Emissions calculations for Harvested Wood Products and Forest Land
Second in the sector-by-sector National Greenhouse Gas Inventory computation: Harvested Wood Products and Forest Land. Data from Natural Resources Canada on harvested wood volume supports a satisfactory estimate of Harvested Wood Products emissions, and a first step toward a Forest Land estimate.
Table of Contents:
Introduction
The previous post introduced the International Panel on Climate Change (IPCC) sectors that Canada uses to report greenhouse gas emissions, and estimated Public Electricity and Heat emissions from Statistics Canada data. This post continues that replication effort for the next-largest source of emissions in 2005: Forest Land. As it turns out, Forest Land emissions are tightly coupled to emissions of the Harvested Wood Products sector, so this post looks at both. The material from this post refreshes the IPCC Sectors / Forest Land page, introduces a IPCC Sectors / Harvested Wood Products page, and introduces emissions estimates for both sectors based on data from Canada's National Forestry Database. The estimate is relatively accurate for Harvested Wood Products, and not as accurate for Forest Land. Proper replication of the Forest Land calculation seems difficult. I might come back to it in future, but I expect to need help from someone with domain expertise.
Forest Land
Canada's forests are enormous, and collectively, they exchange a lot of carbon dioxide with the atmosphere and the soil in any given year. Plants in forested land absorb carbon dioxide from the atmosphere to grow (via photosynthesis), and release carbon dioxide to the atmosphere when they die (via e.g. combustion or decay). When forests burn, much of the carbon from the wood is released to the atmosphere, but some is converted to biochar and effectively locked away. When plants decay, microbes break down the wood and release most of the carbon as , but a fraction of the carbon is converted to stable organic matter called humus that remains in the soil. The long term carbon balance of soil, plant matter, and atmosphere can reach equilibrium in a few hundred years. This equilibrium can recover from disturbances by e.g. fires, insect infestations, and logging. The equilibrium can be indefinitely shifted by e.g. land-use conversion, or environmental changes in temperature, moisture, or atmospheric concentration.
Forest carbon accounting is an active area of scientific research, and Canada is an international leader. The Canadian CBM-CFS3 model, and the team that developed it helped define the IPCC guidelines on reporting carbon for forest land. The CBM-CFS3 model is now being replaced with a new model, the Generic Carbon Budget Model, developed within the context of a global open-source collaboration with Australian greenhouse gas experts and the Linux Foundation. Integrating PlanZero with this more-established project is an interesting direction for future work (github issue).
Natural Resources Canada (NRCan) works together with provincial and territorial ministries of natural resources to manage and model Canada's forests for a variety of purposes, including the estimation of direct and indirect greenhouse gas emissions from forests. One artifact of that management is the National Forest Inventory, and for my goal of understanding how Canada's emissions calculations work, it's a great starting point. Canada's emissions in the Forest Land sector are calculated only from so-called Managed Forest.
- Managed Forests: These are forests under direct human influence, including lands managed for timber harvesting, fire protection, or conservation. This covers roughly 226 million hectares (about 65% of Canada's total forest area).
- Unmanaged Forests: Remote northern forests not subject to active human stewardship are generally excluded from the inventory until human activity (like a new road or mine) expands into them.
Harvested Wood Products
Harvested Wood Products represents one of the largest out-flows of carbon from the Forest Land sector. The IPCC accounting guidelines treat Forest Land as a carbon pool with in-flows and out-flows from natural and anthropogenic drivers, and logging is one of the biggest and best-tracked of those drivers. The Harvested Wood Products sector is, itself, also modelled a set of carbon pools, corresponding to volumes of wood used for different purposes, with different lifespans before the carbon they contain returns to the atmosphere. The product categories used in the National Forestry Database are:
- Logs and Bolts for lumber, plywood, veneer and shingles
- Other Industrial Roundwood for poles, pilings, fence posts, shoring timbers
- Fuelwood and Firewood for residential, commercial, or industrial combustion
- Pulpwood for producing paper and cardboard.
Estimated HWP emissions were calculated in terms of carbon (a) captured by the HWP pool, shown as negative values in the figure, and (b) carbon released by the HWP pool, shown as positive values in the figure. The net HWP emissions are shown as the solid grey line, and the reference emissions for the sector (from the 2025 National Inventory Report, NIR) are shown by the solid black.
The EstForestAndHarvestedWoodProducts class in PlanZero's est_nir.py file
has the details but at a high level,
the HWP pool was implemented as four carbon pools, one per product category.
The "Harvest" data set from the National Forestry Database lists volumes
of wood harvested by year, by province, by species group, and by product category.
(The data I downloaded only covers years up to 2021, but I hope that newer data can be downloaded from
zenodo.org once their service resumes.)
A per-species-group factor estimated carbon mass from harvested wood volume,
and that mass was modelled as being added
to the HWP pool associated with the product category at each given year.
To subtract carbon from the pools, the pools were modelled as decreasing by a certain
factor 1/T every year, with the value of T coming from Annex 6 of the ECCC's National Inventory Report,
specifically from Table 5-3. Logs, bolts, and other industrial roundwood were modelled as lasting
the longest (T=35), and pulp and paper lasting only a short time (T=2).
Firewood and fuelwood was modelled as being combusted in the year of harvest.
The quality of fit of this model is acceptable to me at this point, but it is admittedly imperfect. My estimation of net HWP emissions is too low by a few Mt in the mid-1990s, and too high by various amounts since 2001, suggesting that at least some of the Logs and Bolts volume should be modelled as being released a little more slowly. The time constants provided in Annex 6 Table 5-3 of the 2025 NIR indicate lumber exported to the US does tend to last longer, but I don't yet incorporate data on how much lumber was exported. The jagged shape of associated with Logs and Bolts harvest vs. the smooth shape of the reference total from the national inventory data is, candidly, troubling, and I wish I could explain it. Still, the overall shape of the net HWP emissions estimate matches the reference better than nothing (which I started with), so I feel progress has been made toward understanding emissions in the sector. I have updated the IPCC Sectors / Harvested Wood Products page with this graph and a sector overview.
Approximating Forest Land with Harvested Wood Products
Coming back to the post's original purpose of estimating Forest Land emissions, recall that we were looking at the HWP sector because it was one of the easiest-to-track drivers of carbon removal from Forest Land. How much progress have we made in modelling emissions for Forest Land, simply by counting the carbon content of harvested lumber as being emitted as ?
Unsurprisingly, perhaps even reassuringly, the fit to the reference emissions curve for Forest Land is neither terrible nor great. The removal of HWP carbon is certainly better than the implicit constant 0 that we would have without any modelling: the mean is only about 20 Mt off, both curves have a sharp dip in 2008, and the overall trend over the last 30 years is qualitatively similar. Still, this approximation is missing almost everything about forests, such as the net growth of managed forests, losses due to wildfire and insect damage, and losses due to land-use conversion. I'll use this figure to refresh the IPCC Sectors / Forest Land page, and move on to other sectors for now, but clearly this approximation leaves much to be desired.
Summary
In this second post developing an estimate of Canada's National Greenhouse Gas Inventory, we looked at two IPCC sectors: Harvested Wood Products and Forest Land. They are tightly coupled, in that one of the driving factors in the emissions from Forest Land is the removal of carbon via the harvest of trees. IPCC reporting guidelines recommend modelling both Forest Land and HWP in terms of carbon pools. Physically, emissions from HWP come from product decay over time, and in the model, per-product-category carbon pools decay by a constant factor from year to year. In the model, the HWP pools are topped up by the carbon that has been removed from Forest Land by the harvesting of trees. The PlanZero approximation to net HWP emissions does not account for different decay rates for exported lumber, and does not account for the smoothness of the reference value, but is usually within 10 Mt . I have updated the IPCC Sectors / Harvested Wood Products page with the figure developed in this post and a sector overview.
The official NRCan Forest Land model is too detailed for me to incorporate into PlanZero with the sector-by-sector time-boxed efforts I'm doing with this series. It seems from the national inventory report annex 3.5 on LULUCF methodology that forests are modelled in terms of geography (ecozones, provincial boundaries), ecology (soil at various depths, carbon content of root systems, bark, branches, leaves), and time (rates of growth, carbon transfer matrices in response to various disturbances). Perhaps at some point, PlanZero can be integrated with the GCBM open-source software developed by NRCan to incorporate the reference model directly. For now, I have updated the IPCC Sectors / Forest Land page with the HWP approximation, such as it is.
Non-Fun Fact: Forests and Harvested Wood Products Are Net Emitters
Forest Land and Harvested Wood Products represent a carbon cycle through the logging industry. Their net effect over the last 30+ years in terms of emissions has been to transfer carbon from forests into the atmosphere.
Although I have not modelled the absorption of carbon by forests here in PlanZero, we can see the limits of its effect. The black "Combined" reference line in the figure above shows that according to Canada's official report to the United Nations, based on the gold-standard modelling and data of NRCan, the net effect of Canada's Forest Land and HWP sectors, in every year since 1990, has been to emit more carbon into the atmosphere than has been captured. The IPCC guidelines recommend to think of Forest Land and HWP a carbon pools rather than a carbon sinks because on the time scales required to achieve net-zero, most of the trees that are growing now will later die, and most of the wood products made today will later decay, and most of the carbon in these things will be cycled back into the atmosphere. Furthermore, according to the best science, those pools are being drained in Canada, rather than replenished. We can, and perhaps should, research how to increase the capacity and degree-of-fill of these carbon pools. It seems like it should be very possible to achieve a sustainable forestry sector, given that forests are naturally sustainable. At the same time, we should temper any expectations that even if we can manage to start replenishing these carbon pools, that Canada's forests will substantially counterbalance the scale of carbon emissions currently emitted by other sectors. The carbon pools of Canada's forests were already full at the beginning of the industrial revolution, when the world's fossil fuels were still underground.
The next post in this series, still working in decreasing order of emissions magnitude in 2005, will be fugitive emissions due to the venting of natural gas.
- © James Bergstra, 2026.
- Licensed under CC-BY.
- Statements on this site are not necessarily true or fit for any purpose.
- Design: HTML5 UP