Integrating the statistical modeling of spatial data in ecology

• Andrew M. Liebhold
• Jessica Gurevitch

Members Only Area

Abstract
In a variety of ecological disciplines, there is growing interest in quantifying and modeling spatial patterns. This trend reflects the growing recognition of the importance of spatial processes in shaping ecological interactions. Several different approaches have been employed by ecologists for quantifying patterns in spatial data. One approach, developed originally by geographers, focuses on quantifying spatial autocorrelation and hypothesis testing. Geostatistics is another approach that was originally developed by geologists for quantifying and Interpolating spatial data. A third approach, "point process" statistics, was developed for analyzing patterns in the physical location of plants and animals. All of these methods have unique features that are desirable in ecological applications but the relationships and similarities among these methods have not been fully explored. The diversity of these various approaches is sometimes overwhelming to ecologists and may represent an impediment to the choice of effective statistical methods for the analysis of ecological data. We propose to assemble a working group of 10-12 scientists representing a spectrum of backgrounds, including pure statistics and various ecological disciplines. The group will compare these differing approaches to spatial analysis and summarize their similarities and differences. Ultimately we intend to provide a road map for ecologists to use in the selection of statistical methods for analyzing spatial data.