Predicting the next big flood

The question most people have about the threat of stormy weather and more flooding is the one the scientists find hardest to answer: how bad is it going to be?

The fact is that the science of understanding why rain falls, and where and when, and how it then soaks into the ground or spills into the streets, is extremely challenging.

And Britain, as an island nation, tucked between the Atlantic Ocean and the continental mass of Europe, and caught in a tussle between mild maritime air and icy blasts from the Arctic, has the least predictable weather of all.

So, as the Met Office tells us what we guessed already - that 2012 was one of the top five wettest years on record - a huge effort is under way to improve the forecasting.

This could make the difference between floodwaters that are contained and those that ruin homes and wreck lives.

One initiative is research that involves bands of scientists braving the turbulence of flying through the heart of storms - the only way to measure what is happening inside them. I reported on one of these missions last year.

Further flights are planned this year with a focus on the south west of England where violent downpours have caused devastation in places like Boscastle in 2004.

Inside clouds

The rain that battered that Cornish village was far greater than the computer models had forecast so the scientists will try to get a better grasp of the forces at work inside the clouds.

One of the scientists leading the research is Prof Geraint Vaughan, of the University of Manchester, who told me that however good the models "we can't escape the fact that the atmosphere is chaotic".

One option is for weather centres to run 'ensembles' of models with different variables - groups of computer simulations to see if any come up with similar forecasts.

According to Prof Vaughan: "Sometimes the ensembles all look broadly similar (and different models from different weather centres agree) - then a forecast can be made with some confidence.

"At other times the ensembles diverge and confidence is much lower.

"Knowing how to construct and interpret an ensemble is something we still need to do a lot of research into, but in principle it's a big step forward in gauging uncertainty of the weather forecast."

So a really accurate forecast weeks ahead is unlikely - there are just too many different factors involved.

Water forecasts

The best hope lies in trying to identify the broader changes under way - for example with the retreat of the Arctic sea-ice or the cycle of warming in the Pacific Ocean with El Nino and La Nina.

Understanding their impact might make it possible to offer some odds on the kind of weather likely in a coming season.

But, as Prof Vaughan admits: "I don't think I'd want to plan my holiday on such a basis!"

So, what about trying to forecast not only the rainstorms but also their likely impact?

One project being planned for release later this year is the first of a series of 'hydrological outlooks' for the UK - forecasts of water conditions for the month ahead.

This is being prepared with data from every rainfall station across the country, together with readings from the UK's 1300 river gauges, plus information about soil moisture.

It involves the Met Office, the Environment Agency and the Centre for Ecology and Hydrology (CEH).

On Thursday I reported from the CEH on the latest findings about how much water has soaked into the ground - and saw a bore-hole used for measuring water levels actually spouting out water because the ground was so saturated.

According to Professor Alan Jenkins, of the CEH at Wallingford in Oxfordshire, the outlook will be "a first attempt to give stakeholders what they've been asking for - a guide to water conditions ahead."

"It obviously comes with huge uncertainties but it's about trying to give advance warning.

"As with all models you've got to ask - are they better than nothing? We won't know till we've tried one in anger."

Soil saturation

A key factor that is missing at the moment for any flood forecast is the level of saturation of the soil.

The current estimates are produced by analyzing a range of variables like rainfall, temperature and evaporation.

So scientists are planning a major new research network of moisture indicators - 50-100 devices across the country that would feed real time data on the amount of moisture in the soil.

Similar to a much larger network being rolled out in the US, and Britain's version, known as Cosmos, would cost in the region of £10m.

The rationale is that any flood forecast would be much more accurate if the computer models include not just rainfall and river flow but the level of ground saturation too.

The devices would be so-called cosmic ray moisture probes - small instruments standing about 2m above the ground - which would measure neutron intensity.

Basic physics describes how neutron activity is correlated with the presence of water and each instrument would cover an area of about one square km and down to a depth of about 50 cm.

Prof Jenkins estimates that if the project gets the go-ahead it could improve flood warnings "by several hours".

"Even an additional 2-4 hours of warning can make a big difference - you can get a lot of sandbags out in that time."

None of this will stop the barrage of criticism aimed at the authorities when they get forecasts wrong - but it is the start of the long and painful journey of trying to get it right.

This really matters. The Met Office count of days of extreme rainfall for the past half-century shows a slight increase in the frequency of the heaviest downpours. And warmer air can hold more moisture - which means more rain.

Lives, homes, businesses, travel - there's a lot at stake.

Soil moisture deficit (SMD) is the amount of water needed to bring the soil moisture content back to the maximum amount of water the soil can hold. For example, if you had a potted plant and you watered it too much and water leaked out the bottom of the pot, this would be a negative SMD. If you poured in too little water, you would have a positive SMD (the potted plant needs more water). Just enough and the SMD = 0