13.7.2010, Edinburgh 10th IMSC
Bridging the gap from indirect to direct climate data
experience with homogenizing long climate time series in the early instrumental period
and some general remarks on the problem original vs. processed (faked?) data
Reinhard Böhm
Central Institute for Meteorology and GeodynamicsVienna, Austria
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Sligachan Hotel Isle of Skye
13.7.2010, Edinburgh 10th IMSC
Bridging the gap from indirect to direct climate data
experience with homogenizing long climate time series in the early instrumental period
and some general remarks on the problem original vs. processed (faked?) data
Reinhard Böhm
Central Institute for Meteorology and GeodynamicsVienna, Austria
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Sligachan Hotel Isle of Skye
13.7.2010, Edinburgh 10th IMSC
1550 1600 1650 1700 1750 1800 1850 1900 1950 2000
DAWN OF THE INSTRUMENTAL PERIOD
EARLY INSTRUMENTAL
PERIOD
FULLY DEVELOPED INSTRUMENTAL
PERIOD
Proxies necessary to understand climate
variability and trends
Proxies and instrumental „at eye level“
Systematic networks of meteorological
services
Invention of principal meteorological instruments
1597: T
hermom
eter (G
alileo)
1643: Barom
eter (T
orricelli)
1732: Anem
ometer
(Pitot)
1769: Hygrom
eter (Lam
bert)
1792: Psychrom
eter (H
uton)
1838: Pyrheliom
eter (P
ouillet)
1853: Sunshine
recorder (Cam
pbell)
Some basic facts on indirect and direct climate information
13.7.2010, Edinburgh 10th IMSC
CLIMATEGATE, 2009:
One of the “arguments”:
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
My clear answer:
(based on some 20 years of experience in the field of homogenizing)
Let me argue this briefly in the next minutes
the original data were intentionally not posed to everybody’s free access to conceal the “tricks” applied to the original data (to increase the amplitude of anthropogenic warming)
WE NEED SUCH TRICKS TO TRANSFER „MEASURED NUMBERS“ TO „CLIMATE DATA“
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
My arguments in a nutshell:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases are of the order of the real climate signal
There are a number of mathematical-statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
My Examples will be from Central Europe (HISTALP):
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
region: GAR (4 to 19°E, 43 to 49 deg N)
more than 500 series on 200 sites
7 climate elements
monthly series:
quality improved (homogenised, outlier corrected, gaps filled)
maximum achievable length (100 to 250 years)
element specific network density (190 precip – 60 pressure series)
different modes:
station mode and 3 different grid-modesAvailable in the Web: http://www.zamg.ac.at/histalp
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Examples for:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
KLAGENFURT, 1813-2009: 7 relocations – different urban bias
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Examples for:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
KLAGENFURT, 1813-2009: 7 relocations – different urban bias
5
6
7
8
9
10
1810 1830 1850 1870 1890 1910 1930 1950 1970 1990 2010
°C TEMPERATURREIHE KLAGENFURT 1813-2007einzelne Jahresmittel und 20-jährig geglättet
grün: homogenisiert (angepasst an aktuellen Zustand)rot: Originaldaten
Quelle: ZAMG-HISTALP Datenbank, Auer et al., 2007
Achatzel1 2 Prettner Sl1 Landesmuseum ZAMG-Annabichl
Klagenfurt composite series: red original green homogenized
red: original data green: homogenized
City center Aiport sites
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Examples for:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
400
600
800
1000
1200
1400
1600
1800
1810
1820
1830
1840
1850
1860
1870
1880
1890
1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
mm
KAR ORI 1 to 12
KAR HOM 1 to 12
PRECIPITATION (MM)
Precipitation in Karlsruhe with a bias of nearly 100% in the 1870s and 1880s
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Examples for:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
HISTALP - PRECIPITATION DATASET: BREAK ANALYSISseries 192available data (incl. closed gaps) 26063 yearsavailable data (incl. closed gaps) 312756 monthsmean length of series 135.7 yearsdetected breaks (total) 966mean homogeneous subperiod 22.7 yearsmean breaks per station per year 0.037square mean break 17.9 %maximum break 238 %mean break amplitude 23.4 %corrected overshooting adjustments 1861detected real outliers 529closed gaps 14927mean gap rate (%) 4.8 %
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Examples for:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
An example for systematic biases: EARLY PRECIP- SERIES
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
0
10
20
30
40
50
1800 1825 1850 1875 1900 1925 1950 1975 2000
m a
bove
gro
und
(lo
g)
The development from high to low elevated raingages in the climate network of 154 central European sites with sufficient
station history documentation
source: ZAMG-HISTALP metadata thin: single sites, bold, red:
average 1820: 17m
average 2000: 1.4m
average 1870: 3.5m
0
10
20
30
40
50
60
70
80
90
100
JA
N
FE
B
MA
R
AP
R
MA
Y
JU
N
JU
L
AU
G
SE
P
OC
T
NO
V
DE
C
14
.5m
/1.3
m
(%)
Comparison of mean monthly precipitationPula (HR) 1873-1897:
rooftop (14.5m) versus 1.3m near to ground
parallel measurements 1873-1897tower (14.5m above ground)
near to ground (1.3m above ground)
kk. Hydrographisches Amt in Pula: 2 sites (25 years of parallel measurements)
Early rain-gauges systematically higher above ground early precip-series systematically too dry
13.7.2010, Edinburgh 10th IMSC
Another problem a bit more in detail here:
the EARLY INSTRUMENTAL TEMPERATURE BIAS described in:
Frank D, Büntgen U, Böhm R, Maugeri M, Esper J, 2008.
Warmer early instrumental measurements versus colder reconstructed temperatures: shooting at a moving target.
Quaternary Science Reviews, (2008), doi:10.1016/j.quascirev.2007.08.002
Hiebl J, 2006.
The early instrumental climate period (1760-1860) in Europe. Evidence from the Alpine region and Southern Scandinavia
Diploma Thesis, Inst. f. Geography, Uni-Wien, 103 pages
Böhm R, Jones PD, Hiebl J, Brunetti M, Frank D, Maugeri M, 2010
THE EARLY INSTRUMENTAL WARM-BIAS: A SOLUTION FOR LONG CENTRAL EUROPEAN TEMPERATURE SERIES 1760-2007
Climatic Change 101: 41-67
Büntgen U, Frank DC, Nievergelt D, Esper J 2006.
Summer temperature variations in the European Alps: AD 755-2004.
Journal of Climate 19: 5606-5623
13.7.2010, Edinburgh 10th IMSC NW NE S alle HISTALP sites
S
NENW
The HISTALP-early instrumental potential
13.7.2010, Edinburgh 10th IMSC
248 years of GAR-temperature: the result so far (version 2006)
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s to
185
1-20
00 in
K SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s to
185
1-2
000
in K WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s to
185
1-20
00 in
K SUMMER HALF YEAR (AMJJAS)
HOMOGENISED (NO EIP-CORRECTION)-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s to
185
1-2
000
in K WINTER HALF YEAR (ONDJFM)
HOMOGENISED (NO EIP-CORRECTION)
??
original series (smoothed)
homogenized
13.7.2010, Edinburgh 10th IMSC
the nucleus of the EI-problem:Diverging instrumental and proxy warm season temperatures before 1860
-5
-4
-3
-2
-1
0
1
2
3
1750 1770 1790 1810 1830 1850 1870 1890 1910 1930 1950 1970 1990 2010
°C CRSM-LOW vs TR-Büntgen JJAS
rela
tive
to 1
901
-200
0Described in:
Frank et al, 2007 Hiebl , 2007 Büntgen et al., 2005
pre-1860-proxy temperatures (here: high elevation TRX-series)
pre-1860 HISTALP temperatures
13.7.2010, Edinburgh 10th IMSC
Our solution: strictly staying within the instrumental domain:
•Intensive re-analysis of station history of the more than 30 EI-series•Combined with the evidence from Kremsmünster:•(multianual parallel measurements of a preserved original EI-site with a nearby typical modern installation (AustrianTAWES network)
13.7.2010, Edinburgh 10th IMSC
The Kremsmünster evidenceOrientation NNE
6m above ground
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1 3 5 7 9 11 13 15 17 19 21 23
JAN FEB MARAPR MAY JUN
Temperature Kremsmünster 1995-2002 historic minus modern site (means)
time (MLT)
°C
NE-facing wall(measured)
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1 3 5 7 9 11 13 15 17 19 21 23
JUL AUG SEPOCT NOV DEC
Temperature Kremsmünster 1995-2002 historic minus modern site (means)
time (MLT)
°C
NE-facing wall(measured)
Mean differences historic minus modern installation
13.7.2010, Edinburgh 10th IMSC
Three correction models derived from Kremsmünster
-1,0
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
1 3 5 7 9 11 13 15 17 19 21 23
JAN FEB MARAPR MAY JUN
Temperature Kremsmünster 1995-2002 historic minus modern site (means)
time (MLT)
°C
NE-facing wall(measured)
-1,0
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
1 3 5 7 9 11 13 15 17 19 21 23
JUL AUG SEPOCT NOV DEC
Temperature Kremsmünster 1995-2002 historic minus modern site (means)
time (MLT)
°C
NW-facing wall(modelled)
-1,0
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
1 3 5 7 9 11 13 15 17 19 21 23
JAN FEB MARAPR MAY JUN
Temperature Kremsmünster 1995-2002 historic minus modern site (means)
time (MLT)
°C
N-facing wall(modelled)
13.7.2010, Edinburgh 10th IMSC
4 examples from the intensive new metadata studies :
yellow: still missing information (status: 2007-06-05)
ID name acr. nat alt CRS from - to name od subseriesheight above
ground orientation observing times or
means calc in EI
m N - deg - E
METADATA SAMPLER FOR THE EARLY INSTRUMENTAL PERIOD OF THE HISTALP T01 LONG SERIES SUBSET
TOR IT *) 4 1760-1786Ignazio Somis, *)mostly Università, via Po, but also other locations 10 to 25 28° variable (a)
TOR IT 250 4 1787-1802/05 Accademia delle Scienze, biblioteca 12,5 28° variable (a)TOR IT 4 1802/06-1802/12/21 sunrise, 14, sunsetTOR IT 4 1802/12/22-1851/01/05 sunrise, 12, sunsetTOR IT 4 1851/01/06-1857/07 09, 12, 15TOR IT 4 1857/08-1865/06 (max+min)/2TOR IT 232 4 1865/07-1865/11 Castello del Valentino 2,0 28° 08, 16
TOR IT 276 4 1865/12-1919/02 Palazzo Madama 37,7 20° Tmin, Tmax
44,2Accademia delle Scienze, specola282 0°?
di Napoli G, Mercalli L, 2007. Il clima di Torino. Ca.900 pages, in press
139 TORINO
INN AT 576 2 1777-1784 Jesuiten Colleg Sillgasse (Franz v. Zallinger) second floor -9° 4, 13.30
INN AT 576 2 1784-1828Kapuziner Kloster, Universitätsstrasse (Franz v. Zallinger)
8 15° near min, near max
INN AT 590 2 1828/09-1855/12 6, 13.30INN AT 590 2 1856/01-1859/12 (7+14+2*21)/4INN AT 590 2 1860/01-1870/12 6, 14INN AT 576 2 1865/01-1870/12 (7+14+2*21)/4INN AT 576 2 1871/01-1875/07 6, 14, 22
8 15°
8 -19°INNSBRUCK
Prämostratenser Kloster Wilten (Prantner)59
Old University-Botanical Institute
Auer et al., 2001: ALOCLIM - Austrian long-term climate 1767-2000. Österreichische Beiträge zu Meteorologie und Geophysik 25 147pages plus data and metadata CDZallinger F, 1833. Innsbrucker meteorologische Beobachtungen von 50 Jahren. Ferdinandeum, Wagnersche Schriften 107 pages
STR SA 2 1801-1843 (?) Astronomical observatory, Place de St. Thomas (J.L.A. Herrenschneider) 3,9 0° 6-7 / noon / 21-22
STR SO 2 several years Vorstadt Neudorf (Besson) 1,5STR MO 2 1891/04-before 1915 Institut de Physique du Globe 6 0° (7+13+2*21)/4STR DI 2 before 1915- Institut de Physique du Globe 6 0° (7+14+2*21)/4
STRASBOURG135
1890-1950: Annales de l'Institut de Physique du Globe (before 1918: Elässische Jahrbücher)Herrenschneider, J.L.A., 1825. Résumé des observations météorologiques faites a Strasbourg depuis le commencement de l'an 1811 jusqu'à la fin de 1820. Mémoires de la Société des Sciences de Strasbourg 2 46 pages plus 10 tablesHerrenschneider, J.L.A., 1815. Résumé des observations météorologiques faites a Strasbourg depuis le commencement de l'an 1801 jusqu'à la fin de 1810. Mémoires de la Société des Sciences de Strasbourg 1
AUG DE 1 1812-1822 station Stark 1 (western part of historic centre) 12 WSW and ENE
AUG DE 1 1822-1829 station Stark 3 (near cathedral) 5,5AUG DE 1 1829-1836 station Stark 2 (astronomical observatory) 12AUG DE 1 1838-1872 8,3 W and E true meanAUG DE 1 1872- 8,3 3°
159
Benedictine Monastery St. Stephan
AUGSBURG
2nd yearbook of Bavaria, 24-28
(7+14+2*21)/4
Not important in this context, but very impotant for tx, tn DTR-studies:
0
10
20
30
40
1760 1800 1840 1880 1920 1960 2000
me
ters
ab
ove
gro
und
The development from high to low elevated thermometers in the climate network of 97 central European sites with sufficient station history
documentation
average 1760-1860: 13.7m
average 2000: 2.5m
thin: single sites, bold, red: mean source: ZAMG-HISTALP metadata
13.7.2010, Edinburgh 10th IMSC
4 examples from the intensive new metadata studies :
yellow: still missing information (status: 2007-06-05)
ID name acr. nat alt CRS from - to name od subseriesheight above
ground orientation observing times or
means calc in EI
m N - deg - E
METADATA SAMPLER FOR THE EARLY INSTRUMENTAL PERIOD OF THE HISTALP T01 LONG SERIES SUBSET
TOR IT *) 4 1760-1786Ignazio Somis, *)mostly Università, via Po, but also other locations 10 to 25 28° variable (a)
TOR IT 250 4 1787-1802/05 Accademia delle Scienze, biblioteca 12,5 28° variable (a)TOR IT 4 1802/06-1802/12/21 sunrise, 14, sunsetTOR IT 4 1802/12/22-1851/01/05 sunrise, 12, sunsetTOR IT 4 1851/01/06-1857/07 09, 12, 15TOR IT 4 1857/08-1865/06 (max+min)/2TOR IT 232 4 1865/07-1865/11 Castello del Valentino 2,0 28° 08, 16
TOR IT 276 4 1865/12-1919/02 Palazzo Madama 37,7 20° Tmin, Tmax
44,2Accademia delle Scienze, specola282 0°?
di Napoli G, Mercalli L, 2007. Il clima di Torino. Ca.900 pages, in press
139 TORINO
INN AT 576 2 1777-1784 Jesuiten Colleg Sillgasse (Franz v. Zallinger) second floor -9° 4, 13.30
INN AT 576 2 1784-1828Kapuziner Kloster, Universitätsstrasse (Franz v. Zallinger)
8 15° near min, near max
INN AT 590 2 1828/09-1855/12 6, 13.30INN AT 590 2 1856/01-1859/12 (7+14+2*21)/4INN AT 590 2 1860/01-1870/12 6, 14INN AT 576 2 1865/01-1870/12 (7+14+2*21)/4INN AT 576 2 1871/01-1875/07 6, 14, 22
8 15°
8 -19°INNSBRUCK
Prämostratenser Kloster Wilten (Prantner)59
Old University-Botanical Institute
Auer et al., 2001: ALOCLIM - Austrian long-term climate 1767-2000. Österreichische Beiträge zu Meteorologie und Geophysik 25 147pages plus data and metadata CDZallinger F, 1833. Innsbrucker meteorologische Beobachtungen von 50 Jahren. Ferdinandeum, Wagnersche Schriften 107 pages
STR SA 2 1801-1843 (?) Astronomical observatory, Place de St. Thomas (J.L.A. Herrenschneider) 3,9 0° 6-7 / noon / 21-22
STR SO 2 several years Vorstadt Neudorf (Besson) 1,5STR MO 2 1891/04-before 1915 Institut de Physique du Globe 6 0° (7+13+2*21)/4STR DI 2 before 1915- Institut de Physique du Globe 6 0° (7+14+2*21)/4
STRASBOURG135
1890-1950: Annales de l'Institut de Physique du Globe (before 1918: Elässische Jahrbücher)Herrenschneider, J.L.A., 1825. Résumé des observations météorologiques faites a Strasbourg depuis le commencement de l'an 1811 jusqu'à la fin de 1820. Mémoires de la Société des Sciences de Strasbourg 2 46 pages plus 10 tablesHerrenschneider, J.L.A., 1815. Résumé des observations météorologiques faites a Strasbourg depuis le commencement de l'an 1801 jusqu'à la fin de 1810. Mémoires de la Société des Sciences de Strasbourg 1
AUG DE 1 1812-1822 station Stark 1 (western part of historic centre) 12 WSW and ENE
AUG DE 1 1822-1829 station Stark 3 (near cathedral) 5,5AUG DE 1 1829-1836 station Stark 2 (astronomical observatory) 12AUG DE 1 1838-1872 8,3 W and E true meanAUG DE 1 1872- 8,3 3°
159
Benedictine Monastery St. Stephan
AUGSBURG
2nd yearbook of Bavaria, 24-28
(7+14+2*21)/4
important for EIP-tm-bias-correction (for chosing the convenient corr-model)
13.7.2010, Edinburgh 10th IMSC
„from – to“: the individual lengths of the EIP:
EIP length
1750 1775 1800 1825 1850 1875
BaselGeneve
TorinoMilano
KremsmünsterRegensburg
PadovaWienBern
InnsbruckKarlsruheBudapest
Hohenpeißenbg.München
VeronaStuttgart
StrasbourgUdine
NiceAugsburg
KlagenfurtBologna
LinzTrento
Gr. St. BernardMantova
ZürichGraz
TriesteSalzburg
detection rates
0 25 50 75 100[%]
80% of the introduction of regular screens happened within 15 years
13.7.2010, Edinburgh 10th IMSC
the individual EIP-bias corrections
-1.0
-0.5
0.0
0.5
JAN
FE
B
MA
R
AP
R
MA
Y
JUN
JUL
AU
G
SE
P
OC
T
NO
V
DE
C
EIP minus non EIP (K)
-1.0
-0.5
0.0
0.5
1.0
JAN
FE
B
MA
R
AP
R
MA
Y
JUN
JUL
AU
G
SE
P
OC
T
NO
V
DE
C
EIP minus non EIP (K)
series with sufficient metadata detection
series with insufficient metadata detection
detection rates
0 25 50 75 100[%]
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-20
07
(K)
SUMMER HALF YEAR
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-2
007
(K
)
WINTER HALF YEAR (ONDJFM)
originalhomogenized-2006
EIP-corrected-2008
1971: evening observations from 9pm tp 7pm in AT and CH
Since 1980s: automation of networks ??? – effects not clear yet
WW-2: systematic relocations from historic centers to airports
THE NETWORK BECOMES MORE RURAL!
1830s to 1870s: systematic EI-bias due to insufficient shading of thermometers
THE RESULT: mean over all 32 single EI-series (20-yrs- smoothed, anomalies from 1978-2007 average)
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Total of systematic temperature-biases:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-20
07
(K)
SUMMER HALF YEAR
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-2
007
(K
)
WINTER HALF YEAR (ONDJFM)
originalhomogenized-2006
EIP-corrected-2008
SYSTEMATIC BIAS = +0.55°C
SYSTEMATIC BIAS = +0.45°C
13.7.2010, Edinburgh 10th IMSC-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-2
007
(K
)
SUMMER- minus WINTER-HALfYEARS
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
1750 1800 1850 1900 1950 2000
anom
alie
s to
197
8-2
007
(K
)
YEAR (Jan-Dec)
Smoothed annual means
THE RESULT: mean over all 32 single EI-series (20-yrs- smoothed, anomalies from 1978-2007 average)
Smoothed seasonal asymmetry:
19th century more continental than 20th
This
has severe implications due to the given seasonal asymmetry of many natural proxies
13.7.2010, Edinburgh 10th IMSC
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1750 1800 1850 1900 1950 2000
°C
summer half year (APR–SEP)
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1750 1800 1850 1900 1950 2000
°C
winter half year (OCT–MAR)
bold black: HISTALP-LSS-Tm-2008, EI-corrected (Böhm et al., 2008), bold grey: HISTALP-LSS-Tm-2007, not EI-corrected (Auer et al, 2007),thin red: ISAC-homogenised, mean of 7 N-Italian series (Brunetti et al., 2006b), thin pink: Central European mean of CRUtem2v gridboxes 40-50N, 10-20E (Trenberth and Jones, 2007), thin light blue: Uppsala-Stockholm mean, not EI-corrected (Moberg et al., 2002, Bergström and Moberg, 2002),thin dark blue: Uppsala-Stockholm mean, EI-corrected (Moberg et al., 2003, Moberg et al., 2005), thin green: central England series (Manley, 1974, Parker et al., 1992), thin violet: de Bilt series (van Engelen et al., 1995), thin brown: St. Petersburg series (Jones and Lister, 2002)
OUR RESULT: compared to other EIP-series (20-years smoothed, anomalies from1851-2000 average)
St.PetersburgISAC Northern Italy
Moberg ori corr
CRU-CEU
De-Bilt
CRU-CEN
My conclusion:
we are nearer to the truth now
but to be realistic, we still have an unpleasant
„spaghetti-type“ situation in the early instrumental period
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases are of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
THE REMAINING THREE POINTS:
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases are of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
the size of the problem: breaks – outliers - gaps
histalp T01-gap rate (%)
0
5
10
15
20
25
30
17
60
17
70
17
80
17
90
18
00
18
10
18
20
18
30
18
40
18
50
18
60
18
70
18
80
18
90
19
00
19
10
19
20
19
30
19
40
19
50
19
60
19
70
19
80
19
90
20
00
1 532 3 4 3 6
1: early period2: first networks with continuity problems3: well organised period of meteorological services (IMO - since 1873)4: world war 15: world war 26: recent problems in the Italian network and due to the political situation in Bosnia and Herzegovina
HISTALP-T01: Frequency distribution of detected and removed breaks
0
50
100
150
200
250
300
350
-6.0
-5.5
-5.0
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
deg C (class width 0.5 deg)
pe
rmill
e
minimum: -5.1 deg
maximum: +4.6 deg
HISTALP-T01-2004: frequency distribution of detected and removed real outliers
1
10
100
1000
-28
-25
-22
-19
-16
-13
-10 -7 -4 -1 2 5 8 11 14 17 20
min = -27.5°C max = 16.0 °C
deg C (class width = 1 deg)
rela
tive
freq
uenc
y (p
erm
ille)
Sample: all HISTALP temperature series (135)
breaksgaps
outliers
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
TAKE-HOME MESSAGE 1:
Alexandersson
Vincent
Caussinus-Mestre Peterson-Easterling
Rhodes-Salinger Craddock
INTERNATIONAL JOURNAL OF CLIMATOLOGY 18: 1493–1517 (1998)HOMOGENEITY ADJUSTMENTS OF IN SITU ATMOSPHERIC CLIMATE DATA: A REVIEWTHOMAS C. PETERSON, DAVID R. EASTERLING, THOMAS R. KARL, PAVEL GROISMAN, NEVILLE NICHOLLS, NEIL PLUMMER, SIMON TOROK, INGEBORG AUER, REINHARD BOEHM, DONALD GULLETT, LUCIE VINCENT, RAINO HEINO, HEIKKI TUOMENVIRTA, OLIVIER MESTRE, TAMAS SZENTIMREY, JAMES SALINGER, EIRIK J. FØRLAND, INGER HANSSEN-BAUER, HANS ALEXANDERSSON, PHILIP JONES and DAVID PARKER
Szentimrey
Bradley-JonesPRODIGE
MASH
SNHT
HOCLIS
LIKE A 15 YEARS OLD SINGLE MALT:
Della Marta
Currently some of them are critically analysed and compared within cost action HOME
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
TAKE-HOME MESSAGE 2:
This is much work which should preferably be done by specialised regional groups close to the metadata – this produces the best results, is more effective and saves the time of research groups wanting to analyse the data
No single longterm climate time series is a priori homogeneous (free from non climatic noise)
At average each 20 to 30 years a break is produced which significantly modifies the series
Many but not all of these single breaks are random if a regional (global) sample is averaged
The biases of the order of the real climate signal
There are a number of mathematical-.statistical procedures which - preferably if combined with metadata information from station history files – are able to detect and remove (or at least reduce) the non climatic information
BUT:
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
A challenge not adequatly solved yet :
The ongoing automation of networkse.g.: the change from CAPBELL-STOKES to photoelectric sunshine recorders
-15
-10
-5
0
5
10
15
20
25
JAN
FE
B
MA
R
AP
R
MA
I
JUN
JUL
AU
G
SE
P
OC
T
NO
V
DE
C
%
<600m asl
>2000m asl
NEW minus OLD SYSTEM
An example from 10 Austrian sites with parallel measurements – 4 at high and 6 at low elevation
Overburning at H-elev.-sites due to intermittant cumulus clouds there in summer
Wetting losses at Low elev. sites in winter
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
Another one:
DOING THE STEP FROM MONTHLY RESOLVED TO DAILY SERIES
SPATIAL DECORRELATION TO 50% COMMON VARANCE
(EUROPE, SYNOP NETWORK)
Source:
Scheifinger and Böhm, 2005)42
993
765
722
533
149
120
105
0 200 400 600 800 1000
YEARLY
SEASONAL
MONTHLY
DAILY
YEARLY
SEASONAL
MONTHLY
DAILY
TE
MP
ER
AT
UR
EP
RE
CIP
ITA
TIO
N
mean Dekorrelation-distance (km)
THE MAIN PROBLEM: HIGHER NETWORK DENSITY NECESSARY – IN PARTICULAR FOR PRECIPITATION
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
More about the homogenizing business:
this afternoon (14.30-17.45) in session 8 (St. Trinneann)
Starting with Olivier Mestre‘s and Victor Venema‘s homogenization procedures intercomparison (cost action HOME)
and at the poster session 8 (Pentland) starting here right now
13.7.2010, Edinburgh 10th IMSC
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies to 1
85
1-2
000
in K
SUMMER HALF YEAR (AMJJAS)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
ano
ma
lies t
o 1
85
1-2
000
in K
SUMMER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
851
-20
00 in K
WINTER HALF YEAR (ONDJFM)
ORIGINAL
-3
-2
-1
0
1
2
3
4
1750 1800 1850 1900 1950 2000
anom
alie
s t
o 1
85
1-2
000 in K
WINTER HALF YEAR
EIP-CORRECTED HOM-VERSION - 2008THANK‘S FOR LISTENING ANYWAY
MUCH OF WHAT I HAVE SAID TO AN AUDITORIUM LIKE THIS WAS OF COURSE
CARRYING
EDINBURGH
TO
OWLS