Headwater Catchments Govern Biogeochemistry in America's Largest Free‐Flowing River Network

Abstract

AbstractRiverine chemistry reflects both watershed conditions and instream processing, both of which vary across river networks, yet little is known about the scales at which watershed attributes regulate biogeochemical constituents. We used spatial stream network (SSN) models to quantify both watershed and instream effects on streamwater constituents in the Kuskokwim River (western Alaska), the largest free‐flowing river in the United States. We assessed chemical constituents spanning from labile nutrients (nitrate [NO3−] and orthophosphate [PO43−]) to biologically and chemically conservative tracers (calcium [Ca2+], strontium [Sr2+], and Sr isotopes [87Sr/86Sr]). We also examined the behavior of dissolved organic carbon (DOC) relative to these constituents to understand whether bulk DOC behaved conservatively in a large boreal river network, where future changes in DOC and nutrient loading are expected under a shifting climate. We derived watershed spatial covariates comprising land cover, geology, and geomorphic characteristics at 14 different spatial extents based on stream order and relative to position in the river network to understand the effect of distant and proximal watershed attributes on streamwater constituents. For all constituents but 87Sr/86Sr, watershed attributes in low‐order headwater catchments yielded the best predictive ability for streamwater constituent concentrations across the entire network. Spatial patterning for conservative tracers and bulk DOC showed strong spatial autocorrelation, whereas PO43− and NO3− exhibited spatial patchiness indicative of instream processing. Our work shows that headwater streams are disproportionately important contributors to network‐wide biogeochemical constituents supporting aquatic food webs and that conservation and management of aquatic systems should include these small and often remote watershed areas.