Developing Alternate Anthropogenic Emission Scenarios for Investigating Future Air Quality William G. Benjey, Daniel H. Loughlin*, Christopher G. Nolte and Robert W. Pinder U.S. Environmental Protection Agency, National Exposure Research Laboratory, Atmospheric Modeling and Analysis Division *U.S. Environmental Protection Agency, National Risk Management Research Laboratory, Air Pollution Prevention and Control Division Abstract: Alternate technologies, emission controls and carbon policies were used with the MARKAL energy system model in conjunction with a base inventory and the SMOKE emission model, to develop alternate anthropogenic emission scenarios for the year The emission differences between scenarios were not as large as anticipated, providing guidance for future efforts. Motivation: The U.S. Environmental Protection Agency’s Climate Impact on Regional Air Quality (CIRAQ) program initially examined the effect of climate change in about 2050 with no change in anthropogenic emissions (Nolte et al, 2008). The second phase of the program (CIRAQ 2) addresses air quality effects pf plausible future scenarios. Consequently, an approach to generating scenarios for use with the EPA Community Multiscale Air Quality (CMAQ) was developed to produce future instantiations of emissions for the United States. References: Fishbone, L.G. and H. Abilock, 1981: MARKAL: A linear- programming model for energy-systems analysis: technical description of the BNL version. Int. J. Energy Res.,5,4, Houyoux, M.R., and Z. Adelman, 2004: Quality assurance enhancements to the SMOKE modeling system. Preprints, Tenth Annual Emission Inventory Conference, Denver, CO, U.S. Environmental Protection Agency, 12 pp. IPCC, 2000; Special Report on Emission Scenarios. Tech. Rep., Intergov. Panel on Clim. Change, New York. Nolte, C.G., A.B. Gilliland, C. Hogrefe, and L.J. Mickley, 2008: Linking global to regional models to assess future climate impacts on surface ozone levels in the United States. J. Geophys. Res. 113, D14307, 14 pp. Scenarios and Rationale: Six scenarios were developed using MARKAL with the intent of exploring the effects of substantially different plausible drivers of anthropogenic emissions for the year The scenarios all reflected the A1 population growth assumptions of the International Panel on Climate Change (IPCC) as modified at the county level by the ICLUS-SERGOM (Integrated Climate Land Use-Spatially Explicit Regional Growth Model) population and land cover projection models (IPCC,2000). The principal assumptions varied in MARKAL included: (1) additional controls applied to large emission sources (e.g., electric generating plants), (2) carbon limitation policies (3) optimistic assumptions about the market penetration of high-technology vehicles and (4) combinations of the above variables. All scenarios assumed existing laws and regulations, including ongoing efforts to meet existing standards through State Implementation Plans, the Tier 2 on-road vehicle emission limits and and the Energy Independence and Security Act (EISA) (Table 1). Table 2 provides a summary description of the assumptions applied for each scenario. Results and Comparisons: Examination of all scenarios total annual emissions shows the expected reductions from the 2002 inventory (Figure 4), primarily because of existing planned emission limits. The amounts and patterns of annual total emissions by pollutant and region between scenarios are similar, with some variations to reflect differing assumptions The overall emissions decrease for each scenario (1 – 6) reflecting their increasing application of controls, carbon policies and hybrid vehicles. However, the relative amount of pollutants and the distribution between regions remain nearly the same. Figures 5 (Scenario 1) and 5 (Scenario 6) show the consistent pattern and range of emissions Quarterly emissions demonstrate a similar pattern, except for the third quarter for all scenarios. Example Figure 7 shows that third quarter emissions for the Scenario 6 are relatively Future Approaches: In the short term (a year), addition scenarios may be developed by the following approaches: 1. Examine alternative assumptions about population and economic growth. 2. Examine alternative time frames for new technology availability in MARKAL. Conclusions: This study successfully demonstrated the ability to develop long-range future emission inventories using tools such as MARKAL and SMOKE. While the pollutant levels in the scenario inventories did not differ as much as expected, insights learned in the process will be used to develop scenarios for the next phase of the analysis.. greater, mainly due to increased CO2 emissions that are mapped to CO and VOC by MARKAL for electric generation units (EGU). In addition SMOKE applies seasonal factors by EGU because of seasonal changes in fuels and technologies (coal, gas, etc). Figure 8 shows the difference factors from 2002 for EGU emissions for the third quarter. However, future seasonal operations may be very different from current. For example, increased use of natural gas combined-cycle technologies may decrease the seasonal variability of natural gas emissions. This is one example of the complex interactions that can occur.. Approach: Estimation of future year emissions was accomplished by using factors developed by the Market Allocation (MARKAL) energy economic equilibrium model (Fishbone and Ablilock, 1981). These factors were applied to a detailed base emission inventory developed for use in air quality modeling (EPA 2002ae modeling inventory). MARKAL allows consideration of alternative assumptions about population growth and migration, economic growth and transformation, land use change, technology change and alternative criteria and greenhouse gas policy directions. Figure 1 summarizes the energy demand sectors used by MARKAL. The EPA developed a 9-region (Figure 3) U.S. version of MARKAL to create technology-specific scenarios and pollutant- specific projection factors at the 3-digit source category code (SCC) level for energy system sources within the electricity production, industrial, commercial, residential and transportation sectors. Regional factors were created for CO2, NOx, SO 2, and PM 10. In lieu of pollutant-specific information, CO 2 factors were used for emissions of Volatile Organic Compounds (VOCs) and carbon monoxide (CO) since CO 2 was deemed a reasonable indicator of electric generation and industrial combustion. For mobile source emissions, NOx projection factors were used as indicators of CO and VOC emissions. PM 2.5 factors were assumed proportional to PM 10 factors. Factors for emissions from industrial processes and other sources that are not part of the energy system model were estimated based on national level energy demand or industrial shipment projections, in that order of preference, from an economic model. The projection factors were mapped to source category codes (SCCs) and applied to the 2002ae inventory using the Sparse Matrix Operator Kernel Emission (SMOKE) system (Figure 3) to estimate six alternative scenario emission inventories (Houyoux and Adelman, 2004). Emissions were assumed to grow in place and were not spatially redistributed. SMOKE also applied spatial and temporal allocation to the projected emissions, along with chemical speciation grouping (SAPRC 99 mechanism) and placed the emissions in a 36km resolution North American grid for use by the Community Multiscale Air Quality (CMAQ) model run as a regional climate model. Figure 4 illustrates the flow of data between MARKAL, SMOKE and related components 3. Make additional changes to examine sensitivities to a wide range of inputs.. 4. Examine different times samples from existing scenarios (e.g. 2030). Figure 10 shows the variability of change in time. 5. Examine the response to criteria pollutant response to increasingly stringent CO2 targets, and the impact of more stringent criteria pollutant targets on CO2.. In the intermediate and longer-term (2010 and beyond), National and regional emission scenarios may be nested in the IPCC AR5 scenarios