Newly found weather records show 1930's as being far worse than the present for extreme weather

Plot of NOAA/NCDC state high temperature records by decade with atmospheric CO2 concentration overlaid. From C3 Headlines with thanks – click to visit website

The Heat Was On—Before Urbanization and Greenhouse Gases

By Patrick Michaels on World Climate Report

Sure is hot out! And what better time for a paper to appear in the Journal of Applied Meteorology and Climatology describing the construction of the “all-time” records for various types of weather extremes for each of the 50 United States plus Puerto Rico and the U.S. Virgin Islands. The paper details efforts of the U.S. State Climate Extremes Committee (SCEC) established by the National Climatic Data Center (NCDC) and led by Dr. Karsten Shein. Basically, the SCEC dusted off old records and found other new sources. So now we have “new and improved” data (available here) for the value, the date and the location of the all-time high and low temperature, greatest 24-hr precipitation, greatest 24-hr snowfall and greatest snowdepth for 50 states and two territories. The statewide record extremes have been updated through 2011 and are subject to continuous updating.

This paper is an interesting read for those who perseverate on climate history and how it is constructed from a variety of observations both made from “official” (federal) observing stations as well as those deemed reliable from “non-official” observations (such as 12-oz soda bottles or credible “amateur” observer accounts). The new effort resulted in “the revision of 40 percent of the values” contained in the old dataset at NCDC and “underscored both the necessity of manual quality assurance methods as well as the importance of continued climate monitoring and data rescue activities to ensure that potential record values are not overlooked.”

It also is useful for putting the recent heat wave in perspective. Despite the 24/7 caterwauling, only two new state records—South Carolina and Georgia—are currently under investigation. And, looking carefully at Shein et al. dataset, there appears to be a remarkable lack of all-time records in recent years.

This is particularly striking given the increasing urbanization of the U.S. and the consequent “non climatic” warming that creeps into previously pristine records. Everything else being equal—and with no warming from increased greenhouse gases—most statewide records should be in or near big cities. But they aren’t.

This year there were a huge number (many thousands) of reports of daily high temperature records being set across the eastern two-thirds of the country in recent weeks, and even a large number (a few hundred) reports of all-time records high temperatures being set for a particular location. But if only two new statewide records were set, that’s hardly an historic heat wave when considered in its totality.

In Table 1, below, we list the all-time record daily maximum temperature observed in each of the 52 entries (as compiled by the SCEC) and the date and location where it was recorded. Notice that the vast majority of the all-time records were set more than half a century ago and that there are exceedingly few records set within the past few decades. This is not the picture that you would expect if global warming from greenhouse gas emissions were the dominant forcing of the characteristics of our daily weather. Instead, natural variability is still holding a strong hand.

Table 1. All-time statewide maximum temperatures (from NCDC)

In Table 2, we’ve compiled the top five years when the most records were set. When multiple years tie for the high, each individual year gets a fraction of a “record”. So, for example, 1954 and 1933 each get a half of a record for Colorado.

Table 2.

But this doesn’t stop people from implying that last week’s heat wave as an indication that global warming is leading to unprecedented conditions.

Capital Weather Gang (CWG)—the popular and respected weather blog for conditions in and around Washington DC, and one which is closely watched by the media, was quite vocal all about all-time records of one sort or another being set in our Nation’s Capital during last week’s heat wave.

If the Shein et al. methodology is applied to DC’s temperatures, then CWG’s very public pronouncements (they were picked up on the Drudge Report) are not all going to stand. That’s because CWG relied only on a single record, while largely ignoring the comprehensive set of observations historically taken within the geographical boundaries of the District of Columbia. The single record used by CWG is the “official” version of the Washington DC daily temperature which is a record which has been stitched together from observations made at National Airport (from 1945 through the present), which by the way is not even in the District of Columbia, and from observations taken at a Weather Bureau location at 24th and M street (1889 through 1944, and other locations prior to then). But when the records were concurrent (which they were during the 1940s and 1950s), only one is included (DCA).

If you really wanted to establish all-time records for Washington DC, you’d have to consider all available records that are credible—rather than relying on a data for a single “station.”

That’s what Shein et al. did. Although the SCEC has not yet compiled the all-time weather records for Washington DC, the word is that they are in the process of doing so, and are considering all available observations.

The CWG should do the same when discussing records for “Washington DC”. Or at the very least, they must be very clear that they are discussing a single (changing) location (i.e., Reagan National Airport, the downtown City Office, etc.) rather than Washington DC as a whole.

Here is an example of how things can go awry.

According to CWG, the recent heat wave “Washington D.C.” tied its record for the longest string of consecutive days in which the daily high temperature was 100°F or above. According to the CWG, the record was/is 4 days set in 1930 and 2012. However, in July/August 1953, 5 days in a row with temperatures of 100+°F were observed at the old Weather Bureau observing station at the City Office. These observations were from “an” official weather station within DC but not part of “the” “official” stitched together record. If the NCDC SCEC were compiling all-time strings of consecutive days of 100+°F, they most certainly would consider the old City Office records (including during the time of overlap with DCA observations), something that the Capital Weather Gang opted not to do.

Whether or not additional examination would alter any of the other “all-time” temperature that the Capital Weather Gang identified as being broken in “Washington DC” during the recent heat wave is unknown at this time.

One lesson here is that when considering “all-time” extreme weather records for a particular region, a comprehensive study must be undertaken (as described by Shein et al.) rather than simply deferring to a single station record.

The other take-home is that one has to be very careful about attributing the recent extreme temperatures to dreaded global warming. As noted above, there are surprisingly few all-time state records in recent years. Further, a look at their table indicates that only one of these—Providence RI, in 1975—comes from a city. Somehow—and this seems impossible—the dreaded greenhouse effect cannot raise already climbing urban temperatures to state record levels.

We can thank the SCEC for helping to do most of the dirty work in establishing an accurate dataset of all-time statewide record extremes for the United States that can be relied upon into the future, so that accurate assessments can be made when comparing current extreme weather events to past ones.

Reference:

Shein, K., D. Todey, F. Akyuz, J. Angel, T. Kearns, and J. Zdrojewski, 2012. Evaluating Statewide Climate Extremes for the United States. Journal of Applied Meteorology and Climatology, doi:10.1175/JAMC-D-11-0226.1, in press.

4.5 11 votes
Article Rating

Discover more from Watts Up With That?

Subscribe to get the latest posts sent to your email.

111 Comments
Inline Feedbacks
View all comments
July 16, 2012 8:04 pm

barry, before arguing it it warming, you might want to consider this table:
Using NOAA data (mean temperature), how cold are the last 12 months (Jul 2011 to Jun 2012) compared to the warmest month.
For example, October 2011 in Alabama was -13.6F colder than the warmest October in Alabama – 1919
October 1919 was 74.0F
October 2011 was 60.4F
http://sunshinehours.wordpress.com/2012/07/16/how-cold-is-the-48-states-for-each-month-compared-to-the-record-temperature-noaa-data/

Robert Craigen
July 16, 2012 9:06 pm

That bar graph looks to me as if it is flipping the IPCC the bird.

Albert
July 16, 2012 9:32 pm

I find it extraordinary to read ”newly found records”
I studied Meteorology in 1967 and this ”newly found” was common knowledge.

barry
July 16, 2012 11:17 pm

Sunshine,

No. The mean temperature record for a month for a state has been set 120 times since 2000.

My method is pretty simple. To check your tally, CNTRL F for page search, type in 200, and keep clicking which gives all values for 2000 – 2009 inclusive. Then search for 2010, 2011, and 2012 individually. Just count the clicks (watching for any anomalous cursor jumps).
2000 to 2009 = 86
2010 = 13
2011 = 16
2012 = 30
Total is 145 records broken since 2000.
For my tally above (the last decade from 2003) I clicked for the years 2000, 2001 and 2002, and subtracted that from 145. Doing that again, I get;
2000 = 2
2001 = 10
2002 = 4
Total for 2003 to present, 9.5 years, is 129.
I think that’s the *decade* with the most hot record breakers acording to your chart.
(Be pleased for you and anyone else to check the numbers)

July 17, 2012 6:42 am

barry, the table I present has the regions and the USA 48 at the bottom. They aren’t states.

barry
July 17, 2012 7:30 am

Ah yes.
Actually, that doesn’t matter. All the data applies to regions, states, contiguous 48. Including those metrics might even be a good idea.
But to round off, the results for the states only are:
1930 to 1939 = 99 max temp record-breakers
2003 to 2012 = 106 max temp record breakers
Far as I can tell, the last ten years has seen the most record-breakers by the metric you’ve nominated.

oldfossil
July 17, 2012 9:14 am

Thankyou Barry. The chart at the top of this page cannot be replicated by selecting any other dataset from the ncdc site. Therefore, ergo, QED, it’s cherry-picking.
Now for a serious question, though I don’t expect a serious answer. The way the greenhouse effect works is that the earth radiates most incoming energy, back into space. On the way out, some energy is absorbed by CO2 (and other) molecules, which in turn radiate the energy again. Some of this re-radiated energy will go back in the direction of earth, partially offsetting the outward energy flow. Okay, got that.
But surely then, the same CO2 molecules must also absorb INCOMING radiation, and “bounce” most of it back into space? To my rusty (and admittedly never very sharp) mind it would seem that the reduction of incoming radiation should more or less balance the reduction of outgoing radiation, taking us back to the status quo ante.
Next question. I’m looking at the diagram on the Wiki page http://en.wikipedia.org/wiki/File:TyndallsSetupForMeasuringRadiantHeatAbsorptionByGases_annotated.jpg. This experiment is the foundation of the Greenhouse Gas theory. It proves that CO2 absorbs energy, because when CO2 is introduced into the gas tube, the radiant energy measured by the thermopile DROPS.
A simplified version of the apparatus would look like this:
Heat source —- >>>> —- gas tube —- >>>> —- thermopile.
Let’s change the names and see what happens:
Sun —- >>>> —- atmosphere —- >>>> —- earth’s surface.
According to the Tyndall experiment, introducing CO2 into the atmosphere should reduce, not increase the earth’s surface temperature.
I’m a newbie to climate science with not even undergraduate science and engineering training. So obviously I’ve missed something. The solution can’t be that simple. Can it?

July 17, 2012 10:23 am

barry,
1930 to 1939 = 99 = 9.9 per year
2000 to 2012.5 = 120 = 9.6 per year
The big question is why are all those old ones still standing?

July 17, 2012 12:52 pm

Oldfossil,
Why dont you just write to Karsten Shein at NCDC and tell him he’s making things up. Be sure to provide evidence, too.

July 17, 2012 12:55 pm

Oldfossil,
Thanks to the Wein Displacement, the sun radiates maximally in the ultraviolet, which largely is unabsorbed by carbon dioxide. The earth radiates maxmallly in the far infrared, which is absorbed.
Get it?

barry
July 17, 2012 5:02 pm

Sunshine,
If you start at 2003 and compare a roughly equal period (10 years), then the latest decade pips the 1930s. If you include the three years prior to 1930 so that the time period matches 2000 to 2013, then the more recent period is still ahead.
1927 to 1939 = 109
2000 to 2012.5 = 120
There is no good reason to compare different length of time periods. But the differences are probably not significant. The reason there are still many records standing from the 1930s is what I remarked on in my first post this thread. I don’t know how to do the mathematical analysis, but I am reasonably confident that the same principle would apply over averaged regions as large as an entire state. The last decade in the US is on average hotter than the 1930s, so it is likely – but not inevitable – that the number of max temp records broken will be greater than then. It will still be possible to slice and dice the data in such a way as to present the opposite, but if all the data is examined, that will give a clearer picture of what has happened. Short answer – the real world doesn’t slavishly follow statstical probability estimates. But you’ll get closer to the mark with more data.
You mentioned UHI effect – you realise that NCDC make a downwards adjustment in their long-term data to account for this? You may speculate, of course, but it’s just muddying the water without proper analysis. Anthony Watts co-authored a paper which corroborated the mean temperature time series of the official records using a suite of best-sited weather stations. That paper discussed problems with the max and min temp trends, finding that

“Temperature trend estimates vary according to site classification, with poor siting leading to an overestimate of minimum temperature trends and an underestimate of maximum temperature trends, resulting in particular in a substantial difference in estimates of the diurnal temperature range trends.”

According to that paper – and hinted at by Menne and others – maximum temps may be overestimated in the past or underestimated in recent times. I’d guess the former problem is more likely.

1 3 4 5