The use of evacuation and loss of life models to improve

flood incident management in the UK

Darren Lumbroso, Manuela Di Mauro and Andy Tagg

Page 2

Flood incident management in the UK

Flood warning and emergency

response

Flood detection

Flood forecasting and interpretation

Flood warning dissemination

The role of loss of life and evacuation modelling to support and improve flood incident management

Page 3

Issue with emergency planning for floods in UK

• Lack of communication to general public

• Identification of evacuation routes and safe havens

• Vulnerable groups (e.g. the elderly) and areas (e.g. caravan sites) are often not identified

• Limited risk assessments for dam breaches and flood defence failures in terms of loss of life and evacuation

Page 4

Supporting incident managementLoss of life and evacuation modelling

Time to evacuate

Loss of life, and how its affected by:Road closuresWarning disseminationUse of safe havensAwareness

Quantified reporting, number of:Fatalities and injuriesVehicles swept awayBuildings destroyed

Page 5

Levels of modelling - Micro

Detailed representationof evacuationroutes

Behaviour of each individualis modelled

Page 6

Levels of modelling - Meso

Simplified evacuationroutes

Mortality rates estimatedfor groups of people or zones

Page 7

Levels of modelling - Macro

Evacuation timebased on averagedistanceto safety

One mortality rate applied to the wholearea

Page 8

Typical scale of application?

Regional/Broad scale study/Lower risk

Small area/Detailed study/High risk

Micro

Meso

Macro

Page 9

Micro scale - BC Hydro Life Safety Model (LSM)

• Allows dynamic interaction between receptors (i.e. people, vehicles and buildings) and the flood wave

• Estimates loss of life from:- Drowning- Exhaustion- Building collapse- Cars being swept away

• Allows the effects of interventions (e.g. road closures, number and location of safe havens) on the evacuation time and risk to people to be assessed

Page 10

Life Safety and Evacuation Model

2D flood Modelling

Virtual world- People

- Properties- Vehicles

(Time and space distributed)

Life safety simulator- Loss of life/Injuries

- Routing of people and vehicles(i.e. physical equations and logic)

Results and visualisation-Loss of life

- Injuries- Building collapse- Evacuation time

Overview of the LSM modelling process

Page 11

BC Hydro LSM – Creation of a virtual world

Page 12

Readily available data in the UK

• Population data from census

• Number of vehicles from census

• Topographic data

• Flood depths and velocities from two dimensional hydraulic model

• Property data from a geo-reference National Property Data set

• Road network data

• Historical data (e.g. 1953 flood)

Page 13

Thames Estuary - Pilot sites

Thames Barrier

Canvey Islandpilot site

0 5 10 15 20 km

NKent

Essex

Administrative boundary

Thames Estuary

0 5 10 15 20 km

Greater London

NKent

Essex

Thames Estuary

Thames Barrier

Canvey Island

Thamesmead

Page 14

Thamesmead - Background

Page 15

Thamesmead

River Thames

Boundary of Thamesmead embayment

River Thames

Boundary of Thamesmead embayment

Page 16

Thamesmead – Modelled scenarios

• A range of important inundation scenarios modelled

• Varied number of safe havens

• Varied number of road closures

• Difference in warning rates and number of warning centres

Page 17

Thamesmead – Modelled scenarios

Page 18

Evaluation of different scenarios - Safe havens

Percentage of people that reach a safe haven analysed for different time steps and for different numbers and locations of safe havens

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6 7 8

Time (hours)

Pe

rce

nta

ge

of

pe

op

le r

each

ing

saf

ety

(%

)

4 safe havens

6 safe havens

8 safe havens

10 safe havens10 safe havens

4 safe havens

6 safe havens

8 safe havens

Page 19

Effect of road closures

Page 20

Effect of road closures on evacuation and loss of life

Una

ware

Aw

are

Aw

are

&E

vacuat

ing

Saf

e

Top

pled

Dec

ease

d

No Road Closed1 Road Closed2 Roads Closed3 Roads Closed

1

10

100

1000

10000

100000

Peo

ple

[L

og

]Consequence on People in Roads ClosingEffect of road closures on people

Increase in fatalitiesfrom 6 to 12

Increase in peopleknocked overfrom 15 to 200

Page 21

Canvey Island

River Thames

Boundary of Thamesmead embayment

River Thames

Boundary of Canvey Island

River Thames

Boundary of Thamesmead embayment

River Thames

Boundary of Canvey Island

Page 22

Canvey Island – Flood map

Page 23

Canvey Island - Background

Page 24

Maintenance of flood defences

Page 25

Results for the modelled scenarios

• Approximately 37,000 people at risk on Canvey Island

• Approximately 15,500 buildings

• For design event between 350 and 400 fatalities (approximately 1% of the exposed population)

• Up to 2,000 buildings destroyed

Page 26

Canvey Island – properties

19532001

Page 27

Canvey Island 1953 flood

Page 28

Results of 1953 modelling

• Results agree well with historical data

• Modelling indicates 55 to 155 fatalities

• Number of buildings destroyed similar to that available from anecdotal evidence

• Around 55 fatalities as the result of drowning and 0 to 100 as the result of exhaustion

Page 29

Use of the Life Safety Model • Risk assessments for dams

• To inform emergency management plans• To assess loss of life and injuries in a transparent manner from a variety of causes:

- Drowning- Exhaustion- Building collapse- Cars being swept away

• To provide another measure of risk to assist in decisions related to the construction and maintenance of new and existing flood defence assets

Page 30

Conclusions

• Provides a more scientifically robust method for estimating loss of life than other methods previously used in the UK

• Could be used to improve emergency plans for heavily defended areas (e.g. London, some coastal areas) and dam risk assessments in the UK

• In future could be adapted to incorporate risks from other hazards e.g. toxic spills, nuclear accidents

• The LSM provides an evacuation modelling and loss of life tool that is sufficiently mature to support emergency planning and incident management

Any questions?