Uncovering Relationships Between Geo-Statistical and Spatial Autoregressive Modelling

点击翻译按钮获取中文摘要

The investigator undertakes to complete establishing connections between two separate contemporary areas of statistics, namely geo- and spatial statistics. These connections are developed (1) by analyzing popular scientific models in both of these sub-fields; (2) by evaluating these models within the context of the sizes and shapes of geographic regions to which they are applied; (3) by studying the role geographic direction plays in preserving or altering these connections; and (4) by investigating properties of these models for geographic maps upon which similar or markedly different classificatory color codings tend to cluster. Much of the theoretical research will be pursued with the use of computers. Selected U.S. EPA Environmental Monitoring and Assessment Program (EMAP) data already obtained for the Chesapeake and Delaware estuaries and the northeast forest inventory will be used for empirical analysis and demonstration purposes. The practical importance of this proposed research arises within the U.S. EPA EMAP project, a major national endeavor in which statistics are being applied to environmental issues as well as monitoring and analysis undertakings. Articulation of a common foundation for geo- and spatial statistics will better enable consistent statistical application and analysis of the massive amounts of geo-referenced, or locationally tagged, EMAP data becoming available. These data currently are being collected and geo-referenced in terms of a network of approximately 12,600 hexagons overlayed on the continental United States.