Current Macroseismic Intensity Scales Around the World

Without going into extensive detail, many macroseismic intensity scales around the world are either informal, not well documented, or rely on dated protocols and observations inherent to the scale used. For a summary of most scales, refer to Musson and Cecić (2002) and Musson and others (2010). For the purposes of an IMS, it is useful to note that in addition to the United States and New Zealand, the adoption of an IMS in several key at-risk nations could considerably improve international consistency in earthquake response and hazard and risk analyses. Although adoption of IMS in these countries is beyond the scope of the workshop, it is useful to consider how such a goal would improve upon the current situation.

For example, in the Philippines, the Philippine Institute of Volcanology and Seismology Earthquake Intensity Scale (Philippine Institute of Volcanology and Seismology, 2018) is as nondescript as MMI and still invokes ground failure observations for higher intensities. The Philippine Institute of Volcanology and Seismology Earthquake Intensity Scale was adopted in the Philippines in 1996, replacing the Rossi-Forel Scale used prior.

In China, the Chinese Seismic Intensity Scale (CSIS; State Administration for Market Regulation, Standardization Administration of the People’s Republic of China [SAMR, SAPRC], 2020), now in English, has five general categories of common building types and five building damage grades. The building classes and damage definitions are quite different from EMS–98’s treatment, but the general philosophy of using quantitative ranges of building damage grade and classes is similar. The CSIS also defines instrumental intensity in a very similar fashion to Wald and others (1999) and defers to the instrumental scale over macroseismic observations where seismic stations are present (SAMR, SAPRC, 2020). A challenge for the USGS ShakeMap group, and in developing historical ShakeMaps Atlas (Marano and others, 2023) for important historical events in China, is the outdated use of isoseismals to describe the mesoseismic area in scientific studies rather than reporting site-specific values. Isoseismals are traditionally drawn to separate adjacent areas of equal intensity. They thus have no numerical use because trends in the values are not respected. Often only isoseismal contours are presented rather than tabulating locations with their assigned intensities.

In contrast, whereas the JMA Intensity Scale (JMA, 1996) has corresponding macroseismic observations, official intensity levels are based strictly on instrumentally derived values; macroseismic observations are not considered. This is unique the world over, in that actual damage does not dictate the intensity value: rather, the instrumental value does. Japan Meteorological Agency Intensity values range from 1 to 7, with 5-, 5+, 6-, and 6+ explicitly considered, so there are nine levels in all, consistent with other scales. Of note, Japan has such a dense network of strong-motion accelerometers and intensity meters that these values are widely distributed and are available and in real time. However, because they are based on instrumental data, they are incompatible with macroseismic scales used elsewhere.

Russia, other former Eastern bloc nations, and India use the Medvedev-Sponheuer-Karnik Scale (Medvedev, 1962), known as the MSK or MSK–64. Because MSK–64 is a predecessor and part of the basis of EMS–98, MSK-64 shares many aspects of the newer EMS–98, but differences remain. As well as creating difficult comparisons, varying scales and differing assignment strategies make it difficult to compare and analyze events around the globe directly and uniformly.

Overviews of existing macroseismic intensity scales by Musson and others (2010) indicate that the various scales used worldwide are not all compatible and that the value of an assignment is dependent on the quality of the scale used. Musson and others (2010) go on to provide approximate conversions from several standard scales (for example, MMI, Mercalli-Cancani-Sieberg Scale [MCS], MSK, and JMA–96) to EMS–98, with the implicit assumption that EMS–98 is the most modern and comprehensive scale yet developed.

Short of proposing adoption of a standard IMS, Musson and others (2010) provide strategies for converting among scales and recommend that the macroseismic observations be reinterpreted into the scale of interest (Musson and others, 2010). However, Musson’s conversions do not address many of the informal scales and modifications made in many countries, so caution is warranted. This was particularly problematic for the USGS ShakeMap team when creating its ShakeMap Atlas (Marano and others, 2023), which presents ShakeMaps for 14,000 historical global earthquakes, some of which were constrained primarily by macroseismic observations.

EMS–98 Development

Workshop participants were informed (or reminded) by our EMS–98 forebearers that several key resources and studies contributed and documented efforts toward EMS–98 and to an IMS.

EMS–98 is now the basis for assigning seismic intensities in European nations. As described by Grünthal (2022), the history of the EMS began in 1988 when he initiated a special international Working Group in the framework of the European Seismological Commission (ESC) to review and update the Medvedev-Sponheuer-Karnik Scale (MSK–64), which was used in its basic form in Europe for almost a quarter of a century. The MSK–64 Scale, which itself is an update, is based on the experiences that were available in the early 1960s from the application of the Mercalli-Cancani-Sieberg Scale (MCS) from Sieberg (1932), the Modified Mercalli Scale in its version from 1931 and 1956 (MM–31 and MM–56), and the Medvedev Scale from 1953 (Musson and others, 2010; Grünthal, 2022).

At the 2022 Powell Center meeting, Jochen Schwarz described how the final European Macroseismic Scale (EMS–98; Grünthal and others, 1998) came into being following the EMS–92 draft version for testing rolled out at the 1988 ESC meeting in Prague in 1992. The principal final layout of the European Macroseismic Scale (EMS–92) was created by Gottfried Grünthal, Roger Musson, Jochen Schwarz, and Maximillian Stucchi at a meeting in Potsdam, Germany, on June 17–21, 1992 (Grünthal and others, 1998), and drawings by A.S. Taubajev were widely used. The EMS–92 was used in testing, and the incorporation of the lessons learnt during the worldwide applications of the EMS–92 by several researchers was maintained by the EMS–98 editors. For more details on the contributions to EMS–98, refer to Grünthal and others (1998).

In 1996, the 11th World Conference on Earthquake Engineering in Acapulco, Mexico, featured a special theme session on the scale and its testing and development (Grünthal and others, 1998). In all, the Working Group spent more than 5 years on research and development, followed by 3 years of testing. Later that year, the XXV General Assembly of the ESC in Reykjavik, Iceland, passed a resolution recommending the adoption of the new scale within the member countries of the ESC (Grünthal and others, 1998). The last meeting of that EMS Working Group was at the 1998 ESC meeting in Tel Aviv, Israel.

Continuing EMS–98 Activities Toward an International Macroseismic Scale

Following the adoption of EMS–98, a series of efforts continued to evaluate the use of EMS–98, expand it to other regions, and consider how it could be further adapted for a potential IMS. Early activities involved an “EMS Group” (Jochen Schwarz, Gottfried Grünthal, Roger Musson, and Massimiliano Stucchi) meeting on EMS–98, “Applications, Experiences and Needs,” held in Weimar, Germany, on October 13–14, 2005. Subsequent related publications included Musson and others (2010), a comparison of macroseismic intensity scales, and Schwarz (2011), which was a review of contributions to empirical vulnerability assessment with the damage description in EMS–98.

International Macroseismic Scale Working Group 2013–14

As recounted by Dr. Robin Spence in the Powell Center meeting, the Cambridge Architectural Research and IMS project in 2013–14 (IMS Working Group 2013–14) included several EMS–98 pioneers (Gottfried Grünthal, GeoForschungsZentrum (GFZ), Potsdam, Germany; Roger Musson, British Geological Survey, Edinburgh, Scotland; Jochen Schwarz, Bauhaus-University Weimar, Germany), Thomas Wenk (Wenk Erdbebeningenieurwesen und Baudynamik GmbH, Zurich, Switzerland), Robin Spence (Cambridge Architectural Research, United Kingdom), and Roxane Foulser-Piggott (Cambridge Architectural Research, United Kingdom).

The IMS Working Group 2013–14 was established at the Congress on Recent Advances in Earthquake Engineering and Structural Dynamics 2013 in Vienna, Austria, with a mini symposium “Conclusions from 15 years of international experience with the EMS–98,” organized by Jochen Schwarz, Gottfried Grünthal, Roger Musson, Robin Spence, and Massimiliano Stucchi. The contributions included papers by Maqsood and others (2013) on the “Application of the EMS–98 in the Asia-Pacific Region,” and Stavrianaki and others (2013), who applied EMS–98 to historical earthquakes in Greece. Moreover, Foulser-Piggott and Spence (2013a, b) described extending EMS–98 for more convenient application outside Europe in two papers pertinent to the workshop: “I: Review of field experience using EMS–98” and “II: Development of the International Macroseismic Scale.”

The initial aims and philosophy of the IMS Working Group 2013–14 were to:

• • Extend the European Macroseismic Scale (EMS–98) and develop it into the IMS to enable it to become more generally applicable to all parts of the globe

• • Consider that the structure of EMS–98, based on vulnerability classes, is appropriate for such international application

• • Consider that the main benefit of this development will be to provide a standard scale for international use, reducing the need for interscale conversions

The IMS Working Group 2013–14 background studies in 2013–14 included review of experiences with EMS–98, an investigation of data on global building types, a review of regionally dominant building types, and the quantification of vulnerability modifiers. The aim of the review of field experience using EMS–98 was to identify areas in which the scale could be developed to make it more suitable for international application (Foulser-Piggott and Spence, 2013a, b). Data were collected using an online questionnaire, completed by EMS–98 users in different countries, inside and outside Europe (the 19 respondents are listed in table 2). The questions covered two main areas: (1) key features of EMS–98 study: structure types present, vulnerability, damage grades, and range of intensities; and (2) problems encountered: absence of structure types, insufficient range of vulnerabilities for a structure type, and difficulties in making intensity assignments. Based on these interviews, the authors proposed regional grouping of countries that shared similar predominant building types (table 3). From those groupings, Foulser-Piggott and Spence (2013a) also proposed vulnerability ranges for dominant building types for their proposed regions (table 4 and fig. 5).

Table 2.    

Participants of 2013 online questionnaire about EMS–98 use in different countries.

[Modified from Foulser-Piggott and Spence (2013a)]

Table 2.    Participants of 2013 online questionnaire about EMS–98 use in different countries.

Author	Location of studies
Augusto Antonio Gomez Capera	Colombia
Michel Clotaire	Martinique, Switzerland
Lekkas Efthymios	Japan, Italy, China
Mustafa Erdik	Turkey
Mohammed Farsi	Algeria
Mohsen Ghafory-Ashtiany	Iran
Tatiana Goded	Spain, New Zealand
Fabrizio Meroni	Italy
Roger Musson	United Kingdom
Prabhis Pandhe	India
Vera Pessina	Italy
Antonios Pomonis	Greece
Christophe Sira	French territories overseas (Martinique and Guadeloupe)
Syed Tariq Maqsood	Pakistan
Fabio Taucer	Haiti
Andrea Tertulliani	Italy
Ufuk Yazgan	Turkey
Wen Zengping	China
Motti Zohar	Israel

Table 3.    

Vulnerability ranges for principal building types and vulnerability class ranges for the proposed regions.

[Modified from Foulser-Piggott and Spence (2013a, table 4). —, not applicable; EMS–98, European Macroseismic Scale; A–E are EMS–98 vulnerability classes; RC, reinforced concrete, x or xx, one or two vulnerability class discrepancy]

Table 3.    Vulnerability ranges for principal building types and vulnerability class ranges for the proposed regions.

Class	Description	Zone 1 range	Zone 2 range	Zone 3 range	Zone 4 range	Zone 6 range	Overall range	EMS–98 median	Discrepancy
—	Number of countries or regions	9	1	9	1	9	29	—	—
TL	Timber, light frame	C to E	—	C to D	C	C to E	C to E	D	—
TH	Timber, heavy frame	C to E	—	0	0	B to D	B to E	D	—
WM	Weak masonry	A	—	A to B	A	A to B	A to B	A	—
LBM	Load-bearing masonry	A to C	—	B to C	B	B to C	A to C	B	—
SRM	Structural masonry	B to D	D	B to C	—	B to D	B to D	C	—
RCFL1	Concrete frame, low rise, precode	B to C	B	C to D	C	B to D	B to D	C	—
RCFM1	Concrete frame, mid rise, precode	B	B	C to D	C	B to D	B to D	C	x
RCFH1	Concrete frame, high rise, precode	B	—	B to C	C	B to D	B to D	C	x
RCFL2	Concrete frame, low rise, early code	B to D	C	C to D	C	B to D	B to D	D	x
RCFM2	Concrete frame, mid rise, early code	B to C	C	B to D	C	B to D	B to D	D	x
RCFH2	Concrete frame, high rise, early code	B to C	—	C to D	C	B to D	B to D	D	x
RCFL3	Concrete frame, low rise, recent code	C to D	D	C to E	—	C to E	C to E	E	x
RCFM3	Concrete frame, mid rise, recent code	B to C	C	B to E	—	C to E	B to E	E	xx
RCFH3	Concrete frame, high rise, recent code	B to C	—	B to D	—	C to E	B to E	E	xx
RCSWL	Concrete shear wall, low rise	C to D	D	C to D	—	C to E	C to E	C/E	—
RCSWM	Concrete shear wall, mid rise	B to D	D	C	—	C to D	C to D	C/E	—
RCSWH	Concrete shear wall, high rise	B to C	C	C to D	—	C to D	C to D	C/E	—
RCTU	RC tilt-up	B to D	—	—	—	C to D	B to D	—	—
RCDU	RC dual	C	D	B to D	—	—	B to D	—	—
SFL	Steel frame, low rise	C	C	C to D	D	C to E	C to E	E	x
SFM	Steel frame, mid rise	B to C	—	C to D	D	C to E	B to E	E	x
SFH	Steel frame, high rise	B to C	—	C to D	D	C to E	B to E	E	xx
SBFL	Steel braced frame, low rise	C to D	—	C	—	C to D	C to D	E	xx
SBFM	Steel braced frame, mid rise	C	—	C	—	C to D	C to D	E	xx
SBFH	Steel braced frame, high rise	B to C	—	C	—	C to D	C to D	E	xx
LMF	Light metal frame	C	—	D	D	C to D	C to D	E	x
SRC	Steel reinforced concrete, Japan	C	—	—	—	C to D	C to D	—	—
MH	Mobile home	D	—	—	—	—	C to D	—	—

Table 4.    

Proposed vulnerability modifiers for building characteristics for each vulnerability class.

[Modified from Spence and Foulser-Piggott (2014, table 4). —, not applicable]

Table 4.    Proposed vulnerability modifiers for building characteristics for each vulnerability class.

Secondary feature	Class A	Class B	Class C	Class D	Class E	Class F
Primary feature: plan or vertical irregularity
Open ground floor	−0.50	−0.50	−0.50	−0.50	−0.50	−0.50
Other vertical irregularity	−0.25	−0.25	−0.25	−0.13	0	0
Horizontal irregularity	−0.25	−0.25	−0.25	−0.13	0	0
Primary feature: poor condition (including previous earthquake damage)
Poor condition (including previous earthquake damage)	−0.25	−0.25	−0.25	−0.125	0	0
Primary feature: quality of construction
Country group 3	−0.9	−0.2	−0.7	−0.6	−0.6	0
Country group 4	−1.5	−0.4	−1.2	−1.0	−1.1	0
Primary feature: position in the block (masonry only)
Middle of terrace or block	0.1	0	0	—	—	—
End or corner of terrace or block	−0.3	−0.2	−0.2	—	—	—
Primary feature: strengthening after construction (masonry)
Metal tie-rods present	0.4	0.5	0.5	—	—	—
Primary feature: height of buildings
Low rise (1–3 floors)	0	0	0.3	0.3	—	—
High rise (8 or more floors)	0	0	−0.2	−0.2	—	—
Primary feature: influence of horizontal structure (masonry only)
Rigid	1.6	1.1	—	—	—	—
Flexible	—	—	−1.1	—	—	—

Such efforts, recounted by Robin Spence at the workshop, were described in several proceedings’ papers and reports. Foulser-Piggott and Spence (2013a, b) reviewed field experience using EMS–98 in an initial phase of a project to modify the current version of EMS–98 for intensity levels greater than VI, to make the scale more internationally applicable. They determined that EMS–98 (and its explanatory documentation) would need revision primarily with respect to building damage grades and their descriptions that are more representative of worldwide building types currently missing from existing EMS–98 vulnerability classes.

Spence and Foulser-Piggott (2014; “The International Macroseismic Scale—Extending EMS-98 for Global Application”) took several important steps forward toward an IMS. The authors extended their work by examining dominant building types in the six global regions and assessing their relative vulnerabilities, including the probable failure modes, and reviewing the process of assignment of vulnerability classes to building types in each region. They also added preliminary drawings of the damage characteristics of timber, reinforced concrete, and concrete buildings.

They further presented a series of recommendations for modifications to the EMS–98 document and detailed the additional steps needed and validation required to expand EMS–98 beyond Europe. The most recent effort on their part was a draft of “International Macroseismic Scale Part III: Building Typology, Vulnerability, and Damage (Building Guide)” prepared in 2015. Schwarz and others (2015) further contributed to expanding EMS–98 with vulnerability assessments and damage descriptions for reinforced concrete frame structures following the EMS principles.

After the Vienna congress on recent advances in earthquake engineering and structural dynamics (VEESD 2013), a second IMS Working Group 2013–14 meeting convened in Cambridge, England, in November 2013, and a third meeting convened in Potsdam, Germany, in June 2014. Financial support was provided in part by the Willis Research Network (up to September 2013) and the Global Earthquake Model (GEM) Foundation from March 2014. The IMS Working Group 2013–14 agreed upon principles to govern the development of the new scale:

• • Intensity is considered as a classification of the severity of ground shaking based on the observed effects in a limited area.

• • The internal consistency with EMS–98 must be preserved.

• • The scale must remain robust and simple to use.

Based on a review of the international experience of using EMS–98 for 15 years (in many places outside Europe), the following principal amendments to EMS–98 were proposed for the structure of IMS–2014:

• • Divide the documentation into three parts: core scale, guidelines, and building plus vulnerability guide.

• • The core scale and guidelines will be mostly as in the existing EMS–98 booklet, but with some updating and revisions. These will include discussion of web-based intensity assignment procedures.

• • The building and vulnerability guide will contain an inventory of principal worldwide building types, description of the vulnerability issues for each type, photographic illustrations of examples of the principal building types, and common failure modes. Partly web based, it will be a dynamic document with scope for continuous development.

As reported at the Second European Conference on Earthquake Engineering and Seismology in Istanbul, Turkey, in October 2014, the tasks completed in 2014 were (1) the draft IMS building guide, which contained the general principles and for each separate building type, comprising (i) general description, (ii) regional variations, (iii) the expected vulnerability and vulnerability modifiers (including earthquake-resistant design (ERD), (iv) damage grade descriptions, and (v) limitations; and (2) a draft revised vulnerability table. In their draft IMS building guide, Spence and Foulser-Piggott (2015) described the proposed vulnerability modifiers (table 4) as follows:

• • Features of a building characteristic responsible for modification to the most likely vulnerability level. These are considered to include quality and workmanship, floor and roof type, state of preservation, regularity, position in the block, and strengthening.

• • The effect of each modifier on the expected vulnerability class. These are summarized using one of the following three levels of modification:

  • o Increase or decrease vulnerability by one class—modifier affects vulnerability and may increase or decrease the vulnerability by one class or more.

  • o Increase or decrease vulnerability by less than one class—modifier affects vulnerability but in conjunction with other modifiers may increase or decrease the vulnerability by less than one class.

  • o Minor increase or decrease in vulnerability—although the modifier affects vulnerability it is unlikely to cause a shift in the vulnerability class.

• • Vulnerability shifts resulting from each vulnerability modifier are independent of one another, and in principle are additive.

Table 4 was discussed within the IMS group in 2015 and was not fully agreed upon. The engineers were concerned that the outcome may be variable and that the basic elements for EMS and IMS remain the graphical elements used for indicating most likely vulnerability class, probable range(s), range(s) of less probable, and exceptional classes.

At the workshop, Robin Spence noted how the guidelines could be improved with better descriptions on how regional factors affect the seismic vulnerability of buildings. In the EMS–98, a circle defines the most likely vulnerability class for each building type (figs. 1 and 5) and factors affecting seismic vulnerability; for example, consideration of level of workmanship can increase or decrease the vulnerability class of a building. Although the most likely vulnerability class obtained for a particular building type can be varied according to several factors as described in the EMS–98 document, building damage data for particular building types in different locations could be used to identify issues that have arisen as the building stock has changed over time. For example, some types of masonry buildings in a country or region may be more modern and therefore have a lower most likely vulnerability, or they may have routine issues related to poor workmanship that may increase their vulnerability.

Yet, for any building type, the building distribution across vulnerability classes could be investigated by examining regional damage data. That is, by considering factors such as the quality of workmanship, one can reduce or increase (modify) the basic vulnerability. Importantly, Spence and Foulser-Piggott drafted a table of proposed modifications by the IMS Working Group to the existing EMS–98 vulnerability table summarized in table 5. Some of the factors affecting vulnerability assignment defined in the current EMS–98 would also benefit from the addition of descriptive diagrams, such as presented in Spence and Foulser-Piggott (2014). Table 5 was the last discussed working draft of the IMS group, and substantive details need to be worked out further.

Table 5.    

Essential recommendations made at the International Macroseismic Scale Working Group meeting in Potsdam, Germany, in January 2015.

[No., number; EMS–98, European Macroseismic Scale of 1998; ERD, earthquake-resistant design; RC, reinforced concrete; WG, Working Group; GEM, Global Earthquake Model; RPF, (unknown); RS, Robin Spence; RM, Roger Musson; IMS, International Macroseismic Scale]

Table 5.    Essential recommendations made at the International Macroseismic Scale Working Group meeting in Potsdam, Germany, in January 2015.

No.	Essential recommendation
No. 5	Illustrations of damage grades in core scale will be updated and possibly extended to timber and steel buildings.
No. 12	The example photographs of EMS–98 will be extended trying to cover more structural types, damage grades, damage quantities, and regions.
No. 13	Steel types should be subdivided in two types in function of ERD similar as the existing RC frame and RC wall types. One option would be a separate type for mixed steel-masonry structures.
No. 14	RC types are kept unchanged.
No. 15	Timber types should be developed consistent with the new steel types including nonframe structures.
No. 17	The vulnerability table in part 3 must be consistent with the vulnerability in part 1; in particular, the link between the ranges in part 3 and the range in part 1 must be clearly explained.
No. 18	The masonry types are kept unchanged. One option would be a separate type for mixed masonry-concrete structures.
No. 20	The WG appreciates the contribution made by GEM to support the development of the draft building guide and the internet intensity guidelines and thanks RPF, RS, and RM for their contribution to the continuing development of the IMS.

The IMS 2014 group also identified tasks not completed in 2014, including (1) a guide to web-based intensity assignments, (2) damage grade illustrations, and (3) the compilation of a revised set of damage photographs. They also identified future activity in 2014 that recommended review by an external review panel, editing and revision, preparation of a web-based version of the building guide, publication of draft text (ESC meeting in Prague, 2015), a period of trial application, and lastly, a final version. Robin Spence noted that none of these last steps have to date been taken, but they remain essential before IMS can become operational at an international level.

International Macroseismic Scale–2015 and 16th World Conference on Earthquake Engineering

Another productive collaboration toward an IMS occurred prior to and at the 16th World Conference on Earthquake Engineering (16WCEE) in Santiago, Chile, in 2017. As described by Jochen Schwarz and Thomas Wenk at the Powell Center, original members of the EMS–98 Working Group convened an IMS Working Group in 2015 (IMS–2015) and held a special session at the 16WCEE. The World Conference on Earthquake Engineering is usually held under the auspices of The International Association for Earthquake Engineering. The 16WCEE special session 96 (SS96), “Evolution of EMS–98: Building Types, Vulnerability Classes and Damage Grade Description toward an International Macroseismic Scale,” was organized by Jochen Schwarz, Thomas Wenk, and Gottfried Grünthal.

At the Powell Center IMS Working Group meeting, Jochen Schwarz and Thomas Wenk summarized the contributions toward an IMS from the IMS Working Group, active from 2013 to 2015, as stated in its final decisions of January 23, 2015. Their recommendations are as follows.

Additional Building Types

• • Add dual systems:

  • o Reinforced concrete wall frame dual systems

  • o Reinforced concrete-masonry dual systems

  • o Reinforced concrete frames with masonry infills

• • Add precast concrete buildings, subdivide adobe building types, and add more timber building types

Refinement of the Vulnerability Table

• • Refinements of reinforced concrete frame and reinforced concrete wall structures by adding a distinction by stories

• • Refinements of steel structures by adding a distinction by ERD level

• • Timber structures should be divided into different categories

• • Refinements of the vulnerability class range by number of floors or by building height

• • Improvements of most likely values for masonry and reinforced concrete buildings

Descriptions of Damage Grades

• • Improvements of damage grade descriptions for reinforced concrete frames with shear failure

• • The classification “few,” “many,” or “most” should be extended to all damage grade descriptions

• • Currently, this is mainly used for masonry (for example, damage grade 3 for masonry: “large and extensive cracks in most walls”)

• • SS96: “Recommendations for Descriptions of Damage Grades”

• • Additional damage grade descriptions for building types with well-known seismic response and structural performance

• • Additional damage grade descriptions for nonstructural elements are important, even though they were not mentioned in SS96

Recommendations for Simplifying the Field Work

• • Compilation of a database with predominant building types for each region

• • Use of geographical information system (GIS) databases of building stock of earthquake-affected areas to simplify the field work and the calculation of quantities and intensities

At its final meeting in Potsdam, Germany, on January 23, 2015, the IMS Working Group (Roxane Foulser-Piggott, Gottfried Grünthal, Roger Musson, Jochen Schwarz, Robin Spence, and Thomas Wenk) summarized its findings in 20 IMS Working Group decisions. These 20 decisions replaced minutes of the Working Group meetings and were grouped in three categories: essential recommendations, editorial recommendations, and desirable recommendations.

The latest effort toward an IMS was a paper at the 16WCEE by Abrahamczyk and others (2017), titled “WHE Reports as a Complementary Database Towards the Development of an International Macroseismic Scale.” The Earthquake Engineering Research Institute’s (EERI) World Housing Encyclopedia (WHE) is a comprehensive database covering the national variations of structural systems for most building types in earthquake-affected regions worldwide (fig. 6).

Abrahamczyk and others (2017) first describe updates of the WHE to give a more harmonized assignment of the most likely vulnerability class (and their ranges) to the respective structure types. Figures 7 and 8 show the vulnerability classes for reinforced concrete and masonry, respectively, expanding on the EMS–98 table.

Many WHE authors were also asked to indicate the quality of their data, state their confidence level in determining the vulnerability class, and describe the information that their assignment was based upon. This editorial qualification supports the direct link of the reports in transforming EMS–98 into an IMS (Abrahamczyk and others, 2017). Another aspect of the paper that bears directly on IMS considerations are the maps and summary of the regional variations of the building types, and their efforts to correlate knowable building characteristics (for example, age) with seismic vulnerability features (in other words, structural layout, elements, connections, floor, roof type, and so forth) to the expected IMS vulnerability class for each regional variant (Abrahamczyk and others, 2017). Following the updated EMS vulnerability table, regional particularities are expressed in their most likely vulnerability class, along with the ranges of probable and less probable (in other words, exceptional) cases (table 6; modified from Abrahamczyk and others, 2017).

Table 6.    

Comparison of most likely vulnerability assignments for all building types.

[Modified from Abrahamczyk and others (2017). —, not applicable; A, adobe; b, block walls; m, mud walls; CM, confined masonry; RM, reinforced masonry; RC, reinforced concrete; FD, dual system; W, structural wall; S, steel; Fg, frame designed for gravity loads only; URM, unreinforced masonry; SW, structural wall; P, precast concrete; Fd, frame designed for seismic effects; re, rammed earth; mw, mud walls with timber elements; SM, stone masonry; rs, rubble stone; T, timber]

Table 6.    Comparison of most likely vulnerability assignments for all building types.

Country	Vulnerability class
	A	B	C	D	E	F
Latin America
Argentina	A-b	A-m	—	CM	CM	—
Chile	A-b	RM	—	RM	CM	RC-FD, RC-SW, S
Colombia	—	RC-Fg, URM	—	RC-SW, RM	—	—
Cuba	—	CM*	S, RC-P, RC-Fd, RC-P, S	RC-Fd	—	—
Guatemala	A-b, A-re	—	—	CM	—	—
Peru	A-mw	—	—	A-m*, CM	—	—
Europe and Central Asia
Greece	SM-rs	—	RC-Fd	—	RC-FD	—
India	A-m, SM-rs	URM	URM, RC-Fg	A-b*, T	—	T
Iran	A-b, SM-rs	A-m	CM, S	S	—	—
Italy	SM-rs	A-m, SM-rs	RC-Fg	—	—	—
Kyrgyzstan	A-m	URM, RC-P*	CM, RC-P	—	RC-P	T
Pakistan	A-b, SM-rs, URM*	T	RC-Fd	—	—	—
Romania	RC-Fg*	URM	URM, RC-Fg*	RC-Fd, RC-P	RC-SW	—
Slovenia	SM-rs	—	URM	CM	T	—
Russia	—	URM	URM	RC-P, T	T	—
Nepal	SM-rs	—	URM, RC-Fg	—	—	—

^*

Some reports show a quite different “most likely vulnerability.”

As noted by Grünthal, EMS–98 was authorized to be translated in the Cahiers du Centre in Luxembourg into French, Spanish, Chinese, and Italian. Likewise, the core scale and (or) short versions are available in 30 languages. In this sense, EMS–98 is effectively an international scale. However, Foulser-Piggott and Spence (2013a, b) and Abrahamczyk and others (2017) identified several important considerations needed for facilitating the expansion of EMS–98 to additional countries that contained structures not explicitly considered in EMS–98 vulnerability class table. The IMS–2015 Working Group’s SS96 recommendations for “Additional Building Types” and their “20 Decisions” provide an essential platform for moving forward on an IMS.

Modified Mercalli Intensity in the United States

The MMI used in the United States after 1931 was the scale of Wood and Neumann (1931), with 1956 modifications by Richter (1958).

Musson and Cecić (2002, p. 7–8) provide the best guidance on the complicated and not-particularly-well-documented evolution of MMI:

Musson and Cecić (2002, p. 7–8) also provide the most complete description of the MMI Scale at that time. Later, USGS assignments of MMI evolved, yet the specific policy changes were never formally documented and are still difficult to follow. As amended by Stover and Coffman (1993; fig. 9), the USGS no longer uses ground changes (landsliding, liquefaction, faulting, buckling, and so forth) as indicators. Criteria added since have been detailed by Dewey and others (1995) after the 1994 Northridge, California, earthquake, and Dewey and others (2002) following the 2001 Nisqually, Washington, earthquake. Important changes include the convention that MMI IX is only infrequently assigned in the United States. Whereas MMI X is allowed, in principle, no recent events have been assigned MMI over IX and only two MMI IX values were assigned for the Loma Prieta and Northridge earthquakes, although shaking was intense (Dewey and others, 1995).

The USGS has continued to make modifications to MMI assignment procedures to accommodate changing building inventories (refer to Dewey and others, 2002). Nevertheless, seemingly minor changes of important indicators over time warrant reconsideration. For example, old-style unreinforced masonry chimneys are no longer common in highly active regions in the west, because of prior damage and replacement or retrofit, and to improved building practices. Yet damage to masonry chimneys remains a key indicator of MMI VII and VIII in the United States (for example, Dewey and others, 2002). Likewise, the MMI Scale does not explicitly account for damage to prefabricated, elevated mobile homes. Such homes are now a more ubiquitous indicator of damaging levels of shaking.

Until 2002, most USGS macroseismic data were primarily obtained from postal questionnaires and were supplemented by press accounts and engineering reports for important earthquakes. Postal canvassing was discontinued in 2002 and replaced by DYFI. Currently, two types of macroseismic intensities are used in the United States: a traditional MMI, which the USGS has used routinely since 1931, and the Community Internet Intensity (CII), developed for handling observations contributed to DYFI (Dengler and Dewey, 1998; Wald and others,1999; Wald and others, 2011). For practical purposes, such as use in loss or hazard studies, the two types of intensities should be interchangeable. The CII procedure was designed with the intention that CII intensities would agree on average with traditional USGS MMI intensities (Wald and others, 1999), and detailed comparisons between MMI and CII have confirmed that the two are indeed very similar (Dewey and others, 1995, 2002). Nonetheless, we are still concerned about the quality of higher intensity assignments made with the DYFI system for the reasons mentioned earlier in this paper. More critically, traditional macroseismic assignments have not been routinely made for recent events by the USGS since the retirement of M. Hopper and J. Dewey.

The Working Group noted that in the United States, higher intensities are assigned by expert opinion rather than quantitatively as in EMS–98. Thus, a better description of standard practices is warranted, yet a more likely strategy would be to reconstitute intensity assignment practice and policy in the United States by first adopting EMS–98, and then fully documenting (and illustrating) the procedures for the revised scale for use in the United States and New Zealand. As described later, this would likely entail damage data gathering by engineering and loss assessment teams who gather such postearthquake data, but not necessarily for the purpose of assigning intensities.

Thus, in addition to the lack of dedicated macroseismic professionals, an evolving scale, and fewer obvious indicators (for example, chimney damage), new strategies are needed to assign higher (VII and greater) intensities in the United States beyond the traditional approach.

Modified Mercalli Intensity in New Zealand

The strategy for MMI assignments in New Zealand was updated most recently by Dowrick and others (2008), building on earlier revisions by Dowrick (1996). Details of the New Zealand MMI Scale of 2007 are provided in appendix 2 of Dowrick and others (2008). Dowrick and others (2008) accommodate some of the advanced features of EMS–98 for New Zealand, adopting additional structure types (for example, six construction types of varying vulnerability, dominated by masonry). The New Zealand MMI Scale of 2007 adopted other aspects of EMS–98, including adapting “few,” “many,” and “most” ranges for specific damage grades. Vulnerability was mainly linked to building age and code changes: up to the mid-1930s, 1970, and 1980; however, it did not fully approach EMS–98’s vulnerability classes nor damage degrees. The authors also added more quantitative analyses of damage to chimneys as observed during New Zealand earthquakes. However, the New Zealand MMI Scale of 2007 is more detailed and quantitative than that currently used by the USGS in the United States, so it is simpler to use but less sophisticated than EMS–98.

As mentioned earlier, much attention was given to environment damage (ground failure by landslides or liquefaction), and that has been discounted as described by recent authors and varying intensity scales, particularly EMS–98 (Grünthal and others, 1998), Dewey’s descriptions of the use of MMI in the United States (Dewey and others 1995, 2002), and Hough’s many analyses of historical earthquakes (for example, Hough, 2014).

Goded and others (2018) describe New Zealand’s current use of internet questionnaires for macroseismic intensity assignments. In 2004, GeoNet (New Zealand’s national geological hazards monitoring service, https://www.geonet.org.nz/) implemented an internet-based questionnaire (“Felt Classic”) together with an algorithm to automatically assign intensity values on New Zealand’s MMI Scale (to each felt report captured from the questionnaire).

As reported by Goded and others (2018, 2021), their “Felt Classic” and “Felt Detailed” questions are similar to the traditional ones that had been used for the decades prior to 2004. “Felt Detailed” consists of 40 questions divided into 10 sections: (1) general questions on the earthquake, (2) earthquake experience, (3) earthquake effects, (4) building information, (5) building damage effects, (6) neighborhood effects, (7) tsunami evacuation, (8) tsunami information, (9) information about earthquakes, and (10) demographic information (refer to appendix 1 in Goded and others [2021] for the complete “Felt Detailed” questionnaire). Recent updates in “Felt Detailed” include some questions on injuries, deletion of section 9, and a method to skip questions when not relevant (for example, skipping all sections on building damage or tsunami if the respondent indicated there was none of these). The New Zealand method to obtain community intensity values uses an expert-based approach developed by the Italian Istituto Nazionale di Geofisica e Vulcanologia (INGV; Sbarra and others 2010; Tosi and others 2015) and is based on a score distribution for each answer to the questions in the survey fully described by Moratalla and others (2021).

In New Zealand, values assigned as described earlier are limited to intensity VIII. As noted prior, at MMI VII and above, buildings can suffer considerable damage and the assignment of intensity values involves an engineering analysis of the building’s damage grade and building type. Hence, in New Zealand, the same limiting factors for intensity assignments as in the United States also apply: crowdsourced observations are insufficient for intensities that require structural and damage assessment, and the traditional MMI strategy is not only prohibitive in terms of available resources and necessary expertise, but also because the MMI Scale is not as quantitative and rigorous as EMS–98. Yet, many of the buildings in both nations, particularly the wide variety of regionally specific wood, masonry, and reinforced concrete buildings, have not been described in terms of EMS–98 vulnerability classes nor damage grades. Similarly to the United States, traditional macroseismic intensity assignments following field inspections have not been carried out since the retirement of macroseismologists David Dowrick and Gaye Downes from GNS Science, prior to 2010.

As a sidenote, in addition to those traditionally used for MMI assignments, the additional questions above provide demographic and tsunami content that may prove useful. In the United States, the USGS is implementing the Goltz and others (2022) questionnaire following its DYFI system questionnaire to allow users to answer (optional) additional questions concerning demographics, their behavior, and their experience with any Earthquake Early Warning they might have received.

On the Use of Ground Failure for Macroseismic Intensity Assignment

The Powell IMS Working Group concurred that ground failure or other so-called seismogeological effects should not be considered when using EMS–98 in New Zealand and the United States. Avoiding the use of such effects is consistent with the recommendations in EMS–98.

Vogt and others (1994) detailed the considerations made by the 1998 European Seismological Commission Working Group when they began upgrading the MSK Intensity Scale. They concluded that seismogeological observations could not be considered the same as effects on humans, objects, and buildings, and recommended only using them as a consistency check for intensity assessment. They provided a comprehensive table where the experimental relations between seismogeological effects and intensity degrees—assessed by means of other effects—are presented (Vogt and others, 1994).

Following those earlier recommendations, EMS–98 advises that “one should not take into consideration damage caused by earthquake-related phenomena other than the actual strong shaking. Such phenomena include * * * landslides, slope failure, and liquefaction” (Grünthal and others, 1998, p. 24). Detailed justifications regarding the authors’ concerns about considering seismogeological effects are described in EMS–98 Section 7 “Effects on natural surroundings,” for example: “in many cases, seismogeological effects occur in such a way that they cannot easily be quantified to the same degree as other observations can. This means that there is always a problem in estimating intensity in an unpopulated area; at best a range of intensities can be given. This is regrettable, but it is better to admit this restriction than to assign intensities which are too unreliable to be useful” (Grünthal and others, 1998, p. 24).

In EMS–98, “Relations of seismogeological effects to intensity degrees” (table 7-1 of Grünthal and others, 1998, p. 97) shows that the ranges of intensities over which each effect can occur illustrates the problems that arise in dealing with these effects for assigning or even assisting in assigning intensities. In contrast to the EMS–98 avoidance of ground failure and other secondary effects, the Environmental Seismic Intensity Scale (Michetti and others, 2007) uses such observations as a main feature of the scale. Environmental Seismic Intensity was developed to account for hazards because of landsliding and liquefaction, reasoning that seismic hazard assessment, or SHA, would benefit from a comprehensive consideration of all earthquake-related effects. There also is potentially more consistency with the use of such indicators for historical earthquakes (for example, Serva, 2019). However, the use of ground failure complicates the relationship between intensity and shaking parameters because ground failure entails unobserved vulnerability of the ground that fails, precluding such failure as a direct indicator of shaking. So, rather than consistently use historical assignments that included seismogeological effects, many authors avoid earlier assigned intensities that included them and use care in general with historical assignments given the scanty, often unrepresentative data that were typically used (for example, Hough, 2014).

We note that in their revision of the MMI Scale for use in New Zealand, Dowrick and others (2008) gave considerable attention to ground failure, a trend opposite of that in EMS–98 and in the United States (Dewey and others, 1995, 2002). In fact, Dowrick and others (2008, p. 196) noted that “rainfall before an earthquake can substantially increase the number and size of landslides caused by the earthquake. This increases the scatter in the relationship between intensity and landslides which creates a source of variability in intensities assigned from landslides only, a variability that is seldom present in intensity criteria based on damage to the built environment,” yet then proceeded to study and apply a lesson of landslide distributions for use in future MMI assignments. It would be beneficial to swing the pendulum back in the direction of EMS–98 when fully applying EMS–98 strategies in New Zealand. As far as the authors know, no MMI assignments have been carried out using ground failure data since the publication of Dowrick and others (2008). For example, high MMI levels were assigned to the February 22, 2011, moment magnitude (M) 6.2 Christchurch earthquake by Goded and others (2019) only using building damage information, and not ground failure data. The MMI maps were subsequently compared to landslide and liquefaction observations.

Identifying Common United States and New Zealand Structures Missing in the European Macroseismic Scale of 1998

Spence and Foulser-Piggott (2015) identified several building types as not being adequately represented in EMS–98. They did not specifically address those types missing in the United States and New Zealand, yet their tables support additional focus on timber, reinforced concrete, steel, and masonry structures that are expected to show varying vulnerabilities and thus differing damage characteristics than those in EMS–98. EMS–98 was also limited to single structure types both for all steel and all timber structures without consideration of ERD. Grünthal and others (1998) and Spence and Foulser-Piggott (2014) also noted that steel assignments were very limited at the time of EMS–98 production because of a paucity of available data from European studies. Similarly, wood building types were limited because they are not that common in Europe, yet a large proportion of residential structures are timber in the United States, New Zealand, and Japan.

Assigning Vulnerability Classes and Damage Grades

The group discussed the importance of generating separate diagrams representing damage grades for additional building types found in both the United States and New Zealand. Suitable photographs of each building type and damage grade would need to be acquired using the WHE, ATC–20, EMS–98, and potentially other damage studies. Resources would need to be obtained for new architectural drawings as necessary, and options for internships and contracting to do so were reviewed.

Validation

The need for validating new or modified building type vulnerability classes and damage grades against existing EMS–98 intensities was discussed and deemed to be a challenging endeavor. One strategy discussed using a consistency check that considered assigning EMS–98 intensities for past United States and New Zealand earthquakes in areas where two (or more) building types are collocated and could thus be used to assign intensities independently.

A second strategy for validation would leverage relationships between intensity and ground motion using locations where well-documented juxtapositions between recent building damage surveys and strong ground motion recording are available. With due care, one could examine the performance of existing ground motion to intensity conversion equations for MMI in terms of applicability to new EMS–98 assignments. This would be based on areas with both damage information, which can be used to assign EMS–98 values and ShakeMaps (Wald and others, 1999), or ShakeMapNZ (Horspool and others, 2015), which provide MMI values based on ground motions. The levels of ShakeMap intensities can then be compared to those obtained using the EMS–98 Scale, and the possible relationships between MMI, IMS, and ground motion to intensity conversion equations can be explored to improve internal consistencies.

Using an expert elicitation approach to assigning appropriate vulnerability classes to new structure types also was discussed at the workshop. Several of the participants have had some success using structured expert elicitations for estimating structural collapse fragilities (for example, Jaiswal and others, 2014). In particular, Jaiswal and others (2014) used Cooke’s method (Cooke, 1991), which is based on a performance-based scoring system that reflects the combined measure of an expert’s informativeness about needed parameters in the problem area under consideration, and their ability to provide statistically accurate uncertainties associated with their assessments. However, we recognize that the EMS editors recommend following section “2.5 Remarks on Introducing New Building Types” in EMS–98 (Grünthal and others, 1998) to maintain the foundational ideas leading to the existing vulnerability table.

It will also be valuable to compare any independent, previously assigned EMS–98 intensities to those made considering the additional building types in the United States. Such data may be available from earlier work done by the EMS core group by applying EMS principles to the 1994 Northridge earthquake data, among others.

Available Building Damage Data in the United States and New Zealand

At the October 2022 IMS Workshop, Keith Porter noted the three general desired characteristics of building types of interest for macroseismic observations. They should be plentiful, and neither very fragile nor very rugged.

They are plentiful owing to both knowledge and value. To assign a macroseismic intensity to a place after an earthquake, one must be able to find buildings of the types described here, with whatever damage and whatever attributes matter to assigning a vulnerability class; this is the knowledge reason. Next, the scale should improve people’s lives by helping them to manage the seismic risk to the buildings in which they work, play, learn, and live. The larger a fraction of the building stock this scale applies to, the more it will help improve people’s lives; this is the value reason.

Building types for macroseismic observations should be neither very fragile nor very rugged because building types at either extreme provide little knowledge. At one extreme, a very fragile building type would suffer great damage at very low levels of shaking and only tell scientists that an earthquake has occurred, and that observation would be redundant and uninformative. At the other extreme, a very rugged building type might suffer little if any damage in most earthquakes, which is also uninformative.

These three desired characteristics can be at odds if most of the buildings are very rugged or very fragile. So, the foregoing rules merely help to prioritize which building types are characterized first.

Postevent Damage Assessments in the United States

At the October 2022 IMS Workshop, Keith Porter also discussed the possibility of using existing ATC–20 building tagging as a source of data and a potential choice of building taxonomy for structures in the United States. Most earthquake-damaged buildings in the United States get evaluated for safety using the ATC–20 methodology. The ATC–20 data are either a one-page rapid evaluation form (fig. 10) or a two-page detailed evaluation form (fig. 11A–B). The data contain location, building type, a description including occupancy, five or more evaluation criteria, and a building safety evaluation usually called a tag color: red (unsafe), yellow (limited entry), or green (inspected).

If one must choose between the two forms for EMS–98, the rapid form is likely the better choice, owing to the quantity of observations and sampling bias. More data per observation (the detailed form) seems better, but only if one can get the additional data. Approximately 99 percent of buildings evaluated after earthquakes with ATC–20 get rapid evaluations. The 1 percent that get detailed evaluations tend to be mostly critical facilities and a few yellow-tagged buildings with ambiguous damage. Both factors introduce sampling bias.

Eguchi and others (1997) compiled a database of 117,000 buildings that were tagged after the 1994 Northridge earthquake. The authors constructed the database so that the ATC–20 rapid and detailed evaluation forms used to tag those buildings could be transcribed and saved. However, that database suffered from two fatal flaws. First, because it contained personally identifiable information, the California Governor’s Office of Emergency Services repeatedly declined to make it publicly available for analysis, even though the personally identifiable information could be removed.

Second, the effort of manually transcribing paper forms proved too demanding for the cities that were compiling them. They tended to save only the address and the tag color. The crucial data needed here were mostly lost: building type and the evaluation criteria used to assign the tag color. The same proved true 20 years later in the 2014 South Napa, California, earthquake when Keith Porter acquired Napa’s unpublished electronic data, despite the data having the same sort of personally identifiable information. Again, it contained addresses, tag color, and in many cases a single text field describing damage, but no information about building type or record of the evaluation criteria.

Both those databases reflect the loss of valuable, perishable information. Applied Technology Council–20 data can in principle inform EMS–98, but only if the knowledge of building types and evaluation criteria survive the loss of the paper forms.

For instance, “ATC–20-3: Case Studies in Rapid Postearthquake Safety Evaluation of Buildings” (ATC, 2005) addressed in detail the key structures that could constitute the initial list of U.S. structures missing from the EMS–98 vulnerability table. Applied Technology Council–20–3 covers the following:

• • Wood frame single-family dwellings

• • Multiunit wood frame dwellings

• • Unreinforced masonry buildings

• • Older concrete buildings

• • Tilt-up buildings

• • Older steel frame buildings

• • Modern and difficult-to-classify buildings

Of note, ATC–20-3 also has a chapter “Damage That Is Difficult to Detect” (chap. 14, ATC, 2005), which may be of interest if the application of nonvisible and nonstructural damage is considered as a supplement to standard damage assessments in EMS–98.

The advantage of using ATC–20 is in using the observable damage measures and mapping to damage grades in EMS–98. However, at the workshop, it was noted that whereas the ATC–20 “Detailed Evaluation Safety Assessment Form” (fig. 11A–B) would be potentially adequate for damage grade determination per structure, the summary color-coded tagging results alone were inadequate for such assignments. Some key disconnects between tagging colors and damage include the fact that red tags can be assigned for undamaged structures nearby or in the collapse shadow of nearby damaged structures. Moreover, the three-level system for tagging (green, yellow, red) cannot be easily mapped to standard EMS–98 damage grades, unless they are subdivided in some sensible way, such as dividing red tags into red-tagged buildings with collapse (say, damage grade 5) and those without (damage grade 4), and yellow tags with some greater degree of damage (damage grade 3) or lesser (damage grade 2). Such an approach is proposed here and illustrated for three building types in the United States. The additional information (collapsed or not, degree of damage in the yellow-tagged buildings) requires knowledge of the evaluation criteria.

Another data source, ATC–38 (ATC, 2000) offers a protocol to collect a variety of building attributes and damage observations at all the buildings within a given distance of a strong-motion instrument. It has been applied twice, once after the 1994 Northridge earthquake and once after the 2014 South Napa earthquake. Data from both events have survived: 530 records from the 1994 Northridge earthquake and 73 records from the 2014 South Napa earthquake. The ATC–38 data can provide rich information about building construction, ATC–20 tag color, and several descriptions of the damage. It is costly to apply, however, and is unavailable for most earthquakes.

Postearthquake Field Surveys and Macroseismic Intensity Assignments

We expect that field inspectors—typically either civil engineers, architects, or those trained on the subtleties of ATC–20 (ATC, 2005) or other inspection standards—could contribute their data for automatic intensity assignments consistent with EMS–98. One of our authors, Ayse Hortacsu, with the Applied Technology Council (ATC), has led the development of more than 35 guidelines, tools, and project reports. Ayse Hortacsu served as the project manager for the development of “FEMA P-154” (Federal Emergency Management Agency [FEMA], 2002) “Rapid Visual Screening of Buildings for Potential Seismic Hazards” and the adaptation of the ATC–20-1 “Postearthquake Safety Evaluation Procedures” (ATC, 2005) for the country of Bhutan, and is currently leading the update of the ATC–20-1 “Field Manual” for the United States.

The Powell Center IMS Working group proposes that most aspects of these ongoing inspections, with minor modifications, could be used to help assign EMS–98-like intensities with proper guidance and documentation on vulnerability class, damage grades, and statistical sampling. As mentioned, these need only be done in the more limited high-intensity areas where damage is prevalent and thus where DYFI assignments are not well constrained. As described in the section “Quantifying Uncertainties,” uncertainty assignments of each macroseismic dataset (or observation) would allow the use of DYFI up to MMI VII, and then taper off their influence in ShakeMap by downweighting higher intensity DYFI data in comparison to engineering-based assessments; the latter would get full weighting where and when they are provided, allowing for the best possible quantitative representation of the shaking distribution for any earthquake.

We also plan to engage colleagues at FEMA, for whom multitudes of postearthquake preliminary damage assessments are done for the purpose of loss disbursement, the data from which could readily be utilized for intensity assignments with the right protocol and tools. Necessary protocols include sampling strategies for where to collect building damage data (similar to the approach in ATC–38 to evaluate all buildings within a 1000-foot radius of a seismic sensor) and standardizing which damage characteristics to collect (for example, Loos, Lallemant, and Wald, 2022).

The Powell Center IMS Working Group thus aims for better strategies for such assignments by tapping into existing postearthquake building inspection (tagging) protocols such that engineers could effectively assign intensities as a byproduct of their surveys. As an example to emulate, in Italy, the recent collaboration between the INGV and the National Fire Department facilitated the exchange of macroseismic data in real time. Reports from National Fire Department technicians, who immediately carry out technical inspections of buildings affected by the earthquake, are planned to be integrated with the existing data survey system by the INGV group Quick Earthquake Survey Team, or QUEST. A review of existing approaches to collect macroseismic intensity data on the ground, such as Italy’s example, is beneficial.

As part of this second task, we will also further make the case for expanded data collection protocols, using imagery-based reconnaissance data into an organized standard for automatically assigning intensities. We have experience in this realm, and we have identified that the main barrier to utilizing such remotely sensed data directly is the lack of protocol for turning critical remotely sensed observations into quantifiable and useful data. The protocols for assigning EMS–98 in its revised form (to be developed herein) can anticipate these important uses—working with rapid imagery was not an option when the original developers of EMS–98 developed the scale in the late 1990s.

As is evident through the multiple entities involved in postearthquake damage data collection (emergency managers, engineers, civil protection, and others), several building assessment teams visit affected areas to collect data, sometimes overlapping data or with varying building damage scales (such as ATC–20 or EMS–98). An added benefit of leveraging existing mechanisms for damage data collection for macroseismic intensity assessment is that it could lead to more coordinated and synthesized efforts for building damage data collection, ideally by local inspectors and for official applications (for example, Loos, Lallemant, and Wald, 2022). Although reaching this goal globally is ambitious, developing strategies in the United States and New Zealand to coordinate building damage data collection between reconnaissance, local officials, and organizations like the USGS is a useful step.

Subsequent Workshops

A roadblock to the organization of the type of collaborative effort needed to further IMS development has been challenges in gathering important players and setting aside time to conduct the necessary work in an intimate-enough setting that all partners involved can focus on the goals of the project. The Powell Center provided a means for collaborative effort, such that important players could gather and discuss and refine priorities, define action items, and ultimately begin to implement and build a consensus-based IMS.

Agreement of the first workshop participants was to (1) organize the second Powell Center in the fall of 2023 (the IMS working group funded proposal included two international workshops, roughly a year apart) and (2) convene a small Working Group meeting of the primary EMS–98 participants in the interim, in Germany, in the summer of 2023.

Applied Technology Council–158 Project, Summer 2023

The Applied Technology Council (ATC) is conducting the ATC–158 project with funding by a USGS external grant in fiscal year 2022 with the scope to augment EMS–98 with regional building classes and damage grades for New Zealand. This work is planned to continue during the last quarter of 2022 and first quarter of 2023 and the results will be documented in a final report.

Web-based Assignments of Lower Intensities

From the analysis of numerous “Hai sentito il terremoto?” web-based macroseismic questionnaires, INGV proposed improvements to the EMS Scale for effects on people and objects to better constrain lower intensities, which could be included in the IMS Scale. Recent studies enabled floor and building effects to be quantified, demonstrating that intensities experienced in buildings are affected by the floors (Sbarra and others, 2012; Sbarra and others, 2015). Specifically, the maximum variation in intensity between the highest and lowest floors was determined to be a half degree of the Mercalli-Cancani-Sieberg (MCS) Scale.

This is lower than that expected by Grünthal and others (1998, p. 29), which prescribed “reducing the assigned intensity by one degree for [buildings with] so many floors.” Moreover, Sbarra and others (2014) reported that people’s physical situation (sleeping, at rest, and in motion) has more weight on earthquake perception than observer’s location (higher floors, lower floors, and outdoors). Based on their results, as a first approximation, Sbarra and others (2014) proposed to change the wording “indoors” with “at rest” and the word “outdoors” with “in motion” in the description of the EMS–98 Scale.

The quantification of effects on humans and objects is key to reducing the subjective component in macroseismic intensity assessment (Sbarra and others, 2020) at lower intensities. Analysis of the frequency associated with each diagnostic effect by Tosi and others (2017) also suggests omitting (or at least critically revising) the oscillation of hanging objects and liquids as diagnostics, to ensure a magnitude-independent scale.

Although updating the collection or processing procedures over time can be beneficial as expressed in Tosi and others’ (2015) quotation above, updates over time can lead to a discontinuity in datasets, which presents potential conflicts with one of the benefits of macroseismic data: its long-time window into the world of earthquake occurrence and effects. Some innovative strategies have emerged to provide continuity and modernization, and one example is INGV’s approach of assigning both MCS and EMS–98 values for internet-based assignments (Tosi and others, 2015). The USGS initially correlated DYFI with MMI in its initial development (Dengler and Dewey, 1998; Wald and others, 1999), but a revision of practices regarding an EMS–98 implementation is not planned without due consideration of continuity of assignments.

Quantifying Uncertainties

For real-time ShakeMap operations, dense, low-cost accelerometer deployments are anticipated to continue (the Community Seismic Network, for example; Filippitzis and others, 2021), and such data can be included in ShakeMaps using existing techniques for converting peak ground motion data to instrumentally derived intensities (Worden and others, 2020). As with uncertain macroseismic data, lower quality class C (micro electro-mechanical sensors, or MEMS, for example) could still be utilized in ShakeMap as long as each station is accompanied by uncertainty measures (Worden and others, 2018).

A critical aspect of our overall strategy is our aim to add uncertainty (sigma) assignments to each macroseismic dataset based on its source. For example, DYFI data sigmas are quantitively described by the number of user responses in each area (Worden and others, 2012). The Powell Center Working Group can assign expert-opinion default sigma values to other traditional datasets and establish protocols for assigning them to our new (nontraditional) approaches mentioned above. Why are uncertainties so important? The ShakeMap system, which utilizes both seismic stations and macroseismic data, geospatially interprets these data using a multivariate normal distribution (Worden and others, 2018) that explicitly requires data uncertainty to properly characterize the complex shaking pattern that drives earthquake response and loss estimates. Proper interpolation requires weighting instrumental and macroseismic data appropriately, according to both their relative distance (via spatial semivariograms) and their inherent sigmas in the interpolation algorithm (a nugget effect).

The DYFI procedure has proven fully sufficient for collecting intensity data at low and moderate intensities, but at and above DYFI intensity VII, the DYIF intensities become more uncertain (Quitoriano and Wald, 2022). Although uncertain, in the immediate postevent aftermath, internet intensities can still be valuable if their uncertainties are quantified.

For any data, ShakeMap weighting allows the use of DYFI up to MMI VII, and then taper off their influence in ShakeMap by downweighting DYFI data in comparison to engineering-based assessments; the latter will get full weighting where and when they are provided, allowing for the best possible representation of the shaking distribution for any earthquake.

Damage Costs and Loss Ratios

To make the IMS Scale applicable for loss and other insurance applications, a valuable yet not-ready-for-use aim with EMS–98 damage grades would be to explicitly quantify loss ratios associated with each damage grade (with commensurate ranges or uncertainties). The workshop participants noted that it would be extremely valuable to review and summarize current practice to define loss ratios and to propose a new set of loss ratios and uncertainties for each damage grade and building type.

Accommodating Damage to Additional or Nontraditional Structures (to EMS–98)

Some discussion ensued at the workshop concerning additional types of structures not considered in EMS–98 and nonstructural damage. We summarize the issues and how these elements might be used and updated in EMS–98.