Soil incubation methods lead to large differences in inferred methane production temperature sensitivity

Quantifying the temperature sensitivity of methane (CH4) production is crucial for predicting how wetland ecosystems will respond to climate warming. Typically, the temperature sensitivity (often quantified as a Q10 value) is derived from laboratory incubation studies and then used in biogeochemical models. However, studies report wide variation in incubation-inferred Q10 values, with a large portion of this variation remaining unexplained. Here we applied observations in a thawing permafrost peatland (Stordalen Mire) and a well-tested process-rich model (ecosys) to interpret incubation observations and investigate controls on inferred CH4 production temperature sensitivity. We developed a field-storage-incubation modeling approach to mimic the full incubation sequence, including field sampling at a particular time in the growing season, refrigerated storage, and laboratory incubation, followed by model evaluation. We found that CH4 production rates during incubation are regulated by substrate availability and active microbial biomass of key microbial functional groups, which are affected by soil storage duration and temperature. Seasonal variation in substrate availability and active microbial biomass of key microbial functional groups led to strong time-of-sampling impacts on CH4 production. CH4 production is higher with less perturbation post-sampling, i.e. shorter storage duration and lower storage temperature. We found a wide range of inferred Q10 values (1.2–3.5), which we attribute to incubation temperatures, incubation duration, storage duration, and sampling time. We also show that Q10 values of CH4 production are controlled by interacting biological, biochemical, and physical processes, which cause the inferred Q10 values to differ substantially from those of the component processes. Terrestrial ecosystem models that use a constant Q10 value to represent temperature responses may therefore predict biased soil carbon cycling under future climate scenarios.


Introduction
Understanding and quantifying methane (CH 4 ) production temperature sensitivity are important to improve predictions of how wetland ecosystems will respond to and feedback on climate warming (Davidson and Janssens 2006).The sensitivity of CH 4 production to temperature is often described by a Q 10 value, which is defined as the factor by which CH 4 production increases when temperature increases by 10 • C (van Hulzen et al 1999).Q 10 is usually derived from laboratory incubation experiments where soil samples are placed in controlled conditions at different temperatures and CH 4 production is measured over time (Zheng et al 2018).Such incubationinferred Q 10 values are often incorporated into terrestrial ecosystem models.A constant Q 10 value of 2 is often assumed (Walter andHeimann 2000, Riley et al 2011), although a large variation of Q 10 values has been reported, and a large portion of that variability remains unexplained (Craine et al 2010, Hamdi et al 2013, Meyer et al 2018, Haaf et al 2021).The significant effects Q 10 value can have on modeled CH 4 emissions implies the need to better understand and quantify variations in Q 10 (Riley et al 2011).We note that many of these issues have also been shown to be important for evaluating the temperature sensitivity of soil carbon dioxide (CO 2 ) production (Gu et al 2004, Davidson et al 2006, Fierer et al 2006, Zhou et al 2009).
The limited understanding of CH 4 production temperature sensitivity arises from several factors (Segers 1998).First, CH 4 production processes are complex and involve various microbial activities, including syntrophic interactions and competition for key substrates (Le Mer andRoger 2001, Bridgham et al 2013).Heterotrophic microbes drive the breakdown of complex organic polymers to simple substrates.Fermentation of these substrates result in production of additional substrates including H 2 , CO 2 , and acetate.Acetate can be fermented to form CO 2 and CH 4 by acetoclastic methanogens (AM), and CO 2 can be reduced to CH 4 using H 2 as an electron donor by hydrogenotrophic methanogens (HM).The rates of these processes are also affected by environmental factors such as soil moisture, soil temperature, oxygen concentration, and substrate concentrations (Schlesinger and Bernhardt 2013).
Although laboratory experiments provide a controlled environment with comparatively stable soil moisture and temperature, uncertainties can arise from the time or season of soil collection, conditions under which the soil is stored (i.e.temperature, duration), and pre-treatment periods (Rhymes et al 2021, Schroeder et al 2021, Wilson et al 2021).Previous work assessing the impacts of sampling time and storage on microbial activities, CH 4 production, and inferences of Q 10 values had inconsistent results (Rhymes et al 2021, Wilson et al 2021).For example, Lupascu et al (2012) found sampling time affects CH 4 production but not inferred Q 10 .In contrast, Bergman et al (2000) found that Q 10 values of CH 4 production varies with the time of collection due to substrate availability and seasonal variability in active microbial biomass.These conflicting results may stem from challenges in quantifying controlling factors (e.g.carbon quality, substrate concentrations and composition, and microbial biomass and activity) continuously and accurately (Blagodatskaya and Kuzyakov 2013), thereby hindering interpretation of incubation measurements.
Here we apply observations and a well-tested process-rich model, ecosys, to (1) interpret laboratory incubation observations; (2) investigate controls on inferred CH 4 production temperature sensitivity; and (3) inform incubation strategies.The ecosys model simulates the physical, hydrological, and biological processes that govern ecosystem responses to environmental conditions and has been applied in dozens of permafrost sites (Grant et al 2017, Mekonnen et al 2021, Riley et al 2021).Ecosys represents multiple microbial functional groups that affect complex biogeochemical transformations of carbon and nutrients.Our study site is Stordalen Mire, a permafrost site in northern Sweden, with an extensive research history, including modeling (Chang et al 2019a(Chang et al , 2019b(Chang et al , 2020) ) and a rich observational record (Bolduc et al 2020).We chose this site because permafrost regions contain a large amount of organic carbon which is vulnerable to decomposition by soil microbes, releasing greenhouse gases including CH 4 and CO 2 (Tarnocai et al 2009).We hypothesized that incubation methods (i.e.sampling time, storage temperature, and storage duration) would influence CH 4 production during incubation by altering the initial incubation conditions, consequently impacting inference of the temperature sensitivity.To address our hypothesis, we developed a field-storage-incubation (FSI) modeling approach that mimics the full incubation process, including the timing of field soil sample collection, soil sample storage, and incubation.Then we evaluated our model and FSI approach through comparison with field observations and laboratory incubation measurements.After model validation, we conducted FSI simulation experiments where we varied storage duration, storage temperature, sampling time, initial substrate and microbial biomass conditions, and incubation temperatures to investigate how these factors affect CH 4 production and inferred Q 10 values.

Study site description
Stordalen Mire is a peatland situated in northern Sweden (68.35 • N, 19.05 • E).The climate in this area is subarctic with annual mean temperature of 0.07 • C and mean precipitation of 308 mm y −1   (Bäckstrand et al 2010).The fen, one of the three sub-habitats in the study site, is fully thawed and inundated, with large reported CH 4 emissions (McCalley et al 2014, Holmes et al 2022).On this site, sedges (Eriophorum angustifolium) are the dominant plant species.Due to recent permafrost collapse and increasing inundation, the fen area has increased by 100% from 1970 to 2014, increasingly offsetting the CO 2 sink at the Mire (Varner et al 2022).The Mire has been closely monitored since the 1970s and a comprehensive dataset of the site has been generated (Bolduc et al 2020).Peat cores were collected and transported for laboratory analysis including microbial analysis (Woodcroft et al 2018), biogeochemistry, and incubation experiments (Hodgkins et al 2014).

Measurements
Climate forcing data that is used to drive the ecosys model includes air temperature, precipitation, radiation, wind speed, and relative humidity.Forcing data prior to year 2014 are based on the GSWP3 reanalysis dataset and bias-corrected using long-term Abisko research station measurements at the Stordalen Mire (Chang et al 2019b).From 2014-2017, forcing data were obtained from the European center for medium-range weather forecasts atmospheric reanalyses (ERA5) (supplemental material).Terrestrial gas flux data, including CH 4 and CO 2 fluxes from 2011-2017, were recorded from auto-chamber systems onsite (McCalley et al 2014, Mondav et al 2014, Holmes et al 2022).Water table depth and thaw depth were also recorded and published (Crill et al 2023).Soil organic matter measurements are available at different depths at the autochamber sites (Hodgkins et al 2014).The incubation experimental data (i.e.CH 4 production) used for model testing in this study were also from Hodgkins et al (2014), which provides details on the site, storage, and incubation protocols.

Ecosys model
The ecosys model is a mechanistically-based terrestrial ecosystem model that couples hydrological, thermal, plant, and microbial dynamics and their exchanges with the atmosphere.Ecosys has been tested on multiple ecosystems including the Stordalen Mire site in this study (Chang et al 2019a(Chang et al , 2019b)).Ecosys has also been successfully tested against scenarios of environmental change, including warming experiments (Bouskill et al 2020) and elevated atmospheric CO 2 conditions (Grant 2013).Eleven microbial functional groups that regulate carbon, nitrogen, and phosphorus dynamics are explicitly represented in ecosys.Decomposition rates of different soil organic matter pools are a function of decomposer biomass and substrate concentration, and are also affected by soil moisture.The effect of temperature on decomposition rate is represented with a modified Arrhenius function which considers the inactivation of enzymes under high and low temperatures (Sharpe andDeMichele 1977, Grant 2014).This temperature sensitivity function does not consider other environmental constraints (e.g.soil moisture, substrate levels) beyond temperature, so we term it the intrinsic temperature sensitivity (Davidson andJanssens 2006, Wu et al 2021).We then calculate Q 10 values based on that intrinsic temperature sensitivity and compare them with the Q 10 values inferred from the simulated incubation experiments.Soil moisture and temperature are solved based on heat and water transfer schemes though canopy-snow-litter-soil profiles.Microbial respiration rates are represented based on Michaelis-Menten kinetics with influences from soil water potential, oxygen concentration, nutrient availability, and temperature.Ecosys represents CH 4 production (acetoclastic methanogenesis and hydrogenotrophic methanogenesis), and CH 4 oxidation (table S1 and figure S1).A detailed description of model structure, inputs, and outputs of ecosys can be found in the data availability statement.

The simulation experiment
Here we developed and applied an FSI simulation approach (figure 1).In this approach, we first use ecosys to model a soil profile under long-term field conditions to create a modeled soil core consistent with the field soil core.A soil layer is then extracted numerically from that modeled field soil core and used in the model incubation protocol, which mimics laboratory sampling, storage, and incubation procedures.This modeling approach creates reasonable initial conditions for the incubation simulations and allows us to evaluate the model against field observations and laboratory incubation measurements.
We apply this FSI approach in Stordalen Mire, using field observations and incubation measurements from the fen site.For the field simulations, we run ecosys for the field site from 1980 until 2017 using the climate forcing described above.Simulation results are then compared against field observations to determine the appropriate field simulation scenario.From this baseline field simulation a soil layer is numerically extracted on the modeled date, 15 June 2011, mimicking the field sampling.Then the model is run with that soil layer through the storage and incubation periods.The storage and incubation conditions of that soil layer follow the incubation experiment procedure in Hodgkins et al (2014).However, a preincubation was not performed in the simulation because we forced the modeled soil and headspace to have no oxygen at the beginning of the incubation (baseline scenario).In the storage and incubation phases, modeled plants are removed and therefore no fresh litter inputs occur, and temperature and humidity are held constant (figure 1).Anaerobic conditions in the model are maintained by setting oxygen concentrations in the soil and headspace to zero at the start of storage.High N 2 concentrations in the headspace are set to mimic the N 2 flushing in the initial incubation experiment.Simulated CH 4 production is then compared against incubation experimental results for model evaluation.
In addition to our baseline simulation, we simulated the following incubation scenarios: sampling where C 2 and C 1 are cumulative CH 4 production at incubation temperatures T 2 and T 1 , respectively.

Results and discussion
We next describe our model validation, which involved two steps: (i) comparison with time series of field measurements, and (ii) comparison with laboratory incubation experiments, for which both measured datasets have been previously published as described in methods and data.

Model validation
In addition to ecosys validation at the same site using field observational data from earlier years (i.S2).
Finally, modeled soil organic carbon content (440-460 g kg −1 ) also compares well with observations (466 g kg −1 ).These comparisons give confidence that ecosys reasonably captures the thermal-hydrological state and carbon cycling dynamics at the site.Also, these comparisons give us confidence that the numerically extracted soil for the FSI simulation experiments broadly matches the actual soil samples used in the laboratory incubation experiments.

Time-dependent CH 4 production, biogeochemistry, and microbial biomass during incubations
Modeled CH 4 production by acetoclastic methanogenesis and hydrogenotrophic methanogenesis both decrease with time during the baseline incubation simulation (figure 3(a)).To disaggregate the We hypothesized that DOC and microbial biomass concentrations at the beginning of the incubation experiment were important controllers for subsequent CH 4 production.To explore this idea, we performed two simulation scenarios with increased levels of initial DOC (factor of ten) and microbial biomass (factor of ten) of all microbial functional groups (figures S3 and S4).Both simulations have similar patterns in the relationship between CH 4 production rates, substrates, and active microbial biomass.Increased DOC leads to an initial increase in CH 4 production, driven by sustained high H 2 concentrations from fermentation and growth of fermenters and methanogens.This modeled increase in soil respiration rate from adding initial substrates has also been reported in incubation experiments (Bergman et al 2000, Pegoraro et al 2019).Notably, the effect on CH 4 production of increasing initial microbial biomass is about twice that of increasing initial DOC, averaged over 3 months.A tenfold increase in microbial biomass resulted in sustained high DOC concentrations due to increased hydrolysis of soil organic matter (SOM) to soluble DOC.In this scenario, we did not see an increase in microbial biomass, possibly due to high maintenance respiration rates (caused by the high microbial biomass) that lead to the senescence of microbes.Our results demonstrate that DOC concentration and active microbial biomass are limiting factors of CH 4 production in incubations.
The simulated biomass of active AM, HM, and fermenters all decreased during baseline incubations with the relative proportions of these functional groups remaining relatively constant (figure 3(c)).The methanotroph biomass remains low due to anaerobic conditions.We confirmed that modeled methanotrophs will grow when oxygen is added to the modeling system, as expected.

Storage duration
To test the effect of storage duration (at 4 • C) on CH 4 production, we conducted simulations for two additional storage duration scenarios (2 and 12 months), and compared with the default scenario (i.e. 7 month storage duration).While the results of the 12 month storage scenario resemble those of the 7 month scenario, distinct differences emerged for the 2 month storage scenario, suggesting that storage duration significantly affects cumulative CH 4 production rates during incubation (figures 4(a) and (b)).The 2 month scenario is similar to the scenario with added initial DOC described above (figure S3).The initial DOC concentration is much larger (about 2 times greater) when storage duration is 2 versus 7 months.The additional initial resource in the 2 month storage scenario leads to a rapid increase in CH 4 production rates, and a >40 d period to decline to rates comparable to the other storage duration scenarios (figure 4(b)).The rapid increase and subsequent decrease of CH 4 production rates are regulated by the interplay of substrates and microbial biomass (figures 4(c)-(h)).The rapid decrease in DOC between about days 3 and 20 is consistent with increased fermentation biomass and products including acetate and H 2 (figures 4(d), (e), and (h), C 6 H 12 O 6 + 2H 2 O-> 2CH 3 COOH + 2CO 2 + 4H 2 ).This quick decrease in DOC concentration also aligns with the rapid growth of AM and HM (figures 4(f) and (g)).The slow accumulation of DOC and acetate after 20 d is similar to patterns observed in other storage scenarios, as described above.
Recognizing the significance of initial incubation conditions, we extended the analysis to include the storage phase in both the 7 month and 2 month storage scenarios (figures S5 and S6) to investigate the reason for their distinct initial conditions.Although the initial storage conditions are identical for the 7 month and 2 month scenarios, the continuous depletion of DOC from prolonged storage results in a higher DOC concentration in the shorter 2 month storage scenario compared to the 7 month storage scenario.Additionally, the close correspondence between substrates and microbial biomass during storage mirrors the relationship shown in the incubation phase.This close correspondence might be attributed to microbial decomposition being the sole source of substrates, given the absence of growing plants and root exudates during both phases.
There is ongoing debate regarding whether storage duration affects incubation data analysis.and thereby high initial incubation DOC concentration (figure S7).Subsequently, the CH 4 production is much higher during incubation compared to storage at 4 • C under the same storage duration (figure S5).Based on these results we suggest either limiting storage duration or using a freezing storage temperature to minimize consumption of substrates during storage.Note that we focus on permafrost soils in this study where microbes are accustomed to freezing conditions.Introducing freezing temperature to warmer soil might lead to death of microbes and distorted interpretation of the incubation (Weiser Russell and Osterud Clarice 1945).

Sampling time
Sampling time (i.e. the time in the growing season when incubation material is collected from the field) influences CH 4 production during soil incubations and the effects on CH 4 production rate diminish over time (figures 5(a) and (b)).Notably, CH 4 production is larger during the incubation period when samples are taken in June, with production declining each sampling month until September.This effect is most prominent in the initial ∼20 d of the incubation period.The higher modeled CH 4 production rates for June samples are attributed to the higher availability of H 2 (i.e.substrate for HM, figure 5(e)) and active biomass of AM, HM, and fermenters (figures 5(f)-(h)) during this period.Tracing back to the storage phase, we found higher initial storage microbial biomass for all functional groups when sampling in September (figure S8) compared with sampling in June (figure S5), leading to more rapid DOC consumption and earlier mortality of microbes.As a Z Li et al result, in the initial phase of incubation, substrate concentrations in the September sampling scenario (compared to June sampling) are higher (from dead microbes) but active microbial biomass is lower, resulting in lower CH 4 production.
The conclusion above that earlier sampling time (i.e,June vs.September) leads to a higher cumulative CH 4 production during incubations is also valid for long-term incubations with −20 • C storage temperature (figure S9).However, for short-term incubations less than 10 d, CH 4 production rates are slightly higher when sampling in September.Experimental results from Lupascu et al (2012) show a progressive increase in CH 4 production rates when sampled from June to September for top 20 cm soil incubated at 4 • C. Bergman et al (2000) highlighted the significant impact of sampling time on incubation results, and showed that whether early season sampling time led to larger incubation CH 4 production varied across field sites.
Overall, our results demonstrate that sampling time can significantly affect incubation CH 4 production.However, the intricate interplay between substrates and microbial biomass, coupled with the initial incubation conditions influenced by storage temperature and duration, adds complexity to understanding how sampling time affects CH 4 production during incubation.

Sampling time, storage duration, and incubation duration affect inferred temperature sensitivity
As described in methods, we inferred Q 10 values using cumulative CH 4 production in the same manner as is typically done from laboratory incubations (i.e. the equal-time method).We hypothesized that sampling  S10) would all affect inferred Q 10 values.Across all these cases, we inferred Q 10 values between 1.2 and 3.5 (figure 6).These results align well with published Q 10 values from laboratory incubation experiments on permafrost soils.For example, Lupascu et al (2012) reported values of 1.9-3.5 at a sedge site in Stordalen Mire based on laboratory incubation experiments.Dutta et al (2006) reported a mean Q 10 value of 1.9 ± 0.3 for permafrost soil samples across tundra and boreal forest sites in Siberia, incubating at controlled temperatures of 5, 10, and 15 • C.
In general, inferred Q 10 values are largest for storage durations of 2 months, indicating that when substrate levels are high, the enhancement of CH 4 production due to warming is also higher.For the 2 month storage scenario, Q 10 values initially increase and then decrease with incubation duration, corresponding to the phases of microbial growth and death.We also inferred much higher Q 10 values under the simulation scenarios of higher initial DOC concentration and −20 • C storage temperature (figure S11).Except for the 2 month storage scenario, Q 10 values stabilize after incubating for ∼50 d.This effect of incubation duration on inferred Q 10 values has also been reported in incubation experiments (Reichstein et al 2005, Gudasz et al 2015).Additionally, our inferred Q 10 values tend to be higher at lower temperature pairs (e.g. using 4 • C and 11 • C) compared to using higher temperature pairs, consistent with prior findings (Zhou et al 2009).
Using the modeled CH 4 production results to infer Q 10 values allows a comparison to the intrinsic temperature sensitivity coded in the model.Ecosys uses a modified Arrhenius function (Sharpe andDeMichele 1977, Grant et al 1993) to infer the effect of temperature on microbial activity (termed the intrinsic temperature sensitivity).Because substrate concentrations, nutrient constraints, oxygen concentrations, and pH also affect activity rates in the model (and real world), the Q 10 values inferred from emergent CH 4 production are expected to be different from the intrinsic value (Davidson et al 2006).This issue is important for model development, since it is common practice to use laboratory inferred Q 10 values in biogeochemical models without accounting for all the other factors which affect CH 4 production rates already included in the model.The intrinsic Q 10 values in ecosys range from 2.1 to 7.4 based on the prescribed temperature pairs (figure 6), decreasing as the incubation temperature pairs increase, and are a factor of about 1.4-2.5 higher than the inferred values (calculated using the mean for each temperature pair shown in figure 6).Thus, a model that used an inferred Q 10 value from a laboratory incubation, and that also included other factors affecting respiration rates, could be substantially biasing low their predicted CH 4 production temperature sensitivities.
Further, the fact that the inferred Q 10 values depend on substrate concentrations, microbial biomass, and temperature, as discussed above, and that those factors are often dynamic in real ecosystems, has significant implications for applications in terrestrial ecosystem models.Frequently, incubation experiments last days to months, with the resulting Q 10 value calculated for the incubation duration.However, modelers often apply the inferred Q 10 value to represent transient (often hourly) changes in respiration rates caused by temperature (Gu et al 2004).Misunderstanding of inferred Q 10 values in terrestrial ecosystem models could lead to an underestimation of the response of soil respiration to warming, especially in cold regions (Zhou et al 2009).
Q 10 values inferred from incubation studies have been widely adopted and have contributed to understanding of ecosystem responses to change.While it may be acceptable for empirical ecosystem models lacking representations of substrates and microbes to use a temperature sensitivity response that includes all those external effects, those models miss key mechanisms and are less favorable for simulating carbon-climate feedbacks with environmental change (Sulman et al 2018, Chang et al 2021, Tang et al 2023).Moving forward, particularly for process-based terrestrial ecosystem models that represent microbial activities and substrate dynamics, we recommend the use of mechanistic temperature sensitivity representations of component processes that contribute to the overall emission instead of a constant Q 10 value for CH 4 production or emission (Davidson et al 2006, Tang andRiley 2020).The characterization of such temperature sensitivities can be informed by theoretical studies and incubation experiments.When characterizing temperature sensitivities based on incubation experiments, we suggest using short-term incubations with abundant substrates, avoiding storage (or above-freezing conditions), and measuring other factors known to affect CH 4 production rates (e.g.substrate levels, microbial biomass).The exact form of the temperature sensitivity representation is still under discussion.Some models, such as the modified Arrhenius function by Sharpe and DeMichele (1977), incorporate the high and low temperature inactivation of enzymes in organism activity.The macromolecular rate theory model, which accounts for the change in heat capacity associated with the transition between the enzyme-substrate complex and the enzyme-transition state, has recently been proposed (Alster et al 2016, 2020, Liang et al 2018), as have other approaches based on chemical kinetics (e.g.Tang and Riley 2024).Other studies argued that the emergent temperature response also depends on dynamic interactions between mineral surfaces and substrates, substrate lability, enzymes, and microbes (Tang and Riley 2014).

Conclusions
We applied observations and a well-tested processrich model, ecosys, to interpret laboratory incubation observations and investigate controls on inferred temperature sensitivity (i.e.Q 10 ) of methane (CH 4 ) production.A field-storage-incubation (FSI) simulation approach was developed to mimic the incubation process.The ecosys model and the simulation approach were benchmarked using observations of field CH 4 emissions and incubation CH 4 production.Incubation simulation results show that dynamic CH 4 production rates are regulated by the interplay between substrates (DOC, acetate, and H 2 ) and activities of acetoclastic and hydrogenotrophic methanogens and fermenters.Using the model, we found that storage duration, storage temperature, and sampling time affect CH 4 production through interactions with substrates and microbial biomass.
Our findings explain how the inferred temperature sensitivity of CH 4 production is affected by incubation duration, incubation temperatures, storage duration, storage temperature, and sampling time.The inferred Q 10 values were substantially lower than the intrinsic temperature sensitivity used in the model because other factors (e.g.substrates and microbial biomass) also affect CH 4 production.In other words, Q 10 values of CH 4 production and emission are regulated by a complex interplay of biological, biochemical, and physical processes.This interaction leads to the aggregated Q 10 being different than those of the component processes.Terrestrial ecosystem models relying on a constant Q 10 value to characterize temperature responses may therefore predict biased soil carbon cycling under future climate scenarios.Our framework for simulating incubations and the associated findings provide valuable insights for interpreting incubation observations and lead us to propose hypotheses relating to the effects of storage duration and sampling time on Q 10 that could be tested in incubation experiments.Further, our work emphasizes the need to accurately measure important auxiliary variables such as substrate availability and active microbial biomass during incubation experiments to improve mechanistic understanding and modeling of carbon cycling responses to warming.

Z
Figure 1.A schematic representation of the three-phase FSI modeling approach: (1) field, (2) storage, and (3) incubation.The baseline scenario includes simulation under field conditions until sampling on 15 June 2011, sample storage for 7 months at 4 • C, and sample incubation at 22 • C for 90 d.
CH 4 production during the modeled baseline incubation experiments broadly matched the observations from Hodgkins et al (2014) (figure 2(b); RMSE = 21 umol (g dry) −1 ), with modeled and observed gas production both accumulating more slowly with time, in line with many published incubations studies, which show that respiration rates decline after the first few days as fast-cycling carbon is depleted and slow-cycling carbon becomes the main contributor to microbial activity (Fang et al 2005, Schädel et al 2020).

Figure 2 .
Figure 2. (a) Modeled CH4 emissions closely align with field observations at the fen site for the period 2011-2017.Solid blue lines represent simulation results, while markers denote measurements.Error bars indicate the standard deviation of daily CH4 emission measurements obtained from two autochamber systems.(b) Modeled cumulative CH4 production is consistent with incubation measurements reported by Hodgkins et al (2014).The solid blue line represents simulation results and the markers represent measurements.Error bars in (b) depict the standard deviation of CH4 production measurements in incubation experiments across three soil samples from the same site as the field observations.
Our results of microbial biomass composition align with results in Wilson et al (2021) that microbial relative abundances were not changed significantly during incubations of fen samples at Stordalen Mire.However, accurately measuring active microbial biomass in experimental incubations has been challenging (Blagodatskaya and Kuzyakov 2013), and debates persist about microbial changes during storage and incubation periods (Stenberg et al 1998).Stenberg et al (1998) found (a) microbial biomass decreased significantly at 2 • C, as estimated by the chloroform fumigation-extraction method and (b) small changes in biomass estimated by the substrate induced respiration method.Microbiome 16 S rRNA copy number, an indicator of microbial biomass, has been reported not changing significantly during incubation (Wilson et al 2019, Fofana et al 2022).
Stenberg et al (1998) found that microbial biomass and activity in soil are influenced by storage conditions, with storage duration being less important when soils are stored at freezing (−20 • C) compared to refrigeration (2 • C) temperatures.However, other studies found that microbial biomass carbon and enzyme activities were not substantially affected when stored at 4 or −20 • C (Lee et al 2007, Wilson et al 2021).Our results indicate that 4 • C storage duration significantly affects CH 4 production through its effects on initial incubation substrate concentrations, which then affect microbial activity and biomass.The simulated storage temperature of −20 • C led to negligible microbial activity during storage

Figure 4 .
Figure 4. Storage duration affects CH4 production during incubation through regulating substrate concentrations and microbial biomass.Legend labels refer to storage duration.Comparisons are made for storage durations of 2, 7, and 12 months for: (a) cumulative CH4 production, (b) CH4 production rate, (c) DOC concentration, (d) acetate concentration, (e) H2 concentration, (f) AM biomass, (g) HM biomass, and (h) fermenter biomass.Initial conditions of the storage phase, storage temperature (4 • C), and incubation conditions are the same for all the storage duration scenarios.Except for storage duration, all the simulation conditions are the same as the default scenario.

Figure 5 .
Figure 5. Sampling time affects CH4 production during incubations through regulating the initial substrate concentrations and microbial biomass.Comparisons are made for sampling times of June 15, July 15, August 15, and September 15, denoted as JUN, JUL, AUG, and SEP, respectively.Results show (a) cumulative CH4 production, (b) CH4 production rate, (c) DOC concentration, (d) acetate concentration, (e) H2 concentration, (f) AM biomass, (g) HM biomass, and (h) fermenter biomass.

Figure 6 .
Figure 6.Q10 values vary with incubation duration (x-axis), temperature pairs (y-axis), (a) sampling time, and (b) storage duration.Temperature pairs represent the two temperatures chosen for calculating Q10 values using the equal-time method (equation (1)).Note Q10 values in the first (June 15) and sixth (7 months) groups of columns are identical because those Q10 values are inferred from the same default simulation with sampling time of June 15 and storage duration of 7 months.(c) Microbial growth and maintenance Q10 values calculated based on the modified Arrhenius function used in ecosys.
Alster C J, Baas P, Wallenstein M D, Johnson N G and von Fischer J C 2016 Temperature sensitivity as a microbial trait using parameters from macromolecular rate theory Front.Microbiol.7 1821 Alster C J, von Fischer J C, Allison S D and Treseder K K 2020 Embracing a new paradigm for temperature sensitivity of soil microbes Glob.Change Biol.26 3221-9 Bäckstrand K, Crill P M, Jackowicz-Korczyñski M, Mastepanov M, Christensen T R and Bastviken D 2010 Annual carbon gas budget for a subarctic peatland, Northern Sweden Biogeosciences 7 95-108 Bergman I, Klarqvist M and Nilsson M 2000 Seasonal variation in rates of methane production from peat of various botanical origins: effects of temperature and substrate quality FEMS Microbiol.Ecol.33 181-9 Blagodatskaya E and Kuzyakov Y 2013 Active microorganisms in soil: critical review of estimation criteria and approaches Soil Biol.Biochem.67 192-211 Bolduc B et al 2020 The IsoGenie database: an interdisciplinary data management solution for ecosystems biology and environmental research PeerJ 8 e9467 Bouskill N J, Riley W J, Zhu Q, Mekonnen Z A and Grant R F 2020 Alaskan carbon-climate feedbacks will be weaker than inferred from short-term experiments Nat.Commun.11 5798 Bridgham S D, Cadillo-Quiroz H, Keller J K and Zhuang Q 2013 Methane emissions from wetlands: biogeochemical, microbial, and modeling perspectives from local to global scales Glob.Change Biol.19 1325-46 Chang K-Y et al 2021 Substantial hysteresis in emergent temperature sensitivity of global wetland CH4 emissions Nat.Commun.12 2266 Chang K-Y, Riley W J, Brodie E L, McCalley C K, Crill P M and Grant R F 2019b Methane production pathway regulated proximally by substrate availability and distally by temperature in a high-latitude mire complex J. Geophys.Res.Biogeosci.124 3057-74 Chang K-Y, Riley W J, Crill P M, Grant R F and Saleska S R 2020 Hysteretic temperature sensitivity of wetland CH4 fluxes explained by substrate availability and microbial activity Biogeosciences 17 5849-60 Chang K-Y, Riley W J, Crill P, Grant R F, Rich V and Saleska S 2019a Large carbon cycle sensitivities to climate across a permafrost thaw gradient in subarctic Sweden Cryosphere 13 647-63 Craine J M, Fierer N and McLauchlan K K 2010 Widespread coupling between the rate and temperature sensitivity of organic matter decay Nat.Geosci.3 854-7 Crill P, McCalley C and Abisko Field Crew 2023 Manual active layer and and water table depth measurements from the autochamber sites at Stordalen Mire, northern Sweden (2003-2017) Zenodo https://doi.org/10.5281/zenodo.10420396 Davidson E A and Janssens I A 2006 Temperature sensitivity of soil carbon decomposition and feedbacks to climate change Nature 440 165-73 Davidson E A, Janssens I A and Luo Y 2006 On the variability of respiration in terrestrial ecosystems: moving beyond Q10 Glob.Change Biol.12 154-64 Dutta K, Schuur E A G, Neff J C and Zimov S A 2006 Potential carbon release from permafrost soils of Northeastern Siberia Glob.Change Biol.12 2336-51 Fang C, Smith P, Moncrieff J B and Smith J U 2005 Similar response of labile and resistant soil organic matter pools to changes in temperature Nature 433 57-59