INTRODUCTION TO CASE

STUDY A – QUANTITATIVE
RISK ANALYSIS

TABLE OF CONTENTS

Introduction 1

Case Study A: Flood Control Infrastructure in the Design Phase 1

A1. Introduction to the Quantitative Risk Analysis of Case Study A 2

A2. Summary of Available Information 2

A3. Architecture of the Proposed Risk Model 3

A4. Main Risk Model Input Data 6

A5. Risk Calculation for the Baseline Scenario in the Current Situation and 2050 Horizon 8

A6. Trend Scenario Analysis and Modified Design 9

A7. Exercise Solution Proposal (ORA) 10

Acronyms 10

INTRODUCTION
The objective of this reading is to introduce the case study that we will work with in this section. The
objective includes a quantitative risk analysis based on a proposed risk model, for the current and future
situation, as well as the assessment of the risk impact of design improvement measures.

CASE STUDY A: FLOOD CONTROL INFRASTRUCTURE IN THE

DESIGN PHASE

1
Case Study A involves the analysis of a flood control infrastructure project in the design phase that
could be susceptible to hydrological and geotechnical hazards. The city under study is located in one
of the most vulnerable areas to natural hazards in Latin America and the Caribbean. This is an area with
the highest natural disaster risk in the country.

A1. Introduction to the Quantitative Risk Analysis of Case Study A

The flood control infrastructure works designed for the city basically consist of a series of changes to
the channel of the river that runs through the city, in order to favor the formation of more convenient
hydraulic profiles than the current ones, from the standpoint of potential floods.

A2. Summary of Available Information

Based on the results of the qualitative risk analysis phase, we have selected a failure mode to be
analyzed in the quantitative phase (to simplify this for educational purposes, we have proposed a risk
model with a single failure mode).

The selected failure mode corresponds to the overtopping of the projected protection levee
failure mode. Figure 1 shows the description and summary resulting from the failure mode identification
phase.

Title Overtopping of the Protection Levee

Description

In the hydrological scenario, a high level of water is produced in the river, spilling over the crest of one of the projected protection
levees and overtopping the levee. In addition, the slope of the protection levee is damaged by lack of maintenance, so the flow of
water causes rapid erosion of the body of the levee, increasing the flow of discharge into urban areas and the consequences of this
flood in urban areas.

Diagram

2 1
3

1 Levee deterioration 2 Overtopping over 3 Erosion of the body Levee failure due to
4
due to lack of the body of the of the levee erosion of its body
maintenance levee

2
 Factors That Increase Risk Factors That Decrease Risk

There is a problem with the development of informal No factors have been identified in this regard.
settlements in areas that are close to the river, which are
also built with low-quality materials.

The sense of security that these levees may provide can

increase urban settlements around the river, as well as the
consequences of their failure.

There are doubts about the responsibilities and budget

for the maintenance and dredging of these infrastructures.

Figure 1. Failure Mode Due to the Overtopping of the Protection Levee.

A3. Architecture of the Proposed Risk Model

In order to analyze both the current situation before carrying out projected protection works and their impact on
risk, we need two architectures for the quantitative risk analysis.

SEASON

DAY\NIGHT Q RIVER FATALITIES DAMAGES

FLOODS

Figure 2. Architecture of the Proposed CaseStudyA_Baseline Risk Model.

The model has a total of 6 nodes that we will describe below:

Season Node: This node enables the model to allow for the differences that
would occur if the flood happened in summer or in winter because, depending on
the season, the duration of day and night considered may be different (e.g., due
to differences in the length of the work day between summer and winter months).
In addition, the population could vary considerably in either case (e.g., if the city
were a holiday destination and its population were to increase in a given season).

Day\Night Node: This node includes the probabilities of the flood occurring
during the day or at night, considering these two options when estimating
consequences for the population and taking into account that the population may
change during the day due to daily variations (e.g., labor reasons, cities that are
“sleeping quarters” for residents who normally work outside the area, etc.).

3
 Floods Node: This node enables the discretization of the entire range of return
periods contemplated for floods in the channel, covering a range of 1 to 200 years
for the return period.

River Flow (Q River) Node: This node incorporates the information

corresponding to the floods that may occur, including their probabilities of
occurrence and the peak discharge of the river hydrograph.

Fatalities Node: This node includes estimation results regarding the

consequences to the population that may stem from the flood.

Damages Node: This node includes estimation results regarding the economic
consequences that may stem from the flood.

Figure 3 shows the CaseStudyA_Works risk model proposed for the quantitative analysis after
considering the existence of the protection works and the potential failure of the planned levee.

SEASON

FATALITIESF DAMAGESF
DAY\NIGHT Q RIVER

FAILURE

FATALITIESNF DAMAGESNF
FLOODS WL RIVER

Figure 3. Architecture of the Proposed CaseStudyA_Works Risk Model.

The model has a total of 8 nodes that we will describe below:

Season Node: Similar to the above description.

Day\Night Node: Similar to the above description.

Floods Node: Similar to the above description.

River Flow (Q River) Node: Similar to the above description.

River Water Level (WL River) Node: Based on the discharges due to any
floods that may occur, this node incorporates the level that the river channel
would reach associated with each of these discharges.

Failure Node: This node incorporates the probabilities of levee breakage

due to overtopping based on the levels reached by water in the channel.

4
 FatalitiesF Node: This node includes estimation results regarding the
consequences to the population in the event of a levee failure due to
overtopping, stemming from the resulting flood.

FatalitiesNF Node: This node includes estimation results regarding the

consequences to the population in the event of no failure of the levee, but
which may stem from the flood resulting from water spilling over the levee’s
crest elevation, thus flooding areas near the channel.

DamagesF Node: This node includes estimation results regarding economic

consequences in the event of a levee failure due to overtopping, stemming from
the resulting flood and the costs of reconstructing the levee.

DamagesNF Node: This node includes estimation results regarding economic

consequences in the event of no failure of the levee, which may stem from the
flood resulting from water spilling over the levee’s crest elevation, thus flooding
nearby areas.

For the above two nodes of economic consequences (in monetary units) and social consequences (in
potential fatalities), both for the baseline scenario and for the different scenarios to be analyzed, input
values are the product of the following processes:

Economic Consequences: We obtain these by overlapping the flood spot

obtained by hydraulic modeling for each return period analyzed, on the
different types of land use that exist in the study area. For each land use we
define a water depth-damage curve, in which an economic loss value is assigned
according to the maximum recorded water depth. Thus, by multiplying the
affected area of a given use by the maximum water depth above it, we can
obtain the damage recorded in that area. The sum of all affected areas (and
the replacement of the infrastructure or infrastructures that failed, for example:
levee reconstruction) provides the overall value of the estimated economic
damage due to flooding for each return period.

Social Consequences: The procedure for obtaining these values is broadly

similar to the one we described above. Thus, the basis for the calculation is
to obtain, on the one hand, the population potentially affected by the flood
spot overlap corresponding to each return period, and on the other hand, the
influence of a Reference Fatality Rate (RFR) obtained based on the “SUFRI
Methodology for pluvial and river flooding risk assessment in urban areas to
inform decision-making (2011).” The fatality rate will be a function of both
the characteristics of the flood and the characteristics of the degree of flood
risk governance that there is in the area. Therefore, the potentially affected
population multiplied by the fatality rate will result in the estimated value of
potential flood fatalities for each return period.

5
A4. Main Risk Model Input Data

In the following tables we will describe the information incorporated into the model at each node. This
input data is introduced using Excel files that can be downloaded from the course platform.

NODES COMMON TO BOTH MODELS

Table 1. Input Data for the Season Node.

Season Probability Description

Summer 0.25 Three months of the year correspond to the conditions considered for summer.

Winter 0.75 Nine months of the year correspond to the conditions considered for winter.

Table 2. Input Data for the Day\Night Node.

Season Time Probability Description

In summer, 60% of the day corresponds to time in which the

Day 0.6 population is considered a daytime resident (population
reduced by commuting).
Summer

In summer, 40% of the day corresponds to time in which the

Night 0.4 population is considered a night-time resident (census population).

In winter, 50% of the day corresponds to time in which the

Day 0.5 population is considered a daytime resident (population
reduced by commuting).
Winter

Night In winter, 50% of the day corresponds to time in which the

0.5 population is considered a night-time resident (census population).

Table 3. Input Data for the Floods Node.

MinT MaxT Description

Range of floods to be included in the analysis, represented by their

1 200 associated return period.

Table 4. Input Data for the River Flow (Q River) Node.

6
 Season Time Probability Description

1 5 7

It includes the results of the channel’s hydraulic

5 20 22
modeling and the hydrograph’s maximum dis-
charge in a section defined as a reference. This dis-
10 45 55 charge is used to link each flood episode with the
results of the estimation of consequences stem-
25 70 78 ming from the flood.

Results are included for both the current and the

50 105 125 trend scenarios (year 2050 with climate change im-
pact).
100 150 175

200 180 208

CASESTUDYA_BASELINE RISK MODEL

Table 5. Input Data for the Fatalities and Damages Node.

Maximum Fatalities [Number Economic Damage Description

Estación (m³/s)
Discharge Momento
of People] (Day/ Probabilidad Descripción
[Millions of US$]
Night)

5 0/0 0

20 5/2 0.55

45 10 / 5 1.2
It includes summary results of the estimation
70 1.8 of consequences for the population, in number
20 / 8
of potential fatalities and economic costs due
to flooding.
105 50 / 20 2.6

150 120 / 30 3.6

180 180 / 60 5.6

210 195 / 66 6.1

CASESTUDYA_WORKS RISK MODEL

Table 6. Input Data for the River Water Level (WL River) Node.

Flood Channel Level Channel Level Description

(T in Years) (m.a.s.l.) (m.a.s.l.)
2020 SCENARIO 2050 SCENARIO

1 138 138.2

5 139.2 139.3

10 140 140.2 It includes the results of the channel’s

hydraulic modeling and the maximum
levels reached in the section
25 140.6 140.8 corresponding to the protection
levee, whose crest elevation is at
50 141.2 141.4 141.5 m.a.s.l.

100 141.5 142.0

200 142.2 142.5

7
Table 7. Input Data for the Failure Node.

Channel Level Probability Description

(m.a.s.l.)

141.5 0

141.6 0.2 It includes probability values based on the fragility curve considered for an
embankment levee, in which failure probability is estimated at 100% for
overtopping heights equal to or greater than 0.5 m above the levee.
141.7 0.5

142 1

Table 8. Input Data for the FatalitiesF and FatalitiesNF Node.

Fatalities Fatalities
Channel Level (Failure) (No Failure) Description
(m.a.s.l.) [Number of [Number of
People] (Day/ People] (Day/
Night) Night)

141.5 0/0 0/0

It includes summary results of the estimation
141.6 100 / 80 0/0 of consequences for the population, in number
of potential fatalities, in the event of a flood by
141.7 120 / 100 0/0 levee breakage (failure) or by overtopping of
the levee without failure (no failure).
142 175 / 160 5/2

142.5 240 / 200 10 / 5

Table 9. Input Data for the DamagesF and DamagesNF Node.

Channel Level Economic Economic Description

(m.a.s.l.) Damage (failure) Damage
[Millions of US$] (No Failure)
[Millions of US$]

141.5 15 0
It includes summary results of the estimation of
economic consequences, in millions of US$, in
the event of a flood by levee breakage (failure),
141.6 18 0.25 including US$4 million in reconstruction costs
and the damage resulting from flooding due to
the breach in the levee. It also includes the costs
141.7 20 0.6 of flooding by overtopping without failure (no
failure), which are much lower because there is
no breach in the levee.

142 28 1.2

A5. RISK CALCULATION FOR THE BASELINE SCENARIO IN THE CURRENT

SITUATION AND 2050 HORIZON

Once all the information has been incorporated, we execute the baseline model to estimate the risk in the current
(2020) scenario, the future (2050) scenario, and the scenario after implementing the planned works (2022). Results
obtained are shown in Table 10.
8
Table 10. Risk Results for the Current Scenario (2020), Future Scenario (2050), and the Scenario with
Planned Works (2022).

Model Used Scenario Probability of Social Risk Economic Risk

Failure (1/Year) (Fatalities/Year) (Millions of US$/
Year)

CaseStudyA_Baseline Current (2020) Not applicable 4.82 0.523

CaseStudyA_Baseline Future (2050) Not applicable 6.10 0.632

CaseStudyA_Works Design (2022) 8 x 10 ³ 1.43 0.236

Results can be represented with F-N or F-D curves, in which F is the cumulative annual exceedance probability
of the level of consequences corresponding to loss of life (N) or to economic costs (D). As an example,
the curves below include the current scenario and the scenario after contemplating the projected works.
They show that the works reduce frequent flooding events (most likely, moving the curve downwards),
but there may be events associated with failure that shift the curve to the right (low probability – higher
consequences). However, the total area under the curve is smaller in the scenario with protective works
(social and economic risk is lower than the current one).

___ ___
___
CASE STUDY A: 2022 WORKS

___
CASE STUDY A: 2022 WORKS

CASE STUDY A: 2020 BASELINE CASE STUDY A: 2020 BASELINE

100 100
ANNUAL ECONOMIC DAMAGE PROBABILITY ≥ D, F
_

10-1 10-1
ANNUAL LOSS-OF-LIFE PROBABILITY ≥ N, F_

10-2 10-2

10-3 10-3
10-4 10-4

10-5 10-5

10-6 10-6

10-7 10-7

10-8 10-8

10-9 10-9
10-10
100 101 102 103 104 100 101 102 103 104 105
LOSS OF LIFE, N ECONOMIC DAMAGE (EC_UNITS), D

Figure 4. F-N and F-D Curves for the Current Scenario vs. the Scenario for Designed Protection Works.

A6. Trend Scenario Analysis and Modified Design

For this case study, we will consider three additional scenarios for performing calculations. Participants can
obtain them from the case model with projected works (CaseStudyA_Works, Figure 3):

Trend Scenario (Works + 2050 Horizon): To assess the impact on risk

of planned changes in expected floods due to climate change. It involves
changes to the WL River node and an increase in expected levels.

9
 Scenario with Design Improvement Measures (Modified Design): To
assess the impact on risk of design changes with which to reduce the risk
of overtopping the levee (the improvement consists of changing the crest
elevation and taking it to 141.75 m.a.s.l.) and a reinforcement on the levee’s
berm to reduce the erosion of the body of the levee in the event of overtopping.
It involves changes to the Failure node and the Fatalities and Damages nodes.

Scenario with Nonstructural Measures (Warning Improvements):

To assess the impact on risk caused by improving warning systems for the
population, incorporating in the model the results of the estimation of
consequences regarding potential loss of life for this scenario with nonstructural
measures. It involves changes to the Fatalities nodes.

A7. Exercise Solution Proposal (ORA)

Given the architecture of the proposed risk model, the results obtained for the current situation, and
the scenarios presented in the previous section, please answer the following questions:

1) If we change the selection of input data for the WL River node (“Works and 2050 Scenario”
sheet), what would be the resulting risk for the 2050 scenario if floods for the trend scenario are
taken into account? How do the F-N and F-D curves change when compared with the current
scenario? Would the planned design be appropriate considering future risk?

2) If we change the selection of input data for the Failure, FatalitiesF, FatalitiesNF, DamagesF,
and DamagesNF nodes (“Improved Design” sheets), what would be the risk resulting from
implementing the design improvement measures? How do the F-N and F-D curves change
when compared with the current scenario? Would design improvements be justified in terms
of risk reduction?

3) If we change the selection of input data for the FatalitiesF and FatalitiesNF nodes (“Warning
Improvements” sheets), what would be the risk resulting from implementing nonstructural
measures to improve the warning systems for the population?

ACRONYMS

F-D FREQUENCY – DAMAGE

FM FAILURE MODE

F-N FREQUENCY – LOSS OF LIFE

M.A.S.L. METERS ABOVE SEA LEVEL

ORA OPEN RESPONSE ASSESSMENT TOOL

RFR REFERENCE FATALITY RATE

10