Search

Inland waters are significant sources of carbon dioxide (CO2) to the atmosphere. CO2 supersaturation and subsequent CO2 emissions from inland waters can be driven by internal metabolism, external inputs of dissolved inorganic carbon (DIC) derived from the catchment, and other processes (e.g., internal geochemical reactions of calcite precipitation or photochemical mineralization of organic solutes). However, the sensitivity of the magnitude and sources of CO2 emissions to fluvial network hydromorphological alterations is still poorly understood. Here we investigated both the magnitude and sources of CO2 emissions from lotic (i.e., running waters) and lentic (i.e., stagnant waters associated to small dams) waterbodies of a Mediterranean fluvial network by computing segment-scale mass balances of CO2. Our results showed that sources other than internal metabolism sustained most (82%) of the CO2 emissions from the studied fluvial network. The magnitude and sources of CO2 emissions in lotic waterbodies were highly dependent on hydrology, with higher emissions dominated by DIC inputs derived from the catchment during high flows and lower emissions partially fueled by CO2 produced biologically within the river during low flows. In contrast, CO2 emissions in lentic waterbodies were low, relatively stable over the time and the space, and dominated by DIC inputs from the catchment regardless of the different hydrological situations. Overall, our results stress the sensitivity of fluvial networks to human activities and climate change and particularly highlight the role of hydromorphological conditions on modulating the magnitude and sources of CO2 emissions from fluvial networks.

[1] Rivers are heterogeneous at various scales. River metabolism estimators based on oxygen time series provide average estimates of net oxygen production at the scale of a river reach. These estimators are derived for homogeneous river reaches. For this reason, they cannot be used to analyze how exactly they average over longitudinal variations in net production, reaeration, oxygen saturation concentration and flow velocity. We try to fill this gap by using a general analytical solution of the transport-reaction equation to (1) demonstrate how downstream oxygen concentration is affected by upstream concentration and (possible) longitudinally varying values of net production, reaeration, oxygen saturation concentration and flow velocity within a reach, and (2) derive how the net production estimate depends on varying upstream river parameters. In addition, we derive a new net production estimator that extends previously suggested estimators. The equations derived in this paper provide a general framework for understanding the assumptions underlying net production estimators. They are used to derive recommendations on the use of single station or two stations measurement layouts to get accurate river metabolism estimates. The estimator is implemented in the freely available statistics and graphics software package R (www.r-project.org). This makes it easily applicable to observed oxygen time series. Empirical evidence of the significance of heterogeneity in rivers is demonstrated by applying the estimator to four subsequent reaches of a river using oxygen measurements from the ends of all reaches.

[1] The chemical composition of surface waters of nine streams with drainage sizes ranging from 44 to 8900 km2 within a semiarid catchment in the southwestern United States was examined over the monsoonal season of 2006. Primary goals were to determine the impact of monsoonal rains on linkages between landscape cover features and water chemistry in a semiarid river network and to identify whether drainage size influences the temporal variability in water chemistry of streams. Landscape cover features (geology, soil, and vegetation types) were quantified for the subcatchment upstream of each study site and the riparian ecotone. Processes of binary mixing, dilution, and concentration were identified by end member mixing analysis (EMMA). Results showed that most chemical constituents corresponded to geological features at the basin scale, but other constituents (TSS and PO43−) corresponded more closely to riparian features. Importantly, statistical relationships between land cover and water chemistry differed between baseflow and monsoonal conditions, suggesting that seasonal changes in hydrologic routing and water sources (e.g., overland flow versus groundwater sources) have important consequences for stream chemistry. In contrast with mesic drainages, temporal variability in water chemistry increased with spatial scale. The variety of contributing water sources and the relative effect of dilution and concentration in each spatial scale may be underlying processes that explain differences across spatial scales. Observed patterns in the Gila basin suggest that climate variability interacts with hydrologic routing to influence spatial patterns of stream water chemistry in arid basins. Observations made here may be similar to other systems with strongly seasonal climate patterns, and the underlying determinants of stream chemistry change with drainage size.