A fundamental characteristic of spatial processes is the geographic scale at which they occur. Despite this inherent link between geographic process and scale, there is a surprising lack of methods for explicitly measuring scale in spatial processes. One exception is the use of spatially local multivariate statistical models by researchers from both the social sciences and physical sciences. These models produce a surface of process estimates that can be visually analyzed for heterogeneity and also produce a metric that may be considered an indicator of scale. However, the nature of these metrics has not been rigorously examined and there is an urgent need to better understand if they truly represent scale. Furthermore, several extensions to local models have recently been proposed to incorporate multiple scales of analysis, which generate significantly more information on the nature of geographic processes. Unfortunately, these multi-scale extensions are in their infancy and several issues need to be addressed before they can reach their full potential. Therefore, the goal of this research is to advance the concept of scale by developing and comparing three multi-scale local modeling frameworks. This goal will be achieved by completing three objectives. First, multi-scale extensions will be developed for three distinct local modeling frameworks. These frameworks will be compared in terms of their inferential and predictive capabilities in order to assess the relative advantages and disadvantages of each. Second, a standard definition of scale will be derived by assessing the meaning of indicators of scale from local models. The indicators will be studied in terms of their generalizability, interpretability, and robustness. Finally, a comprehensive open source software suite will be developed and disseminated in order to make multi-scale local statistical analysis available to researchers and practitioners.
|Effective start/end date||4/1/18 → 9/30/21|
- National Science Foundation (NSF): $399,864.00
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.