Issue 71 - Article 9

The Dhaka Earthquake Simulation: lessons for planning for large-scale urban disasters

March 27, 2018
Charles Kelly

Recent urban disasters, including the 2017 flooding of Houston and 2015 flooding in Chennai, highlight the challenges faced after disasters in large urban areas. While attention to disasters in large urban areas has increased since the 2010 earthquake affecting Port-au-Prince, developing a comprehensive process for managing disasters which can affect millions of people is still at an early stage. Conventional disaster risk management planning is important in reducing urban disaster risk, but the scale of large urban disasters (affecting urban areas with populations above a million) and the social and physical complexity of urban environments require looking beyond conventional approaches.

This paper summarises an Urban Crises Learning Partnership (UCLP) effort to use a scenario-based simulation to:

  • Provide humanitarian actors in Dhaka with an opportunity to examine preparedness and identify where action can be taken to improve the ability to prepare and respond to an earthquake.
  • Explore response options, focusing on existing social protection and market systems, to provide cash for food security, shelter and water, sanitation and health.
  • Explore the use of simulations to promote learning and capacity-building.

The context

The primary disaster risk for Dhaka, a city of between 15 and 17 million people, stems from the seismic faults that cross under and near it. Past earthquakes have resulted in significant damage and land form changes. Assessments and response plans make for dire reading. A major earthquake is expected to lead to half a million fatalities, 1.2m injured and the loss of up to 80% of housing, severe damage to roads, water, electrical and other critical lifeline systems and the disruption of food supplies and other critical commodities. Much of the earthquake planning to date has focused on the immediate life-saving response. Less effort has focused on how Dhaka can sustain itself and recover following a major earthquake.

The UCLP Dhaka simulation intended to go beyond immediate relief to consider how social protection systems could deliver support, how the economy would function with significant infrastructure damage, and how cash could be used to address critical needs in shelter, WASH and food security. The development of the simulation recognised the role of local government and neighbourhood organisations in addressing needs following a major earthquake. As such, in addition to a range of international and local NGO participants, the simulation involved participants from neighbourhood organisations. Two Dhaka neighbourhoods were used as examples of the challenges which could be faced in responding to a major urban disaster.

Simulation planning

Planning for the simulation began with discussions between UCLP members, HfH Bangladesh and Oxfam Bangladesh several months before the event. UCLP developed a back-ground paper which was shared with stakeholders and revised to incorporate their inputs. Weekly calls were held between the UCLP core team and Dhaka counterparts on the simulation, supplemented by emails and document sharing via DropBox. I was hired as a consultant in late April 2017 to lead the development and execution of the simulation. I produced a range of documents, including a schedule for the preparations, a narrative plan for the event and a detailed session outline. Contact was made with the Shelter and Recovery Cluster, and with WFP Bangladesh on social protection issues.

One issue raised was whether a community assessment field exercise was necessary as part of the three-day simulation. Some in Dhaka felt that many who would be involved in the simulation were already competent in conducting community assessments, and the field exercise was not necessary. After discussions which highlighted the critical role field assessments play in disaster response, and the complexity of simulating such assessments in a closed workshop environment, it was agreed to include the community assessment fieldwork in the simulation. Oxfam Bangladesh proposed using Kobo Kobo is a software package which can be used to collect and process survey data. It is adapted for use on hand-held devices, including smartphones and tablets. for data collection to make the process quicker and less complicated. Oxfam converted the survey form into Kobo with ease, a process which also served to check the questionnaire content and translation. While Kobo was not used for all the surveys, the software considerably facilitated field data collection and processing during the simulation.

The simulation itself took place over three days, between 23 and 25 May, and was attended by close to 60 participants from multiple organisations. Further details on the simulation can be found in the Dhaka City Earthquake Simulation report, forthcoming from UCLP.


The simulation was an opportunity for staff from NGOs and international organisations to explore the challenges involved in providing relief and recovery following a major earthquake in Dhaka. Brainstormed ideas, such as boating water to Dhaka or using houseboats to provide shelter, are worth investigating, particularly given an expectation that two or more years will be needed for recovery to reach full steam, and long-term but still interim solutions will be needed across a range of sectors.

Attention is also needed to understand how and where NGOs and international organisations should expect to intervene, where the private sector should be left to its own devices, and where the government needs to take the lead. It is also important to distinguish between what can be delivered through large-scale (e.g. $100 million) distributions of funds to groups such as retirees and the disabled following a major disaster, and what can be accomplished by NGOs working through informal, small-scale or emergent social protection systems after an earthquake.

International organisations and NGOs may find significant opportunities in the shelter and settlements sector. The simulation highlighted that any effort to provide shelter for earthquake-affected people in Dhaka, a city where 80% of the population are renters (a number in the millions), requires engagement with the government on policy, as well as prac-tical challenges such as damage assessment and the extent of provisional repairs. International organisations and NGOs could focus on providing a combination of material support, technical advice, small loans and support for self-recovery. This mix of assistance will be critical to re-establishing short- and long-term housing and rebuilding neighbourhoods.

Developing a clearer understanding of where NGOs, inter-national organisations and the government can best support recovery, and how to identify and mobilise informal and emergent groups, are core to planning and preparing for a major Dhaka earthquake. Most of the recovery options identified in the simulation require government engagement before the disaster, if only because many post-disaster recovery interventions will be a significant change from current practice. Good ideas on addressing the challenges of a post-earthquake Dhaka were identified, including the expanded use of vending machines to sell water and floating houses, but there was insufficient time to explore them in any depth, and the general sense was that the simulation tried to pack a lot into a short period of time. The simulation opened the door to post-earthquake planning, but much more needs to be done to ensure that results have practical applicability. Even so, this is probably the best that could be expected after three days.

The simulation did not capture the role of emergent groups in responding to an earthquake in Dhaka (emergent groups comprise people who come together following a disaster to provide assistance, forming groups that did not exist beforehand). See E. L. Quarantelli, Emergent Behaviors and Groups in the Crisis Time Periods of Disasters, Preliminary Paper 206, Disaster Research Center, University of Delaware, 1994, Emergent groups contribute significantly to disaster relief and recovery, but as they usually form or become engaged only after a disaster has happened, including them in pre-disaster planning and preparation can be difficult.

Another important challenge identified in the simulation was the need to shift thinking from delivering relief to supporting recovery. Many of the ideas generated focused on what is generally termed relief. There was much less focus on how to transition towards what will be a very long recovery process, and for a disaster of the scale anticipated in Dhaka it will take years for market systems and services to return to normal. Developing an understanding of what is necessary to make disaster-impacted systems work effectively after a major urban disaster could not be covered in the simulation, but merits significant attention.

Future simulations covering markets, social protection, shelter and settlements or similar lifeline systems post-disaster need to include sufficient time to consider the different phases of disaster relief and recovery. Single-focus simulations that work through the impacts of the disaster on each critical system may be more useful in large urban areas than multi-topic simulations. Single-focus simulations allow a more in-depth exploration of relief and recovery needs in a way that considers the complex post-disaster urban environment. As skills and knowledge are gained, individual topics can be combined into simulations which build across sectors.

An example of market mapping developed during the simulation. The market mapping example demonstrates how simulations can identify skills and knowledge developed before the simulation exercise itself, in this case from capacity-building to manage cash relief payments in Bangladesh.
An example of market mapping developed during the simulation. The market mapping example demonstrates how simulations can identify skills and knowledge developed before the simulation exercise itself, in this case from capacity-building to manage cash relief payments in Bangladesh.

Simulation as a learning tool

The Dhaka event demonstrated that simulations can be used for learning and capacity-building, but they are not the same as training sessions or workshops. A simulation places participants in situations where they are required to use skills and knowledge. Simulations commonly provide incomplete information to participants, and then expect participants to work around the information gaps in addressing the often complex problems created as part of the simulation. In this sense, a simulation is a test. If skills and knowledge have not been built before a simulation, participants may mistakenly feel that they have failed the test, but in fact through no real fault of their own.

Simulations can provide an opportunity for learning-by-doing, but this is best linked to the on-site provision of advice, for instance where an advisor stops work during a simulation to discuss what is being done and how it could be done better. The Dhaka simulation didn’t provide time for this, even where it was clearly needed. The simulation did demonstrate participants’ skills and knowledge, for instance in the market mapping process, but in other areas, such as analysis of community assessment data, skills and knowledge, it did not meet expectations. There was no time to stop the simulation to investigate and address these gaps. Some of the gaps may have simply been due to a lack of time to complete assignments. There was also no baseline in terms of participant skills and knowledge, and so no real basis for assessing improvements. The simulation review indicated that many of the participants felt that they had benefited from the event, but using a combination of focus groups and interviews to assess capacities and gaps may have been as effective as a simulation.


The UCLP Dhaka Earthquake simulation provided an opportunity to assess the skills and capacities of a range of stakeholders in dealing with the results of devastating earthquake damage to Dhaka. The simulation identified a need to expand assistance opportunities and avenues beyond shortterm relief, and the challenges of addressing food security, markets, social protection, shelter and WASH in a major urban disaster. Simulation participants, for the most part, were satisfied with the event. However, the simulation was a one-off and was not fully tied into government preparations for the next big earthquake. Its actual impact on response policy and practice was likely limited.

Simulations can be used to learn about capacities and gaps, but they need to be preceded by training to build skills and knowledge. The UCLP simulation was a good start in preparing for extended relief and recovery needs following a major earthquake in Dhaka. Lessons were learned on organising a simulation, and how simulations can best be used. Nonetheless, to change how major disasters will be handled in Dhaka will require many more such events, and the training to match.

Charles Kelly is a disaster management consultant with over 40 years’ international experience. This article presents the views of the author and does not necessarily reflect the views of any other individual or organisation.


Comments are available for logged in members only.