Sea-Breeze Observations and Modelling: a Revrew

Aust. Met. Mas. 4l (lgg2) 7-lg Sea-breezeobservations and modelling: a revrew

DeborahJ. Abbs and William L. Physick CSIRO Division of Atmospheric Research,Australia (Manuscript receivedApril 1992;revised JuIy 1992) We presentan overviewofsea-breeze characteristics and behaviour and then discuss somemore important aspectsin greaterdetail. Theseinclude the fine-scalestructure of the sea-breezein the headregion, its behaviourin complexterraino and its ability to penetratehundreds of kilometres inland. Theseproperties are then consideredin a discussionof the role of sea-breezesin the dispersionof pollutants.Finally we examinethe historical milestonesof sea-breezemodelling, as it progressedfrom the early linear analytic modelsto the complexnon-hydrostatic numerical models of the nineties.

Introduction The sea-breezehas been studied extensively in the balances.Because the seais a fluid in continual past,especially so sincethe end of World War II. motion, any heatingor cooling is distributed to a Neumann (1973)refers to the Greekphilosophers considerabledepth and so the rise or fall ofwater Aristotle, Plutarch and Theophrastus,and thelr surfacetemperature is only slight.This results,for ideason the seaand land-breeze.They considered example,in a very small diurnal variation in sur- the sea-breezeto be a reflection from obstacles facetemperature. Conversely, the land hasa very such as islands and coastalhills; onshorewinds smallthermal conductivityand soheats and cools from the open ocean were not regardedas sea- rapidly, resultingin a marked diurnal variation. breezesby them asthere are no objectsin the open The resulting land-water temperature diferences sea from which the land-breezecould rebound. and their diurnal reversalproduce corresponding Jehn (1973)refers to Dampier (1697, 1701and land-water pressuredifferences in the atmosphere 1705) who wrote at length on winds, including which result in a system of breezes across the land and sea-breezes.In his writingshe gavea cor- coast.Onshore flow during the day is known as a rect description of the sea-breezewhich was ig- sea-breezewhile offshore flow at night is referred nored by the theoreticiansfor nearly 200 years. to as a land-breeze. Before the more detailed aspectsof the sea- During the day the heatingof the nearJandsur- breeze can be investigated,it is necessaryto faceis distributedupwards by mixing. The rise in describethe basicprocesses involved in its devel- air temperatureresults in expansion,so that aloft, opment.These processes are discussedin most of individual particlesover the land are lifted rela- the elementarymeteorology texts, the following tive to those over the sea. Level for level, the being basedon the accountpresented by Clarke pressurealoft, but not at the surface,becomes (l 95s). higher over the land than over the sea,and this The sea and land-breezecirculations encoun- results in a compensatingoutflow of air from tered near shorelinesare due to the contrasting near-coastalareas over the land to thoseover the thermal responsesof the land and water surfaces sea. As a result the surface pressure falls over a becauseof their different propertiesand energy coastalstrip ofland and risesover a coastalstrip of sea.The pressuregradient thus setup resultsin lowJevel onshore flow with a return flow aloft. Correspondingauthor address:D. Abbs, CSIRO Division of a AtmosphericResearch, Private Bag No. l, Mordialloc,Vic These factors operate simultaneously,not con- 3195.Australia. secutively. .J Australian Meteorological Magazine 4l December 1992

The sea-breezecirculation cell begins near the matic regimes. In mid-latitudes the horizontal coast in the morning and expandsboth landward extent of the sea-breezerarely exceeds 40 to and seaward.If there is no gradient wind the ex- 50 km, while in tropical and subtropical areas it pansion is greater over the water due to a smaller reaches100 to 150km on both the landward and frictional force at the seasurface than at the land seawardsides ofthe coast,resulting in a total ex- surface.The effect ofnon-zero synoptic flow is to tent of 200 to 300 km. Someof the most extensive either accelerateor delay the development of the studies of deeply penetrating sea-breezeshave system. In some cases strong oflshore gradient beencarried out in Australiaby Clarke(1955) and winds result in the sea-breezedeveloping over the Reid (1957).Clarke has recorded cases ofthe sea- seawith little or no inland progressionduring the breeze reaching Kalgoodie, 300 km from the day. In caseswhere the gradient wind is very light coast.Clarke (1983a)also reports results which or has a moderate onshorecomponent, the sea- indicate that the sea-breezepenetrates 500 km in- breezefront tends to form l0 to 30 km inland in land in northern Australia and has also been the afternoon. recorded 320 km seawardofBorneo. The depth ofthe seaair, usuallydefined by the In the following sections,we examinesome as- level of the zero onshorewind component, is often pectsofthe sea-breezein greaterdetail, including lessthan half the vertical extent of the sea-breeze its deep inland penetration and its role in deter- system.At any given locationthe depth of the sea mining air quality in coastalregions. We conclude air varies with the time of day due to the migration with a summary of the more important milestones of the systeminland and to a thickening of the as sea-breeze modelling progressed from the layer ofsea air as the day progresses.On average early linear analytic models to the complex non- the vertical extent of the sea-breezeis greater in hydrostaticnumerical models of the nineties. tropical areasthan in temperatezones. In temper- ate regions the sea-breezeinflow is usually be- tween200 and 500 m deep.These depths increase from 1000 m in moderately warm climates to 1400 m in tropical coastal regions. Maximum heights of 2 km have been observed in India Fine-scale structure of sea-breezes (Defant 1951;Atkinson l98l). The return flow is often difficult to observeas it The similarities between density currents pro- 'masked' is often by the synoptic conditions. duced in the laboratory, and atmospheric ex- Lyons (1972) found that the return flow layer of amples such as sea-breezesand thunderstorm Lake Michigan breezeswas typically twice asdeep outflows, were first shown at Reading University and with peak velocities half those of the inflow by retired schoolmaster John Simpson. Denser layer. Atkinson (198l) summarisesobservations (sea)air flows towards the sea-breezefront at low which show the depth of the sea-breezereturn levels and is swept up and backwards at the lead- flow ranging from less than 0.5 to nearly 3 km. ing edge in the region referred to as the head The sea-breezeis usually accompaniedby a de- (Simpson 1969, 1912). Mixing between cool crease in the temperature and a corresponding moist seaair and warm dry land air takes place at increase in the humidity of the atmosphere. the rearofthe headthrough breaking billows, gen- Atkinson (1981) presentsdata for the sea-breeze erated at the front of the head by the Kelvin- in India, showing a mean monthly temperature Helmholtz shear instability mechanism. Lidar fall due to the sea-breezevarying from 1.l'C to observationsof suchbillows, with a wavelengthof 4.4'C. These same data show that the range of about 2 km, can be found in Nakane and Sasano relative humidity risesis from five to 30 per cent (1986). The mixed air flows back towards the with an averagevalue of about ten per cent. The coastline above the inflow and constitutes the observationsof Simpsonet aL (1977) show that return flow ofthe sea-breeze. the main humidity rise takesplace within 200- Later experiments by Simpson and Britter 250 m ofthe leadingedge ofthe front. They also (1919)examined the relation betweenthe depths found a minimum value of humidity directly of the boundary layer, the inflow and the raised behind the front, indicatingdry air which hasde- head. Nakane and Sasano(1986) measured a scended behind the front and which will sub- 1300m head with a following flow of 300 m, but sequentlybe mixed with the moist seaair. This no mixedJayer height ahead of the sea-breezeis minimum most often occurs about 800 m from given (the Simpsonand Britter experimentsimply the front. The transport of moister air over the it should be about 5000 m, which is surely too land is regardedas a typical characteristicofthe large). For the well-documented sea-breezede- sea-breeze,although this is by no means defini- scribed by Simpson et al. (19'17), also with a tive. Observationspresented by Physick (1982) 300 m deep inflow and with a mixed layer of and Abbs (1986) documentcases in which the re- 1400m, the head was found to be only twice the versewas observedto occur. depth ofthe 300m deepinflow and extendedback Flohn (1969) gives examplesof the average about one to two kilometres from the leading horizontal extent ofthe sea-breezefor various cli- edge. Aircraft measurements through this sea- Abbs and Physick: Sea-breezeobservations and modelling

Fig. I Plot of vertical velocity, hurnidity mixing ratio frontogenesis.The raised head in this study is and potential temperaturemade during a traverse about twice as deepas the 250 m deepsea-breeze ofa sea-breezefront. The traversewas madeat a inflow. height of 600 rn. The sea-breezefront occursat 'Ihe The first fine-resolutionmodelling study of the 0 km on the horizontal axis. captions sea-breezeto in literature was pub- 'rough', 'bubbly' and 'steady, smooth' indicate appear the the degreeof turbulence(Simpson et al. 1977). lished by Sha et al. (1991). Using a non-hydro- staticmodel and a horizontalresolution of 100rn. they were able to simulate all the fine-scale featuresfound in the laboratoryexperiments and verify the empiricalrelations involving the inflow and head depths, and billow amplitudes and wavelengths. They also found that Kelvin- Helmholtz instability (KHI) doesnot occurin the initial or final stagesofthe sea-breeze,and that the KHl-induced mixing in the middle stage(afternoon) is responsiblefor slowing the inland penetration during this period. This finding may explainthe failure of coarser-gridmodels to simulate the observed middle-of-the-day slowing of the 14 June 1973 sea-breeze(Carpenter 1979; Physick1980).

Behaviour in complex terrain breeze (Fig. l) show mixing taking place in the The topographyof the coastlinehas an important head region with a billow wavelength of about controlling effect on the sea-breeze.As well as in- 200 m. Points marked A and B, 50 m apart, de- fluencingthe inland penetrationofthe sea-breeze note adjacentdata points characteristisof undi. it affectsconvergence/divergence which in turn luted seaand land air respectively. affectsassociated weather. It may also afect the Along the front during the day, the leading edge number ofsea-breezesexperienced and there are consistsofa continually changingpattern oflobes indications that it may have an effect on the onset and clefts about one kilornetre across.These arise time. The classicalmodel of the sea-breeze,as out- from convectiveinstability causedby the over- lined in the introduction, has been developedin running ofwarm land air at the ground by denser uncomplicated terrain, usually along a lqng seaair (Britter and Simpson1978), and disappear straight coastline with a flat hinterland. Later at night when radiative cooling rapidly cools the studies conducted in complex terrain indicate near-surfacelayers. that these sea-breezesdo not conform to the In a summary paper of the laboratory exper- simple model. iments on density currents,Simpson and Britter Irregular coastlines ( I 980) point out that no mixing takesplace at the interface betweenthe sea-breezeinflow and the Curvedcoastlines. Neumann( l95l ) showedthat return flow above.This implies that emissionsof the convergenceor divergenceofthe land and sea- pollutantsinto the sea-breezewill be transported breezesdepends on the curvatureofthe coastline. inland virtually undiluted, rise and mix with am- Where the coast is concave seawards the sea- bient air at the front, and head back towardsthe breezediverges in contrastto the convergingland- coast in the return flow. Furthermore, a'clean' breeze, while at convex coastlines the reverse sea-breezewill sweeppolluted ambient air up- occurs.He also showedthat the variation in the wards and remove it seawardsat higher levels percentageof thunderstorms within coastalareas while maintaining unpolluted air at the ground. is connected in part with the convergent or The role ofsea-breezesin air quality is discussed divergent nature of the sea and land-breezes in a later section. involved. A recent aircraft study of sea-breezeson the In a study of the meteorological conditions as- straightSouth Australian coastline by Kraus et al. sociatedwith Los Angelesair pollution, Edinger ( I 990) obtainedd^ta at 3 m resolutionat various and Helvey(1961) discovered that on somesum- levelsthrough the fronts.The variousterms in the mer afternoons, narrow zones of converging frontogenesisequation were calculated on a 2 km winds extend acrossthe inland valleys.The con- length scale,with confluence,shear and diabatic tinuousrecords from a groupofobserving stations processescontributing in equal magnitude to in the SanFernando valley revealedabrupt shifts l0 AustralianMeteorological Magazine 4l December1992 from the usual sea-breezefrom the south-eastto a Fig. 2 Mean arrival time (in the form of isochrones) westerly sea-breezelater in the day. Thesecon- throughout the Latrobe Valley for the period 8- verging winds are the result of the sea-breezes 12 March 1985 for (a) the sea-breezefrom the coast and (b) the sea-breezefrom the south formed along a convex section of the California east coast(Physick and Abbs 1992). coast.As the sea-breezesconverge, both easterlies and westerliesare deflectedupwards in the convergencezone and becomestrong updrafts. They suggestthat this mechanismmay help to act as a lsochrones EastCoast Sea Breeze generatorof upper level pollution by pumping aloft Los Angelesair pollution. Similarly, in the Elsinore Valley, glider pilots have reported the 'smog occurrence of a front' on many summer afternoons.This resultsfrom the convergenceof two air massesof markedly different histories over the land - one a polluted breezefrom Los Angeles,the other an unpolluted breezefrom the coast. They also report that the line along which the two sea-breezesconverge is the site of strong vertical currents. The convergenceof sea-breezesfrom different coastlineshas also been studied by Physick and Abbs( 1991,1992). Their analysisof datafrom the Latrobe Valley surface network (Fig. 2) revealed that sea-breezesfrom the south and east coastsof the regionwere the dominant wind featuresover the regionduring the studyperiod. In this casethe east coast sea-breezetravels farther inland and is the stronger of the two breezes,pushing the southern sea-breeze'back'{ifwardsthe coast after the two meetin the lateafternoon. This behaviour lsochrones South Coast Sea Breeze was reproduced in their numerical modelling study. The first three-dimensional modelling study of the sea-breeze(McPherson 1970) was motivated by a desireto assessquantitatively the effectof coastlinecurvature.,on'the evolution of the breeze. Ih his simulationshe was able to show that the effectof the bay is evident immediately.Both the sea-breezeconvergence zone and the thermal gradientwere distorted around the modelbay and were symmetric with respect to the bay. As the simulation progressed the degree of distortion diminishedand an asymmetryin the thermaland vertical motion fields developed.This asymmetry is due to the Coriolis acceleration. Bay and ocean breezes. In regions where the coastline changesdirection suddenly, such as a gulfor deepbay, it is possibleto producetwo sea- breezes.The first of these we will refer to as the bay-breeze.This breezeis a responseto the temperature diference betweenthe waters ofthe bay (or gulf) and the surroundingcoastline. The second is the ocean-breezewhich originatesalong the 'large scale'coastline. Physick and Byron-Scott(1977) observedthe existenceof sea-breezeson each side of St Vin- examplesof dayswhen the gradientwind is eastto cent'sGulf and an ocean-breezewhich movesup northeast.the resolvedwind at the westerncoast- the Gulf from the Southern Ocean and reaches line is observedto be from the southeastin the Adelaideduring the afternoon.The ocean-breeze afternoon whereasthe Gulf sea-breezecompon- is evident in the wind recordsonly under near- ent and the gradient wind component are both calm synoptic conditions. This is supportedby from the northeast quadrant. Abbs and Phvsick:Sea-breeze observations and modelling ll

Barbato (1978) observedthe generationof a over the coastalridge into the valley, depending bay-breezeand an ocean-breezein Boston.Boston on the direction of the synoptic wind (Sumner is situated in a concave, hemispheric-shaped I 977; Physick 1982). Studies on the west coastof basinwith an escarpmentrising to 100m located the United Stateshave shown that the inland pen- along the west and northwest.He found that the etration of marine air is strongly dependenton bay-breezewas often supersededlater in the day topographic channels (Fosberg and Schroeder by the ocean-breeze.Ocean-breezes of long dul- 1966;Olsson et al. 1,973).In the San Francisco ation penetrated inland with well defined charac- region,Fosberg and Schroeder(l 966) havefound teristics. that the sea-breezemoves from the coast to the A similar situation has beenobserved by Abbs inland valley through the natural passessuch as (1986) for Port Phillip Bay, Melbourne.She ob- the Golden Gate, PetalumaGap and Carquinez servedsituations in which the ocean-breezefrom Strait. They also found that thesepasses and de- BassStrait penetratedinland abovethe Port Phil- flecting barriers such as hills createvortices in the Iip Bay bay-breezebut was not recorded at the sea-breezeflow. During the Atmospheric Studies surface.Under theseconditions the ocean-breeze in Complex Terrain (ASCOT) program in the ascendsthe b ay-breezeand penetratesinland with Geysers area of northern California, the sea- it. This situation occurswhen the ocean-breezeis breezewas found to havea stronginfluence on the warmed from below by the land surface over initiation and evolution of evening drainage which it passes,hence becoming more buoyant. winds (Neffand King 1987). Becauseof this excessbuoyancy the ocean-breeze Neumann and Savijarvi(1986) modelled the is ableto mount the bay-breezeas it movesinland. situation where mountains rise steeplyfrom the On many occasionswhen the ocean-breezedoes shoreline.They wereprompted by two setsof l9th occur at the surface, its onset is accompaniedby century sea-breezeobservations from the moun- an increasein dry-bulb temperatureand a de- tainouscoasts of Chile and Peru,involving sailing ships.In one ofthese,the vesselln lucolay in port creasein humidity. This is dueto the advectionof 'while warmer, drier, land-modified air replacing the ready to sail for four days the water about cool, moist bay-breezeof Port Phillip Bay. It has her was mirror-smooth'. About 4-6 km offshore, also been shown, both observationally and nu- the seawas disturbedeach day by a moderatesea- merically,that the interactionof thesetwo breezes breeze. In their experiments they varied the may producea mesoscalecyclonic circulation. mountain height and the Brunt-Viiisiilii fre- quency, finding that as each is increased the Orography blocking effecton the sea-breezeintensifies. This Does the presenceof coastalorography act as a effect is most pronounced in the vicinity of the barrier to the inland penetration of the sea- shoreline, where the sea-breezebreaks down into breeze?Numerical experimentsby Mahrer and two cells (Fig. 3), one over the sea and another Pielke (1977) using a 900 m high bell-shaped mountain extending from the coast to about 75 km inland found that the sea-breezeand Isotachsof the onshore component (solid lines, in producea Fig.3 mountain circulationsacting together m s-r) andofthe verticalvelocity (dashed lines, more intensecirculation than they do actingsep- in cms-r) of thesea-breeze flow in thepresence arately, and that by early afternoon the sea-breeze of a 1000m high barrier.Note that in the de- had reachedthe lee side of the mountains.This tachedcell overthe lowerslope of the mountain I general was also found for the Kanto Plain the maximumseaward velocity is - 3 m s- and result -28 region of Japan by Kikuchi et al. (1981) who themaximum downward velocity is cms-r showedthat the extendedsea-breeze (over 200 km (Neumannand Savijarvi 1986). horizontally) observed in that region could only be produced numerically when the orography was included.Similarly, strongafternoon winds regu- larly observed over Lake Kinneret in Israel were shownby Alpert et al. ( I 982) to be the sea-breeze from the Mediterranean coast 45 km away. Dur- ing itsjourney to the lake(210 m below sealevel), the sea-breezepassed over hills 400 m high. In t their numerical experimentsfor the Perth region of WesternAustralia, Manins et al. (1992) similarly found that the presenceof the escarpment, 30 km from and parallel to the coast,acts to accel- eratethe inland movement of the sea-breeze. Observationsfrom coastalvalleys running parallel to the shoreline,and opening to the sea at somelocation, have shown that the sea-breezewill 50&920 o ro 20 & 40 I @km either travel up the valley from the mouth or spill SEA LANO t2 Australian Meteorological Magazine 4 I December' I 992

over the windward side of the mountain. For the Simpsonet al. (1977)were able to show,using a sailing ship case, their predictions were in simple numericalmodel, that the inland advance agreementwith the observationsof the ship's ofthe sea-breezeis governedlocally by the density captain. differencesacross the front. Under conditions in which the synoptic flow acrossthe coastline is weak, a sea-breezewill develop at the coast as soon as the land-sea temperature difference is largeenough. As the seaair movesinland its tem- Inland penetration peraturerises more rapidly than that of the land air as the heat is distributedover a much smaller While working as a forecasterin Canberra in the depth. Consequently,the temperature contrast 1940s,Reg Clarke became interestedin a late acrossthe front decreasesand the sea-breezefront afternoon eastedywind which wasoften observed decelerates.In addition Kelvin-Helmholtz insta- in Canberra during the warmer months (Clarke, bility increasesthe top dragthrough strong turbu- personal communication). Certain that it was a lent mixing and this increasein the top friction is sea-breezefrom the east coast (l l2 km away) but also associatedwith the decelerationof the sea- unable to convince any ofhis colleagues,Reg used breezein the middle of the day(Sha et al. I 99 I ). In car and aeroplane observations to prove that this the late afternoonthe solarheating decreases but wasindeed the case.Data from theseexperiments, the temperature of the land air still increases in which it was also claimed the sea-breezepro- gradually.At the sametime, the temperatureof gressedto Wagga Wagga (270 km inland), were the seaair is becomingcooler as the seaair arriv- published much later and favourably compared ing at the front hashad lessheat addedto it since to Reg's numerical predictions of the inland crossingthe coast.As a result, the temperature penetration(Clarke I 983a). difference acrossthe front increasesand the sea- To avoid any orographic effects,and interested breezebegins to accelerate.In the nocturnal stage in the claim of Hounam (1945) that sea-breezes ofthe sea-breeze,horizontal advection ofcool sea occasionallyreached Kalgoorlie (345 km inland), air is instrumentalin maintainingthe frontal den- he later mounted a sea-breezeexperiment in the sity gradient.This advectionofcool seaair ceases featurelessterrain of southern Western Australia once air which crossedthe coast near sunset (Clarke 1955). To prove his assertionthat sea- reaches the front. Physick and Smith (1985) breezeswere notjust coastalphenomena (the pre- suggestthat the further inland a sea-breezeis at vailing view at the time), Reg and his brother sunset,the longerit will survivebefore dissipating tracked several sea-breezesby car from the south in the manner describedbelow. coast through to Kalgoorlie, where they arrived Clarke (1984) used a two-dimensionalhydro- aroundmidnight. Sincethis pioneeringwork, sea- static numerical model to investigatefurther the breezeshave been observed by Reid (1957) at behaviour of sea-breezeswhen the geostrophic Renmark (217 km inland), by Garratt and wind is onshore.He identified five stagesin the 'im- Physick (1985) at Daly Waters (280 km) and by developmentof the sea-breeze.During the Clarke et al. (1981) acrossCape York Peninsula. mature stage' the inland penetration increases Elsewhere, deeply penetrating sea-breezeshave monotonicallywith time. At this time the relative been observedup to 200 km from the coast in flow is through the isentropesand the flow is far California (Carroll and Basket 1979), in Japan from steady state. Convective overturning main- (Kurita et al. 1985),and in Saudi Arabia (Steed- tains the frontal boundaryin a near-verticalpos- man and Ashour 1976). ition. In the'early maturestage', corresponding to One of the most interestingaspects of the pen- late afternoon, the sea-breezesurge is still not a etration ofthe sea-breezeis the changein the rate steady-stategravity current but accelerationofthe ofadvance inland ofthe front, althoughconflict- surgehas commenced.During early eveningthe ing reports on this characteristic appear in the supplyofcool air from the seaeeases, decoupling literature. Sha et al. (1991) identified three dis- of the frictional link with the surface occurs and tinct stagesin the penetration of their modelled the sea-breezecontinues to accelerate. Clarke 'late sea-breezefront. Initially the front moves at a identifiesthis as the mature stage'of the de- speedof approximately 5 km h-t before decelerat- velopmentof the sea-breezesurge. At this time the ing in the early afternoon. Later, the sea-breeze leadingisentropes are still steeplyinclined but are front acceleratesbefore decelerating and decaying beginning to be flattened and the centre of the late at night. The afternoon deceleration of the horizontal vortex shiftsto the leadingedge ofthe sea-breezewas observedby Pedgley(1958) for circulation. The surgefront is still sharp and ac- 'early sea-breezesin Egypt and by Simpsonet al. (1977) tive during this stage.In the next stage,the for their well documented southern England sea- degeneratestage', the layerofcold air flattensand breeze.The eveningacceleration ofthe sea-breeze spreadsinland propagatingas an unsteadygravity front has been observed and modelled by a num- current.During this periodthe sea-breezesteadily ber of authors(Clarke I 955; Simpsonet al. 1977 ; deceleratesand hasthe form ofa degenerategrav- Clarke 1984;Physick and Smith 1985). ity current or a steadilydiminishing vortex com- Abbs and Physick:Sea-breeze observations and modelling IJ pletely detached from the coastline. Figure 4 Clarkeet al, (1981)present four possiblemech- showsa vortex about l0 km in length observed anisms for the production of the pressurejump during the last stagesof a sea-breezesurge. In this associatedwith the morning glory. Their data vortex the air is moving faster than the front. suggestedthat the dominant mechanism for for- Physick and Smith (1985) showed(using a two- mation of the pressurejumps was the interaction dimensional model) that once the vortex has of the east coast sea-breezegravity current with formed, the main heating/coolingprocess that oc- the nocturnalinversion. The sea-breezedevelops curs is the subsidencewarming in the descending a front as it progressesinland assistedby the pre- arm ofthe circulation.This leadsto the eventual vailing light onshoregeostrophic flow, With time, decayofthe density gradient and the vortex itself' the front may develop into a closed vortex which In the late degeneratestage there is no longer a may persistfor many hoursfollowing sunset.This closed circulation near the leading edge. explanation was supported by observationsof deeplypenetrating sea-breezes (Clarke I 965). The numericalmodelling studies of Crook and Miller (1985)and Sha et al.(l 991)confirm that it isposs- Fig. 4 Relativestreamlines for a dyingsea-breeze surge ible for an undular bore to be produced by a Downs observedduring the Coonalpyn ex- gravity peditionin SouthAustralia. The streamlinesare current moving into a stably stratifi.ed det€rminedfrom serial balloon flights centred on fluid. Other suggestedmechanisms were the pro- 2200 CST at a point 165 km inland (Clarke duction ofan internal bore on a katabatic flow, the 1983a). collapseof lee waves or standing eddies in the vicinity ofhigh ground and the head-oncollision of sea-breezesfrom the east and west coast of CapeYork Peninsula. After further analysis of the field data, Clarke (1983b)concluded that the morning glory is pro- ducedby the interactionofsea-breezes from both sides of the Peninsula.The first necessarycon-

€ too dition for bore formation is that a sea-breezesurge should move inland from the east coastof Cape York Peninsula and interact with the weaker sea- breeze on the west coast of the Peninsula. The west coast sea-breezeacts to modify the atmosphere in a coastalstrip about 100 km wide, pro- ducing a nocturnal boundary layer over both sea and land on which the internal bores can propa- gate.For this sequenceofevents to occura modest easterlygeostrophic wind is necessary. This descriptionofthe formation ofthe internal Role in internal bore generation bore was later confirmed by the two-dimensional modelling studiesof Clarke (1984) and Noonan The 'morning glory' of northern Australia is the and Smith (t986). Clarke showed that the colli- most widely documentedexample of internal un- sion ofthe two sea-breezesfrom either sideofthe dular bores occurring in the atmosphere,and the Peninsularaises cool air in a sharp,upward bulge. importanceof the sea-breezein its formation has One or two bores may then propagateaway from been acceptedsince the early 1980s.A complete the genesissite at the samespeed as the original 'morning descriptionof the glory' is presentedin sea-breezes.The evolution ofthe vertical velocity this volume by Christie (1992). Analysis of the and potential temperaturefields associatedwith field data from the 1978 and 1979 expeditions, morning glory genesis are shown in Fig. 5. enabled Clarke et al. (1981) to show that the Noonan and Smith also showedthat sea-breeze morning glory is not a gravity current, a charac- collision must occur in the late eveningfor a sus- teristic of which is lowJevel feeder flow toward tained bore to be formed. This ensuresthat air the leading edgeof the disturbance in a frame of associated with the east coast sea-breezehas reference moving with the front. Their obser- stabilisedat low levelsand so there is a local up- vations showedthat the relative flow was through ward flux of negative buoyancy on collision. If the system and contained very little advected collisiontakes place earlier, the lowlevel air ofthe fluid. The streamlinepatterns show that the dis- east coast sea-breezeis still well mixed and it is turbance has the characteristicsof a travelling this air that is lifted at collision.Under thesecon- wave (or group of waves)but the pressurerise ditions a sustainedbore is not produced. The accompanyingthe disturbanceis sustainedfor three-dimensionalstudies of Noonan and Smith over threehours. Thus, they concludethat the dis- (1987)showed that the sea-breezeconvergence is turbance is an atmosphericundular bore propa- maximisedin the late eveningat about the time gating on the nocturnal or maritime inversion. and place morning glories are observed to AustralianMeteorological Magazine 4l December1992

Fig. 5 Vertical cross-sectionsof the vertical velocity and potential temperature(a) before,(b) at the time of, and (c) after morning glory genesis.In (d) the bore-like disturbance has begun to move westwards (Noonan and Smith 1986).

^1. ="-4

0. jc0. lLU. JUU. :sU. 200. 300. 800. 0:siaNcE ($, OIS-iNCE (KH:

0. 200. 300. 900. 500. e00. 700. 600. 200. 400. 500. 600. CISTFNCE (X}I' 0lsif,NcEiKr'l)

c A

5 =3.

=2.

0. 2aa. 300. q00. 500. 500 zCC, 300. 400. 500. 500. 7AO. 9C0. OISlFNCE(KHI 0lsTiNcS rKr'rl d 5.

317 _

; i. - ; = 2,

0. ?c0. 4CA. 500. 600. ?00. 800. 2C0. 300. {0c. 500. 500. 800. OISTNNC=(KHI OISTFNCEIKH' Sea-breezeobservations originate and a sustained westward propagating Fig. 6 Schematic diagram illustrating fumigation of a disturbanceis produced.The disturbanceis ap- coastal plume by the thermal internal boundary (1982) proximately parallel with the east coast of the layer. This diagram is from Portelli and network for the Lake Erie of the east illustrates the data Peninsula,reflecting the orientation fumigation experiment. coast sea-breezefront. They conclude that the reason that morning glories are confined to the southern part of the Peninsula is due to the fact that by late evening the disturbance produced by the colliding sea-breezeshas significant ampli- tude only in that region.

Sea-breezesand air quality When assessingthe impact of pollutant emitters in coastalregions, it is essentialto considerseveral of the sea-breezecharacteristics discussed in earlier sections. We examine the air quality in both the near and far field. Near-sourcedispersion The thermal internal boundary layer (TIBL) which developsin the stable sea air as the sea-breeze passesover the heatedland can be responsiblefor enhanced surface concentrations downwind of a coastal source. Figure 6 illustrates how a plume with the fresh emissionsfrom the coastal sources, initially travellingin stableair abovethe TIBL is leading to an ever-increasingloading of the atmos- brought to ground (fumigated) further inland by phere in the shorelineregion. convectiveeddies in the growingTIBL. Probably Our discussion so far has concerned elevated the most comprehensivedata set on TIBL be- sources,primarily emitting oxides of sulfur and haviour and plume fumigation was gatheredin nitrogen. We now turn to surface emissions, of 1978 at Nanticoke on the shore of Lake Erie, reactivepollutants, and show that anotheraspect Canada(Portelli 1982;Kerman et al. 1982;Hoff ofthe sea-breezeis frequently responsiblefor high et al. 1982).Over the eightdays of the experiment, levelsof photochemicalsmog, of which ozone is fumigation of the plume from the coal-firedpower the major component. plant (final-riseheight about 350-400 m abovethe Ozone is formed on a time-scale of several ground) usually began between 1100 and 1300 hours through reactionsinvolving non-methane hours local time and continued till about 1700- hydrocarbonsand nitrogenoxides in the presence I 900 hours. The plume wasfirst brought to ground of ultraviolet radiation. The dominant sourceof at distances ranging from 7 to 20 km from the these ozone precursors in populated areas is the stack,but this distancevaried during theday as the internal combustionengine, although various in- TIBL grew and decayed.The fumigation zonewas dustries also contribute. In coastal cities such as also found to rotate about the stack as the wind Athens (Lalas et al. 1987; Gusten et al. 1988), direction changed,primarily due to the Coriolis Melbourne (Evanset al. l98l; Cope et al. 1990; force acting on the sea-breeze. Johnson and Cope 1990), Sydney (Hyde et al. Continuous fumigation of shoreline plumes 1978;Hawke et al. 1983)and Perth(Manins et al. was alsoexamined by Lyons and Cole (1973)for 1992), the local meteorology in summertime is the Milwaukeearea who found that while the on- such that land-breezesor light offshore synoptic shorelake-breeze was cleanat the shoreline, elev- winds advect emissionsfrom the morning peak ated plumes at the shore were brought to ground traffic period out over the sea.Here the mixture is in the TIBL within 3-5 km inland.A similar study able to 'cook' for a few hours unhindered by by Lyons and Olsson(1973) on the westernshore sourcesof fresh nitric oxide (NO) which are likely of Lake Michigan in the Chicago area found that to reduce the ozone levels. When the sea-breeze not only does the same fumigation mechanism developsin the late morning, the ozoneis trans- occur there, but the air quality problem is exacer- ported onshore and over the city and suburbs bated by recirculation of pollutants within the responsiblefor its precursors.Under these con- lake-breezecirculation. Pollutants fumigated into ditions, the highest ozone levels (in late after- the inflow riseat the front and move backover the noon) are usually found within 30-40 km of the lake in the return flow. However the larger pol- coastalsource. In the following sectionon long- lutant particlessubside into the top region ofthe range transport, we find that daily maximum inflow and are brought back over land where they readingsof'city-produced'ozone can occur asfar are fumigated to the ground once again, along away as 200 km from the source. t6 AustralianMeteorologica I Magazine4 t Decembei I 992

Long-range transport far asminimising air pollution on sea-breezedays The deep inland penetration achievedby some is concerned.Further discussionon the meteor- sea-breezessuggests that pollution from coastal ologyand air quality of the LatrobeValley region regionsmay be transportedlarge distancesfrom can be found in Physick and Abbs (1991, the source.Such evidencehas been found in a t992). number of Californian localities including YosemiteNational Park wherethe sea-breezeand upslopecirculation combineto import pollutants from the San FranciscoBay area 200 km away (Carroll and Baskett 1979),and in Palm Springs Sea-breezemodelling where Los Angeles smog arrives from 180 km away around midnight (Grosjean 1983). Los Whilst the dynamicsof the sea-breezemust have Angelesis alsoresponsible at times for high ozone occupiedthe minds of many naturalphilosophers readingsin the SantaBarbera region to the north- during the last few hundred years, one of the west(Moore et al. 199l; Hanna et al. 199l). earliest worthwhile theoretical studies of the In Japan,the highestozone concentrations are phenomenonappears to be that ofJeffreys(1922), not usuallyfound in Tokyo wherethe precursors who consideredit to be a type of antitriptic wind. are produced, but in the mountainous region Sincethen, a number of linear modelsof the sea- 150km to the northwest in the early evening breeze have been produced, including that of hours.The smog is transportedthere by a circu- Haurwitz (1941). He showedthat the incorpor- lation resulting from interaction between the ation of friction into the modelenabled the theor- sea-breezeand thermally and topographically- etical resultsto resemblethe observationsmore induced flow, known as the extendedsea-breeze closely,insofar as it causedthe sea-breezeto de- (Kurita et al. 1985;Chang et al. 1989;Kondo creaseearlier. He also demonstratedthat the ob- 1990).By analysingmeteorological and air qual- servedveering ofthe sea-breezeduring the day rs ity data for the region and aircraft data alongthe due to the Coriolisforce. Schmidt (1947) devel- route of a polluted air massfrom Tokyo, Ueda et oped a linear model in which the velocity fields al. (1988) found that the highestsurface concen- werecalculated from an atmospherictemperature trations of oxidantsoccurred at night-time in the distribution which was specifiedas a function of mountains,at the rear edgeof the cut-off vortex tlme. where the internal circulation brings upper-layer The linear modelsreproduced certain observed oxidantsto the ground.This vortex, about 25 km featuresofthe sea-breeze,such as the depth ofthe in lengthin the Tokyo cases,forms at the leading flow and the rotation of the wind vector. How- edge of the sea-breezeas it encountersa stable ever, ir was not until Pearce(1955) performed atmosphereafter sunset. the first successfulnumerical integration of the So far our discussionhas concentrated on cases nonlinearsea-breeze equations, that modelswere in which the sea-breezeexacerbates pollution constructedwhich exhibitedthe oft-observedsea- problems.We now examinea situationin which it breezefront or convergencezone. is very effectivein replacingpolluted surfaceair In a mesoscalephenomenon such as the seaor with cleanair. The LatrobeValley in southeastern lake-breeze,there is a significanttransport ofmo- Australia contains coal-fired power stations lo- mentum and heat by turbulence,and thus it is cated about 90 km up the valley from the coast. important that this aspectof the numericalmodel Under clear-skysummertime conditions, the be simulated as accurately as possible. Pearce stationplumes become well mixed throughoutthe (t 955)prescribed an artificial heatingmechanism convectivelayer and drift down the valley,meet- of the atmosphereunrelated to the predictedtem- ing the sea-breezeabout 25 km from the coast.Air peratureand wind fields,and also neglectedsur- in the polluted mixed layer risesat the sea-breeze face drag. In one of the first attemptsat a more front, is mixed with incoming seaair at the top realistic simulation of the momentum and heat and rearofthe headregion and continuesits pass- exchange,Fisher (1961) used a gradienttransfer agetowards the coastin the return flow at higher approachand assumeda constantprofile for the levels.Clean air is broughtin at lower levelsas the eddy exchangecoefficients (K, and Ks) through- sea-breezecontinues its movementinland. By the out the day and night, with a maximum value time it reachesthe power stations in late after- occurring at a height of 50 m. It is well known, noon, convectionhas decayed and the plumesare however, that these transfer coefficientsare at no longer brought to ground. They are taken leastan order of magnitudeless in stablethan in further inland during the night at about 500 m unstableconditions. abovethe ground, while at higher levelstowards Estoque(1961) divided the modelledregion the coast the sea-breezereturn flow transports into a constant-fluxsurface layer of 50 m, in which afternoon emissionsout to sea.For the Latrobe Ku (:Ki is a function of stability and wind Valley region, it appearsthat the emitters may shear, and a transition sub-layer,in which Kaa very well be located at the optimum location, as decreasesin a linear fashionfrom its value at the Abbs and Physick:Sea-breeze observations and modelling 1',l interface to zero at the top of the model at 2 km. Virtually all mesoscalemodels used for sea- This model representsa landmark in sea-breeze breeze simulations employ the hydrostatic as- modelling and all advancessince then, mainly in sumption in the vertical equationof motion and parametrisationofthe boundarylayer, have really usea horizontalgrid spacingranging from 2 to l0 beenincremental. McPherson (1970) showed that km. Pielke(1972), Martin and Pielke (1983)and the manner in which Kp, decreasesto zero at the Songet al. (1985)have all examinedthe effectof top of the model is a'significant factor in the the hydrostaticassumption in sea-breezemodels determination of the intensity and extent of by deriving an estimatefor the non-hydrostatic the computed sea-breezecirculation. An expo- pressure,based purely on results from a hydro- nential decreasewas consideredby McPherson static simulation. As well as concludingthat the to give results in reasonableagreement with hydrostaticassumption is adequatefor grid spac- sea-breezeobservations. ings as small as I km, they also investigatedthe The next step forward followed the significant dependenceof the differenceon the scaleof the advancesmade in boundaryJayermeteorology in heating,atmospheric stability and synopticwind the fifties and sixties.In the numericalmodel de- speed,finding that non-hydrostatic effectsgener- veloped by Pielke (1974) to study sea-breezesin ally tend to weakenthe sea-breezewhen compared Florida, the non-dimensionalwind and tempera- to a hydrostatic simulation. Their conclusions ture profiles based on Monin-Obukhov similarity were corroboratedby Yang ( I 99 I ) who compared theory wereused to calculatefluxes in the surface hydrostatic and non-hydrostatic simulations layer.The heightsofthe surfacelayer and bound- from the samemodel. Yang also stressedthe im- ary layer werealso allowedto vary over a diurnal portance of the turbulenceparametrisation and cycle.Surface temperature (Te) had alwaysbeen recommendedthat higher order closureschemes specifiedas a sinusoidalfunction of time in sea- should be used in non-hydrostaticmodels. breezemodels to this stage,but Physick (1976) Perhapsthe final stagein sea-breezemodelling employed a surface energy balanceequation to was reachedrecently with the publication of a calculateTe, thus allowinghorizontal gradients of paperby Sha et al. (1991).Using a non-hydro- Ts and surfaceheat flux to exist acrossthe front. staticmodel and a horizontalgrid spacingof only Development of this model benefitedfrom dis- 100 m, they were able to resolve explicitly the cussionswith Reg Clarke who was designinga mixing process in the head region, whereby model to help explain his observationsof sea- warmerambient air is entrainedinto coolersea air breezesup to 300 km from the coastline.Simi- by thebreaking of Kelvin-Helmholtzbillows. This larity theory and a local schemefor the exchange fine structureof the sea-breezehead as modelled coefficientsthroughout the boundarylayer, devel- by Sha et al. is shown in Fig. 7. oped from analysisof the Wangara data (Clarke 1974),werc usedin his modelbut detailswere not publisheduntil a decadelater in a study of nocturnal wind surges(Clarke 1983a). However some Fig.7 Verticalcross-section showing the fine structure results on deeply penetrating sea-breezeswere of the sea-breezehead. The numberson the presented at a Fire Weather Conference at horizontalaxis are distancefrom the coastline. Monash University (Clarke 1973). The solidline indicatesthe zerovelocity boundary.The front position is shown.Horizontal Monin-Obukhov similarity theory in also the sur- vorticesare labelled, a, b, c (Shaet al. 1991). face layer and an energybalance equation for the computation of surfacetemperature are now stan- dard fare in models used for the simulation of sea-breezes.The main variation betweenmodels occurs in the parametrisationof turbulent ex- changein the transition layer above the surface layer. Theserange from local (Kondo 1990)and non-local (Pielke 1974) K{heory to first-order (Arritt 1987), second-order(Schumann et al. I 987)and third-order (BriereI 987)turbulent kin- etic energy(TKE) schemes.A comparativestudy ofa numberof theseparametrisations by Mahfouf lrorl LAND l lKil) et al. ( I 987)has found that the meanfields of wind and temperatureare relatively insensitiveto the turbulence specification,but that the more de- Conclusion tailed schemespredict the turbulence statistics more accurately, an important finding for air In this review we have presentedearly results quality models in which the dispersionis driven which described the structure of the sea-breeze by winds and turbulencefields from a mesoscaie and its efect on the wind, temperatureand hu- model. midity fields at coastal locations. The various l8 AustralianMeteorological Magazine 4l December1992 effects of orography and coastline curvature on atmospheric bores: Part II Internal atmospheric bores rn the inland penetration of sea-breezesare also pre- northern Australia. Aust. Met. Mag., 31, 147-60. sented.More recentstudies ofthe sea-breezeare, Clarke, R.H. 1984. Colliding sea-breezesand the creation of internal atmosphericbore waves:two dimensional numeri- in general, the result of air quality research for cal studies.Aust. Met. Mag. 32,2O7-26. coastal regions and the role of the sea-breezein Clarke,R.H., Smith, R.K. and Reid, D.G. 1981.The Morning pollutant dispersalhas been discussed. Glory of the Gulf of Carpentaria: an atmospheric undular Advances in computer power have been paral- bore. Mon. Weath Rev.,109, 1726-50. leledby an increasingsophistication in numerical Cope,M.E., Carnovale, F. andJohnson, M.H. 1990.An air quality modelling systemfor oxidant transport and production models of the sea-breeze.These models have al- in the Port Phillip control region.P/oc. Ilth Int ConJ Clean lowed researchersto obtain a greaterunderstand- Air Soc.Aust and N.2., Auckland, 165-70. ing ofthe sea-breezeand arenowjust beginningto Crook,N.A. and Miller, M.J. 1985.A numericaland analytical be used by regulatoryauthorities for air quality study of atmosphericundular bores.Q Jl R. met. Soc.,I I I, problems. 225-42. Defant, F. 1951. Local Winds. Compendium of Meteorology. AmericanMeteorological Society, Boston, 655-72. Edinger,J.G. and Helvey,R.A. 1961.The San Fernandocon- vergencezone. Bull. Am met Soc.,42,626-35. Estoque,M.A. 1961. A theoreticalinvestigation of the sea breeze.Q. Jl R. met. Soc.,87,:i,36-46. Evans,L F., Weeks,I.A. and Ecclestone,A.J. 1981.An aerial References surveyof Melbourne'sphotochemicalsnog. Proc.Int. Clean Air Conf.,Sydney, 93-1,10 Abbs, D.J. I 986. Sea-breezeinteractions along a concavecoast- Fisher, E.L. 196'l. A theorotica.lstudy ofthe seabreeze. J. Met., line in southernAustralia: observations and numerical mod- I 8. 216-33. elling study.Mon Weath.Rev., I14,831-48. Flohn, H. 1969. Local wind systems. World Suney of Clima- Alpert,P., Cohen, A., Neumann,J. andDoron, E. 1982.A model tology (General Climatology), Volume 2, Elsevier, Amster- simulation of the summer circulation from the eastern dam,139-71. past Mediterranean Lake Kinneret in the Jordan Valley. Fosberg,M.A. and Schroeder,M.J. 1966. Marine air pen- Mon. Weath. Rev.,ll0,994-1006 etration in centralCalifornia. Jnl appl.Met.,5,573-89. Arritt, R.W. 1987 The effect of water surfacetemperature on Garratt, J.R. and Physick,W.L. 1985-The inland boundary lake breezesand thermal internal boundary layers.Bound. layer at low latitudes: II Seabreeze influences. Bound. Lay. Lay. Met.,40, l0l-25- Met., 33,209-31. Atkinson, B.W. 1981. Mesosca[eAtmospheric Circulations. Grosjean,D. 1983. Distribution of atmosphericnitrogenous AcademicPress, London. pollutants at a Los Angelesarea smogrcceptor site.Environ Barbato,J.P 1978.Areal parametersof the seabreeze and its Sci.Technol.. 17. 13-19. vertical structure in the Boston basin. Bull Am. met. Soc , Gusten,H., Heinrich, G, Cvitas, G., Klasinc, L., Ruscic,B., 59. t420-r43t. Lalas, D.P. and Petrakis, M. 1988. Photochemicalfor- Briere,S. 1987.Energetics ofdaytime seabreeze circulation as mation and transport of ozone in Athens, Greece.Atmos. determined from a two-dimensional,third-ordsr,turbulence Environ.,22, 1855-61. closuremodel. J. Atmos Sci.,44,1455-74. Hanna, S.R., Strimaitis, D.G., Scire, J.S., Moore, G E. and Britter, R.E. and Simpson, J.E. ,1978.Experitnents on the Kessler,R.C. I 99 l. Overview of results of analysisof data dynamics of a gravity ourrent head. J. Pluid Mech., 88, from the south-centralcoast cooperativeaerometric rnoni- 223-40. toring program (SCCCAMP 1985). Jnl appl Met., 30, Carpenter, K.M. 1979. An experimental forecast using a non- 5lr-33. hydrostatic mesoscale'model.Q. Jl R. met. Soc., 105, Haurwitz, B 194?.Comments on the sea-breezecirculation.,/. 629-5s. Met..,4,l-8. Carroll, J.J. and Baskett,R.,L. rl'979.Dependence of air quality Hawke,G.S., Heggie, A.C. and Hyde, R. 1983.Meteorological in a remote location on local and mesoscaletransports: a factors controlling high ozone levels in the Sydney region. casestudy. Jnl. appl.Met., 18,474-86. ln: The urban atmosphere- Sydney,a casestudy. Edited by Chang,Y.S., Carmichael, G.R.., Kurita, H. and Ueda, H. 1989. J.N. Carrasand G.M. Johnson,CSIRO Australia. The transport and formation ofphotochemical oxidants in Hoff, R.M., Trivett, N.B., Millian, M.M., Fellin, P., Arlauf, centralJapan. Atmos. Envircn., 23,363-93. K.G., Wiebe,H.A. and Bell, R. 1982.The Nanticokeshore- Christie, D.R. 1992.The Morning Glory of the Gulf of Carpen- line diffusionexperiment, June I 978- III. Ground-basedair taria: a paradigm for nonlinear waves in the lower atmos- quality measurements.Atmos Environ., I 6, 429-54. phere. Aust Met. Mag., 4 1, 2l-60. Hounam,C.E. 1945.The seabreeze at Perth. Weath.Dev and Clarke, R.H. I 955. Someobservations and commentson the sea Res..3.20-55 brceze.Aust. Met. Mag., 11,47-68. Hyde, R., Hawke, G.S. and Heggie,A.C. 1978.The transport Clarke,R.H. 1965.Horizontal mesoscale vortices in the atmos- and recirculation of photochemicalsmog acrossthe Sydney phere.Aust. Met Maq.,50, l-25. basin, II: At the coast. Proc 4th Int. Conf. Clean Air Soc Clarke,R.H. 1973.A numericalmodel of the seabreeze. (Energy Aust. and N.2.. Brisbane.157-66. and massflux acrossa long straight coastlinedue to the sea- Jeffreys,H. 1922.On the dynamics of wind. Q. Jl R. met. Soc., breeze mechanisrn.) Proc. First Aust. Conferenceon Heal 48,29-47. and Mass Transfer, Monash University, Melbourne, 23-25 Jehn,K.H. 1973.A seabreeze bibliography, 1664-1972. Report May. No 37, AtmosphericScience Group, The University of Clarke, R.H. 1974. Attempts to model the diurnal course of Texasat Austin, 5l pp. meteorological variables in the boundary layer. Izvestiya Johnson,M.H. and Cope, M.E. 1990. Mesoscalenumerical Akad. Nauk USSR, Atmospheric and Oceanic Physics, 10, modelling of Melbourne pollution events. Proc l}th Int. 600-612. (Translation AMS pp. 360-374.) Conf. CleanAir Soc.Aust. and N.2., Auckland, l7l-6. Clarke, R.H. 1983a.Fair weathernocturnal wind surgesand Kerman, B.R., Mickle, R.8., Portelli, R.V. and Trivett, N.B. atmosphericbores: Part I Nocturnal wind surges.Aust Met. 1982.The Nanticokeshoreline diffusion experiment, June Mag., 3l,, 133-45. 1978 - II lnternal boundary layer structure. ltir?os Clarke, R.H. 1983b.Fair weathernocturnal wind surgesand Environ..16. 423-37. ' Abb; a;d Physick:Sea-breeze observations and modelling l9

Kikuchi,Y , Arakawa,S., Kimura, F., Shiiasaki,K. and Nagano, Pedgiey,D.E. 1958.The summersea breeze at Ismalia.Meteoro- Y. I 98l Numericalstudy on the effectsof mountainson the logical Report No. 19, Air Ministry, Meteorological Offrce, land and seabreeze circulation in the Kanto district. J. met London,l9 pp. Soc,Japan,59,723-38. Physick, W.L. 1976. A numerical model of the sea breeze Kondo, I{. 1990.A numericalexperiment on the interaction phenomenonover a lake or gulf. /. Atmos. Sci.,33,2lO7- betweensea breeze and valley wind to generatethe so-called 35. "extendedsea breeze". J. met Soc.Japan,68,435-46. Physick,W.L. 1980.Numerical experiments on the inland pen- Kraus, H., Hacker, J.M. and Hartmann, J. 1990. An obser- etration of the sea breeze.Q Jl R. met Soc., 106, 735- vational aircraft-based study of sea-breezefrontogenesis. 46. Bound.Lay. Met., 53,223-65. Physick,W.L. 1982.Sea breezesin the Latrobe Valley. .4us, Kurita, H., Sasaki,K., Muroga,H., Ueda,H. and Wakamatsu,S. Met Mag., 30,255-63. 1985. Long-rangetransport of air pollution under tight Physick,W.L. and Byron-Scott,R.A.D. 1977.Observations of gradientwind conditions.Jnl Clim. appl. Met, 24, 425- the seabreeze in the vicinity of a gulf. Weather,32, 3'13- 34. 81. Lalas,D.P., Tombrou-Tsella, M., Petrakis,M., Asimakopoulos, Physick,W.L. and Smith,R.K 1985.Observations and dynam- DN. and Helmis, C. 1987.An experimentalstudy of the ics of seabreezes in northern Aust ralia.Aust. M et M ag., 33, horizontaland verticaldistribution of ozoneover Athens. ) l-oJ. Atmos Environ-,bf 2 I,2681-93. Physick,W L. and Abbs, D.J. l99l Modellingof summertime Lyons,W.A. 1972.The climatologyand predictionof the Chi- flow and dispersionin the coastalterrain of southeastern cago lake breeze Jnl appl. Met., I I, 1259-70. Australia.Mon Wedth Rev..I19. l0l4-30 Lyons,W.A. and Cole,H.S. 1973.Fumigation and plume trap- Physick,W.L. and Abbs,D-J. 1992.Flow and plumedispersion ping on the shoresof Lake Michiganduring stableonshore in a coastal valley. Jnl appl, Met., 3 I, 64-73. 494-510. flow. Jnl appl. Met., 12, Pielke,R.A. I 972.Comparison ofa hydrostaticand an anelastic Lyons,W.A. and Olsson,L.E. 1973 Detailedmesometeorolog- dry shallow primitive equation model NOAA Tech.lvlemo., ical studiesof air pollution dispersionin the Chicagolake ERL OD-13,NOAA EnvironmentalResearch Laboratories, breeze.Mon. lil'eath Rev.,l0l,387-403. Office of the Director, Boulder, Colorado, USA, 47 pp. McPherson,R.D. 1970,A numericalstudy of the effectof a Pielke,R.A. I 974 A three-dimensionalnumerical model of the coastalirregularity on the seabreeze. Jnl appl Met.,9,76'1- sea breeze over south Florida. Mon, Weath. Rev., 102, l I 5-39. Mahfoul J F., Richard, E., Mascart,P., Nickerson,E.C. and Portelli,R.B. 1982.The Nanticoke shorelinedifusion exper- Rosset,R. 1987 A comparativestudy of variousparame- iment, June 1978 - I Experimentaldesign and program terizationsof the planetaryboundary layer in a numerical overview.Atmos Environ. 16. 413-21. l6'7 model. Jnl Clim appl. Met., 26, l-95. Reid,D.G 1957.Evening wind surgesin SouthAustralia lus, I The of topographyon Mahrer,Y. and Pielke,R.A. 977 effects Met. Mag , 16,23-32 land in a two-dimensionalnumerical model. seaand breezes Schmidt,F.H. 1947.An elementarytheory of the land and sea Mon. Weath Rev..105. l15l-62 breezecirculation. l. Met.,4,9-15. G.M. and Azzi,M. 1992- Manins, P.C.,Hurley, P J., Johnson, Schumann,U., Hauf,T., Holler,H., Schmidt,H. andVolkert, H. possible days.A contribution Predictedwindfieids on ozone I987. A mesoscalemodel for the simulationof turbulence, funded by to the Perth Airshed PhotochemicalStudy clouds, and flow over mountains: formulation and vali- Research,Janu- SECWA,CSIRO Division of Atmospheric dation examples.Beitr. Phys.Atmos, 60, 413-46 ary 1992.25pp Sha,W., Kawamura,T. and Ueda,H. t991. A numericalstudy 1983.The of the hydro- Martin, C.L and Pielke,R.A. adequacy on sea/landbreezes as a gravity current:Kelvin-Helmholtz flat terrain. staticassumption in sea-breezemodeling over "/. billows and inland penetration of the sea-breezefront J Atmos Sci.40.1472-81. Atmos. Sci.,48, 1649-65. R.C.and Killus, J.P l99l Moore,G.E., Douglas, S.G., Kesster, Simpson,J.E 1969. A comparisonbetween laboratory and ldentificationand tracking of polluted air massesin the atmosphericdensity currents. Q Jl R. met Soc., 95, appl.Met., 30,'l 15-32. south-centralcoast air basin."/n1 / )b-b ). Y 1986.Structure of a sea-breezcfront Nakane.H. and Sasano. Simpson,J.E. I 972.Etrects of the lowerboundary on the headof Lidar J. met Soc.Japan, revealedby scanning observation. a gravity current./. Fluid Mech., 53,759-68. 64,'187-92, Simpson,J.E., Mansfield, D.A. and Milford, J.R. 1977.Inland complexter- Neff, W D.and King, C.W. 1987 Observationsof penetration of sea-breezefronts. Jl R. met Soc.,I0 3, 41- topogra- Q. rain flows using acousticsounders: Experiments, 76. phy, and winds Bound.Lay Met., 40,363-92. Simpson,J.E. and Britter, R E. I 979,The dynamicsofthe head Neumann. J. 1951.The land breezeand nocturnal thunder- of a gravity currentadvancing over a horizontalsurface "/. storms. J. Met.. 8. 60-1 . Fluid Mech.. 94. 4'77-95 The and land breezesin the classical Neumann, J. 1973. sea Simpson,J.E. and Britter, R.E. 1980.A laboratorymodel of an Greekliterature Bull. Am met Soc.,54,5-8. atmosphericmesofront. Q. Jl R met. Soc.,106,485-500 H. 1986.The seabreeze on a steep Neumann,J. and Savijarvi, Song,J.L, Pielke,R.A., Segal,M., Arritt, R.W. and Kessler, coast.Beitr Phys Atmos., 59,375-89 R.C. 1985.A method to determinenonhydrostatic effects Noonan, J.A. and Smith, R.K. 1986.Sea breezecirculations within subdomainsin a mesoscalemodeI. J. Atmos Sci.,42, over Cape York Peninsulaand the generationof Gulf of 2r r0-20 disturbances.J. Almos Sci., 43, Carpentariacloudline Steedman,R.A. and Ashour,Y. 1976.Sea breez€s over north- 1619-93. westArabia. Tellus, 18,299-306. R.K The generationof north Noonan, J.A. and Smith, 1987. Sumner, G.N. 1977. Sea breezeoccurrence in hilly terrain. 'morning glory' Aust Met Australiancloud lines and the Ileather, 32,200-208. Mag.,35,31-45. Ueda,H., Mitsumoto,S. and Kurita, H 1988Flow mechanism Olsson,L E.,Elliott, W.P. and Hsu, S-J. 1973.Marine air pen- for the long-rangefransport of air pollutantsby the sea etration in westernOregon: An observationalstudy. Mon breezecausing inland nighttime high oxidants.Jnl appl. l{eath. Rev.,I0 l, 356-62. Met., 27. 182-1. Pearce,R.P. I 955.The calculationofa sea-breezecirculation in Yang,X. I 99l. A studyofnonhydrostatic effects in idealizedsea termsofthe differentialheating across the coastline. Jl R. Q. breezesystems. Bound. Lay. Met 54,183-208. met Soc..8/. 351-81. ,