Trends

[SNOW-MAN2006]

hello,

I have made a file (below) that shows that (by using moving averages) we can see that yes the temperatures are rising countrywide. I have arisen this topic though to show that temperatures can vary and that trend can change at any time. Later on today i will be producing a chart for the past 200 hundred years but for now i have produced one for the past 20. The line of best fit shows a slow increase in temperature but we can see that C.E.T temperatures have risen before and dropped after a rise. Trend can never be used accurately because temperatures can change any time and with todays INDEPENDANT (newspaper) claims that by 2100 2/3rd's of the world will be desert. This was produced by using trend and i think that 2/3rds of the earth being desrt in 90 years time is unrealistic and trend can be an influence of unrealistic outcomes. Feel free to prove me wrong but i think that the graph will show my point that trend cannot be trusted as a reliable source to determine future forecast climates/ changes in environment. A trend is just another word meaning 'expected path' and this 'expected path' is not definate.

Trend.xls

SNOW-MAN2006

(hope my speech wasnt to jumbled up)

Edited October 4, 2006 by SNOW-MAN2006

[Shiny_Bottom_1]

A well balanced post there SM.

For the last 20 years, there is clearly a marked rise in CET (almost 1.5oC in 20 years :o ).

It certainly will be interesting to see the rises and falls of the CET over the last 200 years.

One thing i'm fairly sure it will show though, is that our current 'high' has never been reached before.

[parmenides3]

hello,

I have made a file (below) that shows that (by using moving averages) we can see that yes the temperatures are rising countrywide. I have arisen this topic though to show that temperatures can vary and that trend can change at any time. Later on today i will be producing a chart for the past 200 hundred years but for now i have produced one for the past 20. The line of best fit shows a slow increase in temperature but we can see that C.E.T temperatures have risen before and dropped after a rise. Trend can never be used accurately because temperatures can change any time and with todays INDEPENDANT (newspaper) claims that by 2100 2/3rd's of the world will be desert. This was produced by using trend and i think that 2/3rds of the earth being desrt in 90 years time is unrealistic and trend can be an influence of unrealistic outcomes. Feel free to prove me wrong but i think that the graph will show my point that trend cannot be trusted as a reliable source to determine future forecast climates/ changes in environment. A trend is just another word meaning 'expected path' and this 'expected path' is not definate.

Trend.xls

SNOW-MAN2006

(hope my speech wasnt to jumbled up)

Hi, SM2006; I think your effort is very creditable; it will be interesting to see your final product. The graph in the link below, from the Climate Research Unit, is one of the most used; it shows global temperatures, rather than CETs.

http://www.cru.uea.ac.uk/cru/info/warming/

I don't think anyone will argue with you that temperatures, and temperature trends, can fall as well as rise. If you can provide a link to the Independent article that would be useful; though it's always easiest to assume that journalism tends to 'manipulate' the science to make a better story. The last point you make is quite a challenging one: what do the current trends tell us about future ones? This is something I have no doubt others will discuss. Likewise, they might make the point that, whatever the papers make it sound like, the climate modellers would not say 'this is how the future will be', because they know they can't, but they can say; 'given what we have measured and what we do know about the climate, we think that this is likely to happen...'

:)P

Edited October 4, 2006 by parmenides3

[BLAST FROM THE PAST]

One thing i'm fairly sure it will show though, is that our current 'high' has never been reached before.

Chris

I assume you mean over 200 years and NOT never before?

BFTP

[Shiny_Bottom_1]

Chris

I assume you mean over 200 years and NOT never before?

BFTP

:o Yeah :o

[parmenides3]

No need to post the link to the Indy, S-M2006 - someone has opened a separate thread.

:)P

[VillagePlank]

The idea of trends, in my opinion, is that, for instance, we can, and do have a rather large CET variance; that is to say we can get 'extreme' values both in the warm and in the cold, and we need to have some rigorous method of seeing the big picture.

Obviously these extreme values are noise - such as the anomalously warm July just past. We need to filter out the noise. If I've read your graph correctly you've put a poly 2 trend on it (which is linear) which is the equivalent of y=mx+c (the equation of a straight line) of which the most interesting part is indeed the gradient which , once noise has been smoothed (but not filtered out) gives us the clue that the climate, historically (history meaning any time in the past, even one month) is warming.

When you look at a graph without such trend-lines, then you're mind will do pretty much the same job.

It is, as you say, not a viable means of prediction; it is a more edible form of modelling history. A quick caveat is that one must be sure of what trend line definition they are looking at, and what, semantically it actually means.

For those who'd like to see a VB6 method of generating polynomials, please see:

Polynomials.doc

(This code is copyleft; you can use it as you want to, but I won't support it - just please don't pretend it's yours a quick 'From Wilson, NetWeather' will do)

Edited October 4, 2006 by Wilson

[BrickFielder]

I will be interested to see where you are going with this. It is my contention that quite often statistics, choice of graphical axis to emphasize a point and extrapolation of data are misused. As you say it is quite clear that temperatures have been rising countrywide from that graph ,although the axis is chosen to emphasize the change. As well as looking at temperatures over 200 years perhaps to be fair you should produce one over the last 2 years. Even here graphs may not be reflecting true conditions and perhaps we should seperate day time and night time temperatures. I suspect day time temperatures have not varied as much as night time temperatures and your graph actually indicates an increase in night time cloud cover.

[parmenides3]

I will be interested to see where you are going with this. It is my contention that quite often statistics, choice of graphical axis to emphasize a point and extrapolation of data are misused. As you say it is quite clear that temperatures have been rising countrywide from that graph ,although the axis is chosen to emphasize the change. As well as looking at temperatures over 200 years perhaps to be fair you should produce one over the last 2 years. Even here graphs may not be reflecting true conditions and perhaps we should seperate day time and night time temperatures. I suspect day time temperatures have not varied as much as night time temperatures and your graph actually indicates an increase in night time cloud cover.

This link, on the 'basics of environment change' pin, takes you to th UK climate Impacts programme, and a summary of UK weather trends since 1914. http://www.ukcip.org.uk/scenarios/ukcip02_...ervedtrends.asp

On that page, their is a direct link to Met Office climate memorandum 21 - the source material for the UKCIP page, and much more detailed. It is a strictly NOT for Publication document, so please don't bring chunks out of it without asking the Met first.

It has a great deal of interesting correlations and analyses, including comparison of daily max. & min. (yes, Brickf., you are right; summer mins. have gone up more that summer maxs, and mins in general have gone up more than maxes).

For the reason mentioned above, I am not going to post the link here: go to the UKCIP and access it from there: they have permission, I do not.

:)P

[VillagePlank]

This link, on the 'basics of environment change' pin, takes you to th UK climate Impacts programme, and a summary of UK weather trends since 1914. http://www.ukcip.org.uk/scenarios/ukcip02_...ervedtrends.asp

On that page, their is a direct link to Met Office climate memorandum 21 - the source material for the UKCIP page, and much more detailed. It is a strictly NOT for Publication document, so please don't bring chunks out of it without asking the Met first.

It has a great deal of interesting correlations and analyses, including comparison of daily max. & min. (yes, Brickf., you are right; summer mins. have gone up more that summer maxs, and mins in general have gone up more than maxes).

For the reason mentioned above, I am not going to post the link here: go to the UKCIP and access it from there: they have permission, I do not.

:)P

Why should such information be 'not for publication' ?

[parmenides3]

Why should such information be 'not for publication' ?

Dunno. Ask the Met. Have you read it?

:)P

[VillagePlank]

Dunno. Ask the Met. Have you read it?

:)P

A curious response.

[parmenides3]

A curious response.

Why? I don't get it, Wilson. Please explain. :)P

[SNOW-MAN2006]

I have now completed the 200 year trend graph which shows a glitch towards the present day but these glitches have been fimilar in the past 200 years and often lead to colder years afterwards. I hope this chart will show yes that C.E.T's have risen and fallen but also trend isnt relieable.

For link see my next post (2 down) as this link may be corupted

SNOW-MAN2006

Edited October 4, 2006 by SNOW-MAN2006

[parmenides3]

I have now completed the 200 year trend graph which shows a glitch towards the present day but these glitches have been fimilar in the past 200 years and often lead to colder years afterwards. I hope this chart will show yes that C.E.T's have risen and fallen but also trend isnt relieable.

SNOW-MAN2006

Na. Tried it twice; it's not coming up properly, S-M06.

:)P

[SNOW-MAN2006]

hi, try this new link below

Trend.xls

hope it works

SNOW-MAN2006

[Shiny_Bottom_1]

Its coming up fine here SM

Certainly seems to be a trend of ups and downs over 200 years, however, our current 'up' is probably the longest increase, as well as the highest.

[SNOW-MAN2006]

Its coming up fine here SM

Certainly seems to be a trend of ups and downs over 200 years, however, our current 'up' is probably the longest increase, as well as the highest.

Yes it may well be the highest but the gradient in which it has travelled up isnt the highest so a slow rise may be the trend (which we should be weary of) and it seems joint longest rising with around 1900. will this be the year it goes down a bit .... this years C.E.T is currently at 11.36 (september warmer than august!!!) with falling temps over the next few weeks means we could be set for another high C.E.T year. Good thing about this year is that we have had a cold winter which has brought down the C.E.T a bit.

SNOW-MAN2006

Edited October 4, 2006 by SNOW-MAN2006

[Dawlish]

Its coming up fine here SM :blink:

Certainly seems to be a trend of ups and downs over 200 years, however, our current 'up' is probably the longest increase, as well as the highest.

I'd agree Chris. It is the longest period of increase and UK temps have been significantly warmer in that period than any in any previous, similar length, period.

I'm not quite sure of the point you are making here S-M. The warmer periods didn't "lead" to colder periods, they merely preceeded them. You'd have to look for other reasons why it got warmer and colder in the past. I see little (not "none", note) evidence that the present warming could reverse and a mass of evidence that it will probably continue. If your point is that colder has always followed warmer, no-one could disagree. What we have to talk about are the chances (percentage, odds), on the present trend continuing, or not.

Good on yer(!) calculating the moving average, but it seems to me that what you show is the same as we know already. There had never been a moving average, from your scale, of >10 until 1988 and since 1994, your 4-point moving average has never dipped below 10. At the end of your table, the highest moving average is in the last year of your calculations.

Putting aside, for a minute, your hopes for it to be colder; 2 qns.

1. What does my last paragraph really tell you about what is happening to UK annual temperatures since the 1980's?

2. What are the chances of a trend like this reversing?

Paul

Edited October 4, 2006 by Dawlish

[SNOW-MAN2006]

If your point is that colder has always followed warmer, no-one could disagree. What we have to talk about are the chances (percentage, odds), on the present trend continuing, or not.

Putting aside, for a minute, your hopes for it to be colder; 2 qns.

1. What does my last paragraph really tell you about what is happening to UK annual temperatures since the 1980's?

2. What are the chances of a trend like this reversing?

Paul

Hi paul,

That is my point that in the past colder years have followed an increase in temperatures and mabey we are going the same route (just with higher C.E.T's)... and calculating the possibility of his trend continuing is virtually inpossible. We have to analyis the data we have had an look for patterns that have developed before (increase in temperatures lead to fall in temperatures). If you can caluclate the chances of the present trend continuing then do so but with present Technology we cannot even forecast a week ahead accurately. I think that being able to calculate how long this warm spell will continue contains many factors that would take a very long time to analysis (comparing previous data).

1. That yes we are in a current stage of warming Paul but my point to be made is that temperatures have tended to fall away in the past from a rise. We are now rising at a very fast rate (which it has done in the past).

2. The chance is just as equal as it staying the same.

400 year trend:

Trend.xls

SNOW-MAN2006

[VillagePlank]

That is my point that in the past colder years have followed an increase in temperatures and mabey we are going the same route (just with higher C.E.T's)... and calculating the possibility of his trend continuing is virtually inpossible.

It is indeed the case that the gradient of temperature increase has been seen before; especially if you step outside more 'normalised' methods of looking at historical data.

There is a case to be said that says that the climate has undergone some sort of step-function increase in the base temperature value and that now we will continue as normal but with 1C (or whatever) increase in base temperatures. The problems with this approach can be summarised with these two questions (i)What forced the climate to step-change? (ii) Why is there no record of such a step-change in the CET record until recently?

It is, as you imply, worthwhile looking at the absolute temperature values rather than looking at the rate of anomaly increases (such as the hockey-stick) I have to admit I don't truly understand the reasoning behind publishing plots of anomaly increases rather than absolute temperatures. I suspect it is, like most graphs, drawn to support a point of view; the only problem, of course, is that the absolute temperature line supports such an AGW point of view, but it is just not such an exponential curve.

[Mr Sleet]

Hi, SM2006; I think your effort is very creditable; it will be interesting to see your final product. The graph in the link below, from the Climate Research Unit, is one of the most used; it shows global temperatures, rather than CETs.

http://www.cru.uea.ac.uk/cru/info/warming/

I don't think anyone will argue with you that temperatures, and temperature trends, can fall as well as rise. If you can provide a link to the Independent article that would be useful; though it's always easiest to assume that journalism tends to 'manipulate' the science to make a better story. The last point you make is quite a challenging one: what do the current trends tell us about future ones? This is something I have no doubt others will discuss. Likewise, they might make the point that, whatever the papers make it sound like, the climate modellers would not say 'this is how the future will be', because they know they can't, but they can say; 'given what we have measured and what we do know about the climate, we think that this is likely to happen...'

:)P

I've often wondered why the rate of increase ( in global temps) from 1910 to 1940 is the same as that of 1960 -1998 ( trend levels off after that) when the rate of increase of CO 2 was less.

[parmenides3]

I've often wondered why the rate of increase ( in global temps) from 1910 to 1940 is the same as that of 1960 -1998 ( trend levels off after that) when the rate of increase of CO 2 was less.

Rather than think of it that way round, it might be easier to consider it as a continuous period of warming from around 1900 to the present, interrupted by a period of not-warming; the article below offers an explanation;

http://www.stanfordreview.org/Archive/Volu...opinions1.shtml

I'll check up on the CO2 levels for the fist part of the century and offer a response to that part of the question shortly.

:)P

[The Penguin]

Is this not a bit like recording the incidence of heads and tails being thrown with a coin? We know that the result has to be one or the other but no matter how well we record historical results, the record is no use whatsoever in predicting the next outcome. Also, just because the last few tosses show up as heads doesn’t necessarily mean that there’s something wrong with the coin.

[parmenides3]

Is this not a bit like recording the incidence of heads and tails being thrown with a coin? We know that the result has to be one or the other but no matter how well we record historical results, the record is no use whatsoever in predicting the next outcome. Also, just because the last few tosses show up as heads doesn’t necessarily mean that there’s something wrong with the coin.

Only in the broadest sense; a better analogy might be the throwing of multiple dice. The pattern would tend to be random, sometimes producing a higher sum than the 'normal range', other times producing a lower sum. If, say. six dice were thrown and the total bullet count for each throw recorded, 144 times, it is very unlikely that the totals 6 or 36 would appear more than once or twice. On the other hand, the occurrence of counts between 18 and 24 would be high. At no time would the previous throw impact on the likelihood of the next throw being smaller, larger or the same, but over 144 throws, a pattern emerges which allows you to bet on the likeliest outcome, and, with a generous 'house', break even.

Taking the analogy across to temperature records, what the recent record shows is that high numbers, outside the 'normal' range, are consistently being thrown, time and time again; the last twenty or thirty throws have all, bar two or three, been higher than is statistically accountable for. Why? The dice must be loaded... in other words, a mechanism in each of the individual dice adds weight to the one-spot, every time a six is thrown, making another six increasingly likely. This is why, as the series progresses, there is a trend towards higher numbers. Does this make sense?

:)P