Ozone variability in the atmospheric boundary layer in Maryland and its implications for vertical transport model

of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Fashion & Beauty

Published:

Views: 53 | Pages: 11

Extension: PDF | Download: 0

Share
Description
Ozone variability in the atmospheric boundary layer in Maryland and its implications for vertical transport model
Tags
Transcript
  Ozone variability in the atmospheric boundary layer in Marylandand its implications for vertical transport model Xiao-Ming Hu a , * , David C. Doughty a , Kevin J. Sanchez a , Everette Joseph b , Jose D. Fuentes a , ** a Department of Meteorology, The Pennsylvania State University, University Park, PA 16802, USA b Department of Physics and Astronomy, Howard University, Washington, DC 20059, USA a r t i c l e i n f o  Article history: Received 14 April 2011Received in revised form24 September 2011Accepted 26 September 2011 Keywords: OzonesondeBeltsvilleWRF/ChemResidual layerNocturnal boundary layer a b s t r a c t Although much research has focused on daytime ozone (O 3 ) distribution in the atmospheric boundarylayer, there remain many unresolved processes related to O 3  transport in the residual layer. To addresssuch unresolved questions, a  fi eld study was conducted in Beltsville, MD during the summer of 2010 tostudy the spatial and temporal distribution of O 3  and other pollutants using ground-based gas analyzersand ozonesondes. During elevated pollution events in the daytime, the convective boundary layer, whichreached a maximum depth of about 2 km, had nearly uniform O 3  levels of almost 100 parts per billion(ppbv). Due to intermittent and intense vertical turbulent motion, the residual layer became  “ leaky ”  andpermitted vertical transport to enhance ground-level O 3  mixing ratios by as much as 10 e 30 ppbv ina span of 0.5 e 3 h. Model simulations, using the Weather Research and Forecasting model with Chemistry(WRF/Chem), were carried out to investigate the impact of different treatments of vertical mixing on thesimulation of O 3  in the nocturnal boundary layer and residual layer. WRF/Chem model simulationsprovided realistic O 3  vertical distribution during the daytime. During the nighttime, in the residual layer,model outputs resulted in higher O 3  levels compared with the  in-situ  observations. Model sensitivityanalyses showed that increasing the turbulent length scales and improved stability functions yieldedimprovements in the vertical transport of O 3  within the residual layer. One keyconclusion of this study isthat models such as WRF/Chem require improved numerical algorithms to properly account for thenocturnal vertical transport of O 3  in the residual region of the atmospheric boundary layer.   2011 Elsevier Ltd. All rights reserved. 1. Introduction Uncertainties exist in the manner that numerical modelsdetermine vertical transport of pollutants within the atmosphericboundary layer (Neu,1995; Seigneur, 2001; Yerramilli et al., 2010).Such uncertainties remain one of the sources of inaccuracies inmodel simulations (Pleim, 2007a,b; Hu et al., 2010). While muchprogress has been made in simulating pollutant vertical transportin the convective boundary layer (CBL), additional attention is stillneeded to improve transport processes in the nighttime boundarylayer (NBL) and the residual layer (RL) (Salmond and McKendry,2005; Beare et al., 2006; Brown et al., 2008; Hong, 2010;Fernando and Weil, 2010).Previousstudies,relatedtothenighttimevariabilityofozone(O 3 ),havefocusedonthenocturnalboundarylayer(theairlayerextendingfrom the surface up to 200 e 400 m above the ground) and only fewstudies(e.g.,SalmondandMcKendry,2002)haveinvestigatedtheO 3 exchange between the NBL and the RL. Due to the limited mixing inthe NBL, O 3  mixing ratios typically reach minimum levels at theground due to dry deposition and nitric oxide (NO) titration. There-fore, strong and positive O 3  gradients develop throughout the NBL (Stutz et al., 2004; Geyer and Stutz, 2004; Brown et al., 2007; Pughet al., 2011). In contrast, O 3  mixing ratios are often considerednearly invariantwith timewithin theRL(e.g., Vecchi andValli,1999;Neu et al., 1994). However, recent studies (e.g., Salmond andMcKendry, 2002) suggest that this picture is incomplete as O 3  levelscan be highly variable in the RL. Ozone can be transported down tothe surface from the RL, thereby contributing to the maximum O 3 levels observed near the surface during daytime (Neu et al., 1994;Zhang and Rao, 1999; Yorks et al., 2009; Morris et al., 2010). Addi-tional studies are required to estimate the nighttime O 3  verticalvariability in the residual layer and the free troposphere so that theforecasting of air quality during the daytime can be improved. *  Corresponding author. Current address: Center for Analysis and Prediction of Storms, University of Oklahoma, Norman, OK 73072, USA. **  Corresponding author. E-mail addresses:  xhu@ou.edu (X.-M. Hu), jdfuentes@psu.edu (J.D. Fuentes). Contents lists available at SciVerse ScienceDirect Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv 1352-2310/$  e  see front matter    2011 Elsevier Ltd. All rights reserved.doi:10.1016/j.atmosenv.2011.09.054 Atmospheric Environment 46 (2012) 354 e 364  The goals of the present study are to investigate the temporaland spatial variability of O 3  in the nocturnal boundary layer and toestimate the O 3  vertical transport in the residual layer usinga regional air quality model. Results obtained from the presentinvestigation may reduce uncertainties in regional O 3  forecastingand thus improve the  fi delity of air quality models. 2. Methods The present investigation relied on the integration of airbornepollutant data within a regional air quality model.  2.1. In-situ measurements During summer 2010 a research  fi eld campaign was conductedat Howard University ’ s Atmospheric Research Site in Beltsville, MD(39.06  N, 76.88  W).The studysiteis located in suburban Marylandon 30 ha of land and is approximately 20 km northeast of the mainHowardUniversitycampus.Thesuburbansettingcontainsminimaldevelopment with not more than 5% of the land area occupied bybuilding structures, making it an ideal environment for studyinga broad range of air quality processes. Forests dominate the locallandscapewith both deciduous and coniferoustreespecies. A 31-mwalkup tower serves as the platform to mount meteorologicalinstruments such as platinum resistance thermometers, humidityprobes, anemometers, and radiometers. A sonic anemometer(model CSAT-3, Campbell Scienti fi c Inc., Logan, UT) measured high-frequency wind speed and temperature at 28 m above the surface.During summer 2010 several intensive observation periods werecarried out to obtain pro fi les of O 3 , temperature, humidity, andwinds using balloon-borne meteorological and O 3  sondes. Ozone-sondeswerebasedonthe electrochemicalconcentrationcell (ECC),whichareknowntohaveaccuracyinthetroposphereof5 e 10%.Thegoal of the ozonesonde intensives was to determine the temporalandspatialdistributionofO 3 duringregionalairpollutionepisodes.An intensive  fi eld campaign included at least four O 3  pro fi les in24 h. In addition, air pollutants were continuously measuredthroughout the summer 2010. Ambient levels of O 3  (model 49i,Thermo Environmental Instruments, Inc., Franklin, MA), sulfurdioxide (model 43i-TLE, Thermo Environmental Instruments, Inc.),carbon monoxide (model 48i, Thermo Environmental Instruments,Inc.), nitric oxide and nitrogen dioxide (model 42i, Thermo Envi-ronmental Instruments, Inc.) were recorded every minute. Airintakes were located at 5 m above the surface.  2.2. Three-dimensional simulations To study the O 3  vertical transport in the stable boundary layer,the Weather Research and Forecasting model with Chemistry(WRF/Chem)version3.2.1(Grelletal.,2005)wasusedinthisstudy.Three one-way nested domains (Fig. 1) were employed with gridspacing of 36, 12, and 4 km, respectively. Each domain had 44vertical layers extending from the surface to 100 hPa. The lowestmodel sigma levels were at 1.0, 0.996, 0.99, 0.98, 0.97, 0.96, 0.95,0.94, 0.93, 0.92, 0.91, 0.895, 0.88, 0.865, 0.85, 0.825, and 0.8 (cor-responding heights above ground were 0, 32, 81,162, 244, 327, 411,495, 580, 666, 752, 882, 1015, 1149, 1284, 1513, and 1747 m). Allmodel domains used the Dudhia shortwave radiation algorithm,the rapid radiative transfer model (RRTM) for longwave radiation,theWRFSingle-Moment6-class(WSM6)microphysicsscheme,theNoah land-surface scheme. The Grell e Dévényi ensemble convec-tiveparameterizationwasemployedindomains1and2,butturnedoff in domain 3. Three simulations were conducted with the YonseiUniversity (YSU), Mellor e Yamada e  Janjic (MYJ), and the asym-metric convective model version 2 (ACM2) planetary boundarylayer (PBL) schemes, respectively. The National Centers for Envi-ronmental Prediction (NCEP) Final (FNL) Global Forecast System(GFS) operational analyses were used for the initial and boundaryconditions of all meteorological variables.To determine gas-phase chemical reactions, the Regional Atmo-spheric Chemistry Mechanism (RACM) (Stockwell et al., 1997) wasused.Heterogeneousreactionswerenotincludedinthesimulations,which might be important for nighttime NO 3  and N 2 O 5  (Geyer et al.,2001). Such omission might introduce biases in the O 3  simulationduringnighttime.Anthropogenicemissionsofchemicalspeciescamefrom the national emission inventory (NEI) for 2005. Biogenic emis-sions were calculated using established algorithms (Guenther et al.,1994). Initial and boundary conditions for the chemical specieswere extracted from the output of the global model MOZART4(Emmons et al., 2010). The focus of the modeling study was onregionalhigh-O 3 episodesobservedinthemid-Atlanticregionof theUnited States during August 9 e 10, 2010. Model simulations wereconducted for the period ranging from 00:00 Universal Time Coor-dinated (UTC 1 ) August 7, 2010 to 04:00 UTC August 11, 2010.Previous investigations (Ha and Mahrt, 2001) illustrated thatvertical mixing in the residual layer was an important source of model uncertainties. Vertical resolution was shown to affect themodeled vertical mixing strength (Ha and Mahrt, 2001). Therefore,sensitivity WRF/Chem simulations with higher vertical resolutionwere conducted to investigate how much uncertainties could beattributed to coarse vertical model resolution. Sensitivity simula-tions with different treatments of vertical mixing (i.e., differentstabilityfunctionsandasymptoticlengths)intheresiduallayerwerealso conducted to identify the proper treatment. Such sensitivityinvestigationextendedthepreviousstudies(Huetal.,2010;Nielsen-Gammon et al., 2010) that focused on the sensitivity of the WRFmodel for simulating the daytime convective boundary layer. 3. Results  3.1. General observation During August 2010, average maximum O 3  mixing ratiosroutinelyexceeded60 partsper billion (ppbv) duringthe afternoon(approximately 15:00 Eastern Daylight Time (EDT)) (Fig. 2). Ozone Fig. 1.  Map of model domains. 1 UTC  ¼  Eastern Daylight Time (EDT)  þ  4 h.  X.-M. Hu et al. / Atmospheric Environment 46 (2012) 354 e  364  355  mixing ratios during 6:00 e 7:00 EDT remained below 10 ppbv on37% days. During 18% of the days, the O 3  mixing ratios during6:00 e 7:00EDTreachedabove25ppbv.Ingeneral,themeansurfaceO 3  mixing ratios decreased fromabout 22 to12 ppbv from00:00 to6:00 EDT with a depletion rate of 1.8 ppbv h  1 . There were severalcases, however, when this pattern of depletion did not occur.  3.2. Elevated surface ozone during the nighttime During August 18 and 22, 2010, O 3  mixing ratios exhibitedunusual patterns, particularly between 6:00 and 7:00 EDT when O 3 levels remained above 40 ppbv (Fig. 3). Turbulence data were ob-tained between August 18 and August 26, 2010. The standarddeviation of the vertical wind speed ( s w ) remained low( < 0.2 m s  1 ) during the nighttime and reached higher values( > 0.6 m s  1 ) during the daytime on clear days. However, on thenight of August 17 and 21, 2010, the  s w  showed relatively highvalues,  > 0.4 m s  1 (Fig. 3), indicating stronger turbulence andvertical mixing on those nights. Stronger vertical mixing trans-ported O 3 -rich air from the residual layer aloft to the surface,increasing the surface O 3  mixing ratio to higher levels compared to Fig. 2.  (Top) Observed mean time series of O 3  in Beltsville, Maryland during August2010 and (bottom) the time series of O 3  on each day. Fig. 3.  Observed (top to bottom) O 3 , standard deviation of the vertical wind speed ( s w ), NO x , temperature, relative humidity (RH), and wind speed at Beltsville, Maryland on (left toright) August 18 and 22, 2010.  X.-M. Hu et al. / Atmospheric Environment 46 (2012) 354 e  364 356  the other nights. However, higher O 3  mixing ratios at dawn did notcontributetodaytimeelevatedO 3 levelsonAugust18and 22,2010.This condition was associated with clouds that suppressed photo-chemistry. The precipitation during August 18 and 22, 2010 wasassociated with the passage of fronts. The sharp increase of O 3  of 37 ppbv and decrease of temperature around 3:00 EDT on August18,2010wereassociatedwithacoldfrontthatabruptlytransportedO 3 -rich air from aloft to the surface.On August 11, 2010 during the period from 1:00 to 4:00 EDTsurface O 3  mixing ratio increased by about 30 ppbv while NO x mixing ratio decreased by w 25 ppbv (Fig. 4). Thereafter, surface O 3 mixing ratio decreased to 24 ppbv before the daytime O 3  produc-tion began. During the night of August 10 and most of August 11,2010 the air mass came from the north. A cold front passed theresearchsite,travelingfromnorthtosouthonAugust12,2010.IfO 3 increases resulted from the advection of an upstream, pollutedplume then mixing ratios of other pollutants such as CO would behigher. However, CO level decreased as O 3  increased. The duration(several hours) of the nocturnal O 3  increase was similar to the “ leaky inversion ”  observed in New Hampshire (Talbot et al., 2005),which was caused by the vertical exchange of air between thesurface layer under the nocturnal thermal inversion and theresidual layer above it. Since the residual layer had higher O 3  andlower mixing ratios of other pollutants, the vertical exchange of trace gases allowed decreases in surface NO x  and CO and increasesin surface O 3 . During the nighttime, the calculatedvertical gradientof O 3  near the surface was 0.06 e 0.08 ppbv m  1 . Since surface O 3 increased by 30 ppbv, this implied that air was mixed downwardfromatleast400maboveground. Similareffectswereobservedonthe night of August 13 e 14, 2010 when a rapid event of the  “ leakyinversion ”  was associated with surface O 3  mixing ratio increasesranging from 9 to 19 ppbv around 3:00 EDT on August 14, 2010(Fig. 5).  3.3. Residual layer ozone: case study and simulations A regional episode of elevated O 3  occurred on August 9 e 10,2010, with maximum O 3  mixing ratios exceeding 80 ppbv. In thelower atmosphere ( < 1 km), O 3  mixing ratios showed positive andstrongverticalgradientsintheafternoon(Fig.6).Ozoneneedstimeto be formed through a complex chemistry from primarily pollut-ants, which are usuallyemitted at or near the ground, so O 3  mixingratios are higher away from the ground (e.g., Sillman,1999). On thenight of August 8 and 9, 2010 surface O 3  reached about 10 ppbv(Fig. 6). During the nighttime, the measured pro fi le of O 3  showeda strong vertical gradient below 1 km (Fig. 6), indicating theoccurrences of dry deposition and NO titration of O 3  near thesurface. The O 3  mixing ratio in the RL ( w 60 ppbv) resembled thatin the free troposphere, which was lower than that in thedaytime boundary layer ( w 80 ppbv at 13:54 EDT on August 10and w 100 ppbv at 13:46 EDT on August 9, 2010).Model simulations using the YSU, MYJ, and ACM2 PBL schemeswere conducted to investigate the air chemistry over the mid-Atlantic region of the United States. During the afternoon of August 9 and 10, 2010, simulated patterns of surface wind  fi elds byWRF/Chem with the YSU PBL scheme compared reasonably withobservations (Fig. 7). On the afternoon of August 9, easterly windsdominated in the eastern Virginia region whereas southerly/southwesterly winds prevailed in the northeastern United States.On the afternoon of August 10, westerly winds dominated in mostof Pennsylvania. Such wind patterns played an important role inregulating the air quality across the mid-Atlantic region.Simulated O 3  in the afternoon of August 9 and 10, 2010 werecompared with the measured O 3  at the AIRNOW sites (Fig. 8). Themodel results captured the main regions of elevated O 3  levels Fig. 4.  Observed (top to bottom) O 3 , NO x , wind vector, temperature at Beltsville,Maryland on August 11, 2010. Fig. 5.  Observed (top to bottom) O 3 , temperature, and speci fi c humidity (SH) atBeltsville, Maryland on August 14, 2010.  X.-M. Hu et al. / Atmospheric Environment 46 (2012) 354 e  364  357  during both days. On August 9, 2010 high O 3  mixing ratios( > 80 ppbv) occurred in the Washington DC (DC) to New York Citycorridor. On August 10, 2010 the region with high O 3  mixing ratioextended to the west (southeast of Pennsylvania). Surface O 3 mixing ratios in the afternoon were correlated with O 3  mixingratios in the upper layer (540 m above ground) in the morning(Fig. 8). On the morning of August 9, 2010 the plume of O 3  in theupper layers was along the urban corridor of the eastern UnitedStates. Thus, in the afternoon, surface O 3  mixing ratios in the urbancorridor were elevated. On the morning of August 10, 2010 theplume from the Midwest impacted the eastern United States. Thesurface O 3  mixing ratios along the DC-to-New York City corridoralso exceeded 80 ppbv during the afternoon of August 10, 2010. Incontrast, in the morning of August 8, 2010 the upper layers wererelatively clean in the eastern United States. The surface O 3  mixingratio was relatively low in the eastern United States. Therefore, theO 3  in the residual layer contributed to the surface O 3  duringdaytime. The residual layer also acted like a reservoir of otherpollutants, which likely contributed to surface O 3  formation oncetransported to the surface. Note that the downward transport of O 3 and other pollutants from the residual layer is only one of manyfactors that contribute to high surface O 3. On August 9 and August 10, 2010 we noticed two differentregional transport mechanisms. On August 9, 2010 the air masseswith accumulated pollutants (including O 3 ) were transportedalong the urban corridor region of the eastern United Stateswhereas on August 10, 2010 air masses were transported from themid-west to the eastern United States. Air parcel trajectory anal-yses (Figure not shown) con fi rmed such transport paths. Along thecorridor from Richmond, Va to New York City, there are heavyemissions of O 3  precursors. So if transport occurred along thiscorridor, O 3  might accumulate to a high level in the afternoon.Transport from the Midwest (northern Ohio River Valley) alsocontributed to the O 3  episodes in the eastern United States (Hainset al., 2008).To determine whether numerical models realistically representvertical transport of pollutants, model-simulated O 3  mixing ratiowas contrasted with the  fi eld observations made at Beltsville(Fig. 9). The model captured the daytime high O 3  mixing ratios, butoverestimated the nighttime O 3  (except on the night of August 10).Air quality models tend to overestimate ground-level O 3  during Fig. 6.  Observed vertical pro fi les of O 3  at Beltsville, Maryland on August 9 e 10, 2010. Fig. 7.  Simulated (top) and observed (bottom) surface wind  fi eld at 20:00 UTC on (from Left to Right) August 9, and 10, 2011. The observation on the bottom row is from the MADISdataset.  X.-M. Hu et al. / Atmospheric Environment 46 (2012) 354 e  364 358
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks