skip to primary navigationskip to content

What is Stage Three?

The Stage Three Project and its Scientific Background

Ever since the Swiss geologist Louis Agassiz took the concept of Ice Ages, long argued within the walls of learned societies, into the public arena and towards universal acceptance in the early 19th century, glacial periods, unlike interglacials, have had a bad press. They owe this to perceived extreme conditions fit for neither man nor beast that lasted a hundred thousand years or so and were only rarely relieved by brief milder intervals. This reputation survives, especially among archaeologists concerned with the Middle and Upper Palaeolithic, although recent research has shown that even during glacial intervals such extreme conditions were relatively brief exceptions.

This is regrettable, because whether or not climate and other environmental forces played a significant role during the entry of modern human beings into Europe is not a trivial question. In fact, in the context of human evolution the nature of the palaeoclimates is essential if we are to understand why the Neanderthalers became extinct during the middle part of the last glacial period, a time when the European climate deviated much from that presented by the received wisdom. 

The Stage Three Project was born at the Institute for Quaternary Research at Cambridge University. It was designed to assess from the current state of our knowledge of the mild middle part of last glacial period whether its climate might have played a role in the extinction of the Neanderthals. It was formally established in 1996 at an international and interdisciplinary conference under the auspices of the Godwin Institute; its name derive from the Quaternary strato-chronology based on the oxygen isotope record of 60,000-20,000 years ago, Oxygen Isotope Stage Three (OIS-3). 

Conceived as an exercise in Middle and Upper Palaeolithic archaeology, it was clear from the start that the Stage Three Project needed a strong paleo-environmental scientific base (van Andel 2002). This turned out to be easily achieved, because the OIS-3 climate challenged not only archaeologists, the intended ultimate beneficiaries of the Project, but also members from several other disciplines who found questions to ask that were important in their own fields. Just a few years earlier research on high-resolution Greenland ice cores had shown that the relative mildness of the Stage Three climate was deceptive. Seen in detail the interval displayed numerous abrupt alternations between very cold and quite mild events with durations of a few hundred to a few thousand years. This unstable bipolar climate also intrigued many geologists interested in defining, modelling and if possible explaining the mid-glacial climate and its impact on flora, fauna and landscape in general. 

The impact on the landscape of those many brief, sharp climatic fluctuations might have also affected the lives of Middle and early Upper Palaeolithic humans far more than each temperature extreme by itself. How might one test this possibility? 

The Objectives of the Stage Three Project

A. The palaeoclimatic and palaeoenvironmental base

  • What were the climates, environments, landscapes and faunal and floral resources of Europe and adjacent seas during OIS-3?
  • To what degree did the series of alternating sharp climate changes recorded in Greenland ice cores influence the European landscapes and environments?

B. Human beings in a changing glacial climate

  • What was the impact of climatic changes and palaeoenvironments of mid-glacial Europe on the living conditions of and resources exploited by Palaeolithic humans?
  • To what degree, if at all, does the Middle and early Upper Palaeolithic human record reflect the unstable climate of the time?

In practical terms this meant that the climates and landscapes of a typically "warm" and a typically "cold" climatic event in Europe and the adjacent North Atlantic Ocean and Mediterranean Sea had to be modelled, producing two regional simulations: a warm phase simulation representing most of the interval of 60 to 40 ka BP and a subsequent cold phase model between c. 40 and 25 ka BP resembling conditions during the Late Glacial Maximum (25-16 ka BP). For both phases studies of the response of vegetation and mammalian fauna to the temporal history of Stage Three had to be provided. 

Those were ambitious goals that, if pursued in the traditional way, would have required much time and large financial and human resources. Initially a large part of those resource would have had to be invested in the tedious, time-consuming struggle for major funding. To avoid this, and also because it seemed wise first to take stock of what was already known, the Project used existing data with only minor additions where those were critically needed, and relied in large part on the financial resources of its active members. Earlier, van Andel and Tzedakis (1996, 1998) had shown, albeit purely qualitatively, that the existing information would be adequate for reasonable environmental models. 

The power of climate and all other environmental models is a function of the spatial resolution of their output. For our task the grid-spacing of several degrees latitude and longitude of then-current global circulation models (GCMs) did not suffice. A new, nested-model strategy, employing a standard Global Circulation Model (GCM: GENESIS2) in equilibrium with global boundary conditions, provided the lateral boundaries for a regional circulation model (RCM) with a 60x60 km output grid. Besides the higher resolving power, one of the major advantages of this grid was that its higher resolution topographic base provided much more realistic precipitation patterns. The modelling of climate and vegetation was done at the Earth Systems Research Centre of Pennsylvania State University (Barron & Pollard 2002). 

High-resolution (meso-scale) palaeoclimatic and coupled climate-plant biome models were needed to construct plant cover syntheses, but the palynological data were used as input as little as possible, reserving the bulk of them for sensitivity-testing, validation and iterative improvements of the models. 

The limitation to existing data also implied that, while we were likely to formulate new, operationally better questions, better answers were likely to be a a bonus. The analysis of those better questions would most likely be attempted post-Stage Three through follow-up studies by others or some of our own members. 

During Phases 1-3 of the Project the main task was the assembling of data sets in two classes. The climate models used ice-sheet dimensions, isostatically-compensated shorelines and topography, sea-ice distribution and sea-surface temperatures as input. To test and validate outputs, we reserved micro-faunal data (Coleoptera: Coope 2002) and geological proxies such as permafrost patterns (van Huissteden et al. 2003) and aeolian deposits (van Huissteden & Pollard 2003), while coupled climate and plant-biome models were verified with pollen data from long cores (Huntley et al. 2003). 

The various tasks were executed by specialist Panels that planned modelling experiments and provided inputs evaluations to the modellers, tasks that required little interaction between various specialities and then only at a fairly general level. Nonetheless, from the first annual Phase meeting the discussions were marked by a surprisingly high level of meaningful cross-disciplinary dialogue which proved to be a good preparation for the publication phase. 

Archaeology in the Stage Three Project

The ultimate goal of the Stage Three Project was to place the spatial and temporal distribution patterns of Middle, Early Upper and Upper Palaeolithic human sites within the context of the Stage Three climate and landscape simulations to assess the level of human response to climate and resource changes. For the temporal standard of Stage Three climate changes we adopted the GISP2 Greenland ice core. 

For this final effort two sets of data were compiled. The first set was a chrono-archaeological data-base for Europe that holds all dates from archaeological cave, abri and open-air sites in the interval from 60 to 20 cal ka BP that were published up to the end of the year 2000. More than 80 percent of those dates had been obtained by 14C analysis and were therefore expressed in 14C years, but the Greenland ice-core base record is calendrical. Therefore all 14C dates had to be converted to calendar years. For this we used CalPal (Jöris & Weninger 1998, 2000), a broad-based software-engine designed to perform 14C date calibrations with any available chronological data sets. This calibration programme has been the subject of much vituperation from radiocarbon analysts but proved itself to be excellent and far superior to any other programmes so far available. 

The second data-base provided a temporal and spatial ecology and distribution of the mammalian fauna, thus supplementing extant compilations of Late Quaternary mammals which generally tend to emphasise taxonomy and/or biostratigraphy. 

The Interpretation and Publication Phase

The computer runs of Phase 4 (1999-2001) exhausted the reservoir of available data, and its simulations became the final set. This brought a need for extensive, rigorous interdisciplinary discussion about validation and interpretation of the palaeo-climate (Barron & Pollard; 2002; Pollard & Barron 2003; Alfano et al. 2003) and the plant cover (Huntley et al. 2003). It also required landscape syntheses resting on interdisciplinary publications of the results of the palaeo-environmental components of the project. At a workshop in early July 2000 the Panels concerned with the reconstructions defined the results and replaced the specialist panels with writing teams for a planned set of manuscripts that would close the climate-and-landscape phase of the Project. A draft publication programme was developed, mainly for publication in Quaternary Research, and all but a few of the envisioned papers had appeared by the end of 2003. 

In September of the same year another workshop, representing faunal and archaeological objectives and results, reviewed the reconstructions provided by the climate and vegetation workshop, and in its turn prepared a list of publications to represent the gist of the archaeological part of the Stage Three Project. 

As it turned out, however, most environmental articles tended to emphasise the interests of their own sub-discipline, were written in specific technical languages, and so could not be used directly as a base for comparison with the archaeological findings. A new set of papers of the climate, fauna and flora studies would be required focussed on the archaeological targets and in a form acceptable to archaeologists. In the end, this substantial "translation" effort to convert climatic and botanical results for archaeological interpretation, together with the archaeological results themselves, could be properly displayed only in a book on the achievements of the Stage Three Project (van Andel, T.H. & W. Davies (eds.), 2003. "Neanderthals and Modern Humans in the European Landscape of the Last Glaciation - Archaeological Results of the Stage Three Project", a monograph of the McDonald Institute for Archaeological Research at Cambridge University published in December 2003. 

Funding Acknowledged

Grants from The Leverhulme Trust and the McDonald Institute were the only sizeable external financial support directly awarded to the Stage Three Project. In addition, we benefited from two gift accounts, one of the Godwin Institute and the other in the Earth Sciences Department, that made possible the workshops, and annual meetings. We are grateful to the Leverhulme and McDonald Trusts and especially to two private donors who enabled us to use sophisticated data processing methods. 


  • Alfano, M.J., E.J. Barron, D. Pollard, B. Huntley & J. Allen, 2003. Comparison of climate model results with European vegetation and permafrost during Oxygen Isotope Stage Three. Quaternary Research 59, 97-107.
  • Barron, E.J. & D. Pollard, 2002. High-resolution climate simulations of Oxygen Isotope Stage Three in Europe. Quaternary Research 58, 296-309.
  • Coope, G.R., 2002. Changes in the thermal climate in North-western Europe during marine oxygen isotope Stage Three; estimates from fossil insect assemblages. Quaternary Research 57, 401-408.
  • Huntley, B., P.C. Tzedakis, J.-L. de Beaulieu, D. Pollard, M. Alfano & J. Allen, 2003. Comparison between Biome 3.5 models and pollen data. Quaternary Research 59, 195-212.
  • Pollard, D. & E.J. Barron, 2003. Causes of model-data discrepancies in European climate during Oxygen Isotope Stage Three with insights from the last glacial maximum. Quaternary Research 59, 108-113.
  • van Andel, T.H., 2002. Reconstructing climate and landscape of the middle part of the last glaciation in Europe - The Stage Three Project. Quaternary Research 57, 2-8.
  • van Andel, T.H. & P.C. Tzedakis, 1996. Palaeolithic landscapes of Europe and Environs, 150,000 - 25,000 years ago. Quaternary Science Reviews 15, 481-500.
  • van Andel, T.H. & P.C. Tzedakis, 1998. Priority and opportunity: reconstructing the European Middle Palaeolithic climate and landscape. In J. Bailey (ed.), Science in Archaeology: An Agenda for the Future pp. 37-46. London: English Heritage.
  • van Huissteden, K., J. Vandenberghe & D. Pollard, 2003. Palaeo-temperature reconstructions of the European permafrost zone during marine oxygen isotope Stage Three compared with climate model results. Journal of Quaternary Science 18, 453-464.
  • van Huissteden, K. & D. Pollard, 2003. Oxygen Isotope Stage Three fluvial and eolian successions in Europe compared with climate model results. Quaternary Research 59: 223-233.