An inductive approach to characterize physical, chemical, and biological system interactions in a 5th order river basin

Abstract

<p>Research in the river corridor commonly focuses in two study designs. One research strategy focuses on physical, chemical, and/or biological dynamics and feedbacks, emphasizing local variation and interaction over larger-scale context. A second study design focuses on gradients arising in response to non-local controls (e.g., climate, tectonic setting), with an emphasis on broad trends over smaller-scale “noise”. Here, we present a comprehensive set of measurements and calculated metrics describing physical, chemical, and biological conditions collected at 62 sites in the river corridor within a 5<sup>th</sup> order basin including more than 150 variables at each site. The size and scope of this data set allows us to assess which variables have spatial structure in the basin using spatial semivariograms and regressions with discharge and drainage area. We ask how physical, chemical, and biological sub-systems co-vary using a principal components analyses. Next, we explain both spatial structure and local variance simultaneously using support vector machine regression techniques that reveal possible nonlinear, multivariate relationships that may direct future research. Key outcomes from this study include (1) an introduction to an open-source, comprehensive characterization of the river corridor, (2) interpretations of both broad trends and local variance in the river corridor, and (3) a summary of which metrics have the most explanatory power and why within the study system.</p>