The spatial structure of variability in a semi-arid, fluvial ecosystem

The arrangement and composition of flowpath types within a given network are thought to govern its functioning. This concept assumes that different flowpath types are functionally distinct. We investigated this assumption in a fluvial ecosystem by comparing the riparian zone, parafluvial zone (in-channel gravel bars), and surface stream. We hypothesized that differences in advection, uptake, and sorption would render material cycles more (a) open and (b) mutable in the surface stream, whereas the converse would occur in the riparian zone, and an intermediate state would be seen in the intervening parafluvial zone. To test our first hypothesis, we predicted that spatial heterogeneity in solute concentrations would be least in the surface stream, greater in the parafluvial zone, and greatest in the riparian zone. Using a null model, we ascertained that this pattern was shown by all solute species we examined (nitrate, ammonium, total dissolved inorganic nitrogen [DIN], dissolved organic N, total dissolved N, soluble reactive phosphorus, dissolved organic carbon, and chloride). To test our second hypothesis, we predicted that temporal change in spatial heterogeneity would be greatest in the surface stream, less in the parafluvial zone, and least in the riparian zone. Nitrate, DIN, and chloride showed this pattern. In particular, surface stream inorganic N was less spatially variable following months of high rainfall. According to an extant hypothesis, these results suggest that inorganic N processing may be a stable function in this ecosystem. Other solute species did not support our second prediction, perhaps because their retention and release dynamics are influenced principally by geochemistry. Generally, our findings indicate that a geomorphic template can generate spatial patterns in ecosystem function, warranting an expansion of the spiraling framework to a variety of flowpath types.