Implications of Multi-Fractal Scaling of Catchment Dissolved Oxygen Dynamics for Stream Respiration
Abstract
Time series of dissolved oxygen (DO) are widely used to assess primary production (GPP) and ecosystem respiration (ER) in flowing waters, as well as infer organic matter dynamics in catchments. GPP is inferred from the magnitude of diel DO variability, while ER is inferred from nighttime DO levels that reflect the balance of respiration and reaeration. Because gas exchange is typically rapid compared to network-scale stream channel transport, this inference of ER assumes that the input function - that is, the DO temporal signal derived from terrestrial runoff and subsurface flows from the upland parts of the catchment - has been erased. Long time-series of high frequency DO concentrations in streams of varying catchment size reveals, instead, clear fractal scaling in all settings, consistent with patterns observed for both reactive and conservative non-gaseous solutes. While clear breakpoints in damping of variability is evident at frequencies higher than the channel network residence time, yielding multi-fractal solute signals, the persistence of low frequency signals consistent with hillslope storage and transport suggests that assumptions in our ER computations may be incorrect. Crucially, this fractal signal frequently exhibits higher spectral power at low frequency than the obvious and well-known diel signal (i.e., variability at 24-hr frequency), suggesting that DO variance derived from hillslope processes is an unexpectedly large part of the observed signal. This implies that inferences of temporal variation in riverine ER may actually convolve river and hillslope signals, and thus overstate both the magnitude and variability of the ER flux. Using numerical experiments and existing time series, we explore the implications of multi-fractal DO scaling on river respiration, and ways to deconvolve river vs. hillslope signals. In addition to exploring the embedded assumptions in the one-station metabolism method, this work illustrates the intrinsic value of long time series of high frequency solute measurements, particularly where these are coupled to detailed understanding of catchment flow-generating mechanisms.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H11C..05C
- Keywords:
-
- 1804 Catchment;
- HYDROLOGYDE: 1879 Watershed;
- HYDROLOGYDE: 1880 Water management;
- HYDROLOGYDE: 1895 Instruments and techniques: monitoring;
- HYDROLOGY