Air Pollution as a Climate Forcing: A Workshop
Summary B. Gases
Background. This session focused on methane (CH4) and tropospheric ozone (O3), the two primary greenhouse gases with links to urban air pollution. Most of the presentations are available on the workshop website and are summarized in extended abstracts below. The discussion addressed the overarching question for each gas. These questions, as anticipated, could not be answered at this time, and thus a primary objective was to discuss the research agenda needed to address these questions.
Methane. For CH4 the overarching question "Is it possible to achieve a global warming success story with methane, i.e., can we achieve a real world scenario with decreasing CH4?" was addressed through a sequence of "smaller" scientific statements that attempt to build the case for the overriding question.
The abundance of CH4 has risen with industrialization from about 700 ppb in the millennium preceding the last two hundred years of industrialization to 1750 ppb today. This increase has led to a radiative forcing of 0.55 W/m2 with an uncertainty of ±20% (IPCC, 2001), without including the indirect effects of CH4 on stratospheric H2O and tropospheric O3. The rate of CH4 increase appears to have been most rapid (in terms of ppb/yr) during the period 1960-1985. Since 1985, CH4 continues to increase when averaged over two-year periods, but this rate of increase continues to slow, appearing to approach zero growth rate (see Dlugokencky presentation).
The emissions of CH4 are many and diverse; they are both natural and anthropogenic. While total CH4 emissions are well constrained by the observed atmospheric growth rate and the sink strength derived from scaling the lifetime of CH3CCl3, individual source strengths with their attribution to specific human activities remain highly uncertain relative to the accuracy in the CH4 budget needed to understand the recent trends. However, concerted scientific efforts over the past decade [e.g., the recent finding of much lower emissions from paddy rice (see van der Gon presentation)], plus continued work on refining natural sources, have helped reduce these uncertainties*. While absolute emission rates may still have systematic biases, it is possible that the recent year-to-year variations in emissions from wetlands and biomass burning may be derived with reduced uncertainty (see Matthews presentation).
The atmospheric sink for CH4 is primarily reaction with the OH radical (with small soil and stratospheric losses). A primary issue is sorting out the relative contribution of changes in OH (driven in part by anthropogenic CO and NOx emissions) and in methane emission rates on CH4 abundance variations during the past two decades. Differing interpretations of the CH3CCl3 budget project opposite trends in OH, but the data may be consistent with little or no change of OH. The decay of perturbations to CH4 is about 1.4 times longer than the CH4 budget lifetime and it depends on models with no clear method for validating the feedbacks. Emissions of NOx, CO, and VOC impact the CH4 lifetime and become effective sources (CO) or sinks (NOx). One of the key CH4 science problems is reconciling data of the recent past: how do the observed "wiggles" in CH4 correspond to changes in sources and sinks?
In summary, our ability to predict the greenhouse impact (radiative forcing from all atmospheric changes) of a specified reduction in CH4 emissions is very good to excellent. Thus we can predict the effect of a change in sources to an accuracy of ±20% and the effect of a change in sinks (CO, NO2) to ±50%. Our ability to verify reduction in specific, regional CH4 emissions from observations is weak. Regional source inversions are difficult: poor at the global scale, perhaps just possible at regional scale with co-tracers and intensive airborne measurements. While the publication of annual CH4 inventories by many countries in accord with the U.N. FCCC is helpful new information for producing useful source-level bottom up estimates, we still need to have reliable, verifiable, bottom-up methods for source/sink accounting.
What is the air pollution connection with CH4? CH4 is one of the dominant factors controlling background levels of tropospheric O3: +10% in CH4 (175 ppb) yields about 1 ppb increase in tropospheric O3 globally (to an accuracy about ±50%), which tends to increase air pollution similarly. While any methane abundance reduction will reduce tropospheric O3, we must look at how emission sources of CH4 are interwoven with actions related to reducing air pollution. Given the diversity of CH4 sources, reductions in some but not all are tied directly to decisions related to local air quality regulations. Thus, direct recapture of CH4 (coal and landfill reclamation) would reduce emissions and benefit local energy usage; control of CO and VOC emissions (landfills, biofuels, and waste treatment) is a clear indirect reduction in CH4, but control of NOx while reducing O3 would increase CH4. Additional, but less studied links involve control of PM that impacts photolysis and heterogeneous chemistry rates.
How much can we reduce CH4 below 1750 ppb? Technologies exist to reduce methane emissions from many sources, both industrial and agricultural; however, the total potential reduction, across industrial and biogenic sources, is not well quantified. Recent experience demonstrates that emissions from several sources can be readily reduced at low cost. U.S. methane emissions, for example, have been reduced over the past decade and are currently 6 percent below 1990 levels and are projected to remain below 1990 levels through 2020 (see Gunning presentation). The European Union has made similar progress. Such efforts, if international will clearly stabilize methane abundances, avoiding the large increases projected in some of the SRES scenarios; however, it is not clear if large methane decreases can be realized.
Ozone. For O3 the overarching question "Is it feasible to halt the growth of global tropospheric O3 or even achieve a reduction?" was also addressed with a sequence of smaller questions to identify key uncertainties in our knowledge.
The mean tropospheric O3 abundance has risen with industrialization, although measurements are spotty and do not characterize the magnitude or geographic location of these increases very well. The IPCC (2001) analysis of these observations and model estimates derived a best value for tropospheric O3 increase of 9 Dobson Units (out of a mean value of 34 DU today, 1 DU = 1.5 ppb) with an uncertainty of ±40%. Based on a range of model results, this increase corresponds to a radiative forcing of +0.38 W/m2. From the IPCC modeling study, this increase is due primarily to increases in CH4 abundance (~40%) and NOx emissions (~40%) with lesser contributions from CO and VOC emissions. An alternative viewpoint from the Harvard model (see Mickley presentation) has tropospheric O3 increasing by more than a factor of two since the late 19th century. The 18 DU increase results in a radiative forcing of about +0.8 W/m2; however, most of this increase is not reparable through air pollution controls since it assumes dramatically lower pre-industrial levels of biomass burning (-90% of today) and lightning (-70% of today).
Over the past two decades — the only period with reliable observations over several regions of the globe — the O3 abundance in the free troposphere has a general, but irregular increase with relatively large interannual and geographic variations. There are probably multiple causes given the magnitude of the changes during this period in CH4, O3 precursors, and even stratospheric O3. Sources of O3 include direct influx from the stratosphere (about 550 Tg/yr) and production by photochemical precursors including air pollution (~4500 Tg/yr). Sinks for O3 include photochemical loss proportional to UV and water vapor (~4000 Tg/yr) and reaction with the Earth's surface, especially the biosphere (~ 1000 Tg/yr). Indeed, it is difficult to explain the lack of an obvious increase of tropospheric O3 in northern mid-latitudes since 1970, given the large increases in transportation (especially aviation). One of the key, unsolved ozone science problems is defining the history of O3 precursor emissions from 1880 to 2000, using these to stimulate the O3 history, and reconciling this with observations.
The impact of greenhouse gas controls on air pollution will be primarily through the change in background tropospheric O3 (and of course also through the changes in local temperature, water vapor and atmospheric circulation that control individual pollution episodes). As background O3 abundances rise or fall, air quality will do so also (except in extremely polluted environments where air quality levels are likely to be driven only by local emissions). The ability to predict the greenhouse impact of a specified reduction in O3 precursor emissions ranges from very good (±33% for changes in CH4 abundance) to fair (factor of two at best, for NOx, CO, and VOC emissions). Difficulties in quantifying the latter center on the fraction of NOx exported from urban air pollution to the free troposphere. Unfortunately, our ability to verify a reduction in O3 from precursor emissions through observations is poor and not likely to improve.
Air quality controls that reduce all emissions of NOx, CO, VOC will also reduce global tropospheric O3, but optimizing for reductions in local regions may not optimize the export of O3 and hence reduction of the greenhouse gas impact. Reductions in global CH4 abundance (not just emissions) could also significantly reduce tropospheric O3. Co-emitted aerosols are likely to interact with NOx and alter O3 export (the sign of this alteration is uncertain). Based on O3 sources, there are major opportunities for significant global tropospheric O3 reductions: road transport, biofuel use, power plants, air and ship transport, fossil fuel production and the handling of the products.
The path to design a strategy for reducing tropospheric O3 probably should involve an inventory development for most or all sectors, including bottom-up inventories, atmospheric observations, and model-verification. It will need to define and quantify cause-effect functions for, among others, urban NOx/CO emissions and O3 export, and the synergism of combined reductions in trace gases and aerosols.
Can we control the projected growth in tropospheric O3 radiative forcing? Given that the last two decades (with pollution controls in Europe and N. America) have had little increase in mid-latitude O3 in the free troposphere, it seems possible that pollution controls in developing countries can prevent the large rise projected in the IPCC SRES scenarios (up to 20 DU). (These scenarios specifically did not assume any new air quality regulations in developing countries.) The ability to reduce tropospheric O3 by 9 DU in order to cut radiative forcing by 0.4 W/m2, however, is more elusive and requires a significant research effort to answer.
Research Areas. Considering both CH4 and tropospheric O3, and allowing for the considerable overlap in the air pollution sources of both O3 precursors and aerosols, what might be critical, joint-research agendas?
- The Urban-to-Global Connection. The mapping of reactive emissions to resulting impacts needs to be quantified better. For example, how do different urban/industrial emissions impact CH4 and O3 through the indirect chemistry-climate gases (NOx, CO, VOC). We need to develop emission-to-global-response functions (including regional responses) over a range of conditions/locations, recognizing the non-linearity of response, and examining chemical coupling and synergism with co-emitted gases/aerosols. One important approach to reducing uncertainties might be through emission closure: follow the chemical evolution, deposition, and export of key pollutants; design the network (ground-based, airborne, and satellite) for observational verification of emissions change. Such efforts would provide critical science input to the next round of national/international assessments. Synergism exists with aerosol-climate impacts and also with short-lived ozone-depleting substances.
- Biomass Burning Source Strengths. Biomass burning is probably a major source of the year-to-year variations of CH4 and tropospheric O3, as well as variations of carbonaceous aerosols. A major, coordinated effort will be needed: satellites to obtain fire-scars on a sub-seasonal basis; fire counts to help with day-to-day emissions (but this is not quantitative); ground-based surveys of major biome types to calibrate inventories before and after fires, thus helping to define fuel burn emissions.
- Hindcasting CH4 and Tropospheric O3. The ability to match the observed seasonal, year-to-year, and decadal variations of CH4 and O3 over the last 20 years would provide a fundamental test of our understanding. Such an effort would coordinate the field work, data analysis and modeling efforts with a focus on tropospheric chemical change; develop emissions inventories with uncertainties; use historical meteorological fields and stratospheric O3 measurements; include detailed simulations of intensive periods with satellite and field campaign measurements of many chemical tracers; and it would include aerosols. Such a study is a large undertaking and, while building on individual science, it needs to be a community effort so that results would be accepted widely. Accurate hindcasting provides the critical science validation for assessment projections. Combined with better aerosol information, it may also allow empirical determination of climate sensitivity.
* The U.N. Framework Convention on Climate Change requires national inventories only from Annex I countries (developed nations and former Soviet Union) of all anthropogenic greenhouse gas and related emission strengths on an annual basis, following methodological guidance developed by the IPCC. The IPCC has also provided guidance on how to quantify uncertainties in the emission estimates, which countries have begun implementing in the last two years. Many non-Annex I developing countries have also prepared emission inventories, although at less frequent intervals. A scientific evaluation of the uncertainty in the reported plus inferred anthropogenic CH4 emissions, at a level that can shed light on recent trends, has not been done.