Differential response of carbon cycling to long-term nutrient input and altered hydrological conditions in a continental Canadian peatland

Peatlands play an important role in global carbon cycling, but their responses to long-term anthropogenically changed hydrologic conditions and nutrient infiltration are not well known. While experimental manipulation studies, e.g., fertilization or water table manipulations, exist on the plot scale, only few studies have addressed such factors under in situ conditions. Therefore, an ecological gradient from the center to the periphery of a continental Canadian peatland bordering a eutrophic water reservoir, as reflected by increasing nutrient input, enhanced water level fluctuations, and increasing coverage of vascular plants, was used for a case study of carbon cycling along a sequence of four differently altered sites. We monitored carbon dioxide (CO2) and methane (CH4) surface fluxes and dissolved inorganic carbon (DIC) and CH4 concentrations in peat profiles from April 2014 through September 2015. Moreover, we studied bulk peat and pore-water quality and we applied delta C-13-CH4 and delta C-13-CO2 stable isotope abundance analyses to examine dominant CH4 production and emission pathways during the growing season of 2015. We observed differential responses of carbon cycling at the four sites, presumably driven by abundances of plant functional types and vicinity to the reservoir. A shrub-dominated site in close vicinity to the reservoir was a comparably weak sink for CO2 (in 1.5 years: -1093 +/- 794, in 1 year: +135 +/- 281 g CO2 m(-2); a net release) as compared to two graminoid-moss-dominated sites and a moss-dominated site (in 1.5 years: -1552 to -2260 g CO2 m(-2), in 1 year: -896 to -1282 g CO2 m(-2)). Also, the shrub-dominated site featured notably low DIC pore-water concentrations and comparably C-13-enriched CH4 (delta C-13-CH4: -57.81 +/- 7.03 parts per thousand) and depleted CO2 (delta C-13-CO2: -15.85 +/- 3.61 parts per thousand) in a more decomposed peat, suggesting a higher share of CH4 oxidation and differences in predominant methanogenic pathways. In comparison to all other sites, the graminoid-moss-dominated site in closer vicinity to the reservoir featured a similar to 30% higher CH4 emission (in 1.5 years: +61.4 +/- 32, in 1 year: +39.86 +/- 16.81 g CH4 m(-2)). Low delta C-13-CH4 signatures (-62.30 +/- 5.54 parts per thousand) indicated only low mitigation of CH4 emissions by methanotrophic activity here. Pathways of methanogenesis and methanotrophy appeared to be related to the vicinity to the water reservoir: the importance of acetoclastic CH4 production apparently increased toward the reservoir, whereas the importance of CH4 oxidation increased toward the peatland center. Plant-mediated transport was the prevailing CH4 emission pathway at all sites even where graminoids were rare. Our study thus illustrates accelerated carbon cycling in a strongly altered peatland with consequences for CO2 and CH4 budgets. However, our results suggest that long-term excess nutrient input does not necessarily lead to a loss of the peatland carbon sink function.