The PRUDENCE The PRUDENCE projectproject

Jens Hesselbjerg ChristensenPRUDENCE coordinator

jhc@dmi.dkhttp://prudence.dmi.dk

COP10, 13 December 2004

1. Danish Meteorological Institute, Copenhagen, DK2. CINECA, Bologna, IT3. Météo-France/CNRM, Toulouse, FRA4. Deutsches Zentrum für Luft- und Raumfahrt e.V., Weßling, GER5. Hadley Centre for Climate Prediction and Research, Met Office, Bracknell, UK6. Climate Research ETH (Eidsgenössische Technische Hochschule), Zürich, CH7. GKSS Research Center (Institute for Coastal Research), Geesthacht, GER8. Max Planck Institut für Meteorologie, Hamburg, GER9. Swedish Meteorological and Hydrological Institute, Rossby Centre, Norrköping, SWE10. Universidad Complutense, Madrid, SP11. Universidad Politecnica, Madrid, SP12. International Centre for Theoretical Physics, Trieste, IT13. Danish Institute of Agricultural Sciences, Foulum, DK14. Risø National Laboratory, System Analysis Dept., DK15. University of Fribourg, CH16. Finnish Environmental Institute, Helsinki, FIN17. University of Reading, UK18. University of Lund, SWE19. Centre International de Reserche sur l’Environment et Developpement,SMASH, Paris, FRA20. Climate Research Unit, University of East Anglia, UK21. Finnish Meteorological Institute, Associated to FEI (No. 16), FINA. Norwegian Meteorological Institute, Blindern, NOB. Royal Dutch Meteorological Institute, De Bilt, NLC. UQAM, Montreal, CAND. CSIRO, Victoria, AUSE. Czech Republic, Israel, Greece, California, USA………………..F. Munich-Re, Electricité de France, Elforsk, Hamburg Institute of International Economics, Uni-Münster, DG-

Research, STARDEX, MICE

The PRUDENCE Consortium

Overview this presentation

• The scientific objectives

• Aims and products

• Results

PRUDENCE objectives

• provide a series of provide a series of high resolutionhigh resolution (spatially (spatially and in time) climate change scenarios for and in time) climate change scenarios for 2071-2100 for Europe;2071-2100 for Europe;

• assess the assess the uncertaintyuncertainty in European regional in European regional climate scenarios resulting from model climate scenarios resulting from model formulationformulation;;

• in practical terms characterise the level of in practical terms characterise the level of confidenceconfidence in these scenarios and the in these scenarios and the variability in them related to model variability in them related to model formulations and climate natural/internal formulations and climate natural/internal variability;variability;

PRUDENCE objectives

• quantitatively assess the quantitatively assess the risksrisks rising rising from changes in regional weather and from changes in regional weather and climate over all of Europe, and estimate climate over all of Europe, and estimate future changes in future changes in extreme eventsextreme events such such as flooding and wind storms, by as flooding and wind storms, by providing a robust estimation of the providing a robust estimation of the likelihood and magnitude of the likelihood and magnitude of the changeschanges;;

PRUDENCE objectives

• demonstrate the valuedemonstrate the value of the wide-ranging of the wide-ranging climate change scenarios by applying them climate change scenarios by applying them to climate impacts models focusing on to climate impacts models focusing on effects on adaptation and mitigation effects on adaptation and mitigation strategies;strategies;

• assess assess socio-economicsocio-economic and policy related and policy related decisions for which such improved scenarios decisions for which such improved scenarios could be beneficial;could be beneficial;

• disseminate the resultsdisseminate the results of PRUDENCE widely of PRUDENCE widely and provide a project summary aimed at and provide a project summary aimed at policy makers and non-technical interested policy makers and non-technical interested partiesparties

Regional aspects of coupled ocean-atmosphere general circulation models

Regional information

A2 & B2

Temperature change relative to global mean

(Giorgi et al. GRL, 2001)

Regional aspects of coupled ocean-atmosphere general circulation models

Regional information

Time slice experiments utilizing high- and variable- resolution atmospheric GCM’s

Regional aspects of coupled ocean-atmosphere general circulation models

Regional information

Limited area (regional climate) models RCM’s Statistical down scaling (see STARDEX)

Time slice experiments utilizing high- and variable- resolution atmospheric GCM’s

Impact scenarios

today

scenario

GCM

today

scenario

RCM

scaleglobal local

An interface

ImpactImpactmodelmodel

A road to impact scenariosA road to impact scenariosthe Delta Change approachthe Delta Change approach

UNCERTAINTIES IN CLIMATE CHANGE PROJECTIONS

• Uncertainty due to observational limitations

– use multiple means of validation

• Uncertainty in future emissions

– use a range of SRES emissions scenarios

• Natural variability

– use a number of different initial conditions (ensembles)

• Uncertainty in the response of the climate system

– use a range of climate modelling systems

– AND/OR assess confidence in climate projections (better models)

• Need for a large-scale coordinated effort

CO2 EMISSIONS PROFILESunder IPCC SRES scenarios

Source:IPCC

GLOBAL TEMPERATURE RISEdue to four SRES emissions scenarios

Source: Hadley Centre

AGCM exp forcing

HadAM3H

ARPEGE ECHAM5 CCM3

HadCM3 SRES A2

3 ensemble members 150 km BDY 1

2 mem high res.

1 member T106

2 members T80 BDY4

HadCM3 SRES B2

1 member 150 km BDY 2

1 mem high res.

ECHAM4/OPYC3 SRES A2

1 member T106 BDY 3

ARPEGE/OPA SRES B2

1 mem high res.

RCM 50km Input

P5

P9

P1 P10 P6 P2/P12 P8 P7

BDY 1 3 mem 3 mem BDY 1 1 mem 1 mem 1 mem 1 mem 1 mem 1 mem BDY 1 ini cond.

1 mem

BDY 2 1 mem 1 mem 1 mem BDY 3 1 mem BDY 4 1 mem RCM 20 km Input

P5 P9 P1 P6 P8

BDY 1 1 mem 1 mem 1 mem 1 mem 1 mem

DMIHad Rossby Es ETH IPCC MPI GKSS

MeteoT42

100km

12 km

Input

Impact model GCM RCM 50 km RCM 20 km

N.E. storm surge one mem one from each using BDY1

all members

C.E. river catchment probed well probed all N.E. drainage area all all all S.E. agriculture All (2 for adaptive

responses) all (only 2 for adaptive responses)

Only one from contractor/ partner 10 –UCM- for impact and adaptive responses

N.E. agriculture all (only 1 for adaptive responses)

all (only 2-3 for adaptive responses)

all (only 2 for adaptive responses)

ecosystems all well probed simple models and indices

all all all

Mediterranean agriculture and hydrology

One parent GCM and associated RCMs at 50km and 20km. Comparisons can be made between (a) different forcings (SRES A2 and B2) (b) different ensemble members (c) different scales. Attention will focus on the range of scenarios.

Relating to observed trends

• Flooding

• Heat wave

Recent European Summer Climate Trends and Extremes

• Summer precipitation over much of Europe and the Mediterranean Basin has shown a decreasing trend in recent decades

• The intensity of summer precipitation events has shown predominant increases throughout Europe

• The western European summer drought of 2003 is considered one of the severest on record.– Heat related casualties in France, Italy, the Netherlands, Portugal,

the United Kingdom, and Spain reached nearly 20,000.– Many countries are experiencing their worst harvest since World

War II.• In contrast, during 2002, many European countries

experienced one of their wettest summers on record.– Weather systems brought widespread heavy rainfall to central

Europe, causing severe flooding along all the major rivers.– The Elbe River reached its highest level in over 500 years of

record

What can PRUDENCE say?

Christensen&Christensen (2003;2004)Change in JAS mean precip (2071-2100 minus 1961-1990)

Christensen & Christensen, Nature (2003)

Sensitivity due to GCM and RCM resolution

ECHAM Hadley 50km Hadley 25km

Changes in heavy and mean precipitation (1961-90 =>2071-2100)

Schär et al. (2004)

Schär et al. (2004)

Changes in Summer500 hPa Geopotential Heights

( meters) ( meters)

B2 Scenario(2071-2100) minus (1961-1990)

NCEP Reanalysis(1976-2000) minus (1951-1975)

Pal et al. (2004)

Change in Summer PrecipitationB2 Scenario

(2071-2100) minus (1961-1990)

(% change)(% change)

CRU Observations(1976-2000) minus (1951-1975)

Changes in Summer Extremes:B2 Scenario

(% change)( Days)

Max Dry Spell Length(2071-2100) minus (1961-1990)

Max 5-Day Precipitation(2071-2100) minus (1961-1990)

B2

Precipitation Distribution

REF

refB2

B2

ref

MoreDroughts

MoreFloods

DrierSummers

Conclusions I

• In both the A2 and B2 scenarios we find summer warming and drying over most of the European region.

• Maximum dry spell length (drought), maximum precipitation intensity (flood) and interannual variability increase in summer throughout most of Europe

• Shift and change in shape of the precipitation distribution

• The results from the climate change simulations are consistent with trends of summer climate observed over Europe in recent decades

Impacts

• Hydrology

Prudence basins

Baltic Basin

7 RCMs … A2

same GCM boundary

7 RCMs ~50km … A22 RCMs ~25km … A2

same GCM boundary

9 RCMs ~50km … A22 RCMs ~25km … A2

2 GCMs

9 RCMs ~50km … A22 RCMs ~25km … A2

3 RCMs ~50km … B2

2 GCMs

7 RCMs ~50km … A21 GCM ~150km … A2

same GCM boundary

Conclusions 2

• Ensemble information from different models provides valuable information about the degree of uncertainty in the impact signal

• Seasonal shift in hydrological cycle confirmed

Impacts

• Hydrology

• Storm surges

Conclusions 3

• More intensive surge in warmer climate– Up to 30% increase in high percentiles, no

change in mean.– Magnitude of shift is highly dependant on

location.

• Valid for all four RCM simulations, driven with same GCM (HC) and also similar signal for simulation, driven with another GCM

Impacts

• Hydrology

• Storm surges

• Simple indices

Potential shifts in extreme climatic events => risks in society and ecosystems

CLIMATIC CHANGE IN: IMPACT SECTORS ADAPTATION OPTIONS

Maximum 1-day and 5-day precipitation total

Water resources, agriculture

Regulation guidelines, flood gates, land use planning

Maximum length of dry spells

Water resources, agriculture

Increases in water-use efficiency, water recycling

Total number of frost days Ecosystems, transport, heating, building

Preparedness for decreases in energy consumption

Total number of days crossing the 0ºC threshold

Wintertime road maintenance

Timing of salting of roads

Frost-free season Ecosystems, transport Timing of cultivation practices

Snow season Recreation, tourism Artificial snow in ski centres

Maximum ice cover over the Baltic Sea

Wintertime shipping Timing and efficiency of ice-breaking

Changes in frost days and min temperature1961-90 =>2071-2100

Impacts

• Hydrology

• Storm surges

• Simple indices

• Agriculture

Thermal suitability for grain maize (baseline + 2080’s)

green – baseline suitabilityred – suitability extension for all RCMsblue – RCM uncertainty in extension

Suitable area

Observed baseline1961-1990 (CRU)

3 RCMsB2, HadAM3H-driven

9 RCMsA2, HadAM3H-driven

Winter wheat yield in 2080’s (example from 1 RCM)

Modelled 2080’s Difference to CRU baseline

Nitrate leaching from wheat in 2080’s (example from 1 RCM)

Modelled 2080’s Difference to CRU baseline

Conclusions 4

• General productivity increases for agricultural crops in Northern Europe and decreases in Southern Europe has low uncertainty, although the option to cultivate crops during the winter in some Mediterranean countries needs more consideration

• Impacts of nitrate leaching (and possibily other environmental effects of agriculture) may have a completely different spatial structure than the yield effect

Impacts

• Hydrology

• Storm surges

• Simple indices

• Agriculture

• Society

• To compare the future climate of Paris and the present climate of a gridpoint x’, we define 3 distances:

– dT measures the mean absolute distance between the 12 monthly mean temperatures

– dAP measures the relative distance between the annual

mean precipitations (to account for the total water availability)

– dMP measures the mean relative distance between the 12

monthly mean precipitations (to account for the precipitation seasonal cycle)

• 10 European cities: Athens, Barcelona, Berlin, Geneva, London, Madrid, Marseille, Paris, Roma and Stockholm.

Methodology

A global shift southward

Results based on ARPEGE

Impacts

• Hydrology

• Storm surges

• Simple indices

• Agriculture

• Society

• Surprises?

Dec 2001

Sept 2002

Oct 2003

Sept 2004

Thank you all

Utilisation of PRUDENCE data for regional analysis

Ekstrøm et al. (2004)

Assessing uncertainty of regional changes

• Construct a probability distribution function (PDF) of climate change

• Combine PDF from – global annual mean temperature increase– change in regional temperature/precipitation

– per degree of global temperature increase (Jones, 2000)

• (Uniform distributions from within a range)• Normal distribution* of PDF for the scaling

variables, log normal for global increase• Full range of uncertainty

*(estimated from ANalysis Of VAriance (ANOVA) )

(2071-2100) wrt. (1961-1990)

Temperature Precipitation

Ekström et al. (in press)