4

Surface Wind Speeds and Property Damage

Richard D. Marshall, National Institute of Standards and Technology, Gaithersburg, Maryland

INTRODUCTION

To assess properly the performance of buildings and other structures subjected to extreme natural events such as Hurricane Hugo and to establish the severity of the event itself, it is essential to have accurate wind-speed information. It is not unusual to encounter reports of extraordinarily high wind speeds following a destructive hurricane, and Hugo was no exception. To verify their accuracy, these reports must be tracked to their source. This can be a lengthy and difficult process because it usually takes place well after the investigative team has completed its field work and before communication systems have been fully restored. This was the case for the U.S. Virgin Islands, St. Croix in particular, where telephone service was seriously disrupted for more than 4 months. The problem of ascertaining true wind speeds is usually complicated by inaccurate reporting by the news media and, to some extent, by local authorities and government agencies. Such reports, when cited often enough, begin to generate their own credibility.

Ideally, it should be possible to retrieve records (daily logs and/or stripcharts) of wind speed and direction from sites that make routine weather observations. In a typical analysis of surface-wind speeds, the records are adjusted as necessary to standard conditions of wind exposure and are presented in a consistent format of sustained (1-min mean) speeds. In the case of Hurricane Hugo, however, only two such records were obtained from all of the potential recording sites along Hugo's track through the Virgin Islands and across northeastern Puerto Rico. In those areas where surface-wind speeds were not recorded, it has been necessary to resort to other sources of data and to other indicators of wind speed.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA 4 Surface Wind Speeds and Property Damage Richard D. Marshall, National Institute of Standards and Technology, Gaithersburg, Maryland INTRODUCTION To assess properly the performance of buildings and other structures subjected to extreme natural events such as Hurricane Hugo and to establish the severity of the event itself, it is essential to have accurate wind-speed information. It is not unusual to encounter reports of extraordinarily high wind speeds following a destructive hurricane, and Hugo was no exception. To verify their accuracy, these reports must be tracked to their source. This can be a lengthy and difficult process because it usually takes place well after the investigative team has completed its field work and before communication systems have been fully restored. This was the case for the U.S. Virgin Islands, St. Croix in particular, where telephone service was seriously disrupted for more than 4 months. The problem of ascertaining true wind speeds is usually complicated by inaccurate reporting by the news media and, to some extent, by local authorities and government agencies. Such reports, when cited often enough, begin to generate their own credibility. Ideally, it should be possible to retrieve records (daily logs and/or stripcharts) of wind speed and direction from sites that make routine weather observations. In a typical analysis of surface-wind speeds, the records are adjusted as necessary to standard conditions of wind exposure and are presented in a consistent format of sustained (1-min mean) speeds. In the case of Hurricane Hugo, however, only two such records were obtained from all of the potential recording sites along Hugo's track through the Virgin Islands and across northeastern Puerto Rico. In those areas where surface-wind speeds were not recorded, it has been necessary to resort to other sources of data and to other indicators of wind speed.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA SOURCES OF DATA Radar images recorded on film by the WSFO in San Juan were particularly valuable in establishing wind directions and the presence of intense rainbands during the passage of Hugo. WSFO radar began tracking Hugo southeast of St. Croix early on September 18 and provided essentially continuous tracking until Hugo passed out of range late in the day. Also of value in assessing the relative strength of Hugo and, hence, the intensity of surface winds, were estimates of central pressure and eye diameter presented in post-storm summaries prepared by the NHC (Lawrence, 1989). Other sources of information were the flight-level records obtained by NOAA and USAF reconnaissance aircraft, satellite data, wind damage observations, and reports by residents of the affected areas. In the following paragraphs is an assessment of surface-wind speeds for each of the major areas affected by Hugo. Also included is a description of the reported speeds and the final disposition of these reports. To be consistent with standard wind-speed measurement and reporting procedures, the speeds described in the following sections refer either to peak gusts or sustained speeds at a height of 10 m in flat, open terrain, typical of airport exposures. It is important to note that wind speeds measured under nonstandard conditions may differ widely from standard measurements. In order to gain some perspective of the severity of the estimated and measured surface winds, comparisons are made with design wind speeds specified by local and regional building codes in the Caribbean. Finally, the extraordinarily high surface-wind speeds reported for St. Croix and St. Thomas are cast in terms of mean recurrence intervals (MRIs) derived from a recent study of wind statistics for the Caribbean region. ASSESSMENT OF SURFACE WIND SPEEDS St. Croix The first indications of rainbands over St. Croix appeared on San Juan radar at about 0330 GMT on September 18. These rainbands, believed to contain locally intense winds, formed to the north and east of St. Croix and moved over the area around Christiansted from the northeast. However, eyewitnesses said the high winds began as early as 0200 GMT when Hugo was approximately 150 km to the southeast. Radar data confirm these accounts, indicating the storm's eyewall passed over the Frederiksted area at about 0600 GMT. By 0640 GMT all but the eastern end of St. Croix was within the eye and Hugo began to stall 10 to 20 km south of Frederiksted. From 0700 to 0800 the eyewall was located over the central section of St. Croix, with the eastern end of the island experiencing strong winds from the southeast to south. By 0900 Hugo was approximately 20 km west of Frederiksted, and the western

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA portion of the island began to experience strong winds out of the southwest. By 1000 GMT Hugo was approximately 30 km northwest of St. Croix. A preliminary report issued by the NHC on October 26 (Lawrence, 1989) reported maximum sustained surface winds of 120 knots (138 mph) for St. Croix at 0600 GMT on September 18. In the weeks following the storm, reports circulated of gust speeds in excess of 175 knots (201 mph) at Alexander Hamilton Airport and a peak gust of 186 knots (214 mph) at the nearby Hess Refinery. A travel news magazine (Showker, 1990) reported speeds of 174 knots (200 mph) for St. Croix, and the December 4 issue of Time (Gibbs et al., 1989) put the maximum speeds at 191 knots (220 mph). By tracking these reports to their sources, it has been possible to dispel some of the more spectacular claims of high wind speeds. With regard to wind-speed measurements at the airport, facilities of the FAA were shut down well before Hugo made landfall, and no wind-speed observations were obtained. In fact, the control tower was heavily damaged along with the weather instrumentation. FAA personnel stated that, to their knowledge, no other agency or party at or near the airport had the capability of obtaining wind-speed measurements. In tracking the report of 186 knots at the Hess Refinery, it was established that the anemometer site is located at the Hess Marine Agency directly southwest of the refinery. Company employees stated that the anemometer was blown away well before the strong winds arrived and that no actual measurements of wind speed were made. Thus, it appears that on St. Croix no verifiable wind-speed measurements were obtained during the passage of Hurricane Hugo. An aerial survey of St. Croix by three members of the CND investigative team 5 days after the storm indicated that the most severe damage was located along the north coast from the mouth of the Salt River eastward to the end of the island. This observation is consistent with the San Juan radar images, which show very intense convective bands developing to the northeast of Christiansted and moving over the area directly off the ocean from about 0530 to 0630 GMT. The terrain around Christiansted and to the east slopes steeply upward from the coast to a major ridge extending east-west the length of the island. In places this slope exceeds 150 m in the first kilometer and caused local acceleration of the surface winds. Damage around Frederiksted on the southwest corner of St. Croix was less severe, but areas of heavy damage were observed in the south-central portion of the island near the airport and at the oil refinery. Wind damage on St. Croix was strikingly similar to that observed at Darwin, Australia, following Cyclone Tracy (Marshall, 1976). For the Darwin subdivisions experiencing the heaviest damage, the peak gusts averaged 135 knots (155 mph), and it is considered likely that the peak gusts on St. Croix were of similar magnitude. Analyses of numerous stripchart records from Hugo in South Carolina and from previous Atlantic and Gulf Coast hurricanes indicate that the ratio of peak gust to corresponding sustained speed is in the range 1.20 to 1.25. Therefore, if 135 knots is representative of the maximum gusts, the corresponding maximum sustained wind speed on St. Croix would be approximately 110 knots (127 mph). Because of the

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA accelerating effect of the terrain on surface winds near Christiansted, wind speeds based on observed local damage would tend to be overestimated. An independent estimate of surface-wind speeds can be obtained from data collected by reconnaissance aircraft at the 700 mb level. These data indicate maximum speeds of about 130 knots (150 mph) during Hugo's passage over St. Croix (Lawrence, 1989). Data presented by Powell and Black (1989) suggest that the ratio of the 10-min mean speeds at 10 m over rough water to the 700 mb flight-level speeds is approximately 0.7, which in this case yields a 10-min mean speed of 90 knots (104 mph). Again, from the analyses of stripchart records mentioned above, the ratio of the maximum sustained speed to the corresponding 10-min mean speed in flat, open terrain is approximately 1.2. Thus, the corresponding sustained speed for standard exposure conditions is about 110 knots (127 mph). In November 1990, well after the completion of the research for this study, it was learned that two U.S. Navy workboats, the Acoustic Explorer and the Acoustic Pioneer, were moorerd in Krause Lagoon (adjacent to the Hess refinery) during the passage of Hugo. The Pioneer registered a peak gust of 140 knots (161 mph) before its anemometer was blown away. The Explorer registered a peak gust of 146 knots (168 mph). Anemometer heights and superstructure blockage effects are not known. However, given the over-water wind exposure, these observations are consistent with the 135- knot (155-mph) gust speed cited earlier for a standard wind exposure. St. Thomas Based on San Juan radar images, intense rainbands formed over St. Thomas and St. John at about 0930 GMT and persisted until 1130 GMT. During this time interval, the wind directions shifted gradually from northeast to southeast. Based on observations of damage made from the air, it was clear that the wind speeds over St. Thomas were not as high as those over St. Croix. As was the case with St. Croix in the weeks following the passage of Hurricane Hugo, there were persistent reports of wind speeds in excess of 174 knots (200 mph). One such report was traced to Cyril E. King Airport on St. Thomas. Contact with the local FAA staff revealed that procedures similar to those at Alexander Hamilton Airport were implemented when it became clear that Hugo would produce strong winds in the area. Operations at the airport were discontinued early in the afternoon of September 17, halting all official weather observations. Damage to FAA facilities was far less severe than on St. Croix, involving the loss of a satellite antenna and some windows in the control tower cab. Because of misinformation regarding design criteria for FAA facilities located in the Caribbean, some local staff members were under the impression that speeds of at least 174 knots (200 mph) would have been required to cause the observed damage. However, a check of FAA design requirements indicated that these facilities were required to comply only with local code provisions for wind loading. Thus, the required design wind speed would be approximately 70 knots (81 mph), rather than the 200 mph figure that was widely

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA cited. Additionally, the general level of damage at the airport clearly was inconsistent with such high speeds. As on St. Croix, it was not possible to verify wind-speed records on St. Thomas. However, based on the observed damage, the storm track, and flight-level wind data, the field team believes the maximum surface-wind speeds on the island were about 85 knots (98 mph), with peak gusts of 105 knots (121 mph). Vieques Because of the long east-west extent of Vieques and the changing direction of the adjacent storm track, it is difficult to generalize about the wind directions for Vieques. As indicated by rainbands registered by San Juan radar, strong winds out of the northeast impacted the eastern half of Vieques by 0930 GMT. However, it is likely that severe winds were felt in this area as early as 0800 GMT, while Hugo was still off the southwest tip of St. Croix. Based on the estimated track position and a 35-km eye diameter, it is likely that the eye began to move over the eastern end of Vieques shortly before 1100 GMT. An eyewitness at Esperanza on the south coast said the eye arrived after daylight (1100 to 1130 GMT). At the beginning of the lull, which was estimated to have lasted only a few minutes, the winds were from the north. After eye passage, winds commenced from the west. By 1200 GMT the strongest winds over the west end of Vieques were from the north-northwest. The highest measured wind speeds on Vieques were gusts of 85 knots (98 mph) recorded at the U.S. Navy's Isabel Segunda facility by the 9-m anemometer mast just prior to its failure at approximately 0900 GMT. Comparisons of damage on Vieques with damage at Roosevelt Roads Naval Station, where a verifiable wind-speed record was obtained, suggest peak gusts of approximately 115 knots (132 mph) and corresponding sustained speeds of 95 knots (109 mph). Based on information attributed to this same anemometer site, the New York Times reported on February 28, 1990, “During the more than 12 hours it took for the storm to cross Vieques, and its sister island of Culebra, wind gusting in excess of 200 miles an hour...” (Pitt, 1990.) Culebra Strong winds from the northeast began to affect Culebra at about 0900 GMT when Hugo was located approximately 75 km to the south-southeast. Data from San Juan radar indicate that intense rainbands began to move over Culebra from the northeast at 1020 GMT. The wind direction changed from northeast to east as the eye began to engulf Culebra at about 1130 GMT. At this time the circulation, as indicated by the radar reflectivity of the rainbands, became less well defined, making it difficult to establish accurately the position and diameter of the eye. In fact, one eyewitness at Ensenada Honda on the south side of Culebra claimed that no lull was observed during the passage of Hugo. This same source reported a peak gust of 140 knots (161 mph) at 1130 GMT. The anemometer in this case was mounted on the

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA mast of a boat that had been driven aground in Ensenada Honda approximately 1/2 hour earlier. The central pressure continued to rise as Hugo moved past Culebra, reaching 956 mb at 1300 GMT. Based on the observed damage, the storm track relative to the island, and the steady weakening of Hugo after it passed St. Croix, it is estimated that peak gusts of 130 knots (150 mph) and maximum sustained speeds of 105 knots (121 mph) for standard exposure conditions were experienced on Culebra. Puerto Rico The two stripchart records obtained during the passage of Hugo over Puerto Rico on September 18 were recorded at the Roosevelt Roads Naval Station and at the WSFO at San Juan International Airport. The 7-m anemometer at Roosevelt Roads is located adjacent to the main runway and has a clear exposure to wind from all directions. A time history of recorded peak gusts, unadjusted for anemometer height, is plotted in Figure 4-1. The plotted gust speeds correspond to the peak values observed in succeeding 10-min intervals and are shown connected by line segments in the figure to improve readability. Peak gusts of 104 knots (120 mph) were recorded between 1150 and 1220 GMT, and an almost complete penetration of the eyewall was observed at 1250 GMT. Maximum sustained winds at this site were approximately 85 knots (98 mph). Mechanical equipment and a part of the air operations building roof were lost at 1339 GMT, terminating wind-speed and direction recordings at Roosevelt Roads. Because the site satisfies the requirements for standard exposure, the only required adjustment to the data is for the height of the anemometer. This would have the effect of increasing the observed speeds by about 6 percent. The WSFO at Luis Munoz Marin International Airport in San Juan recorded peak gusts of 80 knots (92 mph) between 1350 and 1415 GMT, and a time history of the peak gusts is plotted in Figure 4-2. The maximum sustained wind speed was 67 knots (77 mph). Adjusting this figure to take into account the 6.1-m height of the F420C anemometer would increase the maximum sustained wind speed by about 7 percent to 72 knots (83 mph). It is known from eyewitness accounts that the eye passed over the region from Luquillo east to Cape San Juan. Because the winds there would have been directly off the ocean, speeds slightly higher than those at Roosevelt Roads Naval Station may have been reached. DESIGN WIND SPEEDS While the specified design wind speeds for the affected areas can provide some insight into the wind resistance of local structures, they are perhaps more an indicator of perceived importance of wind loading as a design consideration. In

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-1 Gust speeds measured at Roosevelt Roads NAS, Puerto Rico. FIGURE 4-2 Gust speeds measured at San Juan International Airport.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA addition to the competence of the designer, the actual wind resistance of the end product depends on workmanship, local construction practices, quality of materials, and the degree to which the code provisions are enforced. U.S. Virgin Islands Building construction in the Virgin Islands is governed by the Virgin Islands Building Code (1972). The code is mandatory throughout the territory and was developed to meet the specific circumstances of the territory. The wind-load provisions of the building code specify that the loads listed in Table 4-1 be applied to the horizontal projection of the building area over the height zones indicated. There are no provisions in the building code for local terrain effects or for classification of buildings as to occupancy or function. If typical surface-pressure coefficients of 0.8 and -0.5 are assumed for windward and leeward sides, respectively, the required design pressures can be converted to corresponding reference wind speeds (assumed to be sustained speed over a 10 minute period in open territory). For the pressures listed in Table 4-1, the corresponding sustained wind speed is approximately 70 knots (81 mph). Puerto Rico The Puerto Rico Building Regulation (1968) specifies design wind pressures that are reasonably consistent with the requirements of ANSI Standard A58.1-1982 (Minimum Design Loads for Buildings and Other Structures). The basic wind speed specified by ANSI A58.1-1982 for Puerto Rico is 83 knots (95 mph). A recent revision of the Puerto Rico Building Regulation has increased the basic wind speed to 96 knots (110 mph), applicable to all areas and to all types of structures. However, it is doubtful that this revision could have had any significant impact on the design of buildings that were in place at the time of Hugo. TABLE 4-1 Lateral Wind Loads (Virgin Islands Building Code). Height (ft) Pressure (psf) ≤ 30 25 31 to 50 35 Over 50 45 Note: 1 ft = 0.3048m; 1 psf - 47.88 Pa

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Caribbean Region In 1989, the Caribbean Uniform Building Code (CUBC) became available for adoption by countries in the region. While this code has had negligible influence on the built environment in the areas affected by Hurricane Hugo, it is useful to examine the wind loading requirements as they reflect the results of recent and extensive studies of hurricane statistics in the Caribbean (Davenport et al., 1985). The CUBC specifies reference wind-velocity pressures based on 10-minute mean speeds. For the region being considered here, the corresponding 50-year sustained speed is approximately 89 knots (102 mph). Design wind speeds specified by the local and regional Caribbean building codes are summarized in Table 4-2. For ease of comparison with the estimated and measured speeds in the affected areas, the specified design speeds have been converted to equivalent sustained speeds in knots. SUMMARY OF MAXIMUM WIND SPEEDS Based on the sources of data and analyses described in the preceding paragraphs, the maximum surface-wind speeds for the major areas affected by Hugo have been estimated. These wind speeds, referenced to standardexposure conditions, TABLE 4-2 Code-Specified Design Wind Speeds for the Caribbean Region. Source of Requirement Sustained Speed (knots) Virgin Islands Building Code—1972 Basic Wind Speed 70 ANSI A58.1-1982 (Puerto Rico) Basic Wind Speed (50-year MRI) 83 Caribbean Uniform Building Code—1989 50-year Mean Recurrence Interval 89 100-year Mean Recurrence Interval 99 Note: 1 knot = 0.515 m/s

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA and the speeds obtained from actual measurements are listed in Table 4-3 below. If the CUBC-specified speeds are taken as the most reliable statistical representation of hurricane wind speeds in this region of the Caribbean, then it is possible to assign an MRI to each of the maximum sustained speeds. These intervals are listed in the last column of Table 4-3. Based on the code-specified design speeds listed in Table 4-2 and the probable maximum wind speeds listed in Table 4-3, the following observations can be made: The basic design speed implied by the Virgin Islands Building Code was exceeded on St. Croix and on St. Thomas. On the basis of the CUBC equivalent sustained speeds, the maximum speed on St. Croix corresponds to an MRI of 300 years, and that on St. Thomas corresponds to an MRI of 40 years. As a point of interest, the basic wind speed of 70 knots (81 mph) implied by the Virgin Islands Building Code corresponds to an MRI of 15 years. For Puerto Rico, the basic wind speed of 83 knots (95 mph) as specified by ANSI A58.1-1982 was exceeded on Culebra and Vieques and at Roosevelt Roads (105, 95, and 1.06 × 85 = 90 knots). The corresponding MRIs are 170, 80, and 50 years. The adjusted sustained speed at San Juan (1.07 × 67 = 72 knots) corresponds to an MRI of 18 years. It is interesting to examine the reported peak gust of 191 knots (220 mph) for St. Croix. The corresponding MRI from the CUBC is in excess of 1,000 years. Because of the large estimation errors, the physical significance of these calculated longer intervals is questionable. It is clear, however, that gust speeds of 174 knots (200 mph) or higher in this part of the Caribbean are a very rare event. TABLE 4-3 Probable Maximum Wind Speeds and Corresponding Mean Recurrence Intervals Location Sustained Speed (knots) Gust Speed (knots) MRI (years) Virgin Islands   St. Croix 110 135 300 St. Thomas 85 105 40 Puerto Rico   Vieques 95 115 80 Culebra 105 130 170 Roosevelt Roads 85a 104a 50 San Juan International Airport 67a 80a 18 a Denotes actual measurement, unadjusted for height Note: 1 knot = 0.515 m/s

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA PROPERTY DAMAGE Observed damage in the areas affected by Hurricane Hugo is in general agreement with the surface-wind speeds listed in Table 4-3. This damage ranged from superficial to total devastation. In general, the most damaging winds were located in the northeast quadrant of the storm. This quadrant was also where the most intense rainbands, as indicated by their radar reflectivity, were located. In the following paragraphs, the damage is described in sequential order, following the general path taken by Hugo. St. Croix The strongest winds came from the northeast and caused the heaviest damage along the north coast from the Salt River eastward to the end of the island. Damage in this area was remarkably uniform and was likely made more intense by the local terrain, which slopes steeply upward to a central east-west ridge running the length of the island. This same ridge provided some shielding for structures located on its south (leeward) slope. In the southwest sector of the island, near Frederiksted, the most damaging winds appear to have come from the southwest following passage of the eye. The airport and oil refinery are located in the south-central portion of the island, where the strongest winds came from the northeast. However, because of the relatively flat and uniform terrain, the winds in this area were not as strong as those affecting the north coast. Figure 4-3 shows typical damage to a condominium located on the north coast, approximately 6 km east of Christiansted. The building system consists of concrete floor slabs and concrete masonry walls with cast-in-place corner columns and perimeter beams. Balcony decks and roof are of wood construction. The strongest winds from the northeast were approximately normal to the long axis of the building. All of the roof and most of the walls of the upper story have been removed. The building is situated on a steep slope overlooking the water. There are several new housing developments in the Christiansted area, and one of these, located on the north slope directly east of Christiansted, is shown in Figure 4-4. It is believed that local terrain characteristics caused a significant increase in wind speed, perhaps as much as 20 percent, in this area. This is some of the more intense damage observed along Hugo's path.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Figure 4-20 shows a commercial development near Esperanza on the south coast of Vieques. The large building in the center of the photograph is an old factory building that was being renovated for use as a resort hotel. The building at the top of the figure has lost all of its roof sheeting. Culebra The island of Culebra experienced severe damage from winds that were only slightly less intense than those on St. Croix. Figure 4-21 shows a housing area adjacent to the local airport that was totally destroyed. The quality of construction was poor, and it is doubtful that there was any attempt to comply with the Puerto Rico Building Regulation. In addition, the terrain slopes steeply upward to the west, which is very likely to have caused an acceleration of the intense easterly winds from Hugo. Directly to the south of this area is the Ensenada Honda, where many small watercraft sought refuge from Hugo and were either sunk or driven ashore by the strong easterly and southeasterly winds. Figure 4-22 shows several small craft beached along the north shore of Ensenada Honda. At the west end of Ensenada Honda is the town of Culebra. The aerial view in Figure 4-23 shows wind damage in the central business district and more small craft driven ashore by the strong easterly winds. FIGURE 4-20 Commercial development near Esperanza, Vieques.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-21 Housing development next to airport, Culebra. FIGURE 4-22 Small craft in Ensenada Honda, Culebra.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-23 Culebra business district. Puerto Rico On Puerto Rico, the measured sustained speeds ranged from 67 knots (77 mph) at San Juan International Airport to 85 knots (98 mph) at Roosevelt Roads. Neither location experienced a clear lull from eye passage as did Cape San Juan at the extreme northeast tip of the island, although Roosevelt Roads came close. Figure 4-24 shows two reinforced concrete buildings on the east coast of Puerto Rico between Cape San Juan and Fajardo. The building on the right was oriented so that the strongest winds from the north were blowing directly onto its broad face. Post-storm inspectors found that the windward curtain walls on many floors near the top of this building had failed, and that the interior partitions and leeward curtain walls were subsequently blown out. Figure 4-25 shows a partition wall near ground level that developed a diagonal tension crack under wind load. Figure 4-26 shows a shearwall of the building oriented so that the strongest winds came from left to right. Note the evidence of movement in the horizontal construction joint. Figure 4-27, which is a photograph of the same wall, shows damage to the perimeter beams under the floor slabs caused by transverse displacement of the building. The adjacent shearwalls showed no signs of distress. The house shown in Figure 4-28 is typical of a class of structure in Puerto Rico that performed extremely well in Hurricane Hugo. The walls are either cast-in-place

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-24 Dos Marinas, Fajardo, Puerto Rico. FIGURE 4-25 Diagonal tension crack in partition wall, Dos Marinas.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-26 Horizontal construction joint in shearwall, Dos Marinas.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-27 Damage to perimeter beam. concrete or constructed of concrete masonry units with integral reinforced concrete columns and perimeter beams. The roof slab is 100 to 125 mm thick. As with the highrise reinforced concrete buildings just described, the seismic requirements of the building code governed the design, and the substantial dead loads made this type of structure highly resistant to Hugo's wind forces. Observed damage was limited to inadequate attachment of door and window frames to the concrete walls and perimeter beams. One of these houses is shown under construction in Figure 4-29. A typical form of structural failure encountered in northeastern Puerto Rico is shown in Figure 4-30. The corrugated steel sheet is attached to the purlins by self-drilling/tapping screws. The high-strength sheet is susceptible to low-cycle fatigue at the attachment points and develops fatigue cracks after only a few minutes of wind action. This problem has been investigated and reported by Morgan and Beck (1975). Use of a large washer, shaped to conform to the ridge contour of the corrugated steel sheet, has proven to be a simple and cost-effective solution to this problem. Damage to buildings in San Juan was generally light. Some window and curtain wall failures were observed in commercial structures in the downtown area and in the highrise buildings along the beach north of the airport (see Figure 4-31 and Figure 4-32). However, most of the damage involved light commercial/industrial buildings in the Carolina district. Most of these are pre-engineered metal buildings, and experienced numerous failures of rollup doors and roof sheeting.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-28 Reinforced concrete house under construction, Puerto Rico. FIGURE 4-29 Reinforced concrete house, Fajardo, Puerto Rico.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-30 Corrugated Steel sheet blown off, Luquillo, Puerto Rico. SUMMARY For all of the regions affected by Hurricane Hugo on September 18, only two verifiable wind-speed records were obtained. Because of the lack of data, it was necessary to rely on other sources of information and employ indirect methods to obtain estimates of surface-wind speeds. Maximum speeds on St. Croix correspond to an MRI of approximately 300 years and those on St. Thomas approximately 40 years. The basic wind speed implied by the Virgin Islands Building Code corresponds to an MRI of 15 years. In Puerto Rico, estimated speeds on the islands of Culebra and Vieques, and the measured speeds at Roosevelt Roads, all exceeded the basic wind speed of 83 knots (95 mph) specified by ANSI A58.1-1982. The highest estimated sustained speed was 105 knots (121 mph) at Culebra, and this corresponds to an MRI of 170 years. The maximum sustained speed measured at San Juan International Airport and adjusted for standard wind exposure corresponds to an MRI of 18 years. Widespread loss of roof structures was observed on St. Croix, which suggests that code requirements for wind uplift should be reviewed. In Puerto Rico, both highrise and single-story reinforced concrete buildings that were designed to meet seismic requirements performed well in Hugo. The attachment of nonstructural elements such as doors, windows, and cladding needs to be improved. Loss of corrugated steel sheet roofing continues to be a widespread problem, even though past research has shown that cost-effective solutions are available.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA FIGURE 4-31 Damage to curtain wall, San Juan. FIGURE 4-32 Damage to curtain wall near street level, San Juan.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA REFERENCES ANSI. 1982. Minimum Design Loads for Buildings and Other Structures. ANSI A58.1. New York, New York: American National Standards Institute. Caribbean Uniform Building Code. 1989. Part 2—Structural Design Requirement. Georgetown, Guyana: Caribbean Community Secretariat. Davenport, A. G., P. N. Georgiou, and D. Surry. 1985. A hurricane wind risk study for the Caribbean. Pp. 17-24 in Proceedings, 5th U.S. National Conference on Wind Engineering, November 6-8, 1985, Lubbock, Texas. Gibbs, N., J. Carney, and E. Rudulph. 1989. Rebuilding paradise after Hugo. Time 134(23): 90-92. Golden, J. H. 1990. Meteorological data from Hurricane Hugo. Pp. 3-23 in Proceedings, 22nd Joint Meeting, UJNR, Panel on Wind and Seismic Effects, HIST SP 796. Gaithersburg, Maryland: National Institute of Standards and Technology. Lawrence, M. 1989. Preliminary Report—Hurricane Hugo, 10-22 September 1989. Coral Gables, Florida: National Hurricane Center, National Oceanic and Atmospheric Administration Marshall, R. D. 1976. Engineering aspects of Cyclone Tracy, Darwin, Australia, 1974. Building Science Series 86. Gaithersburg, Maryland: National Bureau of Standards. Marshall, R. D. 1984. Fastest-Mile Wind Speeds in Hurricane Alicia. NBS Technical Note 1197. Gaithersburg, Maryland: National Bureau of Standards. Marshall, R. D. 1990. Hurricane Hugo in the Caribbean—Assessment of Surface Wind Speeds. Workshop on Hurricane Hugo, Six Months Later. Mayaguez, Puerto Rico : University of Puerto Rico. Morgan, J. W., and V. R. Beck. 1975. Sheet Metal Roof Failures by Repeated Loading. Technical Report No. 2, Housing Research Branch, Australian Department of Housing and Construction Pitt, D. E. 1990. Island Still Suffers Hurricane Scars. New York Times. February 28, 1990.

OCR for page 82
HURRICANE HUGO: PUERTO RICO, THE U.S. VIRGIN ISLANDS, AND SOUTH CAROLINA Powell, M. D., and P. G. Black. 1989. The landfall of Hurricane Hugo in the Carolinas: Surface wind distribution. Weather Forecasting 6: 379-399. Puerto Rico Building Regulation. 1968. Planning Regulation No. 7—Ammended 1968. Regulations and Permits Administration, San Juan, Puerto Rico. Showker, K. 1990. After the storm. AAA World, January/February 1990. 10: 1-22. Virgin Islands Building Code. 1972. St. Thomas, U.S. Virgin Islands: Department of Planning and Natural Resources.