Podlaha Wed 130pm Ven II

8/12/2019 Podlaha Wed 130pm Ven II

1/33

European Floods

Cat Modeling 2014, February 12, OrlandoAdam Podlaha


2/33

1Proprietary & Confidential

Agenda

Section 1 Introduction

Section 2 Data availabilitySection 3 Flood modelling

Section 4 Selected chapters of flood modelling in Europe

Section 5 Close


3/33

Flood losses in Europe

Flood insurance coverage in Europe

Section 1: Introduction


4/33


Flood losses are second highest after windstorm

Source: http://catastropheinsight.aonbenfield.com
http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/


5/33


Flood losses are regularly present in Europe

Source:http://catastropheinsight.

aonbenfield.com

And there is no reason

why this shouldnt

continue

Often cross country

events

Catastrophe modelling is

needed
http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/http://catastropheinsight.aonbenfield.com/


6/33


Different flavours of flood coverage in Europe

Different flavours and obscurities present due to Europes fragmentation

Country DescriptionCzech Rep.

(+ SK, PL)

All covered, limits for Ind & Com, no limits for Res, commercial rates for

primary, frequent recent experience with flooding

Austria All covered, very small (few 1,000s )for Res, almost no top-up

available, doesnt affect the primary rate as much, rating system

available

Netherlands No Res cover available, very limited Com & Ind cover, cars covered,

talks about pool in 2013, not happening as of now

Switzerland 2 insurance pools, mandatory cover in some, in some no commercial

rate for primary,no reasons for rating systems

Germany Low insurance penetration in the former West. Germany, higher in

former East. Germany, primary rating system available

UK High penetration in all occupancies, commercial rates, good rating

systems, talks about a pool to happen soon


7/33

Portfolio data

Model development data

Section 2: Data availability


8/33


Better portfolio data in places with recent flooding

..., flood coverage available & functional rating systems (go hand in hand)

Geo-coding

At least postal code available, longitude & latitude data varies from all to none

Primary and secondary modifiers

Limited number of modifiers can be really used for modelling either due tounavailability or the model not being able to use those

Common issues

Multi-locations for industrial occupancy

Wrongly assigned longitude and latitude Inconsistency in what belongs to which occupancy

Portfolio data are improving with events happening, also due to rating needs


9/33


Detailed model develop. data, not always cheaper

Elevation data and other topographic data

Country specific data: from topographical maps or remote sensing (on increase) Continental data: remote sensing based, need to be careful about the specs

Many topographic data still from country specific providers

Hydrological data

The availability is generally very good

The cost varies widely from free to very costly Geo-coding data

The usual: Google, ESRI, Bing, PBS for address geo-coding

Same + others for postal code data, local providers as well

Data for loss estimation from just happening events Remote sensing data (disaster charters, PERILS etc.)

Conclusions: more and better data are available, cost varies widely


10/33

Flooding types

Model types

Availability Two basic approaches

Section 3: Flood modelling


11/33


Flooding comes in many different flavours

Fluvial (riverine) flooding flood plain and off-

flood plain

Storm surge

Tsunami

Pluvial flooding / cloudbursts / rainfall

Flood plain extent (shades of blue) with

off-flood plain buffers (red)

Conclusion: different flood types need different modelling approaches


12/33


Scenario models the 1ststep in risk quantification

1. Scenario models

Historic events Events linked to particular return

period

Hypothetical events

Dam breaks,

2. Fully probabilistic models

Stochastic event set needs to be

generated (assumptions)

Other components same

0.00%

0.01%

0.02%

0.03%

0.04%

0.05%

0.06%

0.07%

0.08%

1 10 100

Loss[%o

fTSI]

Return period [years]

6yr 7yr 56yr

0.00%

0.01%0.02%

0.03%

0.04%

0.05%

0.06%

0.07%

0.08%

0.09%

0.10%

A B C D E

Los

ses[%o

fTSI]

CompanyModelled Observed

Conclusion: Ideally modelling

solution includes both probabilistic

and scenario models


13/33


Brokers, RIs, Insurers & others ahead of vendors

Classic vendors offer limited model coverage (UK, DE, BE, AT)

Brokers, Reinsurers, Insurers and other commercial companies farahead

Why?

Harder to do than other perils (higher and often local based data requirements,

tricky vulnerability comp, time consuming 2D modelling) It all adds to more expensive in both cost and time

New models face complex model acceptance (due to existing models, S2)

Need to involve local institutions due to handling complex issues

The European flood modelling scene is not led by classic model vendors


14/33


Climate model vs. river gauge based model

1. Global climate model based

2. Measured discharge based

Global Climate Model (rainfall)

Rainfall Run-off model (modelleddischarge)

Hydrodynamic modelling (flood extent withinundation depth / velocity / duration)

Measured discharge

Hydrodynamic modelling (flood extent

with inundation depth / velocity / duration)

Conclusion: In fact both approaches are part of one process, the difference is

the starting point and what is modelled / measured


15/33


Have a guess which approach is better suited to:

... capture global patterns in flooding?

... reconstruct historical events well?

... take into account existing hydrological regulations?

...take into account real climate trends?

... take advantage of the best measured data?

... represent best the nature of flooding in a given country?

Conclusion: each approach has its pros and cons, combination is possible

GCM

Flow

Flow

GCM / Flow

Flow

GCM / Flow


16/33

Hazard uncertainty in the Dutch flood model

Cloudburst modelling

How to test and validate a flood model

Vulnerability curves development

Location uncertainty quantification

Section 4: Selected chapters of flood

modelling in Europe


17/33


Hazard uncertainty in the Dutch flood model

How many are there? ...

Many

To mention just few:

1. Safety standard of dyke

rings (affecting the

frequency of dyke ring

being flooded)

2. Correlation of dyke rings(affecting the scale of

events)

3. Maximum flood extent

4. Evacuation (Hazard?)

All based on experts

opinion

Conclusion: various uncertainties can be quantified/ implemented by creating various event sets

where the particular uncertainty varies

Event sets

Probability of flooding

Optimistic Realistic Pessimistic

Correlatio

nmatrix Low (isolated

events)Event set 1 Event set 4 Event set 7

Best estimate Event set 2 Event set 5 Event set 8

High (large

scale events)Event set 3 Event set 6 Event set 9


18/33


Cloudburst modelling

Extreme rainfall events, very localized

Pluvial flooding often (not-correctly) mixed with flash-floods Processing extreme rainfall intensities in a very short time steps (~5min)

Two types of assessment:

Rainfall intensity based

Flood depth based


19/33


Testing and validation of a flood model

Level 3 Guidelines on External Model and Data

The undertaking should perform their own validation

of the material assumptions of the external model that

are relevant to their risk profile and of the process for

incorporating the external model and data into their

own processes and internal model.

In addition to their own validation, undertakings

may also leverage on th e validat ion do ne by the

vendors o rother third party...

1. Sensitivity testing (input change)

2. Back testing(historical events)

3. Stability testing (effect of randomness)

4. Stress testing (model change)


20/33


Sensitivity testingaggregation (Thai flood)

Analysis resolution

Postal code

CRESTA (canton)

Analysed aggregated to the

analysis level, means:

More rows in the postal code run

Less rows in the CRESTA run

Two effects

1. Change of spread

2. Change of mean / median

Which causes what?


21/33


29.8%

26.8%

29.8%

8.3%

1.0%

3.6%

0.1%0.7%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

A B C D E F G H

Shareofmarketloss

Back testingas if 2013 event (Austria flood)

As if 2013 data analysis

Modelled vs. real loss for the 2013 flood for the market and 8 portfolios

Background: To test the model results against values from observed historical events

from hazard and loss point of view

34%

16%

8%

-42%

7%

34%

80%

-15%

14%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

A B C D E F G H marketDifference(%o

frealloss)


22/33


Stability testingNumber of samples (Swiss flood)

No. of samples can be changed in ELEMENTS interface

Depends on the portfolio size

Background: To test the stability of the outputs with different number of samples

-5.0%

-4.0%

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

100005000200010005002502001501005025201052

Differencefrom2

00D

S[%]

Return period [years]

10 DS

25 DS

50 DS

100 DS

200 DS


23/33


24/33


0

50

100

150

200

250

300

350

400

450

0%

5%

10%

15%

20%

25%

30%

35%

0 200 400 600

Lo

ssdatacount

Lossratio

Flood depth (cm)

Loss data Engineering curve Loss data counts

Vulnerability curves development

Two main approaches

1. Claims data

2. Engineering

Assumptions

Claims data present realistic view on risk

Sometimes no sufficient amount of dataLack of construction types differentiation

Possible solution

Use claim information as a basis

Use shape of the engineering curve

Use relations between engineering

curves for different construction types


25/33


Location uncertainty quantification

2 insured locations

Sum Insured 34 mil CZK

Limit 8.5 mil

0

5,000,000

10,000,000

15,000,000

20,000,000

25,000,000

30,000,000

35,000,000

FGU SIR GR OTHP RREC NTY NFAC

Perspective Box Plot Statistics for Event 123002013

with 5000 samplesThere are 328 zero samples

0%

2%

4%

6%

8%

10%

Histogram for Event 123002013 forperspective FGU

with 5000 samples

0%

10%

20%

30%

40%

50%

60%

70%

80%

Histogram for Event 123002013 forperspective GR

with 5000 samples

Section 5: Close


26/33

Summary

About Impact Forecasting

Questions Contact

Section 5: Close


27/33


Summary

Flooding is the second major peril in Europe but first in some countries

Classic 3 vendors offer limited flood model coverage however many models

available from alternative vendors

Different types of flooding, different data available require innovative

modelling approaches

The ability to quantify (and implement) various uncertainties is the key to

understanding the risk

Better synchronisation needed between primary underwriting systems andcatastrophe modelling for reinsurance purposes


28/33


About Impact Forecasting

Impact Forecasting Team

Catastrophe model developers

Independent, transparent, open,

modularand bespokemodels

Natural and man-made perils

Filling gaps and main perils

Global team (60+)

ELEMENTS platform

Runs our models

Runs 3rdparty models

Visualisation: uncertainty & mapping

20+ programmers in last 4 years

Distributed to re/insurers and AB

Integrated with ReMetrica, IWB Lite


29/33



30/33

Q


31/33


Questions

C t t


32/33


Contact

Adam Podlaha

55 Bishopsgate, EC2N 3BD

London, United Kingdom

+44 (0)20 7522 [email protected]

Impact Forecasting

www.impactforecasting.com

Catatrophe InsightJoin our LinkedIn group

Di l i
mailto:[email protected]://www.impactforecasting.com/http://catastropheinsight.aonbenfield.com/http://www.linkedin.com/groups/Impact-Forecasting-6586802http://www.linkedin.com/groups/Impact-Forecasting-6586802http://catastropheinsight.aonbenfield.com/http://www.impactforecasting.com/http://www.impactforecasting.com/http://www.impactforecasting.com/mailto:[email protected]


33/33


Disclaimer

Legal Disclaimer

Aon Limited trading as Aon Benfield (for itself and on behalf of each subsidiary company of Aon Corporation) (Aon Benfield) reserves all rights to the content of this report (Report). This Report

is for distribution to Aon Benfield and the organisation to which it was originally delivered only. Copies may be made by that organisation for its own internal purposes but this Report may not be

distributed in whole or in part to any third party without both (i) the prior written consent of Aon Benfield. and (ii) the third party having first signed a recipient of report letter in a form acceptable to

Aon Benfield. Aon Benfield cannot accept any liability to any third party to whom this Report is disclosed, whether disclosed in compliance with the preceding sentence of otherwise.

To the extent this Report expresses any recommendation or assessment on any aspect of risk, the recipient acknowledges that a ny such recommendation or assessment is an expression of AonBenfields opinion only, and is not a statement of fact. Any decision to rely on any such recommendation or assessment of ris k is entirely the responsibility of the recipient. Aon Benfield will not in

any event be responsible for any losses that may be incurred by any party as a result of any reliance placed on any such opinion. The recipient acknowledges that this Report does not replace the

need for the recipient to undertake its own assessment.

The recipient acknowledges that in preparing this Report Aon Benfield may have based analysis on data provided by the recipient and/or from third party sources. This data may have been

subjected to mathematical and/or empirical analysis and modelling. Aon Benfield has not verified, and accepts no responsibility for, the accuracy or completeness of any such data. In addition, the

recipient acknowledges that any form of mathematical and/or empirical analysis and modelling (including that used in the preparation of this Report) may produce results which differ from actual

events or losses.

The Aon Benfield analysis has been undertaken from the perspective of a reinsurance broker. Consequently this Report does not constitute an opinion of reserving levels or accounting treatment.

This Report does not constitute any form of legal, accounting, taxation, regulatory or actuarial advice.

Limitations of Catastrophe Models

This report includes information that is output from catastrophe models of Impact Forecasting, LLC (IF). The information from the models is provided by Aon Benfield Services, Inc. (Aon Benfield)

under the terms of its license agreements with IF. The results in this report from IF are the products of the exposures model led, the financial assumptions made concerning deductibles and limits,and the risk models that project the pounds of damage that may be caused by defined catastrophe perils. Aon Benfield recommends that the results from these models in this report not be relied

upon in isolation when making decisions that may affect the underwriting appetite, rate adequacy or solvency of the company. The IF models are based on scientific data, mathematical and

empirical models, and the experience of engineering, geological and meteorological experts. Calibration of the models using actual loss experience is based on very sparse data, and material

inaccuracies in these models are possible. The loss probabilities generated by the models are not predictive of future hurricanes, other windstorms, or earthquakes or other natural catastrophes,

but provide estimates of the magnitude of losses that may occur in the event of such natural catastrophes. Aon Benfield makes no warranty about the accuracy of the IF models and has made no

attempt to independently verify them. Aon Benfield will not be liable for any special, indirect or consequential damages, including, without limitation, losses or damages arising from or related to any

use of or decisions based upon data developed using the models of IF.

Additional Limitations of Impact Forecasting, LLC

The results listed in this report are based on engineering / scientific analysis and data, information provided by the client, and mathematical and empirical models. The accuracy of the results

depends on the uncertainty associated with each of these areas. In particular, as with any model, actual losses may differ from the results of simulations. It is only poss ible to provide plausible

results based on complete and accurate information provided by the client and other reputable data sources. Furthermore, this information may only be used for the business application specified

by Impact Forecasting, LLC and for no other purpose. It may not be used to support development of or calibration of a product or service offering that competes with Impact Forecasting, LLC. The

information in this report may not be used as a part of or as a source for any insurance rate filing documentation.

THIS INFORMATION IS PROVIDED AS IS AND IMPACT FORECASTING, LLC HAS NOT MADE AND DOES NOT MAKE ANY WARRANTY OF ANY KIND WHATSOEVER, EXPRESS OR

IMPLIED, WITH RESPECT TO THIS REPORT; AND ALL WARRANTIES INCLUDING WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE HEREBY

DISCLAIMED BY IMPACT FORECASTING, LLC. IMPACT FORECASTING, LLC WILL NOT BE LIABLE TO ANYONE WITH RESPECT TO ANY DAMAGES, LOSS OR CLAIM WHATSOEVER,

NO MATTER HOW OCCASIONED, IN CONNECTION WITH THE PREPARATION OR USE OF THIS REPORT.

Date post:	03-Jun-2018
Category:	Documents
Upload:	ndimuzio
View:	224 times
Download:	0 times

Podlaha Wed 130pm Ven II

Documents