You are on page 1of 69

Probabilistic Seismic Risk

Analysis and Damage


Scenarios
for Civil Protection Purpose
Needs, implementations and
critical issues

Prof. Mauro Dolce


Italian Department of Civil Protection
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
THE ITALIAN
NATIONAL SERVICE OF CIVIL PROTECTION
(Law n. 225 / 1992)
By “Civil Protection” it is meant
The ensemble of the activities put in place to protect
life, goods, settlements and environments
from damage and risk of damage due to calamities

In Italy «Civil Protection»


IS NOT a task assigned to a SINGLE ADMINISTRATION
BUT a function played by a COMPLEX SYSTEM

“NATIONAL SERVICE OF CIVIL PROTECTION”


(SNPC)
Established by the Law n. 225 of 1992
and coordinated by the (National) Department of Civil Protection
of the Prime Minister Office
National Fire-fighters Corps
Interior Police
THE NATIONAL SERVICE Foreign Affairs
Prefectures
www.protezionecivile.gov.it
OF CIVIL PROTECTION
Environment
ISPRA

Health 118

Economy and Finance Revenue Guard Corps

Defence Army
Navy
Economic Development Air Force
Carabinieri
PRESIDENCY Infrastructures
Department TERNA
OF THE
of Civil Transportation Costal Guard
COUNCIL OF ANAS
Protection National Highway
MINISTERS Cultural Heritage
National Railway
University and Research
INGV
CNR
Major Risks Public Education
National Research
Commission Communications
Institutes

Agricultural Policy and Forestry State Forest Corps

Regions
volunteers
Provinces

Municipalities
Mandate www.protezionecivile.gov.it

The National Service of Civil Protection of Italy (Law


225/1992) aims at safeguarding human life and health,
goods, national heritage, human settlements and the
environment from all natural or man-made disasters.

It deals with:
Forecasting and Warning
Prevention and Mitigation
Rescue and Assistance
Emergency overcoming
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
Needs and use of PSRA and SRS
www.protezionecivile.gov.it
for CP purpose
Needs and use of PSRA for CP purpose:
• Prevention policy set up: comparison of losses expected from
different hazard risks for which probabilistic risk analyses are
available
• Prevention strategy set up: comparison of losses expected
from the same hazard risk in different areas and /or for different
elements at rtisk
• Contingency planning: at national and sub-national level

Needs and use of SRS for CP purpose:


• Emergency management: first damage estimates and losses
• Contingency planning: at national level and subnational level
• Contingency planning: at local level
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
4

Using PSRA for the


allocation of seismic prevention funds
Law n. 77 24.06.2009
Article 11: Interventions for seismic risk prevention
Funds allocated for seismic prevention:
• M€ 44 for the year 2010
• M€ 145,1 for the year 2011
• M€ 195,6 for each of the years 2012, 2013, 2014
• M€ 145,1 for the year 2015
• M€ 44 for the year2016

Annual funds (M€)

200

Total 965 M€ 150

in 7 years 100

50

0
2010 2011 2012 2013 2014 2015 2016 9
4
ACTIONS

2012-
ACTION 2010 2011 2015
13-14
a) Seismic microzonation and CLE
4 M€ 8 M€ 16 M€ 16 M€
studies
b) Seismic retrofit or reconstruction of
public buildings and bridges of
strategic interest or critical for the 130 170 124
34 M€
M€ M€ M€
consequence of their collapse.
c) Interventions of seismic upgrading or
reconstruction of private buildings.
d) Other urgent interventions. 4 M€ 4 M€ 8,5 M€ 3,8 M€

Funds are distributed among the 20 Italian Regions


proportionally with their seismic risk.
4
Distribution of funds
among the Regions according to PSRA

HAZARD % People involved in building


collapses in 100 years

EXPOSURE

RISK

DISTRIBUTION OF
11
FUNDS
4

Distribution of funds
among the Regions according to PSRA
Based on the “average” of seismic risk studies carried out by
ReLUIS, by EUCENTRE and by DPC, characterized by:

• Same hazard as given by DPC-INGV S1 (2004-2006) Project,


• Same exposure data (people and buildings) provided by ISTAT
2001 census,
• Different vulnerability models:
1. DPC: empirical damage probability distribution (Goretti
et al., 2008).
2. EUCENTRE: fragility curves drawn from mechanical
models of sample buildings designed according to
standards and uses of buiding age (Borzi et al. 2011).
3. ReLUIS : empirical damage probability distribution
(Zuccaro et al., 2004) recalibrated at regional level. 12
4

Choice of the risk index

Two risk indices related to human losses were considered:


IRG provides a measure of the total risk of loss of life in a Region
IRS provides a measure of the risk of loss of life of each individual

IRG IRG NORMALISED ΣIRG =1

IWRR = IRG + IRS


IRS IRS NORMALISED ΣIRS =1

 = 0.5  = 0.5 Year 2010


 = 0.77  = 0.23 Year 2011, 2012, 2013, 2014
Choice of the risk index www.protezionecivile.gov.it

IRG IRS IWRR


PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
Seismic scenarios for civil protection www.protezionecivile.gov.it

Seismic risk scenarios are used by the Civil Protection


Department with different goals:
• In pre-event situation, supporting prevention activities, they
are employed for enhancing emergency planning, supporting
Major Risks National Committee evaluations, as well as
supporting drills, technical training, information to population
and so on.
• In the immediate post-event, they are important tools for an
early and prompt estimate of the seriousness of the event,
before specific information are collected from affected
territories.
24 August 2016, h. 4.00 a.m.
Operational Committee meeting (event at 3.36 a.m.)
National coordination for the first emergency response
The Operational Committee met permanently until 28 August 2016
Seismic scenarios for emergency www.protezionecivile.gov.it

emergencyotection
SIGE - Information System for Emergency
Management and simulated scenarios
INGV seism. network
In case of an earthquake, of Magnitude 4+, an
automatic procedure is immediately activated by
SIGE DPC using SIGE to produce data, maps, and
information concerning:
DPC
– Description of the area (anthropic, physical and
administrative aspects; characteristics of
buildings and infrastructures; monitoring
networks)
– Exposure (building stock, schools, hospitals)
Maps and reports
– Hazard (seismogenic zones, catalogue,
isoseismals, attenuation)
Emergency
Management – Preliminary evaluation of damage and losses
Seismic scenarios for emergency www.protezionecivile.gov.it

SIGE input data (provided by INGV)


Event characteristics: Few minutes after the event, a
Magnitude (Mw or Ml), epicenter simulation of the impact on
coordinates, hypocenter depth constructions is run by SIGE, before
SIGE outputs detailed information are collected
Expected structural damage: by territories involved in the event.
collapsed and unusable
dwellings
Expected number of
casualties: fatalities and
homeless
Gis maps
SIGE: DPC tool for damage scenarios www.protezionecivile.gov.it

SIGE is based on an empirical approach, using a magnitude-macroseismic


intensity conversion, an intensity attenuation relationship, and Damage
Probability Matrices providing the probability of a damage level given the
intensity and the building vulnerability class.
Hazard
Source type Point source
Magnitude ML (or Mw)
Conversion ML-Mw Gasperini et al.2013
Epicenter intensity Io
Constant for focal depth <20 km (Io=2.1241*Mw-4.106 (CPT11)) MCS
(as a function of Mw)
Two different attenuations depending on Magnitude and focal depth:
GMPE M≤5.5 Io≤VIII) = Gomez Caprera (2007)
M>5.5 (Io>VIII) =Pasolini et al (2008)
Focal Depth Processed for h>20 km
Local amplification Not processed
Shaking parameters Macroseismic intensity MCS
Minimal unit of process Municipal
Uncertainty Processed (half degree of macroseismic intensity in each municipality)
www.protezionecivile.gov.it

Exposure model
Database ISTAT 2001
Minimal unit of process dwellings
Uncertainty Not processed

Vulnerability model
Vulnerability description EMS 98 Vulnerability classes
Damage description D0 – D5 (EMS 98)
Vulnerability model Damage probability matrix (DPM – Di Pasquale et al. 2000)
Processed from DPM, providing min, mean, max values
Uncertainty
corresponding to 16%, 50%, 84%)

Scenario outputs
Scenario boundaries Macrosesimic intensity ≥VI MCS
Collapses: D5
Usability definition
Non usable: D3+D4+D5 damage levels
- No. collapsed buildings (D5)
- No. unusable buildings (D3+D4+D5)
Consequences
- No. people in buildings D5 (Fatalities estimated 10-30% of this value)
- No. people in unusable buildings (D3+D4+D5)
Seismic Risk Scenarios supporting
emergency management www.protezionecivile.gov.it

Recent experiences and critical appraisal


Recent experiences made in recent Italian seismic crises can
help to highlight some critical issues concerning applicability and
reliability of seismic scenarios used for emergency management.

2009 Abruzzo Earthquake Event Characteristics and further Mw updates

2012 Emilia Earthquake Sequence characterised by multiple mainshocks

Impact on population: residential and non


2016 Centro Italia
residential people
2009 Abruzzo earthquake www.protezionecivile.gov.it

On April 6, 2009, at 3.32 a.m. the city of Magnitudes of the mainshock:


L’Aquila and surrounding municipalities • Ml 5.8 initial estimate
were shaken by a strong earthquake. • Ml 5.9 revised estimate days after
Four further events Mw5+ followed until • Mw 6.3 later estimate
April 9. • Mw 6.1 recently revised estimate

L’Aquila 2009
Parameters of events M ≥ 5
Date 06-apr-09 07-apr-09 07-apr-09 09-apr-09 09-apr-09
Time 03:32:40 01:15:36 19:47:37 02:52:59 21:38:16
Ml 5,9 5,0 5,4 5,1 5,0
Mw 6,1 5,0 5,4 5,2 5,0
Depth (km) 8 10 17 11 9
Latitude 42,34 42,46 42,3 42,49 42,5
Longitude 13,38 13,39 13,49 13,35 13,35

Total fatalities 308


Total injured 1600 ca.
Total homeless 65,000 ca.
2009 Abruzzo earthquake www.protezionecivile.gov.it

This damage
scenario was
prepared for the
operational
committee Scenario estimates:
• 200-2200 people in collapsed buildings
meeting 30’ after • 8700-54000 homeless
the event, using • 6700-38000 unusable dwellings
a previous • 61000-207000 damaged dwellings
• VIII-IX Max Intensity (MCS)
version of SIGE.
2009 Abruzzo earthquake www.protezionecivile.gov.it

The macroseismic survey


was carried out in 315
locations.
Maximum MCS intensity
was found to be IX-X in
Onna [AQ] and
Castelnuovo [AQ], due
to strong local
amplification effects.
2009 Abruzzo earthquake www.protezionecivile.gov.it

Impact on built environment is more precisely estimated from post


earthquake inspections carried out during the emergency phase, under
the coordination of DPC.
70.000 residential buildings were inspected. 45% of them were found
unusable, 27% heavily damaged, 17% slightly damaged
ESITO A B C D E F TOTALE SCHEDE
USO N. % N. % N. % N. % N. % N. % N. %
Abitazioni private 35985 50,79% 9330 13,17% 1952 2,75% 426 0,60% 19216 27,12% 3947 5,57% 70856 91,19%
Scuole 347 50,36% 194 28,16% 18 2,61% 10 1,45% 107 15,53% 13 1,89% 689 0,89%
Att. produttive e commerciali 765 57,13% 230 17,18% 53 3,96% 14 1,05% 218 16,28% 59 4,41% 1339 1,72%
Caserme e pubblica sicurezza 145 65,32% 43 19,37% 6 2,70% 0 0,00% 23 10,36% 5 2,25% 222 0,29%
Strutture sanitarie 46 36,22% 37 29,13% 9 7,09% 3 2,36% 31 24,41% 1 0,79% 127 0,16%
Uffici 173 49,57% 69 19,77% 11 3,15% 5 1,43% 59 16,91% 32 9,17% 349 0,45%
Altro 2236 54,23% 578 14,02% 135 3,27% 27 0,65% 974 23,62% 173 4,20% 4123 5,31%
TOTALE ESITI 39697 51,09% 10481 13,49% 2184 2,81% 485 0,62% 20628 26,55% 4230 5,44% 77705 100,00%
2009 Abruzzo earthquake www.protezionecivile.gov.it

Comparison of the scenario results with observed damage, requires


some considerations:

 Damage scenarios are related to dwellings, while observed damage


is related to buildings. A suitable conversion factor to transform
buildings into dwellings is assumed to be 2.5.
 SIGE uses as basic inventory the census provided by ISTAT, relevant
to residential properties. Hence, validation of scenarios must be
limited to residential buildings;
 Observed damage is provided in terms of usability classification (A,
B, C, etc.), while SIGE calculates damage levels which must then be
converted into usability classes. In particular:
• Unusable buildings are defined as: D3+D4+D5
• Collapsed buildings: D5
2009 Abruzzo earthquake www.protezionecivile.gov.it
Comparison of scenarios (damage levels)
 Scenarios associated to the two estimated values of Mw are sensibly
different: losses of the first (Mw=6.3 – SIGE 1) are around twice
those associated to the revised value of Mw (Mw=6.1 – SIGE 2).
 Mw calculated by SIGE using revised Ml value (5.9) is 6.13. Hence,
SIGE 2 scenario results are consistent with ML5.9, and then with the
present best estimate of Ml and Mw

SIGE 1 SIGE 2
(Mw = 6.3) (Mw = 6.1)
Dwellings in D1 124457 64071
Dwellings in D2 66534 34958
Dwellings in D3 26188 14134
Dwellings in D4 7116 3869
Dwellings in D5 1081 532
2009 Abruzzo earthquake www.protezionecivile.gov.it

Scenario estimates and "real" data


SIGE 1 (Mw = 6.3) Observed SIGE 2 (Mw = 6.1)
Item Min Mean Max data Min Mean Max
Total collapsed dwellings 393 1081 3030 185 532 1534
Total unusable dwellings 13406 34385 87486 50000 7082 18535 46724
No. people in collapsed dwellings 520 1419 4170 308+1600* 254 714 2035
No. people in unusable dwellings 16168 44951 135251 65000 8663 22393 63112

First scenario estimates (previous SIGE version, Ml 5.8):

• 200-2200 people in collapsed buildings  308+1600 *


• 8700-54000 homeless  65000
• 6700-38000 unusable dwellings  50000
• VIII-IX Max Intensity (MCS)  IX-X

* Warning: the No. of fatalities and injured is only a subset of the


people involved in collapsed buildings.
2012 Emilia earthquake www.protezionecivile.gov.it

On May 20, 2012 a Ml 5.9 earthquake hit


territories of the Po Valley, in Central Italy.
On May 29 a further mainshock of Ml 5.8
occurred with epicenter 20 km west of the
main event.
Emilia 2012
Parameters of the mainshocks
Date 20-May-09 29-May-09
Time 04:03:50 09:00:02 Daily No. Earthquakes since May 16, 2012
Ml 5.9 5.8
Mw 5.8 5.6
Depth (km) 6 8
Latitude 44.89 44.85
Longitude 11.23 11.09

Total fatalities 7 + 19
Total injured 350 ca.
Total homeless 19,000 ca.
2012 Emilia earthquake www.protezionecivile.gov.it

After May 20 a macroseismic survey in 30 municipalities was carried out.


After May 29 the survey was repeated and widened in 87 municipalities.
Maximum MCS intensity was:
• VII after first mainshock
• VII-VIII after second mainshock
2012 Emilia earthquake www.protezionecivile.gov.it

Actual losses to the built environment include, beyond ordinary


buildings, numerous industrial warehouses and heritage buildings,
however only damage data of ordinary dwelling buildings collected with
the AeDES form can be compared.
45,000 ordinary buildings, 25,000 residential buildings inspected.
ESITO A B C D E F TOTALE SCHEDE
USO N. % N. % N. % N. % N. % N. % N. %
Abitazioni private 10596 41,3% 4517 17,6% 1270 5,0% 127 0,5% 7677 29,9% 1450 5,7% 25637 57,0%
Scuole 471 45,2% 301 28,9% 63 6,1% 19 1,8% 160 15,4% 27 2,6% 1041 2,3%
Attività produttive 588 27,3% 417 19,3% 95 4,4% 15 0,7% 953 44,2% 89 4,1% 2157 4,8%
Attività commerciali 1036 37,1% 633 22,7% 178 6,4% 20 0,7% 659 23,6% 263 9,4% 2789 6,2%
Uffici 462 38,8% 277 23,3% 78 6,5% 10 0,8% 257 21,6% 107 9,0% 1191 2,6%
Depositi 2432 20,4% 1911 16,0% 533 4,5% 30 0,3% 6514 54,5% 529 4,4% 11949 26,6%
Attività turistiche 65 34,9% 38 20,4% 19 10,2% 1 0,5% 56 30,1% 7 3,8% 186 0,4%
TOTALE ESITI 15650 34,8% 8094 18,0% 2236 5,0% 222 0,5% 16276 36,2% 2472 5,5% 44950 100,0%
A B C D E F
2012 Emilia earthquake www.protezionecivile.gov.it
Scenarios with local Magnitude (ML)
Seismic scenarios using Local
Magnitude (5.9 and 5.8 ML Sige 1 Sige 2
respectively) for both Dwellings in D1 230609 151496
mainshocks of 20th and 29th of Dwellings in D2 106751 68058
May. Potential losses are Dwellings in D3 34074 22330
relevant to the two events Dwellings in D4 6527 4782
Dwellings in D5 510 493
independently.
Damage is converted into unusable and collapsed dwellings, and in
people involved to provide information for emergency management
20th of May (SIGE 1) 29th of May (SIGE 2)
Min Mean Max Min Mean Max
Total collapsed dwellings 89 510 1927 142 493 1789
Total unusable dwellings 11597 41111 103027 10927 27605 81789
Number of people in collapsed dwellings233 1152 4236 342 1144 3947
Number of people in unusable buildings
25519 86496 219364 24212 61451 173554
2012 Emilia earthquake www.protezionecivile.gov.it

Repeated mainshocks - Combined scenario (ML)


A combined scenario has been computed by enveloping the
calculated macrosesimic intensities of the two scenarios, not
accounting for cumulated damage and exposure changes.

Bagnolo San Bagnolo San


Vito MCS 6.3 Vito MCS 8

1st 2nd

Sige 1 Sige 2
Combined
(20/5) (29/5) 20-50% increases
Total collapsed dwellings 510 493 770 with respect to
Total unusable dwellings 41111 27605 48293
No. people in collapsed dwellings 1152 1144 1719
the worst single
No. people in unusable dwellings 86496 61451 102152 scenario
2012 Emilia earthquake www.protezionecivile.gov.it
Scenarios with Moment Magnitude (Mw)

For the same events, seismic scenarios are re-processed using as input
data Mw (5.8 and 5.6), rather than Ml (5.9 and 5.8) converted into Mw
(6.1 and 6.0).

20th of May (SIGE 1) 29th of May (SIGE 2)


Min Mean Max Min Mean Max
Total collapsed dwellings 12 80 506 13 45 129
Total unusable dwellings 2366 8010 33035 1518 3570 8198
Number of people in collapsed dwellings 38 217 1167 36 110 314
Number of people in unusable buildings 5227 17874 72240 3480 8206 18727

A scenario combining the two events in Mw is then calculated


following the same intensity envelope criterion assumed for ML
scenarios.
2012 Emilia earthquake www.protezionecivile.gov.it

Correlation with observed damage

 First combined scenario (Sige Comb ML) overestimate losses


 Second combined scenario (Sige Comb Mw) is closer to observed damage,
but the number of fatalities, almost all occurring not in residential
buildings, is badly estimated.
 Although other factors (like local amplification) can strongly affect the
results, the magnitude is responsible of wide uncertainties. Immediate
Mw estimate is recommended due to the uncertainties on the Ml-Mw
conversion laws for single earthquakes.
 Industrial production buildings can play a major role on economic and
social losses in industrialized areas

SIGE Comb SIGE Comb Observed


(ML) (Mw) damage
Total unusable dwellings 48293 8373 24157
No. people in unusable dwellings 102152 18573 19000
Fatalities 344 49 26
2016-17 Central Italy earthquakes M. Dolce
www.protezionecivile.gov.it

August 24th, 2016, Ml 6.0-Mw 6.0 Earthquake


• On August 24, 2016, at 3:36 a.m., a strong earthquake (Ml 6.0, Mw
6.0, depth 8 km) occurred along the Apennines Chain, Central Italy.
• Disruption occurred in three small municipalities, Amatrice,
Accumoli and Arquata.
• intensities attained
the degree X-XI on the MCS
scale and X on the EMS
scale (INGV).

• 299 fatalities.
• 390 hospitalized
injured people.
39
2016-17 Central Italy earthquakes M. Dolce
www.protezionecivile.gov.it

October 30th, 2016, Ml 6.1-Mw 6.5 Earthquake

On October 30, at 7:40 local time,


the strongest seismic event of the
sequence occurred in an area
located between the two zones
previously hit.

The Mw 6.5 occurred 18km


northnorthwestward from the
first epicenter.

No casualty occurred
(28 people injured)
40
2016-17 Central Italy earthquakes M. Dolce
www.protezionecivile.gov.it

Present-day seismicity
On April 26th, 2017, the seismic
sequence
was formed by more than
64,000 events:

• 2 with Mw ≥ 6.0
• 7 with 5.0 ≤ Mw < 6.0
• 61with 4.0 ≤ Ml < 5.0
• 1034 with 3.0 ≤ Ml < 4.0.

Maximum distance between


Mw5+ events was about 50 km
along NNW-SSE strike.
41
2016-17 Central Italy earthquakes M. Dolce
www.protezionecivile.gov.it

42
M. Dolce
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

SHAKEMAP
August 24, 2016

Mappe di scuotimento
INGV degli eventi del
24.08.16, 26.10.16,
30.10.16, 18.01.17 in
termini di Intensità
Strumentale
M. Dolce
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

SHAKEMAP
October 26, 2016

Mappe di scuotimento
INGV degli eventi del
24.08.16, 26.10.16,
30.10.16, 18.01.17 in
termini di Intensità
Strumentale
M. Dolce
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

SHAKEMAP
October 30, 2016

Mappe di scuotimento
INGV degli eventi del
24.08.16, 26.10.16,
30.10.16, 18.01.17 in
termini di Intensità
Strumentale
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

SHAKEMAP
October 18, 2017

Mappe di scuotimento
INGV degli eventi del
24.08.16, 26.10.16,
30.10.16, 18.01.17 in
termini di Intensità
Strumentale
M. Dolce
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

Macroseismic survey
after August 24th, 2016

DPC officers and CNR-IGAG and


INGV researchers conducted field
surveys to assign a macroseismic
intensity MCS to municipalities and
localities in the epicentral area.
Values exceeding intensity X MCS
were found in and near the epicenter.

47
M. Dolce
2016-17 Central Italy earthquakes www.protezionecivile.gov.it

Macroseismic survey
after October 30th, 2016

The maximum observed


(cumulated) intensity is XI in
the MCS scale.

The macroseismic field of


cumulated intensities IMCS≥7
is 70 km long and 30 km
wide (before January 18,
2017).

48
2016 Central Italy earthquake www.protezionecivile.gov.it
Observed damage due to the first earthquake only
Before the second mainshock of October 26th, 26.149 ordinary buildings were totally
inspected in the 4 stricken regions, against a total amount of 30.000 building inspection
requests (corresponding approximatively to 69.955 inspection requests);

RESULT A AF E + EF B+C+D+BF+CF+DF w/o result TOTAL INSPECTIONS


REGION N. % N. % N. % N. % N. % N. %
Abruzzo 3423 65,51% 149 2,85% 716 13,70% 790 15,12% 147 2,81% 5225 19,98%
Lazio 2091 36,28% 456 7,91% 2426 42,10% 682 11,83% 108 1,87% 5763 22,04%
Marche 4102 42,40% 449 4,64% 3156 32,62% 1738 17,96% 230 2,38% 9675 37,00%
Umbria 3414 62,23% 244 4,45% 932 16,99% 732 13,34% 164 2,99% 5486 20,98%
TOTAL 13030 49,83% 1298 4,96% 7230 27,65% 3942 15,08% 649 2,48% 26149 100,00%

Inspected Buildings Projection


A 13030 49,83% 34858
AF 1298 4,96% 3472
E + EF 7230 27,65% 19342
B+C+D+BF+CF+DF3942 15,08% 10546
w/o result 649 2,48% 1736
Total Requests (26/10)
26149 69955
2016 Central Italy earthquake www.protezionecivile.gov.it

Scenarios of the 24.08.16 earthquake associated to the estimated


values of Ml and Mw are sensibly different: losses of the first (Ml=6.0
converted by SIGE to Mw=6.23 – SIGE 1) are around three times
those associated to Mw (Mw=6.0 – SIGE 2).

Comparison of scenarios (damage levels)

SIGE 1 SIGE 2
(ML = 6.0) (Mw = 6.0)
Dwellings in D1 98670 28649
Dwellings in D2 47868 15380
Dwellings in D3 16725 5657
Dwellings in D4 3515 1218
Dwellings in D5 299 97
2016 Central Italy earthquake www.protezionecivile.gov.it

Real data for a direct comparison relevant to the first mainshock only
must be better elaborated.
At the present only data on fatalities and injured can be compared.
The human exposure on the 24.08.16 was considerably increased,
with respect to residents, because of the summer holydays and a
country fair in Amatrice.
Sige 1 (ML = 6.0) Sige 2 (Mw = 6.0)
Min Mean Max Min Mean Max
Total collapsed dwellings 65 299 1129 23 97 474
Total unusable dwellings 5459 20539 56603 2053 6972 26323
No. people in collapsed dwellings 51 351 1542 16 85 604
No. people in unusable dwellings 5502 29617 89133 1602 7201 38928

299 fatalities
390 Injured
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
Comparison of scenarios www.protezionecivile.gov.it

• Besides SIGE, DPC Competence Centres ReLUIS and EUCENTRE


can make SRS’s, as well as PSRA’s, using different platforms
based on different models.
• Under request of the National Major Risk Commission, scenarios for
9 hypothetical earthquakes with different magnitude and
epicentral location and same focal depth (10 km) were produced.
• Scenarios were prepared with three platforms, to evaluate the
epistemic uncertainty related to model peculiarities.
Scenario Mw
• SIGE (Civil Protection Department) Alfa 6.6
• SP Bela (EUCENTRE) Beta 6.6
• SISMA (ReLUIS-PLINIUS) Gamma 6.4
Delta 7.0
Epsilon 7.1
Modelling hypotheses and output quantities Zeta 6.1
were preliminarily harmonized in order to Eta 6.1
Teta 6.1
eliminate unnecessary differences Iota 7.1
Comparison of scenarios www.protezionecivile.gov.it

SIGE SpBELA SISMA


Hazard Hazard SISMA
Source type Punctual Areal (fault); pucntual Punctual
Magnitude
Conversion ML-Mw
ML (or Mw)
Gasperini et al.2013
Mw
1 -
ML
Gasperini et al.2013
Constant for focal depth <20 km (Io=2.1241*Mw-4.106
Epicenter intensity I o (as function of Mw) - Io=(ML-1)/0.51 (Karnik) MSK
(CPT11)) MCS
Per sorgente puntuale ed areale: For
Two different attenuations depending on Magnitude and
areal and punctual source: Akkar e Bommer (2010); Blake (1941) on purpose ri-calibrated through y coefficient
focal depth:
Boore e Atkinson (2008)
GMPE M≤5.5 Io≤VIII) = Gomez Caprera (2007)
1
Boore e Atkinson (2008)

Bindi et al. (2011) For punctual


by default is 4. changed into 5.52 (for Amatrice)

M>5.5 (Io>VIII) =Pasolini et al (2008) γ ricalibrato per Amatrice (=5.52)


source also Cauzzi and Faccioli is considered.

Processed. With punctual source, depth is calculated


Marked
1
Focal Depth Processed for h>20 km through Cauzzi Faccioli. Further 3 relations do not Processed (Blake)
process focal depth.

Local amplification Not processed


Processed using soil maps from Di Capua and
Peppoloni (2009)
-
differences
Shaking parameters

Minimal unit of process


Macroseismic intensity MCS

Municipal
Per sorgente areale l’attenuazione avviene dai bordi del
rettangolo.
Municipal
Macroseismic intensity MCS

Census section
in hazard,
Processed (half degree of macroseismic intensity
vulnerability
Uncertainty Processed through spectrum at 16°, 50°, 84° percentiles; Not Processed

and
evaluated in each municipality)

Vulnerability model

Vulnerability description EMS 98 Vulnerability classes


Vulnerability model
Mechanical model based on prototypal buildings
Vulnerability model

EMS 98 Vulnerability classes


exposure
associated to vulnerability classes;
Damage description D0 – D5 (EMS 98) Limit states converted into damage D0 – D5 (EMS 98) modelling
1
levels D0 – D5 (EMS 98)
Vulnerability model
Damage probability matrix (DPM – Di Pasquale et al.
2000)
Processed from DPM, providing min, mean, max values
Straight comparison between structural capacity and
displacement capacity
Damage Probability Matrices (Cacace-Zuccaro)

Processed through min, mean and maximum values


and in
Uncertainty

Exposure model
corresponding to 16%, 50%, 84% percentile) Processed through probablitic distribution obtained on
building samples for each vulnerability class
associated to DPM

Exposure model
scenario
Database Istat 2001
Exposure model

Istat 2001
Istat 2001 - on purpose ricalibrated output
Minimal unit of process dwellings Building
Uncertainty Not processed 1
Buildings and dwellings
Not processed
Not processed

Scenario outputs Scenario outputs


Scenario outputs
Scenario boundaries Macrosesimic intensity ≥VI MCS Macrosesimic intensity ≥V MCS
Radius=70 km for M≥VI MCS; Radius=100 km for M≥VII
Collapses: D5; MCS Collapses: D5; Non usable:
Usability definition
Non usable: in D3+D4+D5 damage levels Collapses: D5; Non 0.1*D2+0.6*D3+D4+D5

Casualties
- people resident in buildings damaged D5 (Fatalities
can be estimated between 10% and 30% of this value);
1
usable : 0.816 (D2+D3)+D4;

- people resident in buildings damaged D5 (Fatalities


Victims and injured calculated on the basis of a fixed
occupation ratio (80%). Casualties are dependending on
- people resident in unusable buildings (D3+D4+D5) can be estimated between 10% and 30% of this value); residents in D4 and D5 buildings
- people resident in unusable buildings (D3+D4+D5)
Comparison of scenarios www.protezionecivile.gov.it

Scenario results in term of collapses and unusable dwellings (mean values)

Number of collapsed dwellings


12000

10000
High M Large
absolute differences
Low M Large
8000
SIGE
6000 SISMA
% differences SPBELA
4000

2000

0
Zeta (6,1) Eta (6,1) Theta (6,1) Gamma (6,4) Alfa (6,6) Beta (6,6) Delta (7,0) Epsilon (7,1) Iota (7,1)

Number of unusable dwellings


800000

700000

600000

500000 SIGE
400000 SISMA
SPBELA
300000

200000

100000

0
Zeta (6,1) Eta (6,1) Theta (6,1) Gamma (6,4) Alfa (6,6) Beta (6,6) Delta (7,0) Epsilon (7,1) Iota (7,1)
Comparison of scenarios www.protezionecivile.gov.it

Scenario results in term of involved people in collapses and unusable dwellings


(mean values)

Number of resdents in collapsed dwellings

High M Large
30000

25000
absolute differences
20000
Low M Large SIGE
15000 SISMA
% differences SPBELA
10000

5000

0
Zeta (6,1) Eta (6,1) Theta (6,1) Gamma (6,4) Alfa (6,6) Beta (6,6) Delta (7,0) Epsilon (7,1) Iota (7,1)

Number of residents in unusable dwellings


1400000

1200000

1000000

800000 SIGE
SISMA
600000
SPBELA

400000

200000

0
Zeta (6,1) Eta (6,1) Theta (6,1) Gamma (6,4) Alfa (6,6) Beta (6,6) Delta (7,0) Epsilon (7,1) Iota (7,1)
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
Critical issues of PSRA and SRS
www.protezionecivile.gov.it

Hazard data and modelling


 Source Modelling: Point-source vs. finite-source, directivity, modelling of the
space-time stochastic process (PSRA)
 Attenuation Modelling (GMPE’s): choice of the most suitable ones among
hundreds
 Considering and modelling local amplification effects: Not considered,
use of (morphological/geological) proxies, direct microzonation
 Considering strong motion records (shakemaps): availability in short
time, modelling

 Considering and modelling coseismic effects on damage: landslides,


rockfalls, soil fractures, liquefaction, surface fault rupture, etc.
 Considering cascading effects: NATECH (e.g. Seveso Plants, nuclear
power plants, fires,…) or NATNAT (tsunamis, landslide dams, failure of river
levees, snow avalanches …)

 Uncertainties: of basic hazard data (magnitude and location  double, one


is permanent in time, the other reduces with time), of modelling assumptions
06.04.2009 Abruzzo Earthquake
www.protezionecivile.gov.it

Macroseismic survey vs. SIGE intensity

6,5

Updated INGV Mw estimate 6.1


Survey – Max MCS Intensity IX-X
SIGE – Max MCS Intensity VIII-IX
20.05.2012 Emilia Earthquake
www.protezionecivile.gov.it

Macroseismic survey vs. SIGE intensity

Updated INGV estimate Mw 5.8


Survey – Max MCS Intensity VII
SIGE – Max MCS Intensity VII - VIII
20.05.2012 Emilia Earthquake
www.protezionecivile.gov.it

Macroseismic survey vs. SIGE intensity

Updated INGV estimate Ml 5.9


Survey – Max MCS Intensity VII
SIGE – Max MCS Intensity VIII - IX
24.08.2016 Amatrice Earthquake
www.protezionecivile.gov.it

Macroseismic survey vs. SIGE intensity

SIGE, and in general scenario 6,5


simulations, do not estimate correctly
the maximum intensity and the intensity
distribution for single earthquakes, as
they are derived from statistical analyses
of past earthquakes.
Moreover, local ground motion
modifications increase the
discrepancies.

Updated INGV estimate Mw 6.0


MS – Max MCS Intensity IX-X AMATRICE
SIGE – Max MCS Intensity VIII - IX
Critical issues of PSRA and SRS
www.protezionecivile.gov.it

Exposure data and modelling


 Poor information of inventory: exposure dataset related to ordinary
buildings are typically provided by census, cadastral and/or satellite images,
with poor information (geometry, age, structural type), which do not properly
characterize the variety of buildings types and their taxonomy.
 Completeness of inventory: Exposure dataset need to be improved for:
• essential facilities (e.g. hospitals, schools)
• cultural heritage (with peculiar structures like churches, tower, castles)
• commercial and industrial production buildings
• transport infrastructures (e.g., roads, railways, airports, seaports, etc.)
• lifelines (main lines and in urban environments)
 Exposure in multiple mainshock sequences: After the first mainshock
building stock exposure and human exposure vary considerably, due to the
state of damage (change of vulnerability) and to people evacuation from
buildings for damage and fear
 Uncertainties: Uncertainties are relevant to number of buildings, building
taxonomies, human exposure (daily, weekly, seasonal).
DPC-ReLUIS
www.protezionecivile.gov.it

CARTIS Project
DPC is developing, along with ReLUIS, CARTIS a data
collection for improving building taxonomies and their
territorial distribution at national level
32 harbors
DPC-EUCENTRE
www.protezionecivile.gov.it
Seismic Risk Platform
WebGIS for the computation of
seismic risk maps and real
time damage scenarios

50.000 schools

238 dams

Residential buildings
(ISTAT 2001) Road way network
(17.000 bridges
38 airports)
Critical issues of PSRA and SRS
www.protezionecivile.gov.it

Vulnerability data and modelling


 Completeness of inventory of fragility/vulnerability functions: structural
types other than ordinary buildings (residential, office, schools) still need
much research work;
 Consideration of cumulated damage under repeated shocks: strong
aftershocks, sometimes even stronger than the mainshock finds the
vulnerability of the building stock increased by the damage suffered from the
previous shocks;
 Comparability of metrics: fragility curves, DPM, vulnerability functions are
hardly comparable and so are their outputs (repair/reconstruction cost,
damage ratio, damage levels, limit states, etc.).
 Calibration: if not derived from statistical damage data of previous
earthquakes, fragility curves and vulnerability functions should be calibrated
through correlation between predicted and observed damage.
 Uncertainties: uncertainties are relevant to the damage model, the structural
features within the same structural type, etc..
Da.D.O
Damage Database Observation www.protezionecivile.gov.it

With the final scope of supporting further calibration, validation and


elaboration of fragility curves and damage probability matrices (DPM),
the Italian Civil Protection Department is developing, with the support of
Eucentre, a web-platform specifically aimed at damage database storage.
Da.D.O (Damage Database Observation) is about to be opened to the
scientific community in the next weeks. It collects damage database relevant
to 10 seismic events occurred in Italy since 1976 try.
For each event DaDO provides:
 Event characteristics;
 Losses (population);
 Damage Database;
 Macroseismic maps;
Users will be enabled to:
 Download database;
 Download shapefile;
 Use filters and specific queries;
PSRA and SRS for Civil
www.protezionecivile.gov.it

Protection Purpose

Summary
1. Organization and mandate of the Italian Civil Protection
2. Needs and use of PSRA and SRS for CP purpose
3. Application examples:
• PSRA for the allocation of seismic prevention funds
• SRS of 2009 Abruzzo Earthquake
• SRS of 2012 Emilia Earthquake
• SRS of 2016 Amatrice Earthquake
• Comparison of SRS’s based on different approaches
and platforms
4. Critical issues of PSRA and SRS
5. Final considerations
Final Considerations (1) www.protezionecivile.gov.it

• PSRA and SRS are affected by very large uncertainties, most


of them epistemic, due to the lack of knowledge of territorial
information on hazard, vulnerability and exposure.
• They must therefore be used for CP purpose with great care.

• PSRA can be of great use to make rational decision for


seismic risk mitigation and for contingency planning at
national level.
• The choice of the most suitable risk indicator (human losses,
direct economic losses, indirect economic losses, integrated
risk-social vulnerability, etc. at individual or global level), is a
very delicate one.
• It must be well addressed to the specific use.
Final Considerations (2) www.protezionecivile.gov.it

• The most delicate use of SRS is for emergency management,


especially in the first few hours after an event, when
information from the affected territory is nul or very scarce.
• The extremes of the uncertainty interval differ by one order of
magnitude and sometimes are yet insufficient to include the
real value of the assessed quantity.
• Evaluations from SRS should be complemented ASAP with
data from the territory and remote sensing.
• All the above considerations are drawn from the
experiences in a country with a long tradition of seismic
risk assessment and where data are quite available.
• Uncertainty problems become even bigger (or perhaps
smaller?) for countries for which poorer data relevant to
hazard, exposure and vulnerability are available.
Probabilistic Seismic Risk
Analysis and Damage
Scenarios
for Civil Protection Purpose
Needs, implementations and
critical issues

Prof. Mauro Dolce


Italian Department of Civil Protection

You might also like