Methane distribution and oxidation around the Lena Delta in summer 2013

The Lena River is one of the largest Russian rivers draining into the Laptev Sea. The predicted increases in global temperatures are expected to cause the permafrost areas surrounding the Lena Delta to melt at increasing rates. This melting will result in high amounts of methane reaching the waters of the Lena and the adjacent Laptev Sea. The only biological sink that can lower methane concentrations within this system is methane oxidation by methanotrophic bacteria. However, the polar estuary of the Lena River, due to its strong fluctuations in salinity and temperature, is a challenging environment for bacteria. We determined the activity and abundance of aerobic methanotrophic bacteria by a tracer method and by the quantitative polymerase chain reaction. We described the methanotrophic population with a molecular fingerprinting method (monooxygenase intergenic spacer analysis), as well as the methane distribution (via a headspace method) and other abiotic parameters, in the Lena Delta in September 2013. The median methane concentrations were 22 nmolL−1 for riverine water (salinity (S) < 5), 19 nmolL−1 for mixed water (5 < S < 20) and 28 nmolL−1 for polar water (S > 20). The Lena River was not the source of methane in surface water, and the methane concentrations of the bottom water were mainly influenced by the methane concentration in surface sediments. However, the bacterial populations of the riverine and polar waters showed similar methane oxidation rates (0.419 and 0.400 nmolL−1 d−1), despite a higher relative abundance of methanotrophs and a higher estimated diversity in the riverine water than in the polar water. The methane turnover times ranged from 167 days in mixed water and 91 days in riverine water to only 36 days in polar water. The environmental parameters influencing the methane oxidation rate and the methanotrophic population also differed between the water masses. We postulate the presence of a riverine methanotrophic population that is limited by sub-optimal temperatures and substrate concentrations and a polar methanotrophic population that is well adapted to the cold and methane-poor polar environment but limited by a lack of nitrogen. The diffusive methane flux into the atmosphere ranged from 4 to 163 μmolm2 d−1 (median 24). The diffusive methane flux accounted for a loss of 8 % of the total methane inventory of the investigated area, whereas the methanotrophic bacteria consumed only 1 % of this methane inventory. Our results underscore the importance of measuring the methane oxidation activities in polar estuaries, and they indicate a population-level differentiation between riverine and polar water methanotrophs.


Biogeosciences
Manuscript under review for journal Biogeosciences Discussion started: 8 February 2017 c Author(s) 2017. CC-BY 3.0 License. fragments of 350 to 700 bp were included (Schaal, 2016). Binning to band classes was performed with a position tolerance setting of 1.88%. Each band class is referred to as a MISA operational taxonomic unit (MISA-OTU).
Band patterns of MISA-OTUs were translated to binary data reflecting the presence or absence of the respective OTU.

Calculation of the diffusive methane flux
The gas exchange across an air-water interface can be described in general by the following function (Wanninkhof et al., 2009): where F is the rate of gas flux per unit area (mol m -2 d -1 ), c m is the methane concentration measured in surface 180 water and c equ is the atmospheric gas equilibrium concentration based on Wiesenburg and Guinasso (1979). Data on the atmospheric methane concentration were obtained from the meteorological station in Tiksi via NOAA, for CO 2 at 20°C) was converted to k CH4 according to (Striegl et al., 2012), where Schmidt numbers (Sc) are determined by water temperature and salinity (Wanninkhof, 2014). k CH4 / k 600 = (Sc CH4 / Sc CO2 ) 0.5 To estimate the role of methane oxidation and diffusive methane flux for the methane inventory in the Lena Delta we made the following calculations. The area was divided into two squares, which surrounded our station 195 grid (Appendix Figure A1). The median depth from the stations within each of these squares was 13 m. Based on the longitude / latitude of the squares we calculated the area and then the volume of each square (1.3 x 10 11 m 3 Biogeosciences Discuss., doi:10.5194/bg-2017-22, 2017 Manuscript under review for journal Biogeosciences Discussion started: 8 February 2017 c Author(s) 2017. CC-BY 3.0 License. and 2.5 x 10 11 m 3 ). With the median methane concentration and median MOX of all stations within each square, we calculated than the total methane inventory of the investigated areas (in mol, sum of both squares), as well as the total methane oxidation rate (mol / d). The total diffusive flux (in mol / d) of the region was obtained by 200 multiplying the median diffusive flux of all stations with the total area.

Statistical analysis
To test for differences between the different water masses we applied a one-way ANOVA with log transformed

210
We grouped our sampling stations into "riverine water" with a salinity < 5. In this water mass the median salinity was 2.45, ranging from 0.8 -4.8. Median temperature was 9.8°C, ranging from 7.3 -11.4°C. In the "mixed water" the median salinity was 11.4, ranging from 5 -19.7. Median temperature was 6.4°C, ranging from 2.5 -8.8°C. In the "polar water" the median salinity was 27.2, ranging from 21.5 -33.2. Median temperature was 3.0°C, ranging from 1.8 -6.2°C. In September 2013 we observed a sharp stratification with 215 warm freshwater at the surface (0 -5 m), followed by a mixed water body. Below approx.10 m water depth, we found cold and saline water (= polar water). As example of this sharp stratification, the salinity distribution of Transect 1 is shown in Figure 2. The freshwater plume was most pronounced in Transect 4 and 5 and extended far to the north. In Transect 6 only the first near-shore station had riverine water, the following stations were already characterized by polar waters. When applying our water masses (riverine, mixed and polar), we observed significantly different methane 230 concentrations in these water masses, with medians of 22, 19 and 26 respectively ( Table 1).
In "riverine water", methane concentration was significantly correlated with temperature (r 2 = 0.38, Table 2) and negatively correlated with the oxygen concentration (r 2 = 0.73). In "mixed water", we found a weak but significant correlation between methane and TDN (r 2 = 0.27, Table 2). In "polar water" the methane concentration of the water column was significantly correlated with the methane concentration in the surface 235 sediment (r 2 = 0.33). The influence of the sediment methane concentration on the water column concentration was even more pronounced when taking all bottom water samples (="polar water" + one "mixed water" + one sea water results in brine formation with strongly increased salinity, while its melting results in a freshwater input (Eicken et al., 2005). In contrast to sea-ice, the freezing and melting of freshwater-ice does not alter the 335 salinity pattern. In 1999, the river water fraction in ice-cores near our study area ranged from 57% -88% (Eicken et al., 2005), thus at least some additional non-river-freshwater is possible. Even though not much is known about methane concentrations in ice, based on a recent study on sea-ice in the East Siberian Sea (Damm et al.,

2015)
, we assume that this melt water probably has lower methane concentrations than the river-freshwater. This low methane input could than explain the missing relation between salinity and methane concentration.

340
In bottom water, methane concentrations were only influenced by the methane concentration in the sediment below. Thus we assume that this methane mostly originates from a (diffusive) methane flux out of the sediment.
Unfortunately no isotope analysis to validate this assumption was possible. .8 x 10 -3 h -1 ) (Wahlström and Meier, 2014). From our data we suggest more realistic turnover times ranging from 91 d -1 in riverine water, 167 d -1 in the mixed water and 36 d -1 in polar water.

355
In the "riverine water", MOX and fractional turnover rate were correlated with temperature (ranging from 7 -11°C), while in the other water masses no such correlation was found. Also, the influence of the methane concentration on the MOX was most pronounced in "riverine water" (r 2 = 0.98). In polar water, MOX was influenced the by TDN, but compared with riverine water, methane concentration had a much lower influence (r 2 = 0.56). near shore to about 120 km offshore (Transect 1). Figure 2. Salinity distribution versus depth and distance from the shore for Transect 1. Indicated are also the water masses defined as "riverine" with a salinity < 5, "mixed water" between 5 and 20, and "polar water" with a salinity > 20     Figure A1. Map of study area with two grids to estimate the total sampling area.
490 Table 1. The median values of important parameters in the different water masses. A one-way ANOVA was performed to test for significant differences of the log-transformed data between the water masses. 495 Table 3. Linear corelation between the methane oxidation rate (MOX) and the fractional turnover rate (k) versus different enviromental parameters splitted into three water masses. Analysis was performed with log transformed data, shown are the r2-values and the level of significance (p). Empty fields indicate no siginificant corelation