Information Technology Reference
In-Depth Information
As a next step, meteorologists apply wellunderstood principles
of physics (e.g., fluid dynamics) to their data in order to project
where various weather systems will move and how those systems
will affect future weather conditions. Computers play an important
role in this process because:
1.
Forecasting depends heavily upon applying detailed mathe
matical formulae to massive amounts of data.
2.
The accuracy of the forecasts generally increases as the num
ber of observation sites increases.
Meteorologists aim to analyze the weather with as much preci
sion as possible, so the more data they can utilize, the better. More
data, however, mean more computation is needed in applying the
formulae. Without computers helping, computation on this scale
would take years or more, and your daily forecast would be years
out of date! This second phase of forecasting meshes nicely with the
capabilities of computers to perform computations efficiently and in
a timely manner. Computers are effective at number crunching, and
forecasting is worthwhile only if the millions of computations can
be performed quickly.
To be somewhat more precise, forecasting normally proceeds by
dividing a geographical region into relatively small blocks in a
threedimensional grid, and data are collected for at least one point
in each block. As an analogy, think of a house as being divided into
rooms; you could record the temperature and humidity in each
room to get a general picture of the comfort level of each location
in the house. A more precise picture might include variations of
temperature within a room. To be this specific, you might divide
each room logically into four or more sections and take readings in
each area. These additional readings would offer you new informa
tion; for example, they might show the locations of drafts of cold or
hot air.
In weather forecasting, to determine weather patterns, compu
tations are made for each geographical block (or room or section of
a room) following wellknown formulae. General trends can be es
tablished when these formulae are applied repeatedly to determine
changes every six hours or every day. More accurate and detailed
computations can be made if changes are noted every hour or every
fraction of an hour. Throughout the work, the same formulae are
used, but accuracy depends upon dividing the area in question into
