Theoretical and empirical comparative evaluations on measures of map association

Assessing the association between maps is a basic task in GIS. It not only helps to discover the consistency between mapped distributions and to reveal the relationship between various spatial patterns, but also implies attribution. Although literatures are rich in the map association measures, their applications are still tricky because of the complexity in spatial data. In this study, we investigate nine representative measures of map association, where each map could be in three scales of measurement: nominal, ordinal and quantitative (interval or ratio). The strengths and limitations of individual methods are addressed and compared, which include their applicability to different measurement scales, interpretability, and potential applications in complex scenarios. The theoretical comparisons are then demonstrated by two empirical examples. These works provide a rule of thumb for the selection of map association measures for practical purposes.