Methodologies for Estimating Emissions for the US ... - ACS Publications

Aug 31, 2001 - ... 3-yr cycle inventories, and NOx SIP call inventories by county for all ... Timothy Onasch , Ezra Wood , Scott Herndon , John Jayne ...
0 downloads 0 Views 58KB Size
Environ. Sci. Technol. 2001, 35, 4408-4413

Methodologies for Estimating Emissions for the U.S. EPA’s NOx SIP Call, CER Rule, and Other Complexities MOHAMMED A. MAJEED* State of Delaware, DNREC, Air Quality Management Section, Dover, Delaware 19901

Emissions inventories play an important role in many air quality decisions, and the importance of obtaining correct emissions data challenges the emissions inventory community to continuously improve its estimation techniques. The Clean Air Act Amendments of 1990 requires states with ozone nonattainment areas (NAAs) to submit periodically a comprehensive, accurate, and actual inventory of ozone precursor emissions until the areas are redesignated to attainment. They require the states with NAAs to report peak ozone season daily and annual estimates of the inventories for the counties that are in nonattainment. The recently proposed U.S. Environmental Protection Agency’s (EPA) NOx SIP call and the Consolidating Emissions Reporting (CER) rule are more demanding still. The CER rule requires the states to report statewide point source inventories, 3-yr cycle inventories, and NOx SIP call inventories by county for all source types, regardless of the attainment status; the estimating periods are peak ozone season daily, 5-month ozone season, and annual estimates. Furthermore, complexities in emissions inventories such as estimating the emissions with different seasonal and multiple controls exist. These complexities and methodologies for estimating emissions for different time periods are not addressed sufficiently either by the EPA’s procedure manuals or by the Emissions Inventory Improvement Program guidance documents. This paper presents methodologies for estimating emissions for different time periods and multiple controls. These methodologies will help the EPA, state, and local government agencies to meet the evolving demands of emissions inventories and the reporting requirements of the NOx SIP call and the CER rules.

Introduction Section 182(a) of the Clean Air Act Amendments (CAAA) requires states with ozone nonattainment areas (NAAs) to submit periodically a comprehensive, accurate, and actual inventory of ozone precursor emissions until the areas are redesignated to attainment. For example, the CAAA requires states to submit an actual emission inventory every 3 yr (3-yr cycle) for ozone NAAs beginning 1993, where the actual inventory of emissions reflects the actual rate of emissions of pollutant from an emissions unit. The inventories include actual volatile organic compounds (VOC), oxides of nitrogen (NOx), and carbon monoxide (CO) emissions. An emissions * Corresponding author phone: (302)739-4791; fax: (302)739-3106; e-mail: [email protected]. 4408

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 22, 2001

inventory is a compilation of emissions estimates of pollutants of concern from anthropogenic (human-caused) and natural (biogenic and geogenic) sources. The anthropogenic sources can be broadly classified as point, area, off-road mobile, and on-road mobile sources. Point sources are stationary facilities such as electric power plants and industrial processes that emit more than a threshold amount of a pollutant per year. Area sources are all other stationary sources that emit less than the threshold amount of pollutant per year for point sources. Area sources collectively represent individual sources that are small and numerous. For example, gasoline stations and dry cleaners are often treated as area sources. Off-road mobile source categories are aircraft, commercial and military marine vessels, locomotives, and miscellaneous engines. The miscellaneous engines include miscellaneous equipment such as construction, farm, and industrial equipment; lawn and garden equipment; motor cycles; and recreational vehicles. On-road mobile sources include light-duty gasoline vehicles, light-duty gasoline trucks, heavy-duty gasoline vehicles, heavy-duty gasoline trucks, light-duty diesel vehicles, light-duty diesel trucks, heavy-duty diesel vehicles, and motorcycles. Natural sources consist of biogenic and geogenic sources. Biogenic sources include those sources that result from biological activity. Geogenic sources include lightning and oil and gas seeps. The U.S. Environmental Protection Agency (EPA) requires states with ozone NAAs to report annual as well as peak ozone season daily emissions estimates of VOC, NOx, and CO for the counties that are in nonattainment. States have been following the EPA’s approved methodologies found in several EPA procedure documents (1-6) in estimating the emissions of VOC, NOx, and CO. These methodologies, however, did not meet some of the new EPA emissions inventory requirements. To address this concern, the EPA and State and Territorial Air Pollution Program Administrators/Association of Local Air Pollution Control Officials (STAPPA/ALAPCO) created the Emissions Inventory Improvement Program (EIIP). The EIIP has provided information on what is a preferred method and what are the alternative methods available to the states for developing emissions inventories through the EIIP documents (7, 8). The EPA’s recent emissions reporting requirements are becoming increasingly complex and demanding. For example, to properly address the ozone transport issues and to assist the states with ozone NAAs east of Mississippi River in attaining the 1-h ozone National Ambient Air Quality Standards (NAAQS), recently the EPA issued a NOx controlling strategy, widely known as the NOx SIP call, to reduce NOx emissions in the affected states and the District of Columbia. The EPA requires the NOx SIP call affected jurisdictions to report 5-month ozone season (May 1-September 30) NOx emissions estimates (9). States included in the NOx SIP call will collect emissions data from the sources that are subject to controls as a means of compliance. Per this rule, all sources not inventoried as point sources shall be inventoried as area or mobile sources and reported only if they are to be controlled to meet the emission budget. The EPA’s recently proposed Consolidating Emissions Reporting (CER) rule consolidates various emissions reporting requirements that already exist, establishes new ones for PM2.5 and regional haze, and establishes new requirements for the statewide reporting of area source and mobile source emissions including reporting requirements for the NOx SIP call and new reporting requirements for air toxics (10). This rule requires the states to report statewide inventories of point source inventories, 3-yr cycle inventories, and NOx SIP 10.1021/es001817v CCC: $20.00

 2001 American Chemical Society Published on Web 08/31/2001

TABLE 1. Summary of Requirements for Reporting Emission Inventories point source inventory provision

type A sourcea

type B sourcea

NOx SIP call inventory

3-yr inventory

frequency of reporting estimating period areas to which provision applies pollutants & source thresholdb

annual annual statewide SOx g 2500 NOx g 2500 VOC g 250 PM10 g 250 PM2.5 g 250 CO g 2500 NH3 g 250

every 3 yr annual and daily statewide SOx g 100 NOx g 100 VOC g 100 PM10 g 100 PM2.5 g 100 CO g 1000 NH3 g 100 Pb g 5

annual 5-month season statewide NOx g 100 note: all sources not inventoried as point sources shall be inventoried as area or mobile sources and reported only if they are to be controlled to meet emission budget

every 3 yr annual and daily statewide ozone NAAc: VOC g 10 NOx g 100 CO g 100 CO NAAc: CO g 100 PM10 NAAc: PM10 g 70 (serious) PM10 g 100 (moderate) PM2.5 NAAc: PM2.5 g 100 NH3-point/area source point g specified tpyb area < specified tpy on-road mobile off-road mobile biogenic

a Previously, type A and type B sources together constituted the annual inventory and were required to be reported annually. b Threshold units are in tons per year (tpy). c Thresholds apply to NAAs only; the remainder of the state uses type B source thresholds to distinguish between point and area sources.

call inventories. States should submit statewide point source and 3-yr cycle inventories for PM10, PM2.5, and regional haze, consistent with the data requirements for ozone and CO. States use data obtained through current annual reporting requirements (point source inventories) to record emissions from large sources and to track progress in reducing emissions from them. States get 3-yr cycle data from stationary sources and use them with the point source inventories to update their emissions inventory every 3 yr. The rule also takes advantage of data from Emissions Statements available to the states but not reported to the EPA. Emissions Statement requirements apply to all stationary sources located in an ozone NAA that emit or have the potential to emit facilitywide 25 t/yr of NOx or VOC emissions. These requirements also apply to stationary sources located in ozone attainment areas that emit or have the potential to emit 50 t/yr of NOx or VOC emissions. As appropriate, the states may use these data to meet their reporting requirements for point source data. Combining data from these activities gets the most information from sources with the least burden on the industry and less effort by state and local government agencies. These inventories help NAAs develop and meet SIP requirements to reach the NAAQS. The estimating periods for smaller point sources (type B sources; 10) and 3-yr cycle inventories are daily and annual, whereas the states are required to report actual annual emissions for major point sources (type A sources; 10). The CER rule requires the states to report 5-month seasonal estimates every year for the NOx SIP call inventory. The affected source categories and the reporting requirements of the NOx SIP call and CER rules are summarized in Table 1. There are many possible sources of error and uncertainty in the emissions inventory process. The EPA procedures and the EIIP methods of estimating emissions often rely on average values for the parameters that are input to emissions calculations, an inadequate approach in some cases. Errors in emissions estimates can result from inattention to critical details. In fact, these critical details are the missing parts to the EPA’s procedures and EIIP guidance documents that result in over/underestimating emissions. Furthermore, the EPA’s procedures documents and EIIP guidance documents do not provide guidance on estimating emissions for different estimating periods and, therefore, do not meet the demands

of the NOx SIP call and the CER reporting requirements. Also, other complexities related to emissions inventories that have not been addressed by EPA/EIIP are how to estimate emissions with controls for one or multiple seasons and how to estimate emissions with multiple controls for different seasons. The former is a case where emission controls could be different for different seasons (e.g., open burning ban during the peak ozone season). The latter is a case with more than one control program on a piece of equipment due to different requirements [e.g., off-road small spark ignition engines with reformulated gasoline (RFG) and new emissions standards]. While quality assuring the Delaware 1996 Periodic Emissions Inventory (PEI) (11), I came across these complexities in the PEI. To my knowledge, neither the EPA nor the EIIP provided any guidance on how to estimate emissions for these situations. Yet the emissions inventory community has to meet the demands of the NOx SIP call and CER rule and other evolving emissions inventory requirements by continuously improving its emission estimation techniques. The EPA’s on-road mobile source emission factor models Mobile5a and Mobile5b account only for the estimation of emission factors for multiple controls. The EPA/EIIP guidance documents, however, do not provide methodologies for estimating on-road mobile source emissions for multiple seasonal controls. Furthermore, the procedure adopted for applying multiple controls in the Mobile5a/b model and its application to the other source categories (point, stationary area, and off-road mobile) are not made available to the inventory compilers. Therefore, procedures are warranted to fill these gapssnamely, for on-road mobile source category, methodologies to estimate emissions with multiple seasons; for point, stationary area, and off-road mobile source categories, methodologies for estimating emissions with multiple seasons and multiple controls. This paper presents procedures for correctly estimating emissions with multiple seasons and multiple controls, and these procedures are equally applicable to all source categories. As seen in Table 1, if area and mobile sources are controlled to meet the NOx SIP call emissions budget, then these sources will also be inventoried; all source categories will be inventoried to meet the 3-yr cycle inventory requirements. Therefore, it can be said here that all source categories are likely to be affected by seasonal and/or multiple controls in areas subject to the VOL. 35, NO. 22, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4409

NOx SIP call and CER rules. The procedures presented here will help the state and local agencies correctly estimate the emissions of all source categories with multiple seasonal and/ or multiple controls subjected to the NOx SIP call and CER rules. This paper also illustrates these methodologies with some numerical examples.

Methodology Emissions Estimates with One or Multiple Seasonal Controls. Numerous examples exist in our inventories where the controls are not annual but only seasonal. As an example, seasonal fuel switching to a lower NOx emitting fuel on major stationary sources of NOx is often considered a good point source NOx Reasonably Available Control Technology (RACT) control. Some examples of stationary area source controls are residential brush, slash, prescribed, and agricultural burning bans during the peak ozone season (June-August in Delaware) and also regulating the use of cutback and emulsified asphalt during the ozone season (April-October in Delaware). For such cases, traditionally, the annual emissions are estimated as though the seasonal controls do not affect the annual estimates, and the adjustments for controls are applied only to the peak ozone season daily estimates. Obviously, such an approach results in overestimating the annual emissions. There also exist cases in control strategy development where different sets of controls are needed for different seasons. An example of multiple seasonal controls is for an area that is in nonattainment for both ozone and carbon monoxide, thus requiring the use of Federal reformulated gasoline (RFG) for summer months and special oxygenated fuels for winter months. Another example is for the attainment of annual PM2.5 NAAQS that may warrant different sets of controls for summer and winter episodes. The response of sulfate aerosol concentrations and size of changing sulfur dioxide (SOx) emissions and hydrogen peroxide (H2O2) concentrations may strongly depend on the region and season due to regional and seasonal dependencies of the emissions and physical and chemical processes. Over North America, the efficiency of sulfate production is higher in the summer than in the winter due to increased photochemical production of oxidants (12). However, as compared to clear-sky production of sulfate, the in-cloud aerosol production is relatively more important in the winter (13). Consequently, the sulfate production processes might imply that different seasonal controls might be warranted for attaining the annual PM2.5 NAAQS. Here we will develop a methodology for estimating emissions with controls for multiple seasons. Because it is a generalized case, this methodology can be applied to estimating emissions for one or multiple seasons. We will also illustrate examples of estimating emissions for both the cases. Emissions Estimates with Controls for Multiple Seasons. Let us consider the generalized case where the control measures could vary from season to season, i.e., the controls are multi-seasonal. Emissions with controls for one season can be estimated from the generalized case easily. The seasonal, annual, and daily emissions are estimated as follow: Seasonal Emissions. To estimate the seasonal emissions, an estimate of seasonal activity is required. The seasonal activity Ai for any season i can be estimated from eq 1 as

Ai ) SAFi × AD × Di

(1)

where AD is the annual average daily emissions generating activity, SAFi is the seasonal adjustment factor that adjusts the annual average daily activity for the season i, and Di is the number of days in the season i. 4410

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 22, 2001

Now the seasonal emissions Ei for the season i can be estimated using eq 2 as

Ei ) Ai × EFi × CFi

(2)

where EFi and CFi are respectively the precontrolled emission factor and the control factor for the ith season. The control factor CFi is usually expressed as CFi ) 1 - CEi × REi × RPi where CEi, REi, and RPi are respectively control efficiency, rule effectiveness, and rule penetration of a given rule. Rule effectiveness accounts for any noncompliance of a rule. Frequently, some sources in a category do not comply or are late in complying with the limits in the rule. All of these factors are considered in determining the effectiveness of a rule. Rule penetration determines to what portion of a source category the rule applies. Because the stringency of the rules or controls could differ from season to season, the control efficiency (CEi), rule effectiveness (REi), and rule penetration (RPi) are likely differ from season to season. Therefore, the control factor CFi is also likely to vary from season to season. Annual Emissions. The annual emissions EA can now be estimated as the sum of all seasonal emissions: n

EA )

∑E

i

(3)

i)1

where n is the total number of seasons in a year. Daily Emissions. The seasonal daily emissions EDi for the season i can now be estimated using eq 4 as

EDi )

Ei Di

(4)

Example 1 (Emissions Estimates with Controls for One Season). Here, it is illustrated how typically the state and local agencies overestimate the annual emissions by not paying attention to critical details. For the purposes of illustration, residential brush burning from Delaware’s 1996 PEI is chosen, and emissions estimation methodology of VOC will be presented here. Brush burning is prohibited during the peak ozone season for Kent and New Castle Counties. However, there is evidence of illegal brush burning during the summer months; so a rule effectiveness (RE) of 80% was applied. A rule penetration (RP) of 99% was chosen for the yard waste-brush open burning classification since there are some exemptions to the regulation, e.g. campfires, etc. Because of the brush burning prohibition during the peak ozone season, the control efficiency (CE) is set to 100%, and the control factor (CF) is estimated to be 20.8%. The emission factor (EF) for VOC is 19 lb/t. The annual VOC emissions for Kent County in Delaware’s 1996 PEI were estimated as

annual activity level (AA) ) 5913 t of brush/yr annual emissions (EA) ) 5913 t of brush/yr × 19 lb of VOC/t × t/2000 lb ) 56.174 t of VOC/yr Obviously, the annual emissions were overestimated in Delaware’s 1996 PEI because emissions were assumed to be uncontrolled throughout the year. The correct estimates are shown below. Seasonal Emissions. In this example, there exists only one season for which controls apply. Therefore, to estimate annual emissions, we need to determine two seasonal emissionss one for which brush burning is prohibited and the other when no prohibition applies.

Annual average daily emissions generating activity (AD) is determined as

AD )

5913 t of brush/yr ) 16.2 t of brush/day 365 day/yr

Now the seasonal activity A1 during the peak ozone season is determined from AD using eq 1 with SAF ) 1.0 as

A1 ) 1.0 × 16.2 t of brush/day × 92 day/season ) 1490.4 t of brush/season Seasonal emissions during the peak ozone season are estimated using eq 2 as

E1 ) 1490.4 t of brush/season × 19 lb of VOC/t of brush × ton/2000 lb × 0.208 ) 2.945 t of VOC/season Now the seasonal activity A2 for which controls do not exist is determined as

A2 ) 5913 - 1490.4 ) 4,422.6 t of brush/season Seasonal emissions for the rest of the year are determined as

E2 ) 4422.6 t of brush/season × 19 lb of VOC/t of brush × ton/2000 lb ) 42.015 t of VOC/season Annual Emissions. Annual VOC emissions are estimated by adding the two seasonal emissions:

EA ) 2.945 + 42.015 ) 44.960 t of VOC/yr Example 2 (Emissions Estimates with Controls for Multiple Seasons). In the absence of real data, a hypothetical example of seasonal fuel switching to a lower NOx-emitting fuel on major stationary sources of NOx is presented here. Delaware NOx RACT regulation requires major NOx-emitting sources to comply with the provisions and establishes emission standards of the RACT that can be applied to achieve the required presumptive emission standards. The regulation treats fuel switching during the summer ozone season (April 1-October 31) to a low NOx-emitting fuel as a RACT. Fuel switching is limited to the use of natural gas, liquid petroleum gas (LPG), or distillate oil. The NOx RACT demands only 90% availability of the new fuel during the summer ozone season. Let us consider a boiler with a heat capacity of 100 MMBTU/h or less, where MMBTU stands for millions of British thermal units. Let us assume that the source switches from residual oil to natural gas during the peak ozone season, i.e., JuneAugust, and to distillate oil during the other four months of ozone season, i.e., April, May, September, and October. We assume that the daily average emissions are 1.0 t of NOx/day and also that the seasonal adjustment factor (SAF) ) 1.0 for all seasons. The NOx reduction from fuel switching is determined from the heat content and emissions factors of the fuels. The heat content for residual oil, distillate oil, and natural gas are assumed as 150 MMBTU/1000 gal, 140 MMBTU/1000 gal, and 1,050 MMBTU/MMCF, respectively, where MMCF stands for million cubic feet. The NOx emissions factors for these fuels are selected from the EPA document (14)s67 lb of NOx/ 1000 gal for residual oil, 24 lb of NOx/1000 gal for distillate oil, and 140 lb of NOx/MMCF for natural gas. On the basis of these values of heat content and emissions factors, the emissions factors can be expressed in the units of pounds

of NOx/MMBTU. The revised emission factors for residual oil, distillate oil, and natural gas are estimated as 0.447, 0.171, and 0.133 lb of NOx/MMBTU. Now the control efficiencies of switching the fuel from residual oil to distillate oil and residual oil to natural gas are estimated as CEresidual-distillate ) 0.6174 and CEresidual-nat.gas ) 0.7025. With an EPA-recommended rule effectiveness (RE) ) 80% default value, the control factors are estimated as CFresidual-distillate ) 0.5061 and CFresidual-nat.gas ) 0.4380. Note that the NOx RACT requires only 90% of the availability of the new fuel during the summer ozone season. With these data, we can estimate the seasonal and annual emissions estimates as follows. This example requires emissions estimates for three different seasonssfor three peak ozone season months (June-August), four nonpeak ozone season months (April, May, September, and October), and other five months (January, February, March, November, and December). The RACT, however, requires only 90% of the availability of the new fuel during the summer ozone season. Therefore, there will be two emissions estimates, one for the period when the new fuel is available and the other when the new fuel is not available, during the peak ozone season months and the other four summer ozone season months. The emissions estimates for each of the seasons are estimated below. NOx Emissions during the Peak Ozone Season Months (June-August). With 90% availability of the new fuel during the summer ozone season, seasonal emissions for the period when the new fuel is available are estimated as

E1 ) ED × CFresidual-nat.gas × DS × % fuel availability ) 1.0 t of NOx/day × 0.4380 × 92 day/season × 0.90 ) 36.266 t of NOx/season Seasonal emissions for the period when the new fuel is not available are estimated as

E2 ) ED × DS × % fuel nonavailability ) 1.0 t of NOx/day × 92 day/season × 0.10 ) 9.200 t of NOx/season NOx Emissions for the Nonpeak Ozone Season Months (April, May, September, and October). Seasonal emissions for the period when the new fuel is available are estimated as

E3 ) ED × CFresidual-distillate × DS × % fuel availability ) 1.0 t of NOx/day × 0.5061 × 122 day/season × 0.90 ) 55.570 t of NOx/season Seasonal emissions for the period when the new fuel is not available are estimated as

E4 ) ED × DS × % fuel nonavailability ) 1.0 t of NOx/day × 122 day/season × 0.10 ) 12.200 t of NOx/season NOx Emissions for the Rest of Year (January, February, March, November, and December). NOx emissions for the days left (DR) are estimated as

E5 ) ED × DR ) 1.0 t of NOx/day × 151 ) 151 t of NOx/season Annual Emissions. Annual emissions can now be estimated by adding the above five estimates as VOL. 35, NO. 22, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4411

5

EA )

∑E ) 264.236 t of NO /yr i

ED2 ) SAF × AD × EF × CFRVP

(8)

ED3 ) SAF × AD × EF × CFRVP × CFStage I

(9)

x

i)1

Estimation of Emissions with Multiple Controls. There exist cases with multiple control programs, which could be a mix of national, regional, and state or local controls. The benefits from these controls may be significant enough to account for them in the inventory compilation. Numerous examples can be listed of point, area, off-road, and on-road mobile sources with multiple controls. Two examples are the estimation of emissions from the filling of underground storage tanks (UST) for stationary area sources with Reid Vapor Pressure (RVP) and stage I controls and from off-road small spark ignition engines with reformulated gasoline (RFG) and new emissions standards. Mobile5a/b accounts for multiple control programs in estimating emissions factors for on-road mobile sources. Neither the EPA procedure documents nor the EIIP documents, however, provide enough guidance for estimating emissions with multiple programs for point, area or off-road mobile sources. The methodology for estimating daily, seasonal, and annual emissions with multiple controls is presented below. Daily Emissions. The daily emissions (ED) for a source category (SIC/SCC) with multiple control measures (m) at a given instance in time can be estimated using

ED ) SAF × AD × EF × [CF1 × ... × CFi × ... × CFm] (5)

where EF is the precontrolled emission factor, CFRVP is the control factor that accounts for RVP effects, and CFStage I is the control factor that accounts for stage I effects. The VOC benefit due to RVP measure is obtained by taking the difference in estimates ED1 and ED2, and the VOC benefit due to state/local controls is obtained by taking the difference in estimates ED2 and ED3. The difference between ED1 and ED3 gives the total VOC benefit from both the control measures. Alternatively, ED2 and ED3 can be determined by eqs 10 and 11:

ED2 ) SAF × AD × EFRVP

(10)

ED3 ) SAF × AD × EFRVP × CFStage I

(11)

where EFRVP is the new controlled emission factor that accounts for the RVP effects, and CFStage I is the control factor from stage I controls. Seasonal and Annual Emissions Estimates with Multiple Controls for Multiple Seasons. The seasonal emissions Ei for any season i for which the m seasonal controls apply can be estimated using eq 12 as m

where AD is the annual average daily seasonal activity, SAF is the seasonal adjustment factor that adjusts the annual average daily activity for the season of interest, EF is the precontrolled emission factor, and CFi is the control factor corresponding to the ith control measure. Equation 5 can also be expressed as m

ED ) SAF × AD × EF ×

∏CF

i

(6)

i)1

Equation 6 treats any changes in the emissions factor (EF) due to multiple rules, regulations, and standards as changes in the control factor. Although it is possible to directly represent any changes in emission factor, we elect to reflect those changes as changes in the control factor (CF). In other words, all the changes in the emissions factor (EF) are expressed as changes in the control factor (CF). A good example of the applicability of multiple controls is the estimation of emissions from filling of underground storage tanks (UST) for stationary area sources with RVP and stage I controls. The applicable controls for filling of underground storage tanks are the RVP regulations on gasoline fuel and additional state/local regulations on stage I controls. The effect of RVP regulation is to modify the emissions factor, and the effect of state/local regulations is to modify the control factor. Per eq 6, however, the new effect of RVP would appear as a control factor but not as a modified emission factor. To determine the benefit from each control measure, we could apply the control factors incrementally one at a time in eq 6 and then determine the benefit due to each control measure by subtracting the emissions estimates successively. As an example, to determine the VOC benefits of RVP and state/local regulations on stage I, we need three estimatess uncontrolled emissions (ED1), emissions with RVP control alone (ED2), and emissions with RVP and state/local regulations on stage I (ED3). These estimates are determined by eqs 7-9, respectively:

ED1 ) SAF × AD × EF 4412

9

(7)

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 22, 2001

Ei ) Ai × EFi ×

∏CF

j

(12)

j)1

where the seasonal activity Ai is determined using eq 1, EFi is the precontrolled emission factor, and CFj is the control factor for jth control. The annual emissions (EA) can now be estimated as the sum of all seasonal emission using eq 3. Example 3. For the purposes of illustration, the example of off-road small spark ignition engines with two controls, reformulated gasoline (RFG) and new emissions standards, is selected. Because of lack of good emissions inventory data for this case, an example from Delaware’s 1999 rate-ofprogress plan (RPP) is chosen. Reformulated Gasoline. Reformulated gasoline will affect VOC emissions from 2-stroke and 4-stroke vehicles listed in the “other off-road sources” category. Per the EPA memorandum (16), reformulated gasoline will lower exhaust and evaporative emissions for the 2-stroke and 4-stroke components of the off-road inventory. Of the affected categories, 86.51% of the VOC emissions are exhaust emissions, 5.58% are evaporative emissions, and 7.91% are other emissions. The use of reformulated fuel will cause a 3.3% reduction in exhaust emissions and a 3.5% reduction in evaporative emissions. EPA Court-Ordered New Emissions Standards. The EPA memorandum (16) provides guidance for the calculation of emissions reductions resulting from Federal off-road engine emissions standards that have court-ordered deadlines for promulgation. The EPA is under court order to promulgate standards for new off-road heavy-duty compression ignition (diesel) engines, small off-road spark ignition (gasoline) engines, and outboard/inboard marine engines. The EPA is regulating new spark-ignition engines with a power output at and below 19 kW (25 hp) in two phases. The phase I regulations identify exhaust emission standards for hydrocarbons (HC), NOx, and CO for all new small gasoline engines manufactured on or after August 1, 1996. Phase II standards for small engines take effect in 2001. The small gasoline engine regulations will affect small engines used in a broad range of equipment categories, including lawn and garden, utility, small farm and construction, and light industrial applications.

Peak Ozone Season Daily Emissions. The EPA memorandum (16) lists the hydrocarbon reduction in the year 1999 as 22.9%. The emissions standards result in a NOx emissions increase of 79.6%. For the purposes of this illustration, however, only the VOC emissions are estimated here. For Kent County, the uncontrolled peak ozone season daily emissions for this category are estimated as

ED1 ) 1.885 t of VOC/peak ozone season day The RFG will lower the exhaust and evaporative emissions by 3.3% and 3.5%, respectively. The net peak ozone season daily emissions ED2 after the RFG effect is estimated following the EPA memorandum (16):

ED2 ) 1.885 t of VOC/day × [0.8651 × (1 - 0.0333) + 0.0588 × (1 - 0.035) + 0.0791] ) 1.829 t of VOC/day Per the EPA memorandum (16), the new standards will result in a reduction of 22.9% VOC. This would result in a control factor (CF) ) 0.771 ) 77.1%. The net emissions with both the controls ED3 are estimated as

ED3 ) 1.829 t of VOC/day × 0.771 ) 1.410 t of VOC/day Seasonal Emissions for the Peak Ozone. Seasonal emissions for the peak ozone season ES3 can be estimated to be

ES3 ) 1.410 t of VOC/day × 92 day/season ) 129.72 t of VOC/season

Discussion Errors in emissions estimates can result from inattention to critical details. In fact, these critical details are the missing parts to the EPA’s procedures and the EIIP guidance documents for estimating emissions. The methodologies in EPA procedures and EIIP guidance documents result in overestimating annual emissions because the annual emissions are estimated as if the seasonal controls do not affect the annual estimates. Furthermore, the EPA procedures and EIIP guidance documents do not provide ways to estimate emissions for different estimating periods. This paper provided methodologies for estimating emissions for different estimating periodsspeak ozone season daily, seasonal, and annual estimates; these methodologies meet the reporting requirements of the NOx SIP call and the CER rule. Two other inherent problems to emissions inventories that have not been addressed by EPA/EIIP are how to estimate emissions with controls for one or multiple seasons and how to estimate emissions with multiple controls for different seasons. This paper presented methodologies for estimating emissions for these situations and also illustrated these methodologies with numerical examples. In example 1 of residential brush burning, the new methodologies resulted in avoiding an overestimation of 11% of VOC emissions. Methodologies presented in this paper will help the EPA, state, and local governments in correctly estimating emissions to meet the NOx SIP call and the CER rule reporting requirements.

Acknowledgments I thank Darryl Tyler, Raymond H. Malenfant, and Alfred R. Deramo for their review and helpful comments on this

document. I am also thankful to the three reviewers for their expert comments, which resulted in bringing clarity to the document.

Literature Cited (1) U.S. Environmental Protection Agency. Procedures for the Preparation of Emission Inventories for Carbon Monoxide and Precursors of Ozone, Volume I: General Guidance for Stationary Sources; EPA-450/4-91-016; Office of Air Quality Planning Standards: Research Triangle Park, NC, May 1991. (2) U.S. Environmental Protection Agency. Procedures for Emission Inventory Preparation, Volume IV: Mobile; EPA-450/4-81-026d; Office of Mobile Sources: Ann Arbor, MI, 1992. (3) U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors, Volume I: Stationary Point and Area Sources, AP-42, 5th ed.; Office of Air Quality Planning Standards, Research Triangle Park, NC, January 1995. (4) U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors, Volume I: Stationary Point and Area Sources, AP-42, 4th ed.; Office of Air Quality Planning Standards: Research Triangle Park, NC, September 1985. (5) U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors, Volume II: Mobile Sources, AP-42, 4th ed.; Office of Air Quality Planning Standards: Research Triangle Park, NC, September 1985. (6) U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission Factors, Volume I: Stationary Point and Area Sources, AP-42, 5th ed., Supplement B; Office of Air Quality Planning Standards: Research Triangle Park, NC, November 1996. (7) U.S. Environmental Protection Agency. Emissions Inventory Improvement Program (EIIP) Volume II, Point Sources Preferred and Alternative Methods; Office of Air Quality Planning Standards: Research Triangle Park, NC, July 1997. (8) U.S. Environmental Protection Agency. Emissions Inventory Improvement Program (EIIP) Volume III, Area Sources Preferred and Alternative Methods; Office of Air Quality Planning Standards: Research Triangle Park, NC, July 1997. (9) U.S. Environmental Protection Agency. Finding of Significant Contribution and Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone, Rule. Code of Federal Regulations, Parts 51, 72, 75, and 96, Title 40; Fed. Regist. 1998, 63 (207). (10) U.S. Environmental Protection Agency. Consolidated Emissions Reporting; Proposed Rule. Code of Federal Regulations, Part 5, Title 40; Fed. Regist. 2000, 65 (100). (11) State of Delaware. 1996 Periodic Ozone State Implementation Plan Emissions Inventory for VOC, NOx, and CO for the State of Delaware; Air Quality Management Section, Division of Air and Waste Management: November 1999. (12) Salzen, K. von; Leighton, H. G.; Ariya, P. A.; Barrie, L. A.; Gong, S. L.; Blanchet, J.-P.; Spacek, L.; Lohmann, U.; Kleinman, L. I. J. Geophys. Res. 2000, 105 (D8), 9741-9765. (13) Roelofs, G. J.; Lelieveld, J.; Ganzeveld, L. Tellus, Ser. B 1998, 50, 224-242. (14) U.S. Environmental Protection Agency. AIRS Facility Subsystem: Source Classification Codes and Emission Factor Listing for Criteria Air Pollutants; EPA 450/4-90-003; Technical Support Division, Office of Air Quality Planning and Standards: Research Triangle Park, NC, March 1990. (15) U.S. Environmental Protection Agency. VOC Emission Benefits for Nonroad Equipment with the Use of Federal Phase I Reformulated Gasoline. Memorandum from Phil Lorang, Director, Emission Planning and Strategies Division, Office of Mobile Sources, U.S. EPA, Ann Arbor, MI, August 18, 1993. (16) U.S. Environmental Protection Agency. Future Nonroad Emission Reduction Credits for Court-Ordered Nonroad Standards; U.S. Government Priniting Office: Washington, DC, November 28, 1994.

Received for review October 27, 2000. Revised manuscript received August 1, 2001. Accepted August 2, 2001. ES001817V VOL. 35, NO. 22, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4413