• Subscribe

The forecast: no fire for tomorrow

Image of the island of Lesbos in VENUS-C fire-risk prediction tool.
This is a demonstration of the fire-risk map that was built for the opening of the Microsoft Cloud and Interoperability Center in Brussels in March 2011. This video shows the system being used for the island of Lesbos. In this example, the demo was preloaded with weather data from the previous summer. The selected dates show when real wildfire events happened on the island with 50 outbreaks that summer.Blue areas shows low risks of fire and green high risks. Click the image to play the video. Image courtesy Microsoft Europe.

Imagine a weather forecast with a difference: "There may be risk of fire in the northwest, with forest fire outbreaks in the south."

That's exactly what researchers at the University of the Aegean, Greece and the European Microsoft Innovation Center, Germany - partners in the EU Framework Program 7 (FP7) VENUS-C (Virtual Multidisciplinary Environments Using Cloud Infrastructures) project - have done. They combined a fire risk algorithm with a parallel computing cloud infrastructure to create a model that can predict fire risks five days in advance, much like meteorological reports do for the weather.

Forest fires are a big problem for the South of Europe. "Almost seven percent of fire ignitions in Portugal, Spain, France, Italy and Greece resulted in 85 % of burnt areas," said Christos Vasilakos, a researcher of geography and natural disasters at the University of the Aegean, Greece.

Fire prevention is one of the most important aspects in managing natural hazards like wildfires. "Every year thousands of acres of forest burn down threatening peoples' lives and property. VENUS-C is addressing these kinds of issues. It's giving computing resources to relatively small universities that have big ideas and supports them in running their algorithms in a massive data center," said Ulrich Pinsdorf, a program manager at the European Microsoft Innovation Center.

VENUS-C is supporting a wildfire application for prediction and prevention purposes. This is just one of many areas of research that VENUS-C supports within its infrastructure for scientific cloud computing. The application was developed by the University of the Aegean, which has 12 years of experience in wildfire research.

Predicting further into the future

The university started work on creating algorithms that could systematically predict wildfire outbreaks back in 2002. This research was led by Kostas Kalabokidis, a senior geography researcher, who helped start up the University of Aegean's wildfire research program. A breakthrough came in 2009 when research funding was granted through the use of high-performance computing resources given by Microsoft Research and its European Innovation Center. This gave the researchers more computing power; they converted their desktop-based algorithm into a parallel computing one. Now, they could predict fire risks up to five days in the future, instead of just for one day.

Today, their wildfire prediction application combines Microsoft's Bing Maps, Microsoft Silverlight (a front-end application for multimedia, graphics, and animation), and Windows Azure, for managing the cloud computing resources at the back end.

"Around the world, there are many fire risk algorithms being developed in places like the USA, Canada, Russia and Southern Europe. What makes our tool different is it provides a quantitative and systematic approach, based on Geographic Information Systems, whereas, for example, Greece's fire risk scheme is qualitative or empirical based. Our model can predict fires at an hourly rate too," said Kalabokidis.

When making a prediction, their system computes 120 images and converts them into a film that is overlaid onto the map. "The complex calculation takes into account the vegetation of an area, topography, social economic parameters and, of course, the weather," said Kalabokidis. This includes 40 years worth of historical data. Vasilakos uses 'neural networks' to enable his algorithms to 'learn' and find patterns for future predictions.

All these calculations are dispatched to the Windows Azure cloud for processing.

Image of an array of servers used for wild-fre risk calculations in the VENUS-C application.
A user of the VENUS-C wild-fire risk application can manually select how many computers they need in a calculation just by touching the icons on screen. Image courtesy Microsoft Europe.

An operator can manually choose how many computers are involved in each calculation through the Microsoft Silverlight interface. The Windows Azure cloud shows a visual display of an array of 64 powerful servers.

"A government operating this system doesn't need to own, maintain or operate the physical machines," said Pinsdorf. "You can not only simulate what's happening, you can change global and local parameters. Researchers can do a 'what if' analysis. For example, changing local weather conditions will predict alternative scenarios. Governments can clearly plan their actions in a better way."

The tool has already been used to help local Greek authorities deal with real life wildfire outbreaks, so they can know at least sixty minutes ahead of time what's ahead of them. There has even been interest from Russia, which experienced destructive wildfires last year.

There has even been interest from Russia, which experienced destructive wildfires last year.

The VENUS-C project will finish in May 2012, and the University of the Aegean is currently working on predicting the behavior of fire, known as fire propagation, which simulates fire growth and flame length.

Kalabokidis said, "You wouldn't think it, but even countries like Germany, Belgium and Switzerland experience forest fires. European countries have different approaches to fire risk. Now, they're trying to develop a unified wildfire risk strategy. Our project is just one brick in the wall of fire risk prediction."

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2019 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.

Republish

We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.