Fossil tests now show that computers have underestimated climate change
By John Newell,
former science editor,
BBC World Service
A global view of Surface Air Temperatures
Conclusions about global
warming and how to respond
to it are largely based
upon predictions from computer
models. Scientists at the Open
University in the United Kingdom and
elsewhere - who have used data from
fossils to check computer simulations of
temperatures in the late cretaceous period
(65 - 100 million years ago) - have
found that the models are unreliable.
Data from fossil leaves shows that the
best computer models available substantially
underestimate the warming that
can be caused in inland areas in the centres
of continents, as a result of increased
carbon dioxide (CO2) in the atmosphere
causing a "greenhouse" effect. They
imply that, for large areas of the world
at least, the predictions which some advisors
to the United States government
and others regard as exaggerated and
unfounded may in fact represent considerable
underestimates of the amount
of global warming which will be caused
by excessive CO2 emissions today.
The late cretaceous period was one
in which atmospheric CO2 concentrations
were high, as a consequence of increased
volcanic emissions. Earth appears
to have been much warmer than
at present, probably as a consequence
of a greenhouse effect caused by CO2 in
the atmosphere trapping the sun's heat
below it.
Obviously, being able to discover
what effects high CO2 concentrations
had on the climate then would be very
valuable as an aid to predicting what
effects increased CO2 emissions caused
by human activities are having, and will
have, on our climate today. The data
from leaves shows that the best computer
models of such changes are unable
to simulate temperatures in the interiors
Fossil leaf data shows that in greenhouse eras
the models predicted that the winter temperatures
in the centres of continents would be much
lower than they in fact were, as a consequence
of the greenhouse effect.
Air Pollution
of continents during previous warm periods.
Similar to tree rings, the shapes
of fossil leaves are strongly related to
the mean annual temperature when the
plant grew. But the problem has been
finding good, well-preserved assemblages
of such leaves.
An international team led by Professor
Robert Spicer, of the Open University,
with Professor Paul Valdes, of
Reading University, both in southern
England - with colleagues in Russia,
Sweden and the Czech Republic - located
such a site, in the Vilui basin in
Central Russia, at the centre of the large
Eurasian landmass.
In 1999, a major international expedition
visited the area and collected fossil
leaves and other geological evidence
of climate change. The leaves from the
Vilui basin, along with some existing
museum collections, were analysed using
the Climate Leaf Multivariate Programme
(CLAMP, for short).
The analysis yielded a picture of a
warm and wet continental interior during
the late cretaceous period. The climate
in what is now central Russia was
mild, with winter temperatures hardly
falling below freezing. The average temperature
in the coldest winter month was
about 5.1° Celsius. These results were
compared with the climate at the same
period in the same area as predicted by
the Hadley Centre Climate Model, the
most advanced climate prediction
model. This predicts that the winter temperatures
in the Vilui basin in the late
cretaceous time would have been well
below zero - about minus 20 to minus
12° C.
The same kind of disagreement is
found when the predictions from other
computer models are compared with the
results from leaf analysis. It is also
found when the same kind of comparison
is made between computer model
predictions and fossil leaf data in predicting
the climate in the interiors of all
continents during other periods when the
atmosphere was rich in CO2 - often
called greenhouse worlds.
The big differences between the predictions
from computer models and the
results from fossil leaf analysis are only
found in the interiors of continents, not
in coastal areas where models and data
are broadly in agreement.
"A strong characteristic of all
paleoclimatic (ancient climate) models
is that for greenhouse worlds such as the
mid-cretaceous they predict cold continental
interiors, conditions much more
extreme than the geological data (from
fossil leaves) suggests. This must be understood
before we can have full confidence
in future global warming predictions
for regional climates," said Professor
Spicer and the other authors of the
research report.
The models that have been used to
predict climate change and global warming
in today's global greenhouse may
have underestimated the greenhouse effect.
This likely underestimate should
be taken into account in today's international
decision-making, as well as acting
as a driving force for more research.
