University of Southampton OCS (beta), CAA 2012

Font Size: 
Zooming patterns among the scales
Alessio Palmisano

Last modified: 2011-12-18


During the last years digital technologies have been used in Archaeology for the documentation, the management and the representation of the archaeological data. A consequence of this phenomenon is the increasing popularity of Geographical Information Systems (GIS) as powerful tools for the organization and the visualization of the archaeological data in relation with the correspondent spatial information, while less attention is paid to the application of spatial statistics for detecting specific patterns of such datasets.

As a result, distribution maps are widely used by the archaeologist just as fancy and colourful accessories to use for embellish some publications. The uncritical and intuitive readings of those maps and the lack of a rigorous scientific methodology undermine the opportunity to make sense of such dataset.

Therefore, the present paper aims to offer an overview of the existing statistical approaches to settlement patterning in Archaeology by one multi-scalar method (Ripley’s K function) and to show both problems and potentiality of such technique dealing with spatial data.

I will use as case study the Iron Age I period settlements located in the actual West Bank in order to show as point pattern analyses can help us to detect spatial patterns and investigate if phenomena of attraction or repulsion are mainly related to the first or second order of effects. With the term “first order of effects” I mean that the observations throughout a study area “vary from place to place due to the changes in the underlying properties of the local environment” (O’Sullivan and Unwin, 2003, 79). The second order of effects, instead, is due to the local interaction between the observations. Therefore, in my case, we can say that the different spatial patterns (cluster, evenly or random distribution) of the settlements are either due to the environmental variables (first order of effects), or due to the direct interaction between the settlements themselves. Nevertheless, we have to point out that first/second order distinction is often difficult to detect and we have marked first order of effects when the absolute location are important determinant of observations, while we have strong  second order of effects when there is interaction between the locations, depending on the distance between them.

Therefore, I will use two different point pattern analyses for investigating if the spatial patterns of the Iron Age I settlements are due to the direct interaction among them or to the environmental variables: homogenous and inhomogeneous Ripley’ K functions. I will also investigate different kinds of homogenous Ripley’s K functions: global, bivariate and local. 


Spatial Analysis; Ripley’s K function; settlement patterns.