The chemical succession in anoxic lake waters as source of molecular diversity of organic matter.

The aquatic networks that connect soils with oceans receive each year 5.1 Pg of terrestrial carbon to transport, bury and process. Stagnant sections of aquatic networks often become anoxic. Mineral surfaces attract specific components of organic carbon, which are released under anoxic conditions to the pool of dissolved organic matter (DOM). The impact of the anoxic release on DOM molecular composition and reactivity in inland waters is unknown. Here, we report concurrent release of iron and DOM in anoxic bottom waters of northern lakes, removing DOM from the protection of iron oxides and remobilizing previously buried carbon to the water column. The deprotected DOM appears to be highly reactive, terrestrially derived and molecularly distinct, generating an ambient DOM pool that relieves energetic constraints that are often assumed to limit carbon turnover in anoxic waters. The Fe-to-C stoichiometry during anoxic mobilization differs from that after oxic precipitation, suggesting that up to 21% of buried OM escapes a lake-internal release-precipitation cycle, and can instead be exported downstream. Although anoxic habitats are transient and comprise relatively small volumes of water on the landscape scale, our results show that they may play a major role in structuring the reactivity and molecular composition of DOM transiting through aquatic networks and reaching the oceans.

Magnitude and drivers of integrated fluvial network greenhouse gas emissions across the boreal landscape in Québec.

Streams and rivers are now recognized to be sites of intense carbon (C) emissions, yet the lack of C emission estimates that integrate beyond individual river systems has slowed their inclusion in landscape C budgets. Here we apply empirical models of CO2 and CH4 concentrations and gas exchange continuously along entire fluvial networks to derive the total fluvial CO2 and CH4 emissions in large (3000 to 30,000 km2) watersheds located across the boreal biome of Québec (Canada). We assess how total fluvial network C emissions vary with landscape and climate properties, and compare their magnitude to other components of the landscape C budget. The total fluvial network emissions expressed as per unit watershed area ranged from 0.7 to 29.2 g C m-2 yr-1 for CO2, and 4-1780 mg C m-2 yr-1 for CH4, and neither was related to watershed area or drainage density. Rather, watershed slope and terrestrial net productivity were major drivers of the integrated network fluvial emissions. We also show that steeper watersheds had a greater proportion of emissions relative to downstream export of C from the watershed. Integrated fluvial emissions are of the same magnitude as the terrestrial C sink, yet these two fundamental components of the boreal landscape C budget are not tightly coupled.

Fluvial CO2 and CH4 patterns across wildfire-disturbed ecozones of subarctic Canada: Current status and implications for future change.

Despite occupying a small fraction of the landscape, fluvial networks are disproportionately large emitters of CO2 and CH4 , with the potential to offset terrestrial carbon sinks. Yet the extent of this offset remains uncertain, because current estimates of fluvial emissions often do not integrate beyond individual river reaches and over the entire fluvial network in complex landscapes. Here we studied broad patterns of concentrations and isotopic signatures of CO2 and CH4 in 50 streams in the western boreal biome of Canada, across an area of 250,000 km2 . Our study watersheds differ starkly in their geology (sedimentary and shield), permafrost extent (sporadic to extensive discontinuous) and land cover (large variability in lake and wetland extents). We also investigated the effect of wildfire, as half of our study streams drained watersheds affected by megafires that occurred 3 years prior. Similar to other boreal regions, we found that stream CO2 concentrations were primarily associated with greater terrestrial productivity and warmer climates, and decreased downstream within the fluvial network. No effects of recent wildfire, bedrock geology or land cover composition were found. The isotopic signatures suggested dominance of biogenic CO2 sources, despite dominant carbonate bedrock in parts of the study region. Fluvial CH4 concentrations had a high variability which could not be explained by any landscape factors. Estimated fluvial CO2 emissions were 0.63 (0.09-6.06, 95% CI) and 0.29 (0.17-0.44, 95% CI) g C m-2  year-1 at the landscape scale using a stream network modelling and a mass balance approach, respectively, a small but potentially important component of the landscape C balance. These fluvial CO2 emissions are lower than in other northern regions, likely due to a drier climate. Overall, our study suggests that fluvial CO2 emissions are unlikely to be sensitive to altered fire regimes, but that warming and permafrost thaw will increase emissions significantly.

Transport and transformation of soil-derived CO2, CH4 and DOC sustain CO2 supersaturation in small boreal streams.

Streams are typically supersaturated in carbon dioxide (CO2) and methane (CH4), and are recognized as important components of regional carbon (C) emissions in northern landscapes. Whereas there is consensus that in most of the systems the CO2 emitted by streams represents C fixed in the terrestrial ecosystem, the pathways delivering this C to streams are still not well understood. We assessed the contribution of direct soil CO2 injection versus the oxidation of soil-derived dissolved organic C (DOC) and CH4 in supporting CO2 supersaturation in boreal streams in Québec. We measured the concentrations of CO2, CH4 and DOC in 43 streams and adjacent soil waters during summer base-flow period. A mass balance approach revealed that all three pathways are significant, and that the mineralization of soil-derived DOC and CH4 accounted for most of the estimated stream CO2 emissions (average 75% and 10%, respectively), and that these estimated contributions did not change significantly between the studied low order (≤3) streams. Whereas some of these transformations take place in the channel proper, our results suggest that they mainly occur in the hyporheic zones of the streams. Our results further show that stream CH4 emissions can be fully explained by soil CH4 inputs. This study confirms that these boreal streams, and in particular their hyporheic zones, are extremely active processors of soil derived DOC and CH4, not just vents for soil produced CO2.