Can Concentration - Discharge Relationships Diagnose Material Source During Extreme Events?

Floods can carry >90% of the basin material exported in a given year as well as alter flow pathways and material sources. In turn, sediment and solute fluxes can increase flood damages and negatively impact water quality and integrate physical and chemical weathering of landscapes and channels. Concentration-discharge (C-Q) relationships are used to both describe export patterns as well as compute them. Metrics for describing C-Q patterns and inferring their controls are vulnerable to infrequent sampling that affects how C-Q relationships are interpolated and interpreted. C-Q relationships are typically evaluated from multiple samples, but because hydrological extremes are rare, data are often unavailable for extreme events. Because solute and sediment C-Q relationships likely respond to changes in hydrologic extremes in different ways, there is a pressing need to define their behavior under extreme conditions, including how to properly sample to capture these patterns. In the absence of such knowledge, improving load estimates in extreme floods will likely remain difficult. Here we explore the use of C-Q relationships to determine when an event alters a watershed system such that it enters a new material source/transport regime. We focus on watersheds with sediment and discharge time series include low-frequency and/or extreme events. For example, we compare solute and sediment patterns in White Clay Creek in southeastern Pennsylvania across a range of flows inclusive of multiple hurricanes for which we have ample ancillary hydrochemical data. TSS is consistently mobilized during high flow events, even during extreme floods associated with hurricanes, and sediment fingerprinting indicates different sediment sources, including in-channel remobilization and landscape erosion, are active at different times. In other words, TSS mobilization in C-Q space is not sensitive to the source of material being mobilized. Unlike sediments, weathering solutes in this watershed tend to exhibit a relatively chemostatic C-Q pattern, except during the runoff-dominated Hurricane Irene, when they exhibit a diluting C-Q pattern. Finally, we summarize the vulnerability of these observations to shifts in sampling effort to highlight the utility and limitations of C-Q-derived export patterns.