0618-B2
W.A. Kurz and M.J. Apps 1
The rate of carbon accumulation in the global atmosphere can be reduced by decreasing the emissions from fossil-fuel burning and by increasing the net carbon uptake in terrestrial (and aquatic) ecosystems. The Kyoto Protocol seeks to accomplish both. Canada is developing a national forest sector carbon monitoring, accounting and reporting system in support of its international obligations to report greenhouse gas sources and sinks. The system synthesizes data from many sources, including: a new national forest inventory and existing regional inventories; growth and yield information derived from a large network of sample plots; change statistics on change agents such as wildfire, insect disturbances and forest management activities; and land-use change, obtained from a range of sources including remote sensing. Computer simulation models are used to integrate this information to estimate carbon stocks, changes in carbon stocks and the emissions of greenhouse gases. A key component of the system is the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS2) which will enable analyses at four spatial scales (national, provincial, forest management unit and stand). The model will be used to assess carbon-stock changes between 1990 and the present, and to predict future carbon-stock changes, under a wide range of assumptions about management activities and natural disturbance rates. The system is designed to be compliant with the reporting requirements under the Kyoto Protocol and other international agreements. Moreover, applied at the scale of operational planning, the simulation models will enable forest managers to include carbon consequences of proposed management alternatives among their criteria for management choices. The Canadian approach may be of interest to other northern countries developing forest carbon tracking systems.
Forest ecosystems are both large stores of C, and, depending on the stage of stand development, can be either sources or sinks of C. Canada contains about 10% of the world's forests, most of which are owned by the Crown. Reporting requirements of the Kyoto Protocol and other international agreements are increasing the demand for quality information on Canada's forests, land use and land-use change. Accounting for impacts of land-use changes is mandatory for those countries that ratify the protocol. This paper provides a general overview of the information requirements resulting from the Kyoto Protocol and outlines the development of Canada's National Forest Carbon Monitoring, Accounting and Reporting Framework as an example of how these national information requirements can be met by building on existing forest information systems.
Countries that ratify the Kyoto Protocol enter into internationally binding agreements to reduce net GHG emissions and to report their annual greenhouse gas (GHG) balances. The protocol recognizes the role of land use, land-use change and forestry (LULUCF) in contributing both sources and sinks to national greenhouse gas balances. The protocol outlines a two-step process in which countries first have to define and identify the areas subject to certain activities, and then have to quantify the net-GHG sources and sinks on these areas. To account for the impacts of land-use change, specifically the conversion of forests to non-forests and vice versa, Article 3.3 stipulates that countries identify those areas affected by afforestation, reforestation and deforestation since 1990, and then quantify the carbon stock changes (and non-CO2 emissions) on these areas during five-year commitment periods. The first commitment period starts January 1, 2008 and ends December 31, 2012.
For the first commitment period, the protocol (Article 3.4) also provides countries with the option to include certain land use and forestry activities in their national accounts of GHG sources and sinks. Countries must decide by the end of 2006 whether to account for the C stock changes resulting from one or more of the activities specified under Article 3.4 (forest management, crop land management, grazing land management and revegetation). In the context of the protocol, forest management, is defined as a system of practices aimed at meeting a suite of societal goals. Countries that choose to include forest management in their accounts must identify the areas subject to forest management since 1990 and then account for the C stock changes (and non-CO2 emissions) on these areas during commitment periods.
After applying a series of accounting rules and guidelines (that are beyond the scope of this paper) as well as country-specific caps, the resulting net balance of sources and sinks will be applied to a country's national account. If LULUCF activities contribute a net sink, it will contribute to meeting the national emissions reduction targets, while a net source will be added to the country's emissions.
Compliance with the monitoring and accounting requirements of the Kyoto Protocol will require substantial improvements and refinements to most countries' forest and land-use inventories. It requires an inventory of land-use changes since 1990 at a resolution of 0.05 to 1 ha. While most nations will likely choose the upper bound of the size range, estimating past and future land-use changes with a 1 ha resolution will present a formidable challenge for large countries. The development of statistically valid sampling protocols will allow countries to meet these information requirements more efficiently.
The Marrakesh Accords specify that countries define the boundaries of the areas of land encompassing the units of land subject to the activities of Articles 3.3 and 3.4. Countries will likely choose these strata such that estimation of the areas subject to these activities is facilitated and the uncertainties of estimates obtained for these areas are minimized. In Canada, for example, population density and human access to forests differ greatly between regions and pre-stratification of the land area could be based on the extent and type of expected land use and land-use change activities.
Once the areas subject to activities under Articles 3.3 and (if elected) 3.4 have been identified, annual estimates of changes in all ecosystem C stocks (and the non-CO2 emissions) on these areas are required. Since it is not feasible to measure C stock changes on an annual basis, accounting frameworks must support the monitoring, accounting and reporting needs through a combination of measurements and modelling. The remainder of this paper will outline the ongoing development of the Canadian national forest carbon accounting framework.
The purpose of Canada's National Forest Carbon Monitoring, Accounting, and Reporting System is to estimate forest ecosystem C stock changes from 1990 to the present, and to predict, given various assumptions, C stock changes in the next two to three decades. The system therefore must be able to process data on past events into estimates of C stock changes, as well as simulate future forest changes based on assumptions about forest management, natural disturbances, and other ecological processes.
Current development of Canada's forest C accounting system focuses on four spatial scales:
To ensure consistency of the estimates across spatial scales, the model uses the same basic principles and accounting rules at each scale. The spatial resolution and aggregation of the input data, however, differ between scales. At the scale of operational management units, all forest cover polygons can be represented as separate stands. At the national scale, stands with similar attributes in a specified area are aggregated. The impacts of the analysis scale and aggregation on C stock estimates will be explored.
At the core of the system is the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS2) a stand and landscape-level model of forest ecosystem carbon dynamics (Kurz et al. 1992, Kurz and Apps 1999, Apps et al. 1999). The model has been applied to analyze past and future forest biomass and dead organic matter C stock changes in the entire forest area, in the managed forest, in individual provinces or regions, and at the scale of operational units. The CBM-CFS2 has been used to explore numerous possible future scenarios that assume a range of natural disturbance rates, harvest rates, growth rates, decomposition rates, and combinations of these.
The CBM-CFS2 is undergoing extensive revisions in preparation for its future role in the National Forest Carbon Accounting Framework. The model is currently a research tool, but work is in progress to make the model more operational and accessible to analysts. The resolution of the model has greatly increased, allowing the explicit simulation of millions of stands in annual time steps.
At the stand level, the CBM-CFS2 tracks the dynamics of aboveground and belowground biomass carbon pools, as well as the dynamics of several dead organic matter pools that represent standing dead trees, coarse woody debris, litter and soil C. The model accounts for the C uptake through biomass growth, and for the transfer of biomass to dead organic matter pools through litterfall, mortality, management actions and natural disturbances. The model simulates the decay of all dead organic matter pools and the release of carbon associated with slash burning and fires.
At the landscape level, the CBM-CFS2 accounts for the impacts of management actions, land-use changes and natural disturbances such as wildfire and various insects on the age-class structure and distribution of different forest types. The model integrates over space and time, all C stocks and C stock changes and reports a variety of indicators using either full C accounting or specific Kyoto Protocol accounting rules.
The information on the current or past state of the forest in the area to which the model is applied is derived from forest inventory systems. Depending on the scale of the application these can be either spatially explicit or non-spatial summaries of the relevant information. Forest inventories are obtained from forest industry, provincial and territorial resource management agencies, or from federal compilations such as Canada's National Forest Inventory (Bonnor 1985).
The previous national compilations of forest inventory data are based on a large number of source inventories, collected over a number of years employing a wide range of methods and covering areas that are changing over time. It is therefore inappropriate to attempt to infer stock changes from consecutive (past) inventories. (Past C budget modelling efforts with the CBM-CFS2 therefore always combined a single inventory with information about stand (biomass and dead organic matter) dynamics and landscape-level (disturbances, management) changes).
The new National Forest Inventory (NFI, Gillis 2001) of Canada is intended to become a true change inventory allowing, for the first time, statistically valid assessment of forest attributes and their changes over time. The NFI sample protocol is based on a regular grid covering all of Canada's land area. Photo plots (2 x 2 km) will be established at grid intersections (20 x 20 km) and their characteristics will be recorded from aerial photographs or remote sensing information. Additional ground plots will be established at a subset of the sample plots to measure a range of plot attributes, including biomass and dead organic matter pools. The national grid system and sampling protocols have been developed and plot establishment (photo and ground plots) has commenced in several provinces and territories, while others are at the planning stages. It is currently assumed that all sample plots will be established by 2006 and that these will be re-inventoried in 5 to 10 year intervals.
The current design for the forest C accounting framework calls for use of the "best available" inventory information at various spatial scales (Figure 1). By the end of the first commitment period (Dec. 31, 2012), the NFI will provide much of the inventory information required for the C stock analyses at the national scale. The system will be designed such that it can, for specific areas, also make use of provincial or industry forest inventories whose coverage goes beyond the NFI plots. The forest C accounting system ensures that all required areas are accounted and no double counting occurs.
Remote sensing of forest area and its characteristics will play an important role in the national forest C accounting framework. For example, the Earth Observation for Sustainable Development (EOSD) team of the Canadian Forest Service is, in collaboration with the Canadian Space Agency (CSA), developing a very high resolution (30 m pixels, Landsat 7) forest cover map of Canada for the year 2000. Such information can be used to enhance forest inventories and to extend inventories to areas that are currently not or only poorly covered, such as non timber-productive forests and parks.
Remote sensing will also play an important role in the detection of forest changes and land-use change (Leckie et al. 2002). Developing and maintaining a database on land-use changes (afforestation, reforestation and deforestation) as required under Article 3.3 of the Kyoto Protocol is a major initiative involving the close cooperation of the Canadian Forest Service, Environment Canada, Agriculture and Agrifood Canada, Statistics Canada and resource management agencies in all Provinces and Territories.
The Canadian Forest Service, in cooperation with provincial and territorial resource management agencies is developing and maintaining a database on large (< 200 ha) fires, which cover about 97% of the area burned in Canada (Stocks et al. 1996). Similar databases for insect disturbances are being compiled. This information is essential for the simulation of the impacts of fires and insects on carbon dynamics (Kurz and Apps 1999). In combination with information on annual fire conditions it can also be used to estimate direct fire emissions (Amiro et al. 2001).
Over the past decade, government agencies and forest industry in Canada have established over 100,000 permanent sample plots and a bigger number of temporary sample plots. These provide a large body of knowledge about the characteristics and dynamics of forest stands in Canada. In many provinces, the growth and yield information is tied to ecological and site classification systems, thus allowing for the regional extrapolation of estimates.
Yield tables, empirical yield models and other growth and yield modelling approaches are the primary sources of information on stand dynamics currently used in the CBM-CFS2. This ensures that the model is data driven and that the analyses about future C stock changes are consistent with the regional and provincial timber supply analyses and other planning processes.
The disadvantage of empirical yield functions is that typically they are not responsive to changes in environmental conditions. Between-year variations in growing conditions can result in deviations in the annual growth dynamics relative to those observed in the past. Development plans for the CBM-CFS2 include the ability to either incorporate the growth predictions from process models (e.g. Bernier et al. 1999) or to quantify the impacts of environmental variation such that empirical growth curves can be modified in response to those changes.
Growth and yield data usually quantify the dynamics of merchantable volume but carbon stock estimates are required for the total biomass pool. Expanding estimates of volume to stemwood, branch and foliage biomass requires stand-level expansion factors. A group of CFS scientists is currently developing such biomass expansion factors based on individual tree biomass regression equations and permanent sample plot data (Gillis and Magnussen, pers. comm.).
The dynamics of belowground biomass, i.e. coarse and fine roots, are simulated based on functional relationships between aboveground and belowground biomass (Kurz et al. 1996). Motivated by the observation of higher than expected estimates of belowground net primary production in the model (Li et al. 2002), the approach to estimating fine root production has recently been revised (Li et al. in press).
Forest inventories typically do not include much information on dead organic matter C pools including litter and soil C. In the new NFI sample plots, information on these pools will be collected, and some newer provincial and territorial inventories also contain relevant information.
The CBM-CFS2 uses a simulation approach to link the dynamics of dead organic matter pools to the dynamics of forest stands and to the impacts of past disturbances and management actions (Kurz and Apps 1999). The model tracks the dynamics of four pools, characterized by the type of biomass input and their turnover rates. To facilitate the comparison between predicted and observed dead organic matter pools (e.g. Bhatti et al. 2000), model revisions are under way to increase the number of pools, including the introduction of pool of standing dead trees. Planned revisions to the model will also include a review of the decay rates in response to recent findings of long-term decomposition rate experiments (Trofymow et al. 2002).
National compilations of soil C information (e.g. Siltanen et al. 1997) do not contain information on past natural disturbances or management actions or on current stand conditions. Since these do impact the composition and size of dead organic matter pools, a new national compilation of ecosystem carbon stock information that will include, where available, data on past disturbances and management actions as well as on the current stand conditions (Bhatti and Shaw, pers. comm.). The representation of dead organic matter dynamics in the model will be refined as improved regional data become available for calibration and validation of the model.
The National Forest Information System (NFIS) is under development under the guidance of the Canadian Council of Forest Ministers. The purpose of the NFIS is to increase access to up-to-date source data on Canada's forests. These data are compiled and maintained by custodial agencies in Provinces, Territories, the Federal Government, the forest industry and other groups. This internet-based system will provide the standard, protocols and infrastructure to facilitate information sharing and the access to information and results of analyses.
The National Forest Carbon Accounting and Monitoring Framework will be able to draw upon information sources made accessible through the NFIS and in turn will be able to provide access to analytical results through the NFIS.
Although the design and concepts for a national forest C monitoring, accounting and reporting system are evolving, many details and specifics have not yet been defined. This process will involve continued consultation with federal departments, Provinces, Territories, forest industry, universities, and other communities.
Development of Canada's National Forest Carbon Accounting Framework is supported in part through funding from the Climate Change Action Fund, Action Plan 2000, and the Panel on Energy Research and Development. The development of the CBM-CFS2 and the supporting infrastructure and science involves a large number of scientists and other experts in the Federal Government, in Provincial and Territorial Resource Management Agencies, in universities and industry. The suggestions, support and data provided by that community are greatly appreciated - without this continuing support an integrating project of this scale would never be possible.
Amiro, B.D., J.B. Todd, B.M. Wotton, K.A. Logan, M.D. Flannigan, B.J. Stocks, J.A. Mason, D.L. Martell, K.G. Hirsch, 2001. Direct carbon emissions from Canadian forest fires, 1959-1999. Can. J. For. Res. 32: 512-525.
Apps, M.J., W.A. Kurz, S.J. Beukema and J.S. Bhatti, 1999. Carbon budget of the Canadian forest product sector. Environ. Sci. Technol. 2: 25-41.
Bernier, P.Y., R.A. Fournier, C.H. Ung, G. Robitaille, G.R. Laroque, M.B. Lavigne, R. Boutin, F. Raulier, D. Pare, J. Beaubien and C. Delise, 1999. Linking ecophysiology and forest productivity: an overview of the ECOLEAP project, The Forestry Chronicle 75: 417-421.
Bhatti, J.S., M.J. Apps and C. Tarnocai, 2000. Estimates of soil organic carbon stocks in central Canada using three different approaches. Can. J. For. Res. 32: 805-812.
Bonnor, G.M., 1985. Inventory of forest biomass in Canada. Can. For. Serv., Petawawa Natl. For. Inst., Chalk River, Ontario.
Gillis, M.D., 2001. Canada's National Forest Inventory (responding to current information needs). Environmental Monitoring and Assessment 67: 121-129.
Kurz, W.A. and M.J. Apps, 1999. A 70-year retrospective analysis of carbon fluxes in the Canadian forest sector. Ecological Applications 9(2): 526-547.
Kurz, W.A., M.J. Apps, T.M. Webb and P.J. McNamee, 1992. The carbon budget of the Canadian forest sector: phase I. Forestry Canada, Northwest Region. Information Report NOF-X-326, 93 pp.
Kurz W.A., M.J. Apps, E. Banfield and G. Stinson, 2002. Forest carbon accounting at the operational scale. The Forestry Chronicle 78: 672-679.
Leckie D.G., M.D. Gillis and M.A. Wulder, 2002. Deforestation estimation for Canada under the Kyoto Protocol: a design study, Can. J. Remote Sensing, 28: 672-678.
Li Z., M.J. Apps, E. Banfield and W.A. Kurz, 2002. Estimating net primary production of forests in the Canadian Prairie Provinces using an inventory-based carbon budget model. Can. J. For. Res. 32: 161-169.
Li, Z., W.A. Kurz, M.J. Apps and S.J. Beukema (in press). Belowground biomass dynamics in the Carbon Budget Model of the Canadian Forest Sector: recent improvements and implications for the estimation of NPP and NEP. Can. J. For. Res.
Siltanen, R.M., M.J. Apps, S.C. Zoltai, R.M. Mair and W.L. Strong, 1997. A soil profile and organic carbon data base for Canadian forest and tundra mineral soils. Natural Resources Canada, Canadian Forest Service, Northern Forestry Centre, Edmonton, Alberta. 50 p.
Stocks, B.J., B.S. Lee and D.L. Martell, 1996. Some potential carbon budget implications of fire management in the boreal forest. Pages 89-96 in M.J. Apps and D.T. Price Forest ecosystems, forest management and the global carbon cycle. Springer-Verlag, Heidelberg, NATO ASI Series I 40.
Trofymow, J.A., T.R. Moore, B.D Titus, C. Pescott, I. Morison, M. Siltanen, S. Smith, J. Fyles, R. Wein, C. Camire, L. Duschene. L. Kozak, M. Kranabetter, S. Visser, 2002. Rates of litter decomposition over 6 years in Canadian forests: influence of litter quality and climate. Can. J. For. Res. 32: 789-804.
Figure 1: The Carbon Budget Model of the Canadian Forest Sector is operating at 4 spatial scales and uses the best available information on forest inventories, growth and yield, natural disturbances and other ecosystem processes at each of these scales.
1 Natural Resources Canada, Canadian Forest Service, Pacific Forestry Center, 506 West Burnside Road, Victoria, British Columbia, V8Z 1M5, Canada. [email protected]