Fraud and Insanity - Welcome to Meteorology 2015, Part 1

There are several things going on in the weather business that make me cringe these days.  This field has become politicized and even worse of late, in ways I never dreamed of in my past, when I was nose down into my research.  I can only touch on a few of them, but the list is long, especially in the climate business - a can of worms I won't review here.

Most everyone now is accustomed to seeing 7-10 day forecasts, where the forecast high and low temperature, as well as clouds and precipitation are presented, usually in some graphical format (as shown in the example).

These show up on TV weather broadcasts but are quite common on the Internet.  No doubt many people look at these on a regular basis, most without any suspicion that they're being sold a pig in a poke - such an extended range forecast includes a fundamental dishonesty.  Here's the crux of the problem:  if anyone gives the topic even a few moments of thought, it should be obvious that the farther ahead the forecast product extends, the skill and accuracy of that forecast diminishes.  It's a well established principle in meteorology that every self-respecting meteorologist knows - it's an indisputable fact of the science.  A 3-day forecast is less reliable than a 1-day forecast; a 7-day forecast is less reliable than a 3-day forecast.  By roughly 10 days, the forecasts are essentially without any skill: they're no better than simply forecasting the long-term average conditions.  At that range, if you know the local climatology, you know as much as the forecast can offer.

There are scientific reasons for this growth of error with forecast time - basically, in about 10 days, any very small errors at the start of the forecast will grow to the point where they have contaminated the resulting forecasts.  And small errors at the start are inevitable - we don't know atmospheric conditions well enough to eliminate those small errors.

Furthermore, the detail contained in the forecast becomes less and less reliable with increasing lead times (that is, the time from the start of the forecast to the time when the forecast is valid).  A 1-day forecast has one day of lead time, for example.  The farther into the future the forecast goes, those details (such as the location of a front, or the region where precipitation might occur) become "fuzzy" - it's sort of like looking into a crystal ball, where the farther ahead you look, the cloudier the crystal ball becomes.  In the example above, the day-7 maximum (57) and minimum temperature (39) simply aren't known to a precision of 1 degree Fahrenheit, despite what's shown.  At 10 days or so, none of those details can be trusted beyond what you would expect from climatology for that location and time of the year.

Presenting this information with attractive graphics but without any indication of decreasing reliability is just not being honest.  It's a misrepresentation of the information.  Where does that "detail" come from?  For all practical purposes, these forecasts are derived from numerical weather prediction (NWP) models, where the equations governing the atmosphere determine the values on a grid of points in space (and time).  These models not only provide the forecast data used in those pretty graphics.  They also have been used to determine the limits of reliable forecasts I've referred to previously - roughly 10 days or so.

Any source of forecast information that doesn't let the user know about these limitations on extended forecast skill and accuracy is misrepresenting the science of meteorology.  In my view of things, the meteorological community should rise up in protest over this fraudulent practice.  The only hope to reverse this situation lies with the professional meteorologists, who need to stand up for what is right.  My blog can only reach so many and has little influence by itself.  If professionals can somehow muster the courage to speak out against this egregious practice - yes, it takes courage, as many feel their jobs are at risk when they speak out for the truth - we can change the way our forecasts are presented to the users, for whom we can have no expectation that they will figure this out for themselves.  It's our professional duty!

... more to come ...