More on Forecasts
In the US, the National Weather Service (NWS) runs several different forecast models and also collects the real time data of existing weather conditions. There is also a model run by the Canadian Meteorological Centre and another by the European Centre for Medium-Range Weather Forecast
None of the models are perfect because our understanding of the atmosphere is very imperfect. So a meteorologist experienced in the weather patterns of a specific area will look at the output predictions of the model and apply his/her judgment to create a forecast
NWS forecasters do this for their forecast but private-sector companies use the models to create their own forecasts which may, and often do, differ from the NWS forecast. Typically the differences aren't very large but in locations where conditions are complex, differences among forecasts can be large. An example is the Washington, DC/Baltimore, MD metropolitan area which is surrounded by mountains to the west and north, and the Cheasapeake Bay and Atlantic ocean to the east and southeast. All these geographical features affect weather patterns and make modeling weather forecasts difficult.
Here are some links for more information.
This article from the NY Times in 2012 is an excellent and readable summary of the evolution of weather forecasting and the thinking behind NWS forecasts versus private sector forecasts:
This article goes into detail about the NWS, Canadian, and Euro weather models:
Frequently Asked Questions (FAQs)
An Introductory Guide