ABSTRACT
This chapter brings together the work of previous chapters in a clustering context to extend explorations of spatial structure in a multivariate mode. We combine percent development (DvlpIVI) and modeled percent of extended development in¥uence (DvlmIVI) from the previous chapter with the two road densities from Chapters 2 and 3 to provide a more substantial suite of indicators, all of which are taken as speaking to some aspect of human imprint on the landscape. We use these indicators for compound clustering to elucidate similar settings. The ¢rst stage of our compound clustering serves to extract collectives as “clustered localities agglomerated nonspatially” (CLANs) based entirely on the indicator characteristics without regard to spatial situation. The second stage further segregates the CLANs into “clustered localities using map positions” (CLUMPs) of nearby localities belonging to the same CLAN cluster. We then apply patterns of precedence/ propensity as in Chapter 3 and introduce representative ranks of cluster collectives for studying similarity and comparative contrast of spatial settings.
