REGIONs DuE TO CLImATE CHANGE INDuCED PERmAFROsT RETREAT
M.J. WHItIcar
school of earth and ocean sciences, university of Victoria, Victoria, canada J. BHattI, n. startseV
northern forestry centre, st edmonton, aB, canada
Abstract
Thawing permafrost peatlands substantially influence Canadian northern ecosystems by changing the regional hydrology and mobilizing the vast carbon (c) reserves that results in increased greenhouse gas (gHgs) emissions to the atmosphere. With permafrost distribu-tion controlled largely by topography and climate, our Internadistribu-tional polar year (Ipy) study intensively monitored the local C cycling processes and GHG fluxes associated with different hydrologic and permafrost environments at 4 sites along a climatic gradient extending from the Isolated patches permafrost Zone (northern alberta), to the continuous permafrost Zone (Inuvik, nWt). each site encompasses a local gradient from upland forest and peat plateau to collapse scar. Our multi-year measurements of peatland profiles and flux chambers for CH4
and co2 concentrations and stable isotope ratios indicate processes, including methanogenesis, methanotrophy, transport and emission that control the distribution of these gHgs. these re-lationships are modulated by fluctuating local soil water and corresponding ecosystem condi -tions. The gas geochemistry shows that significant surface CH4 production occurs by both hy-drogenotrophic and acetoclastic methanogenesis in submerged, anaerobic peats, e.g., collapse scars, whereas methane oxidation is restricted to aerobic, drier environments, e.g., upland sites and peat-atmosphere interface. the most active methanogenesis and emissions are in areas of actively thawing permafrost contrasting with sites under continuous permafrost. this degree of methanogenesis is being amplified by the increased rate of Arctic warming and the rapid retreat of permafrost in canada’s arctic (approximately. 2.5 km/a).
1. IntroductIon
forests and peatlands in northern regions are the largest terrestrial reservoir of carbon (c) [1]. In canada’s northern region, the 304 M ha of forest and 89 M ha of peatland represent a combined estimated C pool of 186 Pg and annual C flux of 87 Tg
Whiticar et al.
[2, 3]. In continental western canada, peatlands cover ca. 365 000 km2 or 20% of the land area [4, 5]. These large forest and peatland areas are the most significant po -tential source of terrestrial c in canada. these northern ecosystems are also a major contributor of global, atmospheric greenhouse gases (gHgs), including cH4, co2, and n2o. further, the global warming potential of cH4 and n2o are 23 and 296 times that of co2, respectively, and the concentrations of co2, cH4 and n2o in the atmos-phere are increasing significantly [6].
growing evidence indicates that the northern latitudes of canada are warming more rapidly than the mid latitudes [7]. In canada’s northern region this trend has increased to a rate of 1.8°c/century during the past 25 to 30 years and is associated with drier conditions. the Mackenzie Valley region of northwestern canada (fig. 1) has undergone the most warming (1.7°c) over the last century in canada [7]. Moreo-ver, general circulation models predict that for climate warming resulting from a dou-bling of co2, the region will experience increases in mean annual air temperature of up to 5°c.
In the event of such predicted climate change, forest and peatland ecosystems in the Mackenzie valley region will also be affected by changes in the key distur-bance regimes of permafrost and wildfire. Regional studies including ECMBIS [8]
and nrcanpeMV [9] predict that permafrost will partially or completely disappear over large areas in the Mackenzie Valley.
FIG. 1. Left: location map of study in different permafrost areas (with δ2HH2O). Right: record of temperature increase at Anzac, Ft. Simpson, Norman Wells and Inuvik study sites.
IAEA-CN-186/216
climate change may lengthen growing seasons, change mineralization rates, in-crease fire frequencies, and lead to widespread thawing of permafrost. It is estimated that since 1960 about 2600 km2 of permafrost has thawed in the peatlands of boreal western canada [10], and the projections are for this to increase spatially and tempo-rally. Recent analysis projected a 74–118% increase in area burned in northwestern canada by end of this century under 3 × co2 scenarios [11].
It is uncertain how the above-mentioned changes will affect the distribution, composition and c source–sink capacity of forests and peatlands, i.e., balance be-tween organic matter production, heterotrophic mineralization, and gHg emissions, especially cH4 [12–16]. Because of the relatively large changes in climate and asso-ciated disturbances occurring in this region, our Ipy project was designed to improve our understanding of the potential effects of recent climate change, as well as natural and anthropogenic disturbances, on the total c storage and source–sink relationships, and gHg emissions of forests and peatlands in the Mackenzie Valley.
Quantifying the c source/sink nature and gHg emissions budgets of forests and peatlands represents a substantial challenge. Yet even more difficult is predicting future c and gHg balances under climate change and associated disturbances such as fire and permafrost melting. The warming and drying predicted for the Mackenzie Valley is expected to lead to increased area burned [11]. fire also indirectly controls longer term patterns of c accumulation by initiating secondary succession [17–19]
and altering site environmental conditions [20, 21]. Fire has a strong influence on patterns of net primary production [22] and soil respiration [23], which cumulatively influence long term patterns of soil C storage [24, 25]. In addition to soil CH4 release with melting permafrost, increased fires are a potential source of GHGs and were im -portant during enhanced warming periods in the Holocene [12, 26, 27, 15]. consider-ing these interactions, the role of fires in long term C cyclconsider-ing in the region is complex and remains little understood.
In forest and peat ecosystems, soil organic c can be assimilated, immobilized, leached and emitted to the atmosphere as co2 via aerobic, or cH4 via anaerobic, microbial respiration. To estimate the microbial mediated fluxes of CO2 and cH4 to the atmosphere, we need to identify the types of microbes and establish the linkages with GHGs formation. At present, there has been very little research linking specific the types of microbes and gHgs dynamics in northern ecosystems.
during the younger dryas-preboreal (yd-pB) transition (ca. 11.5 kyBp), warming, similar to that expected in this century, led to the retreat of ice cover and permafrost. new cH4 c isotope measurements from greenland ice over this yd-pB period [28, 29] strongly suggest that this cH4 rise was not from marine gas hy-drates (clathrate gun Hypothesis, [30]) or from tropical wetlands. rather, permafrost modulated cH4 emissions are the most probable, yet poorly characterized, cause of the rise.
In the subarctic region of sweden [31] between 1970 and 2000 there was a 22–
66% increase in CH4 emissions on a landscape scale attributed to climate change
Whiticar et al.
influence on permafrost and bog vegetation. It is conceivable that climate change and its associated disturbances could tip c balances from net sinks to net sources over the next 20 to 60 years.
overall, the most important controls of c cycling in forests and peatlands are plant communities, soil temperature, hydrology, and the chemistry of peat and plant tissues [32]. limited data and understanding of how changing environmental condi-tions and disturbances (including fires and permafrost thawing) influences the C cy-cle of forests and peatlands over short and medium timescales (10–100 years), poses a great challenge to predict changes in c sink–source relationships and gHg dynam-ics under a changing climate. given the complexity of climate–disturbance–c cycle interactions, there is a strong need for across-scale and across-ecosystems studies.
Our research focuses on soil profiles of GHG (CH4, co2) concentrations and their corresponding isotope ratios (δ13co2 and δ13cH4) to determine and track the dif-ferent processes of methanogenesis, methylotrophy and microbial respiration, e.g., ref. [33]. this is possible due to the distinctive isotope effects (and resultant isotope ratio signatures) that are characteristic of these various microbial processes, e.g., ref.
[33]. By comparing co-existing isotope pairs, i.e., δ13co2–δ13cH4, it is possible to dis-tinguish between different methanogenic pathways, i.e., acetoclastic fermentation vs hydrogenotrophic (co2-reduction) [33]. this information reveals the state of organic substrate availability and ‘reserve’, which could be expected to shift in type or strati-graphic depth as the climate changes. for example our previous work in fluxnet [34]
and elsewhere has shown that acetoclastic fermentation is a dominant methanogenic process until the labile organic pool (loc, e.g., acetate, formate, tMa, dMs etc.) is exhausted. the methanogenic pathway then shifts, typically in older and deeper soils and sediments, to hydrogenotrophic methanogenesis.
We test with concentration and isotope information whether the acetoclastic methanogenic pathway diminishes during dryer conditions with lowered water tables (less loc), with a resultant drop in longer-term cH4 fluxes. It is important to note that measurements on soil profiles, not just flux chamber studies, are vital to under-stand the spatial and temporal (seasonal and decadal) variations in diagenesis and gHg systematics.
furthermore, c-isotope information is critical to identify the presence and de-gree of aerobic methane oxidation occurring in soils [33]. characteristic isotope shifts are diagnostic of this process [33, 34]. We use isotopes to test the hypothesis that de-creased water table leads to enhanced microbial oxidation of cH4 and enhanced co2 formation.
2. study area
the study area is a 1500 km transect from northern alberta along the Mac-kenzie Valley to the Beaufort sea coast (fig. 1). the transect covers the latitudinal
IAEA-CN-186/216
gradient of Boreal, subarctic and arctic ecoclimatic regions. Its general south-north orientation offers representative permafrost distributions from the southern fringe of the discontinuous permafrost zone, where the Mackenzie river begins at great slave lake, to the continuous permafrost zone on the arctic ocean coast and the Macken-zie delta. In May 2007 intensive monitoring sites were selected throughout the Mac-kenzie Valley and northern Alberta. Each study site was chosen to reflect a gradient from upland to peatland conditions. Individual plots are located within the upland forest (ul), peat plateau (pp), and collapse scar (cs), as shown in fig. 2.
3. saMplIng and MetHods
Passive gas flux chambers using permanent soil collars inserted in the ground were deployed at all sites to determine the net cH4 and co2 fluxes. Soil gas samples are collected using thin metal tubes (sas tubes) permanently inserted in the soil to pre-determined depths from 5 to approx. 200 cm and. soil water or gases are sampled seasonally by aspirating with a syringe (cH4 in water samples is extracted by shak-ing the water with air). cH4 and co2 concentrations are measured in the field using a portable optical spectrometer system (gyro™) developed by Isometric Instru-ments, Victoria. stable isotope ratios are measured by continuous flow-Isotope ra-tio Mass spectrometry (cf-IrMs) methods at the Bf-seos uVic [35]. the carbon isotope ratios are expressed as δ13C in per mil (‰) using standard δ-notation:
FIG. 2. Comparative sampling sites along hydrologic gradient (UL – Upland, PP – Peat Plateau, CS – Collapse Scar) at Anzac, Ft. Simpson, Norman Wells and Inuvik.
Whiticar et al.
sample standard 3 standard
(‰)=R R 10
R d −
where R is the molar ratio of heavy to light isotope, i.e., 13c/12c for the sample and standard. carbon isotope ratios are referenced to Vienna pee dee Belemnite (V-pdB).
4. results and dIscussIon
cH4 emissions varied significantly between study sites with the highest val-ues in the ft simpson collapse scar (cs) even though anzac has the highest annual soil temperature. ft simpson is located in the zone of most active permafrost thaw-ing. the highest cH4 emissions are consistently observed in the submerged or water saturated features of the landscape at all study sites (fig. 3). Both peatland and up-land up-landscape features showed very low efflux or even negative influx of CH4 from the boundary layer into the soil, i.e., cH4 concentration was continually dropping with time in the chamber (Fig. 3). This indicates significant CH4 uptake and micro-bial oxidation in these soils.
Although the peatlands (PP, UP) had no surface flux of CH4, soil cH4 concen-trations in water-saturated layers at depth reached several hundred ppm, indicating that the drained soil effectively oxidized produced cH4.
Methane concentration soil depth profiles in water saturated areas of the CS show sharp increases in cH4 concentration with depth (fig. 4). overall, cH4 produc-tion was greatest in early summer across all sites. In midsummer, cH4 production was minimal in all study sites despite the fact that the water table remained high (within 10 cm from the surface) (fig. 4). this can be attributed to the increased o2 concentration throughout the profile and was observed in mid-July to mid-August.
FIG. 3. Comparison of methane concentration build-up in flux chambers at CS, PP and UL sites along the hydrologic gradient at Anzac.
IAEA-CN-186/216
We hypothesize that the increase in o2 concentration is due to the activity of herba-ceous roots, capable of transporting o2 down the profile of anaerobic soils [36].
The δ13cH4 from collapse scar soil samples in the 4 study areas (fig. 5) show considerable range from –83.7‰ to –47.8‰, the latter a value similar to atmospheric air (–47.6‰). The soil samples show remarkably little variation in δ13cH4 with depth at any specific site, but large differences between sites. In particular, at Ft Simpson the δ13cH4 varied about –52‰, whereas it was more 12c-enriched around –66‰ at the southern anzac site (fig. 5). this isotopic homogeneity is rather unexpected based on the soil methane concentration gradients observed in fig. 4. Isotope effects due to methane production, consumption or even diffusion, should lead to a vertical partitioning of the δ13cH4 [33], but this is not observed in this soil profile dataset.
The surface soil chamber δ13cH4 measurements at anzac are similar to the soil pro-file, whereas at Ft Simpson the chambers δ13cH4 are about 8‰ 12c-enriched relative to the soil (fig. 5).
A possible explanation for the vertical δ13cH4 homogeneity could be that the methane in collapse scars is generated rapidly and in relatively large amounts once the soil becomes anoxic. Vertical migration may smear the isotope signature so that there is no change in δ13cH4 with depth. The consistent δ13cH4 offset between ft.
simpson and anzac (ca. 14‰) could be due to differences in the methanogenic path-way. If so, then Ft Simpson expresses a δ13cH4 signature that is more characteristic FIG. 4. Comparison of soil methane concentration depth profiles in collapse scars (CS) at Anzac, Ft. Simpson, Norman Wells and Inuvik.
Whiticar et al.
of acetoclastic methanogenesis [33]. In contrast, Anzac has a δ13cH4 signature more representative of hydrogenotrophic methanogenesis [33].
The comparison of the soil δ13co2 and δ13cH4 values as shown in fig. 6 [33]
can be useful to determine the methanogenic/methanotrophic pathways and process-es. Although the soil δ13co2-δ13cH4 database is currently limited, it does appear from these data that the northern permafrost sites, including ft simpson, are dominated by acetoclastic methanogenesis. In contrast, anzac, the southern permafrost-free site, is dominated by hydrogenotrophic methanogenesis. a possible explanation is related to the carbon precursors. the northern sites may have more labile organic material pre-sent, e.g. organic acids, that support acetoclastic methanogenesis. these labile sub-strates have only recently, and to a limited degree, become active due to the exposure of the active layer (especially at ft simpson) because of the retreat and loss of per-mafrost at these sites (fig. 1). In comparison the anzac site has been perper-mafrost-free for thousands of years and the labile carbon pool has been exhausted. under these conditions, the methanogenic pathway shifts from acetoclastic to hydrogenotrophic methanogenesis (as observed).
overall, our study reveals that there is a net emission of cH4 and co2 from the soils at all 4 sites (fig. 7). In particular, co2 from the upland soils is the larg-est source of gHg emissions observed up to 1 g co2 m–2 hr–1. the emissions follow FIG. 5. Comparison of soil δ13CH4 depth profiles and chambers in collapse scars at Anzac, Ft.
Simpson, Norman Wells and Inuvik.
IAEA-CN-186/216
a clear trend to lower values as we move down the hydrologic gradient to peat pla-teau (pp) and then collapse scar (cs) (fig. 7). only in the collapse scar at the north-ern sites is there net uptake of co2 by the soils. Methane is only a minor contributor at all sites compared with co2 (cH4 converted by gWp to co2 equivalents [36] and is relevant only in the collapse scar settings (fig. 7).
5. conclusIons
this study indicates that in permafrost-related peatlands, increased cH4 pro-duction by anaerobic methanogenesis is associated with water saturated areas, such as the collapse scars. the greatest cH4 production takes place in the regions with FIG. 6. Comparison of soil δ13CO2 and δ13CH4 at Anzac, Ft. Simpson, Norman Wells and Inuvik.
FIG. 7. Comparison of CH4 and CO2 emissions at Anzac, Ft. Simpson, Norman Wells and Inuvik.
Whiticar et al.
the most recently thawed permafrost. this illustrates the potential for increased gHg emissions under climate warming.
Based on initial stable isotope measurements, the soils with an active layer most recently thawed (ft simpson) appear to be dominated by acetoclastic metha-nogenesis, whereas the ‘aged’ soils, e.g., anzac, appear to have depleted the labile organic material and have transitioned to hydrogenotrophic methanogenesis.
In the upland forest and peat plateau methanogenesis is absent or low. In these aerated, oxic soils there is evidence for methane uptake and consumption by methy-lotrophs in the soil.
At the height of the growing season, increased oxygenation of the soil profile, possibly due to root activity, led to decreasing cH4 concentrations throughout the soil profile.
the effect of temperature (Q10 factor) for cH4 emissions decreased from southern to northern sites, indicating differences in methanogenic microflora adapta-tion and possibly available methanogenic substrates.
ACkNOWLEDGEmENTs
We are grateful to s. gooderham, M. gravel, t. lakusta and p. rivard from the forest Management division, government of nWt. We also thank r. errington, p. eby, c. andrews, t. Varrem-sanders, s. Kull, V. Mucciarelli, c. renz, and p. sug-awara. funding is from government of canada program for International polar year (Ipy) and nserc discovery grants.
REFERENCEs
[1] IntergoVernMental panel on clIMate cHange, “observed climate Variability and Change”, Climate Change 2001, Working Group I: The Scientific Basis, cambridge university press, cambridge, uK (2001) 99–182.
[2] apps, M.J., KurZ, W.a., luXMoore, r.J., nIlsson, l.o., sedJo, r.a., scHMIdt, r., sIMson, I.g., VInson, t.s., Boreal forest and tundra, Water air soil poll. 70 (1993) 39–53.
[3] WeBer, M.g., flannIgan, M.d., canadian boreal forest ecosystem structure and function in a changing climate: impact on fire regimes, Environ. Rev. 5 (1997) 145–166.
[4] VItt d.H., Halsey, l.a., ZoltaI, s.c., the changing landscape of canada’s western boreal forest: the current dynamics of permafrost, can. J. for. res. 30 (2000) 283–287.
[5] ZoltaI, s.c., pollet, f.c., “Wetlands in Canada: their classification distribution, and use”, Mires: swamp, Bog, fen, and Moor, Vol. 4B, regional studies (gore, a.J., ed.), elsevier, amsterdam (1983) 245–268.
IAEA-CN-186/216
[6] BHattI, J.s., laI, r., apps, M.J., prIce, M.a., (eds), climate change and Managed ecosystems, crc press, Boca raton, fl (2006) 464.
[7] HengeVeld, H. WHIteWood B., understanding climate change: a synthesis of climate change science, Meteorological service of canada, ottawa (2005) 72.
[8] coHen, s.J., Mackenzie basin impact study (MBIs): final report, ottawa, envi-ronment canada (1997) 372.
[9] dyKe l.d., BrooKs, g.r., (eds), northwest territories: a baseline for the assess-ment of environassess-mental change, geol. surv. canada Bull. 547 (2000) 189.
[10] roBInson, s.d., carbon accumulation in peatlands, southwestern northwest ter-ritories, canada, can. J. soil sci. 86 (2006) 305–319.
[11] flannIgan M.d., logan, K.a., aMIro, B.d., sKInner, W.r., stocKs, B.J., future area Burned in canada, climatic change 72 1–2 (2005) 1–16.
[12] carcaIllet, c., at al., Holocene biomass burning and global dynamics of the carbon cycle, chemosphere 49 8 (2002) 845–863.
[13] roulet, n.t., peatlands, carbon storage, greenhouse gases and the Kyoto proto-col: Prospects and significance for Canada, Wetlands 20 (2000) 605–615.
[14] tHonIcKe, K., prentIce, I.c., HeWItt, c., Modelling glacial–interglacial changes in global fire regimes and trace gas emissions, Global Biogeochemical Cy-cles 19 (2005) 3008.
[15] turetsKy M., WIeder, r.K., VItt, d.H., Boreal peatland C fluxes under vary-ing permafrost regimes, soil Biol. Biochem. 34 (2002) 907–912.
[16] Bourgeau-cHaVeZ, l.l., KasIscHKe, e.s., BrunZell, s.M., Mudd, J.p., sMItH, K.B., frIcK, a.l., analysis of spaceborne sar data for wetland mapping in Virginia riparian ecosystems, International Journal of remote sensing 22 (2001) 3665–3687.
[17] roWe, J.s., scotter, g.W., fire in the boreal forest, Quat. res. 3 (1973) 444–464.
[18] VIerecK, l.a., Van cleVe, K., dyrness, c.t., “forest ecosystem distribu-tion in the taiga environment”, forest ecosystems in the alaskan taiga: a synthesis of structure and function (Van cleVe, K., et al., eds), springer, new york (1986) 22–43
[19] KasIscHKe, e.s., BruHWIler, l.M., emissions of carbon dioxide, carbon monoxide and methane from boreal forest fires in 1998, J. Geophys. Res. 107 d1 8146 (2003).
[20] o’neIll, K.p., KasIscHKe, e.s., rIcHter, d.d., environmental controls on soil co2 flux following fire in black spruce, white spruce, and aspen stands of interior
[20] o’neIll, K.p., KasIscHKe, e.s., rIcHter, d.d., environmental controls on soil co2 flux following fire in black spruce, white spruce, and aspen stands of interior