We currently make climate change predictions using mathematical models developed using climate records of the last two centuries. The trouble is, the global climate was fairly static during this period, so it is difficult to know how accurate these models are in predicting climates very different from the current state.

The Earth has experienced large shifts in climate in the past and if a snapshot of the Earth surface at one of these intervals was available, we could use it to test the reliability of our mathematical models.

During the last 150,000 years, the Mediterranean experienced several episodes of oceanographic conditions markedly different from today. So it can serve as an ideal testing ground for models.

By looking at abundances of species of planktonic foraminifera recovered from marine sediments, Dr. Kucera is developing a new tool to reconstruct sea-surfaces temperatures in the past. This technique is based on an artificial intelligence algorithm learning the relationship between temperature and abundances of planktonic foraminifer species in the modern ocean.

The research, funded by the Leverhulme Trust, will generate high-resolution maps of Mediterranean sea-surface temperature for several time-slices in the past in order to provide targets for testing of numerical models of ocean circulation.

Dr. Kucera said: "A coherent set of reconstructions is urgently needed in order to test ocean circulation models. This would also provide an important step towards an integrated effort to reconstruct past climate variability with the high precision and reliability that is needed to support the long-range strategic management of our society and its resources."

Dr. Kucera, Department of Geology, Royal Holloway, University of London, will be collaborating with Dr. Eelco Rohling, School of Ocean and Earth Science, University of Southampton and Professor Björn Malmgren, Department of Earth Sciences, Göteborg University, Sweden.

"Reconstructing Mediterranean paleoenvironment from fossil faunal counts" is funded by a grant of £71,994 from the Leverhulme Trust.

Until recently, the main comprehensive modeling target available was the CLIMAP [1976] global compilation -- revolutionary at its time, but now largely outdated, its methods superseded by more complex, computer-intensive techniques.

In addition, several geochemical paleothermometers have since been developed and paleotemperature reconstructions based on faunal counts play a crucial role in deconvolving the effects of various chemical and biological processes on these geochemical proxies.

10-Jun-2002