1 INTRODUCTION

1.1 The Purpose of the Project
1.2 Terms of Reference
1.3 Natural Hazards in the Caribbean
1.3.1 Earthquakes
1.3.2 Volcanic Activity
1.3.3 Tsunamis
1.3.4 Hurricanes
1.3.5 Torrential Rains
1.3.6 Storm Surge
1.3.7 Landslides

2 INVESTIGATION PROCEDURE

2.1 Briefing
2.2 Field Surveys
2.3 Field Reviews
2.4 Desk Studies

3 RESULTS

3.1 Vulnerability to Hurricanes
3.1.1 General Observations
3.1.2 Aspects of Vulnerability
3.1.3 Summary of Results
3.2 Vulnerability to Torrential Rain
3.2.1 General Observations
3.2.2 Summary of Results
3.3 Vulnerability to Earthquakes
3.3.1 Multi-hazard Design
3.3.2 Differences Between Designing Against Earthquakes and Hurricanes
3.3.3 Aspects of Vulnerability
3.3.4 Summary of Results
3.4 Recommendations
3.5 Classes of Vulnerability

4 GENERAL ISSUES

4.1 Impact of Hurricanes on Caribbean Schools
4.2 Hazards versus Disasters
4.3 Codes and Standards
4.4 The Regulatory Environment
4.5 Public Awareness
4.6 Vulnerability Surveys
4.6.1 Qualitative Assessment
4.6.2 Analytical Evaluation
4.7 Disaster Mitigation after the Event
4.8 Maintenance as a Tool for Mitigation

Appendix I Survey Forms

• Hurricane Vulnerability Assessment
• Seismic Vulnerability Assessment - Part I
• Seismic Vulnerability Assessment - Part II

Appendix II Technical Paper  (only introduction available in web format)

• Hassan, Ahmed F. and Sozen, Mete A., Seismic Vulnerability Assessment of Low-Rise Buildings in Regions with Infrequent Earthquakes, ACI Structural Journal, January - February 1997

Appendix III   Photographs of Selected Buildings Designated as Shelters in Anguilla

Throughout the world, including the Caribbean, natural hazards cause as much damage to educational facilities as they do to buildings of less importance. This is both regrettable and avoidable. Educational facilities deserve special attention because of their roles during the active periods of storms and also as post-disaster assets.

It is traditional for schools to be used as hurricane shelters. It goes without saying, therefore, that the damage and destruction of schools would put the sheltered population at risk during severe storms. Also, damage and destruction of such schools would not facilitate the use of school buildings for temporary housing after hurricanes. The longer-term problem of loss of educational facilities is arguably even more severe. If the children are not at school the parents' work is often adversely affected (in part because of "baby-sitting" problems). There is also the inevitable disruption of the pupils' education.

It is often said that safe buildings may not be affordable, especially in relatively poor developing countries. This is a fallacy. Particularly with respect to hurricane resistance, safe buildings are not only technically feasible but also achievable at very modest cost. This thesis has been tested and confirmed on several occasions over the years.

The Caribbean Development Bank (CDB) proposes to assist its borrowing members in reducing the vulnerability of designated shelters to natural hazards. The initial screening of these designated properties is being carried out in some of the CDB member countries through the auspices of the Organisation of American States (OAS) using CDMP [1] and ECHO [2] funding.

It is recognised that the suitability of properties for use as emergency shelters depends on several factors other than structural safety. Some of these factors are location, size, water storage and supply, sanitary facilities, kitchen facilities, standby power and telecommunications within the facility and externally. Most of these issues are already being addressed by other agencies. The focus of this study is on the physical vulnerability of the built facilities to wind forces, torrential rain and seismic forces. This initial screening will provide the relevant local agencies with some of the input required for selection of properties for the follow-on, CDB-funded projects.

The portions of the overall Terms of Reference (ToR) relevant to the subject of the present report are:

• Preliminary assessment and screening of properties identified by participating governments (Anguilla in this case), for the purpose of classifying the properties with respect to their retrofit needs.

• The consultant will prepare a survey form to be applied by local engineers employed by the participating governments (Anguilla in this case). The form will be applied to up to 20 properties per country (six in the case of Anguilla). The form will contain a section on wind resistance, amended from the WIND-RITE form of IIPLR [3], and an earthquake resistance section, based on the work of Ahmed F Hassan (ACI Journal January-February 1997).

• The consultant will prepare a workplan for the application of the survey forms by the local engineer(s). The local engineer(s) will be assigned by each participating country (Anguilla in this case) to work with the consultant. The consultant will train the selected local engineers in the application of the forms and will periodically monitor the progress and quality in the application of the survey forms.

• The consultant will execute a field visit to each of the participating countries. The field visit will be planned and executed with the participation of the local engineers. Prior to the field visit, the consultant will review the completed forms.

• During these field visits, the consultant will continuously articulate the issues being observed in determining vulnerability; the relative importance of various factors such as location, geometry, materials of construction, detailing; the differences between wind and earthquake vulnerability; etc. For each of the participating countries (Anguilla in this case), the consultant will develop a classification (screening) of all identified properties. The classification will achieve a broad-brush appreciation for the vulnerabilities of the individual buildings to wind and earthquake events and the suitability of the properties for inclusion in a CDB retrofitting project.

As articulated above, an important objective of this project is technology transfer and the broadening of the pool of "disaster mitigation" engineers in the region. Therefore, the active involvement of local engineers (assumed to be from the relevant public works departments) was essential.

The primary natural hazards facing the islands of the Caribbean are earthquakes and hurricanes. Some of the islands are also subjected to instances of volcanic activity. There are also the related hazards of tsunamis and storm surge.

Torrential rains can also, by themselves, be considered natural hazards since they often occur without the concurrent occurrence of hurricanes and sometimes result in disastrous flooding of low-lying areas. Likewise, some landslides are naturally occurring, isolated events.

1.3.1 Earthquakes

All of the Commonwealth Caribbean countries, with the exceptions of Bahamas and Guyana, lie close to the known tectonic plate boundaries. The North American Plate is moving westward with respect to the adjacent Caribbean and South American Plates at a rate of approximately 2 centimetres per year. A moderate level of inter-plate activity is generated along these boundaries. Along the northern margin, including areas in the vicinities of Jamaica and the Virgin Islands, moderate earthquakes of shallow depth are generated. Near the plate boundaries there are also intra-plate earthquakes. In the northern Caribbean these intra-plate earthquakes are caused by internal deformation in a slab of the North American Plate. Concentrations of these earthquakes occur at depths of up to 200 kilometres.

Seismic events in the Eastern Caribbean are principally associated with a subduction zone at the junction of the Caribbean Plate and the North American Plate. The North American Plate dips from east to west beneath the Caribbean Plate along a north-south line just east of the main island arc. This leads to a moderate level of inter-plate seismicity. Superimposed on this is a pattern of intra-plate activity. There is a concentration of such activity in the Leeward Islands where the subduction of the Barracuda Rise imposes additional stresses on both the "subducted" North American Plate and the overriding Caribbean Plate. The earthquakes there are generally shallow. In the region north-west of Trinidad there is another concentration of earthquake activity where the strike of the plate boundary changes direction. These earthquakes are of intermediate depth.

Over the past forty-five years a considerable amount of research has been carried out on the seismicity of the Caribbean by the Seismic Research Unit (SRU) of the University of the West Indies (UWI). The engineering community has been requesting more and more assistance from the SRU in interpreting the fundamental research and developing "code" values for seismic forces for use in structural design. The most recent published work in this field is that of SRU's former head, Dr John Shepherd, now at Lancaster University (England).

The Pan-American Institute of Geography and History (PAIGH) is based in Mexico City. The Geophysical Commission of PAIGH is the executing agency for a major project (funded by IDRC [4] ) for preparing Seismic Hazard Maps for Latin America and the Caribbean and headed by Dr J G Tanner. Dr John B Shepherd participated in this project as the Caribbean specialist.

The final report and mapping from this project indicates that Anguilla lies between the iso-acceleration lines representing 250 gals [5] and 300 gals. These are peak ground accelerations with a probability of non-exceedance of 90% in 50 years. The Caribbean Uniform Building Code (CUBiC), which is incorporated in the Anguilla Building Code [6], does not assign a zone factor (Z) to Anguilla; however the PAIGH information would suggest a zone factor of 0.75. This indicates a serious earthquake hazard for Anguilla.

1.3.2 Volcanic Activity

Several of the islands of the Eastern Caribbean are volcanic in origin. The volcanoes there are considered to be either active or dormant. Anguilla does not exhibit volcanic activity.

1.3.3 Tsunamis

A tsunami (or seismic sea wave or tidal wave) is a series of ocean waves generated by any large-scale, short-duration disturbance of the free surface of the ocean. The majority are related to tectonic displacements associated with earthquakes at plate boundaries. However, tsunamis can also be generated by erupting volcanoes, landslides or underwater explosions. In the open ocean, tsunamis may have wavelengths of up to several hundred miles but heights of less than 1 meter. Because this ratio is so large, tsunamis can go undetected until they approach shallow waters along a coast. Their height as they crash upon the shore mostly depends on the geometry of the submarine topography offshore, but they can be as high as 30 meters.

A tsunami travels at an average velocity of 500 to 600 kilometres per hour rising to a maximum of 800 km/h. Therefore within one hour of a major occurrence at Kick 'em Jenny (just north of Grenada), many of the islands of the Eastern Caribbean will be affected. (J B Shepherd and Martin S Smith [7] propose a realistic scenario where a tsunami resulting from Kick 'em Jenny would reach neighbouring Sint Maarten in 100 minutes having a height of approximately 4.35 metres.) This, however, is outside the scope of this report.

1.3.4 Hurricanes

Much is not known about the storms which occurred in the Caribbean in the years before the advent of Columbus. But, of course, the European did not bring hurricanes to the Caribbean. Indeed the very name is derived from the Mayan storm god Hunraken and the Arawak word hurican, which meant the devil wind. The greatest of all recorded hurricanes occurred from 10th to 18th October 1780. Nearly 20,000 people perished as the storm hit virtually every island from Tobago in the south-east through the Windward and Leeward Islands and across to Hispaniola and Cuba. In the last 60 years in the Caribbean another 20,000 people have lost their lives because of hurricanes.

The Caribbean lies in the North Atlantic Ocean, one of the six main tropical areas of the earth where hurricanes may develop every year. Within the 111 years between 1886 and 1996, approximately 1000 tropical storms have been recorded in the North Atlantic. About half of these attained hurricane strength.

The destructive potential of a hurricane is significant due to high wind speeds and torrential rains that produce flooding and occasional storm surges with heights of several feet above normal sea level.

The pattern in recent times has been a reduction of deaths and injuries (because of better warning systems and other preparedness activities) and an increase in property damage (because of commercially-driven unsuitable building practices and locations).

The Caribbean Uniform Building Code (CUBiC), which is incorporated in the Anguilla Building Code, does not assign a reference pressure to Anguilla. However, the wind-pressure contour map in CUBiC suggests that the reference pressure for Anguilla should be about 0.8 kPa [8]. This is equivalent to a 10-minute sustained wind speed of 36.5 m/s [9] (82 mph) or an equivalent 3-second gust speed of 55 m/s (123 mph). These figures relate to a 50-year return period.

1.3.5 Torrential Rains

Although heavy rains often accompany hurricanes, severe rainfall events resulting in flooding are also, and frequently, associated with troughs and tropical depressions. The risk of flooding is therefore not restricted to, nor more likely to occur, during hurricane events.

Drainage systems and structures in the Caribbean are generally designed for rainfall events having return periods of 20 years. This means that such systems are likely to become overloaded and cause some degree of flooding when rainstorms are experienced with return periods greater than 20 years. No guidance on this is provided in CUBiC. However, consideration should be given to moving this common standard to a longer return period. Intensity-duration-frequency relationships for rainfall events are required for rational engineering analysis. These are available from the Caribbean Meteorological Institute for some Caribbean islands.

The damage caused by flooding depends on the type and elevation of facilities in the location. The results of flooding may range from the inconvenience of temporarily submerged driveways to the loss of equipment and finishes inside flooded buildings and consequential disruption of the functions.

Flooding has been the cause of many of the deaths and of much property damage as well. Clearly, location is critical when it comes to flood risk. Low-lying lands, riverbanks and lands adjacent to gullies are to be avoided if possible. If not, deliberate drainage measures must be taken. Usually this is a municipal responsibility, at least in terms of overall control, since what happens to one property can easily be affected by a neighbour's actions.

The design of external works for shelter premises would include consideration of local flooding. It would not be practical to study the overall catchment area for the design of each particular site. However, some general appraisal of the adjacent areas should be carried out as input for the design of on-site drainage structures.

1.3.6 Storm Surge

Storm surge is associated with hurricanes and consists of unusual volumes of water flowing onto shorelines. Storm surge has been responsible for much of the damage caused by hurricanes, especially in large, low-lying coastal settlements.

Storm surge is a complex phenomenon which behaves quite differently from one shoreline to another. The several main components governing their behaviour are:

The increase in coastal settlement has put much of our economic investment at risk from sea damage. Future rises in sea level can only make this condition more acute. Storm surge caused by hurricanes causes the most dramatic damage. Waves cause damage without accompanying surge but they are also superimposed on storm surge.

As well as causing flooding and damage to coastal structures, storm surge may also precipitate flooding further inland through the blockage of the outfalls of drainage systems.

This is outside of the scope of the present exercise.

1.3.7 Landslides

Most landslides in Caribbean islands cannot be regarded as natural hazards since they are brought about by inappropriate farming practices although triggered by natural events. Another man-induced class of landslides is road construction in mountainous areas. Often the cut embankments are too steep for stability. Often the down-slope fills are unstable.

There are, however, a few situations where nature itself produces the events. This phenomenon is outside the scope of this report.

On 11 December 1997 an initial briefing took place in Anguilla. Those present were:

The overall programme was explained and it was decided that Stanford Harrigan (a local contractor) would perform the surveys with Ken Banks’ assistance.

The survey forms delivered to Mr Harrigan were:

• Hurricane Vulnerability Assessment
• Seismic Vulnerability Assessment - Part I
• Seismic Vulnerability Assessment - Part II

The shelters which are daycare and school facilities on the list to be surveyed were:

Stanford Harrigan proceeded with the field surveys in April 1998.

In accordance with the requirements for the work, completed survey forms were sent to Tony Gibbs in Barbados in advance of his Anguilla visit. This was done on 5 May 1998. (The photographs were included with this submission.)

Tony Gibbs undertook the field reviews in the company Stanford Harrigan on 22 June 1998.

To assist with the determination of vulnerability to hurricanes (and in accordance with the ToR) the software package WIND-RITE was utilised. WIND-RITE is a software system developed by the Wind Engineering Research Center at Texas Tech University, Lubbock, TX in conjunction with the Insurance Institute for Property Loss Reduction. This program is used to establish a relative grade for the wind resistance of buildings by assigning various vulnerability weights to: site environment, roof envelope, wall envelope, structural framing, and other secondary considerations.

A relative grade between 1 and 10 is assigned to a building during the evaluation process. A relative grade of 1 indicates that the building is highly resistant to wind-induced damage (less damage), whereas a building which receives a relative grade of 10 is highly susceptible to wind-induced damage (extensive damage). For non-engineered construction, a score of five would be considered very satisfactory.

To assist with the determination of vulnerability to earthquakes (and in accordance with the ToR) the methodology of Ahmed F Hassan [10] was used. This methodology relies mainly on readily accessible data for an existing building such as the dimensions and arrangement of its structural elements and the floor area. Masonry walls, concrete columns and steel columns are given progressively higher weightings in determining their contributions to seismic resistance. The end result is a comparative assessment of seismic vulnerability. A copy of Hassan’s paper is provided in Appendix II.

It would not be unreasonable for the Government to require that its new, formally-designed, school buildings and daycare centres would have the capacity to be largely undamaged by a 50-year-return-period hurricane.

The objective stated above requires simply the application of known technology by the school's architects and engineers to:

• conceptual design (ie site selection, shape, materials, structural systems);
• analysis (ie mathematical determination of forces, stresses and deformations of the building elements); and
• detailing (ie the determination and presentation, principally through drawings, of all of the details of construction).

The next requirement is for faithful attention to be paid during construction to the details incorporated in the drawings, schedules and technical specifications for the project. Supervision and inspection are central to success in this area of execution.

The final requirement is for the property to be adequately maintained.

The above precepts seem simple enough. Yet buildings are often damaged and indeed sometimes destroyed by hurricane events no worse than the "design" storm. That design storm cannot be less than (and could reasonably be greater than) the "50-year" storm. These failures come about because of the failure of the systems (or the inadequacy of the systems) set up to order the construction industry and to maintain public property.

3.1.2 Aspects of Vulnerability

The experience of recent hurricanes has demonstrated that most failures occur to roofs, windows and external doors. In a very few cases there is damage (or collapse) of walls. In even fewer cases there is foundation failure. The experience in Anguilla provides some evidence of this pattern.

Roofs:

• Most of the buildings have gable roofs of moderate pitch while a few have hipped ends to their roofs. None of the roofs are optimally steep. Hipped ends are favourable and gable roofs are acceptable, provided special attention is paid to the gable-end fastening of lightweight roof sheets.

• It is well known that certain areas of roofs are much more highly stressed by wind forces than other parts. Nevertheless most of the roofs observed had constant spacing of fasteners throughout.

• The ridge fixings (without underlying spacer blocks) and the types of washers are in many cases unsatisfactory for confident resistance to wind uplift forces. Valley fixings, where they are used on sheets of trapezoidal profile, have shown themselves to be much more secure - everything else being equal.

• Lightweight roof sheeting has been used for all of the buildings. This sheeting is of several profiles, several thicknesses and several types of material. The tendency, because of commercial pressures and budgetary constraints, is for the sheeting to be somewhat thinner than is desirable.

Windows (and other openings):

• If the wind can get in, then the rain can get in. If the rain can get in, it is surprising what else can get in. Also, if the rain can get in, the use of the building as a hurricane shelter is compromised.

• The convenience of aluminium louvers for everyday use is self-evident. But it must be recognised that such windows have difficulty in keeping out torrential rain driven by strong winds.

• Flying debris usually accompanies hurricanes. Glass windows are very vulnerable to breakage if impacted by such debris. Other materials (aluminium and thin, wooden, louver blades) could also suffer damage in such circumstances.

• There was little evidence in the field surveys that the various properties were provided with "in-place" storm shutters or dedicated prefabricated shutters in secure storerooms. Without shutters most of the buildings would be unsatisfactory shelters during severe hurricanes.

External Doors:

• Doors are much less vulnerable than windows. Their materials of fabrication are usually more robust. The main aspects of vulnerability are the hinges, the locks and the absence of bolts at the tops and bottoms of the "free" edges.

• No braces were noticed for securing doors against wind forces.

3.1.3 Summary of Results

The table below summarises the conclusions of the survey with respect to wind vulnerability. A higher WIND-RITE Rating indicates a higher vulnerability to hurricane events.

RC = reinforced concrete

The WIND-RITE analysis results appear to indicate that the properties have an approximate equally high vulnerability to hurricanes. It should be noted that the WIND-RITE program will produce high vulnerabilities if the building has unprotected openings and light-weight roof sheeting. Since these vulnerable areas can be mitigated against relatively easily, it is not considered prudent to exclude a property from being considered as a shelter based on these results.

In 1995, Hurricane Luis visited Anguilla. The reported damage to the surveyed properties is described in the following table.

It is not known how intense Hurricane Luis was in Anguilla, however the above catalogue of damage is relatively minor to moderate. The path of Luis was closer to Anguilla than Sint Maarten, it must therefore have impacted more severely on Anguilla than Sint Maarten where significant damage was reported. If Luis was experienced in Anguilla as a category 3 or higher hurricane, then the WIND-RITE results could be considered rather pessimistic.

Property becomes vulnerable to torrential rain either directly, if the rain penetrates the building envelope, or indirectly, through the resultant flooding events. Improperly sealed window and door openings or roof covering can result in the immediate ingress of water, while poorly sited buildings, and/or buildings with inadequate or improperly maintained drainage systems can result in the property being vulnerable to flooding.

The not so immediate effect of flooding caused by torrential rains is the adverse effect it can have on the foundations of buildings.

Coastal areas can be particularly vulnerable to flooding especially if high tides accompany heavy rains. In addition, the groundwater level is generally higher in coastal areas resulting in less absorption by the ground and an increased incidence of flooding.

3.2.2 Summary of Results

The table below summarises the conclusions of the survey with respect to torrential rain vulnerability.

When compared with dealing with a single hazard, designing against multiple hazards is more than doubly difficult, especially when those hazards are wind and earthquake. Some favourable features of wind-resistant design are unfavourable for earthquake-resistant design and vice versa, eg:

• Heavy structures resist winds better. Light structures resist earthquakes better.
• Flexible structures attract greater wind forces. Stiff structures (generally) attract greater earthquake forces.

Both hurricanes and earthquakes impose horizontal loads on buildings. Earthquakes also impose significant vertical loads on the overall building. The vertical loading derived from wind is usually significant on parts of a building as determined by aerodynamic considerations.

However, there are many similarities in the effective design and construction of buildings to resist hurricanes and earthquakes:

• Symmetrical shapes are favourable.
• Compact shapes are favourable.
• There must be a realisation that there is a real risk that "design" forces may be exceeded. This is particularly so in the case of earthquakes where the design force is deliberately determined to be less than that expected during the anticipated life of the building. This leads to a requirement for redundancy in the structure and for "toughness" - the ability to absorb overloads without collapse.
• Connections are of paramount importance. Each critical element must be firmly connected to the adjacent elements.

3.3.2 Differences Between Designing Against Earthquakes and Hurricanes

There is a basic difference in the performance expectations in the event of an earthquake as opposed to a hurricane. A building is expected to survive its "design hurricane" with virtually no damage. Even a catastrophic hurricane should only lead to repairable damage. On the other hand the "design earthquake" is expected to cause (hopefully repairable) damage, and a catastrophic earthquake is likely to lead to a situation where the building cannot be repaired and must be demolished. In such an event success is measured by the absence of deaths and serious injuries.

3.3.3 Aspects of Vulnerability

Designing buildings to be safe during earthquakes is a complex process requiring the involvement of specialist structural engineers. In Anguilla, as in most parts of the Caribbean, the problem is compounded because of the absence, until very recently, of mandatory earthquake-resistant standards and the lack of a tradition of conscious engineering attention to this subject. Indeed there is no consensus that the hazard is sufficiently serious to warrant concerted action. Thus successes in the area of earthquake-resistant construction tend to be accidental. This comes about principally when the dominant design criterion overrides the demands of earthquake-resistant design.

The main problem areas are summarised below.

• Structural vulnerability was noted mainly in two-storey construction, and in masonry walls where the presence of reinforcement was moot. (Since concrete block walls often collapse during seismic shaking, all concrete block walls should be reinforced and braced by the adjacent structures.)
• Internal lighting fixtures and utility equipment were generally not fastened to structural elements securely enough to withstand large ground accelerations.
• Closets and heavy furniture were generally not fastened or strapped to the walls where these constituted a danger or contained valuable property.

3.3.4 Summary of Results

The table below summarises the conclusions of the survey with respect to earthquake vulnerability. A higher Seismic Relative Vulnerability Number indicates a lower vulnerability to seismic events.

RC = reinforced concrete

NA = Not Available

The following recommendations are made with a view to reducing the vulnerability of the buildings surveyed to hurricanes, torrential rains, and earthquakes. The activities required are generally:

1. Protect glass windows and doors with approved shutters.
2. Improve timber roof framing connections by installing approved hurricane straps and ties.
3. Improve the roof sheeting connections by installing additional screws in vulnerable roof areas.
4. Improve the water tightness of the aluminium louver windows.
5. Strengthen un-reinforced masonry walls.
6. Improve the drainage around the property.

If these recommendations are carried out, then all of the properties surveyed can be shelters, however it may be prohibitively expensive to improve the drainage around the Island Harbour and West End coastal facilities. It may therefore be appropriate to exclude these two properties from further consideration as shelters.

The following table shows the recommendations for each property.

From the mitigating measures described in Section 3.4, several classes of vulnerability are identified. They are:

1. Vulnerable roof
2. Vulnerable walls
3. Vulnerable dominant openings
4. Vulnerable site drainage

The following table identifies the vulnerable areas for each property.

Hurricanes David in 1979, Allen in 1980, Hugo in 1989 and Luis in 1995 caused significant damage to schools in Dominica, St. Lucia, Montserrat and Antigua respectively. The review of the damage to most of the schools showed that roof failure was the primary cause of problems, leading in some cases to complete collapse of the walls. Most of the roofs were made of wooden rafters or trusses with 26-gauge (or thinner) galvanised sheeting. Few roofs of aluminium sheeting survived. Some roofs remained partially intact but the damage to windows, doors and internal partitions was significant. Some wooden structures were completely destroyed, sometimes by falling trees and sometimes by being blown off their foundations.

In general, therefore, the response of the school buildings to hurricane forces must be judged to have been poor. It is disturbing to note that the reconstruction of many of the schools which suffered from the recent hurricanes and the construction of some new schools have been based on the same design concepts which have led to the failures.

Disasters are often seen as unpredictable, having to do with luck and part of the risks of everyday living. Surely we have progressed beyond the stage when superstition, mythology and fatalism were the public responses to natural hazards. Hurricanes are not natural disasters, they are natural events which sometimes lead to manmade disasters. In these days of widespread technological education, sophisticated research, reliable building materials, computer-based geographical information systems and satellite-assisted warning programmes, hurricanes in the Caribbean should not lead to disasters. The one exception to this would be vulnerable agricultural crops such as bananas.

It is now evident that disasters due to natural hazards are largely preventable and soon the public will demand deliberate actions to protect communities against hazardous events.

The development and maintenance of building standards and codes is a continuing process in many countries. The mandate of national or regional professional and governmental institutions usually includes the development of building standards and the co-ordination of such activities among its various constituent members.

To date the majority of countries in the Caribbean region are without satisfactory, mandatory standards dealing adequately with any of the prevalent natural hazards.

Some government agencies adopt an ad hoc approach to standards based, principally, on the particular individuals involved in the specific projects. In most cases the administrators tacitly assume that their designers and builders would do what is right without being told. In other cases the administrators adopt the approach of not objecting to safe design and construction, provided that these attributes do not interfere with their other aims for the projects.

Many government capital works projects are funded by international lending agencies. Typically there is a reluctance on the part of these agencies to impose structural design criteria on their projects. The funding agencies leave it up to the governments and the governments leave it up to their designers and builders. This laissez-faire approach leads to inconsistent performance, lack of reliability and, arguably, to higher overall life-cycle costs for the built environment.

The Anguilla Building Code has been adopted formally by the Government. This is a significant step in the right direction.

The enforcement of the Code is now the critical function to be performed. The system of check consultants (bureaux de contrôle), routinely used in French territories, is proposed. Check consultants are independent of the design consultants. It is well recognised that quality assurance is more effective where checking is done independently of creating. The system mentioned above formalises the process.

During the past 111 years a total of approximately 1000 tropical storms and hurricanes have been recorded over the North Atlantic area. Of these, about 50 percent were hurricanes, many in the general area of the Caribbean. With such a record, it is not surprising that everyone in this region accepts that hurricanes are a fact of life. However, the frequency of direct hits by hurricanes on any one territory is low. This has led to a considerable lack of consciousness amongst Caribbean people about the dangerous risk to their own properties. Few believe that their island would be hit.

This is the level of assessment being carried out in the present programme. This level of evaluation does not envisage exhaustive testing of materials in place nor sophisticated computation of stresses. It does involve a careful review of all readily available data (such as drawings), an inspection of the building without destructive testing and a non-mathematical review of the data.

4.6.2 Analytical Evaluation

Facilities whose performances are deemed to be doubtful when assessed qualitatively would be subjected to an analytical evaluation. This procedure would also be used as a second stage, prior to implementation of retrofitting, in cases deemed to be self-evidently inadequate by qualitative assessment. Since analytical evaluation is a time-consuming and expensive exercise it would be appropriate to carry it out only when the funds were available for implementing the possible actions indicated by such evaluation.

It is envisaged that analytical evaluations will be undertaken for those buildings selected for the CDB retrofit project.

In the aftermath of a disaster the focus is understandably on getting educational facilities to function again as soon as possible. Also, technical personnel and financial resources are spread very thin at such times. This combination of factors often leads to repairs being carried out in an expedient manner without adequate attention to safety issues. Indeed, post-disaster repairs often leave the buildings even more vulnerable than they were in their pre-disaster, inadequate states.

If such a scenario is to be avoided, very deliberate steps must be taken by the custodians of the educational facilities. Such actions would include:

• clear instructions (on performance criteria for natural hazards) given to engineers, architects and contractors involved in repairs and rehabilitation;
• a willingness to accept (temporarily) smaller functioning spaces or fewer classrooms if the repair funding is inadequate to achieve safe standards for all of the damaged facilities; and
• the employment of a mitigation officer to review and monitor the designs and construction so as to ensure that the agreed performance criteria are being met.

The physical condition of many Caribbean schools is poor. Windows and doors show lack of maintenance and repair. It is considered that a major effort should be taken to bring the condition of the buildings to the standard where a normal maintenance crew can be expected to deal with the routine maintenance requirements of the facility. It is considered, also, that the existing staff and maintenance budget are generally insufficient to provide for proper maintenance.

It is recommended that for public buildings with the heavy use of a school, the annual maintenance budget should amount to about 4% of the contemporary capital cost of the building and equipment, assuming that the facilities are in good condition to start with. For schools, it is estimated that the replacement cost is about US$150,000 per classroom. (This figure includes amounts for common and administrative areas as well as infrastructure.) The maintenance allocation should therefore be no less than US$6,000 per classroom per year.

The maintenance of a school, rather than being a one-time activity as is the construction of the school, is a continuous daily operation of the institution and is an important ingredient in the delivery of education.

A good maintenance system is also a good disaster mitigation system, as the review of damage caused by recent hurricanes and floods has shown. To some extent the damage to buildings was due to lack of sustained maintenance of critical items. Also, a well operated system of maintenance for buildings and equipment has the effect of being a very effective disaster mitigation measure in terms of cost and facility usage. It ensures the most economic way to keep the building and equipment in the best of form for normal use, given the original design and materials. It is essential that a maintenance plan be included in disaster mitigation plans.

It should be noted that the establishment of a secure and orderly system of archiving of construction drawings depicting the as-built condition of the final structure is an essential ingredient of the long-term maintenance schedule of any constructed facility. Because the archive will serve its purpose over the lifetime of the facility, the protection of its contents against the effects of time cannot be overlooked.

REFERENCES

1. Caribbean Disaster Mitigation Project funded by the United States Agency for International Development (USAID) and managed by the OAS
2. European Community Humanitarian Office
3. Insurance Institute for Property Loss Reduction
4. International Development Research Centre, Ottawa, Canada
5. 1000 gals is approximately equal to the acceleration due to gravity
6. This is part of the OECS Building Code project funded by the United Nations Development Programme (UNDP) through the United Nations Centre for Human Settlements (UNCHS) or Habitat.
7. Preliminary Investigations of the Tsunami Hazard of Kick'em Jenny Submarine Volcano, Natural Hazards, 7:257-277, 1993.
8. kPa = kilopascals
9. metres per second
10. Seismic Vulnerability Assessment of Low-Rise Buildings in Regions with Infrequent Earthquakes by Ahmed F Hassan and Mete A Sozen, ACI Structural Journal, January-February 1997, pages 31 et seq