An inventory, polygon by polygon
Follow one 24 × 24 km window in Wabigoon Forest, picked because much of it was harvested after 2000. Thin lines are inventory polygon edges: 1,000 polygons, median forest stand 10 ha.
Trout Lake and Wabigoon forests · Ontario · 2000–2010
The Boreal Ledger
A decade of carbon in two northwestern Ontario forests: what the stands held, what harvest and fire took out, and what growth put back, counted stand by stand with the Carbon Budget Model of the Canadian Forest Sector.
- Forest modelled
- 1.03 Mhaof inventory forest
- Stands
- 61,600inventory polygons
- Period
- 10 yrs2000 to 2010, annual
- Disturbed area
- 91k haharvest, fire, blowdown
Carbon budget, 2001–2010 average · t C/ha/yrTrout LakeWabigoon
Net primary production is growth after the trees' own respiration. A negative total means the forest lost carbon over the decade: harvest and slash burning took out slightly more than growth put back.
Biomass carbon, 2010
Data and model
| Dataset | Provider | Year | Unit | Used for |
|---|---|---|---|---|
| Forest Resource Inventory, Wabigoon Forest (130) | Ontario MNRF | 2010 | stand polygons | Stands, species, age, disturbance history |
| Forest Resource Inventory, Trout Lake Forest (120) | Ontario MNRF | 2013 | stand polygons | Stands, species, age, disturbance history |
| Forest management unit boundaries | Ontario MNRF | Wabigoon 2010 · Trout Lake 2025 | vector | Study areas |
| Operational-Scale CBM-CFS3, driven with cbm3_python | Natural Resources Canada · Canadian Forest Service | toolbox build 1.0.8956 | t C | Carbon pools and fluxes, Boreal Shield West parameters |
| Growth curves | This analysis | — | m³/ha by age | One softwood and one hardwood curve |
The model
How CBM-CFS3 keeps the books
The Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3) is the Canadian Forest Service's model of forest carbon, and the core of the system Canada uses to report forest emissions and removals in its national greenhouse gas inventory. It does not simulate photosynthesis. It reads merchantable volume from a yield curve, converts it to biomass with national volume-to-biomass equations, and every year moves carbon between 25 pools using turnover rates set for the ecozone and decay rates adjusted for its climate.
Harvest, fire and other events are disturbance matrices. In the year they happen, each one moves a fixed share of every pool to other pools, to the atmosphere, or out of the forest as wood, and resets the stand's age. Before the first year, a spin-up cycles each stand through its historical disturbance until the slow soil pools settle, then grows it to its inventory age.
One hectare of Wabigoon Forest
1Seven input tables
sit_inventory.csv sit_classifiers.csv sit_yield.csv sit_events.csv sit_transitions.csv sit_age_classes.csv sit_disturbance_types.csv
Built from the inventory: one row per stand and species group, its growth curve, and the events that hit it. A mapping file ties them to CBM's default species, disturbance types and the Boreal Shield West spatial unit.
→ CSV tables + mapping.json2Standard Import Tool
sit_helper.csv_import(
csv_dir="sit_directory",
imported_project_path=
"sit_trial_stack1.mdb")
Checks the tables against each other and writes them into a CBM-CFS3 project, an Access database.
→ project database3Makelist and simulation
projectsimulator.run(
project_path=
"sit_trial_stack1.mdb",
results_database_path=
"..._result.mdb",
n_timesteps=30,
skip_makelist=False)
Makelist turns each yield curve into biomass curves and builds the simulation list; CBM then spins every stand up and steps the inventory forward one year at a time.
→ results database, one per stack4Read the results
cbm3_results.load_pool_indicators( results_path, classifier_set_grouping=True)
Pools, fluxes, stock changes and disturbed area come back from tables such as tblPoolIndicators and tblDistIndicators, for every stand and year. This page reads 2000 to 2010.
→ the charts and maps belowForest by forest
What goes into each forest, and what comes out
A forest management unit is the area one licence holder plans harvest for under Ontario's forest management plans. Its Forest Resource Inventory maps every stand as a polygon with its species mix, age and any recorded harvest, fire or blowdown. Each forest stand becomes a CBM record; the results come back per stand, so they can be mapped the same way.
Forest management unit 130
Wabigoon
Wabigoon · Leading species
Wabigoon inventory at a glance
Forest area by stand age in 2000
thousand ha · age in yearsLeading species
share of forest areaDisturbed area by year
thousand ha · from the inventoryMethod
From inventory polygons to carbon pools
The inventory is turned into the seven tables CBM-CFS3's Standard Import Tool reads, the model steps each forest forward one year at a time from 2000, and the pool and flux indicators are read back from the results databases for every stand. The runs continue past 2010, but the inventories record no harvest or fire after that, so this page stops at 2010. Each stage is a notebook on GitHub: preprocessing, running CBM-CFS3 and post-processing.
Processing chain
STEP 1Forest standsKeep productive forest polygons
- Forest Resource InventoryPOLYTYPE = FOR
STEP 2Species groupSoftwood or hardwood; mixed stands split when the other group is ≥ 40%
- Species compositionOSPCOMP
STEP 3Age in 2000Inventory age moved to 2000, in 10-year classes
- Age and inventory yearOAGE, YRSOURCE
STEP 4Growth and eventsYield curve per group; each depletion after 2000 becomes an event in its year
- Depletion type and yearDEPTYPE, YRDEP
- Default CBM typesclear-cut with slash-burn, wildfire
STEP 5Run CBM-CFS3Inventory split into stacks, one-year time steps
- Boreal Shield West, Ontarioecological parameters
STEP 6Pools and fluxesRead back per stand and per year, then summed per forest
25 carbon pools · 12 biomass, 13 dead organic matter
Annual fluxes · NPP, decay, harvest, emissions
Per-stand maps · 2000, 2005, 2010
Growth curves
Two curves drive every stand
CBM grows biomass from merchantable volume by age. This analysis gives every softwood stand the same curve and every hardwood stand another, rather than curves by species and site class. Stand age decides where on the curve each stand starts in 2000.
Merchantable volume by stand age
Softwood
Hardwood
Species set the growth curve
The leading species decides whether a stand grows on the softwood or the hardwood curve. Broadleaf here means aspen, poplar and birch.
Age sets the starting point
Inventory age is moved back to 2000. Stands that originated after 2000 are started at 100 years and then reset by the disturbance that created them.
Harvest and fire happen in their year
Each depletion recorded after 2000 becomes a CBM event in that year. Harvest is modelled as clear-cut with slash-burn and fire as wildfire; both reset the stand to age zero.
The model's starting stock
CBM spins each stand up to its age in 2000 and fills its biomass and dead organic matter pools.
Mean biomass carbon in the window: 40 t C/ha
Cut blocks empty out, the rest keeps growing
Harvested and burned stands drop to near zero biomass; their carbon leaves as wood products or emissions, or moves into deadwood and soil.
Mean biomass carbon: 40 t C/ha
Subtract the two years
Stands cut or burned in the decade lose nearly all their biomass carbon; the stands left standing each add a few tonnes per hectare.
Lost more than 20 t C/ha: 1,000 ha · untouched stands gained 3 t C/ha on average
Polygon types
Results
Where the carbon sits, and where it moves
Carbon stock per hectare by pool, 2000–2010
Show the numbers
The yearly carbon budget
Net ecosystem production (growth minus decay)
Harvest removals
Emissions from slash burning and fire
Net change in ecosystem carbon
Show the numbers
A harvested stand
Harvest empties biomass in a year; seven years later it is barely back
Carbon in stands harvested 2001–2003, followed to 2010
Biomass
Dead organic matter
Wabigoon, 2000 and 2010
A decade of cut blocks, side by side
Wabigoon Forest, biomass carbon in 2000 and 2010
Age structure
Ten years older, except the cut blocks
Forest area by age class, 2000 and 2010
2000
2010
Checks on the run
The books balance
Every number on this page was read straight from the CBM-CFS3 result databases. These checks confirm the runs hold together: area, carbon accounting, events and the per-stand extraction.
Disturbance requested and applied
Caveats
Before quoting a number
Keep in mind
- Carbon is in tonnes of carbon (t C), not tonnes of biomass; biomass is roughly twice the carbon.
- Results stop at 2010. The inventories record harvest and fire up to 2009 in Wabigoon and 2010 in Trout Lake, so later years of the runs carry growth and decay only and are left out.
- Every stand uses one of two generic yield curves, so differences between species and site classes are not modelled.
- The exported 2010 biomass product stores biomass carbon (t C/ha) under a Mg/ha label. Relabel it, or roughly double the values, before using it as dry biomass.
- Blowdown, drought and flood map to CBM's "Unknown" disturbance type: they reset stand age but move no carbon in these runs.
- Harvest is modelled as clear-cut with slash-burn, so every harvest also burns its slash; that burning is most of the disturbance emissions.
- Stands that originated after 2000 start the run at 100 years old, then reset when their recorded disturbance happens.
- A few Trout Lake stands share a shortened polygon ID; their carbon is averaged between them on the maps.
Next steps
- Replace the two generic yield curves with Ontario growth and yield curves by species and site class.
- Add harvest and fire after 2010, from the annual reports or national satellite records, to carry the story forward.