The Australian Air Quality Forecasting System. Part II: Case Study of a Sydney 7-Day Photochemical Smog Event

G. D. Hess Bureau of Meteorology Research Centre, Melbourne, Victoria, Australia

Search for other papers by G. D. Hess in
Current site
Google Scholar
PubMed
Close
,
K. J. Tory Bureau of Meteorology Research Centre, Melbourne, Victoria, Australia

Search for other papers by K. J. Tory in
Current site
Google Scholar
PubMed
Close
,
M. E. Cope CSIRO Atmospheric Research, Aspendale, Victoria, Australia
CSIRO Energy Technology, Newcastle, New South Wales, Australia

Search for other papers by M. E. Cope in
Current site
Google Scholar
PubMed
Close
,
S. Lee CSIRO Atmospheric Research, Aspendale, Victoria, Australia

Search for other papers by S. Lee in
Current site
Google Scholar
PubMed
Close
,
K. Puri Bureau of Meteorology Research Centre, Melbourne, Victoria, Australia

Search for other papers by K. Puri in
Current site
Google Scholar
PubMed
Close
,
P. C. Manins CSIRO Atmospheric Research, Aspendale, Victoria, Australia

Search for other papers by P. C. Manins in
Current site
Google Scholar
PubMed
Close
, and
M. Young Department of Environment and Conservation (NSW), Lidcombe, New South Wales, Australia

Search for other papers by M. Young in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The performance of the Australian Air Quality Forecasting System (AAQFS) is examined by means of a case study of a 7-day photochemical smog event in the Sydney region. This was the worst smog event for the 2000/ 01 oxidant season, and, because of its prolonged nature, it provided the opportunity to demonstrate the ability of AAQFS to forecast situations involving recirculation of precursors and remnant ozone, fumigation, and complex meteorological dynamics. The forecasting system was able to successfully predict high values of ozone, although at times the peak concentrations for the inland stations were underestimated. The dynamics for the Sydney region require a sensitive balance between the synoptic and mesoscale flows. Often high concentrations of ozone were advected inland by the sea breeze. On two occasions the system forecast a synoptic flow that was too strong, which blocked the inland advancement of the sea breeze. The peak ozone forecasts were underpredicted at the inland stations on those occasions. An examination of possible factors causing forecast errors has indicated that the AAQFS is more sensitive to errors in the meteorological conditions, rather than in the emissions or chemical mechanism in the Sydney region.

Corresponding author address: G. D. Hess, Bureau of Meteorology Research Centre, GPO Box 1289K, Melbourne VIC 3001, Australia. d.hess@bom.gov.au

Abstract

The performance of the Australian Air Quality Forecasting System (AAQFS) is examined by means of a case study of a 7-day photochemical smog event in the Sydney region. This was the worst smog event for the 2000/ 01 oxidant season, and, because of its prolonged nature, it provided the opportunity to demonstrate the ability of AAQFS to forecast situations involving recirculation of precursors and remnant ozone, fumigation, and complex meteorological dynamics. The forecasting system was able to successfully predict high values of ozone, although at times the peak concentrations for the inland stations were underestimated. The dynamics for the Sydney region require a sensitive balance between the synoptic and mesoscale flows. Often high concentrations of ozone were advected inland by the sea breeze. On two occasions the system forecast a synoptic flow that was too strong, which blocked the inland advancement of the sea breeze. The peak ozone forecasts were underpredicted at the inland stations on those occasions. An examination of possible factors causing forecast errors has indicated that the AAQFS is more sensitive to errors in the meteorological conditions, rather than in the emissions or chemical mechanism in the Sydney region.

Corresponding author address: G. D. Hess, Bureau of Meteorology Research Centre, GPO Box 1289K, Melbourne VIC 3001, Australia. d.hess@bom.gov.au

Introduction

The Australian Air Quality Forecasting System (AAQFS) currently issues 36-h forecasts 2 times per day for two domains—one that encompasses the state of Victoria, and the other most of New South Wales (NSW). The components of the system and measures of statistical verification for it are described by Cope et al. (2004). Briefly, the forecasting system consists of the emissions inventory of the Department of Environment and Conservation (NSW) (hereinafter DEC), in which some components are modeled based on forecast winds and temperatures; the Bureau of Meteorology's operational limited-area forecast model, with 0.05° horizontal resolution and 29 vertical levels; and the Commonwealth Scientific and Industrial Research Organization (CSIRO) specially built chemical transport model with 0.05° horizontal resolution for rural areas, 0.01° resolution for urban areas, and 17 vertical levels. The photochemical mechanism is a highly compressed scheme with seven reactions.

In addition to statistical verification, the performance of the AAQFS has been examined through case studies (Cope et al. 1998, 1999; Hess et al. 2000a,b; Tory et al. 2000). In this paper we explore a photochemical smog case study for the NSW domain, and in Tory et al. (2004) a case study for the Victorian domain is presented.

In January 2001 the Sydney region experienced a prolonged period, lasting seven days, of high ozone concentrations (greater than 80 ppb). We chose to study this event in detail because it was the worst photochemical event for the Sydney region for the oxidant season of 2000/01. The study illustrates the ability of the AAQFS to forecast extreme events, model the impact of recirculation and fumigation of pollutants, and model complex flow over a wide range of scales. The study also illustrates the high demands that are made on the modeling system.

Figure 1 shows a map of the Australian region, giving the location of various places named in the text and indicating the extent of the Sydney region. Expansion of the inset gives the location of various places within the Sydney region and of the DEC monitoring stations. The topographical variation of the region is also shown in Fig. 1 and indicates that there is a basin around Sydney with the highlands of the Great Dividing Range (GDR), oriented north–south, lying to the west.

Synoptic background

The synoptic conditions at 0000 UTC [1100 Australian Eastern Daylight Time (AEDT)] 20–27 January 2001 are depicted in Fig. 2. The first day shown (20 January 2001, Fig. 2a) is the day prior to the event. During the summer period the subtropical ridge is typically located to the south of the continent, and each anticyclone propagates from west to east with a period of about 5–7 days. More often than not, a cold front forms in the deformation region between the anticyclone and a trough to the southeast. These fronts often weaken or “slip away” to the south after they pass through the Great Australian Bight and approach the southeast of the continent. Two examples of the cold fronts slipping away to the south are evident in Fig. 2. Note the frontal progression between 20 and 21 January (Figs. 2a,b) and 22 and 23 January (Figs. 2c,d). Abrupt wind shifts and temperature changes are often associated with midlatitude troughs that propagate up the southeast coast. Unlike the analyzed cold fronts, they tend to be dry and cloud free. Such wind shifts and temperature changes frequently affect the Sydney area during the warmer months. See the dashed lines marking the trough axes on 21, 25, and 26 January (Figs. 2b, 2f, and 2g, respectively). These troughs are often shallow and short lived, and they are not always analyzed on the synoptic charts.

A high pressure cell, which remained quasistationary for the period, was located in the Tasman Sea southeast of Sydney. The gradient wind in the Sydney region during the event period was from the north to northwest. A prefrontal trough propagated up the east coast to the Sydney region on the first day of the event (Fig. 2b). On the next day (Fig. 2c), the high pressure cell in the Tasman Sea strengthened. The high pressure cell moved eastward on the third day (23 January; Fig. 2d), and the region of high pressure leading the front analyzed in Fig. 2d protruded along the southern coast and began to propagate around the southeast corner of the continent. Although no trough was analyzed, a wind shift was present where the coastal ridging impinged on the anticyclone centered in the Tasman Sea. By the next day (24 January, Fig. 2e) the gradient wind had weakened. Another trough propagated up the east coast on day 5 (25 January, Fig. 2f). Both modeled and observed surface winds showed an associated wind shift that reached Wollongong near midnight (see below). The trough passed through the Sydney region on day 6 (26 January, Fig. 2g) and then dissipated. On the last day of the event (27 January, Fig. 2h) the pressure gradient over the Sydney region was very weak. Pressure increased to the north, southeast, and southwest (where ridging led an approaching cold front). We show below that a model wind shift from northeasterly to southerly developed and passed through Sydney at 1100 AEDT and continued up the coast past Newcastle. This was confirmed by observations. Although no associated trough line was analyzed on the synoptic chart, the southerly winds were of sufficient strength to flush the pollutants away from the Sydney region, and bring the 7-day event to a close.

AAQFS performance

The AAQFS has shown good potential for forecasting extreme ozone concentrations. During the 7-day ozone episode in Sydney, the AAQFS went from forecasting near-background concentrations of ozone prior to and after the episode to forecasting concentrations of over 160 ppb during the course of the episode.

We now examine the ozone time series from six of the DEC monitoring stations (see Fig. 1 for their locations), representing both near-coastal stations (the main source region for the primary pollutants) and the inland stations (the main receptor region for the secondary pollutants). Figure 3 shows the time series for predicted and observed ozone concentrations for the entire 7-day period. This type of comparison is a sensitive measure of system performance and places high demands on the accuracy of the entire system: the emissions inventory, meteorological modeling, and photochemistry modeling. There is no metric that directly compares modeled time series values based on volume averages and observed values based on point measurements. Significant differences can result, particularly in regions where concentration gradients vary over very small scales (over hundreds of meters; see Tesche et al. 1990). With this caveat in mind, we present a comparison of the predicted and observed values in Fig. 3, where the predictions have been interpolated to the monitoring station locations.

Measurements of ozone concentration, as shown in Fig. 3, can be categorized by five characteristic profiles in the Sydney region (Hyde et al. 2000), as follows.

  1. A rapid increase in concentration in the morning is followed by several hours of a relatively constant concentration and then a rapid reduction in concentration in the late afternoon (day 5, Liverpool).

  2. A rapid increase in concentration during the morning is followed by a period of relatively constant concentration before a step-change increase in late morning or early afternoon with the onset of the sea breeze. Concentrations decrease after the passage of the sea-breeze front (day 6, Rozelle).

  3. A rapid increase in concentration to a peak is followed by a step-change decrease midmorning or in the early afternoon. A period of relatively constant concentration follows until late afternoon. After this time the concentration is reduced (day 4, Earlwood).

  4. A rapid increase in concentration to a midmorning or early afternoon peak is followed by a rapid reduction to a lower plateau of relatively constant concentration. Another abrupt increase follows with the onset of the sea breeze. Concentrations are again reduced after passage of the sea-breeze front. The effect of a double peak is created during the day (day 1, Lidcombe; day 4, St. Mary's).

  5. A ramping of the concentration to a peak midmorning or in the early afternoon is followed by a gradual reduction (day 7, Liverpool).

The AAQFS was able to reproduce the observed diurnal variation reasonably well on a majority of the days, despite the complexity of the meteorological conditions, which involved often-opposing wind fields from the synoptic flow, topographically forced circulations, sea breezes, and southerly wind changes. Although the general performance was quite encouraging, some important differences between predictions and observations occurred. One of the most significant of these was the underestimation of peak values at some inland stations on some days, coupled with the overestimation of peak values at stations closer to the coast. This can be seen, for example, on days 4 and 5, and occurred when the predicted inland propagation of the sea breeze was blocked by opposing westerly synoptic flow. This led to an underestimation of the peak ozone concentrations at stations located further inland, including Richmond, Vineyard, St. Mary's, Blacktown, Bringelly, Liverpool, Camden, and Campbelltown. On day 5 the sea breeze stalled near the stations of Earlwood, Lidcombe, Lindfield, and Rozelle, and the modeled concentrations kept building there and, hence, were overestimated. On day 4 the sea breeze penetrated beyond Lidcombe but stalled before reaching Liverpool. This resulted in forecast values in coastal areas that were close to the observed values. Northern inland stations, for example, Richmond, Vineyard, St. Mary's and Blacktown, were underestimated on days 1, 2, and 3. These underpredictions may be due to increased dilution resulting from an overestimation of the height of the boundary layer. Errors in the planetary boundary layer (PBL) height can arise from initialization uncertainties, for example, soil temperature and moisture, and/or inadequate initial resolution of the inversion layer structure. Other possibilities are systematic errors in the emission inventory or in the simple condensed chemistry that is employed.

The emissions inventory has been validated only through gross comparisons, and uncertainties remain in the emissions estimates particularly, for example, for hydrocarbons from biogenic sources. Validation studies for the base inventory have included a study in the Sydney Habour Tunnel and a near-road study. Estimates of the nonmethane hydrocarbons (NMHC) to oxides of nitrogen (NOx) ratio for motor vehicles based on these studies range from 0.5 to 3.9 (by weight). In another study, using aircraft traverses of the early morning, weekend day, Sydney urban plume, the ratio was estimated to be on the order of 3.4 (by weight). The total NMHC/NOx ratio, based on the Sydney Metropolitan Air Quality Study emissions inventory, is 1.6 (by weight) and is lower than found in many North American and European cities. Given the uncertainty in the experimental determination of the ratio, the agreement between the inventory and the observations is considered to be satisfactory (Carnovale et al. 1996).

We employ the Generic Reaction Set (GRS), a highly condensed photochemical mechanism (Azzi et al. 1992). Tests of the mechanism under certain conditions of dilution indicate that it can underpredict the ozone peak. The photochemical mechanism is currently being revised to better represent these conditions (M. Azzi 2003, personal communication). Further study is under way to explore these possibilities.

The penetration and arrival time of the sea breeze were correctly forecast at other inland stations, for example, Liverpool, Bringelly, Camden, and Campbelltown, on days 3 and 6, leading to good forecasts of extreme ozone events (ranging up to 140 ppb). On days 6 and 7 the predictions at all stations corresponded well with the observations.

Figure 4 shows the time series comparisons for NOy. The near-coastal stations of Rozelle and Earlwood indicate that AAQFS predicted the diurnal variation of NOy reasonably well, in general, but that there were occasions during stable nocturnal conditions or in the early morning traffic peak (e.g., at 24, 32, and 80 h) when there were significant errors in the predicted peak concentrations. These errors are attributed to errors in the transport trajectory and/or ventilation. For Rozelle and Earlwood these errors led to overestimates of the peaks. This trend reverses itself going farther inland. At Lidcombe the predicted peaks are in good agreement with the observations. Farther inland (at Liverpool and Campbelltown) AAQFS underestimates these peaks. By the time St. Mary's is reached, both the predicted and observed NOy concentrations were close to background levels. On day 4, when the sea breeze was blocked by the synoptic flow, the predicted peak NOy concentrations inland (at Liverpool, Campbelltown, and St. Mary's) were in agreement with observations, but the peak ozone was underpredicted; on day 5, also a day of blocked sea breeze, the NOy peak predictions at Liverpool and Campbelltown were underestimated, but the peak ozone predictions were close to the observed values. The time delay between the peak values of NOy and ozone at the inland stations means that the NOy peak does not determine the ozone peak there.

Key features of the episode

Meteorological conditions and transport phenomena play a key role in determining peak ozone concentrations. The forecasting of these conditions is, therefore, of great importance to the forecasting of air quality. During the study episode, the prediction of PBL heights, vertical mixing, and drainage and sea-breeze circulations and their interactions with the synoptic flows (westerly winds and southerly wind changes) were found to be critical to good air quality predictions.

PBL height, double peaks, fumigation, and long-range transport

A critical factor in predicting ozone concentration levels is correctly forecasting the height of the PBL because this height determines the vertical extent of the turbulent mixing and dilution. In Fig. 5 we show a comparison of the predicted and observed PBL height, based on Bureau of Meteorology's Aircraft Meteorological Data Relay (AMDAR) observations (i.e., measurements from commercial aircraft) for 22–29 January 2001 (measurements were unavailable on 21 January 2001). Particularly important are the values of the PBL height at the times of the early morning traffic peak (0800– 1000 AEDT) and the maximum diurnal height. The plots show severely limited growth of the PBL height during the event period (because of the strong stratification due to subsidence associated with the high pressure cell located over the Tasman Sea). Values of the observed mixed-layer height less than 0, which occasionally are indicated at night, are an artifact of the approximation used to convert pressure to height.

The rise of the modeled PBL height during the morning traffic period was well captured on most days and the maximum daytime mixed-layer height agreed with observations to within about 100–150 m. However, this comparison can be considered only a partial verification. The observations were taken during the aircraft take-off period from Sydney Airport (Mascot), and, hence, the flight paths were limited to the near-coastal region. The PBL growth in this region is often limited by the relatively early development of a sea breeze circulation. No observations of the PBL height were available for the inland region during the event period.

Another feature related to the PBL height is seen in Fig. 3. On some days, some of the inland stations exhibit a double peak in the observations of ozone concentration (e.g., day 3, Bringelly and St. Mary's). This feature has also been reproduced in the forecasts. The first peak develops in the early morning when strong thermal stratification limits the PBL growth (thus, inhibiting vertical dilution of ozone) and the ozone concentration increases with time. During the late morning, surface heating continues to increase until a point is reached when the convective turbulence is strong enough to break through the limiting stratification and the PBL grows rapidly in height. This growth is accompanied by ozone dilution, which creates the trough between the peaks. The second ozone peak occurs when the sea breeze arrives, bringing with it higher ozone concentrations. For stations located closer to the coast the sea breeze arrives before the ozone trough occurs (i.e., before the rapid growth of the PBL), and consequently those time series show only a single peak.

The model provides insights into the dynamical processes governing air quality that are not available through the limited observations of the monitoring stations. The night of 25–26 January 2001 illustrates a number of key features. A southerly wind surge pushed up the coast and was opposed by a northerly background flow. The model predicted that the ozone produced during the previous day (25 January) did not dissipate, and relatively high concentrations of ozone remained overnight (see Figs. 6a,b, e.g., where surface concentrations in excess of 80 ppb are predicted between Canberra and Wollongong at 0100 and 0400 AEDT). This ozone plume was sheared by winds of different directions and speeds over the sea and land—southerly winds over the sea and easterly winds over the land (cf. Figs. 6a and 6b). The predicted ozone, carried inland by the sea breeze and inland component of the southerly surge on 25 January, rotated anticyclonically during the night because of inertial turning and possibly the reestablishment of the northerly background flow. In this slowly rotating flow, high concentrations of ozone were carried inland and then southward toward Canberra (Figs. 6a– c). Above the surface the distribution of ozone was more widespread and extended from Canberra to Muswellbrook (Fig. 6e). By 1000 AEDT on 26 January, the surface heating had generated sufficient mixing for fumigation to occur, and the elevated ozone of relatively high concentrations in places (e.g., >60 ppb about 70 km north-northeast of Canberra) was predicted to mix down to the surface (Figs. 6d,f). The presence of southerly winds prevented the ozone precursors from being advected offshore, and ozone production on 26 January was largely confined to the land near the coast.

Drainage (katabatic) flow and the sea-breeze circulation

Nocturnal cooling of a shallow layer of air near the surface in the near-coastal region often leads to the formation of drainage (katabatic) flows in the Sydney region. The model predicted a drainage flow between the GDR and the coast (see Fig. 1) between 0400 and 0800 AEDT 21 January 2001. This drainage flow (weak flow toward the coast) is evident in Fig. 7a, which shows surface ozone precursors and modeled and observed surface winds at 0600 AEDT. Between Sydney and Wollongong the model predicted the transport of ozone precursors within the drainage flow, from west to east into the urban areas, and out over the sea (Fig. 7a). The light westerly drainage winds contrast with the stronger northerly winds over the sea. Comparison of the observed winds (boldface vectors) with the modeled winds (other vectors) indicates that on a regional scale the model captured the drainage flow pattern well.

Between 0800 and 1000 AEDT the winds over land near the coast were relatively calm, while over the sea northerly winds were maintained. By 1100 AEDT a sea breeze circulation had developed along the coast. The sea breeze transported pollutants back inland that had previously been carried offshore by drainage flows.

On this particular day, in addition to the mesoscale circulation of the sea breeze, a synoptic-scale trough line propagated northward up the east coast of Australia (see Fig. 2b). By 0600 AEDT the edge of the southerly wind change, associated with the trough line, is just visible at the southernmost part of the domain (Fig. 7a). Figure 7b shows PBL height (shaded) and winds at 1200 AEDT. The convergence of the southerly wind surge with the onshore flow of the sea breeze is evident just south of Wollongong. The relatively low PBL height near the coast (∼1000 m) indicates the predicted penetration of the sea breeze in the northern half of the domain, the combined flow of the southerly wind surge and the sea breeze near Wollongong, and the penetration of the surge of cool air behind the wind shift south of Wollongong. The cooler air (implied by lower PBL heights) and the beginnings of the penetration of the sea breeze near Newcastle up the Hunter Valley can also be seen (cf. Fig. 1). The predicted inland penetration of the cooler flow 4 h later is illustrated in Fig. 7c. Note the inland turning of the southerly wind shift, which reinforced the inland penetration of the sea breeze. At this time the combined sea breeze–southerly wind surge front had penetrated significantly farther inland than the sea breeze alone, farther to the north. This daytime inland flow turning was also present in the Tory et al. (2001) study, in which the land–sea temperature contrast was greatest where the cooler southerly wind shift had penetrated. This led to a greater land–sea pressure contrast, and, consequently, greater inland acceleration. The relatively weak and short-lived forcing meant that the modeled southerly wind shift decayed soon as the energy was dissipated inland.

The ozone pattern at 1600 AEDT is shown in Fig. 7d. The two centers of predicted NOy precursor concentrations [see Fig. 7a; the reactive organic compounds (ROC) pattern was similar] led to two centers of ozone production—one due to Sydney urban emissions and the second due to industrial source emissions near Muswellbrook. The model showed that the ozone center that developed over Sydney was advected inland by the combined sea breeze–southerly wind surge front (cf. Fig. 7c, near Lithgow); the more northern center was advected up the Hunter Valley to the west of Muswellbrook by onshore flow funneled up the valley. Figure 8 presents a vertical section oriented perpendicularly to the sea breeze front that passes through the northern ozone center. The position of the section is indicated by the thin line in Fig. 7d. Note the ozone carried inland at the leading edge of the sea breeze (indicated by the region of strongest horizontal potential temperature gradient) and the larger and more highly concentrated ozone plume farther inland. (This is the second ozone center located to the west of Muswellbrook). The superadiabatic layer at lower levels indicates the presence of moist air; this apparent anomaly disappears (except very close to the surface) if virtual potential temperature is plotted instead of potential temperature.

Preliminary sensitivity tests with the AAQFS have been carried out to evaluate the possible causes of forecast error. In Fig. 9 we show the results for changes in the meteorology and in the emissions. For the former, the horizontal diffusion was reduced by a factor of 10 and an experimental effective surface roughness was applied (this gave greater spatial resolution of the effective roughness and resulted in a smoother surface near the coast). For the latter, anthropogenic volatile organic compounds (VOCs) were increased by 50%. The spatial plots of the differences from the base case (Figs. 9a,b) show that the changes in the meteorological conditions led to increases in ozone of more than 40 ppb and covered a larger area; the magnitudes of the differences due to emissions changes were smaller (up to 20 ppb). The changes in the meteorological conditions led to increased near-surface wind speeds in the coastal region ahead of the southerly wind change, which resulted in a few parts per billion increase in the ozone peak at Liverpool (Fig. 9c). Farther inland at St. Mary's the increase was ∼10–20 ppb (Fig. 9d), a significant improvement with the observations there. The increased VOCs led to increases of up to 10 ppb in the ozone peak at Liverpool, but smaller increases at St. Mary's. The use of a more comprehensive chemistry (Carbon Bond IV, not shown) resulted in ozone changes that were similar to the VOC scenario. These tests indicate there is a very high demand on accurate meteorological forecasts if the ozone exceedances inland are to be correctly forecast; errors in the emissions and simplifications to the chemical mechanism seem to be of lesser importance in the Sydney region.

The agreement between the modeled winds and the observations shown in Fig. 7 suggests that the modeled wind shift was a real feature present in the observations. Figure 10 shows the time series of wind speed, wind direction, and screen temperature for four Bureau of Meteorology stations. Although the timing of the predicted wind shift was good, a low wind speed bias was sometimes present in the model results. This bias is partially explained by observations being taken at a point representative of a relatively smooth (grassy) surface being compared with model predictions reflecting the greater surface roughness of an average over the grid square (approximately 5 km × 5 km). The bias also depends on the parameters used for topographical roughness and horizontal diffusion (as discussed above), and on model vertical resolution (the present model has levels at approximately 10, 23, 50, 110 m, …). The predicted screen temperatures were, in general, in good agreement with the observations.

The ozone maxima continued to be advected to the west as the sea breeze and southerly wind surge penetrated farther inland, and by 2200 AEDT both centers had begun to dissipate (not shown).

The drainage flow–sea–breeze circulation cycle is typical of many ozone-event days in the Sydney region. On these occasions drainage flows carry precursors from the urban area and the Hunter Valley off the coast. Ozone develops over the sea and subsequently it is transported inland within the sea-breeze front. Sometimes the picture is made more complicated by the presence of a southerly wind change propagating up the east coast, which combines with the sea breeze, as seen on 21 January. We will discuss the southerly wind change further in the next section.

The days of 24 and 25 January 2001 were modeled by AAQFS with only partial success. The meteorological dynamics governing the westward transport of ozone depended on a delicate balance between the synoptic wind field and the mesoscale sea-breeze circulation. On both of these days the model overpredicted the strength of the synoptic westerly flow, and as a result the sea-breeze ingress was blocked.

For example, on 24 January 2001 there was recirculation of remnant ozone left over from the previous day, both at the surface and aloft (Fig. 11a). The wind pattern was well forecast up until about 1400 AEDT (see Figs. 11a,b), and the sea breeze had begun to travel inland. However, the modeled synoptic flow then strengthened and blocked the advance of the sea breeze between 1400 and 2100 AEDT. The model predicted that some of the ozone ahead of the sea breeze was transported aloft and carried back offshore by the sea-breeze return flow. This is illustrated in Fig. 11c, which shows a vertical section oriented northwest–southeast (perpendicular to the coast), passing through Sydney. Figure 11d shows the predicted and observed surface wind pattern and the ozone spatial distribution at 1700 AEDT. The modeled and observed sea-breeze penetration can be inferred from this figure. The observed winds, with a significant easterly component, indicate locations within the observed sea-breeze flow, and the position of the modeled sea-breeze leading edge can be determined from the location of the converging easterly and westerly pointing vectors. The consequence of the stalled modeled sea breeze and lifting of the ozone is seen in the time series comparisons of modeled and observed ozone concentrations shown in Fig. 3. Concentrations at stations on the coastal side of the blocked sea-breeze front (Earlwood, Lidcombe, Linfield, Randwick, Rozelle, and Woolowarre) were predicted well. Concentrations at stations inland of the blocked sea-breeze front (Blacktown, Bringelly, Camden, Campbelltown, Liverpool, Richmond, St. Mary's and Vineyard) were significantly underestimated.

Southerly wind changes

During the 7-day period southerly wind shifts propagated up the coast from the south on four occasions. The first and second wind shifts (on days 1 and 3) advanced as far north as Wollongong before they turned inland and dissipated. The third pushed beyond Sydney to Newcastle (day 5/6) before it turned inland, and the fourth was a stronger surge on day 7 that cleared the Sydney region of ozone. Although the wind shifts were relatively benign, they each had a noticeable impact on the ozone transport or cycle. The first wind shift enhanced the inland transport of ozone south of Wollongong. The model showed that a mesoscale eddy formed at the leading edge of the second wind shift. This drew in and trapped ozone precursors that later developed into a local ozone center, which was predicted to be transported far inland to the Canberra region during the evening. The third wind shift pushed up the coast overnight and stopped the drainage of precursors to the coastal areas and over the sea. As a consequence, the arrival of the sea breeze brought clean air to the inland stations instead of the ozone-rich air. The fourth wind shift, as mentioned above, was responsible for cleansing the Sydney airshed and brought the 7-day high-ozone event to a close.

Given the impact of the southerly wind changes during the event, it is useful to have an understanding of how they are formed and what leads to their decay. Each wind shift appears to be associated with a trough that propagated up the east coast. This is evident in Fig. 2. The predicted locations of the first and third wind shifts were consistent with the troughs analyzed in Fig. 2b, and Figs. 2f and 2g, respectively. The predicted second and fourth wind shifts were not accompanied by an analyzed trough, although they were associated with local regions of low and high pressure approaching from the south or southwest (see Figs. 2d and 2h, respectively). In all four cases the gradient wind associated with approaching high pressure had an onshore component. We propose that the dynamics of the observed southerly wind surge are similar to the well-documented southerly buster (SB; sometimes called “burster”), which frequently propagates up this part of the NSW coast, and the wind shift associated with the coastal ridging event documented by Tory et al. (2001). The SB is a wind surge, with gusts in excess of 15 m s−1, generated by the interaction of cold fronts with the GDR (e.g., Baines 1980; Colquhoun 1981; McInnes 1993; Reid and Leslie 1999). The GDR is a physical barrier that blocks the cold front and leads to the damming of cold air against the mountain barrier, which then propagates along the coast as a gravity current. The wind shift documented in Tory et al. (2001) was a little more subtle. The interaction of the mountain barrier, and land– sea heat and friction contrasts, on a more gradually changing synoptic pattern led to an abrupt flow transition along the coast. In both systems, near-geostrophic balance is disrupted, which leads to ageostrophic flow acceleration and abrupt changes in wind speed and direction. The intensity of the change decays with distance from the coast. There are similarities between the four southerly wind shifts during the 7-day event and these properties. As mentioned above, the gradient wind shifted to an onshore direction, which suggests interaction with the mountain barrier and the transition from oversea to overland flow was likely to have had an effect on the flow. The model results for days 3 and 7 (23 and 27 January 2001) show that the wind shift tends to be more advanced closest to the coast and is more abrupt than the gradual transition of the gradient wind implied in Fig. 2.

On day 3 the model predicted NOy precursors over the Sydney–Wollongong area, and in the Hunter Valley and along the coast extending south from Newcastle (Fig. 12a). A wind shift moved up the coast. By 0600 AEDT the predicted drainage flow was well established and the wind shift had reached Nowra (not shown). Ahead of the wind shift a mesoscale cyclonic eddy was simulated (see the intersection of southerly and northerly coastal flow in Fig. 12a). Between 0600 and 1000 AEDT this circulation and the northerly coastal flow to the north of it drew the precursors southward and out over the sea to the southeast of Wollongong (see Fig. 12a, which shows the NOy concentrations at 0800 AEDT and the precursors drawn into the eddy leading the wind shift). The presence of significant concentrations of recirculated precursors over the land led to the early formation of ozone at 1000 AEDT (Fig. 12b). At this time two ozone maxima were present; however, shortly afterward the northerly flow merged the northern center with the coastal ozone center. By midday, flow converged on the Sydney ozone center from four directions (Fig. 12c). Synoptic westerlies approached from the west, the southerly wind shift approached from the south, the sea breeze approached from the east, and the weak northerly flow approached from the north. By this time the southerly wind shift had begun to rotate inland in a similar fashion to day 1. It is interesting to note that the model predicted that the ozone plume, which developed from the interaction of sunlight with the precursors trapped in the eddy, was carried inland by the sea breeze and turned counterclockwise with time (presumably because of inertial turning). The ozone pattern at 1800 AEDT is shown in Fig. 12d. By the early hours of the following morning the plume had been transported as far as Canberra (Fig. 11a)—a distance of 100 km.

On the last day of the event (27 January 2001) the general flow patterns were similar to day 1 (Figs. 13 and 7). The predicted NOy precursor concentration pattern and the wind field at 0600 AEDT are shown in Fig. 13a. The southerly wind shift was stronger and broader than on previous days, and had advanced farther up the coast by 0600 AEDT. The model forecast the timing of the southerly wind shift very well (accurate to within 1 h). Figure 13a shows observed southwesterly winds at Newcastle well prior to the model southerly wind shift arrival. Hourly observed wind sequences (not shown) identified a mesoscale wave, which was not resolved by the model. This wave, propagating ahead of the wind shift, accounts for the wind discrepancy at Newcastle. The predicted inland turning of the southerly flow during the day had less of a stalling effect on the northward propagation and the wind shift reached Newcastle by 1500 AEDT. The Sydney ozone plume was advected northward and inland by the winds behind the wind shift. The Hunter Valley plume was carried coastward by an offshore flow preceding the wind shift. It then merged with the Sydney plume and was advected northward beyond Newcastle by the leading edge of the southerly flow. The predicted ozone pattern at 1700 AEDT is shown in Fig. 13b. By 1800 AEDT the coastal winds had turned southerly and transported clean air inland, which brought an end to the 7-day high-ozone event.

Summary and conclusions

The simulation of the 7-day Sydney photochemical smog event demonstrated the ability of AAQFS to forecast situations involving recirculation of precursors and remnant ozone, fumigation, and complex meteorological dynamics. The forecasting system was able to successfully predict high values of ozone, although at times the peak concentrations for the inland stations were underestimated. The dynamics for the Sydney region during the 7-day event were often governed by a sensitive balance between the synoptic and mesoscale flows. On two occasions the system forecast a synoptic flow that was too strong, which resulted in blocking of the inland-advancing sea breeze and provided only partial success in forecasting the correct ozone concentrations on those days. On four occasions forecast southerly wind shifts associated with large-scale pressure troughs propagated up the coast, in good agreement with observations. These wind shifts arrested the seaward progress of overnight drainage flows, enhanced the inland transport of pollutants, and on the last day penetrated far enough northward to cleanse the region, bringing the event to a close.

Further analysis of the ability of AAQFS to forecast photochemical smog at regional and subregional scales is discussed by Tory et al. (2004) and in a forthcoming paper. These papers also outline the work being done to improve the emissions inventory, the meteorological model, and the chemical transport model.

Acknowledgments

The Australian Air Quality Forecasting System was partially funded by the Air Pollution in Major Cities Program (sponsored by Environment Australia through the Natural Heritage Trust).

REFERENCES

  • Azzi, M., G. M. Johnson, and M. Cope. 1992. An introduction to the Generic Reaction Set photochemical smog mechanism. Proc. 11th Int. Clean Air Conf., Brisbane, QLD, Australia, Clean Air Society of Australia and New Zealand, 451–462.

    • Search Google Scholar
    • Export Citation
  • Baines, P. G. 1980. The dynamics of the southerly buster. Aust. Meteor. Mag 28:175200.

  • Carnovale, F., K. Tilly, A. Stuart, C. Carvalho, M. Summers, and P. Eriksen. 1996. Metropolitan air quality study: Air emissions inventory. New South Wales Environment Protection Authority Final Rep., 523 pp.

    • Search Google Scholar
    • Export Citation
  • Colquhoun, J. R. 1981. The origin, evolution and structure of some southerly bursters. Australian Bureau of Meteorology Tech. Rep. 40, 57 pp. [Available from Australian Bureau of Meteorology, National Meteorological Library, GPO Box 1289K, Melbourne VIC 3001, Australia.].

    • Search Google Scholar
    • Export Citation
  • Cope, M. E., P. Manins, D. Hess, G. Mills, K. Puri, P. Dewundege, K. Tilly, and M. Johnson. 1998. Development and application of a numerical air quality forecasting system. Proc. 14th Int. Clean Air and Environment Conf., Melbourne, VIC, Australia, Clean Air Society of Australia and New Zealand, 353–358.

    • Search Google Scholar
    • Export Citation
  • Cope, M. E., G. D. Hess, S. Lee, M. Azzi, J. Carras, N. Wong, and M. Young. 1999. Development of the Australian Air Quality Forecasting System: Current status. Proc. Int. Urban Climatology Conf., Sydney, NSW, Australia, World Meteorological Organization, 595–600.

    • Search Google Scholar
    • Export Citation
  • Cope, M. E. Coauthors 2004. The Australian Air Quality Forecasting System. Part I: Project description and early outcomes. J. Appl. Meteor 43:649662.

    • Search Google Scholar
    • Export Citation
  • Hess, G. D., M. E. Cope, S. Lee, P. C. Manins, G. A. Mills, K. Puri, and K. Tory. 2000a. The Australian Air Quality Forecasting System. AMOS Bull 13:6773.

    • Search Google Scholar
    • Export Citation
  • Hess, G. D., M. E. Cope, S. Lee, P. C. Manins, G. A. Mills, K. Puri, and K. Tory. 2000b. The development of the Australian Air Quality Forecasting System: Current status. Proc. Millennium NATO/CCMS Int. Technical Meeting on Air Pollution Modeling and Its Application, Boulder, CO, Amer. Meteor. Soc., 276–283.

    • Search Google Scholar
    • Export Citation
  • Hyde, R., M. A. Young, and M. Azzi. 2000. Meteorological conditions associated with the occurrence of photochemical smog in Sydney. Proc. 15th Int. Clean Air and Environment Conf., Brighton Beach, NSW, Australia, Clean Air Society of Australia and New Zealand, 421–427.

    • Search Google Scholar
    • Export Citation
  • McInnes, K. L. 1993. Australian southerly busters. Part III: The physical mechanism and synoptic conditions contributing to development. Mon. Wea. Rev 121:32613281.

    • Search Google Scholar
    • Export Citation
  • Reid, H. J. and L. M. Leslie. 1999. Modeling coastally trapped wind surges over southeastern Australia. Part I: Timing and speed of propagation. Wea. Forecasting 14:5366.

    • Search Google Scholar
    • Export Citation
  • Tesche, T. W., P. Georgopoulos, F. L. Lurmann, and P. M. Roth. 1990. Improvement of procedures for evaluating photochemical models. California Air Resources Board Rep. A832-103, 164 pp.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., G. D. Hess, G. A. Mills, and K. Puri. 2000. Verification of the meteorological component of the Australian Air Quality Forecasting System. Proc. 15th Int. Clean Air and Environment Conf., Brighton Beach, NSW, Australia, Clean Air Society of Australia and New Zealand, 221–226.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., C. J. C. Reason, and P. L. Jackson. 2001. A numerical study of a southeast Australian coastal ridging event. Mon. Wea. Rev 129:437452.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., M. E. Cope, G. D. Hess, S. Lee, K. Puri, P. C. Manins, and N. Wong. 2004. The Australian Air Quality Forecasting System. Part III: Case study of a Melbourne 4-day photochemical-smog event. J. Appl. Meteor 43:680695.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Map of the Australian region showing the locations named in the text. (top right) The model domain in New South Wales and the location of four Bureau of Meteorology surface stations: Badgerys Creek (YSBC), Canterbury (Sydney) (CBRC), Bellambi (BELL), and Nowra (YSNW). (bottom) The Sydney region and the location of DEC monitoring stations: Blacktown (BLAC), Bringelly (BRIN), Camden (CAMD), Campbelltown (CAMP), Earlwood (EARL), Lidcombe (LIDC), Lindfield (LIND), Liverpool (LIVE), Randwick (RAND), Rozelle (ROZE), St. Marys (ST.M), Richmond (RICH), Vineyard (VYNE), Westmead (WSTM), and Woolowarre (WOOL). The stippled shading indicates the urban area

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 2.
Fig. 2.

Australian Bureau of Meteorology manual synoptic-scale analyses at 0000 UTC (1100 AEDT) for the day prior to and those days of a 7-day photochemical smog event in Sydney. The day of the month is shown in the lower left-hand corner of the maps (20–27 Jan 2001). The solid lines with barbs indicate cold fronts; the dashed lines indicate the positions of pressure troughs (with the monsoonal trough across the north of Australia); the contours show the mean sea level pressure (hPa) with a contour interval of 4 hPa. The letter S marks the location of Sydney

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 3.
Fig. 3.

The time series of modeled and observed ozone concentrations for the monitoring stations in the Sydney region. The boxes indicate the observations, and the solid lines show the forecast values interpolated to the location of the monitoring stations. The order of the stations from top to bottom represents increasing distance from the coast

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 4.
Fig. 4.

The time series of modeled and observed NOy concentrations for the monitoring stations in the Sydney region. The boxes indicate the observations, and the solid lines show the forecast values interpolated to the location of the monitoring stations. The order of the stations from top to bottom represents increasing distance from the coast

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 5.
Fig. 5.

Comparison of the forecast PBL height (light line) and the observed height (dark line), based on measurements from commercial aircraft (AMDAR) throughout the 7-day episode (local time). The observations are based on the takeoff period from Sydney Airport (Mascot) and, thus, represent the near-coastal region

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 6.
Fig. 6.

Modeled ozone concentrations (ppb): (a) 0100, (b) 0400, (c) 0700, and (d) 1000 AEDT 26 Jan 2001. Vertical cross sections in (e) and (f) are oriented parallel to the coast [along the straight lines shown in (c) and (d)], approximately 100 km inland, passing through Canberra (C), Lithgow (L), and Muswellbrook (M). They show the upper-level O3 at (e) 0700 and (f) 1000 AEDT. Other locations indicated are Wollongong (W), Orange (O), and Newcastle (N). The boldface vectors indicate wind observations. Ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 7.
Fig. 7.

Wind patterns for 21 Jan 2001 (day 1). The wind observations are shown in boldface and the modeled winds in gray: (a) the drainage flow and precursor NOy concentrations at 0600 AEDT, (b) the PBL height patterns at 1200 AEDT, (c) the PBL height patterns at 1600 AEDT, and (d) the ozone concentration patterns at 1600 AEDT. The straight line, approximately perpendicular to the coast, indicates the location of the vertical section shown in Fig. 8. The NOy concentrations increase by 10 ppb, beginning with a minimum contour of 5 ppb; the ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb; the PBL height contours increase by 250 m, beginning with a minimum contour of 100 m. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 8.
Fig. 8.

Vertical section approximately perpendicular to the sea-breeze front (for the position see the diagonal line in Fig. 7d) at 1600 AEDT 21 Jan 2001 (day 1). Contours of ozone concentrations increase by 15 ppb, beginning with a minimum of 35 ppb. Potential temperature is contoured (interval, 1 K), and horizontal winds are represented by wind vectors. Downward- (upward-) pointed vectors represent flow out of (into) the page

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 9.
Fig. 9.

Sensitivity tests of possible forecasts errors in meteorology and emissions: 1600 AEST 21 Jan 2001 showing positive differences between (a) 1.5 × VOCs and base case; (b) roughness (z0 filter) and diffusion and base case; time series for 21–22 Jan 2001 for (c) Liverpool and (d) St. Mary's. Ozone difference contours increase by 4 ppb, beginning with a minimum of 4 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 10.
Fig. 10.

Time series for 21 Jan 2001 (day 1). The observed (solid lines) and modeled (dashed lines) 10-m wind speeds are given (m s−1, lower curves) and the screen temperatures are given (°C, middle curves); the observed (dark vectors) and modeled (light vectors) wind direction are shown at the top of (a)–(d) with north aligned vertically: (a) Badgerys Creek (YSBC), (b) Canterbury (Sydney) (CBRC), (c) Bellambi (BELL), and (d) Nowra (YSNW). See Fig. 1 for station locations

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 11.
Fig. 11.

The wind field and ozone concentration patterns for 24 Jan 2001 (day 4): (a) remnant ozone at 0300 AEDT from the previous day; (b) sea-breeze penetration at 1400 AEDT. (c) Ozone lifted up over the sea breeze at 1700 AEDT 24 Jan 2001. The straight line in (d) indicates the position of the vertical section shown in (c). Contours of ozone concentrations increase by 15 ppb, beginning with a minimum of 35 ppb. Potential temperature is contoured (interval, 1 K) and horizontal winds are represented by wind vectors. Downward- (upward-) pointed vectors represent flow out of (into) the page. (d) The corresponding near-surface wind field and ozone concentration patterns at 1700 AEDT. The wind observations are shown in boldface and modeled winds in gray. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 12.
Fig. 12.

(a) The wind field and NOy precursor concentration patterns for 0800 AEDT 23 Jan 2001 (day 3); the wind field and ozone concentration patterns for (b) 1000, (c) 1200, and (d) 1800 AEDT 23 Jan 2001. The wind observations are shown in boldface and the modeled winds in gray. The ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Fig. 13.
Fig. 13.

(a) The wind field and NOy precursor concentration patterns for 0600 AEDT 27 Jan 2001 (day 7); (b) the wind field and ozone concentration patterns for 1700 AEDT 27 Jan 2001. The wind observations are shown in bold and the modeled winds in gray. The ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

Citation: Journal of Applied Meteorology 43, 5; 10.1175/2094.1

Save
  • Azzi, M., G. M. Johnson, and M. Cope. 1992. An introduction to the Generic Reaction Set photochemical smog mechanism. Proc. 11th Int. Clean Air Conf., Brisbane, QLD, Australia, Clean Air Society of Australia and New Zealand, 451–462.

    • Search Google Scholar
    • Export Citation
  • Baines, P. G. 1980. The dynamics of the southerly buster. Aust. Meteor. Mag 28:175200.

  • Carnovale, F., K. Tilly, A. Stuart, C. Carvalho, M. Summers, and P. Eriksen. 1996. Metropolitan air quality study: Air emissions inventory. New South Wales Environment Protection Authority Final Rep., 523 pp.

    • Search Google Scholar
    • Export Citation
  • Colquhoun, J. R. 1981. The origin, evolution and structure of some southerly bursters. Australian Bureau of Meteorology Tech. Rep. 40, 57 pp. [Available from Australian Bureau of Meteorology, National Meteorological Library, GPO Box 1289K, Melbourne VIC 3001, Australia.].

    • Search Google Scholar
    • Export Citation
  • Cope, M. E., P. Manins, D. Hess, G. Mills, K. Puri, P. Dewundege, K. Tilly, and M. Johnson. 1998. Development and application of a numerical air quality forecasting system. Proc. 14th Int. Clean Air and Environment Conf., Melbourne, VIC, Australia, Clean Air Society of Australia and New Zealand, 353–358.

    • Search Google Scholar
    • Export Citation
  • Cope, M. E., G. D. Hess, S. Lee, M. Azzi, J. Carras, N. Wong, and M. Young. 1999. Development of the Australian Air Quality Forecasting System: Current status. Proc. Int. Urban Climatology Conf., Sydney, NSW, Australia, World Meteorological Organization, 595–600.

    • Search Google Scholar
    • Export Citation
  • Cope, M. E. Coauthors 2004. The Australian Air Quality Forecasting System. Part I: Project description and early outcomes. J. Appl. Meteor 43:649662.

    • Search Google Scholar
    • Export Citation
  • Hess, G. D., M. E. Cope, S. Lee, P. C. Manins, G. A. Mills, K. Puri, and K. Tory. 2000a. The Australian Air Quality Forecasting System. AMOS Bull 13:6773.

    • Search Google Scholar
    • Export Citation
  • Hess, G. D., M. E. Cope, S. Lee, P. C. Manins, G. A. Mills, K. Puri, and K. Tory. 2000b. The development of the Australian Air Quality Forecasting System: Current status. Proc. Millennium NATO/CCMS Int. Technical Meeting on Air Pollution Modeling and Its Application, Boulder, CO, Amer. Meteor. Soc., 276–283.

    • Search Google Scholar
    • Export Citation
  • Hyde, R., M. A. Young, and M. Azzi. 2000. Meteorological conditions associated with the occurrence of photochemical smog in Sydney. Proc. 15th Int. Clean Air and Environment Conf., Brighton Beach, NSW, Australia, Clean Air Society of Australia and New Zealand, 421–427.

    • Search Google Scholar
    • Export Citation
  • McInnes, K. L. 1993. Australian southerly busters. Part III: The physical mechanism and synoptic conditions contributing to development. Mon. Wea. Rev 121:32613281.

    • Search Google Scholar
    • Export Citation
  • Reid, H. J. and L. M. Leslie. 1999. Modeling coastally trapped wind surges over southeastern Australia. Part I: Timing and speed of propagation. Wea. Forecasting 14:5366.

    • Search Google Scholar
    • Export Citation
  • Tesche, T. W., P. Georgopoulos, F. L. Lurmann, and P. M. Roth. 1990. Improvement of procedures for evaluating photochemical models. California Air Resources Board Rep. A832-103, 164 pp.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., G. D. Hess, G. A. Mills, and K. Puri. 2000. Verification of the meteorological component of the Australian Air Quality Forecasting System. Proc. 15th Int. Clean Air and Environment Conf., Brighton Beach, NSW, Australia, Clean Air Society of Australia and New Zealand, 221–226.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., C. J. C. Reason, and P. L. Jackson. 2001. A numerical study of a southeast Australian coastal ridging event. Mon. Wea. Rev 129:437452.

    • Search Google Scholar
    • Export Citation
  • Tory, K. J., M. E. Cope, G. D. Hess, S. Lee, K. Puri, P. C. Manins, and N. Wong. 2004. The Australian Air Quality Forecasting System. Part III: Case study of a Melbourne 4-day photochemical-smog event. J. Appl. Meteor 43:680695.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Map of the Australian region showing the locations named in the text. (top right) The model domain in New South Wales and the location of four Bureau of Meteorology surface stations: Badgerys Creek (YSBC), Canterbury (Sydney) (CBRC), Bellambi (BELL), and Nowra (YSNW). (bottom) The Sydney region and the location of DEC monitoring stations: Blacktown (BLAC), Bringelly (BRIN), Camden (CAMD), Campbelltown (CAMP), Earlwood (EARL), Lidcombe (LIDC), Lindfield (LIND), Liverpool (LIVE), Randwick (RAND), Rozelle (ROZE), St. Marys (ST.M), Richmond (RICH), Vineyard (VYNE), Westmead (WSTM), and Woolowarre (WOOL). The stippled shading indicates the urban area

  • Fig. 2.

    Australian Bureau of Meteorology manual synoptic-scale analyses at 0000 UTC (1100 AEDT) for the day prior to and those days of a 7-day photochemical smog event in Sydney. The day of the month is shown in the lower left-hand corner of the maps (20–27 Jan 2001). The solid lines with barbs indicate cold fronts; the dashed lines indicate the positions of pressure troughs (with the monsoonal trough across the north of Australia); the contours show the mean sea level pressure (hPa) with a contour interval of 4 hPa. The letter S marks the location of Sydney

  • Fig. 3.

    The time series of modeled and observed ozone concentrations for the monitoring stations in the Sydney region. The boxes indicate the observations, and the solid lines show the forecast values interpolated to the location of the monitoring stations. The order of the stations from top to bottom represents increasing distance from the coast

  • Fig. 4.

    The time series of modeled and observed NOy concentrations for the monitoring stations in the Sydney region. The boxes indicate the observations, and the solid lines show the forecast values interpolated to the location of the monitoring stations. The order of the stations from top to bottom represents increasing distance from the coast

  • Fig. 5.

    Comparison of the forecast PBL height (light line) and the observed height (dark line), based on measurements from commercial aircraft (AMDAR) throughout the 7-day episode (local time). The observations are based on the takeoff period from Sydney Airport (Mascot) and, thus, represent the near-coastal region

  • Fig. 6.

    Modeled ozone concentrations (ppb): (a) 0100, (b) 0400, (c) 0700, and (d) 1000 AEDT 26 Jan 2001. Vertical cross sections in (e) and (f) are oriented parallel to the coast [along the straight lines shown in (c) and (d)], approximately 100 km inland, passing through Canberra (C), Lithgow (L), and Muswellbrook (M). They show the upper-level O3 at (e) 0700 and (f) 1000 AEDT. Other locations indicated are Wollongong (W), Orange (O), and Newcastle (N). The boldface vectors indicate wind observations. Ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb

  • Fig. 7.

    Wind patterns for 21 Jan 2001 (day 1). The wind observations are shown in boldface and the modeled winds in gray: (a) the drainage flow and precursor NOy concentrations at 0600 AEDT, (b) the PBL height patterns at 1200 AEDT, (c) the PBL height patterns at 1600 AEDT, and (d) the ozone concentration patterns at 1600 AEDT. The straight line, approximately perpendicular to the coast, indicates the location of the vertical section shown in Fig. 8. The NOy concentrations increase by 10 ppb, beginning with a minimum contour of 5 ppb; the ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb; the PBL height contours increase by 250 m, beginning with a minimum contour of 100 m. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

  • Fig. 8.

    Vertical section approximately perpendicular to the sea-breeze front (for the position see the diagonal line in Fig. 7d) at 1600 AEDT 21 Jan 2001 (day 1). Contours of ozone concentrations increase by 15 ppb, beginning with a minimum of 35 ppb. Potential temperature is contoured (interval, 1 K), and horizontal winds are represented by wind vectors. Downward- (upward-) pointed vectors represent flow out of (into) the page

  • Fig. 9.

    Sensitivity tests of possible forecasts errors in meteorology and emissions: 1600 AEST 21 Jan 2001 showing positive differences between (a) 1.5 × VOCs and base case; (b) roughness (z0 filter) and diffusion and base case; time series for 21–22 Jan 2001 for (c) Liverpool and (d) St. Mary's. Ozone difference contours increase by 4 ppb, beginning with a minimum of 4 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

  • Fig. 10.

    Time series for 21 Jan 2001 (day 1). The observed (solid lines) and modeled (dashed lines) 10-m wind speeds are given (m s−1, lower curves) and the screen temperatures are given (°C, middle curves); the observed (dark vectors) and modeled (light vectors) wind direction are shown at the top of (a)–(d) with north aligned vertically: (a) Badgerys Creek (YSBC), (b) Canterbury (Sydney) (CBRC), (c) Bellambi (BELL), and (d) Nowra (YSNW). See Fig. 1 for station locations

  • Fig. 11.

    The wind field and ozone concentration patterns for 24 Jan 2001 (day 4): (a) remnant ozone at 0300 AEDT from the previous day; (b) sea-breeze penetration at 1400 AEDT. (c) Ozone lifted up over the sea breeze at 1700 AEDT 24 Jan 2001. The straight line in (d) indicates the position of the vertical section shown in (c). Contours of ozone concentrations increase by 15 ppb, beginning with a minimum of 35 ppb. Potential temperature is contoured (interval, 1 K) and horizontal winds are represented by wind vectors. Downward- (upward-) pointed vectors represent flow out of (into) the page. (d) The corresponding near-surface wind field and ozone concentration patterns at 1700 AEDT. The wind observations are shown in boldface and modeled winds in gray. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

  • Fig. 12.

    (a) The wind field and NOy precursor concentration patterns for 0800 AEDT 23 Jan 2001 (day 3); the wind field and ozone concentration patterns for (b) 1000, (c) 1200, and (d) 1800 AEDT 23 Jan 2001. The wind observations are shown in boldface and the modeled winds in gray. The ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

  • Fig. 13.

    (a) The wind field and NOy precursor concentration patterns for 0600 AEDT 27 Jan 2001 (day 7); (b) the wind field and ozone concentration patterns for 1700 AEDT 27 Jan 2001. The wind observations are shown in bold and the modeled winds in gray. The ozone concentrations increase by 15 ppb, beginning with a minimum contour of 35 ppb. The letters indicate the locations of Muswellbrook (M), Newcastle (N), Lithgow (L), Orange (O), Wollongong (W), and Canberra (C)

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 332 149 1
PDF Downloads 138 31 1