“Weather forecast for tonight: dark.” George Carlin was probably the only one who ever gave an accurate weather forecast. Knowing how incredibly unpredictable the weather can be it always surprises me how much people seem to love weather apps. Perhaps it is just me but never in my life have I ever looked at a weather forecast before wearing “appropriate” clothes or grabbing an umbrella.
I love umbrellas and perhaps the only reason I would ever use a weather app was if I needed to find a reason to carry one, but the umbrella is a versatile device: you can use it come rain or shine. On a more serious note, however, weather apps are, quite a lot of the time, little more than entertainment. You would have just as much fun as if someone made an app that predicted all your shots in a game of billiards. Sure, given all the variables to enough degrees of accuracy, you could predict precisely where the cue ball goes and how much you score, but there are so many variables you are bound to go wrong sooner or later.
Model Output Statistics, says the National Oceanographic and Atmospheric Administration (NOAA), is a technique used to objectively interpret numerical model output and produce site-specific guidance of the weather. And there are a huge number of variables involved, all of which are measurable, but all of which are constantly changing. (See an old sample from the NOAA’s technical procedures bulletin above.) Couple this with the fact that it takes time for data to transfer from the detection site to the weather people (some call them meteorologists) to their machines and human interpreters — whom we will talk about presently — and then for the softwares to get updated and then for data to upload and transfer to everyone’s smartphone and you have a surefire recipe for missing the mark, sometimes narrowly, sometimes widely, but missing the mark nonetheless. Continue reading