MOISTURE AND TEMPERATURE SENSITIVITY OF CH4 OXIDATION IN BOREAL SOILS

Summar-We used laboratory experiments to evaluate CH4 uptake kinetics and the influence of soil moisture and temperature on rates of CHd-oxidation by boreal soils at in situ CH4 concentrations. Two upland forest sites (AS2 and BS2) were atmospheric CH., sinks; a bog site (LB) was an atmospheric CH4 source characterized by distinct depth zonation of CH4 production and consumption. Apparent half-saturation constants (K,) for CH4-oxidation showed relatively well-adapted communities. The KS for the high CH4-source soil (LB) was 1.1 PM, about IO-fold higher than values for CHd-sink soils (AS2 and BS2), 37 and 124 nM. Experiments assessing the individual effects of moisture and temperature on CHd-oxidation indicated that moisture was the primary control in CH4-sink soils at AS2 and BS2, while temperature was more important in CH4-source soil at LB. A combination of the highest moist- ure content and lowest temperature for each soil gave the lowest CH4-oxidation rates in experiments evaluating the interactive effects of these two variables. Conversely, a soil moisture content close to the optimum identified in moisture dependence experiments combined with the highest soil temperature consistently gave the highest CH4-oxidation rate. Copyright 0 1996 Elsevier Science Ltd Our laboratory study was aimed at assessing CH4 uptake kinetics and the moisture and temperature sensitivity of CH4-oxidizing microbial communities in boreal soils. Further, we have examined the interactive effects of moisture and temperature on rates of CH4-oxidation in these soils. Our objective was to better understand the physiological characteristics of CH4-oxidizing communities in boreal soils and important


INTRODUCTION
A contemporary increase in the concentration of atmospheric CH4 is well-documented (Houghton et al., 1995) and the magnitudes and ranges of major budgetary terms are becoming better refined (e.g. Fung et al., 1991). Within this context, there is a growing awareness of the need to augment source strength studies with an improved understanding of atmospheric and biospheric CH4 chemistry and their linkages (Prinn, 1994).
Microbial CH4-oxidation in aerobic soils and sediments adjacent to anaerobic zones of production has long been acknowledged as a modulator of CH4 flux to the atmosphere (reviewed by King, 1992;Mancinelli, 1995;King, 1992;Mancinelli, 1995). However, recent field studies point to an additional role of CHd-oxidation in the atmospheric CH4 budget. Well-drained, geographically heterogeneous soils show a widespread, consistent ability to serve as atmospheric CH4 sinks (Seiler et al., 1984;Steudler et al., 1989;Born et al., 1990;Keller et al., 1990;Whalen and Reeburgh, 1990a;Yavitt et al., 1990Yavitt et al., , 1995Crill, 1991;Mosier et al., 1991Mosier et al., , 1993Whalen et al., 1991;Striegl et al., 1992;Koschorreck and Conrad, 1993;Tate and Striegl, 1993;Castro et al., 1993Castro et al., , 1994Castro et al., , 1995 Reiners, 1994). Tropospheric reaction with hydroxyl radicals (OH) was once considered to be the only important sink for atmospheric CH4 (Cicerone and Oremland, 1988). However, studies of CH4-oxidation in well-drained soils point to a sink strength that is about 7% of the annual global destruction by OH radicals (D&r et al., 1993). Methane-oxidation is important in the atmospheric CH4 budget. Reeburgh et al. (1993) estimate that nearly 60% of the 1188 Tg CH4 produced y" from anthropogenic and natural sources is oxidized microbially adjacent to the zone of production or in soils functioning as atmospheric CH4 sinks. Processlevel studies (Whalen et al., 1990;Nesbit and Breitenbeck, 1992;Adamsen and King, 1993;Dunfield et al., 1993;King and Adamsen, 1992) and many of the field investigations cited above have identified soil moisture and temperature as key controls on CH4-oxidation in soils. The relative importance and interactive effects of these regulators remains uncertain. Moreover, the kinetics of CH4 uptake by soil microbes subsisting on atmospheric CH4 are poorly characterized.
The boreal forest is circumpolar and occupies > 8% of the earth's continental surface (Lieth, 1975). This region presently plays a major role in the atmospheric CH4 budget. It is characterized on one extreme by well-drained forest soils that consume atmospheric CH4 (Whalen et al., 1991) and on the other by poorly-drained bog soils that are point sources of CH4 emission to the atmosphere (Whalen and Reeburgh, 1990b). The future role of the boreal I272 S. C. Whalen and W. S. Reeburgh forest in the atmospheric CH4 budget is uncertain. Geophysical climate change models predict significant alterations in temperature and precipitation for this region (Mitchell et al., 1990). Accurate prediction of future trends in soil  in this and other regions as a result of altered climate and land use change requires an improved understanding of the physiology and controls of this microbial process.
Our laboratory study was aimed at assessing CH4 uptake kinetics and the moisture and temperature sensitivity of CH4-oxidizing microbial communities in boreal soils. Further, we have examined the interactive effects of moisture and temperature on rates of CH4-oxidation in these soils. Our objective was to better understand the physiological characteristics of CH4-oxidizing communities in boreal soils and important factors that regulate CH4-oxidation.

MATERIALS AND METHODS
Following from our aims, we selected soils from sites that we expected to represent extremes with respect to CH4-oxidizing communities. Two sites were in well-drained forest soils that were atmospheric CH4 sinks (Whalen et al., 1991). The other site was a peat bog that showed a net CH4 flux to the atmosphere and had distinct soil zones of CH4 production and oxidation (S. C. Moosavi, unpubl. MS. thesis, University of New Hampshire. 1994). Our goal in moisture and temperature dependence experiments was to evaluate the response of CHJoxidation at in situ CH4 concentrations.
Therefore, experiments involving upland forest soils were done at the present atmospheric CH4 concentration of 1.7 j.11 1-l CH4 (i.e. free-air atmosphere) and experiments with bog soil were made in a CH4-amended atmosphere that gave zero-order (substrate-saturated) rates of CH4-oxidation.  (Whalen et al., 1991). Thus, the CHd-oxidizing community is normally exposed to CH4 concentrations at or below the atmospheric level (-1.7 u1 I-'). A complete site description is given in Van Cleve et al. (1991).

Site description and sample collection
Lemeta Bog (LB) is a poorly-drained, open stand of stunted black spruce with an undergrowth of Sphagnum invaded by Labrador tea (Ledum paiustre), cloudberry (Rubus chamaemorus) and lowbush cranberry. The maximum thaw depth is < 1 m with peat extending to permafrost.
Most areas within the bog show net CH4 emission to the atmosphere (Moosavi, 1994). Samples were collected from drained, aerobic peat overlying a water-saturated anaerobic zone of methanogenesis.
Thus, CH4 oxidizers are normally exposed to uptake-saturating concentrations of CH+ This site is fully described in Luken and Billings (1985) and Funk et al. (1994).
Bulk soil samples were collected from the zone of maximum CH4-oxidation (IO-20 cm at AS2 and BS2; 20-30 cm at LB) with a trowel (BS2 and AS2) or serrated knife (LB), after removal of topsoil with a stainless-steel coring apparatus (AS2 and BS2; Whalen et al., 1992) or serrated knife (LB). Samples were returned to the laboratory within 2 h. Individual samples from each site were cornposited (-5 kg total), sieved (AS2 and BS2 only; ~4 mm grain size), homogenized with a garden trowel, and stored at 4°C in foil-covered polyethylene tubs until experiments were begun (within 24 h of sample collection).

Soil physical and chemical properties
Physical and chemical properties were determined for each soil composite according to standard methods (Carter, 1993). Soil moisture was measured gravimetrically by oven-drying at 105°C (AS2 and BS2) or 85°C (LB) and organic content was determined by loss on ignition (SSO'C) of oven-dried samples.
Gravimetric water holding capacity (WHC) was determined as the difference in weight between water-saturated and oven-dried soils. Soil pH was measured potentiometrically on soil-deionized water pastes (ratio I:1 to 1:2).

Experiments
were performed in -l-liter Mason jars modified fov syringe sampling of the headspace gas (Whalen et aI., 1990). Equal-sized ahquots of homogenized soil composites (100 to 200 g fieldmoist samples; mass dependent on site) were dispensed into Mason jars. A soil property (e.g. moisture content) was adjusted within each jar to yield one or two jars exposed to each treatment condition and the microbial communities were conditioned for 8-12 h in a free-air atmosphere to the altered environment.
All experiments involved time-course measurement (every 0.5 to 1 h) of the headspace CH4 concentration in jars incubated statically for 6 to 12 h. Jar headspaces for AS2 and BS2 soils were equilibrated with a free-air atmosphere (ml .7 pl CH4 1-l) to approximate in situ CH,, concentrations and sampling commenced immediately upon sealing. Headspace CII4 concentrations in jars with LB soil were adjusted to concentrations previously determined to give zero-order CH4 consumotion (amended atmosphere experiments; + 600 to @O pl CH4 1-l) and to simulate the in situ CH,, concentration. Zero-order (substrate-saturated) oxidation was indicated by a linear time-course for decrease in the headspace CH4 concentration. These soils were conditioned at the selected moisture or temperature conditions for an additional 2 h after CH4 addition before zero time (To) headspace samples were withdrawn.
The moisture response of CHd-oxidation at room temperature (-25°C) was determined at soil moisture contents above and below those of field-moist soils. Soil moisture was incrementally increased by application of a fine mist of deionized water to the soil with simultaneous stirring. Moisture content was decreased by spreading a thin layer of soil over a foil sheet followed by exposure to ambient or slightly elevated (5°C) temperatures in a convection oven. Soils were periodically consolidated, weighed and redistributed to promote even drying over a 2 to 6 h period. Soil moisture content is expressed as a percentage of water holding capacity (% of WHC) to facilitate comparisons of relative degree of wetness for and among soil samples.
The substrate-saturated temperature response of CH4-oxidation (LB soil) was evaluated at roughly 7°C increments from 5°C to 42°C. We did not expect a pronounced temperature influence on CH4oxidation rates at atmospheric CH4 concentrations (AS2 and BS2 soil). To better detect a response, we eliminated any influence of jar-to-jar variability in microbial community composition and activity by assessing CH4-oxidation in duplicate samples from each site that were exposed to 4, 16 and 28°C (random order) over three successive days.
The large number of samples required to assess the interactive effect of soil moisture and temperature on CH4-oxidation forced us to limit the scope of these experiments. The interactive effect was determined only across the range of soil temperatures (~4 to 20°C) and % of WHCs (-30 to 70%) normally encountered in the field at each site. In contrast, experiments assessing the individual effects of these variables spanned a greater range of values.
Progress curves (17 to 37 h) for CH4 depletion versus time were used to determine values of the kinetic parameters for CHd-oxidation (K, = apparent half-saturation constant; I',,,,, = maximum oxidation rate) by the microbial consortium at each site. Experiments were conducted at room tempera-ture on a single field-moist soil composite from each site.

Methane analysis
Headspace CH4 determinations were made by flame ionization detection gas chromatography with a precision of < 1% (Whalen et al., 1991). Calibration gases are relatable to mixtures obtained from the National Institute of Standards and Technology.

Statistical analyses and calculations
Substrate-saturated rates of CHroxidation (ng CH4 g" dw of soil; dw = dry weight) in moisture, temperature and moisture-temperature experiments involving LB2 soil were calculated from linear regression of headspace CH4 concentration versus time for amended atmosphere experiments. First-order rate constants for CH4-oxidation, k (h-l), were calculated from linear regression of lntransformed concentration versus time data in experiments involving a free-air atmosphere (AS2 and BS2 soils). To facilitate comparison of CH4-oxidation rates among sites, these data are expressed as the rate of CH4-oxidation at 1.5 ul CH4 1-l (ng CH4 g-' dw of soil). This is a typical concentration in the air-filled pore space near the soil surface at AS2 and BS2 (Whalen et al., 1992).
Values for kinetic constants for CHroxidation (KS and V,,,; defined above) were estimated from the integrated solution for the Michaelis-Menten model (Robinson, 1985): where S, and S, are the headspace CH4 concentrations at start of the experiment and at time = t, respectively. Data for CH4 concentration versus time in progress curve experiments was directly fitted to equation (1) by a quasi-Newton method, using the SYSTAT program package (SYSTAT Inc. Evanston, Illinois). Values of K, are expressed as aqueous-phase concentrations (nM) by applying Bunsen solubility coefficients (Yamamoto et al., 1976) to gas-phase CH4 concentrations.
Statistical analyses (tl = 0.05) follow Zar (1984). Experiments assessing community response to multiple treatment conditions for a manipulated variable through time-course analysis of change in substrate concentration often involve compromises with regard to sampling frequency and replication. This is particularly true where rates of activity cannot be determined from time-linear observations, as in experiments assessing k at very low substrate concentrations. We often directed our efforts toward firmly establishing the CH4-oxidation rate in a single sample at each treatment condition by making multiple observations of headspace CH4 concentration closely spaced in time for each time-course experiment used in rate determinations. Values for Y' in linear regression equations used to determine substrate-saturated CHd-oxidation rates or k exceeded 0.95 in nearly all cases. Use of relatively large, homogenized soil samples (100 to 200 g) for each treatment minimized the potential effect of sample-to-sample differences in the size of the microbial community on CH4-oxidizing activity. In cases where CH4-oxidation rate determinations were made on replicate samples, coefficients of variation averaged only 6.8%. In concert, high values of r* for regression equations used to determine CH4-oxidation rates and low coefficients of variation when rate determinations were made on replicate samples suggest that a single rate determination gave a true measure of the response of CH4-oxidation to the manipulated variable. Associations between manipulated variables and rates of CH4 consumption calculated from time-courses were analyzed by best fit relationships (highest r*), which were either linear or polynomial regressions.

Soil properties
Some soil properties differed between the forest and bog (Table I). Bog soil (LB) was lower in pH and higher in organic content and WHC compared with forest soils (AS2 and BS2). However, soils at BS2 and LB2 were most similar with respect to % of WHC at the time of sample collection, as data ranged from about 40-80% at both sites, vs 13-46% at AS2. Gravimetric moisture content increased with increasing percent organic matter, with data averaging about 11, 40 and 500% at AS2, BS2 and LB.

Moisture response experiments
Methane-oxidation rates in moisture response experiments for the two forest soils were similar (Fig. 1). Data ranged from 0.01 to 0.92 ng CH4 g-' dw h-' for AS2 (n = 7) and from 0.01 to 1.35 ng CH4 gg' dw h-' for BS2 (n = 6). Rates of CH4-oxidation were generally lOOO-fold higher in LB soil than in the forest soils, due to the higher concentration of headspace CH4, larger communities of CH4-oxidizing bacteria or both (Fig. 1). Data ranged from 360 to 1330 ng CH4 g-' dw h-' (n = 7). A polynomial function provided the best fit to the moisture (M) response of the CHroxidation rate (V) at each site. These relationships give opti-mum water contents (% of WHC; dV/dM = 0) of 21, 27 and 50% for CH4-oxidation in AS2, BS2 and LB soils, respectively.
The % of WHCs for fieldmoist soils were 13, 45 and 45% for AS2, BS2 and LB, respectively, with the difference between the ambient and optimum % of WHC ranging from -8% at AS2 to + 18% at BS2.
Although the moisture response of CH4-oxidation may partly depend on the CH4 concentration, the study sites can be expected to experience a relatively narrow range of CH4 concentrations.
Soils at AS2 and BS2 are constantly exposed to sub-atmospheric CH4 concentrations ( < 1.7 t.d l-l), while LB soil is subject to CH4 concentrations that are more variable, but saturate the CH1-oxidizing capability of the microbial community. Hence, the data for % of WHC collected through the summers of 1993 and 1994 (Table 1) and the moisture dependence relationships for CH1-oxidation ( Fig. 1) were used for a one-dimensional analysis of the seasonal (June through August) influence of soil moisture on rates of CH1-oxidation in these soils. The moisture response of CH4-oxidation is predicted to range over factors of 2.6, 5.0 and 1.6 at AS2, BS2 and LB.

Temperature response experiments
Site-wise comparison of CH4-oxidation rates in temperature response experiments shows that rates were most similar for the two forest soils, in agreement with results of moisture response experiments. Nonetheless, rates were about four-fold higher at BS2 than AS2, with data ranging from 2.71 to 4.29ngCH4g-' dw h-' (n = 6) and 0.73 to 1.18 ng CH4 g-' dw h-' (n = 6) at the two respective sites (Fig. 2). Rates of CH4 oxidation were 2-3 orders of magnitude higher in LB soil (range: 2727-4473 ng CH4 g-' dw h-t; n = 14) than in forest soils (Fig. 2) also in agreement with moisture dependence data. The data indicate a linear response in CH4 consumption to changing temperature in a free-air atmosphere for forest soils. Slopes for regression equations were significantly different between forest soils, indicating a greater temperature influence on CH4-oxidation for soil from BS2 than AS2. A polynomial provided the best fit to the temperature (7") response for CHd-oxidation (V) under substrate-saturated conditions for LB soil (Fig. 2). However, the temperature range tested was greater than the range investigated for forest soils. The data indicate an optimum temperature, Topt, (dV/dT = 0) of about 23°C for substrate-saturated CH4-oxidation in LB soil. Soils at the depth of sample collection varied in temperature from 621°C at AS2, 418°C at BS2 and 4-13°C at LB during June through August. Thus, Tent for LB soil was 10°C higher than the maximum in situ soil temperature. The relationships from Fig. 2 indicate that the temperature response of CHd-oxidation will range over factors of 1.5 and 1.4 for AS2 and BS2 soil. Data from Fig. 2 suggest a linear relationship between CH4-oxidation rate and temperature (V = 30.7T + 138; r* = 0.96; n = 6) for LB soil over the range 4413°C; the temperature response of CH4-oxidation will range over a factor of 2.1. Thus, soil moisture exerted greater control on rates of CHa-oxidation than temperature in the forest soils over the range of values observed for these environmental measures during this investigation. Conversely, soil temperature was more important than moisture in regulating CH4 consumption in the bog soil. Methane oxidation rate as a function of temperature for boreal soils. Methane oxidation rates for AS2 (top) and BS2 (middle) soils were calculated for 1.5 ul CH4 I-' from first-order rate constants, while CH4 oxidation rates for LB soil (bottom) were directly measured in time-course experiments. A linear (AS2 and BS2) or polynomial (LB) function is fitted to the data from each site.

Moisture-temperature interactions
Results of experiments assessing the interactive effects of moisture and temperature on CH4-oxidation over the general ranges of values observed in situ for these variables at each site (Fig. 3) agree qualitatively with results from experiments evaluating their individual effects (Figs 1 and 2). Rates of CH4-oxidation at each site are within the range observed for experiments described above, with rates for LB soil > lOO-fold higher than rates for forest soils. The data for AS2 and BS2 indicate a modest decline in the CH4-oxidation rate with decreasing temperature over the entire range of % of WHC, but a rapid decrease in the CH4-oxidation rate as soil moisture approached the highest values of % of WHC observed during this investigation. In contrast, the data for LB show a pronounced temperature effect on rates of CH4-oxidation and a less prominent decline in CH4 consumption at high values of % of WHC.
Maximum observed rates of CH4-oxidation in moisture-temperature experiments corroborate well the data from experiments analyzing the individual effects. The maximum rates of CHd-oxidation for AS2, BS2 and LB soil (Table 2) were 0.85, 1.10 and 706 ng g-' dw h-l, respectively. For each soil, the maximum oxidation rate occurred at the highest test temperature, as predicted in Fig. 2 The combined influence of adverse temperature and moisture conditions resulted in a severe reduction in rates of CH.+-oxidation for all soils. The minimum rate of CH4-oxidation was consistently observed with the combination of the highest % of WHC and lowest temperature tested ( Table 2). The ratio of maximum-to-minimum rate of CH4-oxidation in moisture-temperature dependence experiments gives an indication of the potential combined influence of these factors on CH4-oxidation during June through August at each site. Methane-oxidation rates may vary seasonally by factors of 42. I I and 3.9 at AS2, BS2 and LB due to moisturee temperature interactions.

Methane-oxidation kinetics
Rates of microbial activity depend on enzymatic processes and their environmental modulators. Our moisture manipulation experiments were made at headspace CH,, concentrations that approximate in situ amounts. Thus, they provide the first direct evidence for moisture sensitivity of CH4-oxidation rates in boreal soils that represent extreme environments with respect to in situ CH4 concentrations.  Methane-oxidizing bacteria at AS2 and BS2 are normally exposed to a free-air atmosphere, while those at LB are consistently subjected to uptakesaturating concentrations of CH4. These results (Fig. 1) imply a moisture sensitivity of CH4-oxidation by the local microbial communities in all boreal soils across a lOOO-fold range of in situ substrate concentrations.
Results from moisture dependence experiments at forest sites (Fig. 1) corroborate the results of field studies in soils that show only a zone of CH4 oxidation (no methanogenesis) and are atmospheric CH4 sinks. Seasonal investigations have reported an inverse correlation between soil moisture and CH4oxidation rates in forested (Keller et al., 1986;Steudler et al., 1989Whalen et al., 1991Koschorreck and Conrad, 1993;Lessard et al., 1994;Castro et al., 1995Sitaula et al., 1995 and unforested soils (Mosier et al., 1991(Mosier et al., , 1993Koschorreck and Conrad, 1993;Keller and Reiners, 1994), but these data necessarily include any temperature effect. No field data exist for moisture dependence of CH4-oxidation under uptake-saturating conditions. Methane is produced and oxidized simultaneously and in close proximity in these environments (e.g. Sundh et al., 1994) and the observed CH4 flux reflects the net effect of these two processes.
Laboratory studies of the moisture dependence of CH4-oxidation are few, often involve amended atmospheres and give variable results. In accord with our data, Adamsen and King (1993) showed an inverse relationship between gravimetric soil water content (range 20-50%) and CH4-oxidation by coniferous soil composites at a headspace concentration of 10 ul CH4 1-l while Yavitt et al. (1995) generally found that the lowest rate of CH4 consumption at both 51 and 1.7 pl CH4 1-l was associated with the highest soil moisture content (variable among samples) in composites of hardwood forest soils. In contrast, Nesbit and Breitenbeck (1992) reported a relative insensitivity of CH4-oxidation to soil moisture (25 and 75% water-filled pore space) for swamp and forest soils at 100 pl CH4 1-l. Differences in the moisture response of CH4-oxidation among these studies may be related to the physiological characteristics of the extant microbial communities and experimental CH4 concentration. King and Schnell (1994) reported a CH4 concentration-dependent response of NH&N inhibition of CH4-oxidation in forest soils normally exposed to a free-air atmosphere. Our experiments at in situ CH4 level-concentrations probably gave the most reliable measure of the moisture response of CH4-oxidizing microbial communities in the boreal soils tested.
The observed negative relationship between soil moisture and CH4-oxidation rates in many field studies involving atmospheric CH4 consumption is generally attributed to diffusion limitation of substrate supply as soil moisture increases (D&r et al., 1993). Data for AS2 (Fig. 1) suggest that water stress control of CH4-oxidation must also be considered. Increasing the % of WHC from a field-moist 13% to 25% resulted in a 22% increase in the CH4-oxidation rate. Field observations point to water stress control of CH4-oxidation under some circumstances. Dry tropical forest (Keller et al., 1990), desert (Striegl et al., 1992 and grassland (Mosier et al., 1991) soils all showed enhanced CH4-oxidation when soil water content was naturally or artificially increased. Physiological stress is an acknowledged, but poorly understood control of CH4-oxidation (Adamsen and King, 1993). Process-oriented studies of CH4-oxidation in soil should include water stress as well as diffusion limitation in moisture response experiments, as regional changes in climate may include a reduction in soil moisture content.
Present susceptibility of CH4-oxidizing bacteria to water stress may vary among environments, and altered soil moisture regimes under future climatic conditions may not have intuitively obvious effects on microbial consumption of atmospheric CH4. Percent of WHC for AS2 soil varied from 13 to 20% during 1993and from 26 to 46% in 1994. Local June through August precipitation in 1993 was 56% and 92% of the 30 year mean (NOAA, 1993(NOAA, , 1994. Thus, data in Fig. 1 suggest that water stress may be an important control on CH4-oxidation at AS2 only in an abnormally dry year under the present climate. However, a longterm reduction in summer soil moisture predicted for this region in some general circulation models (Mitchell et al., 1990) may lead to chronic water stress and a reduced sink strength for AS2 and similar sites. In contrast, % of WHC at BS2 exceeded the calculated optimum of 27% even during the abnormally dry summer of 1993 (Table 1) and Fig. I predicts that this site will increase in sink strength during a protracted period of reduced precipitation. Other forested areas may react to a long-term re-duction in soil moisture in a fashion similar to BS2. in the relative response of CH4-oxidation to tem- Castro et al. (1995) observed that the highest perature change among investigations are likely remonthly average rate of CH4 consumption in temlated to interactions among gas-phase transport, perate forest soils coincided with the lowest soil soil-water gas exchange and enzymatic activity, but moisture content recorded in 6 years of study and the data clearly point to a lesser role for temperasuggested that this region may increase in sink ture than moisture as a control of CH4-oxidation in strength with a reduction in soil moisture. a free-air atmosphere. A subsurface maximum of CH4-oxidizing activity in the mineral forest soil zone has been commonly observed (Whalen et al., 1992;Adamsen and King, 1993;Koschorreck and Conrad, 1993;Bender and Conrad, 1992;Schnell and King, 1994) and has been attributed to NH: inhibition in surface, organic soils . Avoidance of water stress may be an added benefit, as the only report of a surface maximum in CHd-oxidizing activity (Whalen and Reeburgh, 1990a) involved tundra soil that receives almost daily precipitation. Ecosystem level responses to altered climate may include a shift in the zone of maximum CH4-oxidizing activity to accommodate changes in soil moisture. Downward displacement will avoid cell desiccation as soils dry, but will also increase the diffusional path length for CH4. This potential response of CH1-oxidizing communities to drier conditions climate is unexplored and must be considered in assessing potential effects of altered climate.
The range in response of CH4-oxidation rates to soil moisture and temperature conditions that bracket field observations suggests that for the June through August period soil moisture (diffusion) is the primary environmental control on CH4-oxidation in our forest soils and temperature is of secondary importance. This is consistent with field data (Castro et al., 1994) from a temperate forest where a moisture-CH., flux relationship developed during a watering experiment at nearly constant soil temperature was used to accurately predict (r2 = 0.75) measured fluxes in previous seasons where soil temperatures ranged over -20°C. Other seasonal field studies show a weak temperature correlation for atmospheric CH4-oxidation across ecosystems (Steudler et al., 1989;Born et al., 1990;Crill, 1991;Koschorreck and Conrad, 1993;Castro et al., 1993Castro et al., , 1995Tate and Striegl, 1993;Crill et al., 1994;Sitaula et al., 1995), but these observations include a variable moisture effect.
The zero-order temperature response for CH4-oxidation by LB soil showed a greater relative rate increase for an incremental increase in temperature than soils from forest sites over the in situ range of soil temperatures (Fig. 2). These data point to greater enzymatic control of microbial CH4-oxidation in this high source CH* soil than in forest soil exposed only to atmospheric CH+ Accordingly, the range in response of CH4-oxidation rates to soil moisture and temperature conditions that span field observations indicates that for the June through August period soil temperature exerts a greater influence than moisture on CH4-oxidizing activity. The Tort of 23°C agrees closely with the 20 to 25°C reported by Dunfield et al. (1993) for CH4-oxidation in other subarctic peats incubated at 1000 ul CH4 I-', but is somewhat lower than zero-order values of 38°C and 31°C reported for cultures of Methylomonas rubra (King and Adamsen, 1992) and landfill cover soils (Whalen et al., 1990) respectively. These Topt values for natural soil samples all exceed the highest in situ summer soil temperatures.
Our moisture-temperature experiments (Fig. 3) were designed to assess interactions between enzymatic (temperature) and diffusive (moisture) control of CH4-oxidation as opposed to individual effects (Figs 1 and 2). Data for forest soils confirm the general dominance of diffusive control of atmospheric CH4-oxidation under field conditions and are in agreement with a similar laboratory study of moisture-temperature interactions (Schimel et al., 1993) where moisture controlled CHd-oxidation by upland tundra soil composites ranged from 5 to 15°C and from 25% to 200% of WHC.

Methane-oxidation
rates by soils from AS2 and BS2 increased by only 80 and 61% with a temperature increase from 4 to 28°C (Fig. 2). This linear and limited response of CH*-oxidation in a free-air atmosphere by our forest soils compares favorably with studies that have isolated the temperature effect on CH4-oxidation. Linear increases of about 35% (King and Adamsen, 1992) and 55% (Lessard et al., 1994) in the rate of CHa-oxidation by forest soils have been reported over the temperature ranges 0-30°C and 6-l 7°C respectively. Differences The shape of the three dimensional contour plot for CHA-oxidation in high source CHd bog soil (LB) differs distinctly from similar plots for atmospheric CH4 consumption by forest soils (Fig. 3). A greater temperature response and less precipitous decline at elevated soil moisture contents are evident. This corroborates results of experiments testing the individual effects of moisture and temperature (Figs 1 and 2) and provides further evidence for increased control of CH4-oxidation by enzymatic activity at LB and similar, high source CHI environments. High latitude bogs are important point sources of atmospheric CH4 (Whalen and Reeburgh, 1990) and this region is projected to experience warmer temperatures and a possible reduction in summer soil moisture under an altered climate (Mitchell et al., 1990). Our data (Fig. 3) predict accelerated rates of CHd-oxidation, largely in response to warmer tem-peratures. However, Qlos for substrate-saturated CHd-oxidation (1.3-2.3) in peat (Dunfield et al., 1993;Crill et al., 1994;Nedwell and Watson, 1995) are decidedly lower than Qlos for methanogenesis (5.3-20) in peat (Dunfield et al., 1993;Nedwell and Watson, 1995). Neglecting other potential ecosystem-level effects of altered climate on ecosystems (e.g. change in aerobic-to-anaerobic soil zones), CH4 emission from northern bogs may therefore increase as a result of the greater temperature sensitivity of methanogenesis than CHd-oxidation.
The apparent half-saturation constant for CH4oxidation in LB soil (Table 3) agrees well with similar data for microbial communities in high source CH4 environments.
Soils from a landfill cover (Whalen et al., 1990) and subarctic (Dunfield et al., 1993) and arctic (Whalen et al., 1996) peatland showed KS values in the 1.0-6.2 pM range. Temperate peat had somewhat higher KS values, 14 to 45 FM (Nedwell and Watson, 1995). Forest soil values of KS (Table 3) are in accord with the 30 to 51 nM reported for temperate cambrisols that consume atmospheric CH4 (Bender and Conrad, 1992).
The v,,, for LB soil was lower than previous reports for comparable environments.
Values for kinetic constants undoubtedly reflect complex interactions between the environment and enzyme systems. Moreover, V,,, depends on community vigor and size. Experimental design and curve fitting procedure also influence reported kinetic values (e.g. Robinson and Tiedje, 1982). The relatively narrow range in KS and V,,, values for three cambrisols (Bender and Conrad, 1992) under different management practices and these two boreal soils from different successional stages is therefore noteworthy. These two studies are the first to evaluate kinetic constants for high affinity (Bender and Conrad, 1992) CH4-oxidizing communities. Overall, values for kinetic constants in these boreal soils point to CHd-oxidizing communities that are finetuned to assimilate CH4 at in situ concentrations.
In summary, methanotrophs in subsurface zones of boreal soils are physiologically well-adapted for CH4 consumption over a broad range of in situ substrate concentrations. Soil moisture and temperature affect rates of CH4 consumption in these soils by influencing substrate supply and rates of enzyme activity, respectively. Rates of CHroxidation are more temperature-sensitive in high source CH4 bog soils and more moisture-sensitive in forest soils that consume atmospheric CH+ Our results give insight into the influence changing environmental conditions may have on CH4-oxidation in boreal soils. Ultimately, interactions among these factors will influence the locus and extent of the CH4-oxidizing zone in boreal soils and will determine the future importance of this region in the atmospheric CH4 budget.