What goes into making a forecast?

Do you ever wonder how a meteorologist puts together a forecast?  Or do you think we all just have a big dart board in the weather center and however good our aim is will determine what the weather will be like?  :)

A good forecaster knows that you can't always rely on what the weather computer models tell you.  You have to understand what's going on not only at the surface, but also in the upper levels of the atmosphere.  One of the things we're taught in school is to use the forecast funnel.  Look at the big picture; where are the long wave troughs and ridges, how will it evolve over time and how will this affect what's going on at the surface.  The funnel on the left represents the different scales of the atmosphere: Hemispheric, Synoptic and Mesoscale.  The funnel on the right represents the amount of time you should spend on each.  In the hemispheric scale, we want to look at the overall jetstream pattern and how it will evolve.  The synoptic scale represents the weather producing features; fronts, low pressure, high pressure, etc.  Finally, the mesoscale represnts the processes that are going on closer to the surface - usually 2 to 2,000 km ( is there going to be any dry/moist air moving into my forecast area - cold/warm air, current radar trends, etc). 

One of the things a forecaster wants to do is compare a current analysis- surface, mid and upper, to how well the computer models initalized.  Weather computer models are simulations of the atmosphere.  They use current weather observations as a starting part, such as data from upper air balloons (launched twice a day), surface observations and many, many mathmatically equations, then project the state of the atmosphere in the future.  The problem that forecasters run into is that sometimes models don't initialize very well and unfortunately that will skew the forecast.  Once we've identified whether or not the models have initialized well we then will begin our forecast by looking to see how the jetstream and weather systems will evolve through time.  At times, weather computer models will give a different end result and it's up to forecast to determine which models perform the best given the current situation or use a consensus of them all.  (That's why the further out a forecast is, 5-7 days, the more likely it is going to change)

Another thing forecasters rely on are conceptual models, experience and past weather recognition.  If we see a similar weather pattern begin to set up as one in the past, we begin to ask ourselves what kind of weather did it produce, what was the precipitation or how much did we receive.  From there we can begin to identify if something similar will happen.

Overall, there is a lot of work that goes into putting a forecast together.  A good forecaster will do a hand analysis, compare it to the weather computer models and identify if the weather pattern is similar to one from the past that produced a significant weather event.