Landscape Ecology

, Volume 28, Issue 10, pp 1849–1858

Increasing the chance that landscape- and regional-level hypotheses will reflect important spatial patterns


DOI: 10.1007/s10980-013-9942-5

Cite this article as:
Gutzwiller, K.J. Landscape Ecol (2013) 28: 1849. doi:10.1007/s10980-013-9942-5


Development of new ecological knowledge depends in part on the hypotheses that get tested, and little headway in establishing ecological truth is likely if hypotheses fail to reflect important spatial patterns. Careful exploration for spatial patterns during initial hypothesis formation—before tests of hypotheses and tests of statistical significance are pursued—is therefore essential. Exploratory analyses that rely on quantitative methods alone would preclude detection of important and unexpected spatial patterns that might be suggested or become apparent only through graphical visualization of spatial structure. Geographic mapping of spatial eigenvectors has the potential to improve knowledge about autocorrelation-related patterns by revealing pattern size, shape, geographic location, and directional orientation. Such map information is valuable because it may visually elicit insights about possible landscape- and regional-level drivers that no summary statistic, bivariate scatter plot, or statistical model alone could evoke. Including visual assessments of spatial structure in exploratory analyses may improve the chance that important spatial patterns will be discovered and hence considered during development of initial landscape- and regional-level hypotheses. These benefits in turn have the potential to strengthen the spatial validity of associated inferences.


Distance-based Moran’s eigenvector mapsExploratory analysisGeographic mappingGraphical visualizationPrincipal coordinates of neighbor matricesSpatial autocorrelationSpatial eigenvector analysis

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  1. 1.Department of BiologyBaylor UniversityWacoUSA