Methane consumption and emission by Taiga

. Taiga or boreal forest environments are a poorly understood component of the global CH 4 budget. Results from a 1-year study of CH4 fluxes at a range of representative floodplain and upland taiga sites in the Bonanza Creek long term ecological research area show that soil consumption of atmospheric CH 4 was the dominant process. Methane emission occurred only sporadically in the earliest successional stages in the floodplain system; all other floodplain and upland sites were net CH 4 consumers. Our results suggest that upland and floodplain taiga soils are an atmospheric CH 4 sink of up to 0.8 Tg yr -1. Point-source bogs and fens are the only important CHn-emitting sites in taiga.


Sites
Our study was conducted at sites in the Bonanza Creek Experimental Forest (BCEF), a 5000-ha research area located 20 km west of the University of Alaska Fairbanks campus (Figure 1

Additional Measurements
A soil temperature profile was measured adjacent to one chamber at each site with a portable multithermistor probe (2-cm intervals), and the mean soil temperature to 15 cm was calculated.Soil moisture (w/w, oven-dried at 105øC) [Black, 1965] was determined on 2.8 cm diameter x 10 cm soil cores collected 1 m from each chamber.Soil organic content (w/w; loss on ignition at 550øC of oven-dried sample) and pH [Black, 1965] were measured on similar cores collected from each control plot in late September, 1990.These data are summarized in Soil CH 4 depth distributions were measured at upland sites NB2, SB 1, BS2 and UP3A on samples obtained by inserting a steel tube probe to known soil depths and pumping soil gas into Tedlar sample bags [Born et al., 1990] with a battery-powered diaphragm pump.The bags were sampled and analyzed for CH 4 in the laboratory.
Methane consumption thresholds and capacities were studied at the above four sites using chambers with ambient and amended atmospheres.The ambient atmosphere experiments involved allowing chambers with free air atmospheres (-1.7 ppm CH4) to "ran down" with periodic sampling over a 24 hour period.The amended atmosphere experiments involved adjusting initial chamber atmospheres to concentrations of ~20 ppm CH4 and sampling over a 24 hour period.
Methane fluxes were often below the limit of detection.We report the median and interquartile range (IQR; range of central 50% of data) of the CH 4 flux data as measures of central value and variability.This procedure is recommended by Helsel [1990] for analysis of censored (i.e., many values below detection limit) data.Nonparametric statistics [Conover, 1980;Zar, 1984]

RESULTS
Fluxes from the floodplain sites and upland sites are are presented in Figures 2 and 3, respectively.Table 2 gives -• ----• -------' 1    transition from CH 4 emission to consumption with increasing ecosystem maturity.There is no difference in CH 4 consumption between white spruce (FP4A) and black spruce (FP5A) or between black spruce and poplar (FP3A), but consumption is greater in white spruce than in poplar and in black spruce than in the alder-poplar (FP2A) stand.
Methane emission was never observed at the upland sites (Figure 3; Table 2).As in the floodplain sites, there was no clear seasonal signal for CH 4 flux.The earliest successional stage (bum site, Figure 3a) showed the smallest flux, with only 4% (3 of 76) of the flux determinations above the detectable level.The later successional stages, white spruce (Figure 3d) and black spruce (Figure 3f), exhibited higher rates of CH4 consumption than the middle stages of succession (aspen, Figure 3b; and birch, Figures 3c and 3e The soil pH at the study sites ranged from slightly basic to acidic (Table 1).The highest values of pH were observed in the first stage of succession (FP1A, UP1A), and the lowest values were found at the black spruce sites in both the floodplain (FP5A) and upland (BS2) areas.The soil organic content varied from 2 to 45% and from 9 to 37% in the floodplain and upland sites, respectively (Table 1), with the lowest value for each sequence occurring in the first successional stage.
The soil temperature during the observation period varied from about 2 ø to 21 o C in both the floodplain and upland sites (Table 2).Means varied from 10.5 ø to 15.7 øC for floodplain sites and from 8.9 to 12.9 *C for upland sites.The mean soil moisture content showed a higher range for upland sites (112%) than for floodplain sites (51%), due to the high mean soil moisture at BS2 (Table 2).The effect of soil temperature and moisture on CH 4 flux at each site was assessed by correlation analysis (Table 3) showed significantly increased consumption in the fertilized plot at AS2 (Figure 3b) and no difference between plots at BS2 (Figure 3f) and NB2 (Figure 3e).
The CH 4 concentration in soil atmospheres decreased rapidly with increasing depth at upland sites NB2, SB 1, BS2, and UP3A. Figure 5a  Nitrogen fertilization had no influence on CH 4 consumption at our upland taiga sites but resulted in significantly lower CH4 consumption rates in temperate hardwood and softwood forests [Steudler et al., 1990].The studies differed in form of N added, relative N loading, and frequency of application, so their results are not directly comparable.
Soil CH 4 profiles (Figure 5a) strongly suggest that soils at deciduous and coniferous upland sites are well aerated at least to 60 cm.The organic horizon in these soils is generally limited to depths less than 20 cm.Thus it is likely that methanogenesis is nonexistent or confined to anaerobic microzones [Smith and Arah, 1985]
is permanently seated in the soil, and lucite vertical sections and lids that utilize a waterfilled channel for a seal.Flux chamber bases were permanently deployed in or near the 15 m x !5 m vegetation control plot within each of the floodplain and upland sites (Table1) during September 1989.Bases were deployed in triplicate at all floodplain sites and at upland site UP3A.Bases at site UP1A were deployed in duplicate in areas where ground cover deployed in the vegetation control and low-N plots within these sites to assess the effect of N fertilization on CH 4 flux.Methane flux measurements were made weekly at each upland and floodplain site from late May through September 1990.Syringe samples were collected from each chamber over a 0 are used because of the large number of values below the detection limit and because no single data transformation consistently homogenized variances.Statistical analyses were

Fig. 4 .
Fig. 4. Multiple comparisons of ranked methane flux measurements for BCEF (a) floodplain and (b) upland sites by Dunifs nonparametric procedure.Mean ranks for the site types are arranged in increasing order.Sites underscored by the same line have no significant differences; sites not underscored by the san• line show significantly different fluxes.
with advancing plant community succession (Figure4b).A significant increase in CH 4 consumption rate was observed at black spruce (BS2) relative to the bum site (UP1A).The CH 4 consumption rate at all other sites could not be clearly classified.
and the important biological process influencing the soil CH 4 distribution is oxidation.Keller et al. [1990] andBorn et al. [1990] found similar depth distributions for CH 4 in moist soils from tropical and temperate forests, which also suggests CH 4 consumption only.Yavitt et al. [ 1990] report CH 4 oxidation in the upper 0 of a temperate hardwood forest, and no seasonal variations are evident in the May through October record at other temperate hardwood sites[Steudler et al., 1990].These reports are consistent with our observation that the atmosphere is the major CH 4 source for microbial oxidation in many forest soils.The moist surface zone of microbial activity can be expected to thaw and warm quickly in the spring.in contrast, the gradual thawing and warming of saturated, biologically active tundra soils underlain by permafrost results in a pronounced seasonal signal for CH 4 emission[Whalen and Reeburgh, 1988].The low threshold and high capacity for CH 4 oxidation shown in the CH4 amendment experiment (Figure5½) is in agreement with our previous observations on end-member CH 4 oxidizing environments, namely well-drained subarctic tundra[Whalen and Reeburgh, 1990b] and the surface soil of a retired landfill[Whalen et a1.,1990].These results indicate the presence of a ubiquitous bacterial community capable of rapidly oxidizing CH 4 at concentrations well above and below atmospheric levels.These results also suggest that the rates of CH 4 consumption reported above for forest soils are governed by physical soil conditions (morphology, porosity, moisture content, and temperature) that regulate CH 4 diffusion to microbes, rather than intrinsic biological factors.Further support for physical control of CH 4 consumption by forest soils is shown in Table3; increased CH 4 consumption (negative flux) was correlated with increasing soil temperature and decreasing soil moisture when significant values of z were found for sites that showed CH 4 consumption only.It is likely that soil moisture content exerts the greatest physical influence on CH 4 oxidation rates in forest and other soils.We demonstrated experimentally that CH4 oxidation rates in moist soils from a retired landfill were reduced nearly ten fold when saturated with water[Whalen et al., 1990], and $teudler et al. [ 1990] reported significant decreases in rates of CH 4 consumption by temperate forest soils under conditions of increased moisture.al., 1986; Matthews and Fung, 1987] of ~15 Tg yr -1 [Whalen and Reeburgh, 1990a], largely because of high rates of CH 4 emission from point sources (bogs and fens) [Crill et al., 1988; Hatriss et al., 1985; Moore et al., 1990; Moore and Knowles, 1990] distributed throughout this environment.Using the overall flux results from this study (characteristics resulting from temperature and precipitation changes are likely to influence rates of CH 4 consumption.Water table and temperature manipulation experiments (1 to 10 m 2 scale) are an approach to understanding these relationships.