Staying afloat in a flood of statistics

Weather forecasters get a bad press, even though their record of accuracy is good

Weather forecasters get a bad press, even though their record of accuracy is good. And while the Irish coast had a lucky escape last weekend, the flood warnings will probably prove right next time, writes DICK AHLSTROMScience Editor

IT SOUNDED like time to take to the boats, given last weekend’s dire flood warnings. An impending storm coupled with high tides and heavy rains had councils around the coasts on high alert of flooding. Yet in the end nothing much happened. They were the floods that never arrived.

So how did the weather forecasters get it so wrong? People will certainly cast their eyes upwards and blame Met Éireann, but in fact its forecasts were highly accurate. What saved us from the floods was, simply, good fortune.

“We were lucky last weekend,” says Gerald Fleming, head of forecasting at Met Éireann. Although we had the high tides, with rain, strong winds and a storm surge all together and capable of causing considerable damage, “we didn’t get everything coming into phase together”.

READ MORE

Our weather forecasters actually enjoy a high level of accuracy, in keeping with figures for meteorological services around the world. In part, this global accuracy is because of the data-sharing that now takes place between the various services, according to Fleming.

Meteorological bodies constantly measure the success of their forecasting, and this influences how far ahead they attempt to predict the weather. Met Éireann typically gives a forecast stretching five days into the future, but it also attempts 10-day predictions, according to Fleming. If you ask the public, they will claim that the forecasters always get it wrong, but in fact this is not the case.

“We are getting better all the time,” Fleming says. “We are gaining a day of predictability every decade.”

So what are the numbers? A meteorological prediction for one day into the future is accurate more than 90 per cent of the time. For two days ahead, accuracy decreases to 80 per cent. There is then a gradual decline, with accuracy for five days ahead falling to between 60 and 65 per cent.

Modern forecasting is hugely dependent on supercomputers, powerful machines that are able to suck in weather data from around the world and then spit out a selection of predictions for assessment by forecasters.

Met Éireann is a contributor to the European Centre for Medium-Range Weather Forecasts, based in Reading, England, as are most of Europe’s meteorological services. “Reading has the big computers and the number-crunching capacity,” says Fleming.

Data comes in from satellites and ground and sea locations. It is then fed into complex weather models. To improve accuracy Reading runs an “ensemble” of 51 forecasts, with each of these applying a slightly different set of starting conditions. It is then back to the forecaster. “This is where the human can still add a bit of value,” explains Fleming.

When the differing predictions have been studied it is up to the forecaster to assess which way the weather will go. “You have to make a call on the most likely scenario,” says Fleming. A forecaster looks at probabilities but also looks back a few days. “You try to build that into the forecast.”

The result is a personal view of how the weather will evolve over the next five days.

This is why the conditions last weekend looked set to deliver a “perfect storm”, with all of the ingredients necessary to deliver heavy flooding. There was a deep depression forming out in the Atlantic and coming our way. This promised both high winds and heavy rain, which were due to hit just as particularly high tides reached our shores.

The wind and the very low pressures seen in a major weather depression combine to deliver something called a storm surge, explains Fleming. The wind can push the water along, piling water up ahead of it. If the wind happens to be blowing towards land, sea levels will rise once shallow waters are reached.

Low air pressure also contributes to higher sea levels. Typical air pressure at sea level is 1,013 millibars, a figure that falls sharply during a big storm. For every millibar the pressure falls, the water rises by one centimetre, according to Fleming. A bad storm can see sea levels rise by half a metre because of this effect.

Then there is the rain. Downpours fill up water courses as they flow towards the sea. But if the sea is already above normal levels there is nowhere for this water to go other than sideways on to land.

There was strong wind and low pressure last weekend, so the storm surge moving towards land was considerable. Luckily for those living near the coasts, the winds did not blow towards land at the critical time when tides were at their peak. The sandbags and other sea defences were not needed this time but, as sure as rain, the next big storm is only around the corner.