impact
that the middle income, but less prepared countries of Grenada
Preparedness may cost more in a rich country
than in a poor country due to higher relative prices, eg relative
price of sand bags. However, as climate change is a global trans-boundary
problem, which requires international cooperation, both in terms
of technology transfer and resource sharing to solve it, the issue
of relative cost has to be considered as just one of many factors
influencing the decision-making process.
Other criteria that need to be considered alongside
economic efficiency when making decisions about managing climate
change include: the legitimacy of the decision making process
to reach internationally agreed upon targets and strategies, the
equitable distribution of costs and benefits in managing the problem,
and most importantly the effectiveness of the different strategies
to achieve the desired outcome.
Q. What would be the relative costs and benefits
of using resources, otherwise expected to be allocated to climate
change controls, instead to expand international development assistance?
As mentioned above, using conventional economic
efficiency arguments in isolation to make decisions about resource
allocation for climate change is not necessarily appropriate (even
if the quantification of damage costs can be madesee above).
In addition, the argument that climate change
control activities have to be traded-off against international
development expenditure is unrealistic. For example, the UK Department
of Trade and Industry could subsidise British industry to invest
in mitigation technologiesthis is not money that would
have been allocated to international development initiatives.
In addition, the UK can invest in "climate-proof" international
development, for example, supporting renewable energy projects
for small islands; this would provide energy security and mitigate
Bjorn Lomborg's Copenhagen Consensus suggested
that resources should be invested in international development
and not climate change mitigation, but he didn't take into account
that climate change will exacerbate many existing issues of poverty,
particularly in natural resource dependent communities, low-lying
communities, coastal communities, and any others vulnerable to
changing climate variability, including more intense extreme events
such as wind storms. The climate change issue has the potential
to be an additional stressor that could exacerbate vulnerability
and poverty and as it progresses, it is likely to increase the
economic gradient between north and south.
Q. When are damages likely to occur and how
satisfactory is the economic approach to dealing with costs and
benefits that are distant in time?
If anthropogenic climate change is occurring
nowwhich it is with a high likelihood; cf. IPCC Third Assessment
Reportthen it follows that some of the damage being caused
by extreme weather events now is already attributable to anthropogenic
climate change. The difficulty is knowing "how much",
"where" and "when". It remains scientifically
problematic to attribute any one extreme weather event to anthropogenic
climate change, although new methods are being developed which
exploit the idea of "fractional attributable risk" (cf.
its use in the arguments over tobacco and human health). For example,
the 2003 summer heatwave in Europe caused considerable damage,
both in terms of human health (up to 20,000 premature deaths)
and economic losses (estimates of
15,000,000,000). Although statistically the event
was extremely raresome estimates suggest a 1-in-40,000
year eventthere is evidence from climate models that anthropogenic
climate change is radically reducing the odds of such an event
occurring, and recurring (Stott et al, 2004).
It is certainly possible to make estimatesbased
on modelson how likely such events will be to occur in
the future, and it may soon become an established methodology
to be able to make such statements as "The fractional attributable
risk due to anthropogenic climate change for a certain extreme
weather event, and associated damage, was 0.9ie, 90 per
Economic analysis may be useful to assess in
monetary terms the costs and benefits of balancing adaptation
actions and mitigation activities, however, there is as yet not
an agreed upon method by which this can be undertaken. It is widely
argued that two critical issues make the use of conventional cost-benefit
analysis inappropriate for use as a decision support tool when
dealing with climate change; these are the uncertainties and the
time frame. First, the long time frame over which climate change
impacts will occur raises the problem of selecting an appropriate
discount rate by which to discount future impacts and future adaptations.
Second, the uncertainties associated both with the timing and
type of climate change that will appear, as well as the adaptation
strategies that may be adopted to cope with it, make estimating
costs and benefits too uncertain (see answer above).
Subjective decision-making is important when
managing extremely uncertain events with big impacts. For example,
in 1953 after the East of England storms the Thames Barrier was
constructed. This decision was not based on a cost-benefit analysis
but on a subjective decision (strongly influenced by politics)
to avoid another occurrence of the number of deaths following
a large storm. The Barrier has turned out to be crucial to protecting
London, yet the number of uses of the Barrier probably could not
have been estimated in 1953, and hence the benefits of the Barrier
could not have been estimated. It is unrealistic to think that
pure cost-benefit analysis will guide all future adaptation decisions.
Q. What other associated benefits might there
be from reducing greenhouse gas emissions?
Undertaking mitigation activity now, ie reducing
greenhouse gases will reduce the need to adapt in the future,
as it is assumed that the impacts will not be so severe (so long
as greenhouse gases can be stabilised at an appropriate level).
Investing in mitigation technology in the present time reduces
the uncertainty associated with future adaptation costs. We also
know that there are limits to adaptation, which we will approach
as we do not mitigate. For example, if the sea level rises by
a meter there is little that can be done (given our current technology
and resources) to hold the sea back from the entire coastline
of the UK.
The UK could become an economic leader in exporting
Mitigation now would reduce the likelihood of
low probability but very high impact events resulting from rapid
climate change. Rapid climate change could appear as:
A collapse of the Thermohaline Circulation
(which could lead to mass migration south).
Rapid sea level rise (which could
lead to mass migration inland)
Accelerated climate change through
feedback mechanisms from melting permafrost (which could lead
to mass migration north).
Clearly the consequences of mass migration would
be severe and best avoided.
1 March 2005