A forecasting system at the Bureau of Meteorology has made it to the finalists in the 2012 Australian Government ICT Awards in the ‘Excellence in eGovernment’ category.
The system, NexGenFWS (Next Generation Forecasting System), is an automated program that provides forecasts based on information from a digital forecast database which is fed from Bureau of Meteorology (BoM) numerical weather models, including variables such as temperature and wind speeds.
Essentially, the system means forecasters are able to work directly on a computer screen and can manipulate data in real time to provide automatic forecasts. Previously, forecasts were carried out using numerical weather models and forecasters would have to manually write forecasts.
“It was a very time consuming part of their workload. This is a major paradigm change in that the forecasters can now actually work with the science that are part of the weather model,” Ray Canterford told Computerworld Australia, division head – services at the BoM.
“Once they’ve agreed and modified the information from the model to fit it to local topography and local conditions, they can then just push a button and provide a whole range of forecasts for anywhere across Australia.”
The technology was originally developed from the US, with modifications and improvements made by the BoM to suit Australia’s weather conditions.
“We had what we call a much larger domain area that each office is looking at. This provided some technological challenges as well, [but] we’ve improved some of the software quite significantly,” Canterford said.
For example, the US system operated in about 120 forecast offices in the country, compared to Australia’s seven capital cities.
The system was also adapted to operate in the southern hemisphere and the underlying database within the NexGenFWS server was changed to allow multiple clients to access the system and to reduce contention issues.
The pilot for the system began in October 2008 in Victoria. Due to the nature of the BoM, which operates on a 24/7 basis, the trial could not have any downtime. This required the BoM’s IT specialists and research assistants to develop a robust system which could be integrated with virtually no problems.
“I think mainly the challenges in the Victorian system was that it was new technology. We had to do some additional training for our staff to work with the system,” Canterford said.
“We also had to ensure that our stakeholders, like fire agencies and emergency services, were aware of the type of information they could get.”
After witnessing the success of the Victorian trial, the Federal Government announced it would continue to provide finance for the project in the 2009 budget over five years.
NexGenFWS is now operating in four states in Australia – New South Wales, Victoria, Tasmania and South Australia – with plans for it to be rolled out in Western Australia in October this year and in Queensland in October next year.
While it may seem to have taken some time to work its way across the country, Canterford said this is due to an intensive three-month period of training for forecasters, and the installation of the system also takes another three months to ensure it operates correctly.
“We allow an elapsed time of about 12 months, with periods of intense activity leading up to an installation,” Canterford said.
Eventually, NextGenFWS will provide seven-day forecasts for more than 650 locations across Australia, compared to seven-day forecasts for only seven metropolitan locations before NexGenFWS was implemented.
The BoM is also developing a MetEye application which will be available for the public to trial in the near future. The application will be a complete GIS-based system on Open Layers and will replace the current Forecast Explorer system currently being used, which is a Web-based map viewer of pre-computed images of data.
After three-and-a-half years since the initial trial, Canterford is quick to point out it has been a challenging project which has involved collaboration between a large group of people.
“It’s been a major team effort by our research scientists and our IT specialists working with our forecasters,” he said.
Follow Stephanie McDonald on Twitter: @stephmcdonald0
Follow Computerworld Australia on Twitter: @ComputerworldAU