Memorandum submitted by the Geological
Society of London
The Geological Society of London, founded in
1807, is the oldest geological society in the world. It has 8,800
members world-wide, and is both a learned society and a professional
This submission was prepared by Professor RCL
Wilson (Honorary Secretary, Foreign and External Affairs), with
contributions from Dr Stephen Blake (Open University), Professor
Geoffrey Boulton FRS (University of Edinburgh), Lord Oxburgh FRS
(President elect, and Imperial College), Professor Sir Nicholas
Shackleton FRS (University of Cambridge), and Professor Bob Spicer
(Open University). These colleagues and members of the Society's
External Relations Board have commented on several drafts. The
submission has been approved by the Board.
Changes in atmospheric CO2 and temperature
over the last 400,000 years as recorded in the Antarctic Vostok
ice core (lower left box), and atmospheric CO2 over
the last 1,000 years. The caption for this figure is on page 1.
The figure shows data obtained from cores obtained
by drilling through the Antarctic ice sheet at the Russian Vostok
research station1. Variations in temperature above the present
day mean of -55.5ºC are shown, based on an analyses of the
hydrogen isotopes present in the ice. The plot of atmospheric
CO2 content was obtained by analysing bubbles of air
trapped in the ice. The small box in the upper right diagram shows
the rapid increase in CO2 caused by the industrial
revolution (note that this part of the diagram shows a time span
of 1,000 years, whereas the main diagram spans 400,000 years).
The left hand plot in the lower left box shows
changes in the amount of solar radiation received at the Earth's
surface (insolation) in mid-June at 65º North. Mid-June insolation
in northern high latitudes is thought to be the key to the origin
of glacial periods. If summers in these areas are cool enough,
winter snows will survive, leading to successive annual accumulations
of snow and ice. Variations in high northern latitude summer insolation
are caused by variations in the shape of the Earth's orbit (eccentricity),
the angle of its rotational axis (obliquity) and precessional
changes. The increase in insolation that commenced about 20,000
years ago triggered the end of the last glaciation. Insolation
in northern high latitudes has now almost returned to the level
it was 20,000 years ago, but as yet the climate system has not
responded to this change.
The underlying theme of this submission is that
an essential ingredient of the "critical appraisal"
referred to in the second aim of the case study is to learn from
the past to illuminate the future. The Society has four areas
of concern which it hopes the Committee will investigate; these
are summarised below:
I. The extent to which knowledge and understanding
of past changes in climate and atmospheric composition have contributed
to the development of scenarios of future change considered by
Key aspects of such knowledge and understanding
are that geological data:
provides the only information
we have about how the Earth's climate system has behaved over
meaningful lengths of time (ie hundreds of thousands of years,
in contrast to reliable historical records which only extend back
about 150 years);
provides a record of changes
in atmospheric composition before anthropogenic influences took
shows that the Earth's climate
system may flip rapidly (in decades) from a colder to a warmer
shows that sudden releases of
carbon into the atmosphere (comparable in size to that being caused
by the burning of fossil fuels and other anthropogenic causes)
are only removed slowly over tens of thousands of years.
II. The extent to which Government ensures
that the climate modelling studies used to predict global and
regional climate changes are evaluated by testing their ability
to replicate past climate changes that occurred in both greenhouse
and icehouse worlds.
The Society's key concerns are:
that climate models have inherent
limitations due to their ancestry in models designed to predict
the behaviour of today's weather systems, and because they have
to simplify an exceedingly complex climate system in which processes
operating in the atmosphere, oceans, ice sheets, and solid Earth
are linked together;
the models are unable to fully
replicate past climates during periods when the Earth was much
colder, and much warmer, than it is today. The flaws are most
marked in reproducing conditions in continental interiors during
greenhouse climates that characterised much of the Earth's history
over the last 500 million years;
therefore continued comparison
of model simulations with geological data about past climates
are required to inform the design of improvements necessary to
provide more reliable predictions about future global and regional
climate changes forced by rising atmospheric CO2 levels.
III. The extent to which Government has addressed
the implications of rapid short term changes in weather and climate
caused by large volcanic eruptions.
Historical and geological records show that large
eruptions can cause significant cooling at regional and global
scales resulting in "volcanic winters" that could have
severe agricultural, economic, and political implications.
IV. The extent to which Government seeks
the views of a broad range of scientists in addition to the advice
given by its appointed advisers and consultants.
1. It is natural that the institutions directly
concerned with current weather and climate should be consulted
over climate change. However, reliable climate records go back
little more than 150 yearsa mere blinking of the eye in
the hundreds of millions of years of Earth history. Humanity's
activities may well be triggering changes in the present climate,
but to understand the extent to which similarly large changes
may have occurred in the past without anthropogenic intervention,
it is necessary to look at past climate changes as recorded in
the geological evidenceespecially in the sediments of the
deep oceans and in ancient ice sheets. It is not clear that the
Government has sufficiently taken advantage of the work of geological
specialists in these areas. In the pages which follow we elaborate
the contribution that studies of this kind, combined with others,
may make to understanding changes in the Earth's climate.
2. This submission focuses on the second
aim of the case study, namely to identify "what critical
appraisal there has been of models predicting climate change,
increasing concentrations of carbon dioxide, and other potential
drivers". The underlying theme is that an essential ingredient
of such "critical appraisal" is to learn from the past
in order to illuminate the future. To do this scientists use their
understanding of the Earth's climate system to interpret the information
they collect from fossil climate indicators (often referred to
as proxy climate data).
3. Historical records indicate the extent
of climate change over the last few centuries, but a pattern of
repeated climate change is only revealed by examining the geological
record. The mechanism by which the climate system may change can
be demonstrated by correlating observed effects preserved in the
geological record with postulated causes. Usually this only leads
to a qualitative understanding, which does not yield many insights
concerning changes in the interconnectivity between different
components of the climate system. Palaeoclimatologists obtain
a deeper quantitative understanding by using computer generated
climate models both to aid their interpretation of proxy climate
change data, and to explore the possible effects of changes in
different climatic forcing agents. Increasingly geological techniques
permit a quantitative record of climatically determined parameters
to be reconstructed through time. Although this does not directly
indicate the causes of change, correlations between measures which
reflect different climatic variables can suggest causality. Palaeoclimatologists
can, by using computer generated climate models, attempt to simulate
this geologically inferred behaviour and thereby test their theories
of the causes of climate change. This use of models is often overlooked
in debates about possible future climate change.
4. The only way to predict how the climate
would change in the absence of anthropogenic effects is to understand
the nature of past climate variability over decadal to millennial
time scales. Likewise, the evaluation of climate models used to
predict climate change can most convincingly be made by determining
the extent to which they can fully explain what happened in the
past (ie be used to "retrodict", rather than predict,
5. The first aim of the case study is to
identify "the extent to which the Government has been advised
of potential alternative explanations, how these alternatives
have been assessed, and what conclusions have been drawn".
The Society is not in a position to make such judgements without
being able to question directly ministers, their civil servants,
and advisers, directly. However, the content of the recent Sir
Peter Kent Lecture given to the Society by Michael Meacher2 does
indicate that the Government is well aware of many of the conclusions
that can be drawn from evidence of climate change contained in
the geological record. Nonetheless, we suggest that the committee
investigates the extent to which Government:
ensures that models used to predict
possible future global and regional climate changes are evaluated
by testing their ability to replicate past climatic conditions
based on palaeoclimatic indicators;
seeks the views of a broad range
of scientists in addition to those who provide information and
advice as appointed advisers and consultants.
6. Our comments on the second aim of the
Committee's inquiry are given below under the following headings:
Natural climate variability
How long might it take for atmospheric
CO2 levels to return to normal levels?
Climate modelling and the geological
7. Air bubbles within Antarctic ice enable
changes in the CO2 content of the atmosphere to be
determined for the last 400,000 years (see cover figure). During
the four warm interglacial periods similar to those the Earth
has experienced for the last 10,000 years, atmospheric CO2
content reached 270-280 ppm, and during the coldest parts of glacial
periods, this dropped to 190-200 ppm. Changes in atmospheric CO2
are closely correlated to changes in temperature.
8. What is the link between atmospheric
CO2 content and temperature change? Does increasing
temperature cause carbon dioxide to be released from the ocean,
thereby increasing atmospheric concentration, or does increasing
carbon dioxide cause temperature increase because of the greenhouse
effect? We know that both processes occur. If the greenhouse effect
(which is due not only to the presence of CO2 in the
atmosphere, but also water vapour and methane) did not occur on
Earth, its surface temperature would be 30ºC colder, and
life as we know it would not exist. Equally, we know that warm
tropical oceans exhale carbon dioxide, whilst cold polar oceans
inhale carbon dioxide.
9. It is probable that the close correlations
shown in the cover figure reflect close coupling of temperature
and carbon dioxide, ultimately forced by changes in the amount
of solar radiation reaching the Earth's surface (see second part
of figure caption). If this increases and causes a rise in the
Earth's surface temperature, there is net exhalation of carbon
dioxide from the oceans, atmospheric CO2 increases
and further greenhouse warming occurs. This in turn increases
carbon dioxide exhalation, and so on, in a positive feedback loop.
The reverse process follows reductions in radiation received at
the surface. The question for the past is how do positive feedback
loops involving carbon dioxide and temperature come to be stabilised
at the glacial and interglacial levels shown in the cover figure.
We do not know the answer. The question for the future is what
sort of feedback loops might result from anthropogenically driven
increases in atmospheric CO2. Once again we do not
know the answer.
10. Geological and historical records show
that climate change has occurred over a range of time scales,
from the very slow (over millions of years) to very fast (over
a few decades). The cover figure shows that over the last 400,000
years, climate has varied between cold glacial periods, and warm
interglacial interludes such as the one we enjoy today. These
warmer conditions only persisted for about 10 per cent of this
period of time.
11. During the last glacial period (between
10,000 and about 110,000 years ago) there were 24 warmer interludes3:
each cool-warm cycle lasted for about 1,500 years. The asymmetry
of changes of temperature between the cooler and warmer interludes
mirrors that which occurred over the four glacial cycles shown
in the cover figure: slower stepwise cooling followed by rapid
warming. This suggests that the climate system may flip rapidly
from one state to another.
12. During the present and last interglacial
periods, global temperatures were relatively stable compared to
glacial periods. There is, however, some evidence to suggest that
during the current interglacial there were regular changes in
climate in the North Atlantic area with a periodicity of about
1,500 years4,5. These were pale shadows of much greater changes
that occurred during the last glacial period. One manifestation
of these changes occurs as peaks in the delivery of ice rafted
debris into the North Atlanticduring colder parts of the
1,500 year climatic cycle icebergs almost reached Ireland. These
temperature fluctuations were sufficiently large (about 2ºC)
that, if they occurred in the future, they would have significant
effects on agriculture. The "Little Ice Age" (which
lasted from late medieval times until the 19th century) was the
last of these cooler episodes. Warming since it ended is likely,
therefore, to be due to a combination of "natural variability"
and the anthropogenically driven portion of the greenhouse effect.
ATMOSPHERIC CO2 LEVELS
13. Since the beginnings of industrialisation
at the end of the eighteenth century, the anthropogenic source
of carbon dioxide has released approximately 310 gigatonnes of
carbon into the atmosphere: 260 GtC (1GtC = one thousand million
tonnes of carbon) from fossil fuel and cement production, and
50 GtC from terrestrial ecosystems, mainly due to changes in land
use practices. It has been estimated that whereas 140 GtC of this
anthropogenic output has been absorbed by the oceans, 170 GtC
has accumulated in the atmosphere, with the consequence that the
atmospheric concentration has reached a level unprecedented for
an interglacial period (ie 360 ppm instead of the normal proportion
of 280 ppm). It is increasing at a rate of 0.4 per cent per annum.
Although we expect carbon dioxide and temperature to be coupled
in a feedback loop, as seems to be the case with respect to past
natural variations, we do not understand how and at what level
a future equilibrium will be reached. However, the geological
record does indicate how long it might take for atmospheric CO2
levels to return to their normal level if anthropogenic releases
of CO2 were to cease immediately.
14. Current estimates suggest that in less
than a thousand years, human activity will add 2,000-4,000 GtC
to the atmosphere. Until last year, the only means of evaluating
the possible climatic consequences of this huge release of carbon
was to use modelling studies. However, the recent discovery that
a huge blast of carbon into the atmosphere occurred 55 million
years ago gives new insights not only into the nature of the global
carbon cycle, but also how long it may take for the Earth's system
to return to equilibrium after such a blast.
15. The new work6 has shown that for a brief
period 55 million years ago, temperatures in the deep oceans and
at high latitudes rapidly increased by 5-7ºC. Although the
cause of this warming is not clear, geochemical evidence from
deep-sea sediments indicates that over 1,000 GtC was released
into the ocean and atmosphere. This was probably due to the release
of methane from gas hydrates (a frozen form of methane plus water
that is stable in low temperature and high pressure conditions)
present in sediments along outer continental shelves and continental
slopes. This injection of carbon occurred over less than 10,000
years, but it took 140,000 years for it to be absorbed. This suggests
that if the global carbon cycle operated in a manner similar to
the way it does today, we will have to wait an exceedingly long
time for atmospheric CO2 levels to return to their
normal interglacial levels. It should be borne in mind, however,
that the Earth was much warmer 55 million years ago, so it is
likely that, because the climate system operated in a different
mode, rates of carbon sequestration may not have been the same
as those of today.
16. All sophisticated climate models currently
in use are based on weather forecasting models and by definition
are designed and tuned to reliably simulate the present atmospheric
processes and conditions. Moreover, they do not attempt to simulate
all aspects of climate system processes, but employ a range of
"short cuts" designed to give realistic results for
our present world. Simulation and sensitivity modelling aids our
understanding the causes of past climate change, and the possible
variations that await us in the future, but the nature of the
models has to be borne in mind when interpreting results of experiments
designed to simulate globally warmer conditions.
17. Despite the development of ever more
sophisticated climate computer models, the problem that still
has to be faced is that apart from at the last glacial maximum
18,000 years ago, there is not enough high resolution palaeoclimate
data available from enough locations around the world on which
to base computer driven simulations of past global climate systems.
For this reason, many climate modelling studies use simplified
algorithms rather than trying to reproduce every detail of a past
global climatic episode. In other words, models are used to explore
the sensitivity of the climate system to changes in specific forcing
factors, such as changes in the proportion of greenhouse gases
in the atmosphere, or the distribution of land and sea around
18. The more ambitious simulation models
require that quantitative definitions of various components of
the climate system are fed into the programme, such as incoming
solar radiation, reflectivity (albedo) of land, sea and clouds,
the extent of ice sheets, deep and surface oceanic temperatures,
and the CO2 and water vapour content of the atmosphere
etc. These general circulation models (GCMs) can only be run on
very large supercomputers, and until recently, most of them could
only treat the oceans as a "wet carpet", and did not
simulate shallow and deep circulator patterns. Producing models
of regional and global oceanic circulation is also a complex task
but significant progress is being made. The reliability of the
predictive output of simulation models can be assessed by comparing
their outputs to present day temperature and wind patterns. However,
geologic evidence shows the present world is abnormally cold in
comparison to what it has been for most of the past 500 million
years. If the world is indeed warming, therefore, then it is vital
that the models can operate reliably for such greenhouse conditions.
When existing models are used to produce simulations of past climates,
say, during the Late Cretaceous (99 to 65 million years ago),
or for the Eocene (55 to 34 million years ago) when global mean
temperatures were 10ºC higher than today, they do not perform
well. In particular continental interior climates are predicted
to have been markedly colder and drier compared to what a plethora
of geological evidence shows to have been the case7,8. These errors
are similar, irrespective of which model is used, or which assumptions
are made about the atmospheric or geographical conditions during
those times. The models are demonstrably more reliable for cool
world scenarios than for those for warm worlds. This may reflect
the modern day origins of the models, and the intensive comparisons
with data obtained from glacial periods.
19. The implications for successfully modelling
future possibly warming point to the need to take heed of geological
data. Once we have models that satisfactorily simulate present
and past observations of the behaviour of the atmosphere, oceans
and ice sheets, the next step is to couple them together to refine
simulations of possible past climatic conditions, and to predict
more confidently future climatic changes. Such advances will require
not only more sophisticated simulation models, but a greatly improved
knowledge of past global climate change obtained from the geological
record. Crucial to this appraisal of model predictions (and retrodictions)
is to compare their results with extensive geological observations.
The latter can only be obtained by co-ordinated international
efforts in order to make the increasingly critical tests of model
predictions that Government and the international community require.
20. Historical records reveal that anomalous
weather and climatic effects (volcanic winters9) have occurred
for several years following a number of large volcanic eruptions.
Such effects involve not just the environs of the volcano, but
can have a significant impact on the regional, hemispheric or
21. The connection between lowered temperature
and volcanic eruptions is due to the effects of sulphur dioxide
gas released during explosive eruptions. In sufficiently powerful
eruptions, volcanic rocks and gas are lofted into the upper atmosphere
where the volcanogenic sulphur dioxide gas reacts with water to
form micron-sized droplets of sulphuric acid gas. These aerosol
droplets have the principal effect of reflecting incoming solar
radiation. Consequently, the temperature of the lower atmosphere
is reduced over the period taken for the aerosol to be chemically
and physically destroyed.
22. For example, a powerful eruption of
the Indonesian volcano Tambora in 1815 was followed by one to
two years of unusually cold climate around the world. In Europe,
annual mean temperatures were 1 to 2.5ºC lower than normal,
harvests were either late or failed altogether, grain prices were
at their highest and famine was widespread. During the summer
of 1816, the USA experienced a widespread snowfall. The annual
global mean surface temperature was about 1ºC below normal.
The years 1816, 1817 and 1818 had some of the coldest Northern
Hemisphere summers on record10-12.
23. Two further examples are the 1783 eruptions
of Laki in Iceland that led to a "dry fog" permeating
much of western Europe and large parts of north America, where
average 1783/4 winter temperatures were up to 4.8ºC below
the long term average13, and the 1991 eruption of Pinatubo (Philippines)
with a 0.4ºC fall in mean global surface temperature in 199214.
Eruptions that have a short term influence on climate occur irregularly
but by no means infrequently. At least six such eruptions occurred
in the 20th century.
24. The Tambora eruption in Indonesia that
occurred about 70,000 years ago was one of the largest to have
occurred over the past few hundred thousand years15. It occurred
during a period of rapid cooling from interglacial to glacial
conditions. Ice core records indicate that huge amounts of volcanic
dust and aerosols that it injected into the upper atmosphere may
have remained there in significant amounts for up to six years.
This may have caused global cooling between 3 and 5ºC, and
perhaps as much as 10ºC during growing seasons in middle
to high latitudes.
25. Given the climatic impact of volcanic
eruptions described above, and the likelihood of future eruptions,
the Society recommends that the Committee investigate the extent
to which the Government has explored the agricultural, economic
and political consequences of volcanic winters.
26. The geological record provides relatively
high resolution information about climate changes that have occurred
over the last few hundred thousand years. This shows that there
is a link between global temperatures and the greenhouse gas content
of the atmosphere, but, as yet, the extent to which the latter
is the cause of the former is uncertain. Palaeoclimatic records
also show that global average temperatures have fluctuated by
as much as several degrees in a matter of decades. Although climate
models have yet to successfully simulate changes in the atmosphere-ocean-ice
sheet system that result in such rapid shifts, our understanding
of the system is sufficient to argue strongly for the precautionary
principle to be applied in favour of reducing anthropogenically
induced greenhouse gas emissions. Opponents of this approach argue
that we should postpone taking effective steps to reduce these
until such time as research can convincingly demonstrate the climatic
impacts of emissions, and distinguish natural climatic changes
from human induced ones. This implies a rather naïve faith
in science to deliver certitudes. We will only obtain "proof"
as the global experiment we are conducting unfolds. If this experiment
triggers a rapid reorganisation of the climate system, proof might
come too late for preventative action. In the meantime, continued
investment in research into past climates will be crucial to providing
rigorous tests of the likely effects of anthropogenic intervention
into the workings of the Earth's climate system.
1. Petit, JR, et al, Climate and
atmospheric history of the past 420,000 years from the Vostok
ice core, Antarctica. Nature, 399, 429-436 (1999).
2. Meacher, M, Ice age or global warmingwhich
is it? Geoscientist 10/1, 16-17, (2000).
3. Bond, G, et al, Correlation between
climate records from North Atlantic sediments and Greenland Ice,
Nature, 342, 637-42 (1993).
4. Bond, G, et al, A pervasive millennial-scale
cycle in North Atlantic Holocene and glacial climates. Science,
278, 1257-1265 (1997).
5. Campbell, ID, et al, Late Holocene
1,500 yr periodicities and their implications. Geology, 26, 471-473
6. Dickens, GR, The blast from the past.
Nature, 401, 742755 (1999).
7. Greenwood, RR, and Wing, SL, Eocene continental
climates and latitudinal temperature gradients. Geology, 23, 1044-1047
8. Valdes, PJ, Spicer, RA, Sellwood, BW
and Palmer, DC, Understanding Past Climates: Modelling Ancient
Weather. CD ROM, Gordon and Breach, Reading, UK (1999).
9. Rampino, MR and Self, S, Volcanism and
biotic extinctions. In: Encyclopedia of volcanoes (Ed: H Sigurdsson),
Academic Press, San Diego, 1083-1091 (2000).
10. Lamb, HH, Climate, History and the Modern
World, 2nd Edition, Routledge (1995).
11. Stothers, RB, The great Tambora eruption
in 1815 and its aftermath. Science, 224: 1191-1198 (1984).
12. Briffa, K, et al, Influence of
volcanic eruptions on Northern Hemisphere summer temperature over
the past 600 years. Nature, 393, 450-455 (1998).
13. Sigurdsson, H, Volcanic pollution and
climate: The 1783 Laki eruption. Eos, Transactions of the American
Geophysical Union, 63(32) 601-602 (1982).
14. McCormick, MP, et al, Atmospheric
effects of the Mt Pinatubo eruption. Nature, 373: 399-404 (1995).