Assessing early warning efforts for Typhoon Haiyan in Leyte

Typhoon Haiyan was the deadliest natural disaster ever recorded in the Philippines. It left a massive trail of death and destruction, leaving thousands dead and millions more affected. Many families were left homeless; power lines were destroyed and communications and water facilities damaged. Fishing boats were wrecked and coconut groves crushed, leaving many families without a livelihood. Total damage and economic losses were estimated at $2 billion.

The Philippines has a well-designed and robust disaster risk management system, with a specialised institution for risk reduction and management through the National Disaster Risk Reduction and Management Council (NDRRMC)/Office of Civil Defense (OCD), and the country is accustomed to fearsome storms, floods and other hazards. Even so, it was not fully prepared for a disaster of the magnitude of Typhoon Haiyan. This article outlines the key findings of an assessment of early warning efforts prior to the typhoon, conducted by Deutsche Gesellschaft fuer Internationale Zusammenarbeit (GIZ) in December 2013. See http://www.preventionweb.net/files/36860_36860gizassessmentofearlywarningyol.pdf.  Publicly available documents and interviews with 41 survivors were analysed under three aspects of early warning: risk knowledge, event detection and communication.

Risk knowledge

Hazard maps contribute to risk knowledge by identifying which homes have to be evacuated and where safe places are located. In the context of tropical cyclones, storms, floods, landslides and storm surges, hazard maps are critical tools for preparedness planning. The Philippines weather agency PAGASA has yet to publish a wind/storm hazard map for the provinces that were affected by Haiyan, although such information is available. The only storm surge hazard map on hand prior to Haiyan, generated by PAGASA, indicated possible flooding across a much smaller area than was in fact affected (interestingly, the Philippines Institute for Volcanology and Seismology (PHIVOLCS) produced a tsunami hazard map showing an inundation area closer to that produced by Haiyan). Nine out of 29 respondents said that they did not know that their houses were located in a potential storm surge area, or assumed that they were not. Almost all of the local government officials interviewed by GIZ claimed that they used the hazard maps to identify who would be evacuated and which evacuation centres were safe. Most evacuation centres were located outside the potentially flooded area shown on the PAGASA map, but since a much larger area was affected many were still flooded. For some rivers in Leyte detailed and verified computer flood models exist, but the PAGASA hazard map has yet to receive such verification. Rain-induced landslide areas are also displayed on a hazard map, but as approximately two-thirds of Leyte are marked as landslide-prone many local residents regard the map, which does not appear to be based on a detailed study of individual sites, as exaggerating the extent of the risk.

Several of the evacuation centres were single-storey buildings, typically schools, and were not strong enough to withstand the force of violent waves, so did not protect evacuees from the storm surges caused by the typhoon. The Office of Civil Defense was tasked with checking on the safety of the evacuation centres, though the fact that many of the people who took refuge in them died suggests that these safety checks were not properly done or not done at all. Essentially, the centres became deathtraps.

Event detection

PAGASA and other national and international agencies noted the emergence of a powerful tropical cyclone several days before Haiyan made landfall in the Philippines, and its path was predicted quite well. PAGASA also forecast heavy rains leading to floods and landslides, as well as a storm surge.

Communication of warnings

The NDRRMC/OCD is responsible for the coordination of early warning to the public based on information from PAGASA. There are several ways for warnings to be communicated to the public.

PAGASA declared the highest alert level for provinces along the path of the typhoon. These warnings were published in the daily bulletins of the NDRRMC and were widely disseminated by the media. President Benigno Aquino appeared on television on 7 November 2013 to highlight the warnings and emphasise the dangers of the typhoon. From interviews GIZ concluded that there was enough time to prepare for the typhoon and evacuate to safe areas before it made landfall. However, in the eyes of the public neither PAGASA nor the NDRRMC conveyed the seriousness of the storm surge, instead issuing a rain warning. Many residents and political decision-makers openly admitted that they were not familiar with the term ‘storm surge’; one interviewee said that, while he did not understand what ‘storm surge’ meant, had the authorities used the terms ‘tsunami’ or ‘tidal wave’ people would have evacuated. Media coverage of the tsunamis in the Indian Ocean in 2004 and in Japan in 2011 had raised public awareness of the dangers of tsunamis but not storm surges, although the effects of both hazards are very similar, even if their causes are not.

Although the authorities issued clear orders to evacuate to residents of Tacloban, some did not understand the nature of a storm surge and its dangers. Others stayed in their homes due to fear of looting, underestimated the height and force of the water or simply laughed off the evacuation order. Forced evacuation was carried out in very few areas. According to interviewees, in many places only half of residents left their homes. Government warnings, including those issued by the OCD and the local government, were not sufficient to persuade people to move to safer areas.

The death rate (i.e. the number of deaths in relation to population) in Guiuan in Eastern Samar, where Haiyan made its first landfall, was lower than in Leyte. This could be because settlements in Guiuan were protected from the full force of the winds by a chain of hills, and because the winds affecting the populated parts of Guiuan were offshore, reducing the danger of a storm surge. No such hills protected settlements in Leyte. From Tolosa to Tacloban the wind was onshore, producing a substantial storm surge in a densely populated area. The highest death tolls in Leyte were in the coastal barangays hit by the full force of the storm surge. Inland areas (Dagami, Jaro, Tabon-tabon, Alang-alang, Santa Fe and Pastrana) experienced the same wind speeds as coastal areas, but there was no storm surge and the average death rate was much lower (nine deaths per 10,000 residents, compared to 180 per 10,000 in Tanauan, Palo and Tacloban). The difference is most likely due to the absence of a storm surge.

Lessons

1. Modify the storm hazard map for land use planning and disaster and emergency management

The official storm hazard map should be modified in light of the experience of Typhoon Haiyan. This modified map will constitute a critical technical input into assessments of the risks to life and property, and should be incorporated into land use planning and disaster and emergency management. The modified map will enable the NDRRMC/OCD, local government and other agencies to provide more accurate information to the public and media on areas prone to storm surge hazards. Local authorities should identify and strictly enforce no-build zones in high-risk areas, as well as ensuring that stipulations in conditional-build zones, such as special reinforcements and building exclusively for business use, not residential, are observed.

2. Strengthen dissemination and communication of early warnings

The lack of effective dissemination and communication of early warnings was a notable weakness in the run-up to the typhoon. Critical information was not clearly communicated to those who needed it, and people were unaware of the severity of the hazard, or did not use the information appropriately. People simply did not understand that their lives were at significant risk. If the scale of the impending danger had been communicated properly, and coastal residents had been evacuated to safer ground, fewer lives would have been lost. Timeliness, clarity and coverage are essential if warning messages are to lead to sound decision-making, successful evacuation and other preparatory measures. The capacities of PAGASA, the NDRRMC/OCD and the local government to improve the accuracy of information and the dissemination and communication of early warnings should be enhanced. The OCD and the Philippines Department of Science and Technology (DOST) should include storm surges in the official warning system, as they do with tsunamis.

3. Proper location and appropriate design of evacuation centres

Many evacuation centres were located in storm surge areas. New centres should be located outside danger zones, and the hazards they are designed for clearly indicated. Evacuation routes, shelter locations and emergency services (hospitals, fire stations) should be designated and publicised, and evacuation routes and centres clearly signposted. Evacuation centres should consider the needs of children, women, the elderly and the disabled. In high-risk areas forced evacuation should be considered. While some of the reasons why people may be reluctant to heed evacuation warnings are beyond the government’s control – fear of looting, for instance – that does not mean that efforts to increase people’s understanding of the hazards they may face, and what they should do in response, should not be made.

Gemma Ocon and Olaf Neussner worked with GIZ in the Philippines at the time of Typhoon Haiyan.