How to Accurately Forecast the Weather

For over a thousand years people have attempted to predict the weather and officially since the early nineteenth century. In modern times weather is predicted by collecting data about a current atmospheric conditions in a given location. Then using scientific knowledge of atmospheric processes and with the aid of computers; weather patterns are predicted.

Ancient weather forecasts were not very reliable, becasuse the methods of observed patterns of events underwent constant change. Over hundreds of years developed weather lores proved helpful, but were often incorrect. Such as predicting good weather from a bright and clea sky in a certain hour. The invention of the electric telegraph in 1835, allowed reports of weather phenomena from a large region to be received almost instantaneously and classified by the late 1840s.

Accurate weather forceasting consists of:


At the surface, atmospheric pressure, temperature, precipitation, wind speed and direction and humidity was made by observers, which have now been replaced by accurate computerised sensors. These weather stations, which collect data hourly in METAR reports, or every six hours in SYNOP reports. This is done using standardised instrumentation, that has been callibrated according to the World Meteorological Organisation standards.

Weather balloons that carry radiosondes record wind, temperature and humidity of the atmosphere of the region being investigated in the mid-stratosphere, at an altitude of about 21000 metres. Aircraft Meteorological Data Relay has also collected information and transmitted back to Earth for numerical weather predictions. Satellites are also being used to cover a a much broader spectrum of weather events.

However, satellites are mostly limited to cloud analysis and the infrared data of cloud tops. This helps to explain cloud movement together with wind speed and direction. Geostationary and polar orbiting satellites provide good information on weather soundings of temperature and water content through the whole atmosphere. Although the larger global scanning spectrum comes at a cost, as resolution is often less than in a radiosonde. Regular Radar and pulse Doppler weather Radar are also great aids, as they provide vital information on rain location and intensity, as well as wind speed and direction.


Information is collected from various meteorological sensors gained from the observations is used in conjunction with established numerical model forecasts. This is used to produce a representation in multiple dimensions (usually in 3D) of many weather aspects such as moisture, wind and temperature. This becomes a good estimate of general conditions.


Numerical weather prediction models are computer generated scenarios and simulations of the atmosphere according to specific feedbacks and criterias that fit in the regular data that is collected. The atmospheric behaviour is predicted in the future using fluid dynamics and general physics. The amount of data used and the complexity of equations that simulate the changes of the atmosphere require powerful supercomputers to solve.


Once predictions are finished by supercomputers, the results are fitlered amd modified to remove errors and known biases as well as the ability to be able to incorporate different and changing conditions. Statistics are also used to interpret the final result, which would allow for easy access and understanding for anyone interested in the forecast.

There are numerous techniques, that are incorporated to provide fairly accurate weather predictions. These are:

Physical observation; which includes visible changes such as rain, cloud cover, lightning, winds, and increase in temperature are all indicative of weather changes and can be incorporated into models for weather predictions. This is, however, outdated.
Persistence; his depends upon the presence of persistently same weather conditions. It is useful in short term predictions.
Nowcasting; using latest observational, radar and satellite data, a short range (6 hour frame) and more accurate analysis can be made of upcoming weather events, such as showers.
Use of a barometer; the bigger the pressure change (more than 3.5 hPa), means a larger weather change. Dropping pressure indicates cold fronts, while increasing pressures indicate clearing conditions.

Forecast models; are chosen according to various parameters, such as weather conditions, urgency, etc. This helps to reduce error, however, heavy interpretation is required and extra details are added. Information that supercomputer analysis neglects or cannot solve.
Ensemble Forecasting; is the production of many different forecasts of the same and differing conditions. This is done to analyse any atmospheric uncertainty and to minimise errors, which can be assessed by all the different forecasts.

Analog Technique, a difficult methods, which requires forceasting upon experience from a past weather event, which is expected to be similar in behaviour. This is rather difficult as not weather events are ever the same