4.2 Inferences from finer-scale variation
Using sensors, we were able to observe finer-scale temporal variation in DOC and NO3-N C-Q dynamics than are apparent from grab samples. Of particular note, higher Q generated by ROS or other precipitation events resulted in marked enrichment of DOC and NO3-N (Figure 3c-h). Notably, this enrichment response required rainfall (or rapidly melting snowfall). Diel variation in Q during melting of the snowpack, though of similar or often even greater magnitude, does not generate the same concentration response as even a small increase in Q due to new precipitation. Furthermore, it occurs during periods when DOC and NO3-N concentrations have otherwise shown signs of being depleted. This suggests that recent precipitation-derived increases in Q, as opposed to snowpack derived, are somehow able to mobilize additional solute sources, or short circuit retention along regular transport pathways. The temporal lag between Q and concentration, and the counter-clockwise hysteresis it produces (Figure 3a,b), suggests it is the later arriving water that is more enriched (Evans and Davies 1998). One potential explanation is that stores of carbon and nitrogen in more distal regions of the catchment, being further from the “drain”, are less depleted (e.g., Hood et al., 2006). Rainfall events are able to temporarily establish great connectivity with these regions, and their distance is responsible for the temporal lag. The temporal alignment of high Q and high concentrations makes ROS and other precipitation-driven events “hot moments” for solute transport, generating some of the highest instantaneous flux values observed in the record. However, their relatively infrequent occurrence and short duration mean they are still a much smaller contributor to annual flux budgets than the seasonal melt pulse.
High-frequency measurements also enabled us to observe diel variation. During the melt pulse (Figure 4a), diel Q variation was consistent with greater afternoon melting (Caine, 1992; Lundquist and Cayan, 2002; Kirchner et al., 2020). Following the melt pulse, the timing of peak daily Q shifted to earlier in the day, consistent with effects of riparian ET (Gribovszki et al., 2010; Kirchner et al., 2020). We also observed small, but noticeable diel variation in DOC and NO3-N concentrations. Signal timing was often consistent with daytime production of DOC through gross primary production (GPP), and the coupled assimilatory uptake of NO3-N (Heffernan and Cohen 2010). However, we were confounded in our attempt to estimate rates of GPP using the streamMetabolizer R model (Appling et al., 2018), because the diel DO signal exhibited daytime minima instead of expected maxima (Figure 4c & 5c). To some extent this was the result of effects of diurnal temperature variation on solubility. However, even after accounting for this, the percent saturation still exhibited consistent daytime minima (not shown). One possible explanation is that ecosystem respiration (ER) is stimulated by temperature, and daytime rates of stream ER have been found to be considerably higher than nighttime rates (Tobias et al., 2007; Hotchkiss and Hall 2014). In a narrow, forested stream such as Como Creek, light inhibition by the canopy may limit GPP to a level where daytime enhancement of ER by warmer daytime water temperatures, and the resulting consumption of DO, may become the dominant process. Rates of denitrification are also sensitive to temperature (Rusjan and Mikos, 2010), and this could also potentially explain the diel NO3-N signal absent assimilatory uptake. It is quite interesting to note however, that some of the largest diel variation in NO3-N (Figure 4b) occurred prior to the melt pulse when the stream was still mostly snow covered, there was almost no diel variation in DO, and temperatures were very near freezing. An alternative explanation is that diel solute variation (including DO) is not primarily biological, but potentially an expression of diel Q variation (Nimick et al., 2011; Hensley et al., 2017). This conclusion is supported by the large diel variation in conservative SpC during much of the year. Unfortunately, parameters which could be used to more definitively constrain biologic versus hydrologic influence (e.x. C and N isotopes, concentrations of other dissolved gases), are not collected at the sub-daily resolution required.