Why the bare surfaces of permafrost peatlands (BP) have higher N 2 O

SUR-ROUNDING VEGETATED PEATLANDS SURFACES?

N²O emissions from the BP soil are similar to those observed from northern peatlands drained for agriculture/forestry (Maljanen et al., 2010). When peatlands are drained, the organic matter accumulated with initial high water table conditions (anaerobic conditions), is exposed and the aerobic conditions enhance mineralization of organic matter and microbial process behind the N²O emissions (Martikainen et al., 1993).

The amount of N2O emitted from drained peatlands depends on the quality of peat.

After drainage nutrient-rich peat soils (minerotrophic peatlands) are known to have high N2O emissions in contrast to nutrient-poor peatlands (ombrotrophic peatlands) (Martikainen et al., 1993). BP soil is minerotrophic with high organic N reservoirs and low C to N ratio supporting N2O emissions (Klemedtsson et al., 2005; Marushchak et al., 2011 ). The permafrost peatlands have been uplifted by frost action creating aerobic conditions at upper part of peat profile and exposing the top soil to other processes related to the harsh climatic conditions (e.g frost heave, winter wind erosion) leading to sporadic destruction of the vegetation cover (Marushchak et al., 2011). As mentioned above, permafrost peatlands located in remote areas (i.e. sub-Arctic/Arctic tundra) have low external N input via atmospheric N deposition. Then the availability of mineral N in soil solely depends on mineralization of organic matter. The oxic conditions in the top soil created by permafrost uplifting enables mineralization of organic matter that supplies mineral N for microbial processes including nitrification and denitrification.

Mineralization rates in the BP soil were found to be higher compared to the vegetated soil (VP) surrounded the BP surfaces (Marushchack et al., 2011). The ¹⁵N tracer approach yielded an interesting insight on the role of plants in N²O emissions from the permafrost peatlands. The higher recovery of the ¹⁵N label in N²O in BP (absent of plants) (maximum value ~24%) compared to the relative portion of ¹⁵N label observed in plants from the VP (maximum value ~ 9%) suggested that either the competition for N between plants and microbes limits the N²O release in VP, or that N²O production is restricted by the lower availability of N associated to low N mineralization and nitrification rates in VP soil with higher C/N ratio. Indeed, gross mineralization and particularly gross nitrification rates were much lower in VP com-pared to BP (Table 2 in manuscript), suggesting the importance of the slow N deliv-ery. Even though there was high uncertainty in these results based on methodologi-cal constraints (discussed above), the difference was so large that it likely stands. In any case, the presence of plants limits N2O emissions. This phenomena has been re-ported for restored boreal peatlands with various levels of nitrate addition an plant coverage (Silvan et al., 2005) and is supported by recent results from a mesocosm study which shows that the presence of vegetation limits N2O emissions from tundra soil by ∼90% (Voigt et al., 2017b).

Because BP soils have a constant good supply of inorganic N for microbial processes, the N²O production is largely regulated by soil water content affecting soil oxygen status and microbial processes. In the BP soils the mean water content was ~ 60 % WFPS, where both aerobic and anaerobic processes can take place. Then nitrification and denitrification processes occur simultaneously. The abundance of denitrifiers has been found to be exceptionally high in the BP soil (Palmer et al., 2011).

The results from the ¹⁵N tracer experiment (2010) confirm the occurrence of both nitrification and denitrification in the BP soil where nitrification coupled denitrification accounting for up to 98%, on average 43%, of the N²O produce under intermediate to high soil moisture conditions (≥ 60 % WFPS; like the ones observed in 2010), and support the assumption that denitrification is mostly responsible for the high N²O emissions from the BP soil.

In 2011, the mean % WFPS in the BP top soil was between 8 to 14% lower than in 2010 and previous years (Marushchack et al., 2011). Likewise the N²O emissions were considerably lower than those in 2010 and previously (Repo et al., 2009, Marushchack et al., 2011).The results on SP showed that the N²O emissions in 2011, originated from nitrification and/or nitrifier-denitrification as previously found for non-saturated peat (Liimatainen et al., 2014). These results suggest that the relative contribution of nitrification to the N2O emissions from the BP soil becomes more relevant under drier conditions.

Overall, nitrification and denitrification processes occur simultaneously in the BP soil and the relative contribution of these processes to the N2O production is determined by soil moisture. In wet years (WFPS ≥ 60) (e.g. 2007, 2008, and 2010) N2O emission are high and denitrification is the primary source of N²O. In a dry year (WFPS ≤ 60%) (e.g. 2011) N2O emissions are low and the relative contribution of nitrification/nitri-fier-denitrification increases.

4.5 FUTURE PERSPECTIVES FOR N2O EMISSIONS FROM THE ARCTIC WITH IMPLICATIONS FOR TROPOSPHERIC N2O TRENDS

Because mineral N supply and soil moisture content are optimal for N²O production in the BP soil, the N²O emissions can be regulated by the ambient temperature. This would mean an increase in the N²O emissions from BP soils in future warmer climate.

In fact this has been confirmed by a warming experiment in situ on the peat plateau complexes. The results from this experiments have shown that also the vegetated permafrost peat surfaces can start to emit N2O emissions at temperatures higher than prevailing presently (Voigt et al., 2017a). In addition, recent mesocosms and laboratory studies have proven that arctic soils have a potential for high N²O emissions after permafrost thawing (Elberling et al., 2010; Voigt et al., 2017b). Under this scenario, the results of this work suggest that the expected future increase in the

N²O emissions from natural ecosystems (e.g the sub-Arctic permafrost peatlands) due to a warmer climate might counteract the observed decreasing trend of the isotope composition of tropospheric N²O which presently is largely driven by agriculture as shown by the N²O isotopic composition measured in the agricultural soil of this work and by Perez et al., 2001. Thus, if in the future the decreasing trend of N²O tropospheric δ¹⁵N^bulk isotopic composition diminishes, it cannot be only attributed solely to the decrease in agricultural N²O emissions (with light δ¹⁵N^bulk values) resulting from mitigation actions. In warmer climate also natural ecosystems (with heavier δ¹⁵N^bulk values), such as permafrost peatlands in the Arctic tundra, may emitt more N²O and have an impact on the isotopic composition of N²O in the atmosphere.

In order to evaluate accurately microbial pathways producing and consuming N²O in soils using stable isotope techniques, more laboratory studies, especially soil incubation studies are needed. At present, the results on bulk and SP values of N²O between culture and soil incubation studies do not match (article I; Toyoda et al., 2015) and there is a dominance of culture-based studies on δ¹⁵N²O. We find that par-ticularly soil incubations studies are important, and necessary to use correct end-member values for any isotope partitioning approach. Also further research on the isotopic signature of N2O, particularly SP measurements from soils in situ, is re-quired. Further, there is a need to improve the analysis of ¹⁵N in inorganic N forms, particularly from highly organic soils. The current methods seem to underestimate the isotope values heavily, and result to biases in gross mineralization and nitrifica-tion rates and source partinitrifica-tioning approaches.