Transcript Detection of historical trends - World Conference on Climate Change

Session: Evidence of Climate changes | Sustainability & Climate Change |
Risks of Climate Change
Historical and RCM future trends in
Northern Tuscany (Italy)
M.G. Tanda1, M. D'Oria1, M. Ferraresi1 and P. Molini1
1 Department
of Civil, Environmental, Land Engineering and Architecture,
University of Parma, Parco Area delle Scienze 181/A, 43124, Parma,
Italy
Corresponding Author M.G. Tanda : [email protected]
Northern Tuscany (Italy)
Foto Andrea Ribolini
The land of the marble mountains : here the greatest
artists come to choose the marble blocks for their
sculptures.
Aims and scopes
The climate change is a topic included in the
agenda
of
politics
due
to
its
remarkable
impications.
The availability of the water resources could decrease in the future and
the Water Companies have developed in Europe a new awareness to this
problem in view of developping strategies resilient to climate stresses.
I’ll present some results of a research conducted about the northern
Tuscany region in order to identify possible climate change elements in
the historical temperature and rainfall data.
Then I’ll compare the results of the historical analysis with the forecasts
of the Regional Climate Models (RCM).
Study area
Region of competence of the
Water Company GAIA S.p.A.
Total area = 2900 km2
•
River Serchio catchment
•
Part of the River Magra
basin
•
Streams on the apuanoversiliese slopes
•
Very heterogeneous orography : the elevations range from the
sea level to about 1800 m a.s.l.
•
Annual rainfall from 900 mm to 3000 mm.
Available data
Rainfall data from 259 stations and
temperature observations from 141 stations.
The oldest were collected on 1916.
Rainfall
Temperature
Available
Daily data
Missing data
Aggregate data
Analysis of the historical trends
18 rainfall and 14 temperature stations with a long time series of daily
observations have been selected: 79-97 years of rainfall data and 62-89
mean daily temperature records have been analyzed.
Rainfall station locations
Temperature station locations
Detection of historical trends
The existence of an eventual trend (monotone) in the historical data
has been investigated by means of the Mann-Kendall test (Mann, 1945;
Kendall, 1975) modified in order to remove the data autocorrelation
(Hamed and Rao, 1998) and evaluating at the 5% significance level.
Station: S. Marcello Pistoiese
Historical data
Linear trend
Theil-Sen trend
Assuming a linear trend, slope and
evaluated with the Theil-Sen method
(Theil, 1950; Sen, 1968).
The computations have been carried out
at the month and year scale.
Annual rainfall (mm)
intercept of the trend line have been
Detection of historical trends
H = 1 means there is a 5% significance level !
Annual rainfall
Station
Bedonia
Borgo a Mozzano
Calice al Cornoviglio
Carrara
Casania
Cembrano
Lucca
Massa
Pescia
Pontremoli Verdeno
S. Marcello Pistoiese
Sarzana
Villacollemandina
Bagnone
Arlia
Villafranca Lunigiana
Palagnana
Viareggio
M-K Index
-322
-424
-339
-552
-852
-528
-700
-231
-424
-354
-588
-94
-322
-673
-243
-934
-596
138
A decreasing trend of the annual total
H
0
0
0
0
1
1
0
0
0
0
1
0
0
1
0
1
1
0
rainfall has been detected in 17
stations (on 18).
The decreasing trend seems to be
more remarkable in the first semester
of the year.
In the majority of the cases the trend
doesn’t reach the 5% significance
level.
Detection of historical trends
Rainfall annual trend
Mean gradient per 10 years (mm/10 years)
May
Rainfall annual trend
Mean gradient per 10 years (mm/10 years)
Total annual rainfall
Mean gradient per 10 years evaluated
with the Theil-Sen method.
Positive gradient
Negative gradient
September
Rainfall annual trend
Mean gradient per 10 years (mm/10 years)
Detection of historical trends
Mean annual temperature
H = 1 means there is a 5% significance level !
Station
Arlia
Bagnone
Boscolungo
Calice al Cornoviglio
Castelmartini
Castelnuovo
Garfagnana
La Spezia
Lucca
Massa
Pontremoli
S. Marcello
Pistoiese
Sarzana
Tavarone
Viareggio
M-K Index
-620
1533
652
787
223
H
1
1
1
1
0
1196
1
1378
801
1680
-860
1
1
1
1
320
0
677
869
1035
1
1
1
An increasing trend of the mean
annual temperature has been
detected in 12 stations (on 14).
10 stations (on 14) show at a 5%
significance level an increasing trend.
Detection of historical trends
Mean annual temperature trend
Mean gradient per 10 years (°C/10 years)
January
Mean annual temperature trend
Mean gradient per 10 years (°C/10 years)
Mean annual temperature
Mean gradient per 10 years evaluated
with the Theil-Sen method.
Positive gradient
Negative gradient
August
Mean annual temperature trend
Mean gradient per 10 years (°C/10 years)
Regional Climate Model (RCM) Projections
RCM
GCM
CCLM
HIRHAM5
WRF331F
RACMO22E
RCA4
CNRMCERFACSCNRM-CM5
x
ICHECECEARTH
x
x
MOHCHadGEM2ES
x
MPI-MMPI-ESMLR
x
IPSL-IPSLCM5A-MR
x
x
x
x
x
x
x
x
13 different combinations between General
Climate Models (GCM) and Regional Climate
Models (RCM)
2 future forcing emission
(EURO-CORDEX project, Jacob et al., 2014)
scenarios have been
The space resolution is 0.110, a square grid
considered:
with 12.5 km size
RCP 4.5 and RCP 8.5
In the Euro-Cordex project the simulated time
horizons were:
•
Control period (1950-2005 or 1970-2005)
•
Scenarios period (2006-2100)
Regional Climate Model (RCM) Projections
The results of the climate models
the historical observations
Correction before their use (bias
Cumulative Probability
in the control period differ from
Observed data
Original RCM model
Corrected RCM model
correction)
Temperature (°C)
Distribution Mapping Method
(Teutschbein & Seibert, 2012)
Mean Temperature (°C)
Observed data
Original RCM model
Corrected RCM model
Thirty years control period is
1976-2005
Month
Comparison between the RCM projections and the historical
trends
We have compared the values extrapolated
from the linear historical trends with the
projections of the regional climate models.
Detailed analysis for the projections of:
Box-plot symbol
Maximum value
3th quartile (perc 75%)
•
Decade 2003-2012 (present*)
Median value (2th quart. or 50% perc)
•
Decade 2031-40 (medium term)
•
Decade 2051-60 (long term)
*Notice: the period 2003-2012 acts in simulation up to
2005 e in projection (with the two emission scenarios)
up to 2012.
1th quartile (perc 25%)
Minimum value
Mean Temperature (°C)
Comparison between the RCM projections and the historical
trends
Temperature station of Massa
Comparison between the RCM projections and the historical
Trend falls between 1 and 3
trends
Temperature
th
th
quartile of RCMs
Trend falls between maximum and
minimum of RCMs
Trend under-estimates the RCM
values
Trend over-estimates the RCM values
There is a good agreement, between the values
extrapolated from the linear trend and the outcomes of
the 13 RCM models, only in the first control period
(2003-2012) except for September when there is a
systematic under-estimate.
For the medium and long time horizons a general underestimate, of the values extrapolated from the linear
trend on respect of the projections of the RCM, can be
noticed.
Sporadically, cases of over-estimation of the linear
historical trends with respect of RCM values appear.
Comparison between the RCM projections and the historical
trends
Rainfall
Trend falls between 1th and 3th
quartile of RCMs
Trend falls between maximum
and minimum of RCMs
Trend under-estimates the RCM
values
Trend over-estimates the RCM
values
In most cases the values extrapolated from
the linear trend fall in the range between
the maximum and minimum outcomes of
the 13 RCM models, with a remarkable
number of case of good agreement.
There is a systematic under-estimation of
the trend values in March and April and
more prominently in May.
Over-estimation of the trend values on the
outcomes of the RCM models never appears.
Conclusions
 From the analysis of historical data it has shown, globally in the study area, that a
decreasing trend in annual precipitation and an increasing trend in the average
annual temperature can be detected. The tendency for precipitation is often not
statistically significant, whereas it is verified for the temperature.
 The comparison of the climate change signals extrapolated from local trend
analysis and the RCM models showed a systematic underestimation with
reference to the thermometric data.
 As a results, we decided to use the medium and long-term forecasts produced by
the projection of the RCM climate models in the evaluation of the water resources
modifications due to the climate change in the competence area of the water
company GAIA
 Moreover the linear extrapolation of trends does not take into account the
different, albeit hypothetical, scenarios of emissions that can lead to modification
of the climate variables.
Main references
•
Jacob, D., Petersen, J., Eggert, B., Alias, A., Bøssing Christensen, O., Bouwer L.M., Braun, A., Colette, A., Déqué, M.,
Georgievski, G., Georgopoulou, E., Gobiet, A., Menut, L., Nikulin, G., Andreas Haensler, A., Hempelmann, N., Jones,
C., Keuler, K., Kovats, S., Kröner, N., Kotlarski, S., Kriegsmann, A., Martin, E., van Meijgaard, E., Moseley, C., Pfeifer,
S., Preuschmann, S., Radermacher, C., Radtke, K., Rechid, D., Rounsevell, M., Samuelsson, P., Somot, S., Soussana,
J., Teichmann, C., Valentini, R., Vautard, R., Weber, B. & Yiou, P. EURO-CORDEX: new high-resolution climate
change projections for European impact research. Regional Environmental Change, 2014, 14, 563-578.
•
Mann, H.B. Non-parametric tests against trend, Econometrica, 1945, 13, 163-171.
•
Kendall, M.G. Rank Correlation Methods, 4th edition, Charles Griffin, London, 1975.
•
Sen, P.K. Estimates of the regression coefficient based on Kendall's tau. Journal of the American Statistical
Association, 1968, 63, 1379-1389.
•
Teutschbein, C. & Seibert, J. Bias correction of regional climate model simulations for hydrological climate-change
impact studies: review and evaluation of different methods. Journal of Hydrology, 2012, 456-457, 12-29.
•
Theil, H. A rank-invariant method of linear and polynomial regression analysis, Part 3. Proceedings of Koninalijke
Nederlandse Akademie van Weinenschatpen, 1950.
Acknowledgements
This study has been done in the framework of a research agreement
between the Department of Civil, Environmental and land management
Engineering and Architecture Of the University of Parma (Italy) and the
Water Company GAIA S.p.A. that I thank for the economic support and the
help in the data collection.
THE END !
THANK YOU FOR YOUR ATTENTION