Grantee Research Project Results
Final Report: Methodology for Assessing the Effects of Technological and Economic Changes on the Location, Timing and Ambient Air Quality Impacts of Power Sector Emissions
EPA Grant Number: R831836Title: Methodology for Assessing the Effects of Technological and Economic Changes on the Location, Timing and Ambient Air Quality Impacts of Power Sector Emissions
Investigators: Ellis, Joseph H. , Burtraw, Dallas , Hobbs, Benjamin F. , Palmer, Karen
Institution: The Johns Hopkins University , Resources for the Future
EPA Project Officer: Chung, Serena
Project Period: February 1, 2005 through January 30, 2008 (Extended to January 30, 2009)
Project Amount: $648,733
RFA: Regional Development, Population Trend, and Technology Change Impacts on Future Air Pollution Emissions (2004) RFA Text | Recipients Lists
Research Category: Climate Change , Air
Objective:
The amounts, locations and timing of power sector emissions are sensitive to economic and technological assumptions. In order to more confidently estimate future temporal and spatial scenarios of emissions, a theoretically defensible, transparent and practical methodology is potentially helpful. This project developed a methodology for linking a sequence of models representing market-driven electricity supply and facility location in order to assess potential long-run emissions changes under alternative technology, policy, environment, and economic scenarios.
The methodology incorporates changes in investment in supply and demand side activities in the electricity sector. The Haiku model, developed by Resources for the Future, is used to provide regional technology, demand, and emissions totals and disaggregates national totals to regions. It represents economic behavior associated with the level and timing of electricity demand and associated emissions. The information from this model has coarse regional, spatial, and temporal characteristics that are passed as constraints to an energy facility-siting model, developed by Johns Hopkins University. The facility-siting activity simulates market-driven capacity construction, and builds on detailed empirical analyses of past siting patterns using an expanded dataset. The activity addresses land use patterns, availability of electricity transmission and permitting for air emissions, and yields hourly emissions profiles by facility. Finally, information from this model is used to systematically explore the sensitivity of emissions timing and location as well as ambient air quality results to uncertain drivers in order to assess which assumptions and model components matter most. This is done with a third suite of models, maintained at Johns Hopkins University, which address ambient air quality (tropospheric ozone and particulates) for an example set of scenarios that are simulated using MM5/MCIP/SMOKE/CMAQ. This aspect of the project demonstrates the practicality of integrating the source disaggregation methodology with the SMOKE emissions processing system and subsequently, the CMAQ transport and fate model itself.
Summary/Accomplishments (Outputs/Outcomes):
Methods
The methods developed and employed have been developed separately for each of the modeling components, followed by substantial effort integrating these components. We describe each in turn.
Electricity Supply and Demand:Over the course of the project, development of the Haiku electricity model has been undertaken to provide a detailed characterization of electricity demand including new econometric estimates that take advantage of information about regional and demographic characteristics. The important contribution of this model development is to enable a forecast of how demand changes, by region, in response to changes in demographics and underlying drivers of electricity demand including temperature. Moreover, the model explicitly accommodates behavioral responses to changes in electricity price or policy that can be expected to lead to investments in efficiency improvements in end use of electricity. Hence, the model integrates on the demand side expected responses that complement supply side changes that would be implemented by firms in response to changing economic conditions. Of key interest is the role of climate-driven changes on demand-side and supply-side behavior leading to different predicted electricity market equilibria. A component of that change is an evolution in the profile of supply side resources to meet electricity demand, with associated changes in the magnitude and timing of emissions.
A potentially significant element of the evolution of technology in the electricity sector is the introduction of new institutions including investments in energy efficiency and time-variant pricing of electricity. To address these emerging institutions, a supplementary model was developed that allowed for partitioning of demand across time blocks in a manner that enabled time-block specific responses to changes in prices. This framework enables the characterization of options for improved end-use efficiency with attention to time-of-day and real-time pricing. This is valuable for the examination of environmental policies that may directly or indirectly impose time-specific changes in electricity prices. For example, direct time-sensitive changes in electricity prices could result from episodic controls on emissions of NOx from electricity generation. Other developments to the electricity model include updates to the supply side of the model to incorporate cost projections for fossil, nuclear and renewable technologies including geothermal, wind and biomass.
Emission Downscaling Models: Siting and Dispatch: The downscaling models translate aggregate emissions projections from the national/regional Haiku model into a scale compatible with the needs of air quality simulation models. Downscaling is required both spatially (from regional emissions to facility-level emissions) and temporally (from twelve emissions periods per year to sequences of hourly emissions appropriately correlated with hourly meteorology). The downscaling involves a suite of models that account for historical siting and land use patterns, water, transmission fuel and spatial factors, air quality and siting probabilities, etc. Over the course of this project the suite of downscaling models was made fully operational and was successfully integrated with the upstream Haiku electricity model. In addition the coarse seasonal and temporal time step in Haiku (three seasons, four times of day) was distributed over a greater number of time steps. The improved geographic and temporal characterization of emissions from the electricity sector was successfully integrated with the downstream atmospheric model.
Air Quality Modeling: The air quality modeling takes hourly facility-level emissions information from the Emissions Downscaling Models. It is the final step in the integration of meteorological, emissions and air quality information. It makes operational and tests the air quality modeling system with new domains and emissions inventories, gridded surrogates and an updated chemical mechanism. The air quality modeling has a national domain and nested subdomains that provide improved resolution for the area characterized in detail in the emissions downscaling. The scales employed are:
- US36km_148X112 (national domain – nested in a 108km MM5 domain)
- US12km_160X120 (a nested subdomain covering the PJM electric grid)
- US4km_172X120 (a nested subdomain focused on MD, VA, PA)
Other aspects of air quality modeling include the vertical resolution and background emission inventory (that is taken as given when examining predicted changes from the electricity sector in the study region). Important characteristics of the air quality modeling include:
- Different MM5 vertical resolutions: 15 and 34 layers
- Different MCIP vertical resolutions: 12, 13, 15 and 17 layers
- FDDA vs no FDDA
- NEI01 emissions inventory
- EPA "new" gridded surrogates
- a set of simple routines to window the new surrogates to conform to user-specified subdomains
- generation of new BELD3 inputs for user-specified subdomains (using beld3smk, a component of the spatial allocator)
- cb05_ae4_aq chemical mechanism
- ping and noping selections in SMOKE V2.4 and CMAQ4.6
- different elevated source selection criteria
As of December 31, 2008, there existed one operational integrated emissions scenario (“new baseline”). A second integrated emissions scenario (“accelerated climate change,” or ACC) involving an assumption of an accelerated increase in temperature was created. Both scenarios represent supply and demand conditions for the electric power sector in 2030. In the baseline scenario, Haiku was run with demand functions based on cooling- and heating-degree days representative of 1990s conditions for all years (i.e., every 5th year from the first year through the last, which is 2030). The demand functions include cooling degree days (CDD), heating degree days (HDD), price, and previous periods' quantity demand (lagged demand) as arguments.
In the ACC scenario, Haiku was instead run with demand functions with higher CDDs and lower HDDs as inputs (based on the temperature scenarios). The CDDs and HDDs were linearly interpolated between the 1990s scenario and the 2050s scenario, and in the ACC scenario the interpolated temperature change was assumed to occur on an accelerated basis and achieved by 2030. In this scenario, the power system is therefore experiencing a gradually warmer climate, but warming at a faster rate than GISS anticipates. Both cases impose NOx and SO2 caps consistent with the Clean Air Interstate Rule (CAIR). Thus, total annual SO2 and summer NOx emissions do not change—only their spatial and temporal distribution. A shift towards a more sharply peaked distribution of emissions occurred in the ACC scenario (higher emissions in the high emissions hours, which are correlated with the high temperature conditions favorable to ozone episodes). Haiku's scenarios for 6 Haiku regions in the eastern and northeastern parts of the U.S. (NJ/DE; PA; MD; OH; KY/IN; MI) were downscaled, ultimately siting new plants by county and distributing emissions to hours.
Accomplishments and Research Results
The methods section above describes the major accomplishments with respect to model development and model integration that were implemented as part of the project. Here some of the results that follow from the exercise of the models are described.
In a forthcoming paper (Hobbs et al. IEEE Transactions on Power Systems, in press), emission forecasts and the effects on air quality for the complete integrated modeling framework were developed for an eight-state region. A disaggregation of NOx emissions from the electricity model for 2030 was achieved under two climate scenarios: no climate change (temperature patterns reflecting the 1990s) and accelerated climate change (temperature patterns expected for the 2050s). The accelerated climate change scenario was implemented in a continuous fashion between 2010 and 2030 to allow for incremental changes in supply and end-use technology. Subsequent research will look at the role of “surprises” associated with inter-annual variability.
The downscaling analysis simulated year-to-year variation in emissions by considering several meteorological years from the 1990s climate scenario (based on GISS simulations) as well as several from the 2050s climate scenario, using the 2030 power plant capacity results from Haiku. Average emissions in each climate scenario were constrained to equal the regional totals from Haiku, but year-to-year values (as well as hour-to-hour values within each year) depending on the particular temperatures in each simulated year. This analysis indicates that variations between years in the emissions patterns for a given climate scenario are greater than the differences between the averages in the two scenarios. This is due in part to the role of the NOx emissions cap, so the total emissions over a decade cannot change (however, emissions from year to year can change due to emissions banking). Only its distribution over time and space can change. Specifically, the 10 percent range in peak emissions among years in a scenario is greater than the difference between the average peak emissions for 2030 of the two scenarios (1990s and 2050s climate scenarios) that are modeled. This suggests that year-to-year variability in weather could impact peak emissions more than climate warming, given the role of an emission cap. However, more analysis is needed to ascertain the effect of climate warming upon demand patterns within the day, which was not captured in the version of the Haiku demand model used in this analysis.
The emissions modeling and downscaling was integrated with the suite of air quality models to explore the issues associated with time of day of electricity demand and supply and the change and variability in air quality that follows from a changing climate. This analysis was the centerpiece of a presentation (Ellis et al. 2008) at the EPA’s Region 3 science conference. The analysis for this presentation linked the expectation that climate change contributes to the formation of fine particulates and ozone, how electricity markets would respond to climate change including the operation of electricity generation and the profile of electricity demand. Load density and emissions rates in the PJM power region were analyzed to identify seasonal and diurnal variation in demand. Ozone formation appears as an episodic problem when emissions are linked to the air quality modeling; hence aggregate seasonal limitations on emissions under a seasonal emissions cap do not provide targeted reductions that offset potential violations in ambient air quality.
The modeling indicates that strategic timing of emissions controls is increasingly important for cost-effectiveness and overall effectiveness of pollution control. Moreover, the role of policies to address climate change could have unanticipated impacts on ozone. These policies are expected to lead to more gas generation, but the reduction in NOx that might be achieved by substituting national gas for coal in electricity generation does not map into proportional reductions in local ozone concentrations. Roughly speaking, NOx emissions from a lower plume may be twice as effective in contributing to local ozone formation.
The downscaling market models assumed perfect competition (price-taking behavior) among electricity generators. However, the presence of market power among large generators in transmission-constrained power markets is a reality, and can significantly affect prices and the amounts and patterns of generation and, ultimately, emissions. We have explored the potential significance of these effects in several detailed market simulation analyses for the PJM market region near Pennsylvania (Paul et al., (2009b, submitted to Energy Policy; Chen et al., 2006, Computational Management Science). Among other results, Chen et al., (2006) found that very large generators who also happen to be long in allowances (allocated more than they would use in equilibrium) can sometimes find it profitable to hoard NOx allowances and not use them in order to squeeze other suppliers by raising their input costs. This could lower total emissions if it affected emissions that were not governed by an emissions cap, as well as change the timing of emissions under a cap.
Changing regulation in the electricity sector also will affect the distribution of emissions under the NOx emissions cap. Traditional rate structures are moving toward time of use pricing and real-time pricing, which will have long run behavioral responses. The changing profile of emissions over time of day has implications for the formation of ozone. The model integration indicates the ability to capture these changes and understand them in a way that can inform public policy. However, the atmospheric modeling illustrates more hot episodes, and possibly higher highs during those episodes, could lead to higher ozone peaks. This creates a positive feedback with the electricity sector, where demand could be expected to increase during these episodes due to greater demands for space conditioning. On the other hand, variability across time may also increase, especially in the presence of an aggregate emissions cap.
To better understand and represent the responsiveness of demand to changes in temperature, and the resulting change in electricity market equilibrium and emissions, a new system of econometrically estimated demand functions was developed and incorporated into the electricity market model (Paul et al., submitted to Energy Economics, 2009a). These functions explicitly account for temperature changes and the change in electricity demand that would be expected to result. An important change in the way electricity demand is usually modeled is the incorporation of partial adjustment of changes in demand to changes in price. That adjustment is found to continue over time, providing an explicit link between short-term and long-term demand elasticities. The estimation of a large differentiated panel and its incorporation into an equilibrium model is a new contribution to the literature.
The demand system is useful not only for how it affects electricity supply but also for evaluating how changes in temperature may affect economic decisions about the end-use of electricity. Paul et al. (2009b, submitted to Energy Policy) evaluated the role of investments in energy efficiency. Palmer (2008) gave a well-received presentation on this work to the California conference on Behavior, Energy and Climate Change, and additional work is ongoing. Changes in the pattern of demand, with implications for emissions, are handed off to the downscaling suite of models accounting for transmission congestion and behavior as well as system reliability. This modeling provides specific advice with respect to the use of revenue from an auction of emissions allowances and the potency of investments in energy efficiency. The integrated modeling illustrates how decisions to invest in energy efficiency in one location has spillover benefits affecting electricity price in neighboring regions.
Conclusions:
This project demonstrates the practicality of integrating detailed models of electricity market equilibria with downscaling to represent temporal and spatial detail of emissions and ultimately detailed modeling of ambient air quality. This integrated modeling is used to explore the consequences of a changing climate on air quality associated with changes in the performance of the electricity sector. The changes are not simple or linear. Temperature changes affect both demand and supply side behavior in the electricity sector. These changes may be amplified by changing institutions in the sector, such as the introduction of time-sensitive electricity pricing or episodic controls for emissions. The modeling framework that is developed here is essential to the evaluation of these changing institutions in the future. The level of detail that can be achieved in a market model that links power markets across the country is insufficient to be able to evaluate air quality implications with a high degree of spatial or temporal resolution. The downscaling of effects in the electricity sector is a crucial step in order to be able to conduct this evaluation. That downscaling is achieved, and the results are handed off to a suite of air quality models. The project illustrates the value of the integrated system for scientific and economic analysis and for policy evaluation.
Several extensions to this work would be valuable, and some of these are already initiated under the current project. One is a full analysis of the role of time-sensitive pricing in electricity on the profile of emissions and resulting air quality. Second is an expansive set of air quality scenarios that can be used to span approximately the range of possible outcomes. Third, the potential role for episodic controls on emissions should be explored further in the context of an integrated ability to estimate equilibrium responses and air quality outcomes. Fourth, this modeling framework opens the possibility for analysis of sweeping new technologies such as electric vehicles, which could have a profound impact on electricity demand. Integrated economic and air quality modeling is necessary to identify the public policy that may be necessary to accompany new technologies in order to continue to achieve air quality improvements over time and in the face of a changing climate.
Journal Articles on this Report : 6 Displayed | Download in RIS Format
Other project views: | All 16 publications | 9 publications in selected types | All 7 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Chen Y, Hobbs BF, Leyffer S, Munson TS. Leader-follower equilibria for electric power and NOx allowances markets. Computational Management Science 2006;3(4):307-330. |
R831836 (2005) R831836 (2007) R831836 (2008) R831836 (Final) R828731 (2003) R828731 (Final) |
Exit |
|
Chen Y, Sijm J, Hobbs BF, Lise W. Implications of CO2 emissions trading for short-run electricity market outcomes in northwest Europe. Journal of Regulatory Economics 2008;34(3):251-281. |
R831836 (Final) |
Exit Exit |
|
Hobbs BF, Hu M-C, Chen Y, Ellis JH, Paul A, Burtraw D, Palmer KL. From regions to stacks: spatial and temporal downscaling of power pollution scenarios. IEEE Transactions on Power Systems 2010;25(2):1179-1189. |
R831836 (2008) R831836 (Final) R828731 (Final) |
Exit Exit |
|
Hu M-C, Hobbs BF. Analysis of multi-pollutant policies for the U.S. power sector under technology and policy uncertainty using MARKAL. Energy 2010;35(12):5430-5442. |
R831836 (Final) |
Exit |
|
Paul A, Palmer K, Ruth M, Hobbs BF, Irani D, Michael J, Chen Y, Ross K, Myers E. The role of energy efficiency spending in Maryland's implementation of the Regional Greenhouse Gas Initiative. Energy Policy 2010;38(11):6820-6829. |
R831836 (Final) |
Exit Exit Exit |
|
Sijm J, Chen Y, Hobbs BF. The impact of power market structure on CO2 cost pass-through to electricity prices under quantity competition--a theoretical approach. Energy Economics 2012;34(4):1143-1152. |
R831836 (Final) |
Exit Exit Exit |
Supplemental Keywords:
RFA, Scientific Discipline, Air, climate change, Air Pollution Effects, Environmental Monitoring, Ecological Risk Assessment, Atmosphere, air quality modeling, atmospheric carbon dioxide, ecosystem models, electrical energy, climatic influence, emissions impact, green house gas concentrations, modeling, carbon dioxide, climate models, CO2 concentrations, demographics, electric power sector emissions, ambient air pollution, atmospheric models, Global Climate ChangeProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.