A Maya predictive model
A study on the use of Geographic Information Systems in a multi-scale archaeological project
Previous page Table of content Next page

I. Introduction

Located in Central America, Guatemala and neighboring Belize are countries which are not necessarily well known. Yet, that's where a civilization which has left incredible archaeological remains has developed over several millennia. A people which is passionating by the grandeur of its buildings, its magnificent temples which can still be seen beyond the canopy, and the many houses covered with hills which dot the jungle of these countries. While Belize hosted only 277 409 inhabitants for 22 966 km2 in 2005 (Government of Belize 2006; SIB 2007)⁠, i.e a density of 12 inhabitants / km2, it is striking to see by walking through the forest how much the ancient Maya seemed to be very present.

The densities given by some archaeologists confirm this fact, 280 inhabitants / km2 in the region of the Rio Bec (Fischbeck 2004) ⁠ for example, or 407 to 506 inhabitants / km2 in rural Copal Pocket (Webster & Freight 1990, cited in Wheeler 2008) ⁠. How could such a population support itself while the Maya did not know iron, did not use wheels? What are the secrets which have allowed this civilization to establish its splendor in the Yucatan and to survive for over 1 700 years?

These are the questions that archaeologists are trying to respond to, and Geographic Information Systems play an increasingly important role in their research since the first uses in the 1980s (Aldenderfer & Maschner 1996; Wheatley & Gillings 2002). Many publications have been done on the subject, particularly about the possibility of modeling and the efficiency of predictive models (Mehrer 2006; Dermody 2005; Leuseni 2002), to try to define the rules of proper use of these tools by reconciling knowledge and practical techniques associated with respect for the archaeological aspect.

This dissertation, as part of a program to study Maya environment, is not intended to be yet another theoretical analysis of the relationship between GIS and archeology, but to intimately link the implementation of a predictive project with the identification and solving of difficulties. How can the work of a geomatics engineer with no initial knowledge in archeology fit with the needs of a multidisciplinary team which does not have any GIS administrator? What are the steps to follow for such a project, from the organization of available data within a system usable by everybody to the analysis of results, without forgetting the consultation with the other members of the team? Some ideas to correct biased predictive models are also clarified, highlighting the complexity of choices to do and the wrong conclusions that they may lead to.

This study also highlights the problems of scale when setting up a predictive model: what are the criteria for the selection of data and how to implement the method of the Weights of Evidence chosen here? How can the choice of mask size or the type of sites affect the results? Is it possible to generalize the conclusions obtained at a local level to the whole Maya world system?

Finally, the question of the implementation of predictive models and their analysis is presented. How to produce from these new models information on the studied population, on its livelihoods and its land use? Is it possible to build a new vision of how a people could live by using the tools of geomatics, and to address the issue of current policies in terms of agriculture by looking at the lessons of the past applied to the challenges of the future?

Despite the specificity of the studied program, this dissertation is designed to be applied to any archaeological project interested in the implementation of a population model.

css Copyright © 2018 Sébastien Merlet (Sebeto) xhtml