LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

www.floodrisk.org.uk EPSRC Grant: EP/FP202511/1

Nearshore coastal hydrodynamics

Nicolas Chini & Peter StansbyThe University of Manchester

Nicolas.Chini@manchester.ac.uk

Context and Motivations

How to predict nearshore conditions from offshore operational modelling systemsWill sea level rise (SLR) and climate change (CC) modify the occurrence of extremes?How to deal with long term simulation (~100 years)?

1. Model description2. Long term methodology3. Historical hindcast and 2007 event hindcast4. Nearshore extreme values analysis5. Impact of SLR and CC on extreme values

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

From regional to coastal modelling

Localisation map Local computational grid

From regional to coastal modelling

Local computational grid:- 2877 nodes- from 2000m to 700m- bathymetry 2002 (Seazone Ltd)

Local computational grid

Downscaling : - spatially : bilinear interpolation- temporally: linear interpolation

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Coastal tide-surge modelling

Solving depth-averaged Saint-Venant eq. with TELEMAC2D (EDF R&D)

2 3 4 5 6 7 8 9 10 11-2

-1

0

1

2

3

November 2007 (days)

Wat

er le

vel (

m)

tidal gaugePOLCOMS

Water level at Lowestoft

Coastal tide-surge modelling

Solving depth-averaged Saint-Venant eq. with TELEMAC2D (EDF R&D)

Water level at Lowestoft

2 3 4 5 6 7 8 9 10 11-2

-1

0

1

2

3

November 2007 (days)

Wat

er le

vel (

m)

tidal gaugePOLCOMSTELEMAC2D

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Coastal tide-surge modelling

Solving depth-averaged Saint-Venant eq. with TELEMAC2D (EDF R&D)

Water level at Lowestoft

2 3 4 5 6 7 8 9 10 11-2

-1

0

1

2

3

November 2007 (days)

Wat

er le

vel (

m)

tidal gaugePOLCOMSTELEMAC2D

Coastal wave modelling

Including :bathymetric wave breakingshoalingbottom frictionbathymetric refractionwater depth variation due to tides and surges

Solving wave action conservation eq. with TOMAWAC (EDF R&D)

21 24 27 30 33 36 39 42 450

1

2

3

4

5

6

Day in year 1986

H m0 (m

)

measurementmodel

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Methodology:

1. Creation of the look-up table : simulations of regularly distributed offshore wave conditions and water levels (24000 runs)

2. For any offshore conditions, localisation within the look-up table:

3. Linear interpolation scheme:

Long term coastal wave modellingCreation of a look-up table of simulations for discrete offshore wave conditions (hs, tp, dir) and local water elevations.

Transfer ~100 years of deep water waves conditions towards the shoreline

0 0.5 1 1.5 2 2.50

0.5

1

1.5

2

2.5

Measured Hm0 (m)

Estim

ated

Hm

0 (m)

Significant wave height off Walcott (water depth 14m)

Historical hindcast

Profiles from Strategic surveys (EA)Offshore wave conditions from CETMEF

(1979-2002)

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

9th november 2007 nearshore conditions at Walcott

Offshore wave conditions:

Wind data: METOFFICE off Cromer �

Offshore wave conditionsHm0 max = 4.48mTm02 = 6.4sMean Direction = 150°

Water level conditions: POLCOMS (NOC) - November 2007

WAM (NOC) - November 2007

Significant wave height off Walcott (t=0h is 08/11/2007 at 00:00)

Case1 : one offshore wave conditionCase2 : spatially interpolated integrated wave parametersCase3 : spatially interpolated wave spectrum

0 10 20 30 40 500

0.5

1

1.5

2

2.5

3

Time (hours)

H m0 (m

)

case1case2case3

Long term coastal wave modelling

Scatter plot of water level and significant wave height over 140 years off Walcott without including SLR

GHG emission scenario : A1BFrom 1960 to 2100

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Long term coastal wave modellingSLR influence

SLR = 2mm/yr SLR = 3.5mm/yr SLR = 4.2mm/yr

SLR = 7mm/yr SLR = 10mm/yr SLR = 20mm/yr

Long term coastal wave modellingCC influence

Baseline

Low sensitivity unperturbed High sensitivity

One GHG emission scenario : A1B

RCM simulations with three different parameterisations

For 2 time slices of 30 years:- Baseline (1960-1990)- Future (2070-2100)

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Extreme value analyses for nearshore conditionsAnalysis requests long term temporal series and invalid for future predictionsThe framework is run from 1960 to 2100 to estimate nearshore water levels and waves heights off WalcottFit of GEV distribution to 10-largest annual maxima

Hm0 WL

linear trends test on µ

Estimation of return periods and return levels

Extreme value analyses for nearshore conditionsSLR influence on water levels : extreme values increases linearly with SLRSLR influence on nearshore waves: no linear trends although extremes are slightly increased by SLR

Variation of extreme SWH with SLR

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Extreme value analyses for nearshore conditionsSLR influence on water levels : extreme values increases linearly with SLRSLR influence on nearshore waves: no linear trends although extremes are slightly increased by SLR

Perturbed simulations of A1B scenarios induces an increase in extreme water levels and a decrease on SWH Variation of extreme SWH with SLR

100-years return values induced by CC

Extreme value joint probability

Methodology:

1. Transformation into probability

2. Selection of a level of exceedenceu =10-year return level probability

3. Estimation of Kendall parameter, τ

4. Joint probability estimated using the Gumbel copula, having a parameter :

Overtopping discharge will vary according to the joint probability of extreme water levels and waves heights

where distribution H is estimated using the Gumbel copula:

Scatter plot of wave height vs water levels used to estimate τ

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Extreme value joint probability

Scatter plot of water level vs SWH including joint probability isolines in 2100 without slr and showing the location of the nov 2007 event (red point)

Extreme value joint probability

Scatter plot of water level vs SWH including joint probability isolines in 2100 without slr and showing the location of the nov 2007 event (red point)

The same but with 1m SLR in 2100

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

With SLR, the 2007 event becomes more frequent.

with no SLR, the event RP is about one in 120 years

with a 3.5mm/year, the event RP becomes one in 5 years in 2100

with a 1cm/year, the event RP becomes less than one in 2 years in 2050

0.0 2.0 3.5 4.2 7.0 10.0 20.0100

101

102

103

SLR rate (mm/year)

Retu

rn p

erio

d (y

ear)

20502100

100

101

102

103

Retu

rn p

erio

d (y

ear)

With SLR, the 2007 event becomes more frequent.

with no SLR, the event RP is about one in 120 years

with a 3.5mm/year, the event RP becomes one in 5 years in 2100

with a 1cm/year, the event RP becomes less than one in 2 years in 2050

BaselineLowUnperturbedHighMean

With scenario A1B, the high perturbed run lead to a significant reduction in the return period from 120 years to 9 years.

The ensemble mean reduces the return period to 65 years but this return period remains in the CI of the baseline

0.0 2.0 3.5 4.2 7.0 10.0 20.0100

101

102

103

SLR rate (mm/year)

Retu

rn p

erio

d (y

ear)

20502100

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

Conclusions

- A model has been set up and validated to transfer wave parameters towards the shore including the effect of varying water depth, with a reasonable computational time.

- Regional modelling of water levels is sufficient for predicting nearshore conditions

- Assessment of climate change scenarios and SLR impacts on nearshore wave conditions

- Simulation of the nearshore conditions of 2007 event at Walcott inputs for overtopping modelling

www.floodrisk.org.uk EPSRC Grant: EP/FP202511/1

AcknowledgementThe research reported in this presentation was conducted as part of the Flood Risk Management Research Consortium with support from the:

• Engineering and Physical Sciences Research Council • Department of Environment, Food and Rural

Affairs/Environment Agency Joint Research Programme • United Kingdom Water Industry Research• Office of Public Works Dublin• Northern Ireland Rivers Agency

Data were provided by the EA and the Ordnance Survey.

LANDF RM – Coastal flood risk – from storm surge and waves to inundation

26 May 2011

Nicolas Chini – University of Manchester

−80 −60 −40 −20 0 20−60

−40

−20

0

20

40

60

Significant Wave Height

19700101000000

1

2

3

4

5

6

7

8

GLOBAL CLIMATEMODELLING

climate scenario A1B

REGIONAL CLIMATEMODELLING

offshore waves

tides surges

COASTAL CLIMATEMODELLING

SEA LEVELRISE

nearshore waves

tides surges

EXTREME EVENTS ANALYSIS

joint extreme levels

OVERTOPPING DISCHARGEMODELLING