We all worry about the weather from time to time, but where business is concerned these worries are often heightened. Take the example of a ski resort operator, which is trying to determine the best location for its newest hotel. Ideally, it’d want to build its new property somewhere that receives abundant snowfall during December, so that it can charge premium rates to those that want to enjoy the privilege of shooshing down the slopes.

It’s not just the hospitality industry that wants to predict the weather of course. In agriculture, farmers can make a mint by taking advantage of accurate, long-term weather insights to plant the highest-yielding crops in the right place and at the right time. Then we have insurers, who like to analyze weather history in order to determine risk. The list goes on and on.

See the entire Future Big Data Series on Pinterest and Springpad!

.

Many businesses can benefit from accurate weather forecasts, but unfortunately few fields of expertise generate as much public ridicule as that of the meteorologist. Weather forecasters have been the butt of jokes for years, constantly getting things wrong – and sometimes even spectacularly so, such as in the UK in 1987 when the BBC’s Michael Fish insisted that “there is no gale on the way” just hours before the most destructive storm in a century lashed the country, causing millions of pounds in damage.

We all agree that weather forecasting is a tricky business, but meteorologists do have one MASSIVE factor in their favor – there’s masses and masses of data just waiting to be analyzed – and just as in any other industry, the more data there is the less likely it is that mistakes will be made.

Big Data Gets Under The Weather

One place that could benefit from more accurate weather forecasts is South Korea. Menaced by so-called Hwangsa storms carrying thick clouds of polluted dust from China’s deserts on an annual basis, the country has been busily upgrading its national weather information system in order to understand the climate better and predict where and when extreme weather events might take place.

The Korean Meteorological Administration (KMA) recently boosted its Big Data storage capacity by a whopping 1000% to 9.3 petabytes, becoming the largest storage system in the country. The system collects more than 1.6 terabytes of meteorological data every single day, including barometric pressure and temperature readings, wind speeds, and imagery from balloons, aircraft, satellites and ships.

In order to cope with this Big Data deluge, the Koreans have turned to IBM and its latest data analytics technologies. IBM’s General Parallel File System allows Korea’s weather men to record up to 20 GB of data a second, crunch this data in real-time, and then cross-reference it with maps and historical data to come up with vastly more accurate forecasts than before. KMA is now able to produce detailed weather forecasts for each of the nine regions in the country, with real-time data processing meaning that it can update these predictions immediately as and when conditions change.

In fact, the system is so powerful that the KMA is even planning to offer localized weather forecasts for businesses within South Korea, providing insights for clients as diverse as hotels, golf course operators, farmers, insurance agents and the emergency services.

“The volume of meteorological and satellite data is so vast that it requires the most powerful technologies on earth” said Jung-Uk Tak, of IBM Korea.

“IBM’s smarter computing system helps KMA and NMSC to take a proactive approach to changing weather conditions, supporting its mission of protecting and enhancing lives and business.”

Closer to home, the San Diego-based startup EarthRisk Technologies is also attempting to get a handle on the weather, but for now it has a specific climatic problem in mind – extreme heat.

The startup has designed TempRisk, a web-based application that aims to predict periods of extreme heat up to 40 days in advance by identifying factors linked to the occurrence of such events.

Long-term forecasting is evidently much more difficult that trying to guess what the weather might be like tomorrow, or even next week, and so TempRisk is a hugely ambitious project. Much of the focus for meteorologists (like in Korea) is on more immediate forecasts as these are more important to the majority of people, yet having advance warning of extreme weather could be hugely beneficial to some businesses. For example, energy suppliers could better prepare themselves for a heat wave by purchasing additional fuel for auxiliary generators in anticipation of the extra energy demands.

But how does one go about making long-range weather forecasts? Simply, by gathering up as much historical data as there is, and then analyzing it to try and identify patterns that indicate extreme heat. EarthRisk says that these patterns are like dominos that line up in a specific manner, before setting off a chain reaction of events that leads to unusually warm weather. What TempRisk does is attempt to spot these ‘dominoes’ as they’re lining up, then second guess when they’re about to fall down.

We’ll have to wait and see how successful it is. TempRisk is a specialized tool that will only interest certain types of businesses, and the very idea of using Big Data to dodge weather risks is such a novel concept that it’ll take some selling. But EarthRisk has confidence in its abilities to do just that, so much so that it’s already moved on to its second weather forecasting project, working alongside Colorado State University scientists to try and improve the accuracy of hurricane forecasts.