It offers a solution to a very long-standing problem that is central to the study of ancient warm climates. But it is also controversial, because our result implies that a lot of data gathered using a very well-established technique, over many years, will have to be re-interpreted. Hence, as well as prompting a re-think of past climate states that may be analogous to the future greenhouse effect, it has also sparked a debate over data quality.

Far from being novel, the method is a workhorse for past climate research. It relies on measuring the oxygen isotopic composition of plankton shells from deep-sea sediments to determine the past seawater temperature. Our discovery is that by analyzing the best material available, from time intervals in excess of tens of millions of years ago, we record much warmer temperatures than previously thought.

The Late Cretaceous and Paleogene periods of geological time, between about 40 and 70 million years ago, span the time of the extinction of the dinosaurs. There is widespread evidence that it was a time of generally warm climate when there was no substantial ice at the poles and "tropical"-type organisms such as lizards, turtles, and palms extended into the polar regions. The leading theory to explain this is that there were high levels of greenhouse gases such as carbon dioxide in the atmosphere. However this explanation has had a major drawback, in that most estimates of sea-surface temperature for the ancient tropical oceans are much lower than predicted by greenhouse climate models. A typical value is under 20^oC, which is significantly cooler than the present-day ocean. This so-called "cool tropic paradox" has been used to cast doubt on the validity of climate models and their ability to predict the future greenhouse effect. If we believe the data, it also implies that we are failing to understand the most important features of past climates.

Ancient sea-surface temperatures are routinely measured by laboratory analysis of shells of fossil plankton from deep-sea sediments. However we have long suspected (as have some other specialists) that many previous temperature measurements might be in error because plankton shells from such ancient time periods are generally recrystallized on a microscopic scale. In effect, they may have had material added to them in the cold conditions of the sea floor, biasing the results. Therefore, it was desirable to find and analyze pristine shells from tropical latitudes. Such material is very rare, but our study began when we found some suitable samples from Tanzania.

As we had conjectured, analysis of these shells indicated much higher temperatures, exceeding 30^oC, which is warmer than today and closely in line with climate model predictions for the Late Cretaceous and Eocene under greenhouse conditions. Moreover, enquiries among colleagues led us to of a handful of other localities with similarly good plankton preservation (namely Alabama, Mexico, and the Adriatic Sea). Analyses of these samples produced very similar results. Taken together, they indicate that the "cool tropics paradox" is more apparent than real, having resulted from the analysis of substandard samples. It removes a question mark from the climate models and indicates that ancient warm climates are reasonably well understood and a reasonable analogue for predicted future global warming.

In 1996, a curator at London's Natural History Museum called me to say that a significant collection of Eocene microfossils had been donated to them by BP. He knew I had been working on these types of microfossils for several years, and would be interested. The samples had been collected back in the 1950s in Tanzania, during a phase of oil exploration, and some important taxonomic work had been done on them in the 1960s and 70s. I had expected, from published illustrations, that the preservation would be very good, but the moment of first looking down the microscope took my breath away. The little shells were transparent and glassy, like recently alive specimens from the modern ocean, not chalky and opaque as I had always seen in my studies of deep-sea cores. I realized immediately that chemical analysis of these pristine fossils might solve the tropical temperature problem. So I organized a series of field expeditions with colleagues from the Tanzanian Petroleum Development Corporation, retracing the steps of the British Petroleum geologists and finding new localities, and then bringing back samples for analysis. Our work continues with new funding to drill selected sites, and we are finding that the area, which is in the remote southern part of the country, is a treasure trove for ancient climate research from which many more discoveries are sure to come.

Paul N. Pearson,
Professor,
Department of Earth Sciences,
University of Cardiff,
Cardiff, Wales,
United Kingdom