Spatiotemporal heterogeneity of water flowpaths controls dissolved organic carbon sourcing in a snow-dominated, headwater catchment

Authors: RADKE, ANNA - Idaho State University; GODSEY, SARAH - Idaho State University; LOHSE, KATHLEEN - Idaho State University; MCCORKLE, EMMA - Idaho State University; PERDRIAL, JULIA - University Of Vermont; Seyfried, Mark; HOLBROOK, STEVEN - Virginia Tech

Submitted to: Frontiers in Ecology and Evolution
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/7/2019
Publication Date: 2/27/2019
Citation: Radke, A., Godsey, S., Lohse, K., McCorkle, E., Perdrial, J., Seyfried, M.S., Holbrook, S. 2019. Spatiotemporal heterogeneity of water flowpaths controls dissolved organic carbon sourcing in a snow-dominated, headwater catchment. Frontiers in Ecology and Evolution. 7(46). https://doi.org/10.3389/fevo.2019.00046.
DOI: https://doi.org/10.3389/fevo.2019.00046

Interpretive Summary: Our understanding of how water travels from source areas to stream channels is limited. In the past, hydrologists have observed the timing of inputs to the system and streamflow increases or decreases and then made logical estimates of what the flow path probably was. This is generally works well for most years in terms of estimating streamflow and water supply. It can be misleading for extreme years. In addition, it does not lead to an understanding of how chemical get in to streams, how springs are fed, or what the impacts of extended droughts will be. We investigated flow paths at Reynolds Mountain in the Reynolds Creek Experimental Watershed in soutwest Idaho using a combination of stream chemistry and geophysics data. This site has the advantage that the water inputs cloation is well defined by the location of snow snowdrifts. We found that streamwater is sourced primarily from groundwater (averaging 25% of annual 27 streamflow), snowmelt (50%), and water travelling along the saprolite/bedrock boundary (25%). Multiple subsurface regions in the catchment appear to contribute differentially to streamflow as the season progresses; sources shift from weathered bedrock to deeper bedrock aquifers from the snowmelt period into summer. Unlike most studied catchments, lateral flow of this year’s soil water is not a primary source of streamflow. Thus, water is stored within the catchment for multiple years and the flow for a given year is influenced by previous years flows.

Technical Abstract: The non-uniform distribution of water in snowdrift-driven systems can lead to spatial heterogeneity in vegetative communities and soil development, as snowdrifts may locally increase weathering. The focus of this study is to understand the coupled hydrological and biogeochemical dynamics in a heterogeneous, snowdrift-dominated headwater catchment (Reynolds Mountain East, Reynolds Creek Critical Zone Observatory, Idaho, USA). We determine the sources and fluxes of stream water and dissolved organic carbon (DOC) at this site, deducing likely flowpaths from hydrometric and hydrochemical signals of soil water, saprolite water, and groundwater measured through the snowmelt period and summer drying. We then interpret flowpaths using end-member mixing analysis in light of inferred subsurface structure derived from electrical resistivity and seismic velocity transects. Streamwater is sourced primarily from groundwater (averaging 25% of annual streamflow), snowmelt (50%), and water travelling along the saprolite/bedrock boundary (25%). The latter is comprised of the prior year’s soil water, which accumulates DOC in the soil matrix through the summer before flushing to the saprolite during snowmelt. DOC indices indicate that it is sourced from terrestrial carbon, and derives originally from soil organic carbon (SOC) before flushing to the saprolite/bedrock boundary. Multiple subsurface regions in the catchment appear to contribute differentially to streamflow as the season progresses; sources shift from the saprolite/bedrock interface to deeper bedrock aquifers from the snowmelt period into summer. Unlike most studied catchments, lateral flow of this year’s soil water is not a primary source of streamflow. Instead, saprolite and groundwater act as integrators of soil water that flows vertically in this system. Our results do not support the flushing hypothesis as observed in similar systems and instead indicate that temporal variation in connectivity may cause the unexpected dilution behavior displayed by DOC in this catchment.