Weather--Hi-Temps, PoP's, Precipitation Icons and Me

I'm starting to have my doubts.  In just three weeks, I will be moving into a home that stands about 35 yards above a small body of water next to a marsh, which connects to Broad Creek, which connects to the Calibogue Sound, which connects to the Atlantic Ocean.  At that point, I will be dependent upon weathermen for my survival, so to speak.  No, not Bob Dylan's Weatherman from his Subterranean Homesick Blues (not needed to know which way the wind blows) and not the infamous Weather Underground (although a bunch of them still hang out and offer advice to us all, most notably, Bill Ayers, the professor of Education at University of Illinois Chicago, famous more recently as an Obama campaign fund-raiser/contributor) 

No, I am talking about our trained meteorologists.  Those who slave away staring at computer models, trying to make sense of all those shifting shapes and colors on the LED screen (I hope they at least aren't still using CRT's, I mean if you're going to relentlessly spend tax money and money you someday hope to tax, and more, the least you can do is spend it to provide good, modern equipment to the National Weather Service, so I can have as  much warning as possible of any inclement weather (like a flash flood, for instance), or even a hurricane.  

I sat down to write this while the sun was blazing away, despite the TV forecaster's prediction last night of three days of rain.  I was all set to run him down for being totally wrong, etc. when it finally started to rain about 6PM.  But still, I wondered, were all the local forecasters singing the same song that night, and how about the governmental arm, the National Weather Service?  I started looking around, and I couldn't find any of yesterday's forecasts to review. 

I dug a little deeper and could find little evidence that anyone was looking back.  By contrast, the British government's web site for its Met Office (their national weather service) devotes multiple pages and graphs to rating its forecast accuracy both in terms of hi-temps and lo-temps and PoP's (that's probability of precipitation) on a continuing basis, and includes accuracy goals as well, for key data, such as high and low temps, rain or sun days, etc.

Now, most of you probably believe that these certified meteorologists on TV have high levels of accuracy on their resumes.  How else would they get opportunities like those that land them on your TV?  It turns out that only ratings drive those decisions, not forecasting credentials.  I know, I know, you are shocked and amazed that weather, like the news has always been about ratings.  And don't you believe that the stations pay some person with the proverbial "radio face" to sit in the back room and do the technical work so a face (and, yes, sometimes a figure) can appear on TV and give us a "good-looking" forecast.  In fact, no TV meteorologist posts their accuracy statistics on their résumé.  No station managers use accuracy statistics in the hiring or evaluation of their meteorologists.  They are looking for charm, personality and looks, and I am living 35 yards from the water! 

But do the stations keep those records and evaluate their own  performance?  It turns out that no meteorologist or television station keeps records of what they predicted, and-- therefore they are incapable of comparing their predictions to actual results over the long term.  The station managers will tell you the past is past and people only want to know what the weather is going to be tomorrow.  This latter fact is among the principal reasons why mid to long-range forecasts (going out beyond five or seven days) are so inaccurate.  No one cares about the weather beyond tomorrow and maybe the next day.   

While the TV forecasters probably care about their accuracy, the only thing that yields complaints is when a negative event occurs--meaning they failed to display their precipitation icon on a given day, and the general public was surprised by a sudden downpour.  In other words, people are pleasantly surprised when rain is predicted, but is not forthcoming.  But angry and complaining when the reverse happens.  Given that bias, you might wonder a little why the TV weather personalities don't always predict some rain.  

However, if they did so, to minimize complaints, they would be right only about 27% of the time, since it has only rained on about 27% of all days in the past year on average.  If they wanted a simple way to appear right more often, they would be better off predicting dry days every day and they would be right 73 per cent of the time (100-27% days of rain=73% correct).  But that would also mean lots  of complaints received on 27% of all days. Since TV stations don't want their phone lines tied up with complaints, they tend to have a "wet bias," meaning they will show their precipitation icon for more days when it might rain than they would if they wanted to be perfectly accurate.  It turns out, for example, that the Weather Channel arranges its forecasts such that pleasant surprises outnumber unpleasant ones by a 2 to 1 margin (2012 data).  This "wet bias" is also observable by noticing TWC's threshold for displaying the precipitation icon is just 30%.  If the probability of precipitation equals or exceeds just 30%, TWC will predict rain.  This allows them to avoid unpleasant surprises for their viewers, subscribers, etc.  However, it leaves me wondering who the most accurate forecasters are and where I can find them when that nagging worry about being washed away in a storm pops up in my head.       

(Tomorrow--let's look at what private enterprise has started doing to study forecast accuracy and how to learn about the accuracy of the services you use when trying to divine the future.  The local forecast is for a dull, rainy day, and I believe it.  So what else will there be to do but write?)