Local Temperature Increase

[Ian Bingham]

Seasonal Rise of Temperature in 34 Years (1990-2023 incl.) at Inchmarlo, Aberdeenshire

JAN   -0.35                Winter   +0.3

FEB   +0.35               Spring     -0.1

MAR   -1.8              Summer   +1.1

APR    +0.2               Autumn   +1.1

MAY   +1.3             

JUN   +2.0

JUL    +1.2              I read recently that world average temperature had risen by 1 C in the                                      last 60 years.

AUG   +0.1             Our 0.6 C in 34 years gives a similar result.

SEP   +1.3

OCT   +1.1

NOV   +0.9

DEC   +1.0

_________

Av.    +0.6

Perhaps this points up another aspect if climate change – contrast.  June has risen the most and January has fallen the most.

Ian Bingham

Inchmarlo, Aberdeenshire.

80m asl.

[Graham Easterling]

Warming in Penzance.

Graham

Penzance

[Len]

Ian, I am not sure how you work out your difference/trend for each month.

Do you take a least squares fit through the series and get a straight line from which you deduce your trend?

That is what I have done for my 38 years of seasonal data from 1986 to 2023 here in Wembury, SW Devon.

All seasons show a warming and here are my figures:

Winter(DJF) +1.1 degC

Spring(MAM)  +1.3 degC

Summer(JJA) +1.4 degC

Autumn(SON)  +1.3 degC

If I do a fit to my annual data I get

Annual +1.1 degC

Statistics are a wonderful thing.

Len

Wembury, SW Devon

On Wednesday 6 December 2023 at 11:21:51 UTC Ian Bingham wrote:

[Ian Bingham]

Hello, Len.  As Statistics is hardly my forte, I let the computer do it! 

In Microsoft Excel's chart-drawing facility is a menu for drawing several types of trendline.

I've no idea how the computer does it but I've noticed that if you plot the average point on the chart, the average of the first half and the average of the second half, you get 3 points in a straight line which seems pretty much the same as the computer trendline.

It is interesting that you, in a warmer part of the country, have seen a bigger temperature increase than I have.  Also, like my figures, winter has increased less than summer, It all seems to emphasise the general theme of increased contrast.

Ian.

[Len W]

Ian,

I put a trendline on my time series using MS Excel as well.

Once I have constructed the chart I click on it and a Chart elements menu pops up.

From this you can click on the Trendline box to get a linear least squares fit line.

Len

Wembury, SW Devon

[Ian Bingham]

Len,

Yes, that's what I use.  The trouble is that if you get an outlier it can tip the trendline into a more or less unrepresentative slope.  

So you get rid of outliers and then you have another problem - how far does an outlier have to "out-lie" to qualify as an outlier ?

It's a rocky road !

Ian

[Len W]

Ian,

You are right about what is an outlier or not.

An extraordinary point (outlier) in a time series would perhaps in statistical terms be one that was more than 2 standard deviations from the series mean. MS Excel again will give you a standard deviation for a time series of temperature for example.

An extraordinary point or not, it is real and cannot be removed from the time series.

But you are right, it certainly can affect the trend.

So removing it from the series for convenience, or starting or ending the series before or after it, just to show a convenient trend

is what leads to the adage 'There are three kinds of lies: lies, damned lies, and statistics.'