Nat Hazards DOI 10.1007/s11069-010-9539-0 ORIGINAL PAPER Natural and man-made hazards in the Cayman Islands D. A. Novelo-Casanova • Gerardo Suárez Received: 4 November 2009 / Accepted: 7 April 2010 Springer Science+Business Media B.V. 2010 Abstract In this work, we analyze the various natural and man-made hazards that may affect the Cayman Islands and determine the level of exposure of Grand Cayman to these events. The magnitude, frequency, and probability of occurrence of the natural and manmade hazards that may potentially affect the islands are identified and ranked. The more important natural hazard to which the Cayman Islands are exposed is clearly hurricanes. To a lesser degree, the islands may be occasionally exposed to earthquakes and tsunamis. Explosions or leaks of the Airport Texaco Fuel Depot and the fuel pipeline at Grand Cayman are the most significant man-made hazards. The results of the hazard evaluation indicate that there are four areas in Grand Cayman with various levels of exposure to natural and man-made hazards: The North Sound, Little Sound, and Eastern West Bay (Area 1) show a very high level of exposure; The Central Mangroves, Central Bodden Town, Central George Town, and the West Bay (Area 2) have high level of exposure; The Northwestern West Bay, Western Georgetown-Bodden Town, and East End-North Side (Area 3) are under moderate levels of exposure. The remainder of the island shows low exposure (Area 4). Keywords Natural hazards Man-made hazards Vulnerability Grand Cayman Cayman Islands 1 Introduction It is important to understand that disasters arising from hazards of natural origin are only partially determined by the physical event itself. Reducing disaster risk requires the assessment of the level of the hazard and the various types of vulnerabilities (economical, social, environmental, etc.) of the affected society. When describing a hazard, it is important to know the basic characteristics of hazardous events such as location, time, D. A. Novelo-Casanova (&) G. Suárez Instituto de Geofı́sica, Departamento de Sismologı́a, Universidad Nacional Autónoma de México, 04510 Mexico D.F., Mexico e-mail: [email protected] 123 Nat Hazards intensity, and frequency (Smith 2001). Hazards are often grouped into three main categories, according to their causes: natural (earthquakes, hurricanes, etc.), technological (explosions, release of toxic material, etc.), and anthropogenic (terrorist activity, crowd related, etc.) (Schneiderbauer and Ehrlich 2006). However, hazards may have interrelated causes, and the allocation of a hazard to one class is often difficult. Thywissen (2006) states that every disaster starts with a hazard: known or unknown. Independently of the origin of the hazard, they all have in common the potential to cause the severe adverse effects that lie at the bottom of every emergency, disaster, and catastrophe. One important feature of a hazard is that it has the notion of probability or a likelihood of occurrence. A hazard is a threat not the actual event. Any hazard can manifest itself in an actual harmful event. In other words, if it can be measured in terms of real damage or harm it is not longer a hazard but has become an event, disaster, or catastrophe. Vulnerability is a function of the level of sensitivity or susceptibility of a system (community, household, building, infrastructure, nation, etc.) to be damaged. Rashed and Weeks (2003) pointed out that vulnerability is independent from any particular magnitude of a specific natural event and depends only on the context in which it occurs. Thus, vulnerability cannot be assessed in absolute terms. The performance of a determined community should be assessed with reference to specific spatial and temporal scales. Natural hazards are threatening events capable of producing damage to the physical and social space where they take place not only at the moment of their occurrence, but on a long-term basis due to their associated consequences. When these consequences have a major impact on society and/or infrastructure, they become disasters (Alcantara-Ayala 2001). Hazards are the result of sudden changes in long-term behavior caused by minute changes in the initial conditions (Scheidegger 1994). In this work, we consider the following terms and concepts: • Elements exposed: Elements such as people, buildings, and networks that are subject to the impact of a specific hazard. Other elements that are also exposed include the economy and the natural environment. • Man-made hazard: A threat to a community having an element of human intent, negligence, or error that involves a failure of a man-made system. • Natural hazard: A specific natural phenomenon that may be capable of causing damage to population or property characterized by a certain magnitude and likelihood of occurrence. Common to all natural hazards is the uncertainty associated with the rate of occurrence of a particular hazard, the magnitude, and the spatial extent of its impact. • Vulnerability: We consider the definition provided by the Consultative Group for the Reconstruction and Transformation of Central America (CGRTCA 1999) as part of an analysis on the effects of Hurricane Mitch that occurred in 1998: ‘‘any condition of susceptibility to external shocks that could threaten people’s lives and livelihoods, natural resources, properties and infrastructure, economic productivity, and a region’s prosperity’’. This description states that vulnerability is exacerbated in impoverished areas due to weak infrastructure, a failure to implement prevention and preparedness measures, and an inability to recover from a hazard event. • Risk: The estimated impact that a hazard event would have on people, services, facilities, structures, and assets in a community. It is defined by the following expression (Crichton 1999): Risk ¼ hazard elements exposed vulnerability 123 ð1Þ Nat Hazards Thus, ‘‘risk’’ is the probability of a loss, and this loss depends on three elements: hazard, exposure, and vulnerability. If any of these three elements increases or decreases, then risk increases or decreases, respectively’’ (Crichton 1999). In order to analyze the hazards and the level of exposure of the Cayman Islands to natural and man-made extreme events, we adapted a methodology similar to that developed by the North Carolina Department of Environment and Natural Resources and other research partners during the study entitled ‘‘New Hanover County/Wilmington Project Impact Partnership’’ (http://www.csc.noaa.gov/products/nchaz/htm/methov.htm 2001). In another paper, we discuss the vulnerability of the critical facilities in the islands using the same methodology. Historical and recent data were collected and classified in order to identify and rank the natural and man-made hazards. We identified four areas of high potential impact for the main hazards in Grand Cayman and established a relative ranking of exposure to these hazards in each of these areas. The different degrees of physical vulnerability for each hazard are presented graphically with the aid of maps using a relative scoring system. Spatial maps were generated showing the areas of different levels of exposure to multihazards. These maps are intended to be useful tools for emergency managers and policy developers and to increase the overall awareness of decision-makers for disasters prevention and mitigation plans in the Cayman Islands. It is important to underline that this study presents a first evaluation of the main natural and man-made hazards that may affect the Cayman Islands. A full hazard analysis of the islands should include a complete and quantitative hazards assessment modeling. However, our results constitute the basis of future mitigation risk projects in the islands. 2 The Cayman Islands: geographic and geologic setting Located in the western Caribbean Sea to the northwest of Jamaica, the Cayman Islands (CI’s) are a British overseas territory comprised of three islands: Grand Cayman (GC), Cayman Brac (CB), and Little Cayman (LC) (Fig. 1). These three islands occupy around 250 km2 of land area (Brunt and Davies 1994). GC is approximately 35 km long and 13 km at the widest point wide. The highest elevation is about 18 m above sea level, and the most striking geographical feature is the North Sound, a shallow reef-protected lagoon with an area of about 56 km2. CB lies about 145 km northeast of GC. It is about 19 km long and a little over 1.6 km wide. LC is 8 km west of CB and is 16 km long and 3 km at its widest point (Fig. 1). It is the flattest of the three islands with its highest elevation being 12 m. To the west, an 11-km channel separates CB from LC (Brunt and Davies 1994). The three islands are mostly flat and were formed by large coral heads, covering submerged ice age peaks of western extensions of the Cuban Sierra Maestra range. The highest point is The Bluff, a limestone outcrop 43 m in height on the eastern end of eastern CB. The CI’s lowest elevation is the Caribbean Sea at sea level (Brunt and Davies 1994). The Caymans are formed of two distinct formations of calcareous rock. The older limestone, called bluff limestone, was formed in the Oligocene–Miocene period, about 30 million years ago. This limestone forms the central core of each island. It is a dense karstic limestone. Surrounding this bluff limestone core is a coastal limestone terrace called ‘‘ironshore’’ which is a formation of consolidated coral, mollusk shells, and marl with some limestone. It was formed about 120,000 years ago in the Pleistocene period 123 Nat Hazards 20o N 19.4o N WEST BAY LITTLE SOUND NORTH SIDE NORTH SOUND CENTRAL MANGROVES GEORGE TOWN BODDEN TOWN EAST END 19o N SOUTH SOUND 19.3o N 81.4o W 81.1o W 81o W 80o W Fig. 1 The tectonic boundaries of the Caribbean Plate and the location of the Cayman Islands. The major geologic faults in the northern Caribbean are shown. The Gonave plate is bounded by the Oriente fault to the north that passes just south of the Cayman Islands, and the Walton fault to the south of it, passing through Jamaica (after DeMets and Wiggins-Grandisson 2007) (Folk et al. 1973). Due to the porous nature of the limestone rocks that are present along with the absence of much relief of any kind, all of the Caymans lack rivers or streams (Folk et al. 1973). The Bluff Group is formed of the Brac Formation (Lower Oligocene), the Cayman Formation (Middle-Upper Miocene), and the Pedro Castle Formation (Lower Pliocene). 123 Nat Hazards The Brac Formation, exposed only at the east end of Cayman Brac, is formed of limestones and dolostones, whereas the overlying Cayman Formation is formed almost entirely of finely crystalline dolostones. The uppermost Pedro Castle Formation is formed of limestones and dolostones. The original carbonate sediments, deposited on shallow banks in water no more than 30-m deep, contained numerous corals, bivalves, gastropods, algae, and foraminifera. Boundaries between these formations are unconformities that represent periods of subaerial exposure and extensive karst development. The Brac Unconformity, which separates the Brac Formation from the Cayman Formation, has at least 25 m of relief on it. The Cayman Unconformity, which defines the upper boundary of the Cayman Formation, has at least 40 m of relief and many associated karst features. Diagenesis in the Bluff Group has been extensive. Porosity is high with caves and fossil moldic porosity dominating (Folk et al. 1973). The Cayman Trough (CT) is a depression area on the seafloor of the Caribbean that extends from the Belize margin to northern Jamaica (Fig. 1). At its deepest point, the CT is over 7,500-m deep (Leroy et al. 2000). This margin consists of a 100–250 km wide seismogenic zone of generally left-lateral, strike-slip deformation which covers over 2,000 km along the northern edge of the Caribbean Sea. This left-lateral strike-slip displacement is due to the eastward movement of the Caribbean plate relative to the adjacent North American plate (Leroy et al. 2000). Geological and geophysical data from the region suggest that the CT is underlain by oceanic crust accreted along a short north–south spreading center located between the Oriente and Swan transform faults. Ultrasonic seismic profiles of the CT revealed that the steep northern wall of the trench is virtually sediment free (Rosencrantz and Mann 1991). Based on plate reconstruction schemes estimates of the Caribbean-North America plate motion rates, range from 1 to 2 cm/year (Rosencrantz and Mann 1991; De Mets et al. 1994). Relative plate motion is taken up by strike-slip faulting along the northern boundary of the Caribbean plate and ENE directed convergence along the eastern margin. A study by De Mets et al. (2000) suggests that the motion between the Caribbean and the North America plate is faster than predicted by the NUVEL-1A model averaging velocities of 1.8–2 cm/ year (Stein et al. 1988). Based on these results, we consider 2 cm/year as the average velocity between the Caribbean and North American plates (Fig. 1). Recently, however, De Mets and Wiggins-Grandison (2007) reported that the motion on the Oriente fault might be as slow as 0.6 cm/year. The Oriente fault is the one immediately to the south of the CI’s. In the interpretation of De Mets and Wiggins-Grandison (2007), a relatively small plate, called the Gonave plate, lies between the Caribbean and the North American plate (Fig. 1). The results of De Mets and Wiggins-Grandison (2007) suggest that most of the motion between these two major tectonic plates is absorbed along the Walton fault, well to the south of the CI’s (Fig. 1). They suggest that the Walton fault moves at an average rate of 1.3 cm/year. The deficit existing between the 2.0 cm/year of the Caribbean–North American plate motion is apparently taken up by the Oriente Fault at a rate of approximately 0.6–1.1 cm/year. Thus, these authors argue that most of the motion of this plate boundary has migrated to the southernmost Walton fault and little motion takes place today on the Oriente Fault, immediately to the south of the CI’s. 3 Methodology There are two primary components of the vulnerability assessment methodology used here and developed by the North Caroline Department of Environment and Natural 123 Nat Hazards Resources (2001). The first is a tutorial, which defines a general process for conducting a comprehensive community vulnerability assessment. The second component is the case study in New Hanover County, North Caroline, USA, which illustrates the use of the methodology in a specific community. This case study illustrates the use of geographic information system (GIS) technology as a valuable resource for conducting hazards-related analysis. A major feature of this product is a section focusing on the use of spatial data for hazards planning. Several relevant GIS and remote sensing applications are introduced as potential tools for supporting detailed hazards analysis. The methodology was developed to assist communities in their efforts to reduce their vulnerability to various hazards. It explains the process for analyzing physical, social, economic, and environmental vulnerability at the community level. The foundation for this tool was established by the findings of the H. John Heinz III Center for Science, Economic and the Environment (1999). The methodology is a simple and straightforward approach that lends itself to the quality and quantity of the available data (a full description of the procedures can be found in: http://www.csc.noaa.gov/products/nchaz/htm/methov.htm). Briefly, the steps involved in this methodology are as follows: 1. 2. 3. 4. 5. 6. 7. Hazard Identification Hazard Analysis Critical Facility Analysis Societal Analysis Economic Analysis Environmental Analysis Mitigation Opportunity Analysis Based on the goals of our study, which are to evaluate the level of exposure to natural and man-made hazards of the CI’s, we considered only the first two steps. In Step 1, hazards are identified by the systematization of all available information to determine the main hazards that will be considered in the analysis. This can be either a comprehensive list of hazards that pose some threat to the community or a more limited list of specific hazards on which the community is interested. The hazards are characterized by its probability of occurrence, size of area of impact, and the potential damage. For each identified main hazard, we obtained a total score following Eq. 2 by assigning weights to each factor depending on how critical that factor is: Total score ¼ ðfrequency þ area of impactÞ potential damage magnitude ð2Þ The frequency, area of impact, and potential damage magnitude values are defined by a scale of numbers ranging from 1 to 6, where: extremely low = 1 and very high = 6. The ideal method for assigning score values to the various hazard threats would be a quantifiable probability assessment. Unfortunately, probability data are not consistent usually among the different hazard types nor are they always available. As an alternative, this methodology develops a relative priority matrix to use as a general guide for addressing the different hazards. Designing such a matrix requires determining which factors are most critical for a determined community and assigning weights accordingly. The purpose for this step is to identify the hazards and their potential impacts. It is a subjective exercise where the total scores alone do not have absolute statistical significance. The comparison of scores, however, will provide relative rankings that guide the vulnerability assessment process as well as the establishment of hazard mitigation priorities. In Step 2, the exposure areas are determined for each hazard. The objective of this step is to target priority areas for which a hazard evaluation is needed. The purpose is to identify 123 Nat Hazards geographically the areas that are most likely to be affected by a given hazard. Once the areas exposed to the identified hazards are identified, a prioritization is developed using local data sources. For each identified area, a relative level of exposure to the specific hazard being addressed is established. These locations are considered as areas exposed to different levels of hazard impacts. In many cases, it is possible to establish some relative ranking within the exposure areas. For example, flood zones for 10- or 50-year floods should be ranked higher than 100- or 500-year floods. The different levels of exposure are then represented graphically using the relative scoring system (higher scores for higher exposure areas). For example, hurricane storm surge maps are usually created for five different categories of storms. Category 1 storms are generally associated with the least severe winds and storm surge, while Category 5 storms are considered to be the most severe. In most cases, those areas subject to storm surge in the lower category storms are also projected for inundation in all of the higher categories. When developing a relative priority scoring system for storm surge inundation, Category 1 storm surge areas would therefore have the highest risk of being flooded since they are at risk of inundation in all storm events. The general concept is that locations with no exposure to hazards will have a score of zero and each incremental increase in level of exposure adds one point. 4 Data sources The data for this study were collected via electronic means and from scientific sources in the public domain. The Hazard Management Cayman Islands (HMCI) generously provided part of this information. We interpreted and complemented our data with the use of the following documentary sources (in alphabetical order): • Cadastre Map of the Cayman Islands (Lands and Survey Department, Cayman Islands, 2007) • Cayman Islands’ National Hurricane Plan 2006 (Emergency Operation Center, Cayman Islands, 2006) • Development Plan Map 2006 (Central Planning Authority, Cayman Islands, 2006) • Map of flooding areas during Ivan Hurricane (Lands and Survey Department, Cayman Islands, 2005) • Map of Hurricane Ivan Preliminary Damage Assessment (Lands and Survey Department, Cayman Islands, 2005) • Map of location (latitude and longitude) of critical facilities (hospitals, schools, shelters, fuel deposits, fuel and gas pipeline, government communications infrastructure, power stations, ports, water and sewage treatment plants, water storage plants, airport, police and fire departments, critical government, and Red Cross installations) (Hazard Management Cayman Islands, 2009) • Petroleum Products Location Map (Lands and Survey Department, Cayman Islands, 2007) • Preliminary Post-Ivan Environmental Impact Assessment Report (Department of Environment, Cayman Islands, 2004) • Series of Daily, Monthly, and Annual Precipitation of the Cayman Islands (Cayman Islands National Meteorological Service, 2008) • Terrain and Bathymetry Map, Grand Cayman, Cayman Islands (Lands and Survey Department, Cayman Islands, 2007) 123 Nat Hazards • Topographic Map of the Cayman Islands (Lands and Survey Department, Cayman Islands, 2007) 5 Hazard identification 5.1 Hurricanes The year 2004 had an extraordinarily active hurricane season in the Caribbean basin. Substantial damage was experienced with varying degrees of economic and social impact, depending upon the resilience of the territories, their response capabilities and the size, diversification and strength of their economies, societies and infrastructure (United Nations Economic Commission for Latin American and the Caribbean, UN-ECLAC 2004). The comprehensive assessments performed by the UN-ECLAC (2004) showed that in the CI’s the impact, in terms of the annual GDP, was 138%. During Hurricane Ivan, which impacted the CI’s on 12 September, 2004, GC experienced several storm surges that affected the North and South Sounds (Fig. 1; Young 2004). Ivan also affected CB and LC with tropical storm winds. GC suffered the brunt of the storm in all its intensity: windswept first from the southeast, in a northwesterly direction, and with storm surges that flooded large portions of the coastal areas and deposited huge amounts of sand over roads, houses, and infrastructure from East End to West Bay (Young and Gibbs 2005). Ivan took the lives of two persons on GC, and it temporarily displaced a significant part of the population. About 83% of the GC population was directly affected by the wrath of this hurricane. The remaining 17% of the population was indirectly affected as they sheltered or cared for those directly affected. The whole population experienced the loss of electricity, water, and access to telecommunications for some period immediately following the disaster. Entire Districts such as Bodden Town and the Eastern Districts (Fig. 1) were isolated due to debris or road cuts, and individuals across the island were isolated in their homes due to high water and debris (Young and Gibbs 2005). Hurricane Ivan produced major damage to roads, utilities, and property, especially in the western part of GC and along the southern coast (Young and Gibbs 2005). Although the CI’s have a high level of insurance and hence, a great capacity for rebuilding, the islands faced a shortage of immediate resources and had an important impact on the government’s budget and cash flow. Cruise ships, for example, were able to return seven weeks after the storm (Young and Gibbs 2005). Most public shelters performed extremely well during Ivan, and almost 4,000 people safely rode out the storm. Ivan was an unusually strong hurricane, which produced high winds, storm surge flooding, heavy rain, and strong wave action. The impact on GC was short of catastrophic. Young (2004) showed that the meteorological conditions were certainly capable of causing a major catastrophe, it was only because of the high standards of the infrastructure, especially shelter accommodation and other critical infrastructure, that loss of life and injuries in the islands were kept to relatively low numbers. Hurricane Ivan is by no means the first high-intensity hurricane that the CI’s ever faced. In terms of casualties, the two deaths attributed to Ivan are dwarfed by the reported death toll of 109 deaths during the 1932 hurricane (Craton 2003). The 1903 hurricane had an estimated total of 15 men lost at sea, aboard vessels attempting to weather the storm. A hurricane that occurred in 1785 claimed great loss of life due to the collapse of houses in the strong winds, sailors lost at sea, and people drowning in a reported tidal wave. In 1876, 123 Nat Hazards Fig. 2 Tropical systems passing within 96 km of Cayman Islands since 1853. Tropical storms (TS): blue; Category 1 hurricane (HC1): green; Category 2 hurricane (HC2): yellow; Category 3 hurricane (HC3): orange; Category 4 hurricane (HC4): red; Category 5 hurricane (HC5): pink. Modified from UNECLAC (2004) all churches and most houses were reported to have been totally destroyed, while those still standing were severely damaged and barely habitable (Craton 2003). Fig. 2 shows the paths of tropical storms and hurricanes that passed within 96 km of the CI’s since 1853 (UN-ECLAC 2004). The number of tropical systems passing near the CI’s at 5-year intervals, from 1851 to 2006, is shown in Fig. 3. Most storms occurred from 1930 to 1934 Fig. 3 Number of tropical systems near Cayman Islands at five-year intervals for the period 1851–2006 reaching hurricane strength. Tropical storms: white; Categories 1–2 hurricane: gray; Categories 3–5 hurricanes: black (from Caribbean Hurricane Network: www.stormcarib.com) 123 Nat Hazards and the most severe hurricanes were registered between 1915 and 1919. On average, the CI’s are affected, brushed or hit by hurricanes every 2.23 years. The average number of years between direct hurricane hits (usually within 64 km to include small hurricanes) is once every 9.06 years. The months of September, October, and November are typically the most active for hurricanes in the islands. During these months, storms tend to form in the southern Caribbean and move north, into or close to CI’s. During the 55 years from 1950 to 2004, GC experienced seven tropical storms and seven hurricanes, and CB and LC six tropical storms and five hurricanes. The most important hurricanes that have directly impacted the CI’s in recent years, in addition to Ivan, are: • • • • • Gilbert, September 1988 Mitch, October 1998 Michelle, November 2001 Wilma, October 2005 Dean, August 2007 5.2 Storm surge and storm winds The greatest potential for loss of life related to a hurricane is from the storm surge. Storm surge consists of unusual volumes of water that are pushed toward the shore by the force of the winds swirling around the storm. This advancing surge combined with the normal tides creates the hurricane storm tide, which can increase the mean water level to heights impacting roads, homes, and the critical infrastructure. In addition, wind-driven waves are superimposed on the storm tide. This rise in water level can cause severe flooding in coastal areas, particularly when the storm surge coincides with the normal high tides. Storm surges combined with wave action are responsible for much of the damage usually caused by hurricanes, especially in large, low-lying coastal settlements. In addition to causing flooding and damage to coastal structures, storm surge may also precipitate flooding further inland through the blockage of the outfalls of drainage systems. During Hurricane Ivan, storm surges in GC flooded large portions of the coastal areas and deposited huge amounts of sand over roads, houses, and infrastructure from East to West Bay. Onshore winds produced storm surges of 2–3 m and wave heights in excess of 8 m (Young 2004). According to information from the local meteorological office, by 10 pm local time on 11 September, 2004, the center of Ivan was located 182 km SE of GC. At that time, hurricane winds with a force of over 161 km/h were already being experienced on the island (Young 2004). At 5 am, next day, the storm surge from the North Sound (Fig. 1) was peaking at 3 m. The hurricane made its closest approach at 10 am when the hurricane’s eye passed 34 km SW of the GC with winds of 241 km/h and gusts of 354 km/h. Another storm surge affected South Sound (Fig. 1) when the eye of the hurricane moved northeast. Ivan was a slow moving hurricane, which increased the exposure of the island to hurricane force winds as well as increased the total amount of rain. Key observations from the wind profile data during Hurricane Ivan (12 September, 2004) are (Young 2004; for locations see Fig. 1): • Category 4 winds were sustained over the George Town area for almost 4 h, peaking at 225–233 km/h between 9 and 11 am local time. • Category 4 winds were sustained in the West Bay area for around 3 h, peaking at 209– 217 km/h between 10 and 11 am local time. 123 Nat Hazards • Category 3 winds were sustained over the East End area for 4 h, peaking at 193–201 km/h at around 9 am local time. • Using a standard conversion factor, wind gusts likely reached 289 km/h over George Town. • Tropical storm force winds (63–117 km/h) started in East End at around 1 pm local time on 11 September and ended around 5 am local time on 13 September, a total period of 40 h. • Hurricane force winds ([ 117 km/h) were sustained somewhere in GC for almost 18 h. Young (2004), using the NOAA best track hurricane database from 1851 and a simple wind field model, estimated the peak winds during storms passing close to GC. His results, with an error of 10–15%, demonstrate that Hurricane Ivan has been the most intense storm in terms of wind speed on GC. The other two hurricanes with estimated winds of over 160 km/h were Gilbert in 1988 (172 km/h) and the 1903 storm (193 km/h). Other 21 storms were estimated with winds of 105 km/h or greater. Both storm surge flooding and wave action were strongly influenced by the evolution of Ivan’s wind-field as it passed GC. The geometry of North Sound caused the storm surge flooding through the Red Bay-Prospect neck and across the western peninsula (Young 2004). The main storm surge-flooding peak was recorded along South Sound, likely the result of winds having swung to onshore in that area. The continued onshore winds at South Sound prevented draining of flood waters until the late afternoon; backflow into North Sound occurred somewhat earlier than that, although southerly winds prevented draining of water over Seven Mile Beach late into the evening. Severe wave damage occurred only in a few places where onshore winds and no shallow reef protection came together (Young and Gibbs 2005). 5.3 Floods Climate in the Caribbean basin can be classified as dry-winter tropical with significant subregional variations in annual totals, length of the rainy season, and timing of rainfall maxima (Rudloff 1981). Flooding in this region has been the cause of death and of much property damage in hurricanes. Clearly, location is critical during floods. Low-lying lands, riverbanks, and lands adjacent to gullies should be avoided if possible; otherwise, appropriate drainage measures must be taken. ‘‘Tropical Waves’’ drift through the Caribbean in the summer months, often depositing large amounts of rainfall before they dissipate. Atmospheric hazards, especially excessive rainfalls, are the most important cause of floods. These vary from the semi-predictable seasonal rains over wide geographical areas, which give rise to the annual floods in tropical areas, to almost random convectional storms giving flash flood over some small basins (Smith 2001). Rainfall is seasonal in the CI’s. The season spans from mid-May to November, with May to June and September to October typically being the wettest months. Usually, the dry season takes place on February and March. Rainfall is generally the result of tropical thunderstorms, which develop in the summer months, or localized rain resulting from the evaporation of water in the central mangroves of the main island. Occasional surges of cooler air from continental North America, the leading edge of which is called a cold front, are the main winter system affecting the CI’s from late October through early April. These systems are the major producers of rainfall during the dry-winter months although precipitation is not quite as long lasting or of the same amounts as that brought by the summertime systems (Brunt and Davies 1994). 123 Nat Hazards The average annual rainfall in the CI’s is about 11.4 cm being October the month with higher rain precipitation, and the driest months are March and April (Fig. 4a). Localized rainfall usually results when the summer heat causes evaporation of water in the central mangrove wetlands and rain clouds are formed. These clouds generally drift to the west, depositing rain on the western side of the island. The wettest day on record is January 18, 2003, with 21.5 cm of rainfall (Fig. 4b). Other islands in the Caribbean with mountains register excessive rainfalls compared with the CI’s. For example, the Blue Mountains of eastern Jamaica record around 560 cm of rainfall per year, whereas Kingston, on the southeastern coast, receives about 400 cm. Bridgetown, the capital of Barbados, has an average annual rainfall of 127 cm, while Fig. 4 Average annual rainfall in different periods in Cayman Islands: a Average annual rainfall from 1957–2008; b Largest rainfall amount during a 24-hour period per year from 1971 to 2008 (Data from the Cayman Islands National Meteorological Service, 2008) 123 Nat Hazards Bathsheba on the central east coast reports 254 cm. Recording stations in the Northern Range in Trinidad measure some 302 cm of rainfall per year. Thus, in comparison, the registered rates of rainfall may indicate that flooding from rainfall can be considered a minor hazard to GC (Fig. 4a and 4b). However, if a tropical depression settles over the island, it can rain for days with flooding causing severe problems. Normally, heavy rain only takes place for a few hours, occasionally affecting some low-lying areas in the islands with moderate flooding. This is due to the fact that the soil of the island is mostly formed by limestone outcroppings with little soil (see Sect. 2). Two types of limestone form most of the surface: (1) bluff limestone and iron shore and (2) limestone with coral, mollusk shell, and marl (Alonso-Zarza and Jones 2007). Because limestone is very porous, rain is quickly absorbed by the soil avoiding high flooding. Weather-related hazards are usually directly modified by human activities in and around urban areas (Klein et al. 2003). These human activities may include: (1) changing sediment supply due to changing land use, hydrological modification, or coastal protection and the consequent influence on erosion and deposition; (2) land claim of intertidal areas and deepening of channels for navigation, which often increase extreme water levels and hence flood risk; and (3) increased subsidence due to groundwater withdrawal, which often reduces land elevation. Because of its geographic location, GC is exposed to many weather-related hazards. For these reasons, it is important for the population of GC to minimize or control this type of human activities to maintain flooding from rainfall as a minor hazard. In general, urbanization also increases the magnitude and frequency of floods (Smith 2001). Based on our discussion above and because of the present relatively low rates of average and daily rainfall (Fig. 4) and the soil conditions of the island, we consider flooding from rainfall as a minor hazard to GC and will not be taken into account in our analysis. 5.4 Earthquakes Although there are some reports of earthquakes felt strongly in the CI’s in the past, there is no evidence in the historical record of a major destructive earthquake occurring very close to these islands in the last 300 years. However, an event of magnitude 6.8 occurred on December 14, 2004, approximately 32 km to the south of Georgetown, GC. The earthquake did not cause damage on the island. It provides, however, a stark reminder that the CI’s lie along an active geological fault (Fig. 1). One may assume that the lack of strong events in the past 300 years means that seismic energy has accumulated on the Oriente fault and it may be released in the form of a large earthquake. Based on the available list of historical earthquakes, we estimated that the maximum magnitude of an earthquake on the Oriente fault might be between Mw7.2 and 7.5. This assessment of potential large earthquakes on the Oriente fault is made on the basis of the magnitude observed for the three largest events that have taken place along this plate boundary in the past 100 years: the Guatemala earthquake of February 4, 2006 (Mw7.5) that occurred on the westernmost segment of this plate boundary; the Honduras earthquake of May 28, 2009 (Mw7.3); and the Haitian earthquake of January 12, 2010 (Mw7.0). Wells and Coppersmith (1994) have made a compilation of displacements observed on geologic faults during many large earthquakes in the world. This compilation has led them to estimate relations of the displacement observed on faults showing different types of motion. Assuming that the earthquakes on the Oriente Fault would be of strike-slip type and in the range of magnitude 7.2–7.5, the relations of Wells and Coppersmith (1994) indicate that the expected motion on the fault during an earthquake would be 2–3 m. 123 Nat Hazards Using the average rate of motion of 0.6–1.1 cm/year reported by De Mets and WigginsGrandison (2007) for the Oriente fault, an approximate estimate of the recurrence rates might be obtained. The calculation of the repeat times is done by dividing the maximum displacement expected during an earthquake, which reflects the accumulated plate motion over time, by the rate of motion observed on the fault. The upper bound is given by the slowest rate of motion and the largest earthquake. Thus, if the maximum magnitude expected is 7.5, the motion for an earthquake of this type is 3 m (Wells and Coppersmith 1994). Assuming the rate of motion on the fault is 0.6 cm/year, the resulting recurrence rate is approximately 500 years. On the other hand, if the magnitude of the expected earthquake was 7.2, the motion of the fault would be expected to be approximately 2 m. Assuming the highest rate of motion (1.1 cm/year) in order to determine the lower bound of the recurrence times, the resulting return period for these relatively smaller earthquakes would be approximately 180 years. Thus, the return periods of earthquakes in the range of magnitudes 7.2–7.5 in the Oriente fault would be approximately 180–500 years. These figures, based on observed plate motion rates and expected maximum magnitudes, provide a first-order estimate of the return periods of large earthquakes in the CI’s. 5.5 Tsunami The main sources for tsunami in the Caribbean are earthquakes (generated at the boundaries of the Caribbean Plate or by intra-plate events). No evidence of tsunamis impacting the CI’s exists since they were first permanently settled, 300 years ago. In the CI’s, due to their geological characteristics, lack of rivers, and low topographic relief, there are no sedimentary basins or large marginal deposits along the coast that may potentially slide if they were to be made gravitationally unstable by the occurrence of a large earthquake or a major storm, for example. Thus, local tsunamis originating from large landslides along the coasts of the islands are highly unlikely. Potentially, the CI’s may be affected by tsunamis generated in other parts of the Caribbean Sea and striking the coastlines of the islands. The bathymetry of the islands, however, does not exhibit a continental shelf that shallows gradually toward the coastline producing the amplification of tsunami waves, as it happens on the continents. The CI’s coastline rises sharply from the ocean bottom, and this bathymetry may not give rise to the rapid amplification of tsunami waves. An analogy to the CI’s is the effect observed on Diego Garcia Island in the Indian Ocean during the tsunami of December 26, 2004. Although the island was on the path of the incoming tsunami waves, the island was essentially unaffected. Nonetheless, the coast of Africa, several hundreds of kilometers downstream from Diego Garcia, suffered considerable damage due to the large tsunami wave generated along the coast. The reason for this is that the island of Diego Garcia is a dormant volcano that rises sharply from the ocean floor. The lack of a sloping continental shelf did not generate the large tsunami waves observed on the continental shorelines. Jones and Hunter (1992) identified large boulders that are as heavy as 10 tons along the southern coast of GC. The boulders are derived from the Bluff Formation that forms the rocky coastal terraces of present-day shore. The presence of sponge borings, encrusted gastropods, and other marine flora and fauna indicates that the boulders must have been submerged in seawater before being thrown inland. The authors dated with radiocarbon techniques the coral of these boulders. This dating yields an age of 1,656 ± 32 years AD. This age apparently reflects the time when these large boulders were removed from the seafloor and thrust onshore over vertical distances of 18 m in some cases. The authors 123 Nat Hazards suggest that the most probable cause for the presence of these boulders is a local tsunami or a major hurricane. 5.6 Landslides As discussed above, the islands are generally flat lying with the highest point being 42 m. Coastal cliffs are unlikely to exceed a few meters in this area of subdued topography. There is not reported landslide activity not geological evidence for it in the CI’s. Occasional falls of limestone blocks at low sea cliffs are possible, particularly in CB. Thus, landslides do not represent a main hazard that is likely to have an economic or social impact in the islands (McNally 1988; Richards 1975). 5.7 Volcanoes Several of the islands of the Eastern Caribbean are volcanic in origin. However, there are no active volcanoes on the CI’s. Also, the islands are far from the known active volcanoes in the Caribbean area. The closest volcanoes are those in the Lesser Antilles and in Central America. The long distance of the CI’s to active volcanoes indicates that volcanic hazard is practically non-existent. Thus, we consider that no further evaluation is needed in this respect. 5.8 Man-made hazards Potential man-made hazards in the islands are (in alphabetical order): • • • • • • • • • Acid rain Air pollution Electronic threats Environmental degradation Fires Infrastructure failure Major airplane, cruise ships, medical and agricultural accidents Mined and quarried land Water pollution There are no major industries or toxic chemicals in GC. Nonetheless, the fuel terminals and fuel storage tank, as well as the fuel distribution pipeline, may constitute a potential man-made hazard. The main fuel terminals and tanks as well as the fuel distribution line could, under certain circumstances, exacerbate the impact of a hazard event by adding to the danger. Fuel leaks may result through safety and relief valves, piping ruptures, fire, equipment failures, overfills and overflows of storage tanks, and human error. In these cases, the hazard posed by these facilities can be explosive, flammable, combustible, and toxic, posing a potential risk to life, health, environment, or property that is located nearby. Incidents with an involvement of structural or infrastructural system failures may consist of technical, operational, or organizational failures. Pipelines are transportation arteries carrying liquid and gaseous fuels. Pipelines are buried and sometimes located above ground. Buried and exposed pipelines are vulnerable to breaks and punctures caused by earth movement, material failure, operator error, construction defects, and tampering. If a pipeline fails during land movement, it can shear. 123 Nat Hazards When shearing takes place across abrasive materials or when the fuel comes in contact with an ignition source, the sparks can cause the fuel to explode or burn. 6 Hazard analysis From the results of the previous section, the following main hazards to CI’s are addressed here: • • • • Hurricane (flooding, storm surge) Earthquake Man-made hazard (fuel and gas terminals and fuel pipeline) Tsunami In the case of hurricanes, we consider flooding and storm surge as the main hazards for the following reasons: (1) The two elements are the main causes of hurricane related deaths; (2) Flooding from hurricanes can occur several miles from the coast impacting communities and structures far inland; (3) Storm surges are responsible for much of the damage usually caused by hurricanes, especially in large, low-lying coastal settlements; (4) These two phenomena usually present themselves simultaneously. Using Eq. 2, we developed a relative priority matrix as a general guide for addressing the different hazards. The frequency, area of impact, and magnitude values of the potential damage are defined by a scale of numbers ranging from 1 to 6: Extremely low: 1 Very low: 2 Low: 3 Moderate: 4 High: 5 Very high: 6 Based on this scale, and considering the history of hazards in the CI’s and their past impact discussed in the previous section, we assigned values of 6, 2, 2, and 1 for the frequency of occurrence and of 6, 4, 1, and 2 for the area of impact due to hurricane, earthquakes, man-made hazards, and tsunamis, respectively (Table 1). The values of 5, 4, and 3 for the potential damage of hurricane, earthquake, and man-made hazards are also justified from the history of past impacts. We assigned 2 to a tsunami’s potential damage because most likely only the coastal zones of CI’s would be impacted. It should be emphasized that the results indicated on Table 1 are indicative of the relative level of hazard across CI’s. Although qualitatively defined, the higher the total scores, the greater the relative degree of ranking of the hazard. Table 1 Relative ranking matrix of hazards at Cayman Islands a (Frequency ? area impact) 9 potential damage magnitude 123 Hazard Frequency Area of impact Potential damage Total scorea Hurricane 6 6 5 60 Earthquake 2 4 4 24 Man-made 2 1 3 9 Tsunami 1 2 2 6 Nat Hazards Following Step 2 of our methodology, to outline the exposure areas at GC, we considered the characteristics of each main natural hazard listed in Table 1. We targeted these areas by establishing relative rankings among them. The varying degrees of exposure areas are represented both through a relative scoring system (higher scores for areas exposed to high impact hazards) and graphically in a map. For example, qualitative flood maps were created for five different categories of storms. Storms of Category 1 are generally associated with the least severe winds and floods, while Category 5 storms are considered to be the most severe. Generally, areas subject to floods in the lower category storms are also projected for inundation in all of the higher categories. Thus, in our relative priority scoring system for inundation, storm floods areas of Category 1 would, therefore, have the highest probability of being flooded since they are exposed to inundation in all storm events. Table 2 shows the relative priority scoring system developed for different areas at GC. To facilitate our interpretation, the results are also presented in the corresponding maps (Fig. 5a–5c). The general concept is that locations with no exposure to hazards will have a score of 0 and each incremental increase in exposure adds one point: Table 2 Level of hazards for different areas at Grand Cayman Hazard Area Score Highest Lowest Categories 1 and 2 1 5 Category 3 2 4 Categories 4 and 5 3 3 4 2 Categories 1 and 2 1 5 Category 3 2 4 Categories 4 and 5 3 3 Remainder of Grand Cayman 4 Island 2 Hurricane Flooding (Fig. 5a) Remainder of Grand Cayman Island 5 2 5 2 Storm surge (Fig. 5b) Earthquake Entire Grand Cayman Island 1–4 1 1 1 Very low Near to ocean 1 1 0 Remainder of Grand Cayman Island 4 0 Fuel and gas tanks Adjacent areas 1 1 0 Fuel pipeline Adjacent areas 1 13 5 Tsunami (Fig. 5c) Man-made hazard Remainder of Grand Cayman Island Total 0 123 Nat Hazards a b c Fig. 5 Flood a and Storm surge areas b for different hurricane categories. The arrow indicates the direction of approach of the hurricane; c Tsunami hazard areas for tsunamis coming from the Caribbean Sea No risk: 0 Very low: 1 Low: 2 123 Nat Hazards Moderate: 3 High: 4 Very high: 5 Notice that under our considerations, for the entire GC the score of 1 in Table 2 is given only for earthquake hazard since, with the knowledge we have today, there is potential for this hazard to occur evenly distributed throughout the country. Some variations may exist due to varying soil conditions in GC. In some cases, soft or thick soils induce much higher accelerations of the ground than would normally be observed on hard rock. The information available and the scope of this study do not allow a more detailed seismic zoning based on soil quality. 7 Results and discussion Flood zones resulting from hurricanes were delineated according to hurricane categories on the Saffir-Simpson scale (Categories 1 through 5). The exposure scores for flood hazards were determined by: (1) Using the flood distribution areas during hurricane Ivan as determined by the Lands and Survey Department of Cayman Islands (2005) and (2) Identifying areas with high potential for flooding by analyzing the characteristics of the topography of GC from a Digital Elevation Model (DEM) and using a GIS. The score of 5 was given to the area outlined in Fig. 6a considering that: (1) Zones where coastal flooding and wave action are the highest during Categories 1 and 2 hurricanes are also likely to be inundated during stronger events as well and (2) On average, Categories 1 and 2 hurricanes hit the CI’s every 2.23 years (Young 2004). The second highest score of 4 was applied to Category 3 flood areas (Fig. 6b) because hurricanes of this size usually hit the islands once every 9.06 years (Young 2004). A score of 3 is given to hurricane Categories 4 and 5 (Fig. 6c) that take place approximately every 100 years (Young 2004). Areas located outside the floodplain but appearing on flood-prone soils are rated with a score of 2. However, as Young (2004) indicated, two 100-year hurricanes can occur in successive years; the fact that Ivan occurred in 2004 does not give GC a guarantee of 99 years without a similarly sized storm. The results displayed in Fig. 6c are comparable to the results obtained from field observations by the Lands and Survey Department (2005). This institution showed that up to 70% of GC was submerged by seawater and/or rainwater for significant periods during and after hurricane Ivan. However, around 20% of this flooding was less than 60 cm. Note that in Fig. 6c, we are combining the results for hurricanes with Categories 4 and 5. If we consider a hurricane with Category 4 (results not shown here), only about 70% of GC is flooded as reported by the Lands and Survey Department (2005). Locations that are subject to storm surges from the lowest category storm events are considered the most exposed areas for this type of phenomenon. The identified storm surge areas at GC were mapped from: (1) The storm surge areas during hurricane Ivan reported by Young (2004) and (2) The characteristics of the topography of GC analyzed from a DEM and using a GIS. These zones represent locations that might expect to be impacted by storm surge during future hurricane events. Because of the difficulty in making clear boundaries, a 0.25-mile buffer was established around the identified surge inundation zones. Categories 1 and 2 hurricane storm surge inundation areas are given a score of 5 (Fig. 7a). Category 3 inundation areas are given a score of 4 (Fig. 7b), and Categories 4 123 Nat Hazards a b c Fig. 6 Areas showing their level of exposure due to flooding for hurricanes: a categories 1 and 2; b category 3; and c categories 4 and 5. The arrow indicates the direction of approach of the hurricane and 5 inundation areas are assigned a score of 3 (Fig. 7c) because it is considered that these kinds of events take place approximately every 100 years at CI’s (Young 2004). The results of Fig. 7c are in agreement with observations made by Young (2004). This author points out that the geometry of the North Sound caused the storm flooding through the Red Bay-Prospect neck and across the western peninsula during hurricane Ivan. According to 123 Nat Hazards a b c Fig. 7 Areas showing their level of exposure due to storm surge for hurricanes: a categories 1 and 2; b category 3; and c categories 4 and 5. The arrow indicates the direction of approach of the hurricane Young (2004), the southern parts of North Sound (Fig. 1) were first extensively flooded as a result of the early northeasterly winds. Then, as the winds swung around to the east, water began to inundate the western shore of North Sound and, in places, pushed across the western peninsula to Seven Mile Beach. At the peak of the storm, easterly winds over 123 Nat Hazards western GC and southeasterly over eastern GC caused storm surge inundation from both the western side of North Sound and from South Sound simultaneously. The highest flooding due to storm surges was recorded along South Sound (Fig. 1). This fact is very likely the result of winds having swung onshore in that area. Similar flooding has occurred on a number of occasions in the history of GC (Young 2004). We consider that the seismic risk score is 1 throughout GC. This result is based on the following considerations: (1) There is no evidence in the historical record of a major, destructive earthquake occurring in the vicinity of the CI’s in the last 300 years and (2) Taking into account our analysis of earthquakes presented in Sect. 5, the return period of large earthquakes (magnitudes between 7.2 and 7.5) at CI’s is approximately in the range of 180–500 years. These observations can be extrapolated to the CB and LC because the geological fault where these large earthquakes may be generated lies parallel and to the south of GC. Furthermore, there are no observations that would allow one to assess the effect of local soil conditions in different parts of the islands. Deep and soft soils usually normally produce larger accelerations and seismic intensities than those observed at sites located on rock. Therefore, we consider that the three islands probably have the same response to seismic waves. As indicated in Fig. 5c, areas exposed to the impact for tsunamis were determined based on distance from the coastline along the shore. Because of the low probability of the CI’s to be impacted by a tsunami those areas inland of the coastline up to 0.2 miles are assigned a score of 2 (low exposure). The remaining zones on the islands receive a rating of 1 (very low exposure). As discussed above, we are considering that the CI’s may be affected by tsunamis generated in other parts of the Caribbean Sea and striking the coastlines of the islands. The bathymetry of the islands, however, does not exhibit a continental shelf that shallows gradually toward the coastline producing the amplification of tsunami waves, as it happens on the continents. As pointed out before, the main fuel tanks, the Home Gas facilities, and the associated pipelines constitute the main man-made hazards on GC. We assigned a score of 1 to zones that may be impacted in the case of a major leak or an explosion. These areas are defined as: (1) The circle centered on each of the five major storage tanks with 400 m radius and (2) The area measured within 400 m distance from both sides of the distribution pipeline. These zones of impact for man-made hazards at GC were based on studies of the affected zones from explosions reported in similar circumstances (Arturson 1987; Sceptre Fundraising Team 2006). Based on our previous discussion, GC can be divided basically into four areas with different levels of exposure to natural and man-made hazards. These areas are (Fig. 8): • Area 1: North Sound, Little Sound, and Eastern West Bay. • Area 2: Central Mangroves, Central Bodden Town, Central George Town, and West Bay. • Area 3: Northwestern West Bay, Western Georgetown, Bodden Town, East End, and the North Side. • Area 4: Remainder of country. Using a GIS, the relative individual scores for each area were added to create a total score for that particular region (Table 3). From Table 2, we observe that the lowest and highest potential risk scores are 5 and 13, respectively. These figures represent the maximum values for low and very high exposure zones. Under these conditions, our ranking to establish the level of exposure for the four identified areas in GC is the following: 123 Nat Hazards Fig. 8 Areas showing the level of exposure to natural and man-made hazards at Grand Cayman resulting from the analysis described in this paper Table 3 Levels of exposure to natural hazards per area at Grand Cayman Area Hurricane Earthquake Tsunami Manmade Flooding Storm hazard surge Total Level of score exposure Area 1: North Sound, Little Sound, and Eastern West Bay 5 5 1 0 0 11 Very High Area 2: Central Mangroves, Central Bodden Town, Central George Town, and West Bay 4 4 1 1 0 10 High Area 3: Northwestern West Bay, 3 Western Georgetown, Bodden Town, and East End-North Side 2 1 1 1 8 Moderate Area 4: Remainder of the island 1 1 1 0 5 Low 2 Low: 5–6 Moderate: 7–8 High: 9–10 Very high: C 11 Using these values, the qualitative level of exposure to natural hazards in the four identified areas is established (Table 3, Fig. 8). The most exposed areas to natural and man-made hazards at GC are the North Sound, Little Sound, Eastern West Bay, the Central Mangroves, Central Bodden Town, Central George Town, and the West Bay regions. The Northwestern West Bay, Western Georgetown, Bodden Town, East End, and the North Side zones are considered to be areas with moderate exposure. The remainder of the country has low exposure to hazards. 123 Nat Hazards 8 Conclusions Today, the economy of Cayman Islands is based upon two main pillars: financial services and tourism. Determining the level of exposure of the islands to natural and man-made hazards for development plans and land usage regulations is necessary. Mitigation projects should be performed from a multi-hazard (hurricanes, earthquakes, tsunamis, and manmade hazards) perspective considering the different qualitative levels of physical vulnerability for the four exposed areas identified here: • Area 1: North Sound, Little Sound, and Eastern West Bay: Very high vulnerability. • Area 2: Central Mangroves, Central Bodden Town, Central George Town, and West Bay: High vulnerability. • Area 3: Northwestern West Bay, Western Georgetown, Bodden Town, East End, and the North Side: Moderate vulnerability. • Area 4: Remainder of country: Low vulnerability. We consider that the hazard identification and ranking obtained here are the same for GC, CB, and LC. The hazard risk areas that were identified for GC can be determined in the same manner for CB and LC if the appropriate information is collected and structured. In fact, as discussed above in the case of GC, the earthquake hazard risk would be the same for the three islands. A natural continuation of this work would be a future quantitative risk analysis of the Cayman Islands. The results presented here are meant to provide the basis for this future quantitative assessment of risk. Acknowledgments The authors are grateful to two anonymous reviewers for their thoughtful review of this work and providing helpful comments. We also thank B. Carby for her feedback in the development of this study. Special thanks are due to F. McCleary for his assistance and support during this work. J. Chacon provided information on the meteorological hazards of the Cayman Islands. Our appreciation to the staff of the Hazard Management Cayman Islands that made this study possible and who kindly provided their expertise, time, and data. We thank the staff of the various departments and offices of the government of the Cayman Islands who provided information, comments, and helpful discussions for the development of this study. E. Zúñiga provided help for preparing the figures. References Alcantara-Ayala I (2001) Geomorphology, natural hazards, vulnerability and prevention of natural disasters in developing countries. Geomorphology 47:107–124 Alonso-Zarza AM, Jones B (2007) Root calcrete formation on quaternary karstic surfaces of Cayman Islands. Geol Acta Int Earth Sci J 5:77–88 Arturson G (1987) The tragedy of San Juanico, the most severe LPG disaster in history. Burns Incl Therm Inj 13:87–102 Brunt MA, Davies JE (eds) (1994) The Cayman Islands: natural history and biogeography. Kluwer, Boston, p 604 Cayman Islands National Meteorological Service (2008) Series of Daily, Monthly and Annual Precipitation of the Cayman Islands, Cayman Islands Government Central Planning Authority (2006) Development Plan Map 2006, Cayman Islands Government Consultative group for the reconstruction and transformation of Central America (1999) Reducing vulnerability to natural hazards: lessons learned from Hurricane Mitch. A strategy paper on environmental management. Available via http://www.iadb.org/regions/re2/consultative_group/groups/ecology_work shop_1.htm Craton M (2003) Founded upon the Seas: a history of the Cayman Islands and their people. Ian Randle, Kingston, 532 pp Crichton D (1999) The risk triangle in natural disaster management. Tudor Rose, Leicester 123 Nat Hazards De Mets C, Gordon RG, Argus DF, Stein S (1994) Effect of recent revisions to geomagnetic reversal timescale on estimate of current plate motions. Geophys Res Lett 21:2191–2194 De Mets C, Jansma PE, Mattioli GS, Dixon TH, Farina F, Bilham R, Calais E, Mann P (2000) GPS Geodetic constraints on Caribbean-North America plate motion. Geophys Res Lett 27:437–440 De Mets C, Wiggins-Grandison M (2007) Deformation of Jamaica and motion of the Gonavemicroplate from GPS and seismic data. Geophys J Int 168:362–378 Department of Environment (2004) Preliminary Post-Ivan Environmental Impact Assessment Report, Cayman Islands Government Emergency Operation Center (2006) Cayman Islands’ National Hurricane Plan 2006, Cayman Islands Government Folk R, Roberts H, Moore C (1973) Black phytokarst from hell, Cayman Islands, British West Indies. Geol Soc Am 84:2351–2360 Hazard Management Cayman Islands (2009) Map of location (latitude and longitude) of critical facilities (hospitals, schools, shelters, fuel deposits, fuel and gas pipeline, government communications infrastructure, power stations, ports, water and sewage treatment plants, water storage plants, airport, police and fire departments, critical government and red cross installations, Cayman Islands Government Heinz III Center for Science, Economic and the Environment (1999) The hidden costs of coastal hazards: implications for risk assessment and mitigation. In: White GF (foreword). Island Press, Washington DC, 252 pp Jones B, Hunter IG (1992) Very large boulders on the coast of Grand Cayman: the effects of giant waves on rocky coastlines. J Coast Res 8:763–774 Klein RJT, Nichols RJ, Thomalla F (2003) The resilience of coastal megacities to weather-related hazards: a review. In: Kreimer A, Arnold M, Carlin A (eds) Building safer cities: the future of disaster risk, disaster risk management, series no.3. World Bank, Washington DC, pp 101–120 Lands and Survey Department (2005) Map of flooding areas during Ivan Hurricane, Assessment, Cayman Land Info, Cayman Islands Government Lands and Survey Department (2005) Map of Hurricane Ivan Preliminary Damage Assessment, Cayman Land Info, Cayman Islands Government Lands and Survey Department (2007) Cadastre Map of the Cayman Islands, Cayman Land Info, Cayman Islands Government Lands and Survey Department (2007) Petroleum Products Location Map, Cayman Land Info, Cayman Islands Government Lands and Survey Department (2007) Terrain and Bathymetry Map, Cayman Land Info, Cayman Islands Government Lands and Survey Department (2007) Topographic Map of the Cayman Islands, Cayman Land Info, Cayman Islands Government Leroy S, Mauffret A, Patriat P, De Lepinay M (2000) An alternative interpretation of the Cayman Trough evolution from a re-identification of magnetic anomalies. Geophys J Int 14:539–557 McNally R (1988) World atlas of nations. Rand McNally, New York North Caroline Department of Environment and Natural Resources (2001) North Caroline of Crime Control and Public Safety, University of North Caroline at Wilmington, Federal Emergency Management Agency, NOAA Coastal Services Center, Vulnerability Assessment Tutorial, New Hanover County/ Wilmington Project Impact Partnership. Available via http://www.csc.noaa.gov/products/nchaz/htm/ methov.htm Rashed T, Weeks J (2003) Assessing vulnerability to earthquake hazards through spatial multicriteria analysis of urban areas. Int J Geogr Info Sc 17:547–576 Richards HG (1975) Cayman Islands. In: Fairbridge RW (ed) The encyclopedia of world regional geology, part 1: Western Hemisphere. Dowden, Hutchinson and Ross, Stroudburg, pp 226–227 Rosencrantz E, Mann P (1991) Sea MARC II mapping of transform faults in the Cayman Trough. Caribbean Sea Geol 19:690–693 Rudloff W (1981) World-climates, with tables of climatic data and practical suggestions. Wissenschaftliche Verlagsgesellschaft mbH Stuttgart, Germany, p 632 Sceptre Fundraising Team (2006) The Buncefield explosion. Sceptre Education, 128 pp Scheidegger AE (1994) Hazards: singularities in geomorphic systems. Geomorphology 10:19–25 Schneiderbauer S, Ehrlich D (2006) Social levels and hazard (in) dependence in determining vulnerability. In: Birkmann J (ed) Measuring vulnerability to natural hazards—towards disaster resilient societies, United University Press, pp 78–102 Smith K (2001) Environmental hazards: assessing risk and reducing disaster, 3rd edn. Routledge, London, p 392 123 Nat Hazards Stein R, De Mets C, Gordon RG, Brodholt J, Engeln JJ, Wiens DA, Argus D, Lundgren P, Stein C, Woods D (1988) A test of alternative Caribbean plate relative motion models. J Geophys Res 95:965–981 Thywissen K (2006) Components of risk: a comparative glossary, Report no. 2. Institute for Environment and Human Security, United Nations University Bonn, Germany United Nations Economic Commission for Latin America and the Caribbean (UN-ECLAC) (2004) The impact of hurricane Ivan in Cayman Islands, 82 pp Wells DL, Coppersmith KJ (1994) New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement. Bull Seismol Soc Am 84(4):974–1002 Young S (2004) Impact of hurricane Ivan in Grand Cayman: understanding and quantifying the hazards. GeoSY Ltd, 51 pp Young S, Gibbs T (2005) Impact of hurricane Ivan in Grand Cayman: a technical review of the hazards and their effects. GeoSY Ltd, 47 pp 123
© Copyright 2025 Paperzz