Know the flow

Researchers at University College London are due to start a three-year project to develop a radar-imaging instrument and processing algorithms to better understand the flow of avalanches.

The team, which will be working on the technology with Leeds and Cambridge universities, hopes that with a greater understanding of the fundamental physics of avalanche movement, safety can be improved in areas of the world where they are prevalent.

Radar measuring technology exists, but it is limited to providing the average velocity of an avalanche.

The advantage of the new technology is that it will use phased array multiwaveform radar, so the team will be able to take 100 measurements a second from five different avalanche waveforms to create a 2D animated image that should provide more information about the effects of friction and fluctuations in velocity.

'The advantage of radar is it can see below the powder cloud,' explained UCL's Prof Paul Brennan. 'If you use an optical camera you can only see what's on top which is a very light powder cloud. With this advanced version [of the radar] operating at 5.3GHz we should be able to penetrate the powder cloud and see what's going on underneath and in particular see the velocity distribution of all the components.

'The system currently being used is only a range radar so it can't give a 2D impression of the avalanche and I gather the resolution is only 50m which is very coarse. We are proposing a resolution of 1m that can also image in azimuth so we can generate an animated reconstruction of the avalanche. All in all, the instrument will be much more sophisticated.'

The velocity of the components will be measured using the Doppler effect, which tracks the change in frequency of waves emitted by an object as it moves.

Taking 100 measurements a second during an avalanche, which lasts for around two minutes, Brennan believes the team will be able to track the changes in velocity, and by using multiple waveforms can create a better picture of the flow by analysing the cumulative data provided by each one.

'We will have five different animations of the avalanche, produced from a different waveform and each will have different properties,' said Brennan. 'We can then merge the results from the different waveforms to give us more information and try and resolve some of the ambiguities that you would otherwise have with radar.'

Using this mixture of waveforms means the team can use one waveform with high Doppler properties for getting a good reading of the velocity with another that provides a better image resolution and will give a sharper picture of how the particles move. The animations can then be overlaid to create a picture of the avalanche that provides all the information the researchers require.

'It's all a trade off,' said Brennan, 'depending on the number of frames a second taken. If we take 100 a second I think we can use five waveforms — 10 images a second means we could use 50 waveforms. I think five will be about right.'

The project, which is due to start this year, initially received funding from the Royal Society and recently secured further cash from the Natural Environment Research Council (NERC).

Dr Chris Keylock, of Leeds University's School of Geography, and Dr Jim McElwaine, of Cambridge University's Department of Applied Mathematics and Theoretical Physics, are also involved in the project.

'The two main aims of the project are to get high-resolution data on the flow that we can interpret in terms of process, and to see how well existing models operate,' said Keylock. 'Current models are based on basic principles of mass and momentum with some friction. but there are no existing data that really allow you to get at the nitty-gritty of what process dynamics actually are.

'The radar allows you to image the whole flow which is the real difference. If the physics in the model are better, then presumably that will give more accurate forecasts of the probability of an avalanche reaching a particular point or the probability of it destroying a particular house.'

The team believes the system could be applied to similar particle movements including landslides and the movement of lava from volcanoes.

'The idea is to take the instrument to a test site in Switzerland where we can image some avalanches,' added Brennan. 'If that goes well we will take it to Montserrat where we can try it out on volcanoes. And as we've got a few winters to play with we will probably return to Switzerland to refine the process.'