Journal of Coastal Research West Palm Beach, Florida Winter 2001 ABSTRACT..

Transcription

1 Journal of Coastal Researh West Palm Beah, Florida Winter 21 Charateristis of the Sea Breeze System in Perth, Western Australia, and its Effet on the Nearshore Wave Climate G. Masslink] and C.B. Pattiarathij tgeography epartment Loughborough University Loughborough LEII 3TU United Kingdom :j:centre for Water Researh University of Western Australia Nedlands WA 697 Australia eusss 7.J = b--- ABSTRACT.. MASSE LINK, G. and PATTIARATCHI, C.B., 21. Charateristis of the sea breeze system in Perth, Western Australia, and its effet on the nearshore wave limate. Journal of Coastal Researh, 17(1), West Palm Beah (Florida), ISSN The oastline of Perth, Western Australia, is subjeted to one of the strongest and most onsistent sea breeze systems in the world. Using forty-nine years of wind data olleted at Perth airport (about 2 km inland) it was found that almost 2 sea breezes are experiened per year with an average wind speed at mid-afternoon (15: hrs) of 5.7 mls. At the oastline, sea breeze veloities are 1.4 times greater than 2 km inland and in the summer months, when the sea breeze system is best developed, sea breeze veloities frequently exeed 1 mls. A signifiant feature of the sea breeze is that it blows obliquely-onshore, rather than onshore. The importane of the sea breeze is learly indiated in spetra of the wind speed showing a signifiant spetral peak at the diurnal frequeny. Spetral analysis of hourly inshore wave data also revealed a diurnal spetral peak, suggesting a foring of the wave onditions by the sea breeze. It is onluded that the diurnal sea breeze system an have a major impat on the inident wave limate, and hene nearshore proesses, of sheltered oastal environments in tropial and subtropial regions. AITIONAL INEX WORS: Sea breeze, wind, wave limate. INTROUCTION The Perth Metropolitan oastline (Figure 1) experienes mixed, mirotidal tides with a mean spring tidal range of.6 m (EPARTMENT OF EFENCE, 1996). The offshore wave limate is dominated by a low to moderate energy wave regime haraterized by prevailing south to southwest swell (A VIES, 198) and a mean summer signifiant wave height of 1.5 m and a mean winter signifiant wave height of 2.5 m (LEMM et al., 1999). Closer to shore, the swell is refrated and diffrated by several offshore reef systems and islands, and is greatly attenuated as it propagates aross the inner ontinental shelf. As a result, the inshore wave height is about 3-7/ of that outside the reef system (WNI, 1998), depending on the inshore loation and the inident wave period and diretion. Swell height at the shoreline is generally less than 1 m (STEEMAN, 1993). A highly variable wind wave limate is superimposed on the swell regime, dominated by northwesterly to westerly storm waves in winter and by the wave field assoiated with strong south to southwesterly sea breezes in summer. The oastline of Perth is subjeted to one of the most energeti and onsistent sea breeze systems in the world (PAT TIARATCHI et al., 1997). Along the north-south trending oastline of Perth, a typial sea breeze yle is haraterized by 993 reeived 5 May 1999; aepted in revision 25 April 2. offshore winds from an easterly diretion (8-1 ) in the morning, swithing to obliquely-onshore winds from a southsouthwesterly diretion (18-22 ) in the afternoon (Figure 2). The onset of the sea breeze is often well defined and indiated by an abrupt inrease in wind speed and shift in wind diretion (at 14: hrs in Figure 2). The essation of the sea breeze is less lear beause during the evening wind onditions gradually return to pre-breeze onditions. In the early hours of the morning, land breezes may prevail. Beause of the sheltered nature of the oastline of Perth, loally-generated wind waves, partiularly those generated by strong sea breeze ativity, are a dominant mehanism ontrolling nearshore proesses and morphology (MASSELINK et al., 1997; PATTIARATCHI et al., 1997; MASSELINK and PATTIARATCHI, 1998a, b). Together with the land breeze, the sea breeze is part of a diurnal atmospheri irulation system that arises from the ontrasting thermal responses of the land and the water surfae (Hsu, 1988; ABBs and PHYSICK, 1992). esriptions of sea breezes date to the period of Greek philosophers (NEU MAN, 1984) ontinuing with writings of ampier in the 17th Century (JEHN, 1973) to one of the first sientifi aounts of AVIS et al. (189) and extending to the present (SIMPSON, 1994). Sea breeze ourrene is dependent on latitude (WEX LER, 1946) and onsistent, diurnal sea breeze ativity ours along about two-thirds of the earth oastline, espeially in the

2 174 Masselink and Pattiarathi N t o C") Rottnest 1. o o o l'r" Inshore p- Offshore waverider o buoy,4 Carnu I. w.: man Pt INIAN OCEAN 1 km Figure 1. Loation map of the Perth metropolitan and oastal regions. Bathymetry is in meters. tropis and sub-tropis (SONU et al., 1973). The sea breeze system is diretly driven by the atmospheri pressure gradient that results from the temperature differene between the air over land and over water. At night and in the morning, the air temperature over land is ooler than over the sea, and the land breeze prevails. From late morning to late afternoon, the land is warmer than the surrounding waters, and the sea breeze prevails. Generally, the sea breeze system begins some distane (. 1 km) offshore and expands in the offshore and onshore diretion (LAUGHLAN, 1997). The sea breeze system moves inland as a front, very muh like a density urrent (SIMPSON, 1994), as long as a temperature differene between land and water is maintained. The inland advane an be haraterized by a series of pulses (WALLINGTON, 196). The inland penetration speed is initially relatively rapid (1 3 m/s), but slows down as a result of frition (LAUGHLAN, 1997). In mid-latitude regions, the sea breeze an be deteted up to 1 km inland, whereas at low latitudes inland penetration an be in exess of 2 km (CLARKE, 1955; WALLING TON, 1961; GARRATT and PHYSICK, 1985). The seaward extent of the sea breeze is not very well demarated and has not reeived muh investigation (FINLATER, 1963; Hsu, 197; BANTA et al., 1993). The available measurements seem to indiate that the seaward extent is omparable to the landward extent (SIMPSON, 1994). The seaward extent of the sea breeze is important with respet to the generation of wind waves by the sea breeze beause it diretly determines the feth length; the greater the seaward extent of the sea breeze, the larger the feth length and the higher the waves generated by the sea breeze. In addition to the temperature differene between the air above land and water (i.e., the driving fore), the ourrene Journal of Coastal Researh, Vol. 17, No.1, 21

3 174 Masselink and Pattiarathi AU STRALIA tn Study area Trigg B. Sarbo rough B. Brighton B. Floreat B. ;: Inshore waverlder bu o y Off shor e waverider o b uoy INIAN OCEAN es PI 1 km Figure 1. Loation map of the Perth metropolitan and oas ta l regions. Bathymetry is in meters. tropis and sub-tropis (SONU et al., 1973). The sea breeze system is diretly driven by the atmospheri pressu re gradient that res ults from the te mperature differene between the air over land and over water. At nigh t and in the morning, the air temperature over land is ooler than over the sea, and the land breeze prevails. From late morning to late afternoon, the land is warmer than th e surroun ding waters, and the sea breeze prevails. Generally, the sea breeze sys te m begins some dista ne (. 1 krn) offshore an d expa nds in the offshore and onsho re diretion (LAUGHLAN, 1997). The sea br eeze system moves inland as a fron t, very muh like a den sity urrent (SIMPSON, 1994), as long as a temp erature differen e between land and water is mai ntained. Th e inland advane an be haraterized by a series of pulses (WALLINGTON, 196). Th e in lan d penetration spee d is initially relatively rapid (1 3 m/s), but slows down as a res ult of frition (LAUGHLAN, 1997). In mid-latitude regi ons, the sea breeze an be deteted up to 1 km inland, whereas at low latitudes inland penetration an be in exess of 2 km (CLARKE, 1955; WALLING TON, 1961; GARRATT an d PHYSICK, 1985). Th e seaward extent of the sea breeze is not very well demarate d and has not reeived muh in vestigation (FINLATER, 1963; H su, 197; BANTA et al., 1993). Th e ava ilable measurements seem to ind iate that the seaward extent is omparable to the landward extent (SIMPSON, 1994). Th e seaward exten t of the sea br eeze is importan t wit h respet to the generation of wind waves by th e sea br eeze bea use it diretly determ ines the feth length; t he greater the seaward extent of the sea breeze, the larger the feth length and the higher th e waves generate d by the sea breeze. In ad dition to the temperature differene between the air above land and water (i.e., the driving fore), the ourrene Journal of Coastal Researh, Vol. 17, No.1, 21

4 Effet of Sea Breeze in Western Australia 175 ;; 1 1:) Q) Q) 5 - en 1:) 3 "-5 2 1:) 1:) 1 7/3/95 8/3/95 9/3/95 A B Figure 2. A typial three-day time series of: (A) wind speed; and (B) wind diretion. The data were olleted at 5 m height on City Beah in Marh 1995 (data from Masselink and Pattiarathi, 1998aL and development of the sea breeze is strongly dependent on the speed and diretion of geostrophi winds assoiated with synopti weather patterns (LAUGHLIN, 1997). Unless winds speeds are small «2 m/s), onshore geostrophi winds inhibit Figure 3. Pressure distribution over Australia in surnmer: (A) West Coast trough lying offshore; and (B) West Coast trough moving onshore (after Tapper and Hurry, 1993L the formation of a sea breeze, beause suh onditions do not allow the build-up of a signifiant temperature differene, and hene surfae pressure differene, between land and water. Strong offshore geostrophi winds (> 5 m/s) also hinder the development of a sea breeze irulation beause the onshore pressure gradient due to differential heating may not be suffiient to overome the offshore geostrophi winds. Optimal onditions for sea breeze development are therefore light to moderate offshore winds (2-5 m/s) or alongshore winds indued by the synopti system. The ourrene and intensity of the sea breeze in Perth is strongly related to the presene and position of the West Coast trough (TAPPER and HURRY, 1993). In summer, the Australian ontinent is generally under the influene of easterly airflow and as a result of intense heating of the air aross the ontinent, a low-pressure trough is formed in a northsouth diretion (KEPERT and SMITH, 1992). This pressure trough is generally loated inland of the oast, but its position is subjet to day-to-day variations (WATSON, 198). When the trough lies offshore (Figure 3a), the gradient winds aross the oast are strong east to northeasterly bringing hot, dry ontinental air from entral Australia. Under these onditions, espeially with a large antiylone to the south of Australia, oastal sea breezes are delayed or non-existent. However, when the trough moves onshore (Figure 3b), usually with an approahing front, the easterly flow weakens and beomes more northeasterly. These onditions favor the development of strong and early sea breezes. A signifiant feature of the sea breeze system along the West Coast of Australia is that it blows obliquely-onshore ii.e., south-southwesterly). This is in ontrast to the lassi sea breeze, whih blows perpendiular to the shoreline. In fat, so-alled 'pure' sea breezes, i.e., sea breezes that are not interfered with by geostrophi winds (e.g., KOTINIS-ZAMBAK AS et al., 1989), are virtually non-existent along the Perth oastline. The reason for the obliquely-onshore sea breeze system in Perth may be attributed to the interation between the sea breeze and the geostrophi winds assoiated with the synopti weather patterns. The Perth sea breeze system in sensu strito is southwesterly due to a ombination of pres- Journal of Coastal Researh, Vol. 17, No.1, 21

5 176 Masselink and Pattiarathi sure gradient flow (owing to differential heating) and the Coriolis fore (KEPERT and SMITH, 1992). The Coriolis fore has no influene on the wind diretion at the start of the sea breeze, but may indue a small anti-lokwise shift in the wind diretion of about 2 during the sea breeze (refer to Figure 2). When the West Coast trough is loated inland, the synopti pressure gradient ats in a northeasterly diretion (refer to Figure 3b). The ombination of the sea breeze system (strong southwesterly airflow) and the synopti pressure (weak northeasterly airflow) results in a obliquely-onshore, south-southwesterly sea breezes. When the loation of the trough is suh that the synopti pattern indues southerly winds, the sea breeze enhanes the southerly winds, resulting in very strong sea breezes with wind speeds in exess of 1 m/s. The diretion of land breeze is not muh affeted by geostrophi winds beause both winds generally blow from the same quadrant (south to southwest). espite the dominant ourrene of the sea breeze along the Perth oastline in summer and its signifiant influene on oastal proesses it has not attrated muh sientifi interest. To date, the most omprehensive investigation is by HOUNAM (1945), who analyzed several years of wind data olleted at the Perth Observatory (6 km inland). At this loation, the mean sea breeze diretion is approximately 23, but varies between 2 and 28. In summer, the mean veloity of the sea breeze is 5.5 m/s and sea breezes our more than 69'(i of the time. In winter, sea breezes are experiened up to 2 1 ) of the time with mean veloities of 3 m/s. The onset of the sea breeze is usually around 14: hrs, but an be anytime between 11: and 16: hrs. Aording to HOUNAM (1945), the sea breeze in Perth does not penetrate inland beyond 9 km due to the presene of the arling Ranges, a north-south trending mountain range loated 3-35 km inland. However, TAPPER and HURRY (1993) mention that the sea breeze may penetrate 15 km inland from Perth by about 22: hrs. Using balloon flights, HOUNAM (1945) established that the thikness of the sea breeze layer in Perth averages 6 m. The objetive of this paper is to haraterize and quantify the sea breeze limate of Perth and indiate the impat of sea breeze ativity on the nearshore wave onditions. A omprehensive analysis of forty-nine years of wind data olleted at Perth airport (2 km inland) will be presented. One year of wind data measured at nine other Western Australian oastal loations will also be analyzed to demonstrate strong sea breeze ativity to be a dominant feature along the entire Western Australian oastline. Finally, the effet of sea breeze ativity on the nearshore wave onditions will be addressed using four years of offshore and inshore wave data. ETAILS OF THE SEA BREEZE SYSTEM IN PERTH Time-omain Analysis of Perth Wind ata Wind data olleted every three hours at 1 m height at Perth airport (2 km inland) from were used to determine the long-term sea breeze limate of Perth. The first part of the data set ( ) does not inlude measurements at 21: hrs, but these data were obtained through linear interpolation using the 18: hrs and : hrs data. Linear interpolation to derive the 21: hrs wind speed data does not have any effet on the time-domain analysis and only an insignifiant effet on the frequeny-domain analysis of the wind data. Other oasionally missing data were also linearly interpolated. The data were subjeted to an algorithm that seleted the days during whih sea breezes ourred. A day was onsidered a 'sea breeze day' if: (1) the wind diretion in the afternoon (15: hrs) was from the sea breeze diretion (19 3 ), but the wind in the morning (9: hrs) was not from that diretion; or (2) the wind diretion in the morning and afternoon were both from the sea breeze diretion, but the afternoon wind speed was larger than during the morning. Using these seletion riteria, the number of sea breezes per month and the mean sea breeze speed and diretion were determined. It is aknowledged that the use of suh a seletion algorithm is somewhat subjetive, but is preferred to tedious inspetion of synopti harts and plotted time series of wind speed and diretion. ay-by-day analysis of setions of the wind data indiated that the sea breeze seletion algorithm works extremely well for summer wind data, but is slightly less suessful in identifying sea breezes during the winter. Sea breeze ativity is subjet to strong seasonal variability (Figure 4). In summer (eember-february; southern hemisphere summer), approximately 2 sea breezes are experiened per month with speeds (at 15: hrs) of 6-7 m/s. In winter (June-August; southern hemisphere winter), around 12 sea breezes our per month with speeds of around 5 m/s. On average, 197 sea breezes are experiened eah year with a mean wind speed of 5.7 m/s. The diretion of the sea breeze is onsistently from the southwest and does not exhibit muh variability (standard deviations are less than 1 ). However, the summer sea breeze blows from a slightly more southerly diretion (24 ) than the winter sea breeze (25 ). The larger sea breeze diretion in winter (and also the larger standard deviations of the sea breeze speed and diretion) may indiate that oasionally westerly winds assoiated with the passage of mid-latitude depression have been misinterpreted as sea breezes. Hene, the number of identified sea breezes in the winter is probably over-estimated by. 4 sea breezes per month. Frequeny-omain Analysis of Perth Wind ata For eah year, spetra were omputed of wind speed data measured every three hours for the periods eember-february (summer), Marh-May (autumn), June-August (winter) and September-November (spring). Subsequently, an average spetrum was determined for eah of the seasons (Figure 5). All spetra show three main harateristis. Firstly, a large amount of energy is present at the low-frequeny end of the spetrum (periods longer than 1.5 days). These frequenies orrespond to the time sale of synopti weather patterns, suh as the passage of mid-latitude depression (storms). Seondly, a pronouned spetral peak an be found at the diurnal frequeny (period of one day), representing sea breeze ativity i]; WISEMAN et al., 1998), Thirdly, omplementary spetral peaks our at the first and seond har- Journal of Coastal Researh, Vol. 17, No.1, 21

6 .. E::J Effet of Sea Breeze in Western Australia A Q> 2 Z 1 if) 8 B <1l <1l... (/) C jan feb mar apr may jun jul aug sep ot nov de Month Figure 4. Seasonal variation in: (A) number of sea breezes; (B) sea breeze speed; and (C) sea breeze diretion. The vertial lines indiate the standard deviations assoiated with the averages. Based on three-hourly wind speed data olleted at Perth airport from Summer 1 3 Autumn... N-2. J!2 N E 1 2 ;:: 1 2 > Q> W t95% t95% Winter Spring t95% 1 1 ' '------'------' ' o 2 3 Frequeny (pd) 4 t95% 11 L ' L.- -"-- _ o 2 3 Frequeny (pd) 4 Figure 5. Average seasonal spetra of wind speed. The 95C;(, onfidene limits are indiated by the sale bar and were alulated using 294 degrees of freedom and "pd" refers to yles per day. Based on three-hourly wind speed data olleted at Perth airport from Journal of Coastal Researh, Vol. 17, No.1, 21

7 178 Masselink and Pattiarathi 1 8 A C\J (j) C\J s -,» Q) w , , , r-----, , ,----, ,----_,-,- 7 B 6 5 jan feb mar apr may jun jul aug sep ot nov de Month Figure 6. Monthly variation in: (A) spetral energy in the synopti (solid line; periods larger than 1.5 days) and sea breeze (dashed line; periods smaller than 1.5 days) frequeny band; and (B) relative ontribution of the sea breeze band to the total variane in the wind speed reord. Based on three-hourly wind speed data olleted at Perth airport from monies of the diurnal frequeny (periods of 12 and 6 hours, respetively). These are primarily due to the non-sinusoidal nature of the sea breeze signal. However, the ourrene of land breezes during the night and early morning may also have ontributed to the first harmoni frequeny (refer to Figure 2), The diurnal peak is highly signifiant and present in all seasonal spetra, but is widest in the summer spetrum and narrowest in the winter spetrum. With respet to the energy assoiated with the diurnal frequeny, and hene its importane, it is noted that the area under the spetral peak is of greater relevane than the atual peak value. Monthly spetra of the wind data were omputed and an average spetrum was determined for eah month. The total spetral energy was then partitioned into a low-frequeny band (periods longer than 1.5 days referred to as the 'synopti band') and a high-frequeny band (periods shorter than 1.5 days referred to as the 'sea breeze band') to allow investigation of the seasonal variation in the importane of these omponents (Figure 6). The spetral energy of the synopti band and the sea breeze band show an anti-phase relationship. The synopti band reahes its maximum energy levels in June/ July when storm ativity is prevalent, whereas the sea breeze band dominates from November-January when sea breezes are at their strongest and most abundant. The variane assoiated with the sea breeze band in summer is twie as large as in winter. In summer, the sea breeze band aounts for 6-7(}{ of the total variane in the wind reord, whereas in winter, less than 5(/ of the total variane an be asribed to the sea breeze band. Comparison of Perth Wind ata with Coastal Wind ata The Perth airport wind data were olleted approximately 2 km inland and may therefore not be representative for the sea breeze harateristis at the oastline. In partiular, the strength of the sea breeze an be expeted to derease landward due to frition. In addition, fritional effets (and perhaps the Coriolis fore) may alter the diretion of the sea breeze. To investigate the harateristis of the sea breeze in the oastal region, wind data olleted at the oastline (Swanbourne) and on Rottnest Island, loated 2 km off the oast of Perth, were investigated (refer to Figure 1), These data were olleted at 1-min intervals, but numerous gaps are present in the data, so attention is foused on one month of data olleted in January Both inland (Perth airport) and oastal (Swanbourne and Rottnest Island) wind data were onverted into hourly time series by means of linear interpolation and blok-averaging, respetively, for reasons of onsisteny. The sea breeze seletion algorithm was applied to the January 1994 data, whereby the diretion of the sea breeze at Perth airport was onsidered 19-3 and the oastal sea breeze diretion was taken as The number of reorded sea breezes at Perth airport, Swanbourne and Rottnest Island was 25,28 and 29, respetively. The smaller number of identified sea breezes at Perth airport is asribed to inorret wind diretion data at the end of the data reord, probably a result of a malfuntioning wind vane, beause sea breezes were learly present in the wind speed data, whereas Journal of Coastal Researh, Vol. 17, No.1, 21

8 '<, -- Effet of Sea Breeze in Western Australia (f) l 1 <ll <ll... (f) " 5 A "i B -./ / / \ \ \ (f) l 1 <ll <ll... (f) " 5 C ( / ,---.-.,-- / (f) l 1 <ll <ll... (f) C 5 E : <, /' Time (hrs) 9 F 6 / '-./ -/ : Time (hrs) 24 Figure 7. Ensemble-averaged daily time series of: (A) wind speed and (B) diretion at Perth airport; (C) wind speed and () diretion at Swanbourne; and (El wind speed and (F) diretion at Rott.nest Island. The solid line represents the mean and the dashed line indiates the mean one standard deviation. The data were olleted in January 1994 and only the days during whih a sea breeze was present at all three loations (23 days) were inluded in the ensemble averaging. the wind diretion remained onstant. The number of oastal sea breezes is therefore onsidered similar to that experiened at Perth airport. However, the assoiated veloities and diretion are signifiantly different. The mean sea breeze speed (at 15: hrs) at Perth airport, Swanbourne and Rottnest Island was 6.1 mis, 8.4 mls and 1.2 mis, respetively. In other words, the sea breeze at Swanbourne and Rottnest Island was 1.4 and 1.7 times stronger than at Perth airport, respetively. In addition, the mean sea breeze diretion at Swanbourne and Rottnest Island is 2, whereas at Perth airport the mean sea breeze diretion is 24. For those days in the data reord that a sea breeze was present at all three loations, ensemble-averaged daily time series of wind speed and diretion were omputed (Figure 7). In onjuntion with ross-spetral and ross-orrelation analysis (not shown) it appears that the time history of the sea breeze is similar for all three stations: (1) the onset of the sea breeze, as indiated by a shift in diretion from easterly (offshore) to southwesterly (obliquely-onshore) and an inrease in wind speed, ours within one hour aross all three measurement sites and generally ours around 13: hrs; (2) the maximum speed of the sea breeze is generally attained around 17: hrs; and (3) the essation of the sea breeze, indiated by a hange in wind diretion from south-southwest (obliquely-onshore) to south-southeast (obliquely-offshore), ours around 21: hrs. At Perth and Swanbourne, weak land breezes quikly develop following the essation of the sea breeze, however, relatively strong south-southeasterly winds prevail over most of the night on Rottnest Island. Analysis of Wind ata Colleted at Other Coastal Stations in Western Australia We have found the sea breeze to be a prominent phenomenon along most of the Western Australian oast. Wind data were olleted at nine additional oastal meteorologial loations (Figure 8). Most of these sites were loated some distane inland of the oastline at loal airports, exept for Oean Reef and Cape Leeuwin whih are diretly situated on the oast, and Abrolhos, whih is an island. One year of data (1995) was analyzed. ata were onverted to hourly data for reasons of onsisteny. Table 1 summarizes the time and frequeny domain analysis of the ten Western Australian oastal weather sites. In the northern region (erby, Broome, Karratha, Learmonth) the sea breeze is onshore and from the northwest. Obliquelyonshore and southwesterly sea breezes prevail in the entral region (Carnarvon, Abrolhos, Oean Reef, Perth). In the southern region (Cape Leeuwin and Esperane) the sea breeze blows predominantly onshore and from the southeast. The data were subjeted to the same analysis as the longterm Perth wind data and the number of sea breezes and the mean sea breeze wind speed were determined. Exept for Cape Leeuwin, the mean annual wind speed for all oastal stations is signifiantly less than the wind speed assoiated with the sea breeze. This indiates that at all loations the sea breeze represents a ondition that is more energeti than normal. The number of sea breezes that ourred in 1995 ranges from 137 (Cape Leeuwin) to 34 (Karratha), Journal of Coastal Researh, Vol. 17, No.1, 21

9 18 Masselink and Pattiarathi INIAN OCEAN 2 o Karrarha Learrnonth 25 WESTERN AUSTRALIA Abrolhos Is 3 Oean Re(/ Perth Esperane 15 C. Leuwin SOUTJ-IERN OCEAN Figure 8. Map of Western Australia with ten oastal meteorologial stations. Spetra of the wind speed were omputed for all oastal stations. All loations exept Cape Leeuwin exhibit a pronouned and signifiant spetral peak at the diurnal frequeny with assoiated harmonis (Figure 9). Partitioning of the total spetral energy into the low-frequeny synopti band (periods larger than 1.5 days) and the high-frequeny sea breeze band (periods smaller than 1.5 days) further demonstrates the dominane of the sea breeze for most of the oastal Table 1. Sumrnary of time and frequeny domain analysis ofi99s wind data olleted at ten oastal meteorologial stations in Western Australia. u = mean annual wind speed; N = number of sea breezes; U"b = mean wind speed during sea breeze at 15: hrs. rye of Variane in the Sea Breeze Band ominant (period smaller than 1.5 days) Sea Breeze iretion u N Usl> Sum. Aut. Win. Spr. Ann. erby Broome Karratha Learmonth Carnarvon Abrolhos Oean Reef Perth Cape Leeuwin Esperane Journal of Coastal Researh, Vol. 17, No.1, 21

10 Effet of Sea Bree ze in West ern Austra lia 181 1', , erb y 1' Broome ' Karratha 1' Learmonth ' Carnarv on 1' 1 1' 3 Oean Reef o 1', , Perth ' Q. ",!2 s: "'E ;:, 1» Q; W Frequeny (pd) Cape Leeuwin o o ', , Es perane Figur e 9. Average a nnua l spetra of' wind s peed for te n West ern Aust rali an oast a l st at ions. Th e 95% onfid en e limi ts are ind iated by th e sa le bar a nd we re alu lat ed usin g 24 degrees of freedom a nd "pd" refers to yles per day. Th e spetra were om pu te d usin g hourly wind speed dat a ollete d in sta tions. In gen eral, more than 5'Y,. of th e total variability in th e wind reord an be att ributed to th e sea breeze band. Th e only exeptions a re Cape Leeuwin and the Abrolhos Islands. At both sites, strong southerl y winds prevail throughout the yea r, with th e sea breeze only induing a modest inreas e in the wind speed and a slight hange in the diretion. EFFECT OF SEA BREEZE ON OFFSHORE AN INSHORE WAVE CONITIONS Generation of Wind Waves by the Sea Breeze Loal sea breeze ativity is expeted to have a signifiant imp at on th e inident wave s due to the generation of loal wind wave s that beome superimposed on the bakground wave field. For example, PRITCHETT (1976) noted a 1% inrease in visu ally esti ma ted breaking wav e heights reord ed in the late aftern oon, presumably as a result of sea br eeze ativity. Figure 1 illu strates the effet of th e three se a breeze yles shown in Figure 2 on th e offshore wave field measured south of Rottnest Island in 48 m water depth (refer to Figure 1) and is typial of summer oeanographi ondi tions. uring eah of the sea breezes, the sign ifiant wav e height (dete rmi ned using 4Vm o, where m., is th e zero th moment of the ene rgy spetru m) inre ased by.5 m, whereas the signifiant wave period (determined using m jm" where m, and m. are the zerot h and first mom ent of the ene rgy spetrum, respetively ) dereased by 5 s. The offshore wave ene rgy was partitioned in to swell (f <.15 Hz) and wind wave s (f >.15 liz) to more learly demonstrate the effet of sea breeze ativity. Over th e t hree- da y period, th e signifiant -Iourn al of Coas ta l Resear h, Vol. 17, No. 1, 2 1

11 1H2 Masselink and Pattiarathi I (f) I 2 A '-----.l _ 12 I 8 (f) I- 4 2 C ----, I I I(f) I /.J -'---I. J. _.4.3 N >. <.).2 - w tl.1 ' /3/95 8/3/95 9/3/95 Figure 1. Three-day time series of offshore wave data: (A) signifiant wave height H,,; (B) signifiant wave period T,; (C) signifiant wave height H, of swell (solid line; <.15 Hz) and sea breeze-generated wind waves (dashed line; >.15 Hz); and () frequeny-time spetrum. The ontour lines represent.1,.5 and 1 m 2/Hz and the intensity of the shading inreases with the spetral energy level. The spetra were alulated with 14 degrees of freedom. Wave data were olleted at 1 Hz for 8 minutes every 2 minutes south of Rottnest Island in 48 m water depth. The date labels on the x-axis are plaed at midnight (: hrs). swell height (determined using 4Vm o, where m., was omputed over f <.15 Hz) gradually inreased from.8 to 1.2 m, but was not affeted by sea breeze ativity. The signifiant wind wave height (determined using 4Vm o, where m., was omputed over f >.15 Hz), however, diretly responded to the daily sea breezes and attained maximum heights of m around midnight. The frequeny-time spetrum of the offshore wave energy shows the evolution of the wave field during sea breeze ativity. uring the sea breeze, a progressive inrease in the peak wave period of the wind waves ourred from 3-4 s to 5-6 s, while the peak swell period remained onstant at 12 s. At the end of the three-day period, the arrival of energeti swell and wind wave energy, unrelated to sea breeze ativity, is apparent in the frequeny-time spetrum. The effet of the same three sea breeze yles on the inshore wave field measured 5 km offshore in 17 rn water depth (refer to Figure 1) is shown in Figure 11. At this loation, sea breeze-generated wind waves with signifiant heights of up to.5 m were superimposed on a.3 m swell. uring the sea breezes, the overall signifiant wave height inreased by.2.3 m and the signifiant wave period dereased by 4 s. The measured offshore and inshore wave onditions during the sea breeze an be ompared with those predited using the Sverdrup-Munk-Bretshneider method (CERC, 1984), The wind speed at Perth Un'rlll is taken as a starting point Journal of Coastal Researh, Vol. 17, No.1, 21

12 .....L '----. _ L..._...J Effet of Sea Breeze in Western Australia.8 A I.4 (/) I 12 B I 8 - (/) r- O.8 r--'--'-"-'------""-', , , I.4 (/) I o L_..4.3 N».2 Q) :::l - Q) u:::.1 I <I 7/3/95 8/3/95 9/3/95 Figure 11. Same as Figure 1, but the ontour lines represent.1,.5 and.1 m:!/hz. The spetra were alulated with 14 degrees of freedom. Wave data were olleted offshore of Fremant.le in 17 m water depth. and three types of sea breezes are onsidered: (1) a weak sea breeze with U/'/,rth = 3 m/s; (2) an average sea breeze with UH.,.th = 5 m/s; and (3) a strong sea breeze with U/'ath = 7 m/s. The wind speeds at Swanbourne USII'C/flh and Rottnest Island U R ott fl are approximated by 1.4U/'rth and 1.7U/'rth' respetively (refer to Figure 7). It is assumed that the inshore waves are generated by wind speeds represented by US//'(/flh whereas the offshore waves are generated by wind speeds represented by URuttfl. It is further assumed that the duration of the sea breezes is 6 hours. The offshore wave onditions generated during sea breeze ativity are straightforward to predit beause due to the dominant alongshore omponent of the sea breeze the feth is virtually unlimited. It is predited that an average sea breeze (un'rtlt = 5 m/s) generates offshore waves with H; = 1.2 m and T; = 5.2 s (Table 2), This is in exellent agreement with the offshore wave observations shown in Figure 1 whih indiate H; = m and T; = 5-6 s. Predition of the inshore wave onditions is less straightforward beause it is not quite lear whether the wave generation proess is feth-limited by the presene of Garden Island and Five Fathom Bank (refer to Figure 1). In addition, the ontribution of refration and diffration of offshore waves into the inshore region to the inshore energy level is not known. Under the assumption of unlimited feth, an average sea breeze (U/'rth = 5 m/s) generates inshore waves with H; =.9 m and Til = 4.5 s (Table 2). However, the inshore wave measurements shown in Figure 11 are onsiderably less and indiate H; =.5 m and T; = 3-4 s. If a feth length of 15 km is assumed, orresponding to the approximate distane from the inshore wave rider buoy to Garden Island and Five Fathom Bank, predited wave onditions are H, =.5 m and T; = 3.2 s whih is in lose agreement with the observations. This strongly suggests that the inshore waves are feth-limited. Journal of Coastal Researh, Vol. 17, No.1, 21

13 IH4 Masselink and Pattiarathi Table 2. Inshore and offshore wave onditions generated by three types olsea breezes predited using the SverdruiiMunhBrtehneidr method (CERC, 1984). The duration o] the sea breezes is assumed to be 6 hours. Un'!llI = wind speed at Perth; U.'iU'1I111> =--= wind speed at Su-anbourne: U"'<l1I1I = uind speed at Rott.ne«i Island; H, signifiant uiaoe height; I'll spetral peak ioaoe period. Offshore Waves Inshore Waves Inshore Waves With Unlimited With Unlimited With a Feth of Feth Feth If) km UI'alll US"ulIl, 1I/{<l1I1I H, T/I J( 7'1' H, 7'/) (m/s) (m/s) (m/s) (rn ) (s) (rn ) (s) t m ) (s) : :3.5.: : :3.6 A omparison between the offshore and inshore wave onditions indiates that the inshore wave height was about half the offshore wave height, whereas the wave periods were rather similar i]; WNI, 1998). The redution in wave height was, however, dependent on the wave frequeny; the swell height dereased by 6(,Ir), whereas the redution in wind wave height was 5 l Yr.. The inshore oastal region off the oast of Perth is sheltered by submerged reef systems and islands (Five Fathom Bank, Rottnest and Garden Islands; refer to Figure 1) and the redution in swell height from offshore to inshore is primarily due to wave attenuation proesses (refration, diffration, bed fritional effets). The redution in wind wave height, however, is onsidered the result of the weaker wind speed and the shorter feth length assoiated with the sea breeze in the inshore oastal region. Quantifying the Effet of Sea Breeze Ativity on Inident Wave Conditions Sea breeze ativity has a signifiant effet on the offshore wave onditions and it is useful to be able to quantify this effet. One approah has been disussed above and uses spetral analysis of raw wave data to separate the total wave energy in swell and wind waves. Under the assumption that the wind waves are prinipally generated by the sea breezes, the importane of sea breeze ativity an then simply be expressed as the relative ontribution of wind wave energy to the total wave energy. For example, the ontribution of windwave energy (generated by the sea breeze) to the total wave energy for the entire three-day period shown in Figures 1 and 11 is 4% for the offshore region and 5(}{) for the inshore region. It is reasonable to assume that the majority of wind wave energy in summer is generated by sea breeze ativity. However, for the remainder of the year, winds assoiated with the passage of mid-latitude depressions (storms) should also be taken into aount, espeially in the winter months. In the analysis of the wind limate, the relative ontributions of storm and sea breeze ativity to the overall variane in the wind reord was assessed by omparing the amounts of enorgy present in the synopti (periods larger than 1.5 days) and sea breeze (periods smaller than 1.5 days) frequeny bands, respetively. A similar methodology an be applied to quantify the signifiane of sea breeze ativity for the nearshore wave limate. Spetral analysis was onduted on four years ( ) of hourly signifiant wave height data olleted in the offshore and inshore regions. For eah year, spetra were omputed for the periods eember-february (summer), Marh May (autumn), June-August (winter) and September-November (spring). Subsequently, an average spetrum was determined for eah of the seasons (Figure 12 L The spetra show onsiderably more offshore wave energy than inshore wave energy as a result of the attenuation of the inident wave energy by the offshore reef systems and islands. The offshore wave onditions experiene lear sea breeze foring in summer as indiated by a signifiant peak in the spetrum at a frequeny of one yle per day ii.e.. the diurnal frequeny). The spetra for the remainder of the year do not suggest that sea breezes signifiantly affet the offshore waves. The summer and autumn spetra of the inshore wave height exhibit a signifiant diurnal peak, and an insignifiant, albeit distint, diurnal peak in the spring spetrum. It is apparent that the sea breeze has a larger impat on the inshore waves than on the offshore waves. Monthly spetra of the offshore and inshore wave onditions were omputed and an average spetrum was determined for eah month using the four years of wave height data. The total spetral energy was then partitioned into the synopti and sea breeze frequeny bands to allow investigation of the seasonal variation in the importane of these omponents (Figure 13). espite the substantial wave heights (H, > 1 m) generated by the sea breeze in the offshore region (refer to Figure 1), diurnal foring of the offshore wave height only aounts for 1-2(Yt of the total variane in the wave height reord in the summer and. 5(lr for the remainder of the year. In ontrast, in the inshore region, the diurnal frequenies ontribute 2-4(lr to the total variane in the wave height reord in the summer, 1-2(;' in spring and autumn and. 5lir, in winter. In ontrast to the spetra of the wind speed (refer to Figure 6), the synopti band in the spetra of the wave height is of greater importane than the sea breeze band. This indiates that the variability in wave onditions due to loal and remote storm ativity is greater than the variations in wave height generated by sea breezes. CONCLUSIONS The diurnal sea breeze system in Perth, Western Australia, is a very important ontributor to the overall wind limate. Using forty-nine years of wind data olleted at Perth airport, 2 km inland, it was found that almost 2 sea breezes are experiened per year with an average wind veloity at 15: Journal of Coastal Researh, Vol. 17, No.1, 21

14 Effet of Sea Breeze in Western Australia 18il 1', , 1' Summer Autumn 1 / \ \ \ - \ \ 1 ' \ / ' -,, 1 ' 1' 1 ' 1 1' 1 1' Winter Spring 1 - \,, "" " 1 ' 1 Frequeny (pd) 1' 1 ' 1 1' Frequeny (pd) Figure 12. Average seasona l spetra of offshore Isolid line; 48 m wate r depth ) an d inshore rdas hod line; 17 m water dept.h i wave height. The Bil';' onfidene limits are indiated by the sa le bar and were alulated usin g 96 degrees of freedom and "pd" refer s to yles per day. Based on hourl y signifiant wave height ollet ed from 1B9fi-19B 8. 2 A >, 15 2' Q) Q) tii 1 ;j2 5 4,----,--,_--, --,------r--- r-,,_-_, -,_--, -,_-, 3 B 2 jan feb mar a pr may jun jul Month aug sep ot nov de Figure 13. Monthly va riation in the rela t ive ontrib utio n of th e sea bree ze ba nd rper iods sma ller th an 1.5 days) to the total va ria ne in the wave heigh t reord: IAI offshore 148 m water depth I; and IBI inshore 117 m water dep th ). Base d on hourl y signifiant wave height ollet ed from l B95-19f)8..lourn a l of Coasta l Researh, Vol. 17, No. 1, 2 1

15 186 Masselink and Pattiarathi hrs of 5.7 mjs and a mean diretion of 24. The sea breeze starts around mid-afternoon and usually subsides after 18: hrs. In summer, more than 2 sea breezes our on a monthly basis, but also in winter, sea breezes are not unommon. Throughout the year, more than halfthe variane in the wind speed reord an be attributed to the sea breeze frequeny band (periods less than 1.5 days). Along the oastline, sea breeze veloities are 1.4 times stronger than inland, and in addition, the sea breeze blows from a more southerly diretion (2 ). Some distane offshore on Rottnest Island (2 km offshore), the winds during the sea breeze are 1.7 times stronger than inland, but from the same diretion as on the oast. Analysis of wind data olleted at nine other oastal meteorologial stations in Western Australia indiate similar results, demonstrating that the sea breeze system is highly signifiant state-wide. The strength and persistene of the sea breezes has im portant impliations for the offshore wave limate in the area. Spetral analysis of wave data olleted off the oast of Perth revealed onsiderable amounts of spetral energy present within the sea breeze band, in partiular during summer when the sea breeze system is best developed. The importane of waves generated by the sea breeze is partiularly signifiant in regions that are proteted from the diret impat of swell. The inshore region of the Perth metropolitan oastline is partly sheltered and in the summer up to 4(/6 of the variability in the inshore wave height reord an be attributed to sea breeze ativity. The diurnal sea breeze system is a worldwide phenomenon that is partiularly prevalent along tropial and subtropial oasts, but is also ommonplae in temperate oastal environments (SIMPSON, 1994). In sheltered, low wave energy environments, sea breeze ativity has a signifiant impat on nearshore proesses and morphology, as demonstrated by the investigations onduted along the Perth metropolitan oastline (MASSELINK et al., 1997; PATTIARATCHI et al., 1997; MAS SELINK and PATTIARATCHI, 1998a, b). It is surprising, therefore, that only a handful of studies have been arried out onerning sea breeze effets in other parts of the world, namely the Gulf of California (INMAN and FILLOUX, 196), the Gulf of Mexio (SONU et al., 1973), the Caribbean (HEN RY, 1983; HUNTLEY et al., 1988) and Sri Lanka (PATTIAR ATCHI and MASSELINK, 1996). It is expeted that the importane of sea breeze ativity for nearshore environments is far greater than that suggested on the basis of the amount of attention it has reeived in the literature. ACKNOWLEGEMENTS We would like to thank the epartment of Transport, Fremantle, Western Australia, for providing the offshore and inshore wave data. Roshanka Ranasinghe onduted part of the wind data analysis. We would also like to thank the two reviewers for their omments and referenes to some further sea breeze literature. This is Centre for Water Researh publiation E1465GM. LITERATURE CITE ABBs,.J. and PHYSICK, W.L., Sea-breeze observations and modelling: A review. Australian Meteorologial Magazine, 41, BANTA, R.M.; OLIVIER, L.., and LEVINSON,.H., Evolution of the Monterey Bay sea-breeze layer as observed by pulsed oppler lidar. Journal o] Atrnosphri Sienes, 5, CERC, Shore Protetion Manual. U.S. Army Corps of Engineers, Coastal Engineering Researh Center, Waterway Experiment Station, Viksburg. CLARKE, R.H., Some observations and omments on the sea breeze. Australian Meteorologial Magazine, 11, AVIES, J.L., 198. Geographial Variation in Coastal erelopment. Seond Edition, Longman In., New York, 212p. AVIS, W.M.; SCHULTZ, L.C., and WAI{, R., 189. An investigation of the sea breeze. Harvard University Annals ofastronomial Observatory, 21, EPAUTMENT OF EFENCE, Australian National Tide Tables Australian Hydrographi Publiation 11, Canberra. FINI.ATER, J., Some aerial exploration of oastal airflow. Meteorologial Magazine, 92, GARRATT, J.R. and PHYSICK, W.L., The inland boundary layer at low latitudes: II sea-breeze influenes. Boundary Layer Meteorology, 33, HUNTLEY,.A.; HENI{Y, M..; HAINES, J., and GREENIGE, B., Waves and rip urrents on a Caribbean poket beah, Jamaia. Journal o] Coastal Researh, 4, HENRY, M.., The influene of the sea-land breeze regime on beah erosion and aretion: An example from Jamaia. Caribbean Geography, 1, HOUNAM, C.E., The sea breeze at Perth. Weather evelopment Researh Bulletin, 3, Hsu, S.A., Coastal Meteorology Aademi Press, 26p. Hsti, S.A., 197. Coastal air-irulation system: observations and empirial model. Monthly Weather Review, 98, INMAN,.L. and FILLOUX,J., 196. Beah yles related to tide and loal wind regime. Journal o] Geology, 68, JEHN, K.H., A Sea-Breeze Bibliography, Report No. 37, Atmospheri Siene Group, College of Engineering, The University of Texas, Austin, Texas. KEPERT, J.. and SMITH, R.K., A simple model of the Australian west oast trough. Monthly Weather Review, 12, KOTINIS-ZAMBAKAS, S.J.; NIKOLAKIS,.J., and KANELLOPOULOU, H.A., Analysis of absolute humidity by wind speed and diretion during land- and sea-breeze days at the National Observatory of Athens, Greee. Meteorologial Magazine, 118, LEMM, A.; HEGGE, B.J., and MASSELINK, G., Offshore wave limate, Perth (Western Australia): Marine and Freshwater Researh, 5, LAUGHLAN, G., The User:" Guide to the Australian Coast. New Holland. MAssELINK, G. and PATTIAI{ATCHI, C.B., 1998a. Morphodynami impat of sea breeze ativity on a beah with beah usp morphology. Journal of Coastal Researh, 14, MASSELINK, G. and PATTIARATCHI, C.B., 1998b. The effet of sea breeze on beah morphology, surf zone hydrodynamis and sediment resuspension. Marine Geology, 146, MASSELINK, G.; HEGGE, B.J., and PATTIARATCHI, C.B., Beah usp morphodynamis. Earth Surfae Proesses and Landforms, 22, NEUMANN, J., The Coriolis fore in relation to sea and land breezes-a historial note. Bulletin oj' the Amerian Meteorologial Soiety, 65, PATTIARATCHI, C.B. and MASSELINK, G., Sea breeze effets on nearshore oastal proesses. Proeedings 25th International Conferene on Coastal Engineering, ASCE, pp PATTIARATCHI, C.B.; HEGGI', B.J.; GOUL,J., and ELIOT, I.G., Impat of sea breeze ativity on nearshore and foreshore proesses in southwestern Australia. Continental Shelf Researh, 17, PRITCHETT, P.C., iurnal Variations in Visually Observed Breaking Waves. U.S. Army Coastal Engineering Researh Center, Misellaneous Report No. 76-8, 21p. SIMPSON, J.E., Sea Breeze and Loal Wind. Cambridge, Cambridge University Press, 234p. SONU, C.J.; MURRAY, S.P.; Hsu, S.A.; SUHAYA, J.N., and WA- Journal of Coastal Researh, Vol. 17, No.1, 21

16 Effet of Sea Breeze in Western Australia 187 ELL, E., Sea breeze and oastal proesses. Transations Amerian Geophysial Union, 54, STEEMAN, R., Colletion of' Wave ata and the Role of' Waves on Nearshore Cirulation. Report prepared for the Water Authority of Western Australia, Perth. TAPPEH, N. and HUHHY, L., Australias Weather Patterns: An Introdutory Guide. ellasta Pty. Ltd., Melbourne, 13p. WALLINCTON, C.E., 196. Convetion at the sea breeze front. Proeedings ofthe 2nd Conferene on Cumulus Convetion, Pergamon Press, WALLIN(n'ON, C.E., An introdution to the sea breeze front. Shwizer Aero-Revue, 36, WATSON, 1., 198. A ynami Climatology of' the Australian West Coast Trough. Unpublished Ph.. thesis, Geography epartment, University of Western Australia, Nedlands. WB:XLER, R., Theory and observations of land and sea breeze. Bulletin of the Amerian Meteorologial Soiety, 27, WISEMAN, W.J.; GRYMES, J.M., and SCHROEER, W.W., Coastal wind stress variability along the northern Gulf of Mexio. In: RONKERS, J. and SCHEFFERS, M.B.A.M. (editors), Physis of' Estuaries and Coastal Seas, Balkema, Rotterdam, pp WNI, SEA-ME-WE-3 S3 Oeanographi Study, Perth. WMI Siene and Engineering, Perth, Report No. R931..Iournal of Coastal Researh, Vol. 17, No.1, 21