Multifamily residential buildings in earthquake-prone cities across Central Europe, Balkans, and Central Asia have proved to be vulnerable and have revealed several structural deficiencies during previous earthquake events. Earthquakes in these regions have damaged certain types of multifamily buildings beyond repair or have caused complete collapse, resulting in extensive homelessness and many fatalities.
The vulnerability of some of these buildings can be attributed to the mass-produced nature of multifamily residential buildings as well as lack of maintenance, oversight, and disinvestment over many decades. This study investigates earthquake risk of multifamily buildings constructed before 2000 across 27 cities in 20 countries within Europe and Central Asia (ECA) to better understand their behavior and potential losses when subjected to earthquakes, and to inform recommendations for mitigation. The table on the right shows the cities and countries considered within this study. The intended audience for this report includes government officials, policy and decision makers, engineers, researchers and analysts, among other stakeholders and beneficiaries. Two supplementary technical documents provide additional insight into the inputs, methods, and results, and can be found in the Works Cited section (Mott MacDonald 2020a, 2020b).
TYPES OF MULTIFAMILY BUILDINGS IN ECA
The study evaluated seven broad classes of multifamily buildings found in the cities and countries under consideration. The buildings are categorized into classes based on their behavior when subjected to earthquake shaking, and according to the data available across the different cities. Building classes are characterized by specific features that affect their earthquake performance: horizontal and vertical structural elements and the material type, connections, building height, construction quality, potential deterioration, and period of construction. The diagram on the right shows the seven building types and the acronyms used in this study, as well as the relative vulnerability of major subtypes indicated as low, medium, and high. For example, the unreinforced masonry building type has two considered subtypes: unconfined and confined masonry.
A probabilistic seismic risk analysis of multifamily buildings was conducted in order to predict damage and losses due to earthquakes in the 27 cities. The risk analysis used a combination of hazard, exposure, and vulnerability models, and was performed using an open source software called OpenQuake, developed by the Global Earthquake Model (GEM).
Earthquake Hazard Model: The overall earthquake hazard model is comprised of three existing regional models (ESHM13, EMME14, EMCA15) and modifications that are based on two local seismic hazard models (BIGSEES15, BUL18). The regional models are obtained from the Global Earthquake Model GEM2018 global mosaic.
Exposure Model: The building exposure model was developed by analyzing country-specific, cross-country, and structural typology-specific information sources. For each building type, many data sets were evaluated to understand the number of buildings, number of dwellings, and number of occupants, as well as the associated construction costs. The sources used include census data (mainly 2011 and 2014 data), national energy efficiency programs, building inventory reports, past projects on risk assessment, international building databases, postearthquake reports, technical publications, and satellite images. It should be noted that the reliability of exposure data varies from city to city, based on the information that was available. The relative reliability of the exposure data is shown in the exposure reliability matrix (Table 2) in the Methodology section.
Vulnerability Model: The vulnerability models used in the risk analysis are based on existing damage and loss models from various sources. Over 400 damage and loss models were reviewed and the most appropriate ones for each city and building type were selected. The models were used to quantify several earthquake consequences, including: direct financial losses from building damage, permanent relocation of occupants based on building repair time, and fatalities due to building damage. Similar to the exposure data, vulnerability data varies by city.
The relative reliability of vulnerability models is shown in the vulnerability reliability matrix (Table 3) in the Methodology section.