This case study focuses on a potential fog event in Melbourne, Australia, on 6-7 April 2008. The key aim of this module is to step through the forecast process during a potential fog event from the perspective of an aviation forecaster with the Bureau of Meteorology. This involves consideration of model guidance and observations, identification of potential areas of fog, nowcasting, forecasting fog clearance, and providing TAF updates throughout.

This case is presented in three main parts:

1. A Case Profile which introduces the case setting and the current forecast.
2. A Case Challenge consisting of a series of forecast questions focused on three key time points in the case.
3. A Case Summary with a brief analysis of how the case unfolded.

Geographic Location: Melbourne, Australia

Case Time Frame: 6-7 April 2008

Time Zone: Australian Eastern Time (EST) = UTC + 10 hours.

Phenomenon of Interest: Potential fog impacting on local and international aerodromes.

Forecasting Challenges:

• Forecast the fog potential and classification for the Melbourne region over the next 24 hour period including: forecasting fog onset time and the associated reduction in visibility; forecasting the duration of the fog and fog clearance times at local airports.

• Monitoring for fog risk as the night progresses and updating forecast parameters as necessary.

• If fog and low-cloud have formed, identify areas of fog and low-cloud. Determine the fog and low-cloud onset, associated reductions in visibility, duration and clearance times at airports in the Melbourne area based on identifying processes that will lead to further radiational cooling and/or advection of existing areas of fog/low cloud.

• The forecast process involves integrating all observations, satellite imagery, NWP and official forecast products to forecast and nowcast the basic ingredients required for fog and low-cloud formation and involve: assessment of the synoptic weather considerations; assessment of the mesoscale and local influences; assessment of the regional and local fog climatology.

The setting for this case is Melbourne, Victoria, located in southeastern Australia. The Melbourne metropolitan region is home to 3.9 million people. The Great Dividing Range lies to the west, north and east of the metropolitan area. To the south of Melbourne is Port Phillip Bay, which opens out to Bass Strait, the body of water between mainland Australia and Tasmania.

The highest elevation in the area surrounding Melbourne is to the east with Dunns Hill at 561m (1840ft), and further to the northeast Mount Saint Leonard at 1028m (3373ft). To the northwest of Melbourne is Mount Macedon, 1001m (3284ft ), and to the west there are several smaller ranges, the You Yangs, 348m (1142ft) and the Brisbane Ranges, at 377m (1237ft).

The location of the mountains of the Great Dividing Range can have a large influence on the local winds. In certain situations, they can act to block light southerly winds or they can be the source of night-time, northerly, katabatic, or drainage winds.

Port Phillip Bay and Bass Strait are other geographic features of importance, as they can provide a source of moisture with southerly winds. The sea surface temperature, which for this case is around 19°C in the Bay, can have an influence as well.

Let’s take a look at the location of the local airports, our primary areas of concern for the case.

Melbourne Airport is located 22 km (13.7 miles) to the northwest of the city of Melbourne. The airport is approximately 113 m (370 ft) above mean sea level.

The other local Melbourne airports discussed in this case, Avalon, Laverton and Moorabbin, are all close to mean sea level.

It’s always important to have the local climatology in mind.

In Melbourne, the most prevalent months for fog are the cooler months of April through August, although it can be seen from the graph shown, it is still possible to get fog in Melbourne at any time of the year. In April, the time of our case, Melbourne experiences, on average, at least one fog event.

Fog typically forms in the few hours prior to sunrise. This is when radiational cooling has had the maximum time to take effect, although anytime from late evening is common. Fog clearance times are typically mid-morning.

Average minimum and maximum temperatures for Melbourne in April are 9-12°C and 18-21°C respectively. And the most common wind direction in April for Melbourne Airport is a northerly wind at 9am and a southerly wind at 3pm.

Local fog forecasting aids are used throughout Australia to help in the forecast decision process. These aids have been developed over the years by forecasters or researchers and their properties vary greatly. They range in the different aspects they try to predict, from forecasting onset time or the duration or likelihood of fog. They also differ in terms of the types of weather parameters taken into consideration, such as pressure, moisture, and synoptic type. And the calculation method varies from the use of statistical regression trends to a checklist approach.

The Victorian Forecast Office, the site for our case, has developed and uses the Stern/Parkyn synoptic classification aid. The aid classifies the current forecast synoptic situation into one of 50 different synoptic types. To determine the synoptic type, the aid takes into consideration the gradient wind strength, wind direction and the cyclonicity of the synoptic situation, using pressure values from six locations in southeastern Australia. A climatology of the likelihood of fog for each synoptic type is displayed here. The Stern/Parkyn aid takes this a step further by including the 3pm temperature and dewpoint along with the current month to give a specific likelihood of fog percentage for the individual event.

The current time for our case is 0300 UTC, which is 1pm local time, in Melbourne. Our task is to consider the possibility of fog in the Melbourne area tomorrow morning. We are responsible for four TAF’s:

• Melbourne Airport (YMML)
• Avalon Airport (YMAV)
• Laverton Airport (YLVT)
• and Moorabbin Airport (YMMB)

Our main concern is Melbourne Airport (YMML) as this is an international airport with several early morning flights expected daily.

NOTE: In April 2008, the time of this case study, 24-hour TAFs were issued for international airports in Australia.  During November 2008, all Australian international airports, including YMML, changed to a 30-hour TAF.

The current 00 to 24 UTC Melbourne Airport TAF does not include fog, so we will need to decide if we are happy with this decision, or whether we should issue the TAF with the inclusion of fog. The 06-06 UTC TAF that we are working on is very important to the aviation industry as a lot of planning for the following day is based on it.

TAF YMML 052225Z 060024
17010KT CAVOK
FM15 VRB05KT 8000 HZ SCT008
FM23 04010KT CAVOK
RMK
T 16 20 21 15 Q 1028 1025 1024 1025

The model data we will use for this case study is the US GFS model. A comparison between the GFS model and the local MesoLaps model has already been conducted and there were no significant variations between them for the time period we are considering. So, for simplicity and time efficiency only the GFS model is presented in this case study. There will be two different runs of the GFS model available at different time periods of the case study, the 00 UTC and 06 UTC runs.

Our primary focus and forecast challenge will be forecasting fog potential. Throughout this case, we need to keep in mind the key ingredients required for fog formation. These are: sufficient low-level moisture, light winds and rapid cooling at the surface. 

Let’s now take a look at the current synoptic set-up and forecast before moving on to the case challenge.

It is 0300 UTC, 1PM local time, on Sunday 6 April and the duty forecaster we are relieving gives us a quick briefing:

"Currently an upper ridge is dominating most of the continent and is associated with a surface high pressure system, located near Tasmania. This system is producing light winds over south-east Australia to at least 500hPa.

Mid and upper level clouds currently extend across much of South Australia. The mid and high level cloud over southern Australia (to our west) should continue to advect to the north of Victoria, due to the upper wind flow, but you may also want to monitor this.

In the past 24 hours, the high pressure system dominating our weather has gradually moved southeastwards across the state, and is currently located near Tasmania. Across the state temperatures have ranged from 16-21°C with onshore southerly winds and isolated showers in the south of the state.

Rainfall totals recorded since 9am local time yesterday morning were generally less than 2 millimetres and mostly fell yesterday afternoon. Daytime heating has already produced scattered convective clouds in the south and east of Victoria however the remaining isolated showers are expected to clear this afternoon as the high pressure system continues to strengthen in the Tasman Sea, and the atmosphere stabilises.

The forecast for Melbourne today is fine and mostly sunny, with cool to mild conditions and light to moderate southerly winds.

Overnight minimum temperatures are expected to be below average. Last night minimum temperatures in Melbourne were generally 2-4°C below average, ranging from 7-9°C.

As the high pressure system is expected to remain in the region for at least the next 24 hours, I expect your main concerns will be with the potential for fog or low cloud forming tonight or early tomorrow morning. We currently do not have fog on the Melbourne Airport TAF, but we do anticipate haze and scattered low cloud developing overnight. You may want to consider including fog. 

TAF YMML 052225Z 060024 
17010KT CAVOK
FM15 VRB05KT 8000 HZ SCT008
FM23 04010KT CAVOK
RMK
T 16 20 21 15 Q 1028 1025 1024 1025

I’ll leave you with the current forecasts. Catch you tomorrow morning. Have a good shift!"

Melbourne Forecast
Issued at 10:55 am EST on Sunday 6 April 2008 

Warning Summary
Nil. 

Forecast for Sunday
Fine. A mainly sunny day with light wind. 

Precis:      Mainly sunny.                          
City:        Max 22 

Laverton:    Max 22                 Yarra Glen:   Max 23
Tullamarine: Max 22                 Mt Dandenong: Max 17 
Watsonia:    Max 22                 Scoresby:     Max 22
Frankston:   Max 20                 Geelong:      Max 21 

Forecast for Monday
Fine apart from local morning fog. A mainly sunny day with light wind. 

Precis:      Early fog then fine.               
City:        Min 13   Max 23 

Let’s briefly review the case and highlight some of the key components of the fog forecasting process.

Overnight on Sunday the 6th of April 2008, fog and low cloud formed in the Melbourne region. During the evening, the key ingredients for fog - clear skies, light winds and available moisture at the surface – along with weak upslope flow, produced fog and low cloud on the southern slopes of the ranges to the north and west of Melbourne.

This area of fog and low cloud expanded overnight into the northeastern suburbs of Melbourne. Winds shifted from light southerlies in the evening to light northerlies by sunrise. This change in wind direction allowed the fog in the north of Melbourne to advect southwards over Laverton and Moorabbin airports. The fog gradually dissipated while continuing to advect southwards over Port Phillip Bay. The fog eventually cleared by late morning.

Melbourne Airport managed to avoid being affected by fog, although mist was reported at the airport reducing visibility to 5000 m for a couple of hours. Avalon Airport was not affected by fog, due to a layer of cloud, which lowered to around 1000 ft around sunrise.

Understanding and recognising precursors for fog formation is very important in accurately predicting fog events. Being aware of a potential fog situation will help prevent unforecast fog events, which can cause severe impacts on aviation industry operations. The earlier the precursors can be accurately recognised provides the end users with more planning and preparation time.

Conditions were favourable for fog development overnight, due to the synoptic situation affecting the region. The high pressure system provided all the key ingredients. Light winds, rapid cooling due to clear skies aloft and sufficient low level moisture, enhanced by onshore southerly airflow and isolated showers during the preceding day. Fog potential was evident the afternoon prior to the event through both the model guidance and through short term forecasting based on an analysis of the afternoon observations.

Having a clear understanding of the various processes that can lead to fog formation is important in recognising the type of fog that will form, where fog might form and how it will interact with the local topography.

This case provides a good example whereby a combination of radiative processes enhanced by upslope processes act to produce fog and low cloud. The moist, onshore flow interacted with the hills to the north of Melbourne, resulting in upslope flow and associated cooling enhanced the formation of fog on the slopes.

The model guidance predicted the southerly winds would shift northerly overnight. Advected fog could then be an issue to southern areas if fog formed in the north. Having knowledge of local topography and flow regimes is therefore critical to forecasting fog.

Satellite imagery is especially useful in fog monitoring. As the temperature of fog is often close to the ground temperature, standard infrared channels are not useful in identifying fog. Enhanced infrared imagery is used instead and is often one of the first observational tools to show the presence of fog.

Satellite imagery can also show areas of fog expanding and advecting, which happened in this case in the northeastern suburbs. It can also be used to monitor for fog’s eventual dissipation.

Whenever possible, look for surface observations to verify the presence of fog. In this case, evidence of fog or low cloud appeared in the enhanced IR earlier than forecast and was verified by surface observations.

As fog is a surface based phenomena, surface observations are relied on heavily in both forecasting and identifying fog. Trends in surface observations are essential in short term forecasting and nowcasting of fog. Recognising trends in observations early, helps to provide longer lead times for formation. Surface observations are also crucial in identifying and verifying fog.

Fog was confirmed through several different surface observational tools throughout the case. Ceilometers and visibility meters provided a useful, although limited, insight into the occurrence of fog at a particular location. The five locations around Melbourne were all used extensively throughout this event.

Surface air temperature and dewpoint temperatures can also be used to confirm saturation at the surface, especially for areas without more advanced instrumentation. Saturation at Kilmore Gap, Laverton and Moorabbin confirmed the observed fog on the ceilometer/visibility meters. Avalon reported a dewpoint depression of a couple of degrees all night, which verifies the report of low cloud instead of fog.

Analysing the wind observations was also particularly useful in this case, due to the eventual advection of fog.

The model performed reasonably well in this fog event. The model forecast the key ingredients for fog formation to coincide in the Melbourne region overnight. The forecast of surface winds to shift from southerlies during the evening to northerlies by sunrise was well captured, potentially alerting a forecaster well in advance that should fog form on the southern slopes of the ranges to the north of Melbourne, advected fog would be a possibility.

One area of weakness in the model for this case was under-forecasting the dewpoint temperatures in the Melbourne region. As a consequence of the under-forecasting of the dewpoint, the model would lead you to think that saturation and fog would form later. But an astute forecaster should have noticed this while ground truthing the model during the afternoon prior to the event.

Ground truthing the model performance should always be done before implicitly believing the model output.

The local Bayesian network fog decision aid was run on the actual day of the event; it also suggested a favourable situation for the formation of fog. The extended Stern/Parkyn aid, which includes the additional information of temperature, dewpoint and time of year, forecast the probability of fog to be at 20.5% chance. The synoptic flow was deemed very favourable. The moisture content was deemed at least favourable, with a "worst case scenario" output producing a very favourable situation. Therefore overall, the Bayesian network forecast the likelihood of fog at Melbourne Airport to be at least a 33% chance and the worst case scenario, with very favourable moisture, gave a 66% chance of fog forming.

The national Regano fog aid, had interesting results with only a 5% likelihood of fog when run with the 02UTC observations. This chance gradually increased throughout the afternoon and evening, to reach 31% chance at 14UTC. This demonstrates that forecasting a fog event is more achievable closer to the onset of the event. For more information on the Regano fog aid, see the supporting topics.

Nowcasting fog is difficult. Fog can form or dissipate based on a slight change in conditions. Having a good understanding of the environment the fog is located in is crucial in nowcasting. Knowing the depth, intensity and areal extent of the fog will help with forecasting the duration of the fog event.

The depth of fog was not thick in this event, so along with the increase in solar radiation after sunrise, the fog dissipated relatively quickly. The fog cleared Kilmore Gap with the advection of drier northerly winds around 7:15am. Laverton and Moorabbin both reported fog for around 45 minutes, before improving to mist conditions until mid morning. The mist conditions at Melbourne Airport gradually improved around 8am and the low cloud cleared Avalon by 9am.