Upload
duongtruc
View
213
Download
0
Embed Size (px)
Citation preview
LMNH Unit
Luc Feyen & Ad de Roo
With contributions of:Jose Barredo, Rutger Dankers, Katalin Bodis, Carlo Lavalle,
EFAS teamDG Joint Research Centre
European Commission
Assessing Future Flood Risk in Europe:advances in climate change scenario development -
progress in model predictions -research challenges
Working Group F – Workshop on Climate Change and FloodsKarlstad, 8 September 2009
LMNH Unit
source pathway receptor consequence
� extreme precipitation� prolonged wet period� rapid snowmelt
� overtopping, overflow� flood defence� floodplain
� people� property� environment
� loss of life� economic impact� environmental damage
FLOOD HAZARD FLOOD RISK
Assessment of flood risk
LMNH Unit
source pathway receptor consequence
� extreme precipitation� prolonged wet period� rapid snowmelt
� overtopping, overflow� flood defence� floodplain
� people� property� environment
� loss of life� economic impact� environmental damage
socio-economic system
climate
emissions
FLOOD HAZARD FLOOD RISK
terrestrial system
Assessment of future flood risk
LMNH Unit
Return period (years)
hydrologicalhydrologicalhydrologicalhydrological modelmodelmodelmodelextreme extreme extreme extreme valuevaluevaluevalue analysisanalysisanalysisanalysis
Assessment of future flood hazard
climateclimateclimateclimate modelsmodelsmodelsmodels
sociosociosociosocio----economic economic economic economic scenariosscenariosscenariosscenarios
highhighhighhigh----resolutionresolutionresolutionresolution
climateclimateclimateclimate scenariosscenariosscenariosscenarios
LMNH Unit
� capturing fine-scale climatic structures induced by complextopography or land use patterns is essential
� RCM runs at increasing horizontal resolution
� PRUDENCE, ENSEMBLES, CECILIA, CLAVIER, nationalclimate scenarios (e.g., UKCP09, KNMI'06)
� development of a set of “Representative ConcentrationPathways ” (RCPs)
Climate scenario development
LMNH Unit
� relatively few studies have considered extreme river flows
� focus on catchments rather than regional/continental analyses
� geographical preference for catchments in the UK, Benelux, Germany, Central Europa and Scandinavia
� several studies project an increase in flood hazard …
� … others project a decreasing trend
� application of different scenarios, driving climate models, downscaling techniques and hydrological models
Model predictions of changes in floods
� Overall situation in Europe??
LMNH Unit
Relative change in 100-year return level of river discharge between scenario (2071-2100) and control period (1961-1990) – ensemble run
Dankers, R., and L. Feyen, 2009. Flood hazard in Europe in an ensemble of regional climate scenarios, J. Geophys. Res., doi:10.1029/2008JD011523.
Model predictions of changes in floods
ECHAM4/OPYC3 HadAM3HHIRHAM & RCAO
LMNH Unit
Climate change effect on extremes:change of max 5 day Q and mean
annual min 7 day Q
-> increase extremes everywhere
except Scandinavia and Portugal
LMNH Unit
Number of scenarios (out of total eight experiments) showing a decrease (plot a) or increase (b) in Q100 of more than 5% in the scenario period
Dankers, R., and L. Feyen, 2009. Flood hazard in Europe in an ensemble of regional climate scenarios, J. Geophys. Res., doi:10.1029/2008JD011523.
Model predictions of changes in floods
LMNH Unit
� monetary assessments of climate change impacts on floods are poorly covered
� few studies in UK (e.g., Foresight, 2004; Hall et al., 2005;� Mokrech et al., 2008), showing large increases in flood risk, but� strongly dependent on scenario of economic growth
� studies exclusively account for direct tangible damages
� psychological effects� inconvenience of post-� flood recovery
� loss industrial production� traffic disruption
� emergency costsIndirect
� loss of life � health effects � loss ecosystem services
� damage to buildings� damage to infrastructure � damage to contents
Direct
IntangibleTangible
Predictions of changes in flood risk
LMNH Unit
From flood hazard to flood risk
� convert hazard into damage based on static land use� classification and flood depth damage functions
CORINE Land Cover 2000 (EEA, 2000)
Huizinga (2007)
LMNH Unit
Change in flood risk by 2080s in Europe
� change in Expected Annual Damage (averaged over NUTS2)
A2 B2
EAD of EU27, currently €6.5 billion, is projected to rise to €18 billion (A2) and €14 billion (B2)
PESETA study
LMNH Unit
1 10 100 1000protection ctrl (rp in yrs)
0
50
100
150
200
250
exp
an
n d
am (
bill
ion
€) A2
scenario
control
A2, EU27
∆damage = € 31 billion∆damage = +75%∆damage = € 31 billion∆damage = +75%
↑ protectionrp = 10 → 26 y↑ protectionrp = 10 → 26 y
potentialbenefitscost of adaptation
potentialbenefitscost of adaptation
Potential benefits of flood mitigation
LMNH Unit
1 10 100 1000protection ctrl (rp in yrs)
0
50
100
150
200
250
exp
an
n d
am (
bill
ion
€) A2
scenario
control
∆damage = € 11 billion∆damage = +294%
A2, EU27
potentialbenefitscost of adaptation
potentialbenefitscost of adaptation
↑ protectionrp = 100 → 400 y
Potential benefits of flood mitigation
LMNH Unit
1 10 100 1000protection ctrl (rp in yrs)
0
2
4
6
exp
an
n d
am (
bill
ion
€) A2, Danube in RO
control
scenario
1 10 100 1000protection ctrl (rp in yrs)
0
1
2
3
4
exp
an
n d
am (
bill
ion
€) A2, Danube in CZ
scenario
scenario
1 10 100 1000
protection ctrl (rp in yrs)
0
4
8
12
16
exp
an
n d
am (
bill
ion
€) A2, Danube in AT
scenario
scenario
1 10 100 1000protection ctrl (rp in yrs)
0
10
20
30
40
50
exp
an
n d
am (
bill
ion
€) A2, Danube
scenario
control
Potential benefits of flood mitigation
LMNH Unit
Key Research Challenges
• Major research advances are needed in climatology, hydrology, land-use planning, socio-economic sciences and multi-objective decision-making under uncertainty
• need for sustained, high-quality observations & data sets– E.g. at European scale / transnational river basins, good meteo datasets allowing
downscaling & bias correction are hardly available
• advance scientific understanding of the mechanisms that trigger, and alter the probabilities of, extreme events
– e.g. Vb conditions triggering 2002 Elbe floods
• interaction/feedback between land-use and climate change and variability– Consequences of land use change hardly taken into account in GCM & RCM’s
• interaction/feedback between hydrological cycle and climate– e.g. extreme frequency of thunderstorms in Italy in 2009 <> relation to wet topsoil
conditions?
• improve capabilities of regional climate models to represent and predict variability and extremes at regional and local scale
LMNH Unit
� better quantification of (intangible) damages
� better quantification of cost/benefits of adaptation options
� better quantification of current and future vulnerabilityand of the reliability of protection measures
� formal treatment of uncertainty in the chain “emissions – climate – extreme flows – inundation – damage”
� flood risk mapping and management in the face of these uncertainties
Key Research Challenges
LMNH Unit
Conclusions• Even when scenario results are variable, there are some
emerging trends:– Increase of 100yr flood in several large EU rivers (Danube,
Rhone, Po, Loire, Garonne)– Decrease in snowmelt 100yr floods in parts of Scandinavia– Even with strong decrease in annual rainfall in the
Mediterranean, still substantial or higher flood risk
• Methodology for cost-benefit analysis for flood protection at large scales– Refinements needed & ongoing:
• New RCMs, multi-ensembles, multi-model approaches• Etc etc
• A large number of (research) challenges remain
LMNH Unit
http://floods.jrc.ec.europa.eu Thanks for your attention!