Geoscience Reference
In-Depth Information
and aims to interpolate objectively gridded model output
to a single station based on its climate and weather
history. Multiple regression equations are developed
which relate the actual weather observed at a station
over the course of time with the conditions predicted by
the model. With a long enough history, MOS can make
a correction for local effects not simulated in the model
and for certain model biases. MOS variables include
daily maximum/minimum temperature, twelve-hour
probability of precipitation occurrences and precipi-
tation amount, probability of frozen precipitation,
thunderstorm occurrence, cloud cover and surface
winds.
Various types of specialty forecasts are also regularly
made. In the United States, the National Hurricane
Center in Miami is responsible for issuing forecasts as
to hurricane intensity changes and the track the storm
will follow in the Atlantic and eastern Pacific areas.
Forecasts are issued for seventy-two hours in advance
four times daily. The central Pacific Hurricane Center
performs similar forecasts for storms west of 140°W and
east of the dateline. The US weather service also uses
numerical models to predict the evolution of El Niño-
Southern Oscillation which is important for long-range
forecasts (discussed below). Special events, such as the
Olympic Games, are beginning regularly to employ
numerical weather forecasting into their preparations
and to use regional models designed to be most accurate
at the single point of interest. MOS techniques are also
used to improve these very specialized forecasts.
weather stations (including buoys) that measure wind,
temperature and humidity supplement such data. In
addition, for detailed boundary layer and lower tropos-
phere data, there is now an array of vertical sounders.
These include: acoustic sounders (measuring wind
speed and direction from echoes created by thermal
eddies), and specialized (Doppler) radar measuring
winds in clear air by returns either from insects (3.5
cm wavelength radar) or from variations in the air's
refractive index (10 cm wavelength radar).
Nowcasting
techniques use highly automated computers and image-
analysis systems to integrate data rapidly from a variety
of sources. Interpretation of the data displays requires
skilled personnel and/or extensive software to provide
appropriate information. The prompt warning of wind
shear and downburst hazards at airports is one example
of the importance of nowcasting procedures.
Overall, the greatest benefits from improved fore-
casting may be expected in aviation and the electric
power industry for forecasts less than six hours ahead,
in transportation, construction and manufacturing for
twelve to twenty-four-hour forecasts and in agriculture
for two- to five-day forecasts. In terms of economic
losses, the latter category could benefit the most from
more reliable and more precise forecasts.
3 Long-range outlooks
The atmosphere-ocean system is a non-linear (chaotic)
system making exact long-term prediction of individual
weather events impossible. Small errors in the initial
conditions used to start a model simulation invariably to
grow in magnitude and spatial scale and the entire globe
will generally be affected by a small observational error
at a single point before long. Therefore, long-term
weather prediction and climate prediction do not try
to predict individual weather events for these would
certainly be in error. Instead they generally try to
represent the statistics of the climate rather than the
weather itself and are often associated with probabilities
based on statistical relationships.
Like numerical forecasting at shorter timescales,
long-range (monthly and seasonal) outlooks use a
combination of dynamical and statistical approaches in
order to assess the probability of certain weather
situations. Long-range forecasts rely on the idea that
some types of weather, despite being unpredictable in
their details, may, under certain circumstances, be more
likely than in others. One major recent advance in
2 'Nowcasting'
Severe weather is typically short-lived (<2 hours) and,
due to its mesoscale character (<100 km), it affects
local/regional areas, necessitating site-specific fore-
casts. Included in this category are thunderstorms, flash
floods, gust fronts, tornadoes, high winds (especially
along coasts, over lakes and mountains), heavy snow
and freezing precipitation. Mesoscale models with grid
cells which can be less than 10 km on a side are used
regularly to study such phenomena in detail. The devel-
opment of radar networks (Box 4.1), new instruments
and high-speed communication links has provided a
means of issuing warnings of severe weather within the
next hour. Several countries have recently developed
integrated satellite and radar systems to provide
information on the horizontal and vertical extent of
thunderstorms, for example. Networks of automatic
