The weather system that eventually became Hurricane Elena was initially identified as a well-organized cloud pattern north of the Cape Verde Islands on August 23, 1985. The system moved westward across the tropical Atlantic Ocean at an unusually fast speed (34 mph) and continued through the Caribbean to Cuba. It intensified and was named Elena on August 28 while situated on the northern coast of Cuba.
As the storm moved north-northwest across the Gulf of Mexico, Elena strengthened rapidly and was classified a hurricane on August 29. On August 30, the storm began to decrease its forward speed dramatically. Late on August 30, while located about 200 miles southeast of the mouth of the Mississippi River, Elena made a sudden turn to the east. On August 31 through September 1, this storm moved toward the Florida coast and stalled about 50 miles offshore from Cedar Key. During the afternoon of September 1, Hurricane Elena once again turned westward, increased its forward speed to 10–15 mph, and finally moved ashore at Biloxi, Mississippi, about sunrise on Labor Day, September 2. Figure 1-1 shows the track of Elena, which struck the Mississippi coastal area as a category 3 storm. Figure 1-2 shows that such an event is not uncommon in this area.
WARNINGS AND EVACUATION
Hurricane Elena was a rare Gulf of Mexico storm in that it caused hurricane warnings to be issued along an extensive section of the Gulf Coast from Grand Island, Louisiana, to Sarasota, Florida (population in excess of 5 million). Much of the area was placed under warnings twice. Hurricane
warnings were issued initially at 8 a.m. CDT on August 29 and discontinued at 1 p.m CAD on September 2 after being in effect for 101 consecutive hours. During this time, about 1.5 million people were evacuated to inland areas or local shelters. From southeast Louisiana through northwest Florida, people were within 4 days.
INJURIES AND DEATHS
Injuries and deaths were very low, given the size and scope of the storm. There were 4 storm-related deaths (all in Florida and all due to falling trees) and 98 injuries hospitalization. Another 36 person were hospitalized with stress-related physical problems.
The insured losses from Elena were reported at $543 million by the Insurance Services Group. The total economic loss from the storm has been estimated to be in excess of $1 billion. The vast majority of the significant damage was caused by hurricane-force winds. Although storm-surge damage was evident from Tampa Bay to New Orleans, its effects were minimal compared with wind damage. From Tampa Bay to Gulf Shores, Alabama, wind damage ranged from light to moderate, with windows, power lines, and trees mostly affected. Severe wind damage occurred from east of Gulf Shores to Biloxi, Mississippi (from the coastline to 20 miles inland). Moderate to occasionally severe wind damage was noted from Gulfport to Bay St. Louis, Mississippi. Damage was light from Bay St. Louis to New Orleans. Figure 1-3 shows a wind damage map prepared by a survey team from the University of Chicago. It should be noted that the wind-speed contours on this map are based primarily on damage observations, a procedure subject to considerable error. Based on subsequent analyses of measured wind speeds, these contours appear to overestimate the maximum gust wind speeds by about 10 percent. The detailed analysis of wind-speed data given in Chapter 2 indicates that the wind speeds probably did not exceed the level that has a mean recurrence interval of 50 years, commonly used to design buildings and other structures.
SUMMARY OF FINDINGS AND RECOMMENDATIONS
This report concentrates on three aspects of Hurricane Elena: (a) forecasting and collection of meteorological data, (b) evacuation procedures, and (c) the performance of buildings and other structures in the storm.
Need for In-Depth Study Following Postdisaster Investigation
Early in the investigation, it became apparent that unless the wind conditions could be determined with reasonable accuracy and the wind resistance of common structural systems assessed, this report would simply become a catalog of damage rather than a useful engineering analysis. Since extensive wind damage occurred to structures ostensibly designed to resist the conditions encountered in Elena, a more in-depth analysis was considered essential.
Such an analysis—which includes structural and wind-tunnel tests and a detailed study of building codes—is outside the realm of postdisaster investigations supported by the National Research Council. This ability to fund in-depth studies immediately after hurricanes or tornadoes is essential for advancing our capabilities to reduce the impacts of future disasters. Therefore, the study team recommends that adequately funded investigation teams, similar to those that investigate major air crashes, be established through funding from public agencies such as the National Science Foundation (NSF), the National Oceanic and Atmospheric Administration (NOAA), or the Federal Emergency Management Agency (FEMA).
Forecasting, Warning, and Evacuation
On the positive side, considering the scale of the evacuation, the process appeared to have gone well, benefiting from earlier National Weather Service (NWS), FEMA, and local government activities in preparing evacuation plans.
The NWS appeared to communicate the threat well to the general public through the issuance of hurricane watches and warnings. Local emergency management officials also relied heavily on NWS warnings in making decisions about evacuations. It would appear that such officials could have been better informed by using NWS marine advisories and hurricane probabilities more effectively, which could have given a clearer indication of the threat.
The numerical models used by NWS for forecasting had difficulty in predicting, to an acceptable accuracy, the erratic track of Elena, although NWS believed that these models ought to have been able to do so. The NWS has since reevaluated its forecast models in light of their unsatisfactory performance during Elena.
Need for Surface Wind-Speed Measurement
During and after the storm, the determination of surface wind speeds proved a crucial, albeit controversial, undertaking. Wind speeds are important not only in forecasting storm intensity and in planning emergency response, but particularly in evaluating structural performance in areas struck by the storm. Unfortunately, the current business of gathering surface-level wind speeds is fraught with difficulties and inaccuracies. The NWS itself acknowledges that the chances of accurately measuring maximum surface wind speeds in a hurricane moving onshore are small, using existing measurement systems.
The NWS bases its forecasts on wind-speed measurements made by reconnaissance aircraft flying at altitudes of several thousand feet. Surface-level observations of sea state, limited anemometer observations, and correction factors based on well-known pressure-wind relationships are all applied in an attempt to refine these measurements and adapt them to surface-level estimates. While the resulting data may be useful for meteorological purposes and for emergency response planning, they fall far short of the accurate and detailed record of surface wind speeds that the engineering community requires.
In addition, NWS makes use of damage maps that seek to estimate wind speeds based on how much damage is observed after the event—a practice that engineers warn is approximate at best and tends to overestimate wind speeds. As a practical matter, the surface wind conditions that impinged on area structures during Elena had to be established independently after the storm by engineers using mostly non-NWS data.
For several years, users of wind-speed data, such as engineers and emergency management officials, have been concerned about this lack of accurate ground-level wind-speed data from hurricanes. Recognizing the budgetary limitations of NWS and the fact that its primary mission is to issue warnings and forecasts to help safeguard lives and property in the short term, the study team nonetheless recommends that NWS enhance its record-keeping ability so that it can document actual surface wind speeds for postevent analysis. Clearly, this is consistent with NWS's overall mandate to help the populace manage extreme weather events, since such wind data are instrumental in establishing appropriate wind engineering standards for structures located in areas at risk from hurricanes.
One way to produce the much-needed wind data would be to establish a network of portable anemometer stations along hurricane-prone coasts. In any case, the study team recommends that the use of damage observations to assess wind speeds be restricted to meteorological applications where the inaccuracies inherent in this method are tolerable. Such studies should in no case be put forward as sufficient for engineering purposes unless some scientific justification can be provided.
Structural Performance and Building Codes
While the evacuation effort was certainly successful, the engineering performance of certain classes of buildings was poor indeed. The most important group of buildings to perform poorly was schools. Nearly every school in the affected area suffered roof leakage, and many had serious damage. Two schools being used as shelters suffered considerable damage; in one of these cases it was deemed necessary to evacuate the building at the height of the storm.
It is unlikely that the design wind speed actually used by the local building code for at least 15 years prior to Elena, and implied by earlier codes, was exceeded in Elena. Even buildings possessing a very low factor of safety should not have been damaged. Yet damage was extensive in masonry-walled buildings with light roofs (particularly schools and stores), older metal buildings, and wood-framed residential structures in exposed areas.
Some of these problems may have been associated with poor enforcement of the Standard Building Code used in the area, but most were associated with the contents of the code itself. Although using an appropriate design wind speed, the code, until recently, seriously underestimated the pressures generated by the design wind conditions. In addition, allowable forms of construction in wood and masonry have insufficient wind resistance even to meet the rather deficient wind-loading requirements of earlier codes.
Since the code is designed to set performance standards used by professionals to establish appropriate structural forms, it is unenforceable when
no design calculations are made, that is, in the case of most residential and small commercial structures. A design professional—architect or engineer—is usually employed when design calculations are made. In this situation a building official usually assumes that the professional has checked the wind resistance of the structure and makes no further checks.
The general public, insurers, and emergency management officials generally assume that a certificate of occupancy issued by a building official indicates compliance with the building code and an ability to resist the design conditions specified in that code. Yet it is unlikely that the wind resistance of any building in the area affected by Elena or in any similar area using the Standard Building Code would have been checked by a building official. At best, such an official may have insisted on certain structural details such as the use of hurricane anchors, which experience had suggested should be used.
It is recommended that:
deemed-to-comply provisions be introduced into the code for nonengineered structures, as has been done for many years in the South Florida Building Code;
some form of technical review be conducted for engineered structures;
wood and masonry provisions inconsistent with the wind-loading provisions be removed from the code; and
to ensure uniformity and the unbiased use of modern wind-loading principles, the provisions of American National Standards Institute (ANSI) A58.1 be used for the determination of wind and other design loads.
The building control system based on local enforcement and the use of a model code, the provisions of which are determined in a highly political manner, may have served the communities affected by Elena well in other respects, but failed them with regard to wind resistance. A federally imposed system coupled with insurance availability, such as the highly successful Federal Flood Insurance Program, might have served them better. If the present system of building control cannot be reformed, alternative schemes backed by the federal government or private insurers should be considered.