Air Pollution as a Climate Forcing: A Workshop
Day 2 Presentations
Global Black Carbon Inventories
NOAA Pacific Marine Environmental Laboratory, Seattle, WA, U.S.A.
In collaboration with:
David Streets, Suneeta Fernandes, Sibyl Nelson,
Kristen Yarber — Argonne National Laboratories;
Jung-Hun Woo — CGRER, University of Iowa;
Zbigniew Klimont — IIASA
(The non-inventory portion of these ramblings is my own and should not be blamed on any of my colleagues.)
You may download a MS PowerPoint version (1.2 MB) of this presentation.
Bean-Counting & Uncertainties. We offer a new contribution to the grand tradition of BC inventories (Penner et al. 1993 through Cooke et al. 19991). Our inventory follows the custom of "bottom-up" inventories by applying emission factors to official fuel-use data (International Energy Agency, IEA). We supplement IEA data with information from other sources on biofuel and biomass burning. An innovation in the current inventory is the division of fuel-use sectors into technologies, instead of the use of country-specific, sectoral emission factors. Emissions at different levels of development are then represented by variations in the mix of technologies. Table 1 lists total emission estimates, broken down by region and by major source type, compared with calculations using emission factors from a previous inventory (Cooke et al. 19992). Gridding to the 1°x1° scale, shown in Figure 1, is based on urban, rural, or total population, or land-cover data, depending on the emission source. Within China, India and the U. S., the largest emission sectors are apportioned among provinces or states before gridding.
We estimated uncertainty in BC emissions by propagating uncertainties in emission factors and fuel use, following procedures suggested by IPCC and Cullen and Frey. Work remains to be done on refining these uncertainty estimates for the largest sectors. Table 1 and Figure 2 show the uncertainties, which are a factor of three overall but much higher in some areas. To grid uncertainties, we distribute the variances just like the emissions; this calculation does not include uncertainties in spatial variation, nor does it include a full accounting of missing sources. The inventory, gridding, and uncertainties are works-in-progress; comments, questions, and suggestions are welcome.
|North America||0.39 (0.33-1.56)||0.95||Power gen.||0.12 (0.10-0.49)||1.54|
|Latin America||0.99 (0.73-2.51)||2.41||Industry||0.29 (0.23-0.99)||1.19|
|Europe||0.44 (0.33-1.20)||1.01||Res. coal||0.90 (0.49-3.52)||0.64|
|Former USSR||0.36 (0.22-1.24)||0.87||Res. biofuel||1.24 (0.90-2.38)||2.19|
|Africa/Mid East||1.90 (1.38-4.07)||3.78||Res. other||0.18 (0.16-0.29)||0.50|
|China||1.19 (0.78-3.91)||2.80||Transport diesel||0.51 (0.43-1.22)||2.83|
|India||0.53 (0.40-1.78)||1.45||Transport other||0.19 (0.15-2.47)||0.11|
|Other Asia||0.33 (0.26-0.80)||0.52||Savanna||1.69 (1.18-3.32)||2.85|
|Pacific||0.50 (0.26-1.06)||1.43||Forest||1.18 (0.92-2.61)||3.01|
|Crop residue||0.33 (0.28-0.40)||0.34|
|TOTAL||6.63 (4.68-18.1)||15.2||TOTAL||6.63 (4.83-17.7)||15.2|
|"Previous" values use emission factors from Cooke et al. 1999, applied to 1996 fuel-use data. Low/high uncertainty bounds are given in brackets. The low/high totals are not the same for the two tabulations because we attempt to account for linear dependence between source categories.|
Missing Sources? We were asked to address the question, "Are there missing sources?" The answer is, "Undoubtedly." BC is preferentially emitted by the types of combustion that are likely to be missed by official statistics. We should ask, instead, What are the missing sources? In what regions are they significant? How well can we assess them (and the resulting climate forcing), now or ever? To what extent is it worthwhile to base mitigation strategies on inventories that are uncertain and incomplete? Finally, can the "lost sectors" become mitigation opportunities?
There might be two major contributors to underpredicted BC emissions: fuel use and emission factors. Underreporting of fuel use may occur when a portion of the fuel supply (e.g. wood or coal) does not pass through official channels, or when some "fuels" are not considered at all (e.g. house fires, waste paper). Emission factors may be underestimated if measurements come from better technology or more careful practice than the average. Increased emission factors might be associated with transient operation, poor-quality or adulterated fuels, or badly maintained units. For example, cold-starts and short-duration combustion events contribute to emission "puffs"; diesel "smokers" can emit 50 times more absorption than the average vehicle; and emissions during release of volatile matter from bituminous coal can be 20 times higher than averages, depending on conditions.
- 1. Penner, J. E., H. Eddleman and T. Novakov, Atmos. Env. 27A (8), 1277-1295, 1993.
- 2. Cooke, W. F., C. Liousse, H. Cachier and J. Feichter, J. Geophys. Res. 104 (D18), 22137-22162, 1999.
- 3. IPCC, Good practice guidance and uncertainty management in national greenhouse gas inventories, 2000.
- 4. Cullen, A. C. and H. C. Frey, Probabilistic Techniques in Exposure Assessment. New York, Plenum Press, 1999.