A description of storm events from a respected forecaster in Severe Weather.  It shows the genuine concerns that forecasters have and the atmospheric variabilities that confront them.  This is an interesting report   that shows from hour to hour, the potential of a developing system. On this occasion, the forecaster's analysis and predictions proved correct. Other important factors in the Bureau's analysis is the accurate and correct reporting of conditions by storm spotters which in this case has aided the forecaster in his prediction.                                              

~ed.

On the day in question there was sufficient middle level instability for thunderstorms, in fact castellanus was observed during the morning.   Progged LI between 700hPa and 400hPa were about -2 for the afternoon.  There was a history of embedded thunderstorms in middle level cloud over  SA (albeit a fair way to the west). The potential was there for  thunderstorms initiated from the middle levels. However during the  morning I remember working through various decision trees to establish  whether any surface driven thunderstorms were possible and consequently  whether there was the potential for them to be severe.

The first step was to identify regions of surface convergence. Three regions were possible: the northeast ranges, the Gippsland afternoon seabreeze and a shallow surface trough (2,000-3,000ft) in the southwest. The air being lifted in the northeast and Gippsland coast was too dry, however the southwest possessed higher levels of surface moisture ahead of the trough. At the time middle level cloud covered much of the State, including the southwest, however it was thought that this cloud would clear the southwest in the afternoon allowing the air temperature to rise sufficiently to overcome any convective inhibition.

Now, could the atmosphere support severe thunderstorms. Initial analysis suggested a little more than 1000 J/kg, which wasn't bad although shear was not favourable. However, analysis of precipitable water through the atmosphere identified well above average available moisture in the southwest. Given that the steerage flow would be from the northwest and on the slow side, I was thinking flash flooding was a real possibility. Next step was to document the thought process, fill out a few forms and charts, make a few phone calls.........then wait.

This is the forecaster's time of reckoning. Will the atmosphere co-operate or do its own thing? A few hours go by and the original conceptual model is going to plan - the trough is slowly pushing inland in the southwest; the middle level cloud is clearing; surface dewpoint is still around 12 ahead of the trough and temperatures are in the mid-20s. Why isn't anything happening!

Mt Gambier radar goes into wind finding mode just as a couple of weak cells develop near the border. These cells can just be picked up from the radar in Melbourne, but the resolution is poor. I switch to monitoring the lightning detector - nothing yet. The cells are obviously growing as they push through 8km on the radar then 10km. I'm thinking there's a bit of punch there, however upper reflectivities are on the low side. Trying not to get too excited, however the lightning display starts firing off and I'm telling everyone in the office.

OK, Mt Gambier comes back on line and I switch it to 3-D mode hoping to pick-up the vertical structure of the storms from a closer radar. The image is impressive with 40 dBZ up to about 4km, but I'd like to see 50 dBZ at 6 to 7km. Nothing yet to suggest the storm is anything other than a normal garden thunderstorm. The next two images (every 10 minutes) show the storm is growing, but the reflectivities are still well shy of the levels I had identified earlier in the day as significant - 50dBZ near 7km for the current environment (this changes depending on the freezing level, time of year, available moisture etc). However the storm has maintained its integrity with moderate reflectivities aloft and is displaying characteristics of being an organised multicell. No reports at this stage so wait for the next radar image.

The radar scan shows reflectivities are still below thresholds, however it is still well defined and has remained intact for at least 40 minutes. The lightning detector is going-off and the storm is approaching Dunkeld. Hold brief discussion with colleagues regarding the above characteristics and that the cell is slow moving and showing signs of being organised. Decide to issue Severe Thunderstorm Advice.

Jane rings 5 to 10 minutes later with an observation of flash flooding near Dunkeld from Tim Eckert. Fantastic I say (although not for the people being affected). SES ring a half hour later with reports of flash flooding along the highway between Dunkeld and Colac.The storm continued to track to the southeast with the steerage flow and finally moves out   over Bass Strait near Cape Otway. I finally get to finish work at about 8.30pm after finalising the Advice.

Other cells did develop in the southwest during the evening, however non reached the same intensity on radar. The report from Tim in this instance was crucial as it helped to confirm the existence of a severe thunderstorm and I wasn't left wondering whether I had 'cried wolf' again.

With the winter season almost upon us the forecast process for forecasting severe thunderstorms is a little different. It's almost impossible to wait until a storm is identified on radar before issuing a warning or advice as 'coldies' tend to be so short lived. Hence, warnings or advices are commonly sent out in advance when all the ingredients for severe weather are forecast to be present. We tend to over warn, but still miss some (eg. Aspendale and Leopold tornadoes last year). So keep an eye out for those strong shear days and you just never know!

You will find a PPI and RHI scan through one of the cells within the multi-cell structure. The elevation is 0.5 degrees for the PPI scan. The RHI scan shows reflectivities of 45 dBZ to almost 5km with tops to 10-11km. The '3-D tops' image produces a compilation of the all the scans at the various levels. The local time for the images is about 1710hrs (EST). The 3-D tops image clearly shows the most interesting cell in dark brown - depicting the height of the storm above the surrounding convective activity.

A diagram of the Mt Gambier trace in the morning with surface conditions modified to that at Hamilton around the time of convection (T=28, Td=11). Nice chunk of CAPE!

This is a good example of how even though the atmosphere was unstable in  the afternoon, it was the presence of the surface trough that defined the necessary convergence for convection.

April Forecast Outlook, Discussion & Report page