Model Evaluation of the Impact of Thermally Induced Valley ... · Model Evaluation of the Impact of Thermally Induced Valley Circulations in the ... study was to use these modeling

JAPCA 38:163-170 (1988)

Model Evaluation of the Impact of Thermally InducedValley Circulations in the Lake Powell Area on Long-Range Pollutant Transport

M. Segal, C.-H. Yu and R. A. PielkeDepartment of Atmospheric ScienceColorado State UniversityFort Collins, Colorado

A numerical mesoscale model was used to simulate meteorological fields in the Lake Powellarea during the summer, providing an input to a Lagrangian-type transport/dispersion modelevaluation. The main objective of the study was to use these modeling tools in order toevaluate the local effect of thermally-induced circulations on large-scale transport of pollutedair masses into that area. Results indicated a substantial modification of the transportcharacteristics due to the local terrain-forced circulations. Most noticeable are: (i) trappingof pollutants within the Lake Powell valley, (ii) upward and downward venting of pollutants inconvergence/divergence zones associated with these circulations, and (iii) slowing of cross-valley transport as compared to equivalent situations involved with flat terrain.

Observations in the Lake Powell areahave indicated that a significant visi-bility impairment often exists in thisarea located in southern Utah andnorthern Arizona (a schematic illustra-tion of the area is provided in Figure 1).Since several National Park areas with-in this region are mandated by federallegislaton as Class I areas under thePrevention of Significant Deteriora-tion (PSD) of the federal air qualityprogram, considerable research efforthas been devoted in recent years to un-derstand the atmospheric processes af-fecting visibility in that area. Studiesevaluating the potential contributionof local sources in the area to the visi-bility impairment is reported, for ex-ample, in Bluemental et al.,1 and Yuand Pielke.2 Most of the focus in recentyears, however, has been directed tovisibility oriented studies includingevaluations of the impact of long-rangetransport of polluted air masses intothe area during the summer. Duringthe summer, a shallow thermal lowdominates the lower atmosphere of thesouthwest United States (e.g., Tangand Rieter3). Studies reported, for ex-ample, by Pitchford,4 Marias,5 Fluci-hini et al.,6 Ashbaugh et al.1 suggestthat southwest synoptic flow associat-ed with that meteorological systemmay lead to the long-range transport ofpolluted air masses from southern Cali-fornia and southern Arizona into the

Lake Powell area. Also, these studiessuggested situations involved withsoutheast synoptic flow advecting pol-luted air into the Lake Powell areafrom New Mexico and Texas. This sec-ond common transport situation is as-sociated with flow around the west sideof a subtropical ridge which is situatedto the east of the thermal heat low.

All of the transport studies in thisregion, however, have been based onobserved synoptic wind data. The tem-poral and spatial resolution of suchdata is typically insufficient to evalu-ate the impact of mesoscale generatedcirculations (i.e., mountain/valleyflows with typical horizontal scales ofless than several hundred kilometers)on the large-scale transport. Therefore,in those studies, the transport modifi-cation of a pollutant mass by mesoscalecirculations were not considered. Alsoin those studies, turbulence processeswere not considered. Hence, whilethese synoptic transport evaluationsprovide bulk information concerningair movement over mesoscale domains,additional evaluations are requiredwhen thermally-induced circulationsexist along the path of the pollution(see, for example, Pielke et al.,8 for adetailed evaluation and discussion ofthis aspect).

The Lake Powell area, which is locatedin the Colorado River basin, is affectedby valley-induced circulations. Our

study is designed specifically to providean evaluation of modifications to long-range transport by mesoscale effects inthat area during the summer using themodeling tools outlined in the next sec-tion. Generally, the impact of mesoscalethermally-induced circulations on long-range transport of a polluted air mass isunaddressed in the literature. Therefore,although the present study provides aspecific evaluation for the Lake Powellarea, it also provides some insight as tothe related impact of valley circulationsin the general case.

Modeling Aspects

A numerical mesoscale meteorologi-cal model was applied in a two-dimen-sional domain along a northwest-southeast cross section in the LakePowell area (crossing through Page,Arizona; see Figure 1 for an illustrationof the cross section location) as adopt-ed in Yu and Pielke.2 In this region, theterrain acquires a general two-dimen-sional symmetry (i.e., uniform terrainalong the lake direction), thereby justi-fying using a two-dimensional modelversion for the preliminary study re-ported in this paper. The model com-puted meteorological fields are used asinput for a Lagrangian particlesscheme to predict their transport/dis-persion.

Numerical Mesoscale MeteorologicalModeling

The formulation of the numericalmesoscale model used in the presentstudy is given in detail in Pielke,9

Mahrer and Pielke,10 and McNider andPielke,11 and also is summarized inPielke.12 The model was validated suc-cessfully in studies of mountain/valleythermally-induced flows similar to thesituation considered in the presentstudy (e.g., Segal et al.,13 McNider andPielke,14 Abbs,15 among others).

Copyright 1988—APCA

February 1988 Volume 38, No. 2 163

Glen CanyonNo tonalRecreationArea

Conyon ondsNotionalPork

Figure 1. Illustration of the Lake Powell area. The simulated cross-section is indicated by the line ab.

The terrain cross section used for thesimulations, which consists of elevationvariations larger than 1000 m, is illus-trated in Figure 2. The meteorologicalconditions simulated reflect July con-ditions in that area. Table I providesthe general information related to themeteorological model simulations. Twobackground flow situations were simu-lated: 1) southeast synoptic flow of 2.5ms"1 (i.e., crossing the valley perpen-dicularly) as a background flow; and 2)southwest synoptic flow of 2.5 ms"1

(i.e., along-valley direction).

Transport/Dispersion Modeling

Using the wind and turbulence fieldscomputed by the meteorological mod-el, a Lagrangian approach is applied toevaluate the transport and dispersionof pollutants. This dispersion andtransport considers both the synopticand mesoscale (i.e., valley circulation)flows. The detailed formulation is giv-en in McNider16 and described also inPielke et al.,17 and McNider et a/.18;thus, it is outlined in this paper onlybriefly. The model consists of trackinga release of a large number of particlesrepresenting a pollutant air mass ad-vected and diffused, in the general caseof a three-dimensional domain, usingthe formulation:

Xi(t + 8t) = Xi(t) + [Ui(t) + u'iit)] St

i = 1,2,3 (1)where x,-(t) is the old x,y, and z positionof a particle and x,-(£ + 8t) is its positionfollowing time interval 8t (in thepresent study 8t = 20 sec); ui are the u,v, and w velocities computed by the

meteorological model and u'i are turbu-lence velocity fluctuations parameter-ized statistically based on the meteoro-logical model boundary layer predic-tions.

The turbulence velocities are provid-ed by the relation:

u'iit) = u'iit - 8t)R(8t) + u'\

i = 1,2,3 (2)where R(8t) is the Lagrangian autocor-relation function at time lag 8t and u"iis a random component with a Gauss-ian distribution of zero mean and vari-ance defined by the local turbulencevariance. A validation of the scheme isreported in McNider16 and McNider etal.18

A set of transport simulations as out-lined in Table II were carried out. Inthe cross-valley (i.e., southeast) synop-tic flow situations, a volume of pollut-ant, represented by particles, was lo-cated initially at the southeasternboundary of the simulated domain. Its

initial height was 600 m and its hori-zontal extent 15 km. The volume wasdifferentiated graphically into threelayers indicated by a (bottom layer); /?(the medium layer); and 7 (the upperlayer) as illustrated in Figure 4a.Transport simulations for this situa-tion are listed in Table II (Cases 1-4).They consist of a release of the volumeat various hours of the day with andwithout turbulent effects. This cross-valley simulation provides an illustra-tion for situations involved with pollut-ant air transport to the Lake Powellarea from New Mexico and Texas.

The second situation reflects asouthwest synoptic flow of 2.5 ms"1

(i.e., along the valley direction). Parti-cles were released in this case from themiddle of the valley as illustrated inFigure 9a. The particles volume had aninitial horizontal extent of 30 km anddepth of 600 m. The simulated trans-port cases for this situation are listed asCases 5-7 in Table II. These cases pro-vide an illustration for situations in-volved with pollutant air transport intothe Lake Powell area from the direc-tion of southern California.

The pollutant mass as representedby the particles was assumed in bothsimulations to be initially horizontallyuniform in order to represent long-range transport into the region withoutmesoscale or turbulent dispersion up-stream. This was done so that the influ-ence of the mesoscale and turbulentflows in the Lake Powell area could bebest visualized. In reality in both simu-lated cases, it is expected that interven-ing complex terrain between the sourceand the Lake Powell area would sub-stantially influence the dispersion andtransport in a similar manner as occurswithin the Lake Powell area itself.

Simulation Results

Cross-Valley Synoptic Flow—Meteorological Fields

Representative meteorological fieldsin the simulated cross section for the

Table I. Input parameters for the meteorological model.

Simulated domain top—9600 mLowest terrain height within the simulated domain—~1300 m ASLDomain horizontal extent—150 kmModel horizontal grid interval—5 kmIntegration time step—60 secModel vertical number of levels—28Simulation start hour—2100 LSTSolar declination for the simulated day—July 10Initial boundary layer depth—250 mInitial potential temperature lapse rate dd/dz:

a) 0 K/1000 m for z < 250 mb) 2 K/1000 m for 250 < z < 2450 mc) 4 K/1000 m for 2450 < z < 9600 m

Latitude—37°

164 JAPCA

0500LST .MOO. .1ST

0.2

-0.2

Figure 2. Representative meteorological fields for the cross-valley transport cases (southeasterly synop-tic flow) for the simulated vertical cross section, for nighttime (0500 LST) and daytime (1400 LST); (a) and(b): u component of the wind (cross-valley) is given in ms~1; dashed contours as well as shaded sectionsindicate negative values of u(southeasterly flow); (c) and (d): vertical velocities (cm s~1); dashed contours aswell as shaded sections indicate negative values of vertical velocities (subsidence); (e) and (f): potentialtemperature (K).

nighttime (0500 LST) and the daytime(at 1400 LST) are given in Figure 2.The geostrophic flow is southeasterlywith a wind speed of 2.5 ms"1, as indi-cated in Figure 2a.

At the end of the nocturnal period,drainage flow is indicated to be shallowwith speeds of several ms"1 (Figure 2a).The synoptic flow is perturbed onlyslightly at the valley top due to terraindynamic effects and interaction withthe valley thermally-induced circula-tion. A shallow layer with downwardvertical velocities is modeled along theslopes as shown in Figure 2c (note thatdashed contours as well as shaded sec-tions in all the presented figures reflectnegative vertical velocity values). Thepotential temperature field (Figure 2e)consists of a shallow, relatively coolsurface-based inversion layer along the

slopes which deepens toward the valleybottom where the turbulence is re-duced.

The daytime meteorological featuresare presented for 1400 LST when thethermally-induced valley circulation is

around its peak. Upslope thermally-in-duced flows are somewhat less intensealong the southeastern slope than thosealong the western slope (Figure 2b),which is attributed to a steeper westernslope as well as the coupling of the up-slope flows with an opposing synopticflow along the southeastern slope. A sig-nificant feature in the flow field is theintensification of the flow aloft over thesoutheastern slope due to the returnflow centered at about 3 km in Figure 2bwhich was caused in response to the lowlevel upslope flows (the flow intensifica-tion aloft to a peak wind speed ofaround 7 ms""1 is caused by a superposi-tion of the return circulation and thesoutheast synoptic flow at that alti-tude). The horizontal convergence dueto the coupling between the synopticand thermally-forced flows along thesoutheastern slope is noticeable in thevertical velocity features (Figure 2d).Sinking of air over the middle of thevalley involved with the upslope circu-lations is also pronounced. The poten-tial temperature contours (Figure 2f)indicate the daytime depth of theboundary layer (mixing depth). An up-ward orientation of the contours is asso-ciated with the near-neutral thermalstratification which is typidal for thatlayer. The boundary layer depth, as in-dicated by the temperature structure, isaround 2-2.5 km for the presented hour.Finally, it is worth noting that the de-tails of the near-ground wind and tem-perature structure are not evident in thefigure because of the vertical resolutionused in the plots.

Cross-Valley Flow—Pollutant Transport

Four illustrations of transport fea-tures are provided for this situation.Three releases in the morning hoursand one in the evening, including a casein which turbulence effects were ig-nored, are shown.

Figure 3 provides the dispersion fea-ture involved with a 0900 LST release(Case 1). The volume of the particles islocated initially at the eastern bound-ary (see Figure 4a) of the domain with adepth of 600 m and a horizontal extentof three grid intervals (15 km). The vol-

Table II. List of transport case studies, u indicates horizontal advection is included;w indicates vertical advection is included; T indicates turbulence is included.

Case Release hour (LST) Release location Forcing Figure No.

090005001100 .2100

050005002100 .

Cross-valley transportSoutheast boundary,Southeast boundarySoutheast boundarySoutheast boundary

Along-valley transportMiddle of valleyMiddle of valleyMiddle of valley

u, wu,w,Tu,w,Tu,w,T

u, wu,w,Tu,w,T

910

February 1988 Volume 38, No. 2 165

1200 1500 LST

Figure 3. Features of pollutant particles dispersion at several selected hours for a release at 0900 LST.Southeasterly synoptic flow; only advection is considered (Case 1).

ume is differentiated graphically intothree layers, enabling an evaluation ofthe mesoscale impact on each layerseparately. The particles are advectedby the u (cross-valley wind component)and w (wind vertical velocity) with theturbulent contributions ignored (i.e.,ur = w' = o). The motivation for this

simulation is the evaluation of the im-portance of the advection processes ascompared to the turbulence processesin the studied cases. This objective canbe accomplished by a comparison ofthe dispersion features presented forCase 1 with those obtained in the casesin which turbulence is considered. By1200 LST the simulated pollutant massis vented upward intensely in the con-vergence zone. By 1500 LST upwardventing within the convergence zone isinvolved with the a layer particles,while the /3 and 7 layer particles, whichhad been advected to the middle of thevalley are sinking in the downward ver-tical velocities there. By 1800 LST thea layer particles also become trappedin the sinking motion, with the (3 and 7particles located at the low elevationsof the valley as a result of the down-ward venting. Affected by the finalstage of the daytime upslope circula-tion, the sinking particles were advect-ed westward, while the /3 and 7 parti-cles were advected eastward and vent-ed up the western edge of theconvergence zone. As the nocturnal cir-culation is established following sun-set, the particles effected by the rever-sal in the valley flow and their interac-tion with the synoptic flow, resulttoward the end of the night in the parti-cle distribution shown for 0300 LST inFigure 3d.

Case 2 is equivalent to Case 1 except

for the incorporation of turbulence,and particles were released at 0500LST (Figure 4). By 1100 LST the parti-cles were advected somewhat downslope. The strong vertical mixing with-in the boundary layer is pronounced;however, the vertical velocity impactillustrated in Case 1 is still noticeablein the particle distributions. At 1700LST the particles become mixed withinthe deeper boundary layer along both

slopes; however, the southeastern por-tion is somewhat deeper as this regionis affected by upward venting withinthe convergence zone. The synopticflow over the shallow drainage flowduring the night causes the pollutantmass to reduce its depth and to be ad-vected gradually out of the domain asillustrated in Figure 4d.

Case 3 is the same as Case 2 exceptthe release of the particles was at 1100LST; thus, the impact of the daytimevalley circulation in this case would ap-pear immediately following the release.By 1400 (Figure 5a) the particles wereadvected somewhat in the downslopedirection. The mass is noticeably af-fected by the upward and downwardvelocities at the slope convergence lo-cation and the mid-valley divergencezone following the patterns illustratedin Figure 3. It is worth pointing out thatthe sinking motion above the valleycauses the particles aloft to be venteddownward, otherwise, particles maynot have been trapped within the valley.The dominance of the convergence/di-vergence effects leads by 1700 LST to adivision of the mass into two branches(Figure 5b). Similar nocturnal featuresshown in Case 2 are also obtained forthis case, while the mass is advectedgradually toward the northwest andout of the domain (Figure 5c,d).

Case 4 is the same as Cases 2 and 3;however, particles are released at 2100LST (Figure 6). By 0600 LST the pre-sented particle distribution had beenestablished mostly due to horizontal

3 . 8 0500 LST

3.8

3.4

3.0

2.6

2.2

1.8

1 .4

1 .0

0.6

0.2

-0 .2

1700 LST

<

>]

' •• • * . " ' ] ' ' . " ' .

/ * ' • ' , ' ' ' •

. •" ' • > t• ' • • ' • • • . "

• • , - • ' ' ' V '•. ;•. ' .•.

3.8

3.4

3.0

2.6

2.2

1 .8

1 .4

1 .0

0.6

0.2

-0 .2

0200 LSJ

Figure 4. Features of pollutant particles dispersion at several selected hours for a release at 0500 LST.Southeasterly synoptic flow; advection and turbulence are considered (Case 2).

166 JAPCA

ffOO LSJ

0.2

Figure 5. Features of pollutant particles dispersion at several selected hours for a release at 1100 LST.Southeasterly synoptic flow; advection and turbulence are considered (Case 3).

and vertical transport (as diffusion bythe nocturnal turbulence is small).Venting upwards of the a particles dueto flow convergence at the bottom ofthe valley occurred, while the /? and ylayers largely conserve their initial con-figuration by that hour. At 0900 LST(Figure 6b) vertical mixing mostly af-fects the a and /3 layers, which stillmaintain a noticeable differentiationfrom the y layer particles. By 1200 LSTvertical mixing affects all layers; how-ever, the y particles, which during theearlier hours were advected faster, areat the front of the advected pollutedmass, toward the northwest boundary.The a and /? particles being trapped inthe valley are affected by thermally-induced winds from both slopes at thishour. Significant advection of the pol-lutant mass during the next hours isevident by the pollutant distribution at2100 LST. By that hour the y particleshad been advected out of the domain,while some of the a and 0 particles werestill trapped in the valley.

Along-Valley Flow—Meteorological Fields

A geostrophic wind of 2.5 ms"1 alongthe valley direction (i.e., southwestsynoptic flow) is considered in this sec-tion. A small contribution of the synop-tic flow to a cross-valley component oc-curred in the boundary layer (i.e., dueto Ekman wind turning); however, thecross-slope wind component is primari-

ly induced by thermal effects. The sim-ulation began around sunset. Since inthis case the synoptic flow is along val-ley, unlike the previous simulation, thenighttime and daytime thermally in-duced cross valley winds are anticipat-ed to be altered only slightly by thesynoptic along-valley flow.

The meteorological fields at 0500LST (providing an illustration of the

0600 LST

meteorological fields at night), are il-lustrated in Figure 7. Shallow drainageflow layers are evident from Figures7a,c. Similar to the previous case study(i.e., the cross-valley simulation), ashallow nocturnal surface inversionalso occurs (Figure 7e).

By 1400 LST a classical thermally-induced circulation had developedalong both slopes as evident by the ucomponent wind distribution (Figure7b). The vertical velocity pattern (Fig-ure 7d) illustrates the vertical branchof the mesoscale circulation. The char-acteristics of the potential temperature(Figure 7f) are similar to those report-ed for the cross-valley synoptic flowcase study.

Along-Valley Flow—Pollutant Transport

In this situation a volume of particlesis initially located in the lower eleva-tion of the valley. It might reflect aportion of a larger volume covering thevalley. Since the synoptic flow is south-westerly (i.e., along the valley), its im-pact on cross-valley transport is minoras pointed out previously. Therefore,only the development of cross-valleythermal flows, in addition to turbu-lence, are expected to cause an alter-ation in the spatial distribution of theparticles in the cross-valley direction.

In the along-valley synoptic flow, theinitial location of the pollutant mass isindicated in Figure 9a. Figure 8 pre-sents the development involved withCase 5 where negligible vertical veloci-ty is simulated with a 0500 LST release.At 0800 LST the a and 0 layers weremost noticeably affected as the upslopeflows began. By 1400 LST most of the a

0900 LST

Figure 6. Features of pollutant particles dispersion at several selected hours for a release at 2100 LST.Southeasterly synoptic flow; advection and turbulence are considered (Case 4).

February 1988 Volume 38, No. 2 167

0500 LST

Figure 7. The same as Figure 1 except for the along-valley synoptic transport case (southwesterlysynoptic flow).

particles were advected toward theboundaries of the domain, but were de-celerating in the east, at the conver-gence zone. The a layer particles wereaffected by a less intense upslope andtheir advection toward the boundarieswas slower.

Figure 9 presents a time sequence forCase 6 of the alteration of mass (initiallyreleased at 0500 LST) including diffu-sion by turbulence processes whichshould cause alterations in the basicfeatures as obtained in the previouscase. Following 9 hours from the release(Figure 9b), the development of upslopeflows and a well mixed layer were evi-dent by the generation of two separateand well mixed pollutant volumes alongthe slopes. The a, /3 and y layer particleswere vertically well mixed up to an ele-vation of about 2.5 km above the valleybottom. Additional impact of the day-time upslope on the particle distribu-tion was seen at 1700 LST. It consists ofthe generation of a relatively clear at-

mosphere at the valley bottom, as wellas the advection of some of the particlesout of the domain, mostly from thewestern boundary where the upslopeflows were stronger. In the nocturnalperiod, relatively strong but shallowdrainage flow was capped by a deep butlight reversal flow aloft. This led to a

3.0

2.6

2.2

1 .8

J i . 4K 1 . 0

0.6

0.2

-O.2

0800

trapping of the pollutant mass over thevalley 24 hours following the release(0500 LST second day).

For Case 7, where the release was at2100 LST, the particular redistributionfollowing 12 hours (0900 LST) is illus-trated in Figure 10a. By that time thedevelopment of boundary layer verticalmixing and upslope flows were notice-able in the redistribution of the parti-cles. Following an additional 12 hours(2100 LST), the vertically well mixedparticles over the valley were similar tothose obtained in Case 6 by that hour(Figure 9).

Summary and Conclusions

Two cases of the influence of meso-scale circulation on long range trans-port that are likely to be involved withthe Lake Powell area during the sum-mer were evaluated using a modelingtool. Pollutant mass was representedby particles which were differentiatedgraphically into three layers (a, /?, 7)enabling a more detailed evaluation asto the transport features in each layer.The simulations provide insight as tothe impact of valley thermal circula-tions and dynamic effects in that areaon long-range transport. The resultsalso illustrate expected influences onlong-range transport of thermally- anddynamically-forced mesoscale flows as-sociated with complex terrain in gener-al. Additional work, of course, is need-ed to quantify the influence of suchcharacteristics, such as the aspect ratioof the valley, synoptic flow intensityand atmospheric stability (or theFroude number) on the modification ofsynoptic long-range transport. Three-dimensional simulations of the LakePowell area, and other geographic re-gions which incorporate both meso-scale and synoptic influences, and tur-bulence effects are, of course, needed.Nonetheless, the two-dimensional sim-ulations presented here demonstratethe expected major importance of me-soscale and turbulence effects on long-range transport across complex terrain.

The main conclusions of the present

1400 LST

Figure 8. Features of pollutant particle dispersion at two selected hours for a release atSouthwesterly synoptic flow; only advection is considered (Case 5).

168 JAPCA

3 . 8

3 .4

3 . 0

2 . 6

2.2

l i .85 1.4

1 .0

0.6

0.2

-0.2

3.8

3.4

3.0

2.6

2.2

J 1 .8

* 1 . 4

r.o0.6

0.2

-0.2

0500LST

1700LSJ.

c

•

' • • ' • " ' . -

' • • • I ' •

•

3.8

3.4

3.0

2.6

2.2

1 .8

1 .4

1 .0

0.6

0.2

-0.2

3.8

3.4

3.0

2.6

2.2

1 .8

1 .4

1 .0

0.6

0.2

-0.2

LST

0500 LSJ

Figure 9. Features of pollutant particle dispersion at several selected hours for a release at 0500 LST.Southwesterly synoptic flow; advection and turbulence are considered (Case 6).

study relevant to long-range transportto the area are:• Pollutants advected into the Lake

Powell area by summer day south-east synoptic flow (i.e., a cross-val-ley large-scale wind) for the day-time releases were trapped for thesimulated period (~24 h) in the val-ley.

• Nighttime releases involved withcross-valley synoptic flow, althoughtrapped in the valley, were trans-ported out of the valley eventually(particularly the elevated pollutedlayers).

• Vertical and horizontal dilution ofthe pollution due to the inducedthermal circulations distinguish itfrom the common and simple caseof advection along flat terrain. Con-sequently, as outlined previously,acceleration and deceleration oftheir center of mass result. With asynoptic flow of 2.5 ms"1 and a do-main of horizontal extent of 150 kmas in the present study, with flatterrain and no mesoscale effects,the pollutants mass would havecrossed the entire domain withinabout 16 hours. This rate of pro-gress is faster than found for thevarious cross-valley wind cases (i.e.,southeasterly wind transport).

• In the cases involved with pollutantmass transport within the valleyalong its orientation, the transportwas affected by cross-valley ther-

mally induced flows, in addition tothe turbulence during the daytimehours. The results show that duringthe daylight hours there is a split-ting of the mass into two branchesalong the slopes, in addition to avertically well mixed layer. This re-sulted in a significant dilution ofthe particles over the domain; how-ever, generally, the particles per-sisted over the valley area for atleast 24 hours.The simulations were carried outunder the assumption of two-di-mensional symmetry of the terrainin the studied area (i.e., a uniformterrain along the valley direction),which is an idealization. Introduc-

3.8

3.4

3.0

2.6

2.2

Ji.B1 .4

1.0

0.6

0.2

-O.2

0900 LST

N

ing three-dimensional terrain con-siderations should lead to variousdistortions in the features obtainedwith the two-dimensional simula-tion. However, general conclusionsof the transport/dispersion charac-teristics in this study should not bechanged appreciably.Finally, based on the simulatedcases, the study indicated the im-portance of vertical velocities in-volved with convergence/diver-gence zones on the vertical ventingof pollutants, eventually causingsignificant alterations in the parti-cle distribution from that whichwould occur without this ventingmechanism.

Acknowledgment

This work was supported by the Na-tional Park Service under contractNA81RAH0001, Amendment 17, Item15, and computer calculations wereperformed on the NCAR CRAY-1Computer. NCAR is sponsored by theNSF. W. Malm, D. Henderson and R.Stocker are thanked for their valuablecomments made during the completionof this work. The typing was ably han-dled by Sandra Wittier.

References

1. D. L. Bluemental, L. W. Richards, E. S.Marias, R. W. Bergstrom, W. E. Wilson,P. S. Bhardwaja, "Effects of a coal-firedpower plant and other sources on south-western visibility (interim summary ofEPA's project VISTTA)" Atmos. Envi-ron. 15:1955 (1981).

2. C.-H. Yu, R. A. Pielke, "Mesoscale airquality under stagnant synoptic coldseason conditions in Lake Powell area,"Atmos. Environ, (in press).

3. M. Tang, E. R. Reiter, "Plateau effectson diurnal circulation pat terns ,"Monthly Weather Rev. 112:638 (1984).

4. A. Pitchford, M. Pitchford, W. C. Malm,R. Flocchini, T. Cahill, "Regional analy-sis of factors affecting visual air quali-ty," Atmos. Environ. 15: 2043 (1981).

3.8

3.4

3.0

2.6

2.2

1 .8

1 .4

1 .0

0.6

0.2

-0.2

2100 LST

Figure 10. Features of pollutant particle dispersion at several selected hours for a release at 2100 LST.Southwesterly synoptic flow; advection and turbulence are considered (Case 7).

February 1988 Volume 38, No. 2 169

5. E. S. Marias, J. 0. Zwicker, W. H.White, "Regional haze case studies inthe southwestern U.S.—II. Source con-tributions," Amos. Environ. 15: 1987(1981).

6. R. G. Flucihini, T. A. Cahill, M. L.Pitchford, R. A. Eldred, P. J. Peeney, L.L. Ashbaugh, "Characterization of par-ticles in the arid west," Atmos. Enviro.15:2017(1981).

7. L. L. Ashbaugh, W. C. Malm, W. Z. Sa-deh, "A residence time probability anal-ysis of sulfur concentrations at GrandCanyon National Park," Atmos. Envi-ron. 19:1263 (1984).

8. R. A. Pielke, M. Segal, R. W. Arritt, M.Moran, "Mesoscale Influences on LongRange Pollutant Transport," presentedat CIRA Workshop on Acid Depositionin Colorado, CIRA, Colorado State Uni-versity, August 1986.

9. R. A. Pielke, "A three dimensional nu-merical model of sea breezes over southFlorida," Monthly Weather Rev. 102:115(1974).

10. Y. Mahrer, R. A. Pielke, "A numericalstudy of the air flow over irregular ter-rain," Contrib. Atmos. Phys. 50: 98(1977).

11. R. T. McNider, R. A. Pielke, "Diurnalboundary layer development over slop-

ing terrain," J. Atmos. Sci. 38: 2198(1981).

12. R. A. Pielke, Mesoscale Meteorology Mod-eling, Academic Press, Orlando, FL, 1984.

13. M. Segal, Y. Mahrer, R. A. Pielke, "Ap-plication of a numerical mesoscale mod-el for the evaluation of seasonal persis-tent regional climatological patterns,"J. Appl. Meteorol. 21:1754 (1982).

14. R. T. McNider, R. A. Pielke, "Numeri-cal simulation of slope and mountainflows," J. Clim. Appl. Meteorol. 23:1441 (1984).

15. D. J. Abbs, "Sea breeze interactionalong concave coastline in southernAustralia: Observation and numericalmodel study," Monthly Weather Rev.114:831 (1986).

16. R. T. McNider, "Investigation of theImpact of Topographic Circulations onthe Transport and Dispersion of AirPollutants," Ph.D. dissertation, Univer-sity of Virginia, Charlottesville, VA,1981.

17. R. A. Pielke, R. T. McNider, M. Segal,Y. Mahrer, "The use of a mesoscale nu-merical model for evaluations of pollut-ants transport and diffusion in coastalregions and over irregular terrain,"Bull. Am. Meteorol. Soc. 64: 243 (1983).

18. R. T. McNider, M. D. Moran, R. A.

Pielke, "Influence of diurnal and iner-tial boundary-layer oscillations on long-range dispersion," Atmos. Environ.(submitted 1987).

M. Segal is a research scientist andR. A. Pielke is a professor in the De-partment of Atmospheric Science,Colorado State University, Fort Col-lins, CO 80523. C.-H. Yu, deceased,was a graduate student at the sameinstitution. This paper was submit-ted for peer review November 3,1986; the revised manuscript was re-ceived October 26,1987.

170 JAPCA