Latest News (hidethedecline)

Original Temperatures: Germany

Posted by Frank Lansner (frank) on 25th December, 2013
Latest News (hidethedecline) >>


Data Sources for German temperature data:
1) Meteorological yearbooks from “Deutsches Reich”, British-, American- and soviet zones, DDR and Bundesrepublik Deutschland.
2) File required from DWD via Danish Newspaper
3) ECA&D web site
4) Statistical yearbooks 1900-1991

For Spain, France, Switzerland and Poland etc. as shown in the ORIGINAL TEMPERATURES writings one could rather easily estimate adjustments of temperature data after 1990 by comparing Tutiempo data with ECA&d data. Temperature data at some point after 1990 had been warm adjusted 0,25 to ?? K

However, German institutions have decided not to deliver the bulk of German temperature data after 1991 to the Tutiempo site, and thus we cannot this way detect to what degree the German temperature series has been warm adjusted after 1991.

From 1991 also another important source of German raw temperature data ends: The statistical yearbooks. Until 1991 around 55 stations have yearly temperature data included, but from 1992 the number dives strongly to just 10 stations of which most are coastal or urban.  1990 is also the year WWR ends publishing numerous temperature sets.

In 2012 I asked a major Danish Newspaper to again ask a row of European national meteorological institutes to send us long raw temperature series for publishing and analysis. Some MI´s never answered, some advised us just to use the adjusted data already available online. Some MI´s were happy to help us with genuine raw long temperature series, but we should pay them a huge amount of money without exaggerating. One MI (The Danish) claimed that they were not in possession of longer non-coastal data in digital format. I asked one MI (the Czech) how come they needed so much money for data often already published. They answered that they were not aware what data had already been published.
But to my luck (and surprise), the German MI swiftly send us a file with around 150 long temperature series often beginning even before year 1900. These series normally ends around 2005-6 but some continue to 2012.

The temperature data from the DWD file matches data from meteorological year books and statistical yearbook well, although the years after 1991 are harder to verify.

ECA data for Germany often matches original data at least before 1991. However, the data available from ECA often begins as late as 1947-52 and thus do not include most of the pre 1950 warm period.

Estimating German temperature data
As always, the task is to identify areas of similar temperature trends. This is done by first drawing a sketch that roughly lives up to the following definition of areas:

Fig 1

We have 4 types of areas:
1) Coastal (red)
2) Areas affected by ocean winds (Yellow)
3) Areas in shelter from ocean winds – “Shelter Areas” (blue)
4) Strongly Urban areas.

German altitude rise gradually from the ocean coasts to the southern and eastern mountains. Therefore we have a rather complicated mosaic of areas with ocean wind trends (yellow areas) and areas in shelter of ocean winds, in blue.

Fig 2 Click to enlarge

Results from ALL German areas can be seen HERE.XX Link will come when ready! XX

Discussion of results
Fig 3

The 2 coastal areas R1 and R2 show similar trends. R2 has been laid on top of the R1 as best fit.

Fig 4
The warm peak 2007 in the R1 dataset is mainly from one station, Wilhelmshaven. Such outliers are revealed fast when working with areas of similar trend but I’m not going to mention all outliers in this writing.
Fig 5

A little further from the coast, three neighbour areas of rather similar geography (mostly flat) thus showing quite similar trends as one might have expected.

Now let’s take a look at the hill areas affected by ocean winds.
Fig 6

6 areas affected by ocean air. I have inserted the individual trend lines on top of the Y8 Siegen graph as best fit, and although the 6 areas are spread out over South and central Germany, trends are similar indeed.

Fig 7

We also find a fair match between the 2 coastal areas and the 6 areas strongly affected by ocean wind. Again, trends are rather similar this time despite larger distances and different altitudes.

Let’s then turn to the areas in shelter from ocean air.

Fig 8
2 areas in shelter B13 and B15 are neighbours in Southern Germany, and thus their mutual similarity is not too surprising. The warming trend from around 1920 to 2010 is limited.

Fig 9

Illustrated is 6 areas in rather good shelter from ocean winds illustrated as best fit. It does seem that if we remove influence from ocean winds, then the resulting temperature trend gets colder and also similar from area to area.

Fig 10

Example of an “outlier” temperature series: Jena


Fig 11

Since this writing is not a PHD, I will just for now show one of the most interesting and obvious outliers among German temperature series, the Jena station.
Fig 12

One of the areas of best shelter against ocean winds is the area I call B3 located mostly in Thüringen.

Fig 13

The Jena temperature station is located in the south central B3 area, and as you can see, the Jena station is affected by something else than its surrounding stations.
Fig 14

Also against stations in the next door area B4, the Jena station appears to be an outlier, so what is the history behind the Jena data?

Fig 15

The temperature difference between the general B3 area and Jena show that we do not have a sudden break in the data from the Jena station. Instead we have a 1 K large but fluent change from around 1950 to 1980.
UHI? Some UHI warming signal in the Jena data cannot be ruled out. Jena is not that big a town, but the town is located in a rather narrow and deep valley so that Urban heat will escape slower.

However, I don’t think UHI is the full explanation. The differences between Jena and the B3 area to some degree resembles typical differences between areas affected by ocean heat and areas in shelter. Temperatures of the ocean affected areas catches up in periods with less warming. If ocean air temperature should reach Jena – rather far from the mountains south of Jena – then one might suspect the river Saale that flows in the deep valley from the mountains.
Fig 16

However, if you compare Jena temperature trend even with near by warm trended mountain stations, then these stations do not show as much heat trend as the Jena station.
Summa: Jena is indeed a warm trended station but this warm trend appears to be quite local. The reason for Jenas warm trend is perhaps difficult to explain, but since this station do not represent larger areas, this station is not suited for larger analysis of climate.

UHI example: The Rhein Rhuhr District.

Fig 17

The Rhein Rhuhr area has urban areas over large areas. Even less populated areas  will often have winds transported from urban areas and thus the whole area should hardly be used in climate science, especially not if the goal is to evaluate the role of greenhouse gasses in the atmosphere generally.

Fig 18

To use temperature stations from urban areas is obviously like attending a lottery. What flavour of UHI do we get? Using Urban stations in climate science must be reduced when possible – no matter what you are told.
In the figure above I made best bit of data in the early 20´th century to easily see how much these datasets differentiates over the years. I added 2 datasets from the near by more rural area (B1) to show all degrees of UHI in the area.

Fig 19

Here are the stations with least UHI. Some stations placed in the centre of cities that were already large around year 1900 shows less UHI. The massive Urban warm temperatures were to some degree already present here in year 1900. The worst cases of UHI happens when a stations gradually is surrounded by urban areas.
Fig 20

And finally the stations with massive UHI: Some of these stations have even more heat trend than coastal stations and stations located on hills affected by ocean air.

The BEST Corner - Evaluating Berkeley Earth Surface Temperature for Germany

For several countries I have found, that areas of less warming (Ocean Air Shelter areas) appear to be mostly ignored by BEST. Is this the case for Germany too?
To test this, I will use the following graph:
Fig 21

Raw average of all the German areas except the 16 shelter areas with colder trends.
Fig 22

The temperature trends for Germany used by BEST resemble German areas when the 16 colder trended areas are ignored.

Thus, disregarded the methods and actions taken by BEST, the bottom line is that temperature stations from the German areas with little or no warming 1925-2010 cannot be detected in the overall result from Germany made by BEST.

This is the primary result of this writing, but let us for curiosity take a look at the destinies of a few stations used by BEST that originates from the cold trended areas, the areas in shelter from ocean wind

Some examples of BEST stations from the 16 colder trended areas.
Many datasets used in ORI-TEMPS GERMANY are not to be found in BEST. Here some examples of temperature sets from the 16 colder areas where the temperature stations are used in BEST.

Erfurt (B3):
Fig 23

This graph illustrates the fundamental problem of BEST. Best assumes that there is warming in most places and call it the “Regional Expectation”. Then they calculate a difference from this expected warming, and they adjust for this difference.

Erfurt temperature is changed this way – and what excuse do BEST have to explain the data changes?  Station move? TOBS? 
No, BEST can only explain their changes with the label:  “Empirical Break”.
Erfurt data is changed by BEST so that 1985-2010 are warm adjusted +0,5 K relative to 1900-1925 data, and warm adjusted around +0,8 to + 1 K relative to 1925-45 data.

Fig 24

Now, by BEST magic, Erfurt matches the blue regional expectation graph.

Fig 25

Halle (B3) temperatures before 1950 are cold adjusted around 0,8 K relative to the most recent years. Again the excuses/explanations for changing originally measured temperatures are “Empirical Breaks”.

Fulda (B6):
Fig 26

Kassel (B6)
Fig 27

From the B6 area, Kassel appears to be an outlier (possibly due to UHI) with warmer trend than the other temperature series of B6. Kassel is used by BEST, and this dataset is not adjusted notably.

Mannheim (B8):
Fig 28

Fig 29

Only smaller fragments of data are used from the Mannheim area in BEST.

Kaiserslautern (B8):
Fig 30

Fig 31

Kaiserslautern in BEST.
Bamberg (B11):
Fig 32

Bamberg from B11
Fig 33

Bamberg in BEST

Hamburg (Y7):
Hamburg.. not even in an area I regard as shelter area, but still massive adjustments from BEST:
Fig 34

Kiel (B5):
Fig 35

Fig 36

Kiel in BEST – many period not represented.

Lubeck (B5):
Fig 37

Fig 38
Most Lubeck data is not used in BEST.

Magdeburg (B2):

Fig 39

Magdeburg adjusted by BEST and also cut before 1936.

Nuernberg (B11):

Fig 40

Nuernberg in BEST

Ulm (B13):

Fig 41

Ulm in BEST.

Augsburg (B13);


Fig 42

Augsburg cut before 1947 by BEST, and also, recent years are warm adjusted around 0,8 K relative to pre-1980 values.

Arnsberg (B1):
Fig 43

Arnsberg appears to be the best sheltered station of the (B1) area of Western central Germany and as could be expected the station of the (B1) area with coldest temperature trend of the B1 stations.

BEST do not use Arnsberg data after 1851.

Leipzig (B4):
Fig 44

Leipzig data has slightly more heat trend than the other stations of (B4) and it may be due to UHI error. Leipzig thus ends up around 0,4 K warmer than the other stations of (B4) in after around 1980. What is then od it that from 2005, the other stations are suddenly just as warm as the Leipzig station.

Leipzig stations are only used as smaller fragments in BEST.

Torgau (B4):
Fig 45

Torgau is perhaps the best sheltered station in the (B4) area, and thus to no surprise also one of the stations with coldest trend.

BEST do not use Torgau data after 1868.

Bayreuth (B12):
Fig 46

The DWD data suggests that temperatures year 2000 is similar to the 1994 warm peak, but this is not supported by near the other stations of (B13)
Fig 47

BEST use little data from the Bayreuth area

And a few more OAS datasets found in BEST:
Braunschweig: BEST cuts data before 1970
Regensburg: BEST cuts data before 1950.
Stuttgart: Warm years 1930-50 missing and data before 1930 cold adjusted in BEST.
Darmstadt: BEST cuts data after 1930

Last changed: 25th December, 2013 at 15:59:25



None Found

Add Comment