PREVIOUS   NEXT   CONTENTS   ISSUE   HOME 

4.3 Palaeolandscape reconstruction: pitfalls and prospects

In the introduction, I identified two main approaches to palaeolandscape reconstruction, namely the use of modern analogue environments and the application of general principles to provide first-approximation models that can then be refined. The former, while useful at the smallest scales (of the single site or multi-site area) has limited applicability at larger scales, where the accuracy and precision of such analogue models is likely to be low. What this article has argued, however, is that the latter approach might usefully take over in such cases, drawing on the principles of the Earth system and existing palaeoclimatic and palaeoenvironmental data strategically to produce more detailed and integrated reconstructions. The most obvious means to do this would be to focus on what I have called the 'local' scale, above the site but below the region. This is a good place to start because, as this article has shown:

1. We can expect to generate decent first-approximation reconstructions of the physical landscape at this scale using current terrain maps, as these variables are likely to have changed less in 6 million years than others;
2. It would be fairly straightforward to start to calibrate these physical landscape maps by integrating data from site-specific palaeolandform reconstructions and geographical or geophysical studies of coastal and tectonic systems to (a) identify areas where change has occurred in the last 6 million years and (b) identify what those changes might have been, and thus improve our maps further, and;
3. At this scale, the evidence from extant landscapes (which function according to the same principles as past ones) suggests that climatic and ecological patterns are likely to be heavily influenced by the physical landscape structure, particularly in east and south Africa. Reconstructing these variables would then be possible by combining existing data on specific climatic and ecological conditions at particular sites using principles derived from more detailed studies of the relationships between, for instance, altitude and climate in these regions today to extrapolate conditions in between them;
4. At this scale, it would be simple to keep track of site localities and contexts, and where there are areas of particular strength and weakness in our reconstructions (the geographic focus of such palaeoenvironmental reconstructions allows for them to be coupled with maps showing the density of calibration points for each variable, and hence the level of confidence with which we can reconstruct each area). This could also include tracking the degree to which each area/dataset used has been modified by human impacts and where 'extant' maps might be particularly affected by local land use. These areas might then be subjected to a higher density of sampling for direct palaeoenvironmental data and/or attempts to use older maps to mitigate this bias.

This, therefore, is the procedure I would suggest for future studies of palaeolandscape. Its use of extensive existing datasets, as well as general principles of earth science, would allow us to begin work on new reconstructions rapidly and without massive new expenditure. The existing geographic knowledge of the Earth system and the broad relationships between variables would also help where our capacity to reconstruct palaeoconditions is lacking, as, for instance, in reconstructing seasonality. In effect, this new approach would use modern environments not just as sources of direct analogy, but also as sources of principles and data for better modelling of ancient environments' spatial and temporal structures. These models would be subject to a continuous process of refinement and calibration, driven not just by the discovery and re-analysis of hominin sites, nor purely by the development of detailed understandings of extant environments, but by both together.

 PREVIOUS   NEXT   CONTENTS   ISSUE   HOME