An introduction to climate change - Climate Change: Biological and Human Aspects - page 33

with real data. What is happening can be measured today, although care is needed.
For example, satellite remote-sensing devices need proper calibration and we need
to know what their data actually mean. (I cite this example because in the 1990s
incorrect assumptions were made about some satellite data.)
The past 10 years have seen tremendous progress in understanding climate change
on two fronts. First, there has been a steady improvement in computer models. This
improvement continues but there is still a long way to go, both in terms of reducing
uncertainty across the globe and in terms of spatial and temporal resolution (see
Chapter 5). These inadequacies mean that even the best models can only present a
broad-brush picture of a possible likelihood. The 2001 IPCC report provides some
good illustrations of the limitations. For example, it presents (p. 10) two global
models of projected changes in precipitation run-off from the 1960s to 2050. Both
show decreases in run-off in the Amazon basin, much of the rim of Australia, and
Central Europe, and run-off increases in south-east Asia, north-west Canada and
southern Alaska. At the moment this does seem to tie in with some observations
(IPCC, 2001b). This similarity of computer-model output with reality (even though
we have yet to reach 2050) is not proof of the various models' accuracy but it does
lend them credibility: one corroborates the other. However, the same two models
differ in that one shows a marked decrease in run-off in the eastern USA and the
other an increase. There are also marked contrasting differences in run-off projected
for much of the Indian subcontinent and much of north-west Europe. Such problems
are not trivial even if progress is continually being made, with models becoming ever
more sophisticated and now increasingly including biological dimensions.
Second, there has been a huge growth in understanding of how the natural world
reacts to climate change, be it in terms of a single organism or an ecosystem. And
so a palaeorecord provides us with a proxy as to what might have really happened in
one place at one time under a different climatic regimen. This is, if you will, just one
pixel in a bigger picture of the Earth's climate history. But the past decade has seen
much building up of records covering many periods of time and places. Such records
can help reduce uncertainty and inform us. This is particularly useful because -
although it is difficult to build up a picture, as if pixel by pixel, from biological and
other indicators of past climates - such a picture reflects something far better than a
computer model of the Earth: the Earth itself. Indeed, if at some time the Earth was
warmer than today we can use climatic proxies (be they biological or chemical, laid
down in the geological record) to determine what the planet was like and use that
as an analogue in an attempt to predict what it would be like in a globally warmed
future.
However, with this success the past decade has also seen problems. Policy-makers
in particular have put much faith in computer models. Some bioscientists on the
other hand have found it difficult to relate their work on one or a group of biological
palaeoclimate proxies with others. Finally, communication between these modellers
and bio- and geoscientists might arguably have been better. A 2001 editorial in Nature
(v411, p. 1) summed it up thus:
To get the best return on [climate change research] funding, there must be strong
interdisciplinary collaboration between those taking data from sea sediments, ice