Cover Image
close this bookSevere Tropical Storms Preparation and Response - Case Study Text (DHA/UNDRO - DMTP - UNDP, 1991, 58 p.)
close this folderPart Two: Disaster and Response
View the documentE. Detection and Warning of the Storm
View the documentF. Characteristics of the Threatened Area
View the documentG. Actions at the Provincial and Local Level in Response to Warning
View the documentH. Primary Impact
View the documentI. Secondary Effects
View the documentJ. Shortcomings in Management of the Response

E. Detection and Warning of the Storm

The Warning Network

The Republic of Salacca was widely reported to have one of the better meteorological departments in the region. It had an extensive network of traditional observation posts which reported surface and upper level observations round the clock throughout the year, using high frequency (HF) radio telephone. It also was able to draw upon geostationary meteorological satellite imagery for the region, and high resolution pictures from low orbit NOAA satellites, using terminals at its headquarters. The headquarters received colour digitised weather radar data by dedicated land-line from seven radar stations strategically located along the coastline and in the hinterland. The headquarters was linked into the Global Telecommunications System of the World Meteorological Organization’s World Weather Watch, using both satellite and cables, and had additional telex links with the main regional forecast centre 2000 km to the Southwest. There were additional HF radio links to airports throughout the region.

Staff were well selected and well trained. Specialist training was provided through World Meteorological Organization programmes, with additional bilateral support in some specialties. However, an on-going problem for the department was the low level of civil service salaries. Almost all the staff were forced to supplement their income in some way, and many of the professional staff had interests in a range of small businesses (including office equipment, fish-farming, computer software, and cosmetics, to name just some). Inevitably, many of those at work found it difficult to concentrate full-time on their jobs, although the level of professional skill remained high.

Detection of the Storm, and Issue of Warnings

A low pressure circulation was detected about 330 km east of Salacca early in the morning of September 30th. Satellite pictures showed an unstructured cloud mass, and on the basis of this data the forecasters in the national Meteorological department estimated that the low pressure area would intensify and move to the west. Five hours later, at around midday, the forecasting unit issued a warning advice, predicting widespread rain and flash flooding in the central provinces. Early the following morning (October 1st), at 3.30 am. the national forecasting centre issued a second warning advice, repeated at 11 am. This said that the low pressure area had intensified into an active depression, and was moving West North West. It forecast heavy rain, flash floods, and very rough seas.

A fourth warning was issued on the same day (October 1st) at 6.00 pm. It gave a likely impact point for the storm, but no estimate of wind speeds. That evening, the position of the depression shifted, and it started to move north at around 6 km/hr, intensifying over the next twenty hours.

The next warning was issued late in the afternoon of October 2nd. It reported that the depression had developed into a tropical storm moving northwest. Winds in the centre were reaching 70 km/hr. At this time it was about 230 km east of the coastline. The warning repeated that rough seas and flash floods could be expected, but gave no estimate of possible landfall. That evening, the storm was given the name Zelda, from the standard list issued by the Regional WMO committee.

The next warning was not issued by the Meteorological department for a further 18 hours, at around 11 am on the morning of October 3rd. It gave the storm’s position, 180 km from the coastline, moving northwest at 9 km/hr. The warning stated that maximum winds at the centre were now about 110 km/hr, with 10 metre waves. Possible landfall was specified as somewhere on a stretch of coastline 220 km long. Heavy rain and flash floods were again forecast. At around midday, satellite photographs showed the development of an “eye” in the centre of the storm, and the meteorologists estimated that the maximum winds would increase over the next 24 hours to at least 130 km/hr within 40 km of the centre.

Early on October 3rd, the weather radar on the coast picked up the storm and tracked it continuously thereafter. The seventh warning was issued from the Meteorological Office at midday on 3rd October. It said that the storm had intensified into a Cyclone, and was moving West North West at 8 km/hr. Maximum winds of 135 km/hr were expected. However, the shipping forecast issued that afternoon still specified wind speeds in the Cyclone area at 30 to 50 km/hr.

The next warning was delayed. The reasons for this are still unclear. The warning was not issued until 5 am on the morning of 4th October, when the centre of the storm was only 40 km from the coastline near the town of Grenora. Landfall was expected at 8 am. The warning message stated that strong winds and heavy rain were expected within 60 km of the centre, and that trees, telephone lines and power lines were likely to be damaged. A sudden rise in sea level, and flash floods, were both mentioned, but no further details were given. The same morning, one hour later, the national TV weather bulletins, prepared by the Meteorological Department early that morning, were still forecasting 25 to 50 km/hr winds, and moderate to rough seas.

F. Characteristics of the Threatened Area

The Provinces of Akutan and Kylinia have a combined population of 4.9 million people. The economy is based on petrochemical products (centred on the city of Sotorino), textiles, agricultural products (particularly tree crops), fishing, and tourism. A major new industrial complex is under construction at Freeport, 30 km North of Sotorino, which also has a new deep-water container port, a natural gas terminal, and a refining complex.

Topographically, the area consists of a narrow coastal alluvial plain, backed by steep hills sloping up to mountains. The coastline is mainly sand and mudflats. Numerous small rivers cross the plain. The main route south to Sotorino crosses three major road bridges, and twenty smaller ones. The rail route is poorly developed, running about twenty km further inland.

There are many small fishing villages on the coast, most connected to the main highway by metalled roads. Inland, in the rural areas, a variety of different tree crops are grown. Large estates intermix with smallholdings. Farms on higher ground mostly produce either livestock or grain.

The main provincial centres are Sotorino and Morenia.

G. Actions at the Provincial and Local Level in Response to Warning

In reality, very little action was taken at the Provincial and local level in response to the warnings. Some efforts were made to activate the Provincial emergency operating centres in the main towns along the coastline. Hospitals and clinics were placed on alert by a member of the Ministry of Health emergencies team who had watched TV weather broadcasts and taken the initiative to call in at the Meteorology office forecasting centre. Staff at hospitals and clinics had been instructed to move essential equipment to protected areas, and to check radio equipment, generators, and fuel stocks. Informally, hospital administrators began restricting admissions to emergency cases only.

The Armed Forces, alerted by Air Force meteorologists, began redeploying fighter and transport aircraft from two bases in Akutan Province to a base in the North. Two Army helicopter units were alerted and put on standby, and three special forces communications teams were moved with their equipment to a base about 100 km outside the expected impact zone. Following an agreed plan, three Air Force helicopters were tasked to transport joint government/UN-Disaster Management Team (DMT)/Red Cross damage assessment teams, formed after the last detailed review of emergency preparedness. A further two commercial helicopters were chartered for the use of a joint WHO/Ministry of Health assessment and surveillance team.

There were no detailed plans setting out a clear policy for actions to be taken during the various stages of the warning period. In the absence of these, Provincial Governors faced a number of decisions, all involving substantial disruption, and most likely to incur expenditures far higher than their statutory discretionary powers. From their perspective, these decisions included:

The content and timing of public warnings. There were no standing instructions on the content of warning messages which could be broadcast on local radio stations. There was also particular concern in some areas that tourism would be severely disrupted unnecessarily.

Decisions on evacuation. Uncertainty about where (if at all) the storm would cross the coast made any consideration of evacuation highly difficult. The population in the forecast possible impact area (over 300 km of coastline, and at least sixty miles inland) was over four million people. There were no detailed evacuation plans. Some Governors felt that a population caught on the roads might be more at risk than those at home.

Decisions on whom to give priority for evacuation. In particular few government staff had any knowledge of the risks of storm surge.

Decisions on where to evacuate people to. The main choice was to move people to schools, and churches. However, few of these had been surveyed and designated, and their were no plans to receive evacuees.

Decisions on protection of government facilities. Movement of records and strengthening of buildings would involve substantial extra expenditure, and would lead to considerable disruption of normal activity. Without authorization from central government, there were no allocated funds to cover the purchase of timber and plastic sheeting, hire workers, or purchase additional generators and supplies of fuel.

Decisions on mobilizing local government staff. Only a proportion of staff in local offices reported operationally to the Governor. Politically, his position to order a major redeployment of staff and resources without reference to central government was not strong.

Failures in the Warning Process

The immediate pre-disaster period was characterised by a general failure of the warning process. Broadcast warning bulletins simply did not reach a majority of the population in the area which was potentially at risk. Many did not listen anyway. Those that did were generally unable to interpret the messages. These failed on three accounts:

First, they did not, in general, describe the likely impact in terms listeners could understand (for example, that most roofs would be at least partially blown off; that winds would make it impossible to move around outside well before the height of the storm; that storm surges could reach two metres or more, last for hours, and carry fast-flowing debris as large as trees; that flood waters could reach three metres or more in places, and persist for three or four days; that wind-driven horizontal rain would penetrate almost any unprotected building; that trees, roofing sheets, and other debris would become flying missiles; and so on).

Second, they did not relate terms such as “alert”, “watch” and “warning” to actions people should take.

Third, they did not specify useful measures people could take. Examples included evacuating named low-lying coastal areas, and other areas susceptible to deep flooding; avoiding moving to steep hillsides in certain areas; at home, securing loose items; stowing away glass items; boarding windows; reinforcing roofs; preparing extra food and water (many will need sufficient food and water for five days spent on a roof); wrapping warm clothes and other personal items in polythene sheeting; and preparing mats or other coverings to crouch under during the storm.

It was suggested later by one international economist that the inadequacy of warning messages led directly or indirectly to at least 200 million US dollars worth of avoidable damage to personal property. Both the Meteorological Department and the broadcasting networks immediately challenged this assertion.

The need for improvised preparations also highlighted widespread prior failures in mitigation. Several hundred thousand people inhabited low lying areas near the coast. Many thousands, in fact, lived in houses built on piles beyond the coastal low-water mark. Housing construction was in general lamentably inappropriate for high-wind resistance, and no effective building codes were enforced in the area. Most of the public buildings in the risk area were highly cyclone-prone, and many evacuees would (it later transpired) also be at substantial risk in schools or churches.

Overall, one of the worst aspects of the warning failure, in human terms, was the inadequacy of warning for fishermen and other seafarers. Although the widespread use of small, hand-held Citizen Band radios by small boat owners had (when combined with rebroadcasting from larger vessels) increased the coverage of direct warning messages, the system as a whole was still dependent on the accuracy and reliability of the meteorological department’s warnings. In this case, the text of the messages issued by that department failed to convey the urgency of the situation effectively to the broadcast control centre run for fishermen and coastal seafarers by the Ministry of Trade and Industry. This, combined with the inadequacy of the general weather forecast which was broadcast (often after further editing) by the commercial stations, meant that most fishermen expected a storm they could ride out at sea in traditional sheltered bays and headlands.

H. Primary Impact

Even before landfall, the Cyclone had killed most of those who died - more than three thousand people. Its winds and high waves destroyed hundreds of fishing boats, many of them small, three or four person vessels, which were sheltering in the lee of islands, or had been beached on low shores. Bodies of fishermen continued to be washed ashore for several weeks, and most were never properly identified.

Cyclone Zelda slammed ashore into the town of Grenora at 8.30 on the morning of 4th October. Winds of 30 to 50 km per hour had been building up since the previous evening, and there had been continuous heavy rain for the past 24 hours. But no one expected the force of the impact. Most of the population were in their houses. The next ten hours were a shrieking, soaking, battering confusion of flying debris, falling trees, driving rain, disintegrating houses, breaking glass, and rising flood, as people clutched their children, crouched under carpets, fled across spaces ripped by flying iron sheeting, and clawed their way into neighbouring houses in search of respite. Half way through the afternoon the passage of the eye of the storm brought a still, humid, awesome quiet. Some people, believing themselves safe, carried injured relatives into the streets, searched for the missing, or tried to salvage possessions. Half an hour later the opposite wall of the eye rolled across the town, whipping up the debris of a thousand broken houses and cutting down those still outside. Only at dusk did the wind falter, and then darkness covered streets awash with floodwater, carrying tile, metal, and dented up-turned cars, wooden beams, branches, and paper everywhere.

The Cyclone moved on, inland, deroofing most houses, destroying many buildings completely, and flooding some areas to a depth of three metres or more. Most structures on a path one hundred kilometres wide were rendered uninhabitable. Almost all roads in the area were blocked by flooding and fallen trees. All towns in the area lost power - a population of three million people was without light that evening.

The death toll on land by the evening of the 4th October was approximately three hundred. One hundred and eighty of these had been caught in the storm surge in two villages close to the town of Grenora, as houses built over the beachfront collapsed, and as a group of workers fled across a sand-bar. Another forty were killed in the town of Grenora itself, mainly by the collapse of roofs into public buildings, by flying metal roof-sheeting, by drowning, and falling trees. The rest died in house collapses, flash flooding, landslips, automobiles and buses, and in attempts to wade to higher ground.

By dawn next morning the extent of flooding, and the damage to buildings was obvious, both from the air and from the ground. Damage to housing was almost universal. There was widespread loss of corrugated metal sheeting, loss of asphalt shingles, loss of the roof structure, blown out windows and doors, and collapsed timber and concrete blockwall buildings. Specific aspects of local domestic building practice contributed to additional damage, including the method of fixing wooden rafters in concrete lintels, and the increasing tendency to build low-pitched roofs with overhanging gables. Many of the weaker buildings had simply collapsed into a pile of wooden sticks and broken sheeting.

Damage to government buildings was also extensive. Again, roof structures commonly failed. There was also serious damage to interiors, caused by driving rain and flooding. Government offices in the main towns of Morenia and Galenna were largely unusable.

Damage to public utilities had significant knock-on effects immediately. A key element was the interdependence of these systems. Power needs communications, water needs power, hospitals need water, restoring utilities needs an open road network, and so on. Most of the communications towers and masts failed (microwave, HF, and VHF). These were uniformly under-engineered for high winds, poorly anchored, and on inadequate foundations. Corrosion of unprotected steel structures further contributed to structural failures. The power and electronic equipment associated with microwave towers was severely damaged first by horizontally driving rain, and often also by loss of windows and roofs and the impact of flying debris. Telephone exchanges in the area were severely damaged by water and flying debris. Some were flooded. Rain damage to switchgear for alternative power supplies (together with inadequate maintenance in some cases) prevented rapid recovery of the less damaged facilities. The entire trunk telephone network was inoperative throughout the region.

Roofing blown off of most housing


Main power lines were badly damaged


Overground lines of both telephone and electricity transmission systems failed at numerous points. A major contributory factor was the failure of wooden poles weakened by rot and insect damage. Electricity transmission was especially vulnerable where long spans crossed deep valleys.

Most of the main hospitals experienced failure of roof fixing bolts, hooks, and washers. Flat, metal sheeted roof contributed to vulnerability, with tearing of thin sheet at fixings. Large glass panels blew in. Poor maintenance, resulting in corroded fittings, meant that the cladding of several buildings ripped off, causing debris-damage, a number of injuries, and problems moving vehicles up to the buildings.

Many school buildings were badly damaged. Their shape and design contributed to this. They were narrow and long, with flat pitch roofs and gable ends. Some metal-frame schools suffered corrosion of the frame and roof elements and this greatly increased the chance of roof collapse.

Casualties and their Treatment

Injuries were closely related to building failure in many areas. Subsequent analysis showed that the rate of injuries overall was about 4700 per 100,000 exposed people. The rate of serious injuries, requiring immediate hospital attention, was less than 400 per 100,000. The most common injuries were lacerations of the arms and back from flying debris, and cuts from broken glass on the feet. Most injured people were isolated by flood waters, and unable to reach any site for treatment for at least two days. The vast majority of injured walked by themselves to clinics or first aid centres to seek treatment. About 150 injured people were eventually moved in boats to the nearest roads, and thence to major hospitals. Generally, boats often served best as ferries on flooded sections of roads.

At hospitals and clinics, there were shortages mainly of dressing materials, orthopaedic splints, and major pain-killers, but the quantities required for replenishment were described by one hospital manager as “logistically trivial”. It was later estimated by a senior relief official that one well-prepared medical unit with one helicopter (or indeed, three taxi-cabs in part of the region) could have accomplished all of the significant emergency medical replenishment for the medical system.

With the exception of relatively small number of severely disabling head and spinal injuries and some amputations (a total of about forty people), and several hundred people with disabling injuries to joints and muscles, most of the medical caseload had been dealt with by the end of the first week, mainly by hospitals in the area.

Damage caused by Cyclone to National Hospital

J-S. Tyndale-Bisco Photo, UNDRO NEWS 9/10.88

Schools damaged by cyclone


Damage and Needs Assessment

An overall picture of conditions in the affected area was in fact built up quickly, not by some large pre-established reporting system (for there was none) but by a carefully planned, small-scale action by small teams formed several months beforehand. There were two groups of assessors. The first, a joint programme involving staff from the armed forces, the police, the UN Disaster Management Team, and the Red Cross, limited themselves mainly to overflights of the whole area in helicopters, and visits to the main urban centres of damage. Their objective was mainly to get a sense of the scale and scope of the disaster, the boundaries of the affected area, and the limits on access. These teams ended their missions with high morale, but experienced a number of disturbing incidents in which they were forced to decide between moving injured people, women in labour, and other people in urgent need and completing their primary mission of information collection.

The second team attempted to obtain a more scientific assessment of risks and needs throughout the area. An early sample survey of the Cyclone-affected area was co-ordinated by two staff from the Ministry of Health, using four two-person teams, flown by helicopters provided by a multinational company and by a USAID charter. Sampling sites were selected during overflights, and two person teams then flew to centrally located relief centres, to interview individuals about personal losses, village losses, and immediate needs. Conditions and supply requirements were discussed with local relief officials, and all available ill and injured people were examined clinically. Radio links with the police network, together with the excellent personal and working relationships between the police air unit and local ministry officials, enabled the teams to call up police helicopters for casualty evacuation on the few occasions where this was required. Sources of drinking water were tested for salinity. The teams summarised their findings each afternoon, and these summaries were collated, edited, and issued nightly to the Government EOC and to several major donors. A total of eighteen sites were visited during this first survey, enabling the team to generate an estimate of overall mortality, losses of housing, and overall Cyclone-related morbidity. The relative absence of excess levels of communicable disease was noteworthy, providing further evidence to resist the widespread clamour for mass-immunization. Water sources were found to be mostly usable.

Immediate Needs for Most People

The immediate problems facing the population were shelter and food. Those who were marooned outside for several days became very hungry, their condition exacerbated by constant drenching in the continuing rain. Reconstruction of shelter from debris was certainly possible. However, there were no stocks of nails or rope, few tools, and most survivors were extremely tired. The inclination of many people was to gather their remaining possessions, and to leave the area to seek refuge with relatives or friends elsewhere. But some, especially those with only moderately damaged homes, quickly began to patch up damage using salvaged materials. The occasional distributions of plastic sheet and tarpaulins by the Army, government officials, or NGOs, while widely welcomed, mainly served as coverings for salvaged possessions.

Food, however, remained the main preoccupation. Shipments of cooked grain and bread brought in by the Army, and by people from neighbouring provinces, were appreciated by all. But distributions were inevitably confined mainly to the more densely populated areas. To a large extent, people in less damaged houses shared remaining food stocks and cooking fuel with their neighbours. Merchants quickly began to sell salvaged stocks in markets even in the more damaged areas, and prices, though at least double the pre-impact levels, were still within reach of most people, who had managed to protect household stores of cash.

Managing Public Health

There was widespread concern about the quality of water supplies, and some local rumours (incorrect for the most part) of water-borne epidemics. Within the more densely populated urban areas, staff from the Ministry of Health had quickly moved to monitor free chlorine levels and bacterial contamination in public water supplies, following detailed instructions given in a number of seminars and training programmes the previous year. In the city of Morenia, staff of the water utility corporations were so familiar with their systems that once broken domestic connections had been isolated, delivery of water in quantity to stand-pipes was quickly restored. However, the vulnerability of the water system to both electricity failure and debris damage to chlorination apparatus was clear. A number of hasty improvisations had to be made to connect the pumps to diesel generators supplied by the Armed Services. The system had to resort to batch chlorination, using supplies of high test hypochlorite brought in by helicopter. Before this was done, two hundred thousand people in Morenia were without clean drinking water for two days. The need for structural mitigation measures in this part of the system was obvious.

Deteriorating sanitary conditions in a number of schools and churches housing displaced people were causing particular concern to Ministry of Health staff. A major problem was the lack of sufficient water storage to match the numbers living in and around the buildings. Once collapsible water tanks had been brought in, and supplied with water by tanker, the risk of diarrhoeal disease outbreaks declined somewhat, but the problem was never adequately dealt with. Within a week, most of the centres were emptying fast. Health officials focused much of their epidemic surveillance activity on these centres, and on the poorer sections of the main towns. In addition, all hospitals were requested to report the last known location of any patients showing a range of communicable disease symptoms.

To their own surprise, health officials managed to resist pressure for indiscriminate mass cholera immunization from the Minister of Health himself. The combined appeals of the WHO representative, a USAID consultant, and a visiting senior official from the US Public Health Service, together with a well-informed CNN reporter, finally prevailed.

Distribution of Casualties and Damage

Even at this early stage, the relationship between poverty and vulnerability were becoming clear. Injury and death rates were higher in the poorer sections of most towns. Particular occupations, and social groups, had suffered disproportionately. The high death toll among poorer fishermen was noteworthy.

Damage to industrial resources was widespread. Within the relatively narrow corridor of major damage, most of the larger industries involved processing of agricultural products. Outside stores of raw product were mostly lost. Much machinery was damaged by metal roofing sheets blown from the large expanses of roofs. Sixty kilometres to the south, the new Freeport, with its container port, natural gas terminal, and refining complex, had experienced significant damage from 150 km per hour winds. Most of the completed capital facilities were relatively unaffected. However, plant still under construction was much more seriously damaged, and flying debris from these units caused specific, highly expensive damage to critical catalytic units and control valves. Noteworthy also was the level of damage to computerised control systems at the natural gas plant, mainly from wind damage to roofs, combined with inadequate rain protection for electronic equipment. Little had been done beforehand to provide expedient protection, and it was clear that the roof design of buildings containing this equipment was inappropriate. Equally serious in terms of disruption was the level of damage to the temporary buildings in which construction workers lived, and in which key tools and materials were stored, beside the sites. Many of these were completely destroyed. Lost also were the pay and personnel records of the largest construction firm, and the firm providing security services. It quickly became clear that construction would be delayed by several weeks, at least.

I. Secondary Effects

The initial impact and its aftermath had a number of immediate secondary effects throughout the area.

There was a rumour, which swept the national stock exchange on the afternoon of the 5th October, that the new container marshalling port had been seriously contaminated by the breakage of a shipment of used transformer oil, containing polychlorinated biphenyls (PCBs - a highly toxic chemical, which is extremely hard to clean up). This was erroneous; in fact a container was damaged, and it did contain a toxic substance (tetra-ethyl lead - shipped contrary to regulations). But there was little leakage, and a safety team from the oil refinery provided adequate advice and assistance to the local fire brigade. However, the rumour added to the general atmosphere of uncertainty and concern, and the stock exchange index lost 20 percent of its value that afternoon.

More immediately serious in welfare terms was the impact of continuous rain, and flooding of the main route from the major port of Sotorino inland to Suremia - the neighbouring country - which had just experienced a serious influx of refugees. A large shipment of food aid, including bagged wheat, and drums of oil had just arrived and had been offloaded to the quayside three days before the Cyclone made landfall. A large convoy had already been dispatched. The convoy was halted for six days by flooding and landslips on the road north into the mountains. The tarpaulins of the vehicles proved inadequate, and much of the food was drenched. Meanwhile, on the quayside, where bags were stacked unpalleted, and where the available tarpaulins failed to cover the stacks, the wholly unanticipated levels of rain ruined at least half the 15000 tonne wheat shipment.

Flooding of routes proved very disruptive to international road traffic. This was exacerbated by the loss of one bridge on the main international route, at Oketo. It took nearly fourteen days to reopen this route, when the Army built a pontoon bridge.

The most immediate, and in the longer-term one of the most damaging, effects of the initial impact was the loss of future tourist trade. News of the disaster spread quickly within the travel industry, exacerbated by rumours of serious breakdowns in the government’s handling of tourists’ emergency needs in the affected area. The outcome was an immediate flood of cancellations by travel agents, and the abrupt suspension of negotiations for room space during the next tourist season.

Cost of Damage

It would be several months before even a preliminary comprehensive estimate of the costs of damage had been made. However, within a few days aerial surveillance clearly indicated that up to 30 percent of buildings in the Provinces of Akutan and Kylinia had been damaged to some extent, and some 40,000 housing units (mostly low-income families) had been destroyed completely. Preliminary estimates of losses, compiled by the Government, with help from UNDP, World Bank, and OFDA, were an estimated $450 million in public facilities and infrastructure, $350 million in housing, $240 million in agriculture, $130 million in tourism, and $260 million in manufacturing.

J. Shortcomings in Management of the Response

Overall, the response period was not well-managed by the government. To those on the spot, the responses of most officials seemed half-hearted, unplanned, and lacking any clear sense of co-ordinated action in relation to clearly recognised priorities and goals. From the local perspective, during the first week, every transaction between individual families and officialdom seemed to end in confusion, frustration, and finally outright anger. Few in authority seemed to be doing anything useful for ordinary people.

Failures in Emergency Planning

The root causes of this were not so much incompetence or venality, but rather the failure of a system of planning, anticipation, and control. In the detailed, but largely confidential analyses carried out by the Office of National Audit, UNDRO, USAID, and a number of NGOs, breakdowns in the relief response were systematically traced back to failures in preparedness and planning. These in turn stemmed from a complex set of problems related to funding levels, the lack of institutions providing effective training, the low level of education and lack of motivation of many junior ministry staff, the lack of “champions” of emergency planning within the system, and the general unwillingness of most people to consider that disasters on this scale would ever happen. ‘

Throughout the government system, there was little prior understanding of who would do what, where, when, and how in different kinds of emergency. There were no adequate databases of personnel skills, buildings, or, available equipment resources, and nobody with the sustained motivation to either build them, or challenge their absence. Few line ministries had tested standard operating procedures, and the Emergencies Department lacked the political “clout” to enforce compliance. Overall management structures for emergencies were ill-defined in both national and local plans. The plans contained few job descriptions or outlines of responsibilities. They contained no clear instructions on how vital supplies and equipment were to be made available.

The outcome when the Cyclone struck was widespread confusion within government over roles and tasks; authority and access to resources were not commensurate with assigned emergency functions; organizations were far too dependent on others for equipment or information; and preparations for handling the massively increased demand for information were very poorly done.

Communications proved to be the most important constraint. The disaster destroyed much of the regional communications system, and the surviving communications network just did not reach those who could have used it best. Existing communications were clogged by calls to higher authority for permission to act. Senior officials were hopelessly overloaded, whether or not they had any idea what was going on. The lack of protection for the telephone system, and the lack of effective substitute civilian communications rendered central government mostly impotent. Confusion over the availability and capacity of military communications ensured that even the surviving channels were not used effectively. The solutions to these problems were largely political, but little was ever achieved.

The subsequent inquiries pointed in detail to other specific shortcomings which in many cases had led to serious loss of life, or massive economic disruption. The early warning system of the meteorological department was criticised for lack of standard operating procedures, and weaknesses in the content of warning messages. The Secretariat of the National Emergencies Committee was castigated for failing to link various levels of meteorological warnings to standard operating responses by the Committee and by individual departments.

Planning for pre-impact preparations (on receipt of warning) received special criticism in the analyses. No detailed plans had been made within the Ministry of Internal Affairs or the Provincial governments. Only the health authorities had paid much attention to this issue, although the operations directors of several public utilities were beginning to pay more attention to the requirements. Planning for protection of most critical economic resources during the warning period was non-existent.

Weaknesses in Co-ordination and Assessment

The lack of plans and preparations for activating the national emergency operations centre were also a major factor in the confused response. A central civilian co-ordination centre was not operating until 36 hours after the Cyclone made landfall. The Prime Minister, and other senior Ministers found themselves having to take decisions “blind”, without any real sense of what was happening.

The initial post-impact assessment went better than expected operationally (because of close, detailed planning, and much prior consultation), but then experienced major problems in disseminating and sharing assessment data. This again was mainly a function of the lack of planning in other areas (no one knew who needed data, or for what purpose), but interpretation of results was also complicated by the lack of baseline data. A special weakness also was the failure to link assessment information with the improvised search and rescue efforts of the Navy, several army field units, and a national police special unit. In addition, information on route blockages was never passed on to the NGO groups who were attempting to mobilize convoys of relief items. One final problem in assessment stemmed from the non-availability of video and photographic equipment. It was hard for returning teams to convey an overview of conditions. News footage gathered by ENG teams from the national commercial television stations tended to concentrate on areas of worst damage and gave a very biased picture. Nonetheless, it was this that most politicians, civil servants, and NGO officials were initially responding to.

Key Problems in Relief Provision

The provision of immediate relief suffered from many problems, but two seemed to dominate everything: the failure to organise road clearance, and the inability to distribute fairly. The extent to which deep flooding and fallen trees would prevent access by road was not understood by many planners. It took days to clear roads, even after the flood waters subsided. Much of the clearance was done not as a result of government efforts, but by local people attempting to get out of their villages to seek help. Distributions were described by all as “a complete mess”, with local people milling around road junctions, market squares, and government offices, in search of the latest official or private hand-out. A number of NGO convoys were simply looted, as they halted at some broken bridge or tree-strewn highway. Only the Army, which provided cooked food, tarpaulins, and organized manpower mainly to more isolated rural villages, seemed to achieve what they set out to do.

Shortcomings in the Medical Response

The medical response was probably better planned than other official actions. Emergency medical planning at national level had achieved several key goals, including appointment of a qualified and properly trained emergency advisory committee, development of a realistic basic national emergency health plan meeting international standards, an accurate database of resources, and standard operating procedures for disasters incorporated into routine health activity. Above all, planning was integrated with training, and senior officials took this seriously enough to make it work, even at the local level. Noteworthy in the planning was the emphasis on structural mitigation. Plans for new medical facilities were carefully scrutinized for signs of poor resistance to hazard agents. Noteworthy too was the emphasis on communications development (and protection).

Nonetheless, the medical system still displayed a number of serious problems. First, the impact of population displacement and evacuation were underestimated. The practical problems caused by several thousand people concentrated in one improvised centre were simply not appreciated. There was little provision for detailed supervision of these centres. Lack of information on the location of emergency evacuation sites, and lack of staff, made it impossible in many cases to visit to supervise water and sanitation and to confirm rumoured disease outbreaks.

Second, mortuary provision was totally inadequate in the affected area. Many bodies were taken to clinics (themselves damaged) near the coast. Nothing had been done to prepare for identification of corpses, or to arrange for their disposal. Attempts were made to quickly cremate them, but much of the wood from debris proved too wet to burn.

Third, the medical planners did not anticipate the scale of the requirement for sorting incoming medical items at the main airport. Not enough trained pharmacy staff were designated.

Fourth, the need to produce plans for hospital and clinic reconstruction within weeks had not been anticipated in the emergency plans, and a number of significant mitigation opportunities were lost as a result.