Spatial autocovariance of forty western Oregon catchments: Patterns, causes, and contrasts

Differences and patterns among means that can be attributed to observed variables are typically used in statistical data analyses.  Patterns among variances of response variables are rarely considered as descriptors of ecological pattern.  How autocorrelation and variance differ among systems could provide a greater understanding of processes and drivers of those processes.  Commonly, a scientist will investigate and attempt to explain why an observation at one location is not found at another location.  In essence, that is the aim of this work.  An extensive dataset of in-stream variables sampled continuously in forty headwater stream networks in western Oregon are used to investigate patterns among the spatial covariance structures of these basins.  Robust empirical network variograms are used to describe the autocovariance structure of each network.  Initial analyses have revealed that spatial autocovariance patterns do exist among the catchment areas.  We investigate potential causes for the observed similarities and differences in spatial autocovariance structures among basins using geomorphologic and stream network structure attributes.