Lakes and impoundments (reservoirs etc) emit a greenhouse gas equivalent of 20% of the global fossil fuel carbon dioxide emissions, with methane contributing 75% of these carbon dioxide equivalents. Stratified lakes accumulate the potent greenhouse gas methane in their oxygen-depleted bottom waters. During lake overturn, stored methane may reach the surface layer, thereby running the risk of outgassing. Despite the established importance of lakes for global greenhouse gas emissions, the fate of methane during the overturn period is still a matter of controversy. Two competing hypotheses regarding the fate of accumulated methane have been proposed. According to the first line of thinking, most of the stored methane will be released into the atmosphere as an emission pulse that adds to continuous methane fluxes across the water-air interface. According to the second hypothesis, methanotrophs oxidise most of the methane with oxygen, thereby forming biomass and carbon dioxide, a greenhouse gas which has a 34 times lower global warming potential for a 100 year time-scale. Some previous estimates assumed overturn to occur on very short time scales, e.g. a single day therefore favoring the first hypothesis. Other studies have indicated that lake overturn typically takes place on a scale of weeks to months, even in shallow lakes, which may allow time for methane oxidation. This controversy has not yet been fully resolved and the role and ecological dynamics of methane-oxidising Bacteria during lake overturn remains to be explored.
In a paper published in the journal Communications Biology on 6 March 2020, Magdalena Mayr and Matthias Zimmermann of the Swiss Federal Institute of Aquatic Science and Technology and the Institute of Biogeochemistry and Pollutant Dynamics at the Eidgenössische Technische Hochschule Zürich, Jason Dey, also of the Swiss Federal Institute of Aquatic Science and Technology, Andreas Brand and Bernhard Wehrli, again of the Swiss Federal Institute of Aquatic Science and Technology and the
Institute of Biogeochemistry and Pollutant Dynamics at the Swiss
Department of Environmental Systems Science, and Helmut Bürgmann, once again of the Swiss Federal Institute of Aquatic Science and Technology, present the results of a field study covering the entire three months of the autumn lake overturn of shallow eutrophic Lake Rotsee, Switzerland, which was designed to assess whether the methane-oxidising Bacteria assemblage grows fast enough to oxidise the methane mobilised from the bottom water before outgassing and whether the standing methane-oxidising Bacteria assemblage is activated, or a new assemblage successively takes over in the changing lake.
Lake Rotsee in Lucerne Canton, Switzerland. Tobias Grosch/Wikimedia Commons.
Methane-oxidising Bacteria in lakes have mostly been investigated during the stratified season, when a structured methane-oxidising Bacteria assemblage forms an efficient methane converter preventing the methane accumulating in the bottom water from outgassing. During lake overturn environmental conditions change compared to the stratified situation. How the methane-oxidising Bacteria assemblage responds to lake overturn remains unknown, but its growth rate and resulting methane-oxidation capacity, which are critical for the amount of methane emitted, can be expected to change in this period. The lake cools down and oxygen and methane, which are vertically separated during stratification, become simultaneously available as water with different substrate concentrations mix in the expanding surface layer. This has been shown to lead to an increase in the methane-oxidation rate and abundance of type I methanotrophs linked to the autumn lake mixing period, but the dynamics of different methane-oxidising Bacteria taxa have not been studied.
Mayr et al. used 16S rRNA gene amplicon sequencing, particulate methane monooxygenase gene mRNA sequencing and quantitative polymerase chain reaction, catalysed reporter deposition-fluorescence in situ hybridisation and potential methane-oxidation rate measurements to investigate succession, growth and methane-oxidation capacity of the methane-oxidising Bacteria assemblage during lake overturn. The rates of physical mixing and the transfer and transformation of methane were analysed with a process-based model in a parallel study.
Mayr et al.'s study provides detailed insights into the dynamics of freshwater lake methane-oxidising Bacteria during the critical overturn period. They show that a new and highly abundant methane-oxidising Bacteria assemblage, representing up to 28% of 16S rRNA gene sequences, thrived in the expanding mixed layer. In parallel with methane-oxidising Bacteria abundance, the methane-oxidation capacity of the lake increased substantially, thereby limiting methane emissions to a small percentage of the stored methane.
To determine the response of the methane-oxidising Bacteria assemblage and methan-eoxidation activity to lake overturn, Mayr et al. sampled the shallow eutrophic Lake Rotsee at eight time points covering stratification, overturn and inverse stratification. The starting point of their sampling campaign on 4 October 2016 represented a typical stratified situation, which in Rotsee starts to develop around 15 May. The large amounts of methane that have accumulated during summer in the hypolimnion (underlying layer) as are referred to as ‘stored methane’.
Most of the oxygen and stored methane in the water column were vertically separated from each other prior to lake overturn. The lake water cooled down from October to January, resulting in a gradual expansion of the mixed layer until December. By 12 December a cooler mixed layer formed on top of warmer bottom water, which has a slightly higher salinity.
From October to December, a total of 4.2 tonnes of carbon in the form of stored methane gradually entered the expanding mixed layer in Rotsee. Nevertheless, median methane concentrations in the surface layer stayed low, ranging from 0.1–1.1 micromoles per litre. During overturn, oxygen levels in the mixed layer dropped down to about 175 micromoles per litre, likely due to methane and other reduced substances from the hypolimnion triggering abiotic and biotic oxygen consumption. From December to January, oxygen concentrations increased again.
Mayr et al. analysed the methane-oxidising Bacteria assemblage with a set of independent and mutually supportive methods. They used 16S rRNA gene sequencing to phylogenetically identify known groups of methane-oxidising Bacteria, to estimate their proportion among the bacterial community and to assess methane-oxidising Bacteria assemblage dynamics. The sequencing of particulate methane monooxygenase gene mRNA sequencing and the quantification of particulate methane monooxygenase gene mRNA transcripts with quantitative polymerase chain reaction provided an independent assessment of the methane-oxidising Bacteria assemblage as well as confirmation of transcriptional activity of the methane monooxygenase.
On 4 October the proportion of methane-oxidising Bacteria marker genes peaked at the upper boundary of the methane-enriched hypolimnion, although methane-oxidising Bacteria were present throughout the water column. The hypolimnion harbored a higher proportion of methane-oxidising Bacteria marker genes than the oxygen-rich mixed layer. Therefore, the hypolimnion represents a potential reservoir of methane-oxidising Bacteria for the following lake overturn. The composition of the methane-oxidising Bacteria assemblage differed strongly between the mixed layer and hypolimnion. The mixed layer was dominated by a Gammaproteobacterial type Ia methane-oxidising Bacterium previously detected but uncultivated (i.e. known only from marker gene sequences that have been recovered from environmental samples, but not successfully cultured in the lab). According to particulate methane monooxygenase gene mRNA gene-based analysis a Methylocystis strain was also dominant; however, the corresponding quantitative polymerase chain reaction test did not detect these Bacteria.
Methylocystis Bacteria, found in Mayr et al.'s October samples by particulate methane monooxygenase gene mRNA gene-based analysis, but not by quantitative polymerase chain reaction. Scale bars are 10 μm in (a) and (b), and 0.5 μm in (c). Dedysh et al. (2007).
In the hypolimnion, other uncultivated type I methane-oxidising Bacteria closely related to lacustrine Crenothrix prevailed, with two strains present, one of which shared 95.6–96.7% of its tested gene sequences with Crenothrix, and the other 96.1–97.4%. In between the umixed layer and the methane-enriched hypolimnion the abundance of additional, unidentified, methane-oxidising Bacteria sequences peaked.
By 27 October Mayr et al. observed an increase of methane-oxidising Bacteria abundance at the upper boundary of the methane-enriched hypolimnion. Although the mixed layer depth increased throughout October, it did not yet reach the methane-enriched hypolimnion. However, the deepening of the mixed layer brought oxygen-rich water closer to the methane-enriched hypolimnion.
By 9 November the mixed layer had finally reached and eroded a small part of the methane-enriched hypolimnion. The region occupied by the methane-oxidising Bacteria peak on 27 October was replaced by the strain which already prevailed in the mixed layer in October and showed up again as the dominant methane-oxidising Bacteria in the expanding mixed layer. On the other hand, Crenothrix-related methane-oxidising Bacteria remained restricted to the hypolimnion. From 9 November onwards, Mayr et al. no longer observed the methane-oxidising Bacteria peak formation at the oxygen-methane interface, which was typical for the stable stratification period. Concurrently, the interface was constantly moved downwards as the mixed layer deepened.
From 9 November to 12 December the mixed layer deepened from 9 to 12m, thereby incorporating large amounts of stored methane. It must be assumed that the methane-oxidising Bacteria assemblage from the hypolimnion was continuously transported into the mixed layer. However, the Crenothrix-related methane-oxidising Bacteria that dominated there were only detected at very low abundances in the mixed layer. Instead, a methane-oxidising Bacterium with 97.2% rRNA gene fragment sequence identity to Methylosoma difficile was rapidly increasing in abundance, peaking at a median percentage of 16% of all Bacterial sequences in the mixed layer on 12 December. At this date, all methane-oxidising Bacteria sequences together reached up to 28% of the total Bacterial sequences, with the Methylosoma difficile-related strain as the most important contributor to the increase in methane-oxidising Bacteria abundance. This taxon was already present in October, but at very low proportion with a maximum relative abundance of 0.2% of Bacterial rRNA gene sequences.
Crenothrix Bactria, prevailent in the hypolimnion layer of Lake Rotsee, but never colonised the mised layer. Völker et al. (1977).
In January, an inverse stratification formed and reduced the depth of the mixed layer methane-oxidising Bacteria proportion and mRNA copy numbers of the Methylosoma difficile-related strain decreased and the methane-oxidising Bacteria assemblage shifted again, with a strain closely affiliated to Methylobacter now dominating.
In October, type I methane-oxidising Bacteria cell numbers peaked at the upper boundary of the methane-enriched hypolimnion. Further, potential methane-oxidation rates showed a strikingly similar pattern. Even within the oxygen-depleted hypolimnion, potential methane-oxidation rates never fell below 0.3 micromoles per day. Throughout the lake overturn, methane-oxidising Bacteria cell numbers and potential methane-oxidation rates increased substantially within the mixed layer. By 12 December the potential methane-oxidation rates in the entire mixed layer exceeded the maximum rates observed during periods of stable stratification, when maximum rates are usually confined to a narrow layer at the oxygen-methane interface. The water volume capable of these high rates had thus expanded considerably. In combination, this led to an overall increase of the potential methane-oxidation capacity in the mixed layer from 0.01 tonnes of carbon per day to 0.3 tonnes of carbon per day between October and December, corresponding to a doubling of the rate every 12.7 days. At the same time, methane-oxidising Bacteria cell numbers in the mixed layer increased with a net doubling time of about 9 days. After lake overturn was complete, i.e. between 12 December and 23 January, potential methane-oxidation rates and methane-oxidising Bacteria cell numbers generally decreased at the water depths formerly occupied by the mixed layer. The methane-oxidation rates and methane-oxidising Bacteria abundance now increased with depth and peaked in the remaining methane-enriched bottom layer near the sediment.
The winners of the lake overturn were taxa of the mixed layer methane-oxidising Bacteria assemblage, which contributed substantially to the increase of potential methane-oxidation rates. The profound shifts in the composition of the methane-oxidising Bacteria assemblage occurred simultaneously with changes of physical and chemical conditions in the mixed layer.
In October and November a previously detected but as yet uncultured Gammaproteobacterial strain dominated in the oxygen-rich mixed layer and increased in abundance at same time as the potential methane-oxidation increased from 27 October to 9 November. The early dominance of this previously uncultured strain was followed by a pronounced increase of the Methylosoma-related strain until 12 December, which led to dominance of that strain during the lake overturn. Prior to lake overturn the Methylosoma-related strain was only detected at very low abundance. During the dominance of the Methylosoma-related strain the potential methane-oxidation rates peaked in the mixed layer. The growth phase of this strain was accompanied by a drop in water temperature to below 10 °C, and increasing nutrient levels in the mixed layer. The Methylosoma-related strain peaked concurrently with the lowest measured oxygen concentration in the mixed layer during Mayr et al.'s study period. A strain related to Methylobacter tundripaludum (98.6% identity) steadily increased during the later phase of the overturn and dominated the methane-oxidising Bacteria assemblage on 23 January. At the same time lake water cooled to below 4 °C and oxygen levels replenished, while elevated nutrient levels persisted. The microbial cell numbers in the mixed layer were lower in December, but at the same time methane-oxidising Bacteria percentage and absolute methane-oxidising Bacteria cell numbers peaked.
The losers of the lake overturn, the Crenothrix-related strains, remained restricted to the eroding hypolimnion, until only a tiny fraction was left on 23 January. The Crenothrix were never able to establish in the mixed layer. They resided in water layers with high methane concentrations and low oxygen concentrations. A more flexible pattern was shown by the Methylobacter tundripaludum-like strain (99.3% identity), which was present at low abundance in the hypolimnion and temporarily gained momentum in the mixed layer in November and December.
In this study, Mayr et al. provide, to their knowledge, the most comprehensive analysis so far of the response of a methane-oxidising Bacteria assemblage to autumn lake overturn. They asked if the methane-oxidising Bacteria assemblage grows fast enough to oxidise the stored methane before outgassing, and their results unambiguously show that this was the case: Mayr et al. documented a substantial methane-oxidising Bacteria bloom during the lake overturn period when methane-oxidising Bacteria abundance increased until methane-oxidising Bacteria represented up to 28% of the Bacterial 16S rRNA gene sequences in the mixed layer. This growth in turn established a high methane-oxidation capacity in the expanding mixed layer. Detailed analysis revealed that potential methane-oxidation capacity always exceeded the methane input from the hypolimnion and limited the emitted methane to a small percentage of the stored methane during lake overturn.
Mayr et al. further asked whether the standing methane-oxidising Bacteria assemblage is activated, or a new assemblage successively takes over. They found that a temporal succession of methane-oxidising Bacteria taxa underpins the increasing methane-oxidising Bacteria abundance and methane-oxidation capacity. Together Mayr et al.'s results show that growth of a new methane-oxidising Bacteria assemblage and its methane-oxidation capacity were the main mechanism effectively limiting outgassing of hypolimnion-stored methane during lake overturn.
Beyond these main findings Mayr et al.'s work provides a number of insights into the ecology of freshwater methane-oxidising Bacteria. There is some evidence that the successional patterns of methane-oxidising Bacteria, similar to vertical niche preferences are related to specific adaptations of the successful taxa. Dominance of the unknown Gammaproteobacteria occurring under, and being correlated with, elevated oxygen concentrations supports previous observations of this taxon being positively correlated with oxygen during stable stratification in four Swiss lakes. Together, this supports the view that this methane-oxidising Gammaproteobacteria taxon thrives under high oxygen concentration and continuous but low concentration methane supply. Currently, lack of a cultured representative or genome precludes further insights into the ecology of this prevalent taxon. The overturn winner (Methylosoma) was detected at low abundance in October in Rotsee prior to overturn and was not observed in three other Swiss lakes in a previous study. This methane-oxidising Bacteria taxon thus appears to be seasonally restricted or conditionally rare, but was the most spectacular profiteer of the lake overturn and highly relevant to methane mitigation during Mayr et al.'s study period. The only cultured representative of the genuus Methylosoma, Methylosoma difficile is microaerophilic (requires oxygen to survive, but requires environments containing lower levels of oxygen than that are present in the atmosphere), which is in line with Mayr et al.'s observation that this strain peaked concurrently with the lowest measured oxygen concentrations in the mixed layer. The prevalence of Methylobacter in January could perhaps be explained by an adaptation of this strain to low temperatures: Previous observations of psychrophily (a capacity for growth and reproduction in low temperatures) within the Methylobacter genus. provide some support for this hypothesis. The Crenothrix-related strains prevalent in the hypolimnion were not able to establish in the mixed layer, although Crenothrix was presumably transported to the mixed layer continuously. The reason for this may be that Crenothrix is adapted to oxygen-deficient, high methane conditions. Further, lacustrine Crenothrix may benefit from oxygen released from phytoplankton or from temporally and spatially limited nano- to micromolar oxygen intrusions to the hypolimnion.
During the phase with the highest methane-oxidation capacity in the lake, the initial methane-oxidising Bacteria assemblage of both, mixed layer and hypolimnion, had been almost entirely replaced by a new methane-oxidising Bacteria assemblage. Different methane-oxidising Bacteria reached a high degree of dominance at different stages of the lake overturn, indicating that the changing environmental conditions favored temporal niche partitioning of methane-oxidising Bacteria taxa. It is notable, however, that despite the well-mixed situation and the often-observed dominance of a single taxon, a diverse methane-oxidising Bacteria assemblage was nevertheless present throughout the overturn period. Answering the question whether the same or similar methane-oxidising Bacteria are dominating the fall overturn every year will require a longer-term monitoring effort. Presence of methane-oxidising Bacteria taxa with specific adaptation that allow them to take advantage of the rapidly changing conditions in the overturning lake is likely essential to establishing the methane-oxidation capacity that ultimately limits methane outgassing.
Although methane becomes available in the mixed layer, dilution and rapid oxidation keep concentrations low as opposed to the situation in the hypolimnion. Methane-oxidising Bacteria in the mixed layer thus likely require a relatively high methane affinity for growth. Mayr et al. therefore speculate that the methane affinity of the mixed-layer methane-oxidising Bacteria assemblage may be a critical factor for the amount of diffusive methane outgassing to the atmosphere during the overturn period. In addition to the inherent methane affinity of methane-oxidising Bacteria taxa and assemblages, other traits like growth rates and ability to access nutrients likely affect the build-up of the methane-oxidation capacity in the lake and require further exploration.
The net doubling time of methane-oxidising Bacteria in the expanding mixed layer was about9 days. Nevertheless, methane-oxidising Bacteria substantially increased their abundance and total potential methane-oxidation capacity in the mixed layer volume reached about 0.3 tonnes of carbon per day by 12 December. While potential methane-oxidation rates may overestimate in situ rates under limiting methane concentrations, Mayr et al.'s oxidation capacity estimate is conservative as only potential methane oxidation to carbon dioxide was measured and incorporation into biomass is not included. Even though the methane-oxidising Bacteria growth rates seem low compared to observations in methane-oxidising Bacteria isolates with doubling times of several hours, they may be in line with the environmental conditions: temperatures in the mixed layer dropped from 18.4°C on 4 October to 6.1°C on 12 December, thus slowing process and growth rates are expected. Further, persistently low methane concentrations likely reduced effective growth rates. The doubling time represents a sum overall taxa. Because of the evident succession within the methane-oxidising Bacteria assemblage, growth rates of taxa becoming dominant were certainly faster as other taxa stagnated or decreased in abundance at the same time. Finally, the results presented by ayr et al. are net growth rates; gross growth rates are likely higher as mortality rates remain unknown.
While some estimates of global methane emissions from lakes assume that all stored methane is emitted to the atmosphere, Mayr et al.'s data are in line with a number of field studies that suggest a considerable proportion of stored methane is oxidised. One study estimated that 46% of the stored methane is emitted during autumn lake overturn, but most estimates are higher, claiming oxidation of 75–94% of the stored methane. Rotsee methane emission data from eddy-covariance flux measurements and modelling obtained in parallel with Mayr et al.'s study showed that about 98% of the stored methane in Lake Rotsee was oxidised. Mayr et al.'s data demonstrates that this outcome was based on the robust response of methane-oxidising Bacteria in the mixed layer where a succession of methane-oxidising Bacteria maintained a high methane-oxidation capacity throughout the overturn.
Mayr et al. therefore caution against adding methane storage to lake emission estimates. Certainly, lakes remain important sources of atmospheric methane: ebullition bypasses biological oxidation and rapid overturn events due to, for example, fast lake cooling and strong winds will lead to increased outgassing. It does, however, appear reasonable to expect that processes as observed in Rotsee may occur in a considerable proportion of temperate lakes and might be important to global methane flux estimates. Most lakes in temperate regions are holomictic and have a mean lake depth of under 25 m, similar to Rotsee. Further, small, relatively shallow lakes represent 28% of the estimated global lake area and methane storage is frequently observed especially in small lakes. The number of lakes with hypolimnetic methane storage may increase in future due to lack of recovery of lakes from eutrophication, and continuing eutrophication of other lakes, while global warming could strengthen summer stratification and, in turn, methane accumulation. A comprehensive understanding on the fate of stored methane in transitionally stratified lakes and information on the adaptability of methane-oxidising Bacteria to different environmental conditions and their methane-oxidation kinetics are therefore critical.
In summary, Mayr et al. showed in this comprehensive study that successional changes in the methane-oxidising Bacteria assemblage of the mixed layer of an overturning lake secured high methane-oxidising Bacteria abundance with high methane-oxidation capacity that strongly reduced the atmospheric emissions of stored methane.
See also...
Although methane becomes available in the mixed layer, dilution and rapid oxidation keep concentrations low as opposed to the situation in the hypolimnion. Methane-oxidising Bacteria in the mixed layer thus likely require a relatively high methane affinity for growth. Mayr et al. therefore speculate that the methane affinity of the mixed-layer methane-oxidising Bacteria assemblage may be a critical factor for the amount of diffusive methane outgassing to the atmosphere during the overturn period. In addition to the inherent methane affinity of methane-oxidising Bacteria taxa and assemblages, other traits like growth rates and ability to access nutrients likely affect the build-up of the methane-oxidation capacity in the lake and require further exploration.
The net doubling time of methane-oxidising Bacteria in the expanding mixed layer was about9 days. Nevertheless, methane-oxidising Bacteria substantially increased their abundance and total potential methane-oxidation capacity in the mixed layer volume reached about 0.3 tonnes of carbon per day by 12 December. While potential methane-oxidation rates may overestimate in situ rates under limiting methane concentrations, Mayr et al.'s oxidation capacity estimate is conservative as only potential methane oxidation to carbon dioxide was measured and incorporation into biomass is not included. Even though the methane-oxidising Bacteria growth rates seem low compared to observations in methane-oxidising Bacteria isolates with doubling times of several hours, they may be in line with the environmental conditions: temperatures in the mixed layer dropped from 18.4°C on 4 October to 6.1°C on 12 December, thus slowing process and growth rates are expected. Further, persistently low methane concentrations likely reduced effective growth rates. The doubling time represents a sum overall taxa. Because of the evident succession within the methane-oxidising Bacteria assemblage, growth rates of taxa becoming dominant were certainly faster as other taxa stagnated or decreased in abundance at the same time. Finally, the results presented by ayr et al. are net growth rates; gross growth rates are likely higher as mortality rates remain unknown.
While some estimates of global methane emissions from lakes assume that all stored methane is emitted to the atmosphere, Mayr et al.'s data are in line with a number of field studies that suggest a considerable proportion of stored methane is oxidised. One study estimated that 46% of the stored methane is emitted during autumn lake overturn, but most estimates are higher, claiming oxidation of 75–94% of the stored methane. Rotsee methane emission data from eddy-covariance flux measurements and modelling obtained in parallel with Mayr et al.'s study showed that about 98% of the stored methane in Lake Rotsee was oxidised. Mayr et al.'s data demonstrates that this outcome was based on the robust response of methane-oxidising Bacteria in the mixed layer where a succession of methane-oxidising Bacteria maintained a high methane-oxidation capacity throughout the overturn.
Mayr et al. therefore caution against adding methane storage to lake emission estimates. Certainly, lakes remain important sources of atmospheric methane: ebullition bypasses biological oxidation and rapid overturn events due to, for example, fast lake cooling and strong winds will lead to increased outgassing. It does, however, appear reasonable to expect that processes as observed in Rotsee may occur in a considerable proportion of temperate lakes and might be important to global methane flux estimates. Most lakes in temperate regions are holomictic and have a mean lake depth of under 25 m, similar to Rotsee. Further, small, relatively shallow lakes represent 28% of the estimated global lake area and methane storage is frequently observed especially in small lakes. The number of lakes with hypolimnetic methane storage may increase in future due to lack of recovery of lakes from eutrophication, and continuing eutrophication of other lakes, while global warming could strengthen summer stratification and, in turn, methane accumulation. A comprehensive understanding on the fate of stored methane in transitionally stratified lakes and information on the adaptability of methane-oxidising Bacteria to different environmental conditions and their methane-oxidation kinetics are therefore critical.
In summary, Mayr et al. showed in this comprehensive study that successional changes in the methane-oxidising Bacteria assemblage of the mixed layer of an overturning lake secured high methane-oxidising Bacteria abundance with high methane-oxidation capacity that strongly reduced the atmospheric emissions of stored methane.
See also...
Follow Sciency Thoughts on Facebook.