supporting response and recovery decisions
Supporting Equitable Disaster Recovery Through Mapping and Integration of Social-Vulnerability Into Rapid Post-Disaster Impact Assessments [current]
Collaborators: Earth Observatory of Singapore (EOS), Nanyang Technological University (lead); Kathmandu Living Labs (KLL); Stanford Urban Resilience Initiative (SURI); Humanitarian OpenStreetMap Team (HOT); World Bank Global Facility for Disaster Risk Reduction (GFDRR) and World Bank Big Data Program; and NASA Jet Propulsion Lab and Advanced Rapid Imaging and Analysis Center (NASA-JPL/ARIA). Funded by the Innovation Fund, through the Global Partnership for Sustainable Development Data and the World Bank's Development Data group.
Several key decisions are made soon after an earthquake that define how housing recovery evolves. These decisions require rapid estimates of the post-disaster impact. But how do we estimate impact? Typical post-earthquake assessments focus on direct economic losses to physical assets (number of houses damaged, total reconstruction cost). However, there's not only disaster-induced damage to the built environment, but also disaster-induced need from the most vulnerable communities. But the way we estimate direct losses shifts the focus to those with the most valuable assets - perhaps not those who are disproportionately affected because they started with less to begin with. Because of this way that we measure impact, disasters (through no bad intent) have the power to indirectly highlight and potentially amplify social inequality. In this work, we will be addressing these ideas by working with Kathmandu Living Labs to study the impacts and recovery since the 2015 Nepal earthquake using impact data collected the year after the earthquake.
This project has two main components which aim to 1) advance methods of rapidly estimating the spatial distribution of post-earthquake building damage and 2) define a metric for measuring impact that incorporates disaster-induced vulnerability.
Rapidly estimating regional building damage with data fusion
After an earthquake, multiple agencies produce different data sources on building damage throughout the post-disaster timeline. This ranges anywhere from damage models and remote-sensing estimates to field measurements of damage. Each of these data sources is represented at different spatial scales (building, pixel, etc.) and uses its own damage measurement system (e.g. damage ratio, numerical scale).
The goal of this work is to develop a framework to fuse damage data sources with these various formats into a more accurate map of damage to be used for post disaster needs assessments (PDNA), recovery aid requests, and planning purposes.
An outline of the framework is included in the “AGU Poster” link below.
Quantifying the relationship between recovery and direct impact metrics
Within the rapid timeframe of post disaster needs assessments (PDNA), social impact information is not collected in a detailed manner, instead left as general characteristics of prevalent socially vulnerable groups. These characteristics include, but are not limited to, gender, income, age, and so forth. However, there has been little empirical evidence that these characteristics actually cause slower recovery, or lower resilience capacity, for communities and households. In fact, important factors that might actually slow down recovery might not be assessed at all, such as geographic isolation, financial ability, etc.
Social scientist, Jamie McCaughey, is leading a follow-on household survey, to assess individual rates of recovery four years after the earthquake. We will then analyze the relation between recovery with pre-existing vulnerability characteristics and post-event impact information.
Platform for one "area-based" crowdsourcing approach
The Uses and Development of Crowdsourced Building Damage Information
Collaborators: Humanitarian OpenStreetMap Team (HOT); World Bank Global Facility for Disaster Risk Reduction (GFDRR) and GIScience at Heidelberg University
The Haiti 2010 earthquake brought about one of the first large-scale initiatives of crowdsourcing for building damage information. Crowdsourcing for rapid damage estimates has the power to inform multiple post-earthquake decisions, where loss estimates are needed at a regional level. Our team at the Stanford Urban Resilience Initiative tested three different methods for crowdsourcing this information, deviating away from the typical "building-by-building" approach to an "area-based" approach to address these information needs.
Report is forthcoming. Click the link below for more information.
Validating inSAR-based Damage Proxy Maps for the 2011 February Earthquake in Christchurch
Collaborators: World Bank Global Facility for Disaster Risk Reduction (GFDRR); NASA Jet Propulsion Lab and Advanced Rapid Imaging and Analysis Center (NASA-JPL/ARIA); and Tonkin + Taylor
After the February 2011 earthquake in Christchurch, New Zealand, NASA-JPL/ARIA produced their first inSAR-based damage proxy map (DPM). Using in-depth building-level field surveys of liquefaction-induced damage, we are in the process of analyzing the validity of these maps. This validation exercise can improve the performance of the DPM for future events.
Work is ongoing and report is forthcoming.