July 2013 CET Forecasts / Competition

[Milhouse]

The first 18C+ month for some time seems to be odds-on now after corrections, this month looks like it's really going to smash the run of below average months.

 

On Temperature, Sunshine and Rainfall this month is performing very well. Definately the best summer month since July 2006 and could even end up warmer than the memorable August 2003.

[GRHinPorts]

maybe you need to study statistics a little more and it does not allow for chaos theory,if is very highly unlikely then it will happen.In these parts march was not the second coldest,was third and 1947 was much colder and had a warmer summer.1989 had a cold april and a hot summer 1995 I think was similar and even 2006 had a cold spring

 

To me it often seems that the years with the hotter Summers are preceded by cold weather in the late Winter or Spring before them. Is this just natures way of averaging out the whole year so that we always (nearly always) end up with an annual CET in the 9.5 - 10.25 range?

[SomeLikeItHot]

18.7 to the 24th (20.0) - on the daily list not on the front page yet.

 

Min today of 16.1 and max around 24.

 

After that using met forecasts for the CET region, with the GFS ensemble mean for the last 2 days.

 

25th 18.8  (20.1)

26th  18.8 (18.5)

27th 18.8  (18.7)

28th 18.8 (18.8 )

29th  18.7  (17.8 )

30th 18.7  (17.0)

31st  18.6 (17.5)

Edited July 25, 2013 by SomeLikeItHot

[BornFromTheVoid]

Made a spreadsheet of the CET monthly anomalies compared to the 1901-2000 average, then subtracted the anomaly of one month from the month before, to get the changes in anomalies from one month to the next on the CET record.

 

Here's a table of the largest increases and decreases in anomaly from one month to the next, and the average change for those months over the entire record.

 

 

So the biggest shift in CET anomaly from one month to the next occurred between December 1874 and January 1875, when the CET went from -0.2C (anomaly of -4.9C) to 6.4C (anomaly of +2.5C), giving a change in anomaly of +7.4C.

 

You can see that 1807 appears on the table 3 times. That's thanks to a run of months, from August to October, with CET's (anomalies) of 16.9C (+1.2C) in August, down to 10.5C (-3.0C) in September, and then back up to 11.4C (+1.2C) in October.

 

The average column (showing absolute average change, so ignores whether it's + or -) on the table clearly shows that winter months tend to see the biggest anomaly swings from one month to the next, with the average anomaly change from January to February being 1.8C, while summer sees a much lower change from one month to the next, with July to August and August to September anomaly changes averaging just 1.0C. This is mostly down to the standard deviation of anomalies in Winter being quite a lot larger (1.9C) compared to summer (1.1C)

 

Using the 30 year mean of absolute month to month anomaly changes, we can see if there are any long term trends in the variability from one month to the next.

 

 

There appeared to be a rapid increase in variability up to about 1875, since which it's been in decline, with a slight hint of things reversing in the last few decades. The 1800s clearly showed the highest degree of month to month anomaly variability, which agrees the the table showing it having the most records too.

 

 

Anyway, this was a little awkward to explain, so if anyone has any questions or wants more data, just ask!

Edited July 25, 2013 by BornFromTheVoid

[Eugene]

10th March- 9th April 2013 CET was 2.4C and 3 months later a July with a CET of 18.0+ is likely that's a 16C difference with a space of just 3 months which is pretty remarkable.

 

 

Yes that's impressive, definitely too much to ask for a 16C drop 3 months later, December maybe, 16C lower than July or more 

[hillbilly]

did any of you happen to see the documentary on ROGUE WAVES or the one on INFINITY.Basically the maths that we use to examine statistics cannot be used for every instance,there are always going to be exceptions that cannot be explained by maths and these are much more common than you may think.I put these last 6 poor summers down to that factthat an exception occurred and where there should have been a good summer or two average ones in there,they went missing and we ended in a run of 6 poor as a result,just as we had mild winters from 95 to 08,a similar occurance.

[Roger J Smith]

This is probably not the right place for a detailed discussion of probability theory, but the argument being made about rogue waves is in any case a separate argument from temperature data sets, proving that rogue waves do not form part of a predictable extension of non-rogue waves has no actual relevance to the statistical distribution of temperature anomalies. I think it's fairly well established that long-period temperature records show the sort of distribution around a mean value that is typical of random variation. Sometimes there are indications of modality (frequency peaks) that can be ascribed to circulation types. Conspicuous outliers are rare in long-term climate data sets, although not unknown. The warmest May (1833) is the most "outlying" extreme month in the 354-year CET data set and is something like half a standard deviation above the second highest value.

 

You get these outliers from time to time but I don't think they are as different from the pack as rogue waves are from storm waves. It's a case of never say never, but that is covered by statistical probability, not a refutation of it. Anything is "possible" in a frequency distribution (where not excluded by physical or logical laws, for example, a frequency distribution of rainfalls cannot include negative values). It is possible for the July CET to be 21 or 22 or even 10,000 given certain circumstances. If the mean is slowly changing, then the distribution will change with it. As BFTV has just shown, variability can change within data sets. But it would be less likely that modality would change, in other words, one century with a bell curve of temperatures and the next with a bimodal distribution.

 

However, the bottom line is that statistical probability is not a theory that can be refuted by any anecdotal evidence. It exists as a set of guidelines to what is likely, not what will actually happen. You can't use it to predict this week's lottery numbers but you can make a pretty close guess on how often numbers will show up in a decade of lottery draws.

[Lettucing Gutted]

For a rapid change in the other direction from exceptional warm months to exceptional cold months we have to go all the way back to 1985/1986 which had a 6.3C December followed by a -1.1C February and a 5.8C April.  For the second such example we have to go back to 1978/1979 which had a 8.5C November followed by a -0.4C January.

Edited July 25, 2013 by Craig Evans

[hillbilly]

This is probably not the right place for a detailed discussion of probability theory, but the argument being made about rogue waves is in any case a separate argument from temperature data sets, proving that rogue waves do not form part of a predictable extension of non-rogue waves has no actual relevance to the statistical distribution of temperature anomalies. I think it's fairly well established that long-period temperature records show the sort of distribution around a mean value that is typical of random variation. Sometimes there are indications of modality (frequency peaks) that can be ascribed to circulation types. Conspicuous outliers are rare in long-term climate data sets, although not unknown. The warmest May (1833) is the most "outlying" extreme month in the 354-year CET data set and is something like half a standard deviation above the second highest value.

 

You get these outliers from time to time but I don't think they are as different from the pack as rogue waves are from storm waves. It's a case of never say never, but that is covered by statistical probability, not a refutation of it. Anything is "possible" in a frequency distribution (where not excluded by physical or logical laws, for example, a frequency distribution of rainfalls cannot include negative values). It is possible for the July CET to be 21 or 22 or even 10,000 given certain circumstances. If the mean is slowly changing, then the distribution will change with it. As BFTV has just shown, variability can change within data sets. But it would be less likely that modality would change, in other words, one century with a bell curve of temperatures and the next with a bimodal distribution.

 

However, the bottom line is that statistical probability is not a theory that can be refuted by any anecdotal evidence. It exists as a set of guidelines to what is likely, not what will actually happen. You can't use it to predict this week's lottery numbers but you can make a pretty close guess on how often numbers will show up in a decade of lottery draws.

I can see now that you roger and maybe one or two others know what you are talking about and put me with my mad theories into the mad corner.Very interesting.Am sure there are many other large swings within a few months but not in recorded history.

[Scorcher]

If we hadn't had broken the all time record for April in 2007, April 2011 would possibly be considered as incredible as May 1833 as it was 1.2C warmer than April 1865 (now 3rd warmest) which is exactly the same gap between May 1833 (15.1C) and the 2nd warmest May, May 1848 (13.9C).

[Summer Sun]

Well it looks like we are going to smash the 2006 mean for July which was 17.8C this remains the hottest month on record for the UK according to the met office

 

July 2006 was the hottest month on record in the UK with a mean temperature of 17.8 C and saw the record July maximum temperature of 36.5 C at Wisley, Surrey on the 19th

[AderynCoch]

Impressive that we're still managing 20C daily means. The higher minima are doing their job of keeping the CET up despite the lower maxima.

[AderynCoch]

Well it looks like we are going to smash the 2006 mean for July which was 17.8C this remains the hottest month on record for the UK according to the met office

 

17.8C was the UK mean. The CET (Central England Temperature) was much higher at 19.7C. We're not going to get close to that, though it will still be an impressively warm July.

 

Does anyone know what the current UK mean is?

Edited July 25, 2013 by AderynCoch

[Summer Sun]

17.8C was the UK mean. The CET (Central England Temperature) was much higher at 19.7C. We're not going to get close to that, though it will still be an impressively warm July.

 

Does anyone know what the current UK mean is?

 

Ah thats me getting mixed up with the mean and CET

[GRHinPorts]

Sorry if this is off topic but I see people are mentioning the 1833 May CET figure of 15.2. I just wonder if both this figure and also the June 1846 CET figure of 18.2 might be seen as highly dubious so far above the norm are they.

 

May 1833 is 1.5 degrees above anything we have seen in all our generally considered warmer lifetimes. June 1846 eclipses every other June in sight with even June 1976 only "managing" 17.0 and that is the only time June has reached such a figure since 1900! 

 

Is it at all possible that back in 1833 and 1846 the measuring equipment of the day could not accurately measure extreme heat? i.e. both months were hot, but not that hot.

[Roger J Smith]

I think if that were the case we would see more examples from the 19th century, perhaps the actual reason other than random chance might lie in different albedo created by whatever agricultural trends dominated in central England around 1833 to 1846. The fact that these anomalous months are both in late spring suggests maybe a background signal of some sort arising from environmental conditions. It also seems to me that only the 17th century portion of the CET is ever really questioned as to exposure and screening issues and thanks to Mr. Maunder there were very few warm months and none that linger in the top five today.

 

As to what those trends might have been, not really my area of expertise, some crop that presents a dark absorbing surface in May-June that has seen a decline in production since those days would be what to look for. Alternately, perhaps with the high solar angle and the clarity of the sky in those early days of the Industrial age, a clear spell of weather might bubble up a little faster. I'm very familiar with Toronto records and the extremes of warmth from start of data (1840) to about 1895 were never out of line with the concept of a gradually warming climate. The summer of 1854 looks about like 1936 minus three degrees on the hottest days. So whatever was happening in Britain was not a global phenomenon affecting variability.

 

A third possibility that I could study further at my leisure would be a slowly-moving background signal that was peaking in May in the 1830s and June in the 1840s, perhaps went dormant for a while, traces could be inferred in July 1868 then August 1911. With the general warming of climate that background signal has probably been overwhelmed by a higher jet stream and might not be easy to find but one could go back from 1833 and look at earlier trends to see if it was there on some time scale. Seems like a 150-250 year sort of cycle. A fourth and final point would be that solar activity was high after the Dalton ended around 1827 and the 1837, 1848, 1860 and 1870 peaks were similar to 20th century intensity, albeit not as sustained. That warm May in 1833 was followed by a very mild winter.

[GRHinPorts]

I think if that were the case we would see more examples from the 19th century, perhaps the actual reason other than random chance might lie in different albedo created by whatever agricultural trends dominated in central England around 1833 to 1846. The fact that these anomalous months are both in late spring suggests maybe a background signal of some sort arising from environmental conditions. It also seems to me that only the 17th century portion of the CET is ever really questioned as to exposure and screening issues and thanks to Mr. Maunder there were very few warm months and none that linger in the top five today.

 

As to what those trends might have been, not really my area of expertise, some crop that presents a dark absorbing surface in May-June that has seen a decline in production since those days would be what to look for. Alternately, perhaps with the high solar angle and the clarity of the sky in those early days of the Industrial age, a clear spell of weather might bubble up a little faster. I'm very familiar with Toronto records and the extremes of warmth from start of data (1840) to about 1895 were never out of line with the concept of a gradually warming climate. The summer of 1854 looks about like 1936 minus three degrees on the hottest days. So whatever was happening in Britain was not a global phenomenon affecting variability.

 

A third possibility that I could study further at my leisure would be a slowly-moving background signal that was peaking in May in the 1830s and June in the 1840s, perhaps went dormant for a while, traces could be inferred in July 1868 then August 1911. With the general warming of climate that background signal has probably been overwhelmed by a higher jet stream and might not be easy to find but one could go back from 1833 and look at earlier trends to see if it was there on some time scale. Seems like a 150-250 year sort of cycle. A fourth and final point would be that solar activity was high after the Dalton ended around 1827 and the 1837, 1848, 1860 and 1870 peaks were similar to 20th century intensity, albeit not as sustained. That warm May in 1833 was followed by a very mild winter.

 

Hmmm....I dont really know enough to debate this further. All I can say is I think it would be truly exceptional if we were ever to experience a May like 1833 or  a June like 1846.

 

Nothing in any of our lives is remotely comparable to what those years delivered for that time of the year. They are simply unimaginable.

 

By comparison this July is pretty run of the mill since we only have to think back to 2006 for a rather similar month. 

[Scorcher]

 

Nothing in any of our lives is remotely comparable to what those years delivered for that time of the year. They are simply unimaginable.

 

 

 

I agree there isn't really anything that extreme, but as I mentioned in a previous post I think April 2011 (11.8C) comes closest as there have only ever been 2 Aprils that have recorded a CET over 11C (3rd place, 1865 is well below 11C). The April record stood for 142 years before it was broken in 2007.

[Weather-history]

Sorry if this is off topic but I see people are mentioning the 1833 May CET figure of 15.2. I just wonder if both this figure and also the June 1846 CET figure of 18.2 might be seen as highly dubious so far above the norm are they. May 1833 is 1.5 degrees above anything we have seen in all our generally considered warmer lifetimes. June 1846 eclipses every other June in sight with even June 1976 only "managing" 17.0 and that is the only time June has reached such a figure since 1900!  Is it at all possible that back in 1833 and 1846 the measuring equipment of the day could not accurately measure extreme heat? i.e. both months were hot, but not that hot.

I have seen some values from June 1846, at one station in London reported maxima over 80F for 9 days out of the first 14 days for that month and Greenwich, 91.1F for the 20th. No question it was a very hot month.

[summer blizzard]

Its worth bearing in mind that the first half of May 08 was stupidly warm in CET terms, its not hard to imagine what would happen if that stayed the full month.

[SomeLikeItHot]

18.8 to the 25th (20.4)

Min today of 13.4 and max of around 24 so with that and the Met forecasts for 5 days + GFS ensemble mean for the last day we will get.

26th 18.8 (18.7)

27th 18.8 (18.3)

28th 18.8 (18.6)

29th 18.7 (17.5)

30th 18.7 (17.8 )

31st 18.6 (17.3)

That 18.6 at the end there is 18.64 so only just. Given that I think pre correction 18.8-18.6 will be the outcome, with a post correction range of 18.3-18.8.

Edited July 26, 2013 by SomeLikeItHot

[Summer Sun]

18.8c to the 26th

Edited July 27, 2013 by Summer Sun

[SomeLikeItHot]

18.8c to the 26th

 

Cooler min today of 12, but the general track looks the same as my post yesterday, falling to 18.7 likely on the 28th or 29th and then perhaps to 18.6 on the 31st before end of month before corrections.

[BornFromTheVoid]

Minimum today of 12.0C with maxima of around 26C, so remaining on 18.8C on tomorrows update.

 

After that, the 06z GFS operational has the CET at

18.8C to the 28th (17.9)

18.8C to the 29th (19.0)

18.7C to the 30th (15.8]

18.7C to the 31st (18.4)

 

18.6 to 18.8C before corrections, 18.2 to 18.7C after correction methinks.

 

So finishing somewhere between the joint 4th and 14th warmest July on record.

 

[summer blizzard]

More impressive is just how on que this 18C month is, I worked out since 1950 that we get one on average every 7 years.