Sea breeze characteristics on two sides of a shallow gulf: study of the Gulf St Vincent in South Australia

Transcription

1 METEOROLOGICAL APPLICATIONS Meteorol. Appl. 23: (2016) Published online 3 March 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: /met.1547 Sea breeze characteristics on two sides of a shallow gulf: study of the Gulf St Vincent in South Australia Zahra Pazandeh Masouleh,* David John Walker and John McCauley Crowther School of Civil, Environmental and Mining Engineering, The University of Adelaide, Australia ABSTRACT: A study of the long-term behaviour ( ) of sea breeze characteristics in the coastal city of Adelaide, Australia has been undertaken, applying a sea breeze day selection algorithm that employed four quantitative filters based on the temperature of the sea and land surfaces and the behaviour of surface- and upper-level winds. Despite the similarity of the daily surface wind circulation on non-sea breeze days on both sides of Gulf St Vincent, the westerly wind component of selected sea breeze days for Adelaide has been accompanied by an easterly wind direction on the opposite side of the Gulf, as would be expected. In the last three decades of the study period, where a higher frequency of data was available, the sea breeze has been shown to start at least 4 h after sunrise and normally cease before sunset and thus was of longer duration in summer than in the other seasons. It also appears that the maximum sea breeze intensity in summer months has a comparatively stronger southerly component than in other seasons. KEY WORDS sea breeze; selection algorithm; Adelaide Received 18 February 2015; Revised 23 September 2015; Accepted 5 October Introduction The phenomenon of coastal sea breeze has been noted from the earliest times. Aristotle, for example, planned the city of Alexandria in such a way that the alignment of the main streets made use of the cooling effect of the sea breeze while still providing shelter from other less favourable winds (Watson, 2005). The theory of sea breezes, however, has been studied extensively only from 1955, with the work by Clarke (1955) on sea breeze observations (Abbs and Physick, 1992). Conceptually, the sea breeze phenomenon can be simply explained as follows: due to different thermal and radiative properties, the land surface temperature increases more rapidly in the morning than the adjacent sea surface temperature, eventually producing an unstable temperature gradient over the land in the lower levels of the atmosphere. The ensuing convection drives a cooling breeze known as a sea breeze that blows initially from the sea towards the land (Simpson, 1994). Onset of the sea breeze is followed by a drop in the surface temperature and increase in the relative humidity (RH) of the land (Sumner, 1977). In the presence of a light gradient wind, the sea breeze forms a circulating flow with an offshore component, known as the return flow, in the upper levels of the atmosphere. The sea breeze system can extend to a height of 1000 m above mean sea level ( 900 hpa) on the South Australian coast (Stone, 1969; Finkele et al., 1995). Recently, improved observational and numerical techniques of this phenomenon have demonstrated the presence of Kelvin-Helmholtz billows within the sea breeze circulations (Plant and Keith, 2007). These turbulent disturbances develop during low static stability periods and have been detected in some * Correspondence: Z. P. Masouleh, School of Civil, Environmental and Mining Engineering, The University of Adelaide, North Terrace, Adelaide SA 5005, Australia. zpazandeh@civeng.adelaide.edu.au studies as close as 20 m from the surface (Britter and Simpson, 1978; Buckley and Kurzeja, 1997; Rao et al., 1999). Having a short wavelength of between 0.5 and 3 km (Sha et al., 1991) and a period of less than an hour, the detection of these billows in wind records requires a higher observational frequency than what is generally available from meteorological stations. In a coastal city, there is a potential for enhanced sea breeze circulation due to the urban heat island effect. Previous work has shown that this influence is governed by the combination of the city width and its distance from the coast (Yoshikado, 1994; Cenedese and Monti, 2003; Miller and Keim, 2003; Freitas et al., 2007). Considering the availability of local meteorological data, this study introduces a selection algorithm to detect potential sea breeze days since the start of the data recording in Adelaide, in South Australia, and investigates the characteristics of the wind on selected days at a location on the opposite side of the adjacent Gulf. The daily cycle of wind on sea breeze days is discussed later in the present study. Before the study site is described, it is useful to review the definition and identification of sea breezes as this is central to the present study. 2. Identification of sea breezes Sea breezes can be characterized by an intensified wind from the offshore direction in the afternoon. During the night, due to the greater cooling rate of the land, a reverse wind can blow from the land towards the sea, a land breeze. The dynamics of the sea breeze circulation are affected by synoptic conditions, and the presence of a moderate opposing gradient wind can modify the climatology of a sea breeze, delaying its inland penetration (Frizzola and Fisher, 1963). It has also been noted that a wind parallel to the shoreline can have an impact on the behaviour 2016 Royal Meteorological Society

2 Sea breeze characteristics on two sides of a shallow gulf 223 of the sea breeze (Azorin-Molina and Chen, 2009). Bigot and Planchon (2003), for example, argue that a strong opposing synoptic wind with a speed >6 8ms 1 can disrupt sea breeze generation. Identification of sea breeze days has been the objective of many previous studies (Stone, 1969; Borne et al., 1998; Furberg et al., 2002; Bigot and Planchon, 2003; Dunsmuir et al., 2003), and it is evident that the availability of meteorological data, and the topography and climate of the study area will help determine the accuracy of the selection method (Azorin-Molina et al., 2011b). In most of the sea breeze day selection methods, the surface wind characteristics have been taken into account as the presence of an intense offshore wind prevents the formation of a sea breeze, and so the diurnal reversal of wind direction from offshore to onshore was used as an identifier of a sea breeze day in the research carried out by many authors (Stone, 1969; Steyn and Faulkner, 1986; Pattiaratchi et al., 1997; Borne et al., 1998; Tijm et al., 1999; Furberg et al., 2002; Miller and Keim, 2003; Azorin-Molina and Chen, 2009). In addition to that, a rapid change in the intensity of wind was considered in some cases as an identifier (Masselink and Pattiaratchi, 2001; Azorin-Molina and Martín Vide, 2007; Azorin-Molina et al., 2011a). An abrupt drop in the surface temperature of the land along with a sudden increase in the RH can also be a feature of the onset of a sea breeze. Where data from recording hygrometers and thermometers are available, the formation of a sea breeze is observable. This method was introduced in studies by Stone (1969), Physick and Byron-Scott (1977), Physick and Byron-Scott (1977), Sumner (1977), Abbs (1986) and Azorin-Molina et al. (2011b). In the study by Borne et al. (1998), a steady-state condition of a 700 hpa wind was included in the sea breeze day criteria as the climate at this level was considered to be unaffected by the sea breeze circulation, whereas Dunsmuir et al. (2003) included the change of a 900 hpa wind as an indicator for sea breeze occurrence as the sea breeze has been observed to extend to this level. A day with a sky cover condition known as broken sky (cloud cover >5/8) is assumed to reduce the probability of sea breeze occurrence as it slows down the heating process of the land. This factor, as employed in studies by Borne et al. (1998), Furberg et al. (2002), Miller and Keim (2003) and Steyn (2003), has been used as a selection criterion to eliminate the days with low potential for sea breeze occurrence. In practice, researchers use a combination of several of these identifiers. For example, Borne et al. (1998) used a set of six filters to determine a sea breeze day, considering the change in the characteristics of the upper air-level and surface-level wind. They attempted to evaluate the accuracy of the method by comparing the days they identified as sea breeze days with the observationally based selection by an experienced researcher from the Department of Geography and Physics, University of Lund, Sweden. Although the activities associated with the start of a sea breeze were neglected in their method, the comparison showed an agreement of >75%. Furberg et al. (2002), followingthe workof SteynandFaulkner (1986), selected sea breeze days as those days with a higher temperature over the land as the main factor together with a dominant offshore wind during the night and early morning, and an onshore wind during the daytime. Azorin-Molina et al. (2011b) introduced two selection algorithms where each was based on different aspects of sea breeze climatology. Their manual technique considered the shift in the wind direction as well as changes in temperature and RH for a sea breeze occurrence, while their computer-based method examined different features of sea breeze processes over a day. Their methods were evaluated using the results of previous independent researchers. The manual approach results were compared with the output of a method developed by a researcher from the Regional Climate Group ( at the University of Goteborg (Sweden), while the computer-based approach was examined against a selection technique developed by Prtenjak and Grisogono (2007). In the above studies of sea breeze identification, all the selection techniques were evaluated using the results of independent methods; however, for small water bodies such as lakes and gulfs, there have been observations of the simultaneous occurrence of the sea breeze on the opposite side of the water (Moroz, 1967; Physick and Byron-Scott, 1977), which can be used to test the accuracy of the selection method. One of the conclusions reached by Azorin-Molina et al. (2011b) is that the selection of sea breeze days is strongly related to the criteria employed. For a study where long-term patterns and behaviour are being sought, this is an issue that has been addressed by employing strict criteria for the recognition of a sea breeze and maintaining them over the length of the study. 3. The study site The city of Adelaide, South Australia, with a population of just over 1.2 million, is located in southern Australia (see Figure 1(a)). European settlement of the city started in the 1830s with extensive population growth in the 1920s. The climate is often described as Mediterranean, with warm, mostly dry summers and mild, wet winters (Erell and Williamson, 2007). The city is located on the eastern shore of Gulf St Vincent with a shoreline that is oriented approximately north-northwest/south-southeast. Gulf St Vincent extends over a distance of 170 km, and at the deepest point, reaches 40 m depth, while in the northernmost third, it is shallower than 15 m (Kaempf, 2006). Mount Lofty Ranges (Figure 1(b)) on the eastern side and the Adelaide Hills to the southeast of the metropolitan area are the main barriers to sea breeze penetration towards the south and the east. In winter, the typical prevailing wind is westnorthwesterly, and so overnight and early morning surface winds tend to be blocked by the Mount Lofty Ranges; the surface flow then turns parallel to the Ranges and flows northeast over the plains (Tepper and Watson, 1990). In summer, the prevailing synoptic flow is southeasterly, and gully winds at nights and early mornings on the western flanks of the Mount Lofty Ranges are common and are attributed to the following: 1. non-linear amplifications by either a hydraulic or wave resonance mechanism occurs usually in the presence of an elevated inversion and/or critical layer (a wind reversal) at about m; 2. a shallow, stable air mass formed by a sea breeze from the Southern Ocean that moves inland during the day on the eastern side of the Mount Lofty Ranges. This air mass is trapped inertially on the eastern flank of the Ranges and builds in sufficient height overnight to cascade down the western slopes and gullies as a gravity current (Sha et al., 1995; Grace, 1995). As Physick and Byron-Scott (1977) demonstrated, the afternoon wind of sea breeze days in the Adelaide region is a combination of a sea breeze from Gulf St Vincent (gulf breeze)

3 224 Z. Pazandeh Masouleh et al. (a) (b) Figure 1. (a) Map of the study area (Google Earth), (b) topography of the Adelaide plain and Mount Lofty Ranges (from GEODATA 9 Second Digital Elevation Model). and a breeze from the ocean, known as a continental breeze. The resultant wind has a south-southwest direction (Figure 1(a)). Being located at 35 S, and due to the presence of Coriolis forces, Adelaide experiences comparatively stronger sea breezes with greater lateral extent than locations at higher latitudes (Yan and Anthes, 1987). The breeze from the Southern Ocean reaches the Adelaide coastline during the late afternoon. The deflection of the maximum sea breeze to the left is partially due to the Coriolis force (Ahrens et al., 2012). The sea breeze ceases before the dissipation of the land sea temperature gradient because in the surface layer, the wind speed is reduced by frictional forces (Haurwitz, 1946). Owing to the presence of frictional drag, wind speed at the surface level is reduced by up to 40% and 60% of the geostrophic wind over the ocean and land, respectively, with a deflection of in the direction. White et al. (1992) have also suggested that due to the presence of a friction difference over the land and the sea, the wind hodograph in sea breeze situations would have a more eccentric shape with a larger angle from the shoreline direction. In the current study, these features have been observed on the hodograph of average sea breeze cases. It has also been noted that locally generated sea breezes can occur at different locations in Gulf St Vincent, and these will have a similar effect on the afternoon wind regime in Adelaide, on the opposite side of the Gulf. To observe this, the characteristics of the wind at the coastal town of Edithburgh have been included in this study. The location of Edithburgh is also shown in Figure 1(a). 4. The sea breeze identification algorithm To identify the sea breeze days for Adelaide, a set of criteria (or filters) was developed to capture the possibility of a sea breeze development over the ocean and its subsequent penetration inland. It is worth noting that the selection algorithm has been applied to the meteorological records of the Adelaide Airport station, and the days that satisfied all the criteria are the days on which the fully developed sea breeze arrived at Adelaide s shores. Moreover, as the availability of meteorological data, and the topography and climate of the study area determine the accuracy of the algorithm, the selection of sea breeze days is strongly related to the criteria employed. Because the analysis of long-term variations in sea breeze characteristics was the main objective of the study, it was necessary to obtain consistent meteorological observations for a longer time with the highest possible temporal resolution. The Australian Bureau of Meteorology supplied meteorological data from 1955 for Adelaide Airport, which were used in the present study, including 3 h observations of wind speed and direction (10 m above ground level) and air temperature (1.2 m above ground level) and 6 h observations of the weather at upper air levels (using radiosondes). The sea breeze is considered to be blowing from the sector 190 to 310 (Crooks and Brooks, 1987). The following filters were introduced and applied to the Adelaide Airport meteorological data to select a sea breeze day: 1. presence of a positive temperature difference between the land and sea surfaces; 2. absence of a wind speed at 700 hpa with an offshore component >7.5 m s 1 in the early afternoon, between 1200 and 1400 CST (0230 and 0430 UTC), respectively (Australian Central Standard Time); 3. one of the following conditions in surface wind observations: presence of a calm condition or an offshore wind in the early morning with a rotation to the sea breeze sector in

4 Sea breeze characteristics on two sides of a shallow gulf 225 the afternoon followed by a calm or offshore wind in the evening and late night; or on days where morning or evening winds are predominantly a light breeze from the sea breeze sector, the afternoon wind speed should exceed 1.5 m s 1 ; 4. moreover, for a day to be considered a sea breeze day, the afternoon wind direction was required to be from the sea breeze sector for two subsequent readings (at least 3 h flow). Although the vertical temperature gradient over the land initiates the formation of sea breeze circulation, the presence of a temperature difference between the sea and land surfaces is the fundamental driver of sea breeze development; therefore, greater temperature over the land has been included as the first condition for a day to be selected as a sea breeze day. The temperature at the Adelaide Airport station was used as the land surface temperature. The more inland station of Kent Town ( 8.5 km further inland) has shown an average temperature difference of 0.64 C higher than that of the airport (since the start of its operation in 1977). The second criterion excludes the days with strong opposing gradient winds (Finkele, 1998) and considers only the days for which there is a greater potential for a sea breeze to form. The last filter considers the fact that the main characteristic of a sea breeze is the rotation of the wind to the onshore direction accompanied by an increase in the wind speed. Filters 1, 3 and 4 have been applied to the surface wind observations. It should be noted that the current algorithm detects only days with fully developed sea breezes in the afternoon and does not capture any complex behaviour in the sea breezes. Moreover, as the wind data were from a meteorological station located along the Adelaide shore line, further penetration of onshore wind was not considered in the selection process Sea surface temperature The records from the Advanced Very High Resolution Radiometer (AVHRR) from National Oceanic and Atmospheric Administration (NODC, 2009), introduced by Townshend (1994), were included to consider the surface temperature at longitude E and latitude 35 S in Gulf St Vincent. The data collection started in 1981, but for the period 24 August December 2008, there were only 2081 records available ( 21%). In order to produce an estimated value for sea surface temperature, the data were interpolated using Matlab s piecewise cubic interpolation spline and Piecewise Cubic Hermite Interpolating Polynomial (PCHIP). The results were averaged for each Julian day over the period and were considered the sea surface temperature for each Julian day from 1956 (Figure 4). The correlation between available data and estimated values was 0.98 with a standard error of 0.65 C Method assumptions Wind speeds of <2ms 1 were taken as a light breeze (Beaufort scale); therefore, weaker afternoon winds were ignored in the selection process. An increase in velocity of >1.5 m s 1, as explained by Azorin-Molina and Martín Vide (2007), was applied in previous sea breeze studies. Owing to the low frequency of observations (3 h surface and 12 h upper air-level records), the selection method is likely to underestimate the frequency of sea breeze days as some potential sea breeze days with shorter duration might not be detected. However, the selected cases provide a set of days with fully developed sea breeze conditions that continue for at least 3 h Characteristics of the wind on the opposite shoreline of Gulf St Vincent It has been observed by Physick and Byron-Scott (1977) that the afternoon sea breeze on the western side of the gulf comprises a southsoutheasterly continental sea breeze and an eastnortheasterly local gulf breeze. Therefore, sea breeze and non-sea breeze days based on the selection algorithm using Adelaide data were examined and evaluated using the wind data from the meteorological station at Edithburgh, located on the western side of Gulf St Vincent. As regular observations of suitable frequency at this site started from 1993, the corresponding period was examined for both sites m s m s Edithburgh Adelaide Airport km 10 mi Figure 2. Hodograph of wind speed vector on sea breeze (full line) and non-sea breeze days (dotted line) for both sides of the gulf ( ). The circles show the wind speed at the interval of 1 m s 1, and the numbers denote the local standard time. An example of a wind vector at 1800 CST on non-sea breeze days is illustrated.

5 226 Z. Pazandeh Masouleh et al (a) U-component (c) V-component Sea breeze days Wind speed (m s 1 ) Non-sea breeze days Wind speed (m s 1 ) (b) (d) Local standard time Local standard time Figure 3. Average sea breeze days U-component (a)andv-component (c)ofadelaide Airport (line)andedithburgh (dashline).same fornon-sea breeze days (b) and (d), respectively. The wind hodographs for average sea breeze and non-sea breeze days are shown in Figure 2, where each point on the hodograph refers to the end of the vector of the average wind speed at that time. The hodographs were plotted over the map outline of the area to make comparisons and evaluations easier. The average was calculated for each component of the wind over the period of 16 years for sea breeze and non-sea breeze days separately. There is an evident difference between the behaviour of the sea breeze and non-sea breeze days at Edithburgh. It is worth noting that the sea breeze days for all locations have been chosen based on the application of the detection filters to the Adelaide Airport observations. The local circulation of wind on non-sea breeze days is similar for Adelaide and Edithburgh, whereas the sea breeze day winds show a different characteristic at the Edithburgh station. Figure 3 shows the averaged west-to-east (U) and south-to-north (V) components at Edithburgh and Adelaide, respectively, for sea breeze and non-sea breeze days, as selected for Adelaide. For sea breeze days, the night time U-component (between and the next day) is from the east (i.e. negative) for both locations but stronger at Edithburgh due to its more exposed location. During the early part of the sea breeze day, the U-component changes direction from offshore to onshore and increases in strength at the Adelaide Airport, whereas at Edithburgh, the U-component becomes more negative and more onshore as expected for a sea breeze on the opposite side of Gulf St Vincent (Figure 3(a)). Later in the sea breeze day, the U-component weakens in magnitude at both locations, becoming offshore at Adelaide but not at Edithburgh. This difference may be explained by the relative strength of sea breezes and land breezes on the larger land mass (Xian and Pielke, 1991) east of Adelaide compared with the narrow peninsula to the north and west of Edithburgh. The V-components have a similar pattern at both locations but are stronger at Edithburgh due to its greater exposure to the southerlies (Figure 3(c)). The V-component on sea breeze days becomes negative at Adelaide but not at Edithburgh, probably again because connected with the greater local land mass and the offshore land breeze at Adelaide. The fact that the observed sea breeze direction is not perpendicular to the shoreline is due to the location of the gulf in the vicinity of the ocean and the presence of the continental sea breeze coming from the south, which shifts the resultant wind southerly as has been previously noted by Physick and Byron-Scott (1977). For non-sea breeze days, the U-components at both Edithburgh and Adelaide are similarly positive and at peak during the afternoon, albeit more strongly at Adelaide because it is less sheltered from the westerlies than Edithburgh (Figure 3(b)). The V-components for non-sea breeze days have a similar pattern at both locations but are stronger at Edithburgh due to its greater exposure to the southerlies (Figure 3(d)). Thus, the hodographs and wind components at Adelaide and Edithburgh show quite similar behaviour on non-sea breeze days but are significantly different on sea breeze days due to Edithburgh s location on the opposite side of Gulf St Vincent. The differences are consistent with the circulation patterns expected on a sea breeze day at Adelaide. Other differences may be explained by the variations in exposure and local land mass Sea breeze frequency The selection algorithm was applied to the data set using Visual Basic codes developed for this study. As there were periods of missing data in the observational records that were used in the detection method, the sea breeze cases for each time period are presented as the percentage of sea breeze occurrence.

6 Sea breeze characteristics on two sides of a shallow gulf % (a) 60% 50% 40% 42% 30% 30% 24% 27% 20% 10% 10% (b) 0% Spring Summer Autumn Winter Year Figure 4. The frequency of sea breeze event for each season; the box shows the lower and the upper quartiles. For the period of study, August 1955 June 2008, which included 95% data coverage for surface- and upper air-level observation, 26.6% of the days (4893 days) were identified as sea breeze days. The percentages of sea breeze cases in each season are plotted in Figure 4. Seasons are defined as follows: March May as autumn (fall), June August as winter, September November as spring and December February as summer. As expected, summer months have the most sea breeze days with an average of 42% and a maximum of 58% of the days, while on average only 10% of winter days were observed to have sea breezes Daily cycle of wind in a sea breeze day The frequency of meteorological observations in Adelaide Airport has changed since 1985 from 3 h to 30 min, allowing the analysis of the start and cessation times of sea breeze more accurately for the period of Therefore, the time that onshore wind intensity increases or an offshore wind shifts to the sea breeze sector (from 180 to 320 ) has been considered the start of the sea breeze, and the rotation of wind to offshore direction or a decline of its strength has been assumed as its end. The characteristic parameters of a sea breeze day, i.e. the average time of start, end and maximum sea breeze intensity, are plotted in Figure 5. As the number of sea breeze cases in winter months (June, July and August) is relatively low (<10% of the winter days), these months were excluded from the figures. Sea breezes start comparatively later in autumn than in summer and spring, which is related to 1 h difference in the time of sunrise and the fact that, compared to spring, the Gulf water remains warmer in autumn. It is also observed that the summer sea breezes cease relatively later than in the other two seasons, attributed to the 14 h duration of daytime compared with autumn and spring with 11.3 and 12.9 h, respectively. While most sea breezes of spring and autumn reach their maximum intensity sometime between 1200 and 1600 CST, there have been 22% of cases in summer where the maximum sea breeze reaches Adelaide Airport after 1600 CST. The most common time of sea breeze start, end and maximum for each month are plotted in Figure 6. For each month (excluding winter), the maximum sea breeze intensity was averaged over the period of The result suggests a maximum wind velocity of 7.1 m s 1 in January to 4.9 m s 1 in May, while the direction from which the wind blows (c) Figure 5. The frequency of time of sea breeze start (a), cessation (b) and maximum (c) for at Adelaide Airport station. Time of day January February March April May September October November December Months Figure 6. Monthly average of sea breeze start, end and maximum for for Adelaide Airport station. is constantly from the southwest (varying from 225 in January to 256 in September). The maximum wind velocity and corresponding direction are shown in Figure 7. It is evident from Figure 7 that the maximum wind intensity on sea breeze days has a considerably stronger southerly component during the warmer months of December February, whereas in September, the maximum wind blows from the direction perpendicular to the shore line. 5. Summary and Conclusions Following earlier research on sea breeze identification (Stone, 1969; Borne et al., 1998; Furberg et al., 2002; Bigot and Planchon, 2003; Dunsmuir et al., 2003), the present research has investigated and introduced a new algorithm to detect previous

7 228 Z. Pazandeh Masouleh et al. January February March November December April May September October Figure 7. Average of maximum sea breeze direction and speed for at Adelaide Airport station. The numbers on the arrows are the wind speed in m s 1. sea breeze days on the coastline of Adelaide in South Australia from the long-term record of near-surface and upper-air meteorological data from 1957 to Over the 51 years of the study period, the average percentage of sea breeze days in summer was 42% and in winter was 10% compared with an annual average of 27%. The characteristics of the winds on select days were compared for coastal sites on both sides of the Gulf as a separate check on the selection process. As the selection algorithm was applied to the observations at the Adelaide Airport, it represents the sea breeze days in Adelaide. Despite the similarity in wind hodographs of non-sea breeze days at both Adelaide and Edithburgh sites, the characteristics of the U-component of the wind at Edithburgh on sea breeze days are very different. This provides further reassurance that the proposed set of filters does indeed select the days with a locally generated gulf breeze and rejects synoptic-scale changes. References Abbs DJ Sea-breeze interactions along a concave coastline in Southern Australia: observations and numerical modeling study. Mon. Weather Rev. 114: Abbs DJ, Physick WL Sea breeze observation and modelling: a review. Aust. Meteorol. Mag. 41: Ahrens CD, Jackson PL, Jackson CEJ Meteorology Today: An Introduction to Weather, Climate, and the Environment. Nelson Education: Toronto, Canada. Azorin-Molina C, Chen D A climatological study of the influence of synoptic-scale flows on sea breeze evolution in the Bay of Alicante (Spain). Theor. Appl. Climatol. 96: Azorin-Molina C, Chen D, Tijm S, Baldi M. 2011a. A multi-year study of sea breezes in a Mediterranean coastal site: Alicante (Spain). Int. J. Climatol. 31: Azorin-Molina C, Martín Vide J Methodological approach to the study of the daily persistence of the sea breeze in Alicante (Spain). Atmósfera 20: Azorin-Molina C, Tijm S, Chen D. 2011b. Development of selection algorithms and databases for sea breeze studies. Theor. Appl. Climatol. 106: Bigot S, Planchon O Identification and characterization of sea breeze days in Northern France using singular value decomposition. Int. J. Climatol. 23: Borne K, Chen D, Nunez M A method for finding sea breeze days under stable synoptic conditions and its application to the Swedish west coast. Int. J. Climatol. 18: Britter RE, Simpson JE Experiments on the dynamics of a gravity current head. J. Fluid Mech. 88: Buckley RL, Kurzeja RJ An observational and numerical study of the nocturnal sea breeze. part ii: chemical transport. J. Appl. Meteorol. 36: , doi: / (1997) 036<1599:AOANSO>2.0.co;2. Cenedese A, Monti P Interaction between an inland urban heat island and a sea-breeze flow: a laboratory study. J. Appl. Meteorol. 42: Clarke RH Some observations and comments on the sea breeze. Aust. Meteorol. Mag. 11: Crooks G, Brooks B Seafront Wind near Adelaide. Bureau of Meteorology, Meteorological Note; 176, 14 pp. [Available from Bureau of Meteorology, GPO Box 1289K, Melbourne, 3001, Australia.] Dunsmuir WTM, Spark E, Kim S-K, Chen S Statistical prediction of sea breezes in Sydney harbour. Aust. Meteorol. Mag. 52: Erell E, Williamson T Intra-urban differences in canopy layer air temperature at a mid-latitude city. Int. J. Climatol. 27: Finkele K Inland and offshore propagation speeds of a sea breeze from simulations and measurements. Bound.-Lay. Meteorol. 87: Finkele K, Hacker J, Kraus H, Byron-Scott RD A complete sea-breeze circulation cell derived from aircraft observations. Bound.-Lay. Meteorol. 73: Freitas E, Rozoff C, Cotton W, Dias PS Interactions of an urban heat island and sea-breeze circulations during winter over the metropolitan area of São Paulo, Brazil. Bound.-Lay. Meteorol. 122: Frizzola JA, Fisher EL A series of sea breeze observations in the New York city area. J. Appl. Meteorol. 2: Furberg M, Steyn DG, Baldi M The climatology of sea breezes on Sardinia. Int. J. Climatol. 22: Grace W Downslope winds. Meteorological Study No. 41, Bureau of Meteorology; 83 pp. Haurwitz B Comments on the sea breeze circulation. J. Meteorol. 4: 1 8. Kaempf J In-situ field measurements for Adelaide Coastal Waters Study. Final Technical Report, Adelaide Coastal Waters Study: Adelaide, SA. Masselink G, Pattiaratchi CB Characteristics of the sea breeze system in Perth, Western Australia, and its effect on the nearshore wave climate. J. Coast. Res. 17: Miller STK, Keim BD Synoptic-scale controls on the sea breeze of the Central New England Coast. Weather Forecast. 18: Moroz WJ A lake breeze on the eastern shore of Lake Michigan: observations and model. J. Atmos. Sci. 24: National Oceanic Data Center (NODC) and Rosenstiel School of Marine and Atmospheric Science AVHRR Pathfinder Level 3 8 Day Nighttime SST Version 5.1. PO.DAAC, CA /PATHF-8DN51 (accessed 20 August 2014). Pattiaratchi C, Hegge B, Gould J, Eliot I Impact of sea-breeze activity on nearshore and foreshore processes in southwestern Australia. Cont. Shelf Res. 17: Physick WL, Byron-Scott RAD Observations of the sea breeze in the vicinity of a Gulf. Weather 32: Plant RS, Keith GJ Occurrence of Kelvin-Helmholtz billows in sea-breeze circulations. Bound.-Layer Meteorol. 122: 1 15.

8 Sea breeze characteristics on two sides of a shallow gulf 229 Prtenjak MT, Grisogono B Sea/land breeze climatological characteristics along the northern Croatian Adriatic coast. Theor. Appl. Climatol. 90: Rao PA, Fuelberg HE, Droegemeier KK High-resolution modelling of the Cape Canaveral land-water circulations and associated features. Mon. Weather Rev. 127: Sha W, Kawamura T, Ueda H Numerical study on sea/land breezes as a gravity current: Kelvin-Helmholtz billows and inland penetration of the sea-breeze front. J. Atmos. Sci. 48: Sha W, Physick W, Grace W A numerical experiment on the Adelaide Gully Wind of South Australia. Math. Comput. Model. 21(9): Simpson JE Sea Breeze and Local Wind. Cambridge University Press: New York, NY; 234 pp. Steyn DG Scaling the vertical structure of sea breezes revisited. Bound.-Lay. Meteorol. 107: Steyn DG, Faulkner DA The climatology of sea breezes in the lower Fraser Valley, B. C. Climatol. Bull. 20: StoneDJ Sea Breezes across Adelaide. Department of Geography, University of Adelaide: Adelaide, SA; 128 pp. Sumner GN Sea breeze occurrence in hilly terrain. Weather 32: Tepper G, Watson A The wintertime nocturnal northeasterly wind of Adelaide, South Australia; an example of topographic blocking in a stably-stratified air mass. Aust. Meteorol. Mag. 38: Tijm ABC, Holtslag AAM, van Delden AJ Observations and modeling of the sea breeze with the return current. Mon. Weather Rev. 127: Townshend JRG Global data sets for land applications from the advanced very high resolution radiometer: an introduction. Int. J. Remote Sens. 15: Watson P Ideas: A History from Fire to Freud. Phoenix: London, UK; 1118 pp. White ID, Mottershead DN, Harrison SJ Environmental Systems: An Introductory Text. Chapman & Hall: Cheltenham, UK. Xian Z, Pielke RA The effects of width of landmasses on the development of sea breezes. J. Appl. Meteorol. 30: Yan H, Anthes RA The effect of latitude on the sea breeze. Mon. Weather Rev. 115: Yoshikado H Interaction of the sea breeze with urban heat islands of different sizes and locations. J. Meteor. Soc. Jpn. 72: