Integrating and geolocating post earthquake building damage surveys: The 7.8 Mw Jama-Pedernales earthquake, Ecuador

Currently, it is fairly widespread to use smartphones or tablets on field surveys to collect and geolocate damage data. However, geolocation is not a straight forward process and may give inaccurate results such as, for example when the size of the object to be surveyed is relatively small or the coverage of the satellite constellation (e.g. GPS) is inadequate due to obstacles and shadows present in urban areas. Moreover, the pressure that surveyors and technicians suffer during and after the impact of a natural hazard may make the whole geolocation process even more difficult. In this paper, we describe a methodology to overcome the issues of inaccurate records in five damage data surveys collected after the 7.8 magnitude earthquake that struck the coast of Ecuador in April 2016, together with the three administrative sources used to interpret the damage. We started off by homogenizing the various states of damage as charted in field and aerial surveys, including satellite imagery. We then resolved geolocation inaccuracies by using a set of algorithms that take into account the spatial context and the size of the building. These algorithms also flag the quality of the sources to ultimately compute a figure of the spatial distribution of the damage suffered by residential buildings, together with harm done to productive and social infrastructure. Without these preliminary proceedings, the geolocation inaccuracies of the damage data surveys would not have allowed for adequate and detailed risk assessment.