Good morning everyone. Today's
lecture is a topic that affects all of us, weather forecasting. Humans have
tried to predict weather for millennia, but only in the last couple of decades
have we been able to predict weather with any accuracy. Thanks to more
sophisticated computer models, stronger understanding of mathematical theories,
and a growing volume of statistical data, meteorologists can now predict
weather several days in advance. This can be especially important when
understanding the trajectory of typhoons or other powerful weather patterns.
The first people who tried to
predict the weather were Babylonian priests, who based their predictions on
cloud formations and astrological information. In 340 BC, Aristotle wrote about
weather patterns in Meteorologica, and this is where we get the modern
term “meteorology.” Philosophers and scientists from India, China, and the
Middle East continued to predict weather in the following centuries. However,
none of them improved on the Babylonian idea of using readily observable
information, such as cloud formations, lunar cycles, or the movement of wind.
Weather is a global event, and
one of the great obstacles for weather prediction is having accurate data from
far away. Until the telegraph was invented in 1835, information couldn't move
faster than weather, so any predictions being made were no better than those in
ancient Babylon. As the telegraph became more popular, a kind of prediction
became possible; mainly that if there was a weather event, say, 100 miles away,
information could be relayed to towns further away.
Those telegraph predictions
were still not very useful and it would take two British men, Francis Beaufort
and his protégé Robert Fitzroy, to provide the critical instruments necessary
to collect proper weather data. Beaufort, an officer in the British Navy, is
known for developing the first of these, the wind force scale, which
empirically measures the relationship between wind speed and weather
conditions. If any of you are studying New Zealand history, you probably
already know that Robert Fitzroy was the captain of the HMS Beagle and the Governor of New Zealand from 1843 to
1845. But he also invented the barometer, which measures atmospheric pressure.
This was the second important innovation in predicting weather.
In the early twentieth
century, mathematicians proposed that, with sufficient data, accurate weather
predictions could be achieved. However, computers had not yet been invented and
human beings were not able to quickly and accurately analyze the necessary
amount of data for a reliable forecast. In 1950, the mathematician John Von
Neumann became the first scientist to perform a computerized weather forecast.
His program was capable of making 5,000 calculations per second, but his
predictions were no more accurate than a random guess.
If a butterfly flaps its wings
in Brazil, it could result in a tornado across the world. You might have heard
a similar description of what’s called the 'butterfly effect,' which is used to
describe chaos theory. This is one of the reasons weather prediction is so
difficult. The Earth's atmosphere is a
dynamic system, meaning it is very sensitive to conditions, and the initial
conditions of a dynamic system influence its future behavior. Thousands upon
thousands of tiny factors, such as a butterfly flapping its wings, can
contribute to changes in atmospheric conditions.
Forecasters at this point knew
they needed to account for chaos theory, but computers still weren't fast
enough to analyze the amount of information needed to create an accurate
forecasting model. In fact, in the 1970s, when chaos theory was identified as
an important factor in weather prediction, forecasters were still missing the
high temperature by over 2 degrees on average when trying to predict it three
days in advance.
However, with the advent of
supercomputers, meteorologists were able to create programs that could handle
massive amounts of data. Prediction failures are now half of what they were in
the '70s, and today a three-day prediction will be off by less than 1 degree on
average.
It is not just temperature
that forecasters have gotten much better at. Predicting typhoons has become
much more accurate. In the 1970s, if a forecaster attempted to predict landfall of a typhoon three days
in advance, she would miss by an average of 560 kilometers. Today the accuracy
has been refined so that a three-day prediction will only miss by 160
kilometers. This has been a great help for governments deciding which coastal
regions should be evacuated.
Meteorology
has come a long way since the astrologers of Babylonia. Advances in
communication and measuring equipment, along with powerful new computers, have
helped make predictions much more accurate. Although weather predictions are
still not as accurate as they could be, the rate of success has slowly and
steadily improved, and on average a meteorologist's predictions are 350 percent
more accurate than they were just 25 years ago.