Download Climate Change Detection: The Importance of Homogenized Time

Workshop on Climate Change Adaptation Indicators
JERUSALEM, ISRAEL
24 – 26 March 2015
Climate change detection - The importance
of homogenized time series
Yizhak Yosef
Climate Department
Israel Meteorological Service
Outline – Part I
• Short introduction to homogenization
• The integrated homogenization model used in the IMS
• Results
– Case study for Negba minimum temperature 1950-2012
– Quick over view on five adjusted series 1950-2012
• Conclusions and summary
2
Outline – Part II
Analysis of extremes in a changing climate –
Extremes Indices
Part III – Avner Furshpan
Climate Change in Israel – IMS findings report,
March 2015
HOMOGENIZATION
Most long-term climatological time series have been
affected by number of non-climatic factors that make
these data unrepresentative of the actual climate
variation occurring over time.
Homogeneity testing is preformed to ensure that time
fluctuations in the data are only due to the vagaries of
weather and climate.
Aguilar E, Auer I, Brunet M, Peterson TC, Wieringa J. 2003. Guidelines on climate metadata and homogenization. WCDMP-No. 53,
WMO-TD No. 1186. World Meteorological Organization, Geneva.
Long-term climate change
Apparent
Real
Local
Directly or indirectly climatic
• Station relocations
• Changes in vegetation,
• Change in circulation
• Screen design
soil or drainage
• Change in planetary
• Instrumentation:
• Re/deforestation,
albedo, ice and snow
Replacement, Change in
artificial lakes
• Change in atmospheric
calibration, maintenance
• Building construction
transparency (aerosols,
• Change in observing times
• Urbanization
CO2)
• Changing in averaging
• Industrialization
• Extraterrestrial changes
methods
(e.g solar constant)
• Change in the close vicinity
of the station
Raino Heino (1996), Metadata and their role in homogenization, Proceedings of the First Seminar for Homogenization and Quality Control
in Climatological Databases, Budapest, Hungary, 6-12 October 1996.
Common Homogeneity Procedure
Quality Control for all the data (base and reference)
Collecting relevant Metadata
Composing reference series
Homogeneity tests (absolute & relative)
Break-points detection
Adjustment
6
AnClimv5 (Štěpánek, P. 2008)
ACMANTv2 (Domonkos, P. 2012)
HOMER_2.6 (Mestre, O. et al, 2012)
RHtestsV4 (Wang, X. L & Feng, Y. 2013)
R Development
Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing,
7
Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/.
Homogenization Model at IMS
Absolute
AnClim,
RHtestsV4
Relative methods
Composing one reference
series based on weighed
average - AnClim
HOMER
ACMANT
Q.C, Cluster
Analysis
Applying metadata
k
k

Qi  yi    2j [ x ji  x j  y ] /   2j 
j 1
 j 1

Full H.T
RHtestsV4
AnClim
applying
metadata
Summarizing all the results and Establishing the Break-Points locations
Adjustment
Yosef, Y., I. Osetinsky, and A. Furshpan (2014), Homogenization of monthly temperature series in Israel – an integrated approach for optimal
break-points detection. Proceedings of the 8th Seminar for Homogenization and Quality Control in Climatological Databases, Budapest,
Hungary, 12-16 may 2014, WCDMP. WMO, Submitted.
Key factors causing inhomogeneity in the
Israeli climatological stations temperature data
 Instrumentation (calibration, upgrading to electronic
sensors, types of sensors).
 Relocations
 Changing screen design and maintenance.
9
Focusing on 5 stations
 Long climate records.
 Data use after a systematic Quality Control procedure.
 Availability of reliable metadata.
 Good representation of various climate regions in Israel.
 The average of the mean temperature of these 5
stations represents quite well the average temperature of
all Israel.
Case study
Negba
10
Using Cluster Analysis to choose reference stations
Beit Jimal
Gat
11
Dorot
Qevuzat
Yavne
Negba
Hafez Hayyim
Mazkeret
Besor
Batya
Farm
Break-point detection using RHtestV4
Negba: Monthly differences between the base and reference minimum temperature series
Thermometer replacement
False detection
Relocation & Change in design
Relocation
Switching sensor in 2012
The method is based on, Wang, X. L., 2008: Accounting for autocorrelation in detecting mean shifts in climate data series
using the penalized maximal t or F test. Journal of Applied Meteorology and Climatology, 47, 2423–2444.
Metadata were used only after the detection phase, to validate the results
12
Adjusted series and correction factors vs. Base (raw data)
Negba Tmin Annual Adjustment
2.0
1.5
1.0
0.5
base
2012
2010
2008
2000
1998
1996
o C/100y
TN
TX
Base
2.65
0.74
1.6
0.35
1.5
After
1.0
adjustment
0.5
13
base
adjusted
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
-1.5
2010
-1.0
Correction
Factors
0.4
0.97
0.4
-0.1
-0.33
-0.15
2008
-0.5
Break
Points
1952
1953
1960
1963
1967
1974
2006
0.0
2004
Temperature Anomaly [0C]
Negba Tmax Annual Adjustment
2.0
1994
adjusted
*Difference from 1961-1990
2.5
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
-1.5
Correction
Factors
0.39
0.63
0.47
2006
-1.0
2004
-0.5
2002
Break
Points
1965
1970
1977
0.0
2002
Temperature Anomaly [0C]
2.5
TX series results: Adjusted vs. Base
35.0
33.0
31.0
29.0
27.0
25.0
23.0
21.0
19.0
14
Eilat adj
Eilat base
Negba adj
Negba base
Beit Jimal adj
Beit Jimal base
Jerusalem adj
Jerusalem base
Zefat adj
Zefat base
2012
2010
2008
2006
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
17.0
TX series results: Adjusted vs. Base
15
TX series results: Adjusted vs. Base
16
TN series results: Adjusted vs. Base
22.0
20.0
18.0
16.0
14.0
12.0
17
Eilat adj
Eilat base
Negba adj
Negba base
Beit Jimal adj
Beit Jimal base
Jerusalem adj
Jerusalem base
Zefat adj
Zefat base
2012
2010
2008
2006
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
10.0
TN series results: Adjusted vs. Base
18
TN series results: Adjusted vs. Base
19
Causes of the inhomogeneities identified
unknown
23%
instrumentation
48%
enviroment
6%
screen
(design &
maintenance)
13%
relocation
10%
Frequency distribution of break magnitudes
(No. of breaks, monthly resolution)
Conclusions and summary
• Long time series have been affected by a number of non climatic factors
(relocation, instrumentation, screen design etc.).
• Some changes may cause sharp discontinuities while other gradual change
biases in the data.
• In average there are 4-5 break-points in each of the examined stations.
• The annual average correction factors interval for the 5 stations are:
 Tx: [-1.23, 0.71] [oC]
 Tn: [-0.48, 1.09] [oC].
• Before analyzing trends for climate change it is important first to do quality
control and then to homogenized (find the break-points and correct the bias)
the series.
An inhomogenized time series can lead to wrong interpretation of the climate
Part II – Extreme Indices
World Meteorological Organization (2009), Guidelines on analysis of extremes in a changing climate in support of informed decisions
for adaptation, Tech. Rep. 72, Geneva, Switzerland.
Analysis of extremes in a
changing climate
Commission for Climatology (CCI)
Climate Variability and Predictability (CLIVAR)
Joint WMO-IOC Technical Commission for Oceanography and Marine Meteorology (JCOMM)
Expert Team on Climate Change Detection and Indices (ETCCDI)
Extreme Indices
In order to gain a uniform perspective on
observed changes in weather and climate
extremes a core set of 27 indices for
moderate/soft temperature and precipitation
extremes was defined.
1.
ID
FD0
Indicator name
Frost days
Definitions
Annual count when TN(daily minimum)<0ºC
UNITS
Days
2.
SU25
Summer days
Annual count when TX(daily maximum)>25ºC
Days
3.
ID0
Ice days
Annual count when TX(daily maximum)<0ºC
Days
4.
TR20
Tropical nights
Days
5.
GSL
6.
TXx
Max Tmax
Annual count when TN(daily minimum)>20ºC
Annual (1st Jan to 31st Dec in NH, 1st July to 30th June
in SH) count between first span of at least 6 days with
TG>5ºC and first span after July 1 (January 1 in SH)
of 6 days with TG<5ºC
Monthly maximum value of daily maximum temp
7.
TNx
Max Tmin
Monthly maximum value of daily minimum temp
ºC
8.
TXn
Min Tmax
Monthly minimum value of daily maximum temp
ºC
TNn
Min Tmin
Monthly minimum value of daily minimum temp
ºC
10. TN10p
Cool nights
Percentage of days when TN<10th percentile
%
11. TX10p
Cool days
Percentage of days when TX<10th percentile
%
12. TN90p
Warm nights
Percentage of days when TN>90th percentile
%
13. TX90p
Warm days
14. WSDI
Percentage of days when TX>90th percentile
Annual count of days with at least 6 consecutive days
when TX>90th percentile
Annual count of days with at least 6 consecutive days
when TN<10th percentile
%
Warm spell duration
indicator
Cold spell duration
indicator
Diurnal temperature
range
Max 1-day
precipitation amount
Max 5-day
precipitation amount
Simple daily intensity
index
Number of heavy
precipitation days
Number of very
heavy precipitation
days
Number of days
above nn mm
Consecutive dry days
Monthly mean difference between TX and TN
ºC
Monthly maximum 1-day precipitation
mm
Monthly maximum consecutive 5-day precipitation
mm
Annual total precipitation divided by the number of
wet days (defined as PRCP>=1.0mm) in the year
mm/day
Annual count of days when PRCP>=10mm
Days
Annual count of days when PRCP>=20mm
Days
9.
15. CSDI
16. DTR
17. RX1day
18. Rx5day
19. SDII
20. R10
21. R20
22. Rnn
23. CDD
24. CWD
25. R95p
26. R99p
27. PRCPTOT
Growing season
Length
Annual count of days when PRCP>=nn mm, nn is user
defined threshold
Maximum number of consecutive days with RR<1mm
Consecutive wet days Maximum number of consecutive days with
RR>=1mm
Very wet days
Annual total PRCP when RR>95th percentile
Extremely wet days Annual total PRCP when RR>99th percentile
Annual total wet-day
Annual total PRCP in wet days (RR>=1mm)
precipitation
Days
ºC
An internationally
coordinated core set
of 27 indices
(moderate/soft extremes)
Days
Days
Days
Days
Days
mm
mm
mm
For detailed calculation see: World Meteorological
Organization (2009), Guidelines on analysis of
extremes in a changing climate in support of
informed decisions for adaptation, Tech. Rep. 72,
Geneva, Switzerland.
http://www.clivar.org/panels-and-workinggroups/etccdi/etccdi.php
27 core
indices
16 indices for
temperature
11 indices for
precipitation
• The indices describe different aspects of moderate temperature and precipitation
extremes, including frequency, intensity and duration.
• The indices are based on daily temperature and/or daily precipitation amount.
• Some are based on fixed thresholds.
• Other indices are based on thresholds that vary from location to location and
thresholds are typically defined as a percentile of the relevant data series.
Core Indices
SECTOR
Indicators
SECTOR
SECTOR
…
Thank You for your attention!
Email: [email protected]
29
Different screen design
1952
1977
2004
Jerusalem
– Center
(Generally building)
From Conventional Weather Station to AWS
Zefat Har Kenaan – Change in the close vicinity of the screen
1939
1957
1952
2005