Supporting response and recovery in Ecuador after the April 16, 2016 M7.8 earthquake

La versión en español de la publicación del blog se puede encontrar aquí.

On April 16, 2016, a magnitude 7.8 (Mw) earthquake struck the northern coast of Ecuador, causing heavy damage and casualties in a widespread area. This event had a hypocentre at 0.37 N and 79.94 W and a depth of 19 km. At the time of writing of this post, hundreds of buildings have been reported collapsed and more than 3,000 people have been affected. With such a striking impact, it is fundamental to respond efficiently and prepare for the recovery phase. Supported by the SwissRe Foundation, GEM has been collaborating with local partners from Ecuador for several years, within the scope of the South America Risk Assessment (SARA) project. Within this initiative, several datasets and models relevant for earthquake loss estimation have been produced. This post aims at sharing some of these outcomes with the scientific and disaster risk reduction communities, hoping that they might prove useful in the identification of the regions where a higher level of damage is expected, as well as in the strategic planning of the recovery process.

Residential building stock

Ecuador is divided into 25 provinces, 224 counties and 1,024 parishes (INEC 2010). The largest urban settlements are Guayaquil, Quito (the country’s capital), and Cuenca. An exposure model including the number of buildings and economic value (replacement cost), was developed at the at province, canton and parish levels as part of the SARA project. The residential building inventory in the country was developed based on census information and expert judgment. From census data, information regarding the material of the exterior walls, the material of the floor, type of dwelling and type of area (urban or rural) was obtained at the smallest administrative region available (parishes). The distribution of the total number of buildings, the distribution for masonry, reinforced concrete and wooden structures are presented in the figure below. It has been estimated that the residential building stock of Ecuador has about 3.02 million buildings with a total replacement cost of 62.6 billion USD.

Distribution of the number of residential buildings in Ecuador

hereThe residential building stock is characterised by low-rise construction, mainly unreinforced masonry and confined masonry with low ductility. In urban areas, in addition to masonry construction, it is common to find reinforced concrete structures, with RC flat-slab system and infilled frames. On the other hand, in rural areas it is more common to observe wattle and daub (quincha), together with earthen or adobe construction. This model can be downloaded in the csv format here and in the OpenQuake-engine format xml here.

For additional information about the development process of this model please visit the SARA wiki.

Fragility functions

Within the scope of the SARA project, a fragility model was derived for each building class in the region. In this context, several numerical models were generated (in order to account for the building-to-building variability) and a large set of ground motion records were selected and scaled (in order to consider the record-to-record variability). These two components were combined using nonlinear dynamic analysis, and four damage states (slight, moderate, extensive and collapse/complete damage) were considered. The resulting fragility models for unreinforced masonry and non-ductile confined masonry (both with two storeys) are depicted below. According to early reports, these building classes were heavily damaged.

Fragility models for unreinforced masonry and non-ductile confined masonry with 2 storeys for Ecuador

These fragility functions have been verified against a number of past seismic events in South America (e.g. Armenia (Colombia) 1999, Pisco (Peru) 2007, Maule (Chile) 2009), and can be downloaded in the OpenQuake-engine format here.

Ground shaking in the region

For the definition of the ground shaking in the region we recommend leveraging upon the ShakeMaps produced by the Prompt Assessment of Global Earthquakes for Response (PAGER) system developed by the United States Geological Survey (USGS). This system employs best available magnitude, location and source mechanism to improve the finite-fault modelling. Furthermore, it uses real-time strong motion data to constrain the ground shaking throughout the affected region, by correcting the ground shaking with the inter-event component (bias), and reducing the total variability in regions in the vicinity of recording stations. The ShakeMap generated shortly after the seismic event is presented below.

ShakeMap in terms of peak ground acceleration generated by PAGER

These datasets supported by the USGS can be used to generate sets of cross-correlated ground motion fields, which in turn can be used with the OpenQuake-engine to calculate damage or losses. A set of such fields can be downloaded here, and the corresponding mean ground shaking in terms of peak ground acceleration and spectral acceleration at 0.3 seconds is presented below.

Mean ground motion fields in terms of peak ground acceleration (PGA) and spectral acceleration at 0.3 seconds (SA) using ShakeMap data from PAGER.

It is also possible to calculate ground shaking for the purposes of assessing damage and losses using the geometry of the rupture and one (or multiple) ground motion prediction equations from the OpenQuake-engine. The fault geometry proposed by the USGS can be downloaded here in the OpenQuake xml format.

Useful information about this event can also be found in the portal of the Instituto Geofisico of the Escuela Politécnica Nacional. In addition, the GEM Hazard team has also compiled a number of datasets describing the seismicity in the region in a dedicated blog post.

Economic losses, fatalities and damage distribution

For the rapid assessment of human and economic losses, PAGER issues a report every time that an event near populated areas and above a certain magnitude threshold occurs. These reports contain not only information about the hypocentre and magnitude of the event, but also estimates of loss. These loss models employ country-based vulnerability models based on empirical data and global datasets with the spatial distribution of population and capital stock. This indicates the expected order of magnitude of the expected human and economic losses and associated probabilities. The results computed considering this event in Ecuador are presented below.

Estimation of human and economic losses by PAGER (version 3)

In order to provide estimates of building damage in the affected region, we have also used the aforementioned datasets (exposure, fragility and ground motion fields from ShakeMaps) with  the OpenQuake-engine to calculate the expected number of collapses at the various parishes, as presented below. It is important to mention that complete damage represents a damage state beyond which the building will not be recovered due to excessive damage, and not necessarily the explicit collapse of the structure.

Distribution of mean number of buildings with complete damage (left) and fraction of buildings with complete damage.

According to our estimates, the provinces where a greater number of collapses is expected are Manabi, Esmeraldas, Guayas and Los Rios.

Socio-economic datasets

The distribution of social vulnerability within Ecuador is an integral part of managing, planning, and mitigating against the country’s earthquake risk. The figure below depicts the spatial variation of the social vulnerability of Ecuador’s parishes (i.e. parroquias). The social vulnerability scores for each parish provides a comparative assessment within the country’s extent. Comparatively, the parishes mapped in red along the classification continuum exhibit higher levels of social vulnerability. Of special interest and potential management concern is the clustering of moderate-to-high and high levels of social vulnerability along the coastal areas, the Andes and the Amazon regions. Note that those parishes in rural areas experience high levels of social vulnerability; whereas main urban centres and parishes surrounding them show the lowest levels of social vulnerability. These results corroborate the fact that major urban areas have better access to basic needs such as lifelines (water, electricity or sewage systems), education and health services.

Distribution of a social vulnerability index at the third administrative level in Ecuador (parroquias)

To determine some of the underlying factors contributing to the trends outlined in the previous figure, and to further explore the urban-rural bias in the observed social vulnerability; the population, education, access to lifelines and health sub-components of the social vulnerability index have been mapped. Several spatial patterns are noteworthy. For example, high levels of population vulnerability occur in major urban areas (Quito, Ibarra, Ambato, Cuenca) where many conditions contribute to social vulnerability such as population density, population with a disability, age dependence, renters, and female population are most pronounced. Similarly, high social vulnerability is observed in the Amazon region where inherent social conditions experiences by native indigenous population and critical access to basic needs are indicatives of social vulnerability. The education component represented by access to formal education indicates that several parishes located in the coastal, Andean and Amazon regions have high levels of vulnerability.

Population and education sub-components of the social vulnerability

The health subcomponent is based on access to basic healthcare, and it can be observed that parishes surrounding major urban centers present the higher levels of social vulnerability in this component as the population experiences difficulties accessing healthcare services. Lastly, access to water and sanitation reflects high levels of social vulnerability, especially in those parishes in rural areas and in the Amazon region.

Healthcare and access to lifelines sub-components of the social vulnerability

The socio-economic datasets for Ecuador can be downloaded here. The contents of this post will be updated as new information becomes available. For additional information about the datasets, tools and models described herein, please contact us at