Carbon stored in peatlands is estimated to be between 200 and 450 Pg [Roulet et al., 2007; Gorham, 1991] and accounts roughly for one third of the world's soil carbon pool. Therefore, ecological controls on peatland carbon balance are of great interest. Particularly, the water balance of peatlands has long been considered a key control for all the physical, chemical and biological processes in peat [Lafleur et al., 2003, 1997; Shurpali et al., 1995]. However, the overall influence of soil water on ecosystem respiration (ER) in peatlands has been questioned recently [Lafleur et al., 2005a] as various, sometimes contradictory observations have been reported in the literature. The ER response to varying water table (WT) depths is further complicated by covariation of water table with soil temperatures [Lafleur et al., 2005a]. Therefore, a key question in peatland research is “how vulnerable is peatland carbon, to changes in peat water contents (θ) and water table”?
1.1. Effects of Peat Water Content θ on CO2 Production
 Field chamber measurements, as well as in situ and laboratory incubations, both support [Sulman et al., 2009; Strack and Waddington, 2007; Bubier et al., 2003a, 2003b; Oechel et al., 1998; Silvola et al., 1996a] and refute [Strack and Waddington, 2007; Updegraff et al., 2001; Bubier et al., 1998; Bridgham et al., 1991] the dependence of ecosystem respiration on WT fluctuations. Some studies suggest that soil respiration in peat profiles increases with increased drying and aeration at depth [Moore and Dalva, 1993]. Others claim that CO2 production rates decrease as θ deviates above or below an optimum value [Silvola and Ahlholm, 1989]. This optimum value was found to vary from ∼0.6 m3 m−3 estimated by modeling [Frolking et al., 2002] to ∼0.9 m3 m−3 determined experimentally [Waddington et al., 2001]. Although the latter number seems high, it is consistent with Orchard et al. , who suggested that stimulating effects of increasing θ selectively allows various microbial communities to become active under different moisture conditions. Silvola et al. [1996a] found that initial WT decline from peat surface resulted in a pronounced increase of CO2 emission, which became less pronounced with further WT decline down to ∼30 cm depth, and even slightly decreased with water table dropping below ∼30–40 cm. Sulman et al.  found that the initial WT drawdown caused an increase of CO2 emission, which became less pronounced with WT decline below ∼25–30 cm depth.
 Strack and Waddington  reported increased respiration in hollows but no significantly different respiration in hummocks with lowereing of the water table. Their study was one of the few that considered peatland microtopography and treated separately respiration in hummocks from respiration in hollows. Other recent field experiments have reported a lack of correlation between the WT depth and ecosystem respiration [Lafleur et al., 2005a; Moore et al., 2003; Updegraff et al., 2001; Scanlon and Moore, 2000], mostly explained by the small contribution of deep peat [Blodau et al., 2007]. Limited CO2 production from deep peat was attributed to (1) the low proportion of readily available organic carbon [Updegraff et al., 1995; Nadelhoffer et al., 1991], (2) accumulation of recalcitrant humic compounds [Hogg et al., 1992], (3) unavailability of suitable electron acceptors [Lafleur et al., 2005a; Frolking et al., 2001; Waddington et al., 2001], and (4) low soil temperatures at depth [Blodau et al., 2007].
1.2. Advances in Modeling of Hydrological Effects on Respiration
 Process-based models are well suited for investigating the effects of hydrology on ecosystem respiration because it is possible to distinguish the effects of moisture from those of soil temperature and nutrients within the peat profile. The ecosys model [Grant, 2001], which couples ecosystem hydrology, soil thermal regime and carbon balance, has been shown to simulate reasonably well hourly dynamics of water table, θ and soil temperatures at various depths in hummocks and hollows [Dimitrov et al., 2010a, 2010b]. The model was applied to simulate bog ecosystem respiration in this study.
 Ecosys has an advantage to other models for peat carbon balance, such as PCARS [Frolking et al., 2002], PDM [Frolking et al., 2001], InTEC V3.0 [Ju et al., 2006] and MWM [St-Hilaire et al., 2008], in that it can explicitly simulate CO2 production in soil by diverse heterotrophic and autotrophic microbial populations that drive substrate hydrolysis, oxidation-reduction reactions and nutrient uptake, which in turn drive microbial growth [Grant, 2001]. Oxidation-reduction reactions in ecosys are determined by demand for and supply of electron acceptors, such as O2, NO3−, NO2−, N2O, H2, and reduced C [Grant and Pattey, 2003, 1999; Grant and Roulet, 2002; Grant and Rochette, 1994], so that a range of aerobic and anaerobic reactions are simulated. In contrast, most of the above peatland models simulate decomposition and respiration by prescribed first-order decay rates as hydrological effects on these processes are formulated through empirical multipliers.
1.3. Objectives and Hypotheses
 The main objective of this research is to understand and model the effects of subsurface hydrology on soil respiration in bog hummocks and hollows, and on bog ecosystem respiration. A secondary objective is to explain and reconcile the contrasting effects of water table on ecosystem respiration observed in peatlands, as summarized above, through modeling the effects of subsurface peat hydrology, and fibric peat thickness and high macroporosity, on microbial and root respiration.
1.3.1. Soil Respiration in Hummocks
 It has been shown previously that water table drawdown in ecosys model creates desiccation in the near-surface peat in hummocks [Dimitrov et al., 2010a, 2010b; Dimitrov, 2009], which was consistent with field observations of rapid drainage through the high macropore fraction of the fibric peat (Figure 1) [Lafleur et al., 2005a; Silins and Rothwell, 1998]. In this study we hypothesize that in hummocks decrease of near-surface microbial respiration through reducing microbial habitat with near-surface drying would be offset by the increase of microbial and root respiration at depth with water table drawdowns and improved aeration. Hereafter, this response is referred as a respiration offsetting mechanism in hummocks and is addressed by ecosys as follows. Increase of aqueous concentration of active microbial biomass with desiccation in the model results in microbial competitive inhibition [Grant, 2001; Lizama and Suzuki, 1991], which slows substrate decomposition (hydrolysis), thus reducing uptake of decomposition products by microbes in the most productive near-surface peat. Slower uptake decreases the active microbial biomass at near-surface, further slowing decomposition and thus promoting low respiration rates. As water table recedes, gaseous O2 diffusion increases at depth, which results in increased aqueous O2 concentrations, and shift from anaerobic to aerobic microbial respiration in above water table. The transition from anaerobic to aerobic respiration results in higher energy yield, increasing microbial biomass growth, hence respiration. Vascular root respiration and moss rhizoid respiration also increase with increased root O2 uptake and increased root growth and densities, and a larger respiring root biomass.
1.3.2. Soil Respiration in Hollows
 We hypothesize that in hollows lack of near-surface desiccation [Dimitrov et al., 2010a, 2010b; Dimitrov, 2009] due to the thin fibric peat (Figure 1), together with improved aeration at depth would result in increased respiration with water table drawdown. However, this increase would be less than that in hummocks due to the waterlogged deep peat with predominating anaerobic respiration.
1.3.3. Aboveground Plant Respiration
 Shrubs would compensate for the near-surface drying through their deeper roots, thus maintaining conservative shrub water potential, productivity and aboveground respiration. However, a decrease in aboveground moss respiration in hummocks with declining moss water potential and productivity caused by the near-surface drying would largely offset increased hollow soil respiration.
1.3.4. Ecosystem Respiration
 As described above, increase of deep peat respiration with water table drawdown would be offset by concurrent decreases of near-surface soil respiration and aboveground plant respiration on hummocks, thus resulting in ecosystem respiration that is relatively unaffected by variations of water table.