Forecasters use a number of different weather models which can be fed with data from a number of different Agencies ( like GFS and ECMWF ). There are high resolution models that will analyse the weather over a period of time for a very small region, or you might have a low resolution model for more long range ( in terms of time ) outlooks. There are just so many variables in the model ( wind speed, temperature, humidity, air pressure for instance ) that no two forecasts ever seem the same. There's also an old adage in the industry that if you gave the same data to 20 forecasters, you'd get 20 different forecasts ! The trend in the industry is to use an 'ensemble' which is a sort of merging of a number of different models to produce an optimised average.