Dealing with microtopography of an ombrotrophic bog for simulating ecosystem-level CO2 exchanges

Peatlands contain approximately 25% of the global soil carbon (C), despite covering only 3% of the earth's land surface. In order to evaluate the role of peatlands in global C cycling, models of ecosystem biogeochemistry are required, but peatland ecosystems present a number of unique challenges, particularly how to deal with the large variability that occurs at scales of one to several metres. In models, spatial variability is considered either explicitly for each individual unit and the outputs averaged, referred to as flux upscaling, or implicitly by weighting model parameters by the fractional occurrence of the individual units, referred to as parameter upscaling. The advantage of parameter upscaling is that it is much more computationally efficient: a requirement for hemispheric scale simulations. In this study we determined the differences between modelling a raised bog peatland with hummock-hollow microtopography using flux and parameter upscaling. We used the McGill Wetland Model (MWM), a process-based ecosystem C model for peatlands, configured for hummocks and hollows separately and then a weighted mixture of both. The simulated output based on flux and parameter upscaling was compared with eddy-covariance tower measurements. We found that net ecosystem production (NEP) for hollows was much larger than that for hummocks because total ecosystem respiration (TER) for hummocks was greater while gross primary production (GPP) did not differ significantly between the two topographic features. However, despite differences in components of NEP between hummocks and hollows, there was no statistically significant difference between the NEP based on flux and parameter upscaling using the MWM. Both flux and parameter upscaling show equivalent capability to capture the magnitude, direction, seasonality and inter-annual variability. The root-mean-square-errors (RMSE) are 0.66, 0.45, and 0.49 g C m⁻² day⁻¹, respectively for GPP, TER and NEP based on the flux upscaling, while 0.67, 0.44, and 0.48 g C m⁻² day⁻¹, respectively based on the parameter upscaling. The degree of agreement (d*) is 0.96, 0.97, and 0.88, respectively for GPP, TER and NEP based on the flux upscaling, while 0.96, 0.97, and 0.89, respectively based on the parameter upscaling. This result suggests that differences in processes caused by peatland microtopography scale linearly, which means an ecosystem-level model set-up (i.e. parameter upscaling scheme), is sufficient to simulate the C cycling.     

Greenhouse gas emissions from Canadian peat extraction, 1990-2000: a life-cycle analysis

This study uses life-cycle analysis to examine the net greenhouse gas (GHG) emissions from the Canadian peat industry for the period 1990-2000. GHG exchange is estimated for land-use change, peat extraction and processing, transport to market, and the in situ decomposition of extracted peat. The estimates, based on an additive GHG accounting model, show that the peat extraction life cycle emitted 0.54 x 10(6) t of GHG in 1990, increasing to 0.89 x 10(6) t in 2000 (expressed as CO2 equivalents using a 100-y time horizon). Peat decomposition associated with end use was the largest source of GHGs, comprising 71% of total emissions during this 11-y period. Land use change resulted in a switch of the peatlands from a GHG sink to a source and contributed an additional 15%. Peat transportation was responsible for 10% of total GHG emissions, and extraction and processing contributed 4%. It would take approximately 2000 y to restore the carbon pool to its original size if peatland restoration is successful and the cutover peatland once again becomes a net carbon sink.     

Methane Fluxes from a Wetland using the Flux-Gradient Technique The Measurement of Methane Flux from a Natural Wetland Pond and Adjacent Vegetated Wetlands using a TDL-Based Flux-Gradient Technique

Methane emissions were measured from a bog andlake in the Experimental Lakes Area in Northern Ontario in 1992and 1993, prior to and following flooding. Bog fluxes were smallin 1992 (0.27 mg m^-2 d^-1) but increased 5-fold in 1993 afterflooding. Over the bog, there was a diel cycle of nighttimeemission and daytime uptake in 1992 in contrast to constantemission in 1993. Lake emissions decreased after flooding butwere much greater than bog emissions in both years (average = 7.3 mg m^-2 d^-1). Seasonally, the bog flux was correlated withground temperatures after flooding. In 1992, lake fluxes werecorrelated with air temperature on a daily basis. In contrast,seasonal lake fluxes were correlated with water and sedimenttemperatures in 1992, but only with sediment temperatures in1993. These results are explained with respect to the effects offlooding on lake and bog dynamics.