Friday, November 5, 2010

When is a record, REALLY a record?


This week we have experienced several weather records. On Monday, Seattle had the wettest November 1 since record keeping began (1.45 inches) and on Wednesday it had the warmest November 3 on record and tied the ALL TIME daily record for any day in November. A reporter called up this week asking whether "all this weird record breaking" is a sign of something...like global warming.

The truth is that there are records and then there are records.

Breaking daily records happens all the time.

Heavy rain is associated with storms and for the relatively short records available for major stations (roughly 60 years at Sea Tac) is not hard for a few days to be missed. Then if you get a decent storm on that day...you break the all-time daily record! Really not the remarkable.

Let me demonstrate this to you. Here is a plot of the daily record precipitation at Sea Tac Airport (blue line) and the daily average (green). You notice how spikey and irregular the plot of record precipitation is. In November some days have daily records of 3-4 inches, while some are substantially less than an inch. Breaking the record on one of the low days is no big deal...any moderate storm could do it. Is that a record to be touted in the press? I don't think so.


On the other hand a record like Wednesday's high temperature is more significant. We tied the all time daily record for the month--of ANY day. But perhaps not as significant as one might think since the temperatures really plummet after the first week of the month. So most of the month really doesn't count.

Want to see that? Below is a plot of the daily records for November, and the temps so far from the National Weather Service web site. The first four days have some fairly toasty records..with a peak of 74F on November 4 (occurred in 1949) and now November 3 as well. But from the fifth on there is not a single day the temperatures got in the 70s and the records rapidly drop into the lower 60s. Bottom line: we have had a number of years where the first week of November has had some pleasantly warm temperatures, but then it ends decisively after November 5. The record mimima are really uneven after November 10th.

Finally, it is useful to keep in mind that the temperatures are rarely normal. Normals usually represent the average of above normal and below normal temperatures. Want to see proof? Here is a plot of the actual temperatures the last two weeks (dark red line) and the average highs and lows. The high temperatures are rarely normal but are found below (early part of the period) and above (later part) of "normal." The two week period probably has an average maximum close to normal. You see how much that tells you.

Kind of like kids in a some high schools...very few are average! (sorry, couldn't help myself)

10 comments:

Dano said...

So, yes or no: are this sort of wx and all increasing record hi temps indicators of man-made climate change? Yes or no: this the sort of thing that happens in a warming/changing world?

That is what the reporter needs to know. That is helpful for the public.

(word verification agrees: orpro)

Best,

D

Andrew Taylor said...

You might need to explain to your younger readers what the picture at the start of this post represents.

Hint: it's a phonograph record. However, the "vinyl" records that some of us still have are flexible and almost unbreakable.

It was only the pre-vinyl (12 inch, 78 rpm) records that were actually easily breakable. I recall purchasing my first (and last) 78 rpm record in about 1958 [when pop singles were available in either breakable [78rpm] or modern unbreakable [45rpm] format.

WAVEGURU said...

Cliff kinda touched on one of my pet peeves. Average verses normal. It is very normal to be well above or below average, and it would be very abnormal to have even several days in a row that were right on the average. Normal is not average and average is not normal?

lemon said...

It might be interesting to see if there's a trend in how many record highs (and lows) are broken each year. Like maybe in the 80s we averaged 20 new record highs per year, but now, due to climate change, we're seeing more?

catman306 said...

Does anyone keep track of rain during total rainfall events rather than daily rainfall totals? A rainfall event may last several days. The quantity and totals of such 'events' might be a good way to track climate change.

Laurel said...

My understanding is that climate change is supposed to create erratic weather--more extreme weather--both colder and warmer. That's why it's now called climate change and not global warming. So if extremes are spiking in both directions, it will tend to average out and look the same, and so how can we judge it by averages? Just curious. If one observes the introduction of new warmer climate species and other indications in nature, something seems to be going on...

Wx Enthusiast said...

lemon, this doesn't exactly answer your question, but here are interesting facts about existing daily record highs and lows that are similar to what you asked.

Note that observations at Sea-Tac airport began in 1948 (63 years of records including 2010).

- Over half of all calendar days during the year (190 days) have a current record high that was set since 1987 (in the past 24 years).
- Exactly half of all calendar days (183 days) have a record high minimum set since 1990 (the past 21 years).

It's important to note though that these numbers include 37 record highs and 35 record high minimums that were ties of records set before the years mentioned above, so the numbers are not quite as impressive because of that. I have not yet figured out these numbers with tied records.

- More than half of all calendar days (201 days) have an existing record low maximum set between 1948 and 1972 (25 years); they have not been broken or even tied since. 29 additional days have existing records set in 1972 or earlier, and tied since.

The most impressive statistic, by far, involves record lows. Over half of the calendar days during the year - 191 days - have a current existing record low that was set between 1948 and 1956 (nine years of records)... none of which have been broken or tied since. 36 additional calendar days have existing records that were originally set between 1948 and 1956, and only tied since, never broken. That means that 227 days of the year (62%) have daily record lows set during that nine-year period that have not been broken since.

December, with 5 days, is the only month of the year with less than ten record lows still existing from this time period. Seven months (January, March, April, June, July, August and September) each have over 20. In July and August combined, 51 of the 62 days in these two have current record lows that still exist from that nine-year period (43 have never been tied, the other 8 have been tied but not broken).

Lindsey said...

"Finally, it is useful to keep in mind that the temperatures are rarely normal."

Thank you for stating this, Dr. Mass! So why do the meterologists who should know better constantly refer to "normal" temps in their official weather Discussions? . . .

Mattias said...

Wx Enthusiast-

Those are interesting stats, but it's important to remember the Urban Heat Island effect. Back when records began being taken at Sea-tac, the area was much more rural than it is now. All that extra concrete there now holds the daytime heating through the night and prevents the area from getting as cold at night as it used to. I'm not saying Global Climate change couldn't have anything to do with the decrease of record lows in more recent years, but UHI is pretty important too.

This is why using statistics from cities can be pretty misleading over long periods of time.

Wx Enthusiast said...

Mattias- Note I didn't say anything about the cause of those numbers, I only stated statistics. They are what they are.