Resumo (PT):
The use of Geographic Information Systems (GIS) as an aid for identifying areas with
archaeological potential is used since the 80's in Canada and the United States by several
departments of cultural management[1].
We analyzed the physical and topographical characteristics (independent variables) of 99
archaeological sites from the Iron Age (dependent variables) taken from a universe of 198
archaeological sites in order to know their characteristics. To analyze the studied area territory
were also examined physical and topographical constraints of 120 sites that do not coincide
with any archaeological site known (called non-sites).
It was used goodness-of-fit tests (Chi-square test and
Mann-Whitney), between the 99 archaeological sites
and 120 non-sites, to exam 14 physical and
topographical variables in order to select which do not
have statistical significance, thus were excluded 6
variables[2].
With the remaining 8 physical and topographical
constraints it was produced a predictive model for
locating potential archaeological sites of the Iron Age.
For the development of the predictive model it was
used, a statistic method called binary logistic
regression[3]. This method, applied with GIS is able to
produce a map of prediction of Iron Age archaeological
areas with values ranging between 0 and 1(see Image
1).
Defining the optimal value of detection it was created
a map locating 97.5% of the archaeological sites of our studied universe and get an index gain of 0.95 (on a scale from -1 to 1)[4]. In this way the GIS is assumed as a good tool to assist the planning by helping the definition of potential archaeological areas.
References:
[1] D. Wheatley and M. Gillings, Spatial Technology and Archaeology. London: Taylor & Francis,
2002.
[2] K. M. Cunningham, "Site Suitability Analysis of Stone Circle Sites in McKenzie County, North
Dakota, at Site 32MZSWC," Resource Analysis, vol. 10, p. 11, 2008.
[3] R. E. Warren and D. L. Asch, "A Predictive Model of Archaeological Site Location in the
Eastern Prairie Peninsula," in Practical applications of GIS for archaeologists: a predictive
modeling kit, T. Francis, Ed., ed Philadelphia, 2005, pp. 6 - 36.
[4] K. Kvamme, Quantifying the Present and Predicting the Past. Colorado: U.S Department of
Interior Bureau of Land Management, 1998.
Idioma:
Português
Tipo (Avaliação Docente):
Científica