Google+

Weather Predictions of Yesterday and Today

by: Bradley Verdesca, Production Staff

Go ahead and take a quick look outside at the sky for me. Was it the same as it was 24 hours ago? How about 2 hours ago? Now, can you tell me what the weather will be in the future? This is a question that scientists have spent centuries trying to answer. Some questions have been answered through pattern recognition with no help from technology at all. For example, many scientists agree that prehistoric sites like Stonehenge in England and Chaco Canyon in New Mexico were constructed based on solar and lunar patterns. Publications like the farmers almanac use past weather patterns to provide seasonal weather predictions. Sure, they can tell you when the longest days will occur or possibly which week  will be the hottest this year, but tools like these are nearly useless when trying to predict conditions such as temperature inversion layers for the greater southeastern region or what day in the week a flash flood will occur.

The biggest advancements in our daily weather predictions have come from the help of mathematicians and their use of complex algorithms. 200 years ago in 1814 a French mathematician named Pierre-Simon Laplace hypothesized that the movement of every particle in the atmosphere is predictable as long as meteorologist knew the position of every particle. In theory, the laws of physics that govern these particles is fairly simple, unfortunately, the number of particles in the Earth’s atmosphere is estimated to be around 100 tredecillion. A one followed by 44 zeros, like this:

100,000,000,000,000,000,000,000,000,000,000,000,000,000,000![1]

A man by the name of Edward Lorenz, in light of Laplace’s postulation of prediction based on current conditions, was going about his career as a meteorologist in the late 50’s when he discovered even more about forecasting. Through research by himself and his team he found two things. First, weather is nonlinear, meaning that it abides by exponential rather than by arithmetic relationships. This means that even a slight change at input can have a dramatic effect on what the algorithm spits out.  Second, it’s dynamic. Meaning its behavior at one point in time influences its behavior in the future. This is the start of a branch of mathematics called Chaos theory which is described in Lorenzs’ breakthrough paper in 1972 titled “Predictability: Does the Flap of a Butterfly’s Wings in Brazil Set Off a Tornado in Texas?”[2] As you can see, predicting the future is, well, complicated.

Today’s technology is miles ahead of where we were in 1972. In 2008 the National Center for Atmospheric Research installed a supercomputer named Bluefire in their research facility in Boulder, CO. This computer has a peak speed of more than 76 teraflops[3]. But that isn’t enough. On January 5, 2015, NOAA announced that their computers will undergo significant upgrades this year and by October of this year each of their supercomputers will have a peak capacity of 2.5 petaflops for a total capacity of 5 petaflops or 5 thousand trillion floating point operations per second[4]. In basic terms this will help improve accuracy as well as increase amount of days out it can predict for smaller regions.

As stated above, the accuracy of inputs can greatly affect the accuracy of the outputs made by our forecasting models. The era of Big Data is upon us and can positively influence our forecasting methods. For example, Samsung has a thermometer (temperature), barometer (pressure), and hygrometer (humidity) built into their Galaxy S4 phone[5]. While not as accurate as a weather observation system like those found at airports, the large number of readings have the possibility of creating or confirming prediction trends made by NOAA and the NWS.

Just this month there was a satellite launched from Cape Canaveral by Space X that will continue the advancement of weather prediction. Named the Soil Moisture Active Passive satellite, this remote sensing technology will better track soil moisture levels. NASA says “The amount of water that evaporates from the land surface into the atmosphere depends on the soil moisture. Soil moisture information is key to understanding the flows of water and heat energy between the surface and atmosphere that impact weather and climate.”[6]

Even with all of our recent advancements our prediction of weather can still be wrong. Recently, experts predicted a record snowstorm for the city of New York. Some even forbode of snow upwards of 2 feet causing massive delays and possibly shutting down the city. In the end most people saw around 6 inches of powder. When we are planning our flights we need to use the best available information and a lot of luck to get the best imagery possible.