Geoscience Reference
In-Depth Information
Africa, and no El Niño in the Pacific. As a result, the
number of Atlantic hurricanes and tropical storms is quite
variable (
Figure 7.18
).
Researchers in the United States
have had considerable success in predicting the number
of hurricane-strength storms in the Atlantic, including the
near-record year of 2005 with fifteen, suggesting that the
right factors are being included in the model.
Frequency / five years
60
10
20
30
40
50
1886 - 90
1891 - 95
1896 - 1900
1901 - 05
1906 - 10
1911 - 15
1916 - 20
We all know from experience how much daily and
seasonal variations in weather influence our lives. Clearly
it is useful to have an idea of the weather which is in store
for us. But how is it possible to foretell the weather? In
the past we relied heavily on folklore. 'Red sky at night,
shepherd's delight; red sky in the morning, shepherd's
warning,' says one country adage. 'When there's sheep-
backs [cumulus clouds] in the sky, not long wet, not long
dry,' goes another. These sayings sometimes contain a
grain of truth - that is presumably why they have survived
- but not enough for them to be reliable.
During the last century, as we started to understand
atmospheric processes in more detail, methods of fore-
casting became more sophisticated. The main approach
used today involves understanding the basic physical
processes of weather formation and expressing them as
mathematical equations. Unfortunately, although we
know many of the basic laws, and can express them
mathematically, the equations that result are difficult to
solve. Only with the development of super-computers
has it been possible to tackle this mind-stretching task; in
1921 the first attempt to forecast weather twenty-four
hours ahead in this way, without computers, took several
months!
To forecast values of pressure and wind, the globe is
subdivided into a grid consisting of about 60,000 squares,
each with 217 points from one pole to the other and 288
around most latitude circles. For each point, the upper
atmosphere is subdivided into thirty levels and the values
of the critical atmospheric properties are determined,
mainly from satellite information. The physical equations
of motion, continuity and thermodynamics are applied to
each grid point at each level to predict the new value a
short time period ahead. The new data set then provides
the starting point for the next set of predictions, and so
on every fifteen minutes until the twenty-four-hour or
forty-eight-hour forecast is produced. Clearly a vast
amount of calculation is required, though, with the speed
of modern computers, six-day forecasts take only about
fifteen minutes of computing time. Realistic results are
1921 - 25
1926 - 30
1931 - 35
1936 - 40
1941 - 45
1946 - 50
1951 - 55
1956 - 60
1961 - 65
1966 - 70
1971 - 75
1976 - 80
1981 - 85
1986 - 90
1991 - 95
1996 - 2000
2001 - 05
storms in the Atlantic Ocean, 1886-2005.
produced in this way, and all meteorological services now
use computer methods to predict future weather patterns.
Models are constantly being revised and improved as
computers and computing skills develop. Current models
use a four-dimensional variational scheme that uses the
statistics of the model and observation errors directly
to generate a model state that fits closely both the
observations and the model equations (
Figure 6.16
).
We might get the impression from this technique that
we can forecast the weather for the distant future, but that
