# May 2018 was Warmest in Recorded History

• kyphysics

Suppose $X$ is a normally distributed random variable, and ##n## and ##m## are positive integers. What is the probability that
$$Max(X_1,X_2,\ldots,X_n) > Max(X_{n+1},\ldots,X_m)$$
where ##n## is the number of prior years we've been recording temperatures consistently and ##m## is the number of years in current memory we consider "recent times". Let's say a decade vs a century? That is the probability that such a headline occurs over the m subsequent years after n years of data.

Now do the same for 12 such normal RV's and ask what is the likelyhood we won't break a record of any type in the last P% of a given string of data.
One doesn't have to do the calculation though. Take your data over the history of the sampling and shuffle the time sequence and see how often the same interesting fact occurs after say 20 or so permutations. If its very rare then this "interesting phenomenon" might be significant. But one's intuition on a single event is not accurate.

... for perspective I just ran a little simulation. The number of times out of 12 where 10 copies of a normal random variable exceeded a previous 100 seems to range between 0.8 and 1.6. Very rough approximation using sum of 10 uniform RV's to simulate a normal RV and using Excel's notoriously bad RNG.

Now I'm counting out of 12 months but I should count out of a much larger pool of records which can be quoted as "evidence" over a given year. Highest annual minimums for example, Regional maximums exceeded.

My point here is that this headline is meaningless as "news" and is either implicitly or explicitly intended as propaganda. I'm making no judgement as to its virtue, just its merit. (Ok... I am judging its virtue because the intent is undermined if the method is without merit!)

If record highs got less and less common as time went on, and n got larger, you could be confident that they were randomly occurring. ("Shuffling the time sequence" is an effective method of burying that bit of highly relevant info.) Get back to us after you've had a look at the actual distribution of record highs over time.

BRVivian, BillTre and TTauri
Meh . Warmest May just after the coldest April ?

Beware of hyperbole.

To see whether climate is really up to something,
I'd rather see the phase margin between solar irradiance and a 30 day rolling average temperature.
I have yet to find that plotted.
Intuitively it seems Dec 21(minimum irradiance in Northern hemisphere?) to mid February(~Coldest week?) is about 48 degrees lag. That sounds pretty good to me.

NOAA says coldest day is earlier
the western half of the Lower 48 typically reaches its climatological coldest day in December, whereas most eastern stations reach their minimum in January. In addition, areas with higher snowfall Normals, such as the Northeast and high-altitude regions in the West, tend to reach their climatological coldest day much later, which is likely because of the increased reflection of solar radiation at the Earth’s surface due to the presence of snow cover.
https://weather.com/storms/winter/news/coldest-day-year-average-map-noaa-20141112

I'd be interested to know if lag has changed much in last hundred years.

old jim

If record highs got less and less common as time went on, and n got larger, you could be confident that they were randomly occurring. ("Shuffling the time sequence" is an effective method of burying that bit of highly relevant info.) Get back to us after you've had a look at the actual distribution of record highs over time.

You mean like so?

https://www.ncdc.noaa.gov/cdo-web/datatools/records

Looks like in the USA over the last year, monthly record highs are happening roughly four times as often as record monthly lows.

Globally it looks like the same sort of picture:

https://www.ncdc.noaa.gov/sotc/global/201705

...where April 2018 was the 400th consecutive month with temperatures above the 20th century average. (latest monthly briefing at https://www.ncdc.noaa.gov/sotc/briefings/201805.pdf )

It's always a problem to mistake local or regional weather for global climate, but I'd like to learn how/where any information could be "buried" in say the last 50 years worth of samples.

CalcNerd and BillTre
If you look at NASA's data here: https://en.wikipedia.org/wiki/File:Global_Temperature_Anomaly.svg
And then compare the 1979 to present data to the MGT data from the weather satellite scan: http://www.drroyspencer.com/wp-content/uploads/UAH_LT_1979_thru_May_2018_v6.jpg you lose all faith in anything NASA has to say. NASA supplies data to most everyone and garbage in = garbage out.

Before we proceed with addressing your claim, I'm curious, are you aware that your argument from authority is based on an authority who is as biased against science as it is possible to be and who has a questionable history of publishing papers?

In addition to having had papers withdrawn (and in one case the journal's editor resigned over publishing one of them), he is one of the signers of the Cornwall Alliance's declaration on global warming stating that it's impossible for significant climate change to happen or for it to ever be harmful because in his personal opinion, his gods would never allow that to happen

https://cornwallalliance.org/landmark-documents/evangelical-declaration-on-global-warming-2/

https://en.wikipedia.org/wiki/Cornwall_Alliance

You're citing his unsubstantiated personal blog claims as superior to the global scientific consensus and his one line of evidence against all lines of evidence from any other source.

That's an extraordinary claim that implies a global conspiracy. I just want to make sure that you're aware of the history of the authority you're basing such a claim on.

James Demers, CalcNerd, Bandersnatch and 5 others
Before this thread devolves into further nonsense, it is closed.