Flooding in Norway under a near - future 2021-2050 climate: Rainfall vs.
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Flooding in Norway under a near
future 2021-2050 climate: Rainfall vs.
snowmelt floods and their uncertainties
Deborah Lawrence1, Torill Engen-Skaugen2,
Lars Roald1, Gusong Ruan1, Stein Beldring1
1Norwegian Water Resources and Energy Directorate (NVE)
2 Norwegian Meteorological Institute (met.no)
Climate change impacts on flooding
and their uncertainties
Demand for information regarding climate change
impacts on flood frequency for use in adaptation
(dam safety, transport, areal planning)
Differing climate scenarios, models and methods can
produce dissimilar results: How can these be
summarised and used in practical applications?
⇒ Ensemble methodology for analysing hydrological
projections and their uncertainties
Some factors contributing to flooding in Norway
which are likely to be impacted by climate change:
Runoff from rainfall
and/or snowmelt
Hønefossen, May, 2008; Photo:NVE
Intense precipitation
in localised areas
Manndalen, May, 2010; Photo: NVE
Excess rainfall in areas with
limited infiltration
Photo: NVE
Modelling chain for hydrological
impact projections
SRES Emission scenario
Global climate model
Regional climate model
Adjustment of P,T to local 1 x1 km grid
Hydrological models
Flood frequency estimation
140
140
120
mm/måned
120
mm/måned
100
100
80
80
60
60
40
20
00
jan feb
jan feb mar
mar apr
apr mai
mai jun
jun jul
jul aug sep
sep okt
okt nov des
55550 Hafslo
55550 Hafslo 54130 Lærdal
Ensemble approach for probabilistic
hydrological projections
3 GCM/RCMs (with SRES A1B emissions)
Echam5/HIRHAM5
BCM/RCA3
HadCM3Qref/HIRHAM
2 Methods for transferring RCM output
to 1 x 1 km grid
Delta change
Empirical adjustment method (met.no)
25 calibrated hydrological models
for 115 catchments
Flood frequency analysis for
200-year flood
Catchment areas
from 3 to 15,000 km2
⇒ Construct pdfs from
150 results for each catchment
Flood frequency estimation applied
to annual maximum flood series
Viksvatn (Hestadfjord) - 83.2
IS92a ECHAM4/OPYC3 NorClim/HIRHAM 25x25 km
'Empirical Adjustment' to 1 x 1 km
900
850 1981-2010 GEV from annual max series
800
2021-2050 GEV from annual max series
Peak daily discharge (m3/s)
750
700 2021 - 2050 Annual maximum series
650 1981 - 2010 Annual maximum series
600 % change in
550 2021-2050
30-year flood
200-year
500
450
flood
400
350 X X
300
250
200
150 1981-2010
100 200-year flood
1 10 100 1000
Return period (years)
Projected change in 200-yr. flood
between 1961-1990 and 2021-2050
Median of 90th percentile
150 models of 150 models
Change (%) Change (%)
Range of uncertainty in projections
Median of Range
150 models 10 to 90%
Change (%) Range (%)
Uncertainty – Relative magnitude of
sampled sources
Downscaling method
100
Percentage below given value
90 Delta change
80
70 Empirical adjustment
60
50
40
30
20
Uncertainty source
10
0 with largest magnitude:
15 17 19 21 23 25
Mean annual peak runoff (mm/day) GCM/RCM
Percentage change in 200-year flood
EA/DC
• Differences in GCM/RCM HBV parameter
tend to be more significant
in inland and northern areas N = 115
GCM/RCM = 50
• Western Norway dominated EA/DC = 38
HBV = 27
by uncertainty from HBV
and methods for adjusting
P, TUncertainty from flood frequency
analysis relative to model uncertainty
Viksvatn - 83.2
A1B HadCM3Qref/HIRHAM (25 x 25 km)
'Empirical Adjustment' til 1 x 1 km
900
850 1961-1990 Gumbel
800
2021-2050 Gumbel
Peak daily discharge (m3/s)
750
700 2021 - 2050 annual maxima
650 1961 - 1990 annual maxima Model uncertainty
600
550
500
X
450
400
350 X
300
250 35% increase
200 in 200-year
150 flood
100
1 10 100 1000
Return period (years)Seasonal analysis - Rainfall-induced
peak flows in annual maximum series
Red – Type 1: > 67% of annual
maximum in mar-july (snowmelt
dominance)
Green – Type 2: 34 – 66%
(transitional)
Blue – Type 3: < 33%
(rainfall dominance)
1961 - 1990 2021-2050
⇒ Change towards increasing occurrence of
autumn/winter peak flows in annual seriesRelative magnitude of rainfall-
induced peak flows in annual series
Ratio of Max (Aug – Feb)
Max (all months)
1961 - 1990 2021-2050
⇒Largest peak flow in some ’snowmelt’ areas
is actually an autumn/winter rainfall floodProjected increase in probable
maximum precipitation (PMP)
Median of
three scenarios
Percentage change
in PMP between
1961-1990 and 2021-2050Further comments
•Ensemble approach is an effective method for
summarising hydrological projections and the
relative contributions of uncertainty sources
•Only ’sampled’ uncertainties are represented, and
other factors should not be disregarded
Implications for dam safety analyses in Norway:
•Flood frequency analysis indicates regions of
increase vs. reduction in low frequency events
•Change of seasonality can have implications for the
season considered in the flood calculations
•PMP is projected to increase in all regions, so that
PMF estimated with event-based modelling
can be expected to increaseYou can also read
