Monday, February 01, 2016

On My Soapbox Because I'm Not "Normal"

Please allow me a moment to get on my soapbox, then I'll go back to talking about the weather.

Too many times the phrase "normal temperature for this time of year" is thrown around.  There's nothing "normal" about a particular days high temperature.  For example, when you hear the "normal high temperature for February 1st is 52 degrees", it's a completely wrong way to describe what one may expect for that day.  In reality, there's much more variance involved

The National Weather Service bases that number on an AVERAGE of a 30 year period.  At this time, I believe that 30 year average is from 1980 to 2010.

Here are the high temperatures (in the second column) for every February 1st from 1980 to 2010


The so called "normal" high for this day is 52 degrees.  Notice during this period, it has never been "normal".  Not once was it 52 degrees... NOT ONCE!  It's an "average"...  plain and simple!  The average high temperature for February 1st is 52 degrees.
So I'm leading the charge to change the Climo summary from the NWS.  LOL.  In reality it's not a big deal, just a pet peeve of mine! 

Ok, I'm getting off my soapbox and I'll return to weather.  Data points to coldest air so far this season next week.  I'm not ready to buy that at this time.  I think it gets cold, but to what degree is in question.  See what I did there? LOL

What about snow chances?  When arctic air gets involved, it's always possible.  So it's really not going out on any limb to say we must watch for any moisture which gets involved.  More later.


Unknown said...

Excellent points, Todd. Same thing with body temperature; it is commonly believed that 98.6 is "normal" and that above (or below) is abnormal and means you are sick. My "normal" body temperature is about 97.8, almost a full degree lower. When I hit 99, I feel sick and am sick.

Matt Young said...

I think that "normal" in this case is used as a synonym for average. It is really just a semantics argument. It can be considered "normal" as the running average tells us there is an "average chance" that the temperature will be about X.

I calculated the average of that data to actually be 53, with a standard deviation of about 13 (it is possible that I missed some of the data). To me, a normal high of 52 in this context means that it is going to be between 39-65F with 95% confidence. I am not aware of the statistical model that is used to match the temperature distribution over time though, so perhaps the confidence is too high or too low.

I do think they should report the deviation from average. I think it would be more accurate to do so.

Anonymous said...

I agree with the semantics argument expressed by Mr. Young. Here's what you do, Todd. Why not post the "normal" or "average" on your broadcast, but then include a standard deviation. In the winter, that could be huge (big temperature swings). In the summer, I would think it would be small (swings of a few degrees). That will give viewers a better idea of what the temperature could be rather than just "normal" or "average".

The NEW Arkansas Weather Blog