This category covers the burning of biomass where it occurs in the open. It includes all fires in natural and managed ecosystems including forests, woodlands, shrublands, grasslands, savannah, plantations, and all fires in agricultural lands. This section does not address any process that converts biomass into another form of energy such as steam, controlled combustion in appliances such as stoves, furnaces and boilers. Such processes are covered in Group 3 Power Generation and Heating.

Fires in natural and managed ecosystems include wildfires regardless of the ignition source and all fires conducted for land management including fuel reduction for wildfire mitigation, biodiversity management, forest slash removal and seed bed preparation following logging operation, and the removal of forest biomass following land clearing for conversion to agriculture and other land uses. Ignition sources include lightning, deliberate arson, accidental ignitions (e.g., burning cigarettes, glass, welders, power transmission lines) and authorized ignitions for management purposes. A broad definition of forest class is applied in the Toolkit. It includes temperate and tropical rainforests; temperate, boreal, and Eucalyptus forests; temperate shrublands; tundra; peatlands; and forest plantations.

Savannah woodlands and savannah grasslands are ecosystems comprising open and sparse woodlands dominated by an extensive understory of grasses. The annual climate is a short productive wet season followed by a long drought during which the grasses senesce and the fuels dry. Fire is an essential and integral component of these ecosystems and has been traditionally used by indigenous populations for many millennia to manage animal and plant food sources (Russell-Smith et al. 2009a). Fire return interval in these regions is short, typically less than three years (Archibald et al. 2010). Fires in these ecosystems comprise the major proportion of global annual fire affected land area (Giglio et al. 2006).

Fires are also commonly used in agriculture. Post-harvest field burning is practiced to remove residues prior to soil preparation and sowing, to control weeds, and release nutrients for the next crop cycle, however, also negative effects on soil may occur and, ideally, this practice is used judiciously. It is applied extensively, but not exclusively, in cereal production (wheat, rice, maize, and coarse grains) in many regions. Pre-harvest burning is used in some crops, particularly sugar cane, to remove debris and pests and to facilitate manual and mechanical harvesting.

Emission Factors

Emission from this source category will vary depending on ambient atmospheric conditions, fuel type and structure, composition, and contamination with PCDD/PCDF precursors. When biomass fuels are wet or compacted, burning efficiency is poor, combustion temperature is low, and consequently, PCDD/PCDF emissions are high. PCDD/PCDF emissions are also assumed to be high when the biomass has been treated with pesticides that act as PCDD/PCDF precursors or as catalysts for PCDD/PCDF formation, in these cases the biomass is said to be “impacted”. At the other extreme, burning of dry, virgin biomass of small size would burn efficiently and have small emission factors.

Emission factors for PCDD/PCDF are listed in Table II.6.3, and for dioxin-like PCB are listed in Table III.45.1 in Annex 45. Detailed information on the derivation of default emission factors can also be found in Annex 45. Revised or newly added PCDD/PCDF emission factors are highlighted in red.

Table II.6.3 PCDD/PCDF emission factors for source category 6a Biomass Burning

6a	Biomass Burning	Emission Factors (µg TEQ/t biomass burned)
Classification		Air	Water	Land	Product	Residue
1	Agricultural residue burning in the field, impacted, poor burning conditions	30	ND	10	NA	NA
2	Agricultural residue burning in the field, not impacted	0.5	ND	0.05	NA	NA
3	Sugarcane burning	4	ND	0.05	NA	NA
4	Forest fires	1	ND	0.15	NA	NA
5	Grassland and savannah fires	0.5	ND	0.15	NA	NA

Guidance for Classification of Sources

Class 1 includes the open burning of agricultural biomass in the field under conditions that may favor increased PCDD/PCDF formation and release. Although little experimental data exist, it is assumed that prior application of chlorinated pesticides to crops would increase PCDD/PCDF formation and release. Other contributing factors include unfavorable burning conditions such as large piles or humid materials. In general, the biomass may be cereal, legume oilseed or fibre crops and can be burned as stubble, cut and left in the field or bundled into piles. This class potentially includes a wide range of fire intensity from relatively cool, low-intensity fires with mostly smouldering combustion to hot, high intensity, fast-moving fires with efficient combustion.
Class 2 addresses the same type of biomass and geometry of the fuel; however, the fuel and the burning conditions would constitute best environmental practices such as the absence of precursors or other conditions that favor PCDD/PCDF formation. These fires also range from cool to hot fires.
Class 3 addresses the pre-harvest burning of sugarcane in the field. It is assumed that the leaves will be burned off and the stems will remain in the field for harvest either manually or by machine. These fires are usually intense, fast moving and of relatively short duration.
Class 4 includes all types of forest fires, including those in which whole trees are burned, canopy fires and forest litter burns.
Class 5 addresses fires in savannahs and grasslands. Fires in savannahs frequently consume low shrubs in addition to grass and litter from trees.

Activity rates

The activity for this source category is the mass of fuel consumed as tons dry matter. National data are not available in this form, and therefore the activity is calculated using other information, such as the total area in each emission class multiplied by the density of combustible fuel (e.g., tons dry matter consumed per hectare burned). Combustible fuel density is determined from measurements of above ground biomass of potential fuel classes and the fraction of this mass that is actually burned. Consequently combustible fuel density varies with vegetation class, fire class and season. A compilation of average fuel densities for most vegetation classes relevant to the Toolkit classes is provided in Table II.6.4.

The best source of information on where fires occur in a country may be fire departments. Departments and research institutions for agriculture and forest may be the best sources for the estimation of biomass grown in a given region.

Some of these data are collated into intern

ational databases, e.g., FAO agricultural statistics which can provide an alternative source of activity data in the absence of readily available local sources. There are also published international compilations of fire-affected area derived from remote sensing as both regional summaries and detailed spatial statistics (Van der Werf et al. 2006). These sources can be valuable supplements to national statistics and are useful for QA/QC assessment of inventory emission estimates.

For orientation, the information on burning efficiency for natural ecosystems such as forests and major crops in agriculture is presented in Table II.6.4. This table also gives advice as to the type of fire.

Table II.6.4 Summary of information on biomass fuel consumed in open fires (compiled from IPCC 2006, Volume 4, Chapter 2, Tables 2.4 and 2.5, and Russell-Smith et al. 2009b)

Ecosystem	Class	Fire category	Fuel burned (t dm/ha)
Natural Ecosystems
Tropical Forest	Primary		43
	Secondary		23
	Tertiary		32
Boreal		Wildfire	21
		Surface fire	3.2
		Post logging slash	23
		Land clearing	52
Eucalyptus		Wildfire	33
		Prescribed fire	10
		Post logging slash	115
		Land clearing	78
Other temperate forest		Wildfire	11
		Post logging slash	48
		Land clearing	25
Shrublands			10
Tropical savannahs	Woodland	Early dry season	2.8
	Woodland	Late dry season	4.2
Other savannah	Woodlands	Early dry season	0.6
	Woodlands	Late dry season	2.4
Savannah grasslands	Tropical/Subtropical	Early dry season	1.6
	Tropical/Subtropical	Late dry season	4.8
	Grassland	Late dry season	3.5
	Tropical Pasture	Late dry season	8.3
Other natural ecosystems	Peatland		21
	Tundra		5
Agricultural Systems
Wheat			3.6
Maize			8
Rice			4.4
Sugarcane			5.2

Table II.6.4 should be supplemented with country-specific data where available. Some examples of the amount of material expected to be involved in a biomass fire based on the land area involved are shown below.

In the UK, values were derived as follows:

• Heather moorland – material consumed in fires – 8 tons per hectare.
• UK forest – material consumed in fire – 23 tons per hectare.

For estimating the releases of PCDD/PCDF from biomass fires, countries may wish to apply the approach as shown for France – and adjust the numbers according to their climatic conditions and vegetation. In the French inventory, the following approaches have been used to estimate the PCDD/PCDF releases from forest fires (Béguier 2004):

• Forests/vegetation has been classified according to climatic zones resulting in temperate and Mediterranean (south of France) zones;
• In the temperate zone, forests typically have 20 kg of biomass per square meter (20 kg/m²) corresponding to 200 t/ha. In the Mediterranean zone, the biomass is 4 kg/m² or 40 t/ha;
• In the temperate zone on average, 20% of the vegetation is removed by the fires hence the fuel burned is 40 t/ha. In the Mediterranean zone, the fires are more efficient with an average of 25% above ground biomass removed; thus, the fires in southern France will generate PCDD/PCDF from 10 t of biomass per hectare affected by the fire.

The Philippines report that on average 43 t/ha is consumed in forest fires in their region.

Fuel loads for crops can be estimated from crop production data that are regularly collected and reported by agricultural companies and agencies. For pre-harvest burning of sugar cane, the following approximation to estimate the amount of biomass burned can be used: about 300 kg of biomass is burned for each ton of sugar produced (Choong Kwet Yive 2004).

For post-harvest burning, Southeast Asian countries used the following approximation to estimate the mass of rice straw burned: harvest residues of 25% w/w are being generated from rice. In other words: 250 kg of rice straw are generated per ton of (polished) rice produced.

Level of Confidence

6a	Biomass Burning	Level of Confidence
Classification
1	Agricultural residue burning in the field, impacted poor burning conditions	Medium	Value is extrapolated from common knowledge the processes
2	Agricultural residue burning in the field, e.g., cereal crops, not impacted	High	Relatively large number of consistent results in narrow range, relatively wide geographic coverage
3	Sugarcane burning	Medium	Relatively many consistent results published in peer-reviewed literature, largest range of results within this sub-category, limited geographic range
4	Forest fires	High	Relatively large number of consistent results in narrow range, relatively wide geographic coverage
5	Grassland and savannah fires	Medium	Limited number of very consistent results, small range of results

This source category includes the deliberate combustion of waste materials for disposal where no furnace or similar is used – for example the burning of domestic waste and other waste in piles in the open, the burning of waste in dumps – both deliberate or accidental, and fires in buildings, cars and other vehicles. In this source category, there is no recovery of the calorific content of the fuel.

As with the source classes under category 6a, releases of unintentional POPs in solid residues such as combustion ashes are regarded as releases to land rather than release to residue since the ashes are disposed on land and are typically not collected for further disposal. Therefore, to avoid double-counting, an EFLand is provided instead of an EFResidue.

Emission Factors

Emission factors for PCDD/PCDF are available for five classes as shown in Table II.6.5. Revised or newly added emission factors are highlighted in red. For dioxin-like PCB, available emission factors are listed in Table III.46.1 in Annex 46. Detailed information on how default emission factors have been derived can also be found in Annex 46, along with guidance on activity rates.

Table II.6.5 PCDD/PCDF emission factors for source category 6b Open Burning of Waste and Accidental Fires

6b	Open Burning of Waste and Accidental Fires	Emission Factors (µg TEQ/t material burned)
Classification		Air	Water	Land	Product	Residue
1	Fires at waste dumps (compacted, wet, high organic carbon content)	300	ND	10*	NA	NA
2	Accidental fires in houses, factories	400	ND	400	NA	NA
3	Open burning of domestic waste	40	ND	1*	NA	NA
4	Accidental fires in vehicles (µg TEQ per vehicle)	100	ND	18	NA	NA
5	Open burning of wood (construction/ demolition)	60	10	10	NA	NA

* Based on a few field measurements and consistent with the biomass burn EFLand where the release in the ashes is 5%-10% of the EFAir.

Guidance for Classification of Sources

Class 1 refers to spontaneous or intentional fires occurring in a municipal or domestic waste repository. Waste at such sites may often include refuse from offices, small factories or workshops and restaurants. In some cases, these fires have the purpose of reducing the volume of waste in the repository. Typically, the waste will be relatively high in organic carbon. The combustible material will tend to be compacted and moist, and will burn poorly and slowly; hence the higher emission factor than for class 3. Typically, ignition occurs from either sparks occurring at the surface area, from self-ignition inside the waste body or intentionally for management reasons. It should be noted that fires of this type are very uncommon in modern engineered landfills, particularly those with compaction, daily soil cover, runoff water recycling or leachate and landfilll gas collection.
Class 2 includes accidental fires involving buildings, such as homes and factories. Consequently, emission factors must be given per event and they depend strongly on the materials burned and on the nature of the fire. There is limited information on releases from these fires and a single indicative figure is given to cover all accidental fires excluding fires in vehicles. Chemical fires may lead to very high releases where certain precursor chemicals are involved. However there is insufficient information to assess releases from chemical fires as a distinct category so releases are included in this class of accidental fires. It should be noted that specific incidents may give rise to local contamination and potential “hot spots”.
Class 3 includes burning of domestic waste in open piles, pits, barrels, with no pollution controls. The waste is typically characterized by a large fraction of organic/agricultural waste and is loosely arranged (not compacted).
Class 4 includes fires that involve cars and other vehicles. Limited data are available for deriving emission factors for such events and vehicles involved can vary considerably so emissions are expected to vary as well. Consequently the emission factors here are for rough estimates only.
Class 5 includes open burning of wood and other materials used in construction and remaining after demolition. Such wood may be painted or treated with preservatives and plastics, including PVC, or may be present in the other materials burned.

Level of Confidence

Through recent studies, more data have been generated including results from developing countries, targeted to generate emission factors which are closer to real country situations. These studies and results cover a larger geographic area than before, and have been published in peer-reviewed literature. On the other hand, extrapolation still needs to be done to estimate the activity and the processes, which especially for source category 6b are not stable. Therefore, although the studies are of good scientific quality, the results are scattered due to heterogeneity of the processes, fuels, and other variables.

A new practical approach has been developed to allow inventory developers to better characterize the activity rate, i.e., estimate the mass of waste that is burned in the open air. The new method is presented in Annex 46.

6a	Biomass Burning	Level of Confidence
Classification
1	Fires at waste dumps (compacted, wet, high Corg content)	Medium	Few studies, consistently large range of results, published in peer-review literature
2	Accidental fires in houses, factories	Low	Very few results, large range of data; process is not stable
3	Open burning of domestic waste	Medium	Few studies, consistently large range of results, relatively large geographic area, published in peer-review literature
4	Accidental fires in vehicles (µg TEQ per vehicle)	Low	Very few studies with consistent data; process is not stable
5	Open burning of wood (construction/ demolition)	Low	Value is extrapolated from common knowledge the processes