Canadian Forest Service Publications

Variability in regional scale estimates of carbon stocks in boreal forest ecosystems: results from west-central Alberta. 2002. Banfield, G.E.; Bhatti, J.S.; Jiang, H.; Apps, M.J. Forest Ecology and Management 169: 15-27.

Year: 2002

Issued by: Northern Forestry Centre

Catalog ID: 20987

Language: English

Availability: Order paper copy (free), PDF (request by e-mail)

Mark record


Aboveground biomass, forest floor, and soil carbon (C) stocks were estimated for a transitional boreal region in western Alberta using available forest inventory data, model simulation, field observed plot data, and soil polygon (area averaged) information from the Canadian soil organic carbon database (CSOCD). For the three C pools investigated, model simulation provided a regional estimate, while forest inventory, plot, and soil polygon data provided an estimate of the spatial variation. These data were used to examine the variation of the C estimates, in both temporal (e.g. climate change) and spatial (e.g. soil physical characteristics) dimensions. Using the carbon budget model of the Canadian forest sector (CBM-CFS2) the regional average aboveground biomass C was estimated at 43 Mg C ha-2, similar to the estimate from the 1994 Canadian forest inventory (50 Mg C ha-2). Model simulation over the period 1920–1995 elucidated the major role that disturbances (harvest, fire and insects) play in determining the C budget of the region. Decreases in stand replacing disturbances over the period resulted in an accumulation in biomass C.

Regional estimates of forest floor C using aggregated plot data, CSOCD (forested area only) data, and CBM-CFS2 simulations were in close agreement, yielding values of 2.9, 3.4 and 3.3 kg C m-2, respectively. Regional estimates of total soil C using the three methods were more divergent (14.8, 8.3, and 15.6 kg C m-2, respectively).

An exponential relationship between clay content and biomass for mature coniferous stand types was found (r2=0.68), which is reasonable considering that as a site variable, texture affects tree growth through the modification of nutrient and water availability. The relationship was used to predict the range of potential values for biomass C at maturity across the region. Forest inventories of biomass seldom provide enough data across the range of ages and stand types to develop stand growth curves that capture the variation in growth across the landscape. Consequently, growth dynamics must be inferred from a large area to provide enough biomass-to-age data, which results in a loss in the ability to use it to predict C pools and fluxes at a small scale. Using relationships between site factors (such as soil texture) and biomass C provides a means to modify inventory-based biomass-to-age relationships to assess the variation across the region as well as make predictions at a higher spatial resolution. This is relevant where both spatial extent and a finer scale are required, but site-specific biomass-to-age relationships are unavailable.