1.COVER & Table of Contents › projects › sos99 › scienceplan.pdfTABLE OF CONTENTS Introduction 1 Proposed Research 7 Primary PM and PM Precursor Emissions 11 PBL Dynamics 13

NASHVILLE 1999 FIELD STUDYSCIENCE PLAN

DRAFT

October 1998

TABLE OF CONTENTS

Introduction 1

Proposed Research 7

Primary PM and PM Precursor Emissions 11

PBL Dynamics 13

Ozone Production Efficiency 19

Characterization of Loss Processes 23

VOC Contribution to Ozone and PM Formation 27

Fine Particulate Matter Formation and Characterization 31

Nighttime Chemistry and Dynamics 37

Instrumented Aircraft 41

Ground-Based Measurements 61

Tracer Release 67

References 69

Appendix A Ð Southern Center for the Integrated Study ofSecondary Air Pollutants (SCISSAP) A-1

Appendix B Ð The National Parks Service Enhanced Monitoring Network B-1

Appendix C Ð TVA/EPRI/NPS Enhanced Monitoring Site GSMNP C-1

Appendix D Ð SEARCH D-1

Nashville 99 Science Plan

1

INTRODUCTION

The SOS Paradigm

The oxidant-management approaches being usedduring the 1980s were based largely on scientificfindings, air quality models, and related air qualitymanagement tools from research conducted insouthern California and the urban megalopolis in thenortheastern United States. Few scientific studies hadbeen conducted in the South. In the late 1980s,however, studies began to emerge that pointed to theSouthÕs unique air quality management problems.These problems included: 1) a high frequency of airmass stagnation, warm temperatures, high humidity,and intense solar insulation that characterize theregionÕs summer climate; 2) abundant naturalemissions of biogenic hydrocarbons from the SouthÕsample rural and urban forests; and 3) ananthropogenic emission mix dominated by islandcities and rural point sources. Because of theseunique characteristics, it became apparent that airquality management approaches developed in otherparts of the Nation might not be appropriate for theSouth. To address this concern, the SouthernOxidants Study (SOS) was formed in 1988.

The SOS is a coordinated, long-term research programfocusing on the formation, accumulation, andmanagement of photochemical oxidants in the South.Since its founding, the SOS has focused on twogeneral purposes:

1 . Using the southern United States as a naturallaboratory for policy-relevant scientificinvestigations, improving scientific and publicunderstanding of the chemical and meteorologicalprocesses that cause ozone and otherphotochemical oxidants to accumulate in theatmosphere near the ground; and

2. Evaluating alternative strategies by which leadersin various Federal, State, municipal, industrial,

and commercial organizations can manage theaccumulation of ozone and other photochemicaloxidants in the atmosphere, and decrease theinjurious effects of these airborne chemicals invarious urban and rural areas.

SOS Research

Prior to the late 1980Õs, biogenic hydrocarbons werebelieved to play little or no role in ozone formation ineither the rural or urban atmospheres. Two paperswere pivotal in changing this viewpoint: 1) Trainer etal., (Nature, 1987) showed that biogenic hydrocarbonsplayed a major role in rural ozone episodes in theeastern United States; and 2) Chameides et al.,(Science, 1988) showed that biogenic emissionsrepresented a significant source of volatile organiccompounds (VOCs) in Atlanta and that theseemissions decreased the efficacy of a VOC-basedozone abatement strategy. The findings of Trainer etal. and Chameides et al. were a major driving forcebehind the inception of SOS and, consequently, thestudy of the role of biogenic VOC in ozone formationwas identified as a prominent research theme for theprogram.

Research conducted as part of the SOS program hasbeen broad in scope, necessitated by the complexnature of the problem. The program participants havebeen responsible for significant advances in methodsdevelopment, emissions characterization, modeldevelopment and evaluation, and have providedimportant new insights into the processes that controlozone accumulation in the atmosphere (Fehsenfeld etal., 1993; Chameides and Cowling, 1995).

Field programs conducted during the early years ofSOS relied primarily on ground-based measurementsin the rural environment and were focused ondeveloping a better understanding of the role of

Introduction

2

biogenic emissions in ozone formation and exploringozone/NOY relationships. There were, however, twoSOS studies that used highly instrumented aircraft toprovide a more comprehensive view of regional ozoneformation and transport. In each case, theinstrumentation mounted on the aircraft were used tocharacterize the spatial variability of ozone and itsprecursors over the region. Further, the flight pathswere designed to provide information on ozoneformation in urban and power plant plumes. In 1990an effort led by NOAAÕs Aeronomy Laboratory(Trainer et al., 1995) made measurements aroundBirmingham Alabama while in 1992 the TennesseeValley Authority participated in a major field programin Atlanta Georgia (Imhoff et al., 1995)

The results of these studies demonstrated the extremevalue of the three dimensional information providedby these kinds of platforms. The experience obtainedin the conduct of these studies and the analysis of thedata collected provided critical new information onozone production efficiency in plumes and formed thebasis for the Nashville/Middle Tennessee Ozone studythat was conducted in 1994/1995.

Nashville/Middle Tennessee Ozone Study Ð1994/1995

The SOS Nashville/Middle Tennessee Ozone Studywas conducted in two phases. Phase I, a preparatorystudy, was conducted during a three-week period inthe summer of 1994. During this period, four ground-based monitoring stations, two wind profilers, and twoinstrumented aircraft (the NOAA WP-3 and the TVABell-205 helicopter) were operational. This three-week period also provided opportunity for: 1)evaluation of protocols for the ground-based andaircraft-based measurements systems to be used in1995; 2) accumulating baseline data to aid in planningand calibrating instruments for the 1995 study; and 3)a rigorous intercomparison of the NOy measurementmethods employed by various SOS groups in the 1992SOS Atlanta Intensive and SOS-SERON field studiesduring 1992-1994.

Phase II was carried out over a six-week period (June19-July 28) in 1995. During this second phase a

three-tiered (Level I, II, and III) network ofprogressively more sophisticated surface chemistrystations was established. The 108 Level I surfacemonitoring stations, operated as part of variousregulatory and research networks, provided broadspatial ozone coverage across the Nashville/MiddleTennessee region and parts of eleven neighboringstates. Six Level II surface monitoring stationsprovided continuous short-term (1 to 5 minute), high-sensitivity measurements of ozone (O3), sulfur dioxide(SO2), nitric oxide (NO), total nitrogen oxides (NOY),and carbon monoxide (CO) along with temperature,relative humidity, wind speed, wind direction, andsolar radiation. Integrated, one-hour, pressurizedcanister samples for VOC analysis were collectedmidday at each station. Two Level III stations, onelocated approximately 20 km NE and the otherapproximately 20 km SE of the Nashville urban centerprovided detailed research-grade photochemistrymeasurements (e.g., organic nitrates, peroxides,aldehydes, VOC speciation). A network of five windprofilers was also employed.

Six instrumented aircraft were employed during thestudy -- NOAA WP-3, DOE G-1, NOAA Twin Otter,TVA Bell 205, NOAA CASA 212, and NASA C-130.The first four of these aircraft made in-situ chemistryand meteorology measurements and the last two hadremote sensing capabilities.

The six-week 1995 intensive study was structuredaround a series of five aircraft-based experiments:

1. Urban Plume - Measurements of ozone formationin the Nashville urban plume under free-flow andstagnation conditions.

2 . Power Plant Plume Studies - Measurement ofozone destruction and production in fossil-fuelfired power plant plumes.

3 . Subregional Characterization - Detai ledatmospheric chemistry and meteorologicalcharacterizations to evaluate urban-ruralinterchange and provide a detailed observationaldata set for model evaluations.


3

4. Regional Characterization Ð Measurements usinginstrumented aircraft that extended north to theGreat Lakes, south to the Gulf of Mexico, west toMissouri and Arkansas, and east to theAppalachian mountains to provide context for andcontrast to the measurements in theNashville/Middle Tennessee area.

5 . Aircraft Intercomparisons - Side-by-sideintercomparisons for the in-situ sampling aircraftand overflights for the remote-sensing aircraftsupporting the quantitative assessment of datainter-comparability

Findings From the 1994/1995 Study

The analysis of the measurements that were madeduring the course of the 1994/1995 Nashville MiddleTennessee Ozone Study has already produced severalimportant findings.

Air Stagnation produced high ozone concentrations ina very confined areaThe episode that resulted in the highest observedground-level, hourly surface ozone concentration (138ppb) occurred during stagnation conditions (July 12,1995). The Nashville urban plume remained localizeddirectly over the urban area covering approximately600 km2. During this episode ozone concentrationsoutside the boundaries of the urban plume were in the60 to 70 ppb range indicating that local productionresulted in an ozone concentration increase ofapproximately 70 to 80 ppb over the background.Under stagnation conditions, the ozone produced overthe Nashville urban area during the day wassubsequently redistributed throughout the region bynighttime winds.

Ozone production efficiency scales inversely with NOXsource strengthOzone production per unit of NOX emission appears tobe greatest for the Nashville urban plume and forpower plants with smaller NOX emissions that arelocated in areas rich in natural VOC emissions. Ozoneproduction was found to be less efficient for NOXemitted form rural power plants with higher NOXemissions.

Inefficient ozone formation in plumesCross-plume pollutant profiles obtained by the NOAAWP-3 were combined with detailed wind fields toinvestigate O3 formation at various downwinddistances for several pollutant sources with verydifferent NOX emissions. The analysis of this datafrom the measurements made in 1994 and 1995indicate that NOX removal occurred rather quickly andthat O3 production in these plumes was much lessefficient than ozone production from more dispersedNOX sources.

Reduced formation of H2O2 in plumesPeroxide concentrations were lower in urban andpower plant plumes than in background air. Thelower peroxide formation rates inside the plumes arethought to be a result of higher NOX concentrationsthat maximize radical consumption by NO2 to formnitric acid as opposed to radical combination reactionsthat form peroxides.

Importance of biogenic VOCsIsoprene chemistry dominated ozone formation in theforested rural areas of the Nashville / MiddleTennessee region. The simultaneous measurement ofperoxymethacrylic nitric anhydride (MPAN),peroxypropionic nitric anhydride (PPN) andperoxyacetic nitric anhydride (PAN) provides anopportunity to apportion photochemically-producedozone into a fraction resulting from oxidation ofbiogenic hydrocarbons (BHCs) and a fractionresulting from oxidation of anthropogenichydrocarbons (AHCs). A significant portion of thephotochemically-produced ozone in the southernportion of the study region resulted from oxidation ofBHCs. The contribution of isoprene to rural ozoneformation was found to decrease significantly withincreasing latitude (i.e. from south to north). Thisdecrease paralleled a similar decrease in isopreneemissions with latitude predicted by current biogenicemission inventories.

Role of carbon monoxide and methane in regionalozone formationAn analysis of the VOC data suggests that carbonmonoxide and methane make a significantcontribution to ozone formation in regions whereisoprene levels are depressed (i.e. in the boundary

Introduction

4

layer in urban plumes, in the upper Midwest and thefree troposphere). The more reactive VOCs have highremoval rates and their absence suggests a lack ofnearby sources of these VOCs. The more limitedcontribution to ozone formation from the morereactive VOCs may also be due, in part, to theeffectiveness of current emission abatement strategiesthat have targeted these compounds. The apparentimportance of methane and carbon monoxide in O3formation in the regions described above necessitatetheir consideration in future ozone managementstrategies. The role of methane and carbon monoxideshould be expected to be further enhanced as emissionof more reactive VOCs continue to decline.

Intercomparison of measurementsA comprehensive intercomparison of ground-basedNOy measurement systems employing both gold/COand molybdenum catalytic reduction systemsindicated good agreement between the two methodsand among the various groups making the NOymeasurements. Intercomparisons between aircraftand surface measurements indicated generally goodagreement for O3, SO2, and CO. However, there weresome significant differences in the NOymeasurements, both between aircraft flying parallelflight tracts and between airborne and ground-basedNOY measurements.

These results, along with many others, are describedin detail in a special section of Journal of GeophysicalResearch Ð Atmospheres Volume 103, No. D17. Foran overview of the 1994/1995 Nashville / MiddleTennessee Ozone Study see Meagher et al., 1998.

Changes in Air Quality Regulations

During 1997, EPA introduced three regulatoryinitiatives to address the most egregious air qualityproblems in the Nation.

1. The EPA implemented a new National AmbientAir Quality Standard (NAAQS) for ozone. Thelevel of the new standard was set at 80 ppb,averaged over 8-hr. (U.S. EPA, 1997a). Thisaction caps nearly three decades of effort tomanage ozone with mixed results.

2 . The EPA revised the NAAQS for particulatematter (PM) by adding a new annual PM2 . 5(particles with aerodynamic diameters less than2.5 micrometers) standard set at 15 mg/m3 and anew 24-hr PM2.5 standard set at 65mg/m

3. Theannual PM1 0 (particles with aerodynamicdiameters less than 10 micrometers) standard wasretained and the form of the 24-hr PM10 standardwas modified (U.S. EPA, 1997b). In taking thisaction, the EPA cited epidemiological evidencelinking significant human health impacts(mortality, hospital admissions, and respiratoryillness) to ambient particulate levels below theprevious standard.

3. New Regional Haze Regulations have beenproposed (U.S. EPA, 1997c). These regulationsare designed to protect and improve visibility inthe 156 mandated Class I areas (National Parksand Wilderness areas) of the country. The focus,once again, is on reducing PM2.5 concentrations inthese areas.

These actions are expected to result in a significantincrease in the number of areas as nonattainment forozone (Chameides et al., 1997) and PM. Thisanticipated expansion in the number of nonattainmentareas serves to heighten the need for effective ozoneand PM management strategies, based on soundscience.

SOS Science Questions

In response to the insights and events described abovethe SOS community has articulated a set of six policy-relevant science questions to serve as a focus forfuture SOS research. They are as follows:

1 . By what specific aerosol measurement methodscan SOS achieve maximally beneficialcharacterization of aerosols in both urban andrural areas of the SOS region?

2 . What are the linkages (similarities anddifferences) between the chemical, biological, andmeteorological processes that govern formationand accumulation of ozone and fine particulateaerosols?


5

3. What is the relative efficiency of production ofozone and fine particulate aerosol in urban plumescompared to that in power-plant plumes?

4 . In what ways are the rates and efficiencies ofozone and fine particulate aerosol production andaccumulation different in large urban areascompared to small urban areas -- especially incases where an air mass spends an extendedperiod traversing a very large urban area?

5 . How do nighttime meteorological and chemicalprocesses influence the rates, efficiencies, andareal extent of ozone and fine particulate aerosolformation and accumulation?

6 . What meteorological and chemical factorsdetermine the regionality and/or locality of ozoneand fine particulate aerosol accumulation events?In particular:a) How do the processes that govern ozoneaccumulation in isolated urban areas surroundedby high-isoprene-emitting forests differ fromthose in urban areas in which urban plumesoverlap from one urban or non-forested area toanother?b) How important are urban heat-islandphenomena in determining the regionality and orlocality of ozone and fine particulate aerosolformation and accumulation processes?

Introduction

6


7

PROPOSED RESEARCH

Study Overview

The ramifications of the conclusions from the SOS1994/1995 Nashville Middle Tennessee Ozone Studyfor policies to control ozone on local and regionalscales in the Southeastern United States and theirapplicability to other locales and regions in the UnitedStates require additional testing and verification. Inaddition, these unexpected conclusions force us toconsider other factors that may influence thephotochemical processing of these compounds. Inparticular, the role played by particulate matter (PM)in tropospheric chemistry needs to be addressed.

The expansion of our measurements to investigate PMand PM-related processes is timely. Aerosolsparticipate in a variety of chemical and physicalprocesses in the troposphere. On a regional scale,these processes are associated with regional air qualityas related to visibility and the effects of fine particleson human health. In this regard, there is a naturalsynergism between the ozone-related researchdescribed previously and the study of processesleading to or involving fine particles. The newFederal air quality regulations force us to recognizethat a basic scientific understanding of the chemistryand physics of the atmosphere is prerequisite to thedesign of effective control strategies for thesepollutants and that the concentrations of the pollutantsin the atmosphere are often co-dependent because ofinteracting chemical reactions.

With this in mind, we are proposing to expand theSOS measurements program to elucidate:

1 . how chemical processing on aerosols influencesozone formation;

2. how the atmospheric oxidation leading to ozoneformation leads to aerosol formation; and

3. how atmospheric chemistry influences the growthand/or the chemical composition of aerosols.

This expansion of SOS to include the study of theformation and fate of fine particles in the atmosphereis being led by SOSÕ affiliated Southern Center forIntegrated Study of Secondary Air Pollutants(SCISSAP). The Nashville-99 field study will beclosely coordinated with the SCISSAP measurementnetwork (see Appendix D). The Field Studymeasurements will augment those of SCISSAP,providing process-level information on PM formationand distribution and an opportunity to evaluateemerging measurement technologies related to PM.

Study Themes

The past measurements and those that we proposebelow are intended to provide a better understandingof the basic chemical, meteorological, and transportprocesses that determine ozone and fine particledistributions and new information to assist policy-makers in devising optimal ozone and PMmanagement strategies. These studies areencompassed in three broad themes.

Local vs. Regional Ð Regional ContrastsThe first area of investigation addresses whetherozone or fine particle pollution is a regional or a localproblem. The range provided by the WP-3D and theG-1 allows the composition of the atmosphere in aparticular location to be placed in a regionalperspective. Even though the concentration of ozoneand fine particles can be elevated over large areas, it isstill an open question how much of either ozone orfine particles are produced locally and how much areproduced remotely and then transported to a particularlocale.

Proposed Research

8

Although the study will be centered in theNashville/Middle Tennessee area flights are plannedfor the Mountain West and the Upper Midwest wheremeteorological conditions and the mix of ozone andPM-precursor emissions are expected to besignificantly different than in the Southeast. Thesemeasurements will build on previous NOAA studiesin Colorado and the Midwest and the recent BNLstudy in Phoenix.

Ozone and PM formation in PlumesThe second area of investigation relates to if and howozone or fine particles observed in a particularlocation can be attributed to a particular source ofprecursor compounds located among a complexmatrix of precursor sources. An ideal setting toaddress this question is an isolated urban area set in arural background with several large point sources ofpollution (e.g., fossil fuel burning power plants)imbedded at various distances with a wide range ofpollutant emission. Under certain general flowconditions the plumes of the power plants merge witheach other and/or the urban plume; in others theydonÕt. Hence, the synergism associated with theinteraction of power plants plumes, urban plumes, andthe regional background can be investigated under avariety of conditions and in various combinations. Asthe dimensions of the urban complex grows themagnitude of the problem becomes greater. A secondaspect of interaction involves the relation of ozonepollution to fine particle pollution. To date, ozonepollution has been largely treated as a local or sub-regional problem, while fine particle pollution hasbeen viewed as a local problem from the health-effects perspective and a regional problem from thevisibility perspective. The co-variation of thesepollutants has never been extensively investigated,particularly on the regional scale.

The study of the evolution of the pollutant mix inplumes is particularly useful in the determination ofpollution formation kinetics. The advection of plumesinto a reasonably uniform background airmassprovides a convenient ÒclockÓ allowing thequantification of the chemical rates that are critical tomodel development and evaluation.

Diurnal Cycle in Chemistry and MeteorologyIn the overwhelming majority of cases, intensive fieldstudies have targeted the study of daytime chemistryand dynamics. This interest is driven by a recognitionof the important role that photochemistry plays insecondary pollutant formation and a concern for theimpacts associated with daytime pollution exposuresfor both humans and plants. The pollutant mix can besignificantly affected by non-photochemical reactionsthat occur at night. The results of the 1995 SOS fieldstudy in Nashville highlighted the importance ofnighttime mixing processes in the redistribution ofpollution throughout the region.

This field study offers the opportunity to documentthe entire diurnal cycle of chemistry and meteorologyover one or more complete consecutive diurnal cyclesusing both surface-based and airborne observations.Information will be developed on processes that arenot well described in current models. Themeasurements will help us understand how ruraldaytime chemistry, which establishes the residualconvective mixed layer above the nocturnal boundarylayer, ultimately is coupled to the daytime urbanphotochemistry that establishes peak ozone and PMlevels in urban areas. Similarly, we will learn howdaytime urban pollution affects rural air quality on thefollowing day.

Focus Areas

In order to address the scientific themes identifiedabove, we propose a research program combining acomprehensive suite of aircraft and ground-basedmeasurements of the chemical and dynamicalproperties of the PBL and lower free troposphere.The proposed study focuses on six research areas.

Primary PM and PM precursor emissionsThe current research-grade ozone precursor inventoryfor the study area will be expanded to include bothnatural and anthropogenic sources of PM fine and PMprecursors, with particular emphasis on animal andcrop-agricultural sources of NH3.


9

PBL dynamicsStudies are proposed to characterize the effects ofvertical and horizontal transport on the concentrationsof ozone and aerosols and their precursors.

Ozone production efficiencyThe influence of NOX source strength and ambientVOC distributions on ozone production efficiency (thenumber of molecules of ozone produced per NOXmolecule emitted) will be investigated through studiesof urban and power plant plume chemistry.

Characterization of loss processesFollow-on studies will be conducted to determine thesource of the high NOX loss rates observed during the1995 field experiments. Both mass balance and tracertechniques will be employed.

VOC contribution to ozone and PM formation

The relative role of anthropogenic and biogenic VOCsin ozone and aerosol formation and aerosolcomposition will be investigated.

Fine particulate matter Ð formation and characterizationThe effect of atmospheric chemistry on thecomposition and morphology of ambient aerosols willbe investigated. The connections between theprocesses that control ozone and aerosol formationwill be studied. The relative importance of primaryand secondary aerosols, with their varied sources, toambient aerosol mass will be determined.

Nighttime chemistry and dynamicsThe processes that control the formation anddistribution of ozone and aerosols during the daytimeand nighttime will be contrasted. A particular effortwill be made to characterize the fate of plumes fromlarge power plants that are emitted above thenocturnal boundary layer.

Proposed Research

10

Nashville Science Plan

11

PRIMARY PM AND PM PRECURSOR EMISSIONS

Emissions inventory improvement has been ahallmark of SOS research from its inception in 1988.Before 1988, ozone precursor emissions inventoriesconsisted almost entirely of Òtypical summer dayÓestimates mainly of anthropogenic sources ofÒreactive hydrocarbons and nitrogen oxides.Ó Naturalemissions of isoprene and other reactive VOCs byforest trees, emissions of NO by soil microorganism inwell-fertilized row crops and pastures, and lightningin thunderstorms in some states with high frequenciesof lightning events, typically were omitted from ozoneprecursor inventories in many State ImplementationPlans. Carbon monoxide emissions also wereregarded as a separate environmental issue and thusnot included in some ozone precursor inventories.Although the phrase Ònonmethane hydrocarbonsÓ wascommonplace in ozone-management circles, the SOSNashville/Middle Tennessee Ozone Study and otherSOS investigations have demonstrated conclusivelythat both CO and methane, and oxygenated as well asnon-oxgenated VOCs, must be recognized in ozoneprecusor inventories that are to be used especially insmall-grid-scale atmospheric modeling studies andother research projects. Thus, SOS has developed thenotion of Òresearch grade emissions inventoriesÓwhich need to be compound specific, hourly andspatially resolved, grid based, environmentallycorrected, and arranged to include all natural sourcesand anthropogenic sources of precursor emissions.Furthermore, the research grad inventories should notbe based on unrealistic assumptions about Ònegligiblyreactive compoundsÓ or Ònegligible sources.Ó

As SOS and SOS-SCISSAP have been preparing forthe transition from an essentially Òozone onlyÓprogram to an Òozone and PMfineÓ program, the arrayof natural and anthropogenic precursor species ofconcern has had to be augmented still further toinclude sulfur dioxide (mainly from point sources) andgaseous ammonia (mainly from animal agriculture,but also including use of anhydrous ammonia in crop-based farming operations, and human waste treatment

plants). Thus, Carlos Cardelino of the GeorgiaInstitute of Technology, who has led SOS efforts todevelop research grade emissions inventories forSOSÕ since 1992, is joining with Patrick Zimmermanof the South Dakota School of Mines, Lowry Harperof the USDA-Agricultural Research Service inGeorgia, Wayne Robarge of North Carolina StateUniversity who is responsible for the ÒNitrogenBudget ProjectÓ for the State of North Carolina, theCensus of Agriculture in Washington DC, and theDepartments of Agriculture in the 10 states of the SOSregion in developing more accurate quantitativeestimates of ammonia, amine, and volatile amino acidemissions from a wide range of crop and animalagrcultural operations.

These studies are based largely on domestic and wildanimal and insect (especially termite) populationestimates for the several southern states. So far, thework has been mainly a literature survey based onpublished scientific studies and official records ofanimal waste treatment ÒpermitsÓ approved by stateregulatory agencies. These studies include bothpublished and as yet unpublished estimates ofemissions factors for volatile emissions from bothEuropean and North American animal rearingfacilities, animal waste storage lagoons and otheranimal waste treatment facilities, and landapplications of animal manures.

Ammonia Emission Estimates

Of the atmospheric gases that are part of the nitrogencycle, ammonia is the one that is more close related toagricultural systems. Current estimates indicate thatlivestock wastes and fertilizer losses are the dominantammonia sources in the U.S. They represent about75% of all NH3 emissions (NAPAP, 1990). Toestimate livestock waste emissions we will useactivity data from the States Statistician (USDA,1996) and from the 1992 Census of Agriculture

PM and PM Precursor Emissions

12

published by the Department of Commerce(DOC)(DOC, 1995). In agricultural publications byDOC, the number of head of cattle and calves,poultry, sheep, pigs and hogs, horses, goats, and minkare reported at the county level. The most recentlivestock waste emission factors were published as areport for EPA (Battye et al., 1994). This reportcontains ammonia emission factors in lb. NH3 perhead.

To estimate fertilizer loss emissions we will useactivity data from the Agriculture statisticians officeof the states involved or from the CommercialFertilizer Data Base compiled by the TennesseeValley Authority (TVA, 1990). Fertilizer databasescontain fertilizer usage at the county level for a variety

of fertilizers among which are those that emit NH3.Activity levels of fertilizers that emit NH3 are reportedin tons of Nitrogen. As with the livestock wastefactors, the most recent emission factors for fertilizerlosses appeared in the 1994 report (Battye et al.,1994). In this case, the emission factors are reportedas Kg of NH3 per ton of Nitrogen.

During the coming year, we propose to deriveammonia emission estimates for the SMRAQmodeling domain with a grid resolution of 54kilometers. In addition, we propose to develop a moredetailed ammonia emission inventory (with a gridresolution of a few kilometers), for theNashville/Middle Tennessee modeling domain.


13

PBL DYNAMICS

Pollutant concentrations at the surface are stronglyinfluenced by the physical behavior of theatmosphere, particularly the planetary boundary layer(PBL). Polluted air may exist in a "reservoir" layerabove the immediate PBL and be entrained as the PBLgrows, increasing pollutant concentrations near thesurface ("fumigation"); conversely, the overlying layermay be clean and entrainment may dilute theconcentrations near the surface. The reservoir layersthemselves are created by the collapse of the PBLupwind on the previous afternoon or by venting ofpolluted air out of the PBL by clouds and otherdetrainment mechanisms. Pollutants are transported atall levels, both within and above the PBL, and travelconsiderable distances even during periods of relativestagnation at the surface. In the daytime, verticaltransport (entrainment and venting) tends to dominate,especially during stagnation episodes; at night,horizontal transport is dominant. The overallobjective of the PBL dynamics component of thefield study is to understand the effects of vertical andhorizontal transport through the diurnal cycle on theconcentrations of ozone and its precursors.

Vertical Transport and Mixing

Vertical transport can be broken down intoentrainment, detrainment, and venting. Entrainment isthe capture of overlying air by the growing convectivePBL, this air is then mixed throughout the depth of thePBL. Detrainment describes the opposite process,where air is left behind by the collapsing PBL in theafternoon or is transported horizontally out of the PBLwhen the PBL depth decreases downwind. Venting isthe ejection of air from an active PBL by clouds.Entrainment is better understood than detrainment andventing.

Another important process is vertical mixing withinthe daytime mixed layer. Although potential

temperature is constant in the daytime mixed layer,other species such as water vapor, or ozone may notbe Ôwell mixed,Õ (i.e., constant with height). DuringNashville-99, both the gradient of ozone in the mixedlayer and the turbulence energy, as well as the surfacefluxes driving the mixing processes will be measured.

Objective: Entrainment / morning transitionD e s c r i p t i o n : Quantify the contribution ofentrainment to the concentrations of pollutants nearthe surface.Measurements:

Profilers - continuous, uniform daytimeoperation at high resolution for zi; half-hourly4-5 minute high-resolution RASSDoppler lidar - VAD scans for high resolutionwind profilesOzone lidar - continuous daytime operation,staring vertically, for vertical ozone andaerosol profilesCeilometers - continuous operationAirborne in-situ - constituent profiles duringthe morning as available; turbulence intensityas availableAirborne ozone lidar - ozone profiles duringthe morning as available; horizontaldistribution of ozone as available

Entrainment is quantified as fluxes of heat, moisture,and chemical constituents at the BL top. Since thesefluxes cannot be directly measured, they must beinferred from other measurements. Two basicapproaches are available, using entrainment velocitycoupled with a constituent profile or using budgetmethods. Both approaches have large uncertaintiesand both have limitations, but they arecomplementary. The approaches and some results arediscussed by Angevine et al. (1998).

For the entrainment velocity method, measurements ofBL height zi and its rate of change dzi/dt are key.

PBL Dynamics

14

These will be provided by the profiler network.Profiles of temperature and water vapor will be takenfrom the Nashville NWS soundings. Profiles of ozoneand other constituents will be provided by the ground-based and airborne ozone lidars and by the airbornein-situ chemistry measurements. The cloud altitudesmeasured by the ceilometers help to removeambiguities in the profiler zi measurements. Theprimary uncertainty in the entrainment velocitymethod is the rate of subsidence at the BL top, whichas a small mean of a strongly fluctuating velocity isquite difficult to measure. The profiler network andmodels may provide some estimates of subsidence.

The budget method requires measurements of the BL-mean value of a quantity and of its sources and sinks.For temperature, these are provided by theprofiler/RASS network and by the surface fluxnetwork. Key uncertainties are direct radiativeheating and advection, both of which can be estimatedby models. Since reliable surface flux estimates overthe urban area will be unavailable, the budget methodcan only be applied to the rural surroundings. Theozone budget can also be computed usingmeasurements from the ground-based ozone lidar, butreducing the uncertainties in the source terms toacceptable levels may be quite difficult.

Objective: Detrainment & venting / eveningtransitionDescription: Characterize the processes by whichpollutants are removed from the active PBL and madeavailable for horizontal transport.Measurements:

Profilers - continuous, uniform daytimeoperation at high resolution for zi and residualreflectivity layersSurface flux network - half-hourly estimatesof surface flux through the afternoon andeveningDoppler lidar - VAD scans for high resolutionwind profilesOzone lidar - continuous operation throughthe afternoon and evening for vertical ozoneand aerosol profilesCeilometers - continuous operation

Airborne in-situ - constituent profiles in theresidual layer during the evening as availableAirborne ozone lidar - ozone profiles duringthe evening as available; horizontaldistribution of ozone as available

Losses of PBL pollutants can occur by detrainment ofboundary layer air into the free troposphere either 1)during the evening transition when the CBL collapsesor 2) through horizontal transport in regions where themixing depth decreases downwind. Another lossmechanism is penetrative convection that vents thePBL of pollutants either temporarily, if air parcelssettle back into the PBL, or permanently, if air parcelsare detrained through cloud boundaries. Airbornechemistry measurements are also important indetermining the mix of species that remains aloft,from which the origin and time of emission may bedetermined.

Detrainment during the evening transition is to firstorder approximated by assuming that the BL is well-mixed to its maximum height, and that all the airbetween that height and some nominal nocturnal BLheight is detrained and available for horizontaltransport by 1800 LST. A greater level of detail of thetransition is needed, however. Wind and turbulenceprofiles from the profiler network and the Dopplerlidar during the afternoon will show the evolution ofvertical mixing and the time and space scales of itsdecay. The surface flux network will provideestimates of the decay of the surface fluxes, whichmay be modulated by clouds. The ground-basedozone lidar will show the formation of layers ofdiffering ozone content. Airborne i n - s i t umeasurements and the airborne ozone lidar may alsobe used to show the formation and evolution of theselayers if the aircraft can be flown well into theevening hours.

Venting of the active daytime BL by clouds isextremely difficult to estimate. The amount of cloudis shown by the ceilometers. Budget methods asdescribed above for entrainment may be feasible. If astrong contrast exists between the BL and above-BLvalues of some constituent, airborne i n - s i t umeasurements may be used to derive a budget of thatconstituent in the free atmosphere immediately above


15

the BL. Residual boundary-layer traces may bedetected by backscatter measurements from radarwind profilers and by aerosol and ozone lidars.

Objective: Daytime PBL characterizationDescription: Document the height and intensity ofmixing of the PBL during the day, including its spatialvariation.Measurements:

Profilers - continuous, uniform daytimeoperation at high resolution for zi and residualreflectivity layersSurface flux network - half-hourly estimatesof surface flux through the afternoon andeveningCeilometers - continuous operationS-band profiler - continuous, uniform daytimeoperation for cloud top and precipitationcharacterizationDoppler lidar - primarily vertical staringAirborne in-situ - constituent profiles in andabove the PBL as availableAirborne ozone lidar - ozone profiles asavailable; horizontal distribution of ozone asavailable

It is now recognized that vertical mixing processesoften dominate the diurnal cycle of ozone in the PBL(Kleinman et al., 1994; Neu et al., 1994). Losses bydry deposition at the surface and through the top ofthe mixed layer by detrainment are regulated byvertical mixing. Gains due to sources, production, andadvection are distributed in the vertical by mixingprocesses during the day. Therefore, the need exists tomeasure the temporal evolution of the boundary layerand understand its role in the production andmodulation of ozone and precursor gases. The growthof the morning convective layer allows the chemicalcomposition of the residual layer from the previousday, and perhaps from a different location, to mix tothe surface at the same time that photochemicalproduction of ozone is beginning. Ultimately, thegrowth rate of the PBL, local emissions emitted withinthe PBL, and the chemical composition of the residuallayer all contribute to determining whether there is netproduction or destruction of ozone in the PBL. Forexample, because of the parabolic relationship

between ozone production and NO, a slower growingboundary layer can result in either increasing ozoneby limiting the vertical dilution of precursorconcentrations or decreasing ozone by titration if NOis in excess. After the morning inversion is broken,the rapid growth of the midday boundary layer cancause ozone concentration to decrease even as thephotochemical production rate is reaching amaximum. Later in the afternoon, when ozoneconcentration often reaches its peak in the diurnalcycle, the mixing depth may become quasi-steadythrough the competing effects of subsidence andentrainment. The maximum mixing depth that isachieved depends both on surface forcing (land use,topography, partitioning of latent and sensible heatfluxes) and aloft processes (entrainment, subsidence,advection).

Another important parameter controlling pollutantconcentrations is the mixed layer depth. The mixingdepth depends both on surface forcing (land use,topography, partitioning of latent and sensible heatfluxes) and processes occurring aloft (entrainment,subsidence, advection). One of the interesting resultsfrom SOS-95 was the horizontal variability in mixingdepth observed in lidar transects and by the windprofiler array. In particular, Banta et al. (1998)speculated that the gradient in the peak afternoonmixing depths from the northwest to the southeast ofNashville was due to a difference in vegetationcharacteristics (forest versus open agricultural fields)and that the impact of these varying surfacecharacteristics was strongest under stagnantconditions. A further, possible explanation for thevariability in mixing depth observed during the mid-July stagnation episode was nocturnal advection of theurban residual layer toward the southeast, which led toless stability over a deeper layer downstream of thecity and, ultimately, a deeper daytime mixed layer inthe southeast region (White et al., 1998a). Anotherencouraging result from SOS-95 was that the mixingdepths determined by analyzing aerosol backscatterprofiles from the airborne lidar and reflectivityprofiles from the ground-based wind profilers were ingood agreement, indicating that the turbulenceinterpretation of the mixing depth provided by theprofilers is the correct interpretation for air pollutionapplications (White et al., 1998b).

PBL Dynamics

16

Additional factors affecting chemistry in the PBLarise when cumulus convection occurs. Clouds servetwo potentially important roles in the PBL (Neff,1998). First, clouds present an environment in whichaqueous-phase chemistry can occur. Second,penetrative convection vents the PBL of pollutantseither temporarily, if air parcels are recirculated intothe PBL, or permanently, if air parcels are detrainedthrough cloud boundaries. Clouds provide specialchallenges for remote sensors, for example, byinhibiting the performance of lidars and complicatingthe clear-air interpretation of radar returns. ForNashville-99 we propose to increase the number ofcloud sensing devices including commerciallyavailable laser ceilometers and a new, verticallypointing 3-GHz radar (White et al., 1998c).

As can be seen from these descriptions, a variety ofmeasurement systems are necessary to characterizethe full boundary layer evolution, including high-resolution, short-range instruments for the nocturnalPBL as well as the morning and evening transitionperiods, and long-range instruments for the daytimePBL. In addition, to study the horizontal distributionof mixing depth and to characterize the influence ofinhomogeneous surface conditions, a network ofsurface-based integrated observing systemssupplemented by aircraft is required. The network ofwind profilers and the airborne ozone/aerosol lidarwill be used to monitor the temporal and spatialevolution of the mixing depth. Aircraft with gustprobes (e.g. G-1) will fly beneath the airborne ozonelidar to obtain simultaneous measurements of mixedlayer depth and turbulence. Surface fluxmeasurements at the profiler sites will help todetermine to which extent the horizontalinhomogeneity in mixing depth observed during theSOS-95 campaign in the Nashville area depends onsurface forcing. The ground-based and airborneozone lidars will give a detailed picture of thetemporal evolution and spatial distribution of ozoneconcentration from near the surface to the lower freetroposphere. The profiler network as well as theDoppler lidar will provide continuous monitoring ofthe wind field. In addition, the Doppler lidar will yieldimportant turbulence quantities, such as profiles ofmomentum flux, TKE, and vertical velocity variance.The combination of highly resolved ozone lidar data

with vertical wind speed measurements from theDoppler lidar offers the opportunity to determine theturbulent ozone flux within the PBL directly using theeddy correlation method. Unfortunately, directmeasurements of subsidence, entrainment, and cloudventing are extremely difficult to obtainexperimentally. Divergence fields derived from SOS-95 profiler data were useful in predicting the sign ofthe mesoscale vertical motion, but questions regardingquantitative accuracy still remain (White et al.,1998a). One of the reasons for the planned profilerconfiguration in Nashville-99 (see Section III) is tominimize errors in the profiler divergencecalculations. Recent work applying Large EddySimulation (LES) models and lidar observations ofturbulence statistics shows promise in using the lidarobservations to tune the model, then applying themodel to estimate parameters that are not easilyobserved, such as entrainment.

Horizontal Transport

Objective: Daytime horizontal transportDescription: Document the spatial and temporalvariation of horizontal winds during the day.Measurements:

Profilers - continuous, uniform daytimeoperation at high resolution for zi and residualreflectivity layersAirborne in-situ - horizontal distribution ofwinds in and above the PBL as availableAirborne ozone lidar - ozone profiles asavailable; horizontal distribution of ozone asavailable

The spatial and temporal variation of the horizontalwinds in the daytime mixed layer control peakpollutant concentrations and the direction of transportof urban and power-plant plumes. Some specificissues include: causes of the narrowness of the urbanplume (not sampled at the New Hendersonville site),changes from NOx to VOC sensitivity over the urbanarea at low wind speeds, and transport aloft of ozoneand its precursors.

Wind profilers and the airborne systems (in situ andozone lidar) will document the spatial and temporal


17

variation of horizontal winds and pollutant transportduring the day. The airborne ozone lidar will traversepower-plant plumes farther downstream than duringthe SOS-95 campaign, to characterize the ozoneenhancement region of the plume and study its effecton background concentrations.

Objective: Nighttime horizontal transportDescription: Characterize the vertical and horizontalvariation of winds at night, with special emphasis onthe low-level jet and transport above the nocturnalPBL.Measurements:

Profilers - continuous, uniform daytimeoperation at high resolution for zi and residualreflectivity layersDoppler lidar - VAD and shallow angle RHIscansCeilometers - continuous operation to definecloud layersS-band profiler - continuous, uniformnighttime operation for cloud top andprecipitation characterization

During periods of high ozone and/or aerosol loading,the ability to document processes affecting horizontaltransport, including recirculation and stagnation, iscritical. Of particular importance is the transport aloftof ozone and its precursors (Bigler-Engler and Brown,1995; Trainer et al., 1995). Wind profilers wereinvaluable in assessing transport patterns during SOS-95. For example, in addition to describing the generalmeteorological conditions for the 1995 fieldcampaign, McNider et al. (1998) used a combinationof wind profiler data, GOES imagery, conventionalsondes, and Lagrangian particle models to describethe evolution of ozone episodes. The effects ofinertial oscillations can be seen in the generation ofazimuthal shear with height of the horizontal windwhich leads to enhanced dispersion (McNider et al.,1993) as well as in the generation of the classical low-level jets at nighttime that can produce enhancedtrapping near the surface in the presence of stablelapse rates (Neff, 1990). The importance of thenighttime low-level jet accelerations to pollutanttransport on days with light synoptic forcing,demonstrated during the SOS-95 campaign will againbe a focus for 1999. The previous study showed that

the location of pollutant layers could be determinedfrom trajectories derived from hourly profiler winds.Specific issues include: significance of the inertialoscillation on nights of somewhat stronger winds,documentation of the purging of the urban pollutionÔdomeÕ overnight during stagnation periods, plumeshape and diffusion at night as a function of windspeed, mixing time scales versus reaction rates,maintenance of vertical shear in the Òresidual layerÓ,the magnitude and interpretation of vertical gradientsof pollutant species, especially those subjected to drydeposition, the horizontal variability of the flow nearthe surface and aloft in the low-level jet, and the roleof nighttime transport in maintaining or contributingto background pollution levels.

Historically, horizontal transport has beencharacterized using two approaches (Neff, 1998). Thefirst approach is routine monitoring of the flow fieldusing arrays of instruments that are commonlyavailable, cost effective, and operate unattended or arealready part of operational networks. Presently, thisincludes sodars, wind profilers, rawinsondes, andsurface meteorological stations. However, thetechnique of using NEXRAD data to derive VADwinds (Michelson and Seaman, 1998) may eventuallybe used to supplement the sparse wind observingnetwork currently in place in the U.S. The secondapproach is required for detailed process studies andinvolves using research-grade instruments that mayrequire special attention such as lidars, scanningradars, and supplemental sounding systems includingrawinsondes, tethered sondes, and aircraft. Bothapproaches were used in SOS-95 and will be usedagain in Nashville-99. An additional focus ofNashville-99 will be nocturnal transport, which wasnot documented completely in SOS-95. We proposeto deploy a scanning Doppler lidar and a high-resolution Doppler sodar to measure velocity profilesand a tethered sounding system to measuretemperature and perhaps other scalar profiles at a sitefavorable for studying nocturnal boundary-layerphenomena. The lidar will provide both VAD andRHI scans.

PBL Dynamics

18

Modelling Data base

Mesoscale meteorological models are key tools in theanalysis of the meteorological and chemical results ofthe field campaign. They provide the ability to fill infor missing data with physically consistent fields,initialize chemical models, and predict unmeasuredvariables and cases. There are two objectives underthis heading, the measurement data base and theexternal model database. The actual modelverification and use objectives are beyond the scopeof this science plan.

Objective: Measurement databaseDescription: Produce a set of meteorological datathat will allow for the verification and improvement ofmesoscale meteorological models to represent themeasured behavior of the atmosphere during thecampaign and to predict unmeasured variables andcases.Measurements: All

Still at issue is the proper balance between models andobservations for air-quality assessment (Seaman,1998), where in many cases, the observations are usedto initialize and then nudge prognostic models back to"reality." This issue assumes greater importance asmeteorological models are run with finer resolution

and used as pre-processors for photochemical gridmodels (Lyons et al., 1995). The instrumentsdeployed during Nashville-99 will provide acomprehensive set of measurements for detailedcomparisons of numerical simulations withobservations. The data set will also serve in testingand improving the PBL parameterization schemescontained within the models (e.g., Zamora et al. ,1998). Forecasts of the mixed layer depth are mostimportant. Realistic modelling of mixed layer depthsrequires a detailed specification of the mesoscalesubsidence rate and the surface energy budget.

Objective: External model databaseDescription: Archive fields from runs of MM5 andEta models and analyses made by outsideorganizations for future use as initial and boundaryconditions and as verification for our own modelingefforts.Measurements: NCEP Eta analysis MM5 analysis

Mesoscale numerical models such as the ColoradoState University RAMS system, the NCEP Eta modeland the Penn State/NCAR MM5 system are allcapable of providing the necessary inputs to both thechemical and dispersion models.


19

OZONE PRODUCTION EFFICIENCY

Ground-Based Measurements

Much of the research that has been conducted toinvestigate the relationship between O3 production andNOx and VOC levels has taken place in rural orremote areas, that is where the mixing ratios of NOxare only a few ppbv or less. The results of thisresearch have demonstrated a number of key points.Photochemical production of O3 cannot occur in theabsence of NOx; indeed, at very low levels of NOxphotochemistry destroys O3. At somewhat higherlevels of NOx it has been determined that O3production becomes highly non-linear with respect toNOx and that O3 production peaks and then falls off(Liu et al., 1987; Lin et al., 1988). The level of NOxwhere this roll-off occurs also appears to depend onthe levels and relative composition of the VOCspecies present. Much less research has focused oninvestigating O3 photochemistry for high NOx, that isin urban plumes and point source (power plant)plumes. Modeling studies indicate that O3 productionefficiency, that is the amount of O3 produced per NOxoxidized, should be reduced at high loading of NOxdue to scavenging by NOx of the radical species thatare required for O3 production. In fact, some previousresearch (Ryerson et al., 1998) indicates that the O3production efficiency can be large for urban plumesand for small power plant plumes, while for muchlarger power plants (and consequently much largerNOx emission rates) O3 production efficiency appearsto be sizably lowered. Since the exact mechanism thatcauses this latter effect is unknown at present, theseobservations clearly demonstrate that ourunderstanding of the photochemistry that occurs inthese high NOx environments is incomplete. Giventhat the air quality standard for O3 has recently beenmade more stringent, it becomes imperative that theprocesses that lead to O3 production be wellunderstood in order to implement effective controlstrategies.

The multi-year study of O3 photochemistry in thesoutheastern United States by the Southern OxidantsStudy has focused on the contribution of rural areas tolarge scale O3 pollution episodes, especially withrespect to the impact that biogenic emissions of NOxand VOC have on production of O3 in urban areas.More recently, studies of suburban and power plantplume chemistry in the Middle Tennessee regionsurrounding Nashville have been conducted,principally from aircraft platforms. Thesemeasurements have been valuable for probingphotochemistry over a wide spatial and temporaldomains, and under some conditions there has beensufficient data collected at high NOx to explore O3production. Unfortunately, those groundmeasurements that have been taken have lacked thedata at high NOx because the measurement sites havebeen too far away from the urban ÒcenterÓ. This wastrue, in particular, for the two Nashville intensivestudies in 1994 and 1995. On those infrequentoccasions when the urban plume did pass over the siteat Hendersonville, the photochemistry had alreadypeaked with respect to O3 production. For the 1999Nashville field study, in addition to the aircraftmeasurements, ground-based sampling will beconducted in order to acquire a data record ofsufficient length at a location that has been chosen tohave a high probability of near-field exposure to theoutflow from the urban environment for a largefraction of the measurement period. The site and theproposed suite of measurements are describedelsewhere in this document.

A number of approaches for the determination of theO3 production rate (as a function of NOx and VOC)are proposed for the ground-based portion of the 1999study. Determination of the O3 production raterequires the measurement of NO and the radicalspecies that oxidize NO to NO2, which in turn is

Ozone Production Efficiency

20

photolyzed to produce O3 . High qualitymeasurements of NO are routinely made, butdetermination of the peroxy radical species is verydifficult. We intend to obtain data that will allowcomparison of estimated HOx and RO2 (via modeling)to those from direct and indirect measurements.

1. One direct measurement of these species that willbe performed is the Chemical Amplifier (CA)technique. This instrument utilizes the cyclingthat occurs between OH and HO2 when the peroxyradical oxidizes NO in the presence of CO toproduce an ÒamplificationÓ in the amount of NO2produced, which is then sensitively determined.An improved CA instrument will be provided andoperated by personnel from TVA.

2. Measurements of OH and HO2 will also be takenduring the study. While this is not adetermination of the complete suite of peroxyradicals, reliable HO2 data can provide a lowerlimit to the rate of O3 production. Thesemeasurements are based on the laser-inducedfluorescence detection of OH, where HO2 ismeasured as OH after being titrated by an excessof NO in the instrument inlet system. Thisinstrument will be provided and operated bypersonnel from the Pennsylvania State University.

3. In many cases an indirect determination is madethat infers the radical levels by calculating thedeviation of the ratio of NO2 to NO from theexpected photochemical steady state (PSS) ratiothat is induced with O3 as the sole NO oxidantunder given conditions of light intensity. Thisapproach requires, in addition to NO data, highquality measurements of NO2, O3, and thephotolysis rate of NO2. The suite ofmeasurements that are proposed for the Level IIIsite can provide the necessary data.

4. Photochemical models can be used to derive theperoxy radical levels by estimating theconcentration of radicals produced from knownphotochemical reactions of VOC species that aremeasured at the site. This approach requires aconsiderably larger suite of measurements thanthat of the PSS above, but the results are generally

more robust (Frost et al., 1998). Many of themeasurements required by the model that weintend to use will be available from the urbanground site.

It is expected that the above approaches forestimating the peroxy radical levels will be synergisticin that the more truly independent methods that areused to determine a given quantity, the more robustwill be the result. At the least, redundantmeasurements are more likely to result in useful datashould any given instrument fail.

Aircraft measurements

Important new information on the transformation andloss of pollutants in urban and power plant plumeswas obtained during the 1995 Nashville/MiddleTennessee Ozone Field Study. As discussed above,analyses of these data indicate rapid removal of NOXin power plant plumes and ozone productionefficiencies that were much lower than previouslyreported. By making detailed chemical measurementsat multiple down-wind distances using highly-instrumented aircraft it was possible to describe theevolution of the pollutant mix to a degree that had notpreviously been possible. Two methods were used tocorrect for concentration changes resulting from non-chemical processes (plume dilution, venting, surfacedeposition etc.). These methods have been describedin detail by Ryerson et al. (1998).

1. Mass balanceThe mass balance approach for NOx conversion andO3 production uses the integral of trace species ofinterest over the entire volume section of the plume ateach flight transect. The change in the integratedamount is determined from plots of net mixing ratio inthe plume (plume minus surrounding backgroundlevels) versus distance (or time) downwind from thesource. Knowledge of the emission rates (assumedconstant) from the source, measurements of the windfield affecting plume transport, and estimates ofdeposition (via calculation) and venting loss processes(estimated from longer-lived species, e.g., SO2) areused with the measured change from the plot to


21

estimate the conversion of NOx as a residual. Theproduction of O3 is determined in a like manner.Comparison of NOx conversion and O3 production fordifferent points on a plume and among differentplumes provides insight into the efficiency by whichO3 is generated by the different NOx sources, theozone production efficiency, or OPE. In addition toOPE determinations, the mass balance approach candetermine effective boundary layer removal rates forvarious trace gases. An improvement expected for the1999 study over previous studies will be the in situaircraft measurement of NO2 rather than derivation ofNO2 from PSS calculations.

2. Pollutant ratiosDuring the transport of plumes, mixing effects withinthe boundary layer significantly alter the plumecomposition downwind. These mixing effects can betaken into account by using the ratio of the level ofthat species (minus the background) to that of a co-emitted tracer species. Measurements of the ratio as afunction of downwind distance (or time) provide adetermination of the removal of one species relative tothe other. When a longer-lived species, such as SO2,is used for the tracer compound, then the ratio of O3 toSO2 versus the ratio of NOx to SO2 providesinformation about the photochemical change of O3relative to NOx which yields the OPE. In contrast tothe mass-balance approach described above whereestimates of loss rates are needed, the concentrationratio method accounts for these effects explicitly,depending on the nature of the tracer.

In 1999 both of the above approaches will be used.However, since the meteorological requirement for thesuccessful application of the mass balance approachare quite restrictive it is desirable that a tracer beavailable to allow for the quantification of lifetimes ofplume constituents. Current plans call for the use oftwo types of tracers: 1) tracers of opportunity and 2)artificial tracers.

1. Tracers of opportunityThese are compounds that are present in either theurban or power plant emissions in sufficientabundance that they can be accurately quantified fardownwind (>50 km). These are semi-conservedtracers whose deposition and transformation rates are

sufficiently low and well quantified that theirmeasurement provides a means for normalizing theconcentrations of more reactive pollutants. Thesuccessful application of this method also requires thatthe emission ratios be well quantified and that the timeresponse characteristics of the aircraft instrumentationbe well characterized. Three tracers of opportunityare being considered: SO2, CO and CO2. Thecharacteristics of each are described below.

SO2 - This compound can be used as a tracerfor power plant plumes. Large power plantsare required to make continuous in-stackmeasurements of SO2 with their ContinuousEmission Measurement Systems (CEMS).The deposition velocity and atmosphericoxidation rate for SO2, although low (~1 cmsec-1 and 0.01-0.05 hr-1 respectively) arenevertheless significant and make SO2 a lessthan ideal tracer. However, lifetimescalculated using SO2 as a tracer comparedfavorably with estimates using the massbalance approach (Ryerson, et al., 1998).

CO - This compound can be used as a tracerfor urban plumes. Urban emissions arereasonably well quantified and measurementsof CO and NOY made in the Nashville urbancore during 1995 agreed well with theexpected emission ratio based on availableinventory data. Carbon monoxide has a verylow deposition velocity (< 0.1 cm sec-1) andphotochemical production and loss areexpected to have only a minor effect on themeasured concentration.

CO2 - This compound can be used for eitherpower plant or urban plumes but applicationto power plant plumes appears mostpromising. This compound is also monitoredon a continuous basis with the CEMS in largepower plants. Fast response instrumentationare available to quantify this compound withsufficient sensitivity and precision to detect itin power plant plumes above the very largebackground (~360 ppm) that is present in theatmosphere. Background variability may

Ozone Production Efficiency

22

be a greater problem in the use of CO2 as atracer.

2. Artificial tracerDuring the Nashville 1999 field campaign the use ofthe above tracers of opportunity will be augmented bythe use of an artificial tracer. The tracer will beintroduced into TVAÕs Cumberland power plant at arate proportional to plant emissions. Canister samplescollected on the NOAA WP-3 will be analyzed for thetracer, providing a direct measure of plume dilution.

To be useful for this purpose the tracer must have thefollowing characteristics:· The tracer must have low atmospheric reactivity

and surface deposition rates.· The tracer must be detected with high sensitivity

and selectivity.

· The tracer must be safe to handle andenvironmentally benign.

Tracers that have been traditionally used foratmospheric dispersion studies (SF6 andperfluorocarbons) were not considered since thesecompounds have long atmospheric lifetimes andsignificant global warming potentials.

A hydrofluorocarbon, HFC 152a (1,1,-difluoroethane)has been selected for use and this application isundergoing safety review by TVA. This compoundappears to satisfy all of the above criteria. It has anexpected atmospheric atmospheric lifetime, based onremoval by reaction with hydroxyl radical, ofapproximately 1.5 years and a small global warmingpotential.


23

CHARACTERIZATION OF LOSS PROCESSES

Analysis of power plant and urban plume data forthe measurements made in 1994 and 1995 suggeststhat NOx is rapidly depleted in these plumes in theSoutheast. Cross-plume pollutant profiles obtainedby the NOAA WP-3D were combined with detailedwind fields to calculate NOx fluxes as a function ofdownwind distances for multiple plumes andindicated that NOx removal rates were significantlylarger than those expected from model simulationsbased on the best available knowledge. This impliesthat either the plume chemistry results in a NOxreaction product that cannot be effectively detectedwith the NOy measurement systems employed orthat surface deposition, boundary layer ventingand/or uptake of NOy species on aerosols may beoccurring at a rate much faster than presentlypredicted.

Ozone is a by-product in the oxidation of naturaland anthropogenic hydrocarbons in the presence ofnitrogen oxides (NOx =NO +NO2). NOx acts as acatalyst in this process. In the study region, thesoutheastern United States, NOx is mainly directlyemitted of from combustion sources and thus ofanthropogenic origin. Through photochemistry theprimary emitted NO and NO2 are oxidized and canbe converted into reservoir species. Those arecompounds as organic nitrates that are less reactiveand therefore have a longer lifetime in theatmosphere. Enhanced stability allows the transportof reactive nitrogen over long distances to theremote atmosphere where these species can bereconverted to more photochemically active forms.In this way even remote regions without any localpollution sources can be effected. The ultimate fateof the reactive nitrogen oxides is the conversion t oHNO3 or nitrate aerosol and their removal from theatmosphere through dry deposition or washout byrain.

A goal of the SOS ozone research is to gain theability to attribute the local ozone abundances t othe different precursor sources. The precursor

distributions are shaped by emissions,photochemical conversions that either produce orremove these trace gases from activephotochemistry, other losses like wet deposition orrainout and transport. Any reliable assessment ofhow the ozone production can attributed t odifferent precursor sources is dependent on how wellone can account for the precursor distributions.

The local ozone production is dependent on theavailability of the catalyst. Measurements fromprior SOS experiments suggest that throughout mostof the southeastern U.S. the production of ozone isNOx limited. Therefore, it is most important t ounderstand what happens to the NOx emitted. Howlong that is for how many oxidation cycles are thenitrogen oxides available as catalyst? This,combined with the above cited preliminary resultfrom SOS 94/95 - more rapid loss of NOx thanpresently understood - pose a severe question thatneeds to be addressed in the upcoming SOS 99 study.Is the observed rapid loss real and if so, what is thecause of the rapid loss of NOx that we do notunderstand at this time?

Several potential causes need to be considered for aseemingly too rapid loss:

· Measurement artifact· Incorrectly quantified physical removal of NOx

from airmass (deposition, venting of PBL)· Loss on aerosol

Measurements of the nitrogen oxides are made bywidely accepted techniques (chemiluminescencedetection of NO, photolysis of NO2 to NOcombined with chemiluminescence detection of NO,conversion of all reactive nitrogen componentspecies to NO combined with chemiluminescencedetection of NO). In the laboratory and in groundbased measurements these instruments have beenchallenged with the compounds that are thought t obe the major components of NOy (ie NO, NO2,

Characterization of Loss processes

24

HNO3, PAN, PPN, and other organic nitrates).Many of the individual species have been measuredconcurrently in filed studies. By comparison withthe measurement of total NOy the budget and withit the current understanding can be tested. With therich, complex hydrocarbon mix present in theseatmospheres a variety of organic nitrates can beexpected, and indeed, a number of nitrates havebeen identified and measured in the southeast. Theair mix though is much to complicated to accountfor all the hydrocarbon precursors and reactionproducts. This leaves the possibility thatunidentfied reaction products may be formed thatare specific to this region.

SOS sponsored in 1994 at Nashville a ground basedintercomparison of these techniques with generallygood agreement (Williams et al., in press). Togetherwith the results of other ground-basedintercomparisons (NASA GTE, Fehsenfeld et al,Fehsenfeld et al.) one has to conclude that these arethe best measurements available and that they havebeen thoroughly tested.

The inflight intercomparison of NOy measurementsduring SOS-95 though, revealed significantdifferences. Therefore, it has to be investigatedwhether the deployment of these instrument fromaircraft in the southeastern US with its complex airmatrix can lead to yet unappreciated measurementartifacts. Sampling of reactive trace gases from anaircraft is a difficult task and the impact of the usedinlets on the sampled airmass thoroughly evaluated.Another explanation for the seemingly fast losscould be the removal of NOx through physicalprocesses, either deposition or cloud venting at thetop of the Planetary Boundary Layer (PBL). Tracegas deposition velocities for nitrogen oxides havebeen measured and agree reasonably well. Ventingat the top of the boundary layer is difficult t oquantify (see above) but was thought not t osignificantly exchange air from the PBL to LFTduring the observations of SOS 1994/5. Ambientaerosol might supply enough surface area forreactions or uptake and help removal or conversionof nitrogen oxides from the plumes.

Objectives

¥ Make measurements of reactive nitrogenoxides

¥ Validate them by intercomparison andsupport measurements

¥ Investigate whether loss of nitrogen oxidesin urban and power plant plumes is inagreement with current understanding

¥ Investigate whether loss can be attributed t ohigher deposition or more effective ventingof PBL

¥ Investigate whether loss correlates withaerosol loading

Plans for 1999

A first step is the verification of the rapid NOxdepletion that was indicated in the SOS 94, 95measurements. Measurements will be made bymultiple investigators and thoroughly compared. Inaddition, two of the platforms will attempt to assessthe reactive nitrogen oxide budget by measuring themajor individual species (NO, NO2, HNO3, PAN ,organic nitrates) and total reactive nitrogen oxides(NOy). It is planned to measure under variousmeteorological conditions and to contrast themeasurements in Tennessee with the Midwesternregion north of the Ohio river and measurements inthe Mountain West. Tracers will also be employedas has been described above.

The measurements on the NOAA WP-3D will beenhanced by the deployment of a new instrument t omake fast in-situ measurements of HNO3 bychemical ionization mass spectrometry. This willallow to follow the downwind evolution of theprimary gases (NOx) and its photooxidationproduct.

The possibility of lidar measurements on the NOAAWP-3D to determine the local PBL depths is beingevaluated.

More emphasis will be put on investigating thevertical structure of plumes downwind. Crosswindprofiles of plumes at various heights will show howfar vertical mixing leads to constant volume mixing


25

ratios through the full PBL height. Overflight ofthe plumes just above the PBL top will indicatewhether venting will effectively remove gases froma plume that travels in the PBL into the lower freetroposphere.

The addition of two instruments thatmeasure the aerosol size distribution from the

nanometer (nm) to micrometer (mm) range willallow to follow the development of the aerosolsurface area that could be available for surfacereactions in plumes. Any uptake of reactivenitrogen species should result in a modification ofthe aerosol distributions in the plumes along thedownwind distance.

Characterization of Loss processes

26


27

VOC CONTRIBUTION TO OZONE AND PM FORMATION

Volatile organic compounds (VOCs) play a key rolein the formation of ozone and PM. Emissions ofanthropogenic VOCs have declined significantlyduring the last two decades (U.S. EPA, 1997d) due toemission management programs that target thesecompounds. As anthropogenic emissions havedecreased biogenic emissions of reactive VOCs, suchas isoprene and terpenes, play a greater role in ozoneand PM formation especially in heavily vegetatedareas such as the Southeast. Analysis of VOC datacollected during the 1995 SOS study indicated that

carbon monoxide and methane can also make asignificant contribution to VOC reactivity, especiallyin rural areas without significant biogenic emissions(Figure 1). These latter compounds are usually notincluded in VOC emission management programsbecause of their low reactivity. A betterunderstanding of the relative role of anthropogenic,biogenic and low reactivity VOCs (i.e. methane andCO) to atmospheric reactivity is essential to thedevelopment of effective pollutant managementstrategies.

Figure 1. VOC reactivity distribution as a function of altitude. Data are from canister samples collected aboardthe DOE/BNL, G-1 during the 1995 Nashville/Middle Tennessee field study. Figure provided by L. Kleinman.

Although the role of VOCs in ozone formation is wellunderstood their role in PM formation is lessappreciated. VOCs can contribute to PM formation intwo ways:

1. Generation of free radicals Ð Obviously VOCsare a key ingredient in the pollutant mixresponsible for sustaining the atmosphericpopulation of free radicals such as HO and HO2.

VOC Contribution to O3 and PM

28

The radicals can oxidize SO2 and NO2 directly,with subsequent formation of sulfate and nitrateaerosols.

HO + SO2 + (O2, H2O) Õ H2SO4HO + NO2 Õ HNO3H2SO4 + NH3 Õ sulfate aerosolHNO3 + NH3 Ö nitrate aerosol

These radicals also influence the formation ofozone and peroxide compounds that areresponsible for the oxidation of SO2 in clouds,which is an additional source of sulfate aerosols.

2 . Formation of organic aerosols Ð Low vaporpressure products of atmospheric photo-oxidationcontribute to fine aerosol mass in the atmospherethrough condensation onto pre-existing particles.The aromatic compounds found in gasoline havebeen shown (Odum et al., 1997) to react underatmospheric conditions to form aerosols.Biogenic VOCs are another potential source ofatmospheric aerosols. Laboratory studies suggestthat isoprene photoxodation does not lead tosignificant aerosol formation. However, pinenesand other monoterpenes form aerosols throughreaction with ozone and hydroxyl radicals (U.S.EPA, 1996).

During the 1999 field campaign the role of variousclasses of VOCs (anthropogenic, biogenic, lowreactivity) in promoting the formation of ozone andPM will be evaluated using the approaches describedbelow. Both ground-based and airborne sampling willbe employed. The airborne measurement capabilitywill afford the opportunity to conduct measurementsin different regions (Southeast, Midwest, MountainWest) with a variety of VOC pollutant regimes.

Quantification of VOC EmissionsThe accurate quantification of VOC emissions iscrucial to the successful interpretation of ambientmeasurements performed during the study. The VOCemission mix contained in the Nashville emissioninventory will be evaluated against data collectedduring a tunnel study conducted in the Nashvilleurban area during the 1995 field campaign.Additional checks on the VOC emission mix and

source strength will be provided by VOC samplescollected in the urban core during the 1999 study.VOC canister samples will be collected daily at thedowntown sampling site, located on top of an 18-storey building. In addition, canister samples will becollected for VOC analysis at several elevations abovethe city center using the TVA helicopter. Thesevertical profiles will be repeated several times duringthe study.

Ambient VOC reactivityAmbient VOC concentrations will be determined byseveral means during the study. 1) Daily canistersamples (1-hr integrated) will be collected andanalyzed for VOCs at the upwind, background site(Dixon) the downwind urban site (Cornelia Fort) andthe downtown site (Polk Building). 2) The TVAhelicopter, the DOE G-1 and the NOAA WP-3 willeach collect canister samples for VOC analysis. 3)The DOE G-1 and the NOAA WP-3 will be equippedwith systems (atmospheric pressure massspectrometry and gas chromatography respectively)for real-time VOC analysis.

The VOC data from the sources described above willbe used to evaluate the contribution the various VOCsmeasured to ozone formation. The contribution ofindividual compounds to overall VOCs reactivity canbe estimated in several ways:

1. HO reactivity Ð The rate of reaction with the HOradical provides a convenient measure of thecontribution of a particular VOC to propagation ofthe free radical chain that is responsible for ozoneformation.

2 . Incremental reactivity Ð The use of ozonereactivity scales for VOCs (Carter, 1994) providesa more direct means of weighting the contributionof individual components of a VOC mix withrespect to ozone formation.

3. Model simulations Ð Photochemical grid modelscan be used to determine the relative sensitivity ofozone formation to the concentration of indivudalVOCs.


29

Biogenic/anthropogenic contribution to ozoneformation:Recent analyses (Roberts et al., in press) havedemonstrated the use of relationships betweenobserved photochemical products, PAN (peroxyaceticnitric anhydride) and PPN (peroxypropionic nitricanhydride) and MPAN (peroxy-methacrylic nitricanhydride) and ozone to estimate the contribution ofanthropogenic and biogenic VOCs to regional ozoneproduction. In this approach MPAN is used as amarker for chemistry driven by biogenic VOCs andPPN as a marker for chemistry driven byanthropogenic VOCs.

Measurements of these organic nitrates will beperformed at the downwind urban site (Cornelia Fort)

and onboard the NOAA WP-3. Analysis of these datawill provide a measure of the contribution of biogenicand anthropogenic VOCs to regional ozone formationindependent of source estimates or chemicalmechanisms.

VOC Contribution to Aerosol MassThe careful measurement of the organic fraction of thefine aerosol mass will provide a direct measure of theVOC contribution to aerosol formation. Thesemeasurements will be performed at the downwindurban site (Cornelia Fort) and are described in moredetail in the section ÒFine Particulate Matter ÐFormation and CharacterizationÓ.

VOC Contribution to O3 and PM

30


31

FINE PARTICULATE MATTERFORMATION AND CHARACTERIZATION

Airborne particles may have many sources andcontain hundreds of inorganic and thousands oforganic components. Atmospheric particles aredistributed by size into fine and coarse modes with asplit at about 2.5mm. Course mode particles in therange 2.5-10 mm are largely deposited in the nasal-pharyngeal areas of the respiratory system, whereasparticles smaller than 2.5 mm are more likely to reachthe lungs. Fine mode particles, principally those inthe 0.1-1.0 mm range, scatter visible light efficientlyand are major contributors to the regional haze that isprevalent in the Southeast during the summer.

The atmospheric particles that populate the fine andcoarse modes are characterized by different sources,composition, and atmospheric behavior. Coarse Modeparticle mass is dominated by primary particles suchas fugitive dust and fly ash. In contrast, much of thefine particle mass in the atmosphere is formed through

the chemical transformation of gaseous precursorssuch as SO2 and NH3. Aerosol studies conductedduring the field study will focus on improving ourunderstanding of the processes that control theformation and distribution of fine particles in theatmosphere.

Chemical Composition and Morphology ofAtmospheric Aerosols

The careful characterization of atmospheric aerosolsin terms of chemical composition and morphology notonly provides important clues regarding their sourcesand possible mechanisms of formation but alsoprovides valuable information for those studyinghealth effects of particulate matter (PM).

Figure 2. Composition of PM2.5 aerosol mass in the U.S. Source U.S. EPA, 1996.

The chemical composition of ambient PM2.5 aerosolmass varies significantly across the country (Figure

2). Sulfate constitutes a significant fraction of the fineparticle mass in the East with nitrate and carbon

WestCentral

East

Sulfate Nitrate

Carbon Minerals

Unknown

PM Formation and Characterization

32

(elemental and organic) playing lesser roles. In theWest, on the other hand, the fine particle mass isdominated by nitrate and carbon with sulfate making asmaller contribution. While carbon-containingaerosols comprise only about 25% of PM2.5 mass inthe East they are the dominant fraction of PM2.5 in theCentral and West regions of the country. Thesedifferences in the chemical composition of PM2.5reflect differences in emissions in the two regions. Asthe regulations lower the emission limits for SO2 andNOx, carbon-containing aerosols will represent anincreasingly larger fraction of PM2.5 throughout thecountry.

For aerosol samples collected in the eastern U.S., asignificant fraction (typically 20-30%) of the PM2.5mass cannot be readily identified. This ÒmissingÓmass is believed to be comprised of semivolatilematerials such as water, ammonium nitrate andorganics (Andrews et al., submitted for publication).Of particular concern are the semi-volatile organics,which are not efficiently captured by traditional filtertechniques. Eatough et al. (1995) estimated thatbetween 20 and 60% of the fine particle organic massis lost when these techniques are employed. Theproper quantification of these materials is particularlyimportant in light of their potential health impacts.

Another issue of concern in the determination of fineparticle chemical composition is the length of timeover which the sample is taken. Filter samples aretypically collected over a 24-hr period, whichdiminishes their usefulness for source apportionmentdue to the variation in transport winds that can occurover such an extended period. In addition, chemicalcompositions determined from such samples may notaccurately reflect the makeup of coexisting aerosols.

Thus, being cognizant of these issues, systems havebeen selected that permit the determination of fineparticle chemical composition with: 1) a timeresolution compatible with air mass trajectoryanalysis, and 2) under conditions that allow thereliable quantification of the organic fraction of theaerosol mass. The sample sizes necessary forspeciation of the organic fraction will require longer(6-12hr) sampling times.

Filter-based samplersThese will be deployed both at the ground sites and onthe in situ sampling aircraft. The airborne particlesamplers will operate with short sampling times (


33

be analyzed using both gas chromatography/massspectrometric and liquid chromatography/massspectrometric techniques.

Particle Analysis by Laser Mass Spectrometry(PALMS)PALMS is a real-time in-situ technique for analyzingthe chemical composition of individual atmosphericaerosol particles (Murphy et al., 1995; Murphy et al.,1998). PALMS has the unique feature of detecting awide variety of elemental and chemical species inindividual particles. The advantage of deployingPALMS at the ÒsuperÓ chemistry site is that it can beused to estimate the contribution of different particleclasses (i.e. sulfate, nitrate, organic material, soot, seasalt, crustal material, and/or metals) to the overallpopulation and the extent of internally-mixedparticles. By placing a differential mobility analyzeror a PM 2.5 impactor in front of the PALMS inlet,compositional differences as a function of particle sizecan also be investigated. After completion of the fieldcampaign, the individual particle spectra may begrouped according to sampling conditions such astime of day, meteorological conditions, and trace gasconcentrations and compared with other particlecomposition measurements. Two limitations of thistechnique are that quantitative information of thecomponents present and molecular speciation oforganic compounds is not obtained. However, thetechniques described above will be used to providespeciation of the organic fraction of the fineparticulate mass. Thus, PALMS provides importantcomplementary information to traditional bulkcompositional analysis of atmospheric particles.

Aerosol Formation Rates

As stated above, particulate matter formed as a resultof gas-to-par

Documents

1.COVER & Table of Contents › projects › sos99 › scienceplan.pdfTABLE OF CONTENTS Introduction 1 Proposed Research 7 Primary PM and PM Precursor Emissions 11 PBL Dynamics 13