USF Libraries
USF Digital Collections

Evacuation decision making and public response in Hurricane Hugo in South Carolina

MISSING IMAGE

Material Information

Title:
Evacuation decision making and public response in Hurricane Hugo in South Carolina
Series Title:
Quick response research report ;
Physical Description:
14 p. : ; 28 cm.
Language:
English
Creator:
Baker, Earl J
University of Colorado, Boulder -- Natural Hazards Research and Applications Information Center
Publisher:
Natural Hazards Research and Applications Information Center
Place of Publication:
Boulder, Colo
Publication Date:

Subjects

Subjects / Keywords:
Hurricane Hugo, 1989   ( lcsh )
Evacuation of civilians -- Decision making -- South Carolina   ( lcsh )
Hurricanes -- Forecasting   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
non-fiction   ( marcgt )

Notes

Additional Physical Form:
Also issued online as part of a joint project with the Louis de la Parte Florida Mental Health Institute (FMHI) Research Library’s disaster mental health initiative.
Statement of Responsibility:
by Earl J. Baker.
General Note:
Cover of digitized version erroneously says North Carolina in title.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001985165
oclc - 39126935
usfldc doi - F57-00033
usfldc handle - f57.33
System ID:
SFS0001114:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader cam 2200277Ia 4500
controlfield tag 001 001985165
005 20090309173511.0
008 980515s1990 cou s000 0 eng d
datafield ind1 8 ind2 024
subfield code a F57-00033
035
(OCoLC)39126935
040
DDB
c DDB
043
n-us-sc
049
FHIM
086
UCB6/57.15/39
2 codocs
1 100
Baker, Earl J.
0 245
Evacuation decision making and public response in Hurricane Hugo in South Carolina /
by Earl J. Baker.
260
Boulder, Colo. :
b Natural Hazards Research and Applications Information Center,
1990.
300
14 p. ;
28 cm.
440
Quick response research report ;
v #39
500
Cover of digitized version erroneously says North Carolina in title.
530
Also issued online as part of a joint project with the Louis de la Parte Florida Mental Health Institute (FMHI) Research Librarys disaster mental health initiative.
650
Hurricane Hugo, 1989.
Evacuation of civilians
z South Carolina
x Decision making.
Hurricanes
Forecasting.
710
University of Colorado, Boulder.
Natural Hazards Research and Applications Information Center.
029
AU@
000007722055
4 856
u http://digital.lib.usf.edu/?f57.33



PAGE 1

b..VI\CIJA110NDe.C.ISIl:>NMAI<.lNGt\C!...,tJure.'RINt'JORt""c'AROLINA tr.U",I "4?-r.",iLjofColoY"&do-' 1 ISL-lute"fSeie'l"l(:elt'J-atvV'a.\HCLrc:."" A c1;l1:YlfOY"\'"TIa.lic>VI&Y1.tl!.r.Q\) I e.-I<. III S6."""Re-;:I'11!C\-\1'1::)R 7 -#::59

PAGE 2

NaturalHazardsResearchandApplicationsInformationCenterCampusBox482UniversityofColoradoBoulder,Colorado80309-0482EVACUATIONDECISIONMAKINGANDPUBLIC RESPONSEINHURRICANEHUGOINSOUTHCAROLINAByEarlJ.BakerDepartmentofGeographyFloridastateUniversityTallahassee,FL32306QUICKRESPONSERESEARCHREPORT #39......_., __ .,,......t"'\" This publication is partofthe Natural Hazards Research&Applications Information Center's ongoing Quick Response Research Report Series. http://www. colorado. edu/hazardsTheviewsexpressedinthisreportarethoseoftheauthorsandnotnecessarilythoseoftheNaturalHazards Center. ortheUniversityofColorado.InstituteofBehavioralScience#6 (303)492-6818TELEFAX:(303)492-6924

PAGE 3

EVACUATION DECISION MAKING AND PUBLIC RESPONSEINHURRICANE HUGOINSOUTH CAROLINAPreparedbyEARLJ.BAKERDEPARTMENfOFGEOGRAPHYFLORIDA STATE UNIVERSITYTALLAHASSEE,FL 323061990Theresidential samplesurveyportionoftheresearchreportedinthis documentwasfundedbytheUniversityofColorado NaturalHazardsResearchand Applications InfonnationCenterasa "Quick Response"project;bytheFederalEmergencyManagement Agency andtheU.S.AnnyCorpsofEngineersaspartoftheir post-stonn evaluation offederally funded evacuation study products; andbyHazardsManagementGroup,Inc.aspartof the finn'sresearchand developmentactivities.

PAGE 4

EVACUATION DECISION MAKINGANDPUBLIC RESPONSE IN HURRICANE HUGOSummary Although the worst of Hugo affected one of the more sparsely populated reaches of the South Carolina coast, it was necessary for residents in a much larger area to assess and respond to the threat. Forecast information providedbythe National Hurricane Center was good during most of the crucial .response period, making decision making much easier than might have been the case. Inundation maps and evacuation clearance time calculations produced in pre-storm studies proved useful and generally accurate. Computerized and graphical decision aids were utilized extensively and contributed an impression of "high-tech" performance and credibility to elected officials, but some users had dangerous misconceptions about the tools' functions and capabilities. There was very little clear-cut incorporation of forecast uncertainties into response decisions. Evacuations went well, evidencedinpartbythe low loss of life fromflooding. In many surge-prone areas, however, evacuation was not as completeasiswidely believed, and Jtad the maximum force of Hugo struck any of the major population centers of South Carolina, many homes would have been flooded with occupants still in them. ForecastsandWarnings After Hugo left the Caribbean, forecasts indicated that the storm would follow a northwesterly course. From Monday, September18through Tuesday the 19th, long-range forecasts had the storm approaching various locations along Florida's east coast.OnWednesday, September20the 6 am forecast moved the forecast track farther north, anticipating landfall between Beaufort and Charlestoninapproximately60hours. Charleston's probability of being affectedbyHugo was put at 12%. During the day Wednesday the forecast track was altered slightly, taking Hugo northwesterly, then more northerly just before landfall, whichinthe 6 pm advisory was anticipated at Beaufortinless

PAGE 5

2 than36hours: A hurricane watchwasissued for the area from St. Augustine, Florida to Cape Hatteras, North Carolina at that time, indicating that landfall was expected in36hours or less. Throughout Wednesday communities on Florida's east coast andinGeorgia and South Carolina monitored the storm, watching for a more westerly track or an increaseinforward speed. No significant response actions were implemented, althoughinsome locations suchasBeaufort officials suggested as a precautionary measure that residentsgoto friends and relatives farther inland if they would feel more comfortable doing so. Wednesday evening officialsinCharleston county recommended that residents evacuate.Itwas considered a marginal call, but officials felt the11pm news programs would be their last opportunity to reach residentsviathe mass media until the following morning. The next morning at 6 am the center of Hugowasforecast to reach Beaufortin24hours, and the National Hurricane Center issued a warning from Fernandina Beach, Florida to Cape Lookout, North Carolina. The watch remainedineffect south to St. Augustine and north to Cape Hatteras. Charleston's landfall probability was 30%, and Hugo's sustained winds were reportedas109mph, just 2 mph short of being a category 3 storm. The Governor of South Carolina ordered evacuation of barrier islands, beaches, and peninsulas, except for the city of Charleston. OfficialsinCharleston county changed their recommendation to an order. Local governments disseminated theord"er,assistedbythe National Guard. Most if not all locationsinSouth Carolina evacuated for a category 3 hurricane. Coastal Georgia also began to evacuateinresponse to the warning, as did parts of North Carolina.Atnoon on Thursday Hugo was upgraded to a category 3 storm, with winds reported and forecast to remain at115mph. The track moved slightly north, taking the center over Charleston. At 3 pm Thursday a special advisorywasissued to report that Hugo's winds had unexpectedly increased to126mph and that the forward speed had increased (from17mph to23mph). The hurricane warning area was extended to Oregon Inlet, North Carolina, and the watch was extended to Cape Henlopen, Delaware. The track was shifted a bit farther north (with landfall predicted near

PAGE 6

3. Georgetown, South Carolina), and the areal extent of the wind field was expanded. The intensification prompted evacuees in at least one Myrtle Beach shelter to be relocated farther inland. At 6 pm Hugo was reported having intensified to138mph sustained winds, making it a category4hurricane(133mph the threshold). Forward speed decreased slightly to20mph. The evacuationwasnearly complete by that time, however, andfewpreparations were altered. Decision Making Public officials face a tremendously difficult responsibilityindeciding whether and when to recommend or compel their citizens to evacuate during a hurricanethreat. Unnecessary evacuations are expensive, disruptive, and unpopular, but waiting too late to leave can be disastrous. In Hugo decisions were made much easier due to the consistency and validity of the forecasts providedbythe National Hurricane Center. Near1983Federal agencies began studies which provided the foundation of hurricane evacuation plans in South Carolina. The National Hurricane Center simulated numerous hurricane scenarios to indicate the areas that would be inundated by storm surges. The Corps of Engineers and its contractors calculated the number of people who would need to evacuate, the length of time necessary to evacuate, and the public shelter space neededindifferent storm categories. The state of South Carolina mapped the surge-prone areas of each county, and FEMA paid for much of the information to be computerized. In making response decisions, officials consider the strength of the storm which might affect their location and look up the length of time necessary to evacuate for that storm category..They then consider the length of time remaining before the stormisexpected to arrive and compare the two times to determine when evacuation must begininorder to provide sufficient time for evacuees to reach safety. All the coastal counties in Georgia, South Carolina, and North Carolina had graphical devices providedbythe Corps of Engineersand/orcomputer software made available by FEMA or a private firm to facilitate the necessary computations. Officials appeared to employ the aids with varying proficiency, but at least some used them effectively.

PAGE 7

4 Those computations, however, are not the difficult part of decision making. The difficult partisanticipating how strong the storm willbe, whether itwill"hit"a particular location, when itwillreach a location, and how large the stormwillbe. Each of those parametersisforecastbythe National Hurricllne Center, but eachissubject to error. Accounting for the forecast uncertainty and incorporating it into the decision processisthe difficult part of decision making. The National Hurricane Center since1983has included probabilities with its advisories to indicate the likelihood of a storm passing within65miles of certain locations during various time periods. These probabilities account for uncertaintyinthe direction and forward speed of forecasts. Commercial software in usebysome of the countiesinthe threatened area provided a graphical depiction of the position forecast uncertainty and calculated intensity uncertainties. Few officials exhibited evidence of systematically employing uncertainty information, particularly NHC probabilities,inresponding to Hugo. Discussions of the decision process always centered upon the forecast itself and on clearance times, with only a general concern that the forecast or the clearance time calculations could beinerror. In South Carolina coastal officials relied very heavily upon the Charleston office of the National Weather Service for advice and judgment, and that interactionwasmore influential than any other input.' The Charleston office interpreted NHC forecasts for local officials andinsome instances offered second opinions. Fax connections to local governments and conference call capabilities would have facilitated theNWSoffice's ability to interact with local officials. Computer software available to counties in South Carolina included a module originally developedbythe CharlestonNWSstaff indicating appropriate response actions, based upon theNWSstaffs judgments about acceptable risk and other factors. When Hguowasbelieved to have increased to a category 4 storm at 6 pm on Thursday the CharlestonNWSwasinfluentialinthe decision to not attempt to evacuate a larger area. The Governor's officeinSouth Carolina worked more closely with the ColumbiaNWSoffice. Decisions,inretrospect, appeared "correct" largely because the information being assumed (forecasts and clearance times) proved accurate. But had Hugo increased forward speed earlier, for example, the retrospective might have been different.Itshould also be noted that the worst of Hugo

PAGE 8

5 did not hit the populated locations of Beaufort, Charleston, or Myrtle Beach.Ifsuch had been the case, 20% to 30% of the homes in the inundated areas of those locations would have flooded with occupants stillinthem. Although evacuation notices were timely they were disseminated successfully onlyinthe most hazardous beachfront and island locations. A claim following Hugo that the relativelylowloss of lifeinsuch a severe stormwasattributable to the improved planning conducted since1983isnot supported by fact, except perhaps that the studies provide a better indication of the areas needing evacuation. Most locations didn't evacuate in Hugo until the National Hurricane Center issued a warning, which has been the norm for at least two decades. Luck--the fact that the right side of the eyewall crossed the coastinone of the least populated reaches of South Carolina's coast--wasprobably the greatest factor resultinginsofewdeaths. On average, however, the improved studies and planswillresultinlower deaths over time. The evacuation proceededassmoothlyascould be expected, and the public evaluated the warning and evacuation performancebypublic officials very favorably (Baker, 1990a). Traffic tie-ups on Interstate26leading west from Charleston prompted officials to devise a scheme to employ all lanes for westbound traffic.Bythe time the plan was completed, however, trafficwasmoving more smoothly and the ideawasnot implemented. A public school usedasa public shelterinMcClellanvilleinCharleston county flooded to a depth of approximatelysixfeet with several hundred evacuees inside during the height of the storm, but there were no fatalities. Building drawings providedbythe school board listed the elevation of the ground floor of the buildingasapproximately20feet, whereas the actual elevationwascloser to10feet.Noground survey was conductedaspart of the hurricane evacuation studies to verify the actual ground elevation of the building. The greater planning failure, however,wasthat during the study's review and local involvement process no one questioned the20foot elevationina location clearly shownasflood-prone on study-generated surge maps and on flood insurance maps. With hindsight locals in the McClellanville area felt it was"obvious"that the school sitewasnot20feet high.

PAGE 9

6Use and Evaluation of Evacuation Study Products Surge Maps The single most extensively utilized technical product generatedinevacuation studiesinthe areas threatenedbyHugo were surge inundation maps. The surge heights predictedbySLOSH when the programwasrun after Hugo's landfall, using parameters determined afterward, matched observed surge heights at the coast very closely. SLOSH maps generated earlier and used for evacuation planning were composites of different storm scenarios, and officials seemed satisfied with their validity in Hugo, althoughwehave seen no specific verification data. Usersinseveral locations felt the surge zones should be mapped at a larger scale to show greater detail. Clearance Times Another heavily employed planning product was clearance time calculations, which were generatedbya consultant for the Charleston district of the Corps. These,inconjunction with storm forecasts, provide the basis for timing of response actions such as evacuation. They are based upon assumptions about road and street networks and public response patterns. In the Charleston area, as Hugo threatened, theNWSoffice suggested to the county emergency preparedness director that the clearance times calculated for the evacuation study were unrealistically pessimistic. After polling municipal officialsinthe county, the director was unable to find a consensus regarding the any sort of perceived biasinthe times, so the study-generated times were employed. Clearance times observedinHugo appeared very close to the calculated times, although the analysiswascursory. Decision Aids The response decision making process was discussed earlier, anditwasnoted thatNRCforecasts were usedinconjunction with planning study clearance times. Graphical tools and computer software were employedinsome locations to facilitate the computations. None of the tools appeared to have been used very effectivelyinmost locations to assess forecast uncertainties, however, and some

PAGE 10

7users had gross misconceptions about the tasks performed by the aids. At least two local preparedness officials credited the aids with making accurate predictions of when the storm would arrive. None of the tools made predictions; they simply facilitated computations based upon assumptions which were input. The tools appeared to be accurate simply because in Hugo the NHC forecasts and study generated clearance times were accurate. Users were generally pleased with the computer software available to them, but in many locations itwasmore a device for impressing elected officials and the media than for sophisticated decision making. Public Response Approximately three months following Hugo's landfallinSouth Carolina, telephone interviews were conducted with residentsinMyrtle Beach, the Charleston area, and Beaufort. Parts of each sample were composed of households which had been interviewedinprevious studiesinthe area regarding hypothetical hurricane threats. In Myrtle Beach some respondents had also been previously interviewed to document their responseinDiana. Not all earlier respondents could be contacted, however, andineach area the samplewassupplemented with newly selected households. In Myrtle Beach and Beaufort a total of150post-Hugo interviews were completed, and in Charleston, on the peninsula and west of the Ashley,200households were interviewed. Includedinthe Beaufort sample were27respondents on St. Helena island who had been interviewedinan earlier study. In addition, a combined100interviews were completedinMt. Pleasant, Sullivan's Island, and Isle of Palms, where no hypothetical response data had been gathered previously. The supplemental hQuseholds weren't selected using the same criteriaasthose originally employed, but they are generally comparable. The great majority of respondentsinall areas wereincategory1,2,or 3 surge zones.Nosupplemental respondents were outside category 4 zones. The old and new Beaufort samples are differentinpart because the new sample included no St. Helena respondents. In both the Charleston and Myrtle Beach supplemental samples, a greater portion of respondents lived within a mile of water (beach, harbor, river, etc.) thaninthe original samples.

PAGE 11

8EvacuationRatesInnone of the primary sample locations was evacuation complete. In Beaufort 72% left,inCharleston 62%, Mt. Pleasant/Sullivan's 81%, and Myrtle Beach, 79%. There were, however, variations within these areas. From Sullivan's Island and Isle of Palms 96% left, and there were probably comparable successes in other high-risk barrier islands.Itwouldbeinaccurate to say that everyoneinthe sample should have evacuated orthatofficials indicated that they should. A small minority in fact lived outside areas advised or ordered to leave. Nevertheless, Hugo generated a category 4 type storm surge, and evacuees and officials had no assurance until shortly before landfall thatitwould not cause such effects in their locations. Taking just the category1,2,and 3 surge areas,itisunlikely than more than 75% to 80% evacuated from most areas other than barrier islands and beach fronts. Most but not all respondents believed they had been told by officials to leave(71% in Beaufort, 64%inCharleston, 72%inMt. Pleasant/Sullivan's, and73%inMyrtle Beach). Veryfewrespondents interpreted the evacuation noticesasbeing mandatory. Of the total sample only 30% in Beaufort and Myrtle Beach and 15% to 20% in Charleston said they heard an order to evacuate.Ifmore had heard officials say they should leave and if more had interpreted the notices as orders, more would have left. Overall 89% of those who said they heard an order evacuated, compared to 70% who said they only heard a recommendation. Surprisingly, 61% saying they heard neither evacuated. Thisisrelatively high for people not hearing official evacuation notices. Many did, however, hear from other sources that they should leave and were aware that neighbors were leaving. Those saying they lived within a block of most types of water bodies were most likely to evacuate (84%).Ofthe respondents saying they lived more than a block but less than a mile from water, 73% left, compared to 65% who said they lived more than a mile from water. The exception to this trend were respondents living within a block of rivers, of whom only 67% left. Respondents were asked whether they thought their homes would have flooded if Hugo had struck their location directly.Ifthey believed their homes would have flooded, 83% left, compared to 65%ofthose who felt their homes would not have flooded.

PAGE 12

9Evacuation Timing When a watch was issued Wednesday at 6 PM fewer than 10% of the eventual evacuees from most areas had left. The percentage was slightly higher in Beaufort (17%) where officials indicated they had suggested to residents earlier in the day that a visit to friends or relatives in safer locations mightbeprudent.Bymidnight, following the earlier voluntary evacuation notice in the Charleston area and statements by the Governor, additional evacuees had left, between 10% and 20% of the eventual totals. When the warning was issued at 6 AM Thursday morning, and the governor ordered evacuation from the most vulnerable coastal areas, 50% of the evacuees from Sullivan's Islandand the Isle of Palms said they had already left. In Charleston 30% of the evacuees said they had gone when the warning was issued, followed by 25% from Mt. Pleasant, 22% from Beaufort, and 10% from Myrtle Beach, farther to the north and away from the storm. Throughout Thursday morning most evacuees departed, andbynoon between 75% and 90% had left from all the survey areas except Myrtle Beach, from which only 35% had gone.By4PMalmost everyone who left had already done so except in Myrtle Beach where departures continued until 7 PM. Almost two-thirds of the Myrtle Beach evacuees said they left between noon and 7 PM.Type of RefugeVeryfewevacuees went to public shelters (9% in Beaufort,7%in Charleston,2%in Mt. Pleasant/Sullivan's, and 13% in Myrtle Beach). Across the four sites more people went to motels than shelters, ranging from 15% in Myrtle Beach to 26% in Mt. Pleasant/Sullivan's. More than half the evacuees from all areas (56% to 66%) went to the homesoffriends or relatives. Shelter useisusually associated with income, and such was the case in Hugo.Inhouseholds reporting annual incomes below $10,000, 25% used public shelters. In no other income group did more than 8% go to shelters.

PAGE 13

10Non-whites--primarily blacks--were much more likely to use public shelters than whites(31%vs.5%). There was a difference even within most income groups 39%vs.9% for incomes less than $10,000/year, 27%vs.3%for incomes between$10,000and $25,000/year, and 22%vs.3%for incomes from $25,000 to$4O,OOO/year.Respondents livinginmobile homes were slightly more likely to use public shelters than other residents (14%vs.8%). Evacuees living within a mile of water bodies other than rivers were less likely than other groups to use public shelters. Of the evacuees stayingintheir own county 25% went to public shelters, compared to only 2% of those going out-of-county. Breakdownsbycounty for in countyevacuees going to public shelters are unreliable due to the small sample sizes involved. Evacuation Destinations In all primary sample locations between 64% and 78% of the evacuees went to out-of-county destinations. Roughly a fifth of all evacuees reached their destinationsinless than30minutes, indicating very short trips. Between 28% (Mt. Pleasant/Sullivan's) and 49% (Myrtle Beach) took an hour or less. Beaufort (16%), Charleston (20%), and Mt. Pleasant/Sullivan's (29%) all had substantially more evacuees requiring overfivehours to reach their destinations than Myrtle Beach (3%). NumberofVehicles The number of evacuating vehicles per household ranged from1.1inCharleston to 1.4 in Beaufort and Mt. Pleasant/Sullivan'S. This represented 59% of all available vehiclesinCharleston to71%inBeaufort.

PAGE 14

11Comparison of Public Response to Behavioral Assumptions Derived Before Hugo Evacuation Rates In hypothetical response surveys conducted before Hugo veryfewpeople said they would refuse to evacuate even if ordered (1% to 4%). The behavioral analysis indicated thatwithout disseminating evacuation orders door-to-door35% would not leave from Beaufort, 35% from Charleston west of the Ashley River, 20% from the Charleston peninsula, 20% from moderate-risk areasinMyrtle Beach, and less than 10% from high-risk islands and beachfronts. Evacuation rates in Hugo were extremely close to those indicatedinthe behavioral analysis. Exact comparisons aren't possible without further disaggregating the Charleston sample east and west of the Ashley and without more precise determination of respondents' evacuation zones.Itisclear, however, thatinareas other than high-risk barrier islands and beaches, the evacuation was not extensive enough to be called a complete success, primarily because too many people did not believe they were being orderedbyofficials to evacuate. The limited usefulness of the hypothetical response dataisalso apparent. Evacuation Timing In responses to hypothetical hurricane scenarios, 40% to 50% of those interviewed before Hugo said they would evacuate when a watch was posted,beforeofficials indicated they should leave. The behavioral analyses indicated that a variety of response curves were plausible, depending upon various warning scenarios, but suggested that no more than 10% to 15% of the evacuees were likely to leave before evacuation notices were issuedbyofficials. In Hugo officialsindifferent locations said various things at certain times, but overall the behavioral analysis figures were very close to the mark.Ifanything there was slightly more early responseinHugo than behavioral analysis guidelines suggested.

PAGE 15

12TypeofRefuge. In hypothetical response surveys 37%inBeaufort, 40% in Charleston west of the Ashley, 49% on the Charleston peninsula, and 35%inMyrtle Beach said they would go to public shelters if they evacuated. The behavioral analyses cautioned that hypothetical shelter useisnormally twice actual. Shelter use assumptions in the behavioral analysis labelled "cautious" (i.e., attempting not to underestimate demandinnormal circumstances) were 15% for Beaufort, 15% for Charleston west of the Ashley, 30% for the Charleston peninsula, 20% for moderate-risk areas of Myrtle Beach, and 5% to 10% for high-risk barrier island and beaches. (The exception to the last case was St. Helena island, a socially dose-knit, predominantly black community, where shelter use was projected at 40%, compared to 62% who said they would use shelters.) Shelter useinHugo was generally lower than the numbers citedinthe behavioral analyses, particularlyinmoderate-risk to low-risk predominantly white areas. The behavioral analyses did, however, point out thatinearly evacuations for severe storms more evacuees would leave the local area, causing shelter use to be lower, and that if officials took actions to discourage shelter use, it would be lower. Both conditions appeared to pertaininmost locations during Hugo, especiallyinCharleston and Beaufort where shelter use was lowest and deviated most from the norms citedinthe behavioral analysis. In Myrtle Beach 18% of the original hypothetical response interviewees used public shelters, almost exactly the figure indicatedbythe behavioral analysis. Behavioral analyses should provide a numerical adjustment for special circumstances affecting shelter use rather than simply a directional adjustment, and more situational guidelines rather than place specific estimates should be providedinthe analyses. Those practices areinfact the norminmost contemporary behavioral analyses, the process having evolved since the South Carolina studies were completed. Behavioral analyses should also project demand for in-county and out-of-county public shelter separately. The bulk of public shelter demand was assumedinthe behavioral analyses to be in county, but a numerical distintionwasnot madeinthe report. (In that sense, the behavioral analysis model actually projected demand more accurately than the overall comparison cited above indicated.) The issue of non-white demand for shelters being greater than white demand across income groupsisa factor that needs further consideration.

PAGE 16

13South Carolina officials have estimated that256,000people evacuatedinthe state during the Hugo threat, and Red Cross records indicated that94,000were registeredinpublic shelters, almost halfininland shelters. Those figures seem to imply that 37% of South Carolina's evacuees wenttopublic shelters, whichisalmost certainly not the case. The figures also appear to conflict with the survey data indication that only 2% of the evacuees who went out-of-county went to public shelters.Itispossible that there were substantially more than 256,000 evacuees, including many from low-risk areas not includedinthe sample survey, or that there were fewer than 94,000 peopleinpublic shelters seeking refuge from the storm.Itisalso possible that thoseinshelters include people seeking refugeafterthe storm. Destinations In hypothetical response surveys 45% from Beaufort, 50% from Charleston, and 50% from Myrtle Beach said they wouldgoto out-of-town destinations when evacuating. The behavioral analyses indicated that 40% from Beaufort, 45% from Charleston west of the Ashley, 35% from the Charleston peninsula, and 60% to 70%inMyrtle Beach (the latterfora severe storm with a timely evacuation) would leave the local area. The behavioral assumptions were very closeinMyrtle Beach, butlowfor the other areas. Here too the analyses indicated that early evacuations would see more people going inland, but no numerical guidelines were given except for Myrtle Beach. The effect of actionsbypublic officials, which was largely responsibleforthe large out-of-town evacuationinHugo,wasnot addressed explicitlyinthe behavioral analyses as itwasinthe discussion of shelter demands. Again, the norm in contemporary behavioral analysesisto provide explicit numerical guidelines for such scenarios. Vehicle Use Hypothetical response data indicated that about 65% of all available vehicles would be usedinevacuating households, and the behavioral analyses recommended using that figure for Charleston and Beaufort and using 70% to 75% for Myrtle Beach. Actual usewaswithinfivepercentage points.

PAGE 17

14Response Outside South Carolina Officials' perceptions of public response, especially quantitative estimates thereof, are notasreliable as scientifically designed and administered sample surveys. No such surveys were conductedinNorth Carolina or Georgia, therefore no firm conclusions are possible Jor public responsein Jhose states. The evacuationinmost of North Carolina appears to have been only partially implemented, making comparison between actual response and that projectedinbehavioral analyses for that area meaningless. A full-scale evacuationwasimplementedincoastal Georgia, and officials estimated that there were175,000evacuees,6,000of went. to public shelters. 1f both figures are even close to actual response, the public shelter demand figures projectedinthe Georgia behavioral analysis were grossly exaggerated (15% for barrier islands andupto 65% for other areas). The Georgia behavioral analysis appears to have relied far too heavily on hypothetical response data. Although official estimates of other responsesinGeorgia aren't available, the Georgia behavioral analysis projections for early evacuation (during a watch, prior to recommendations or orders) also appear flawed. Conclusions Public response to the Hugo threatwasextremely good and demonstrated once again the impact public officials can have on evacuation behavior. Over 90% of the respondents felt that officials had handled the evacuationwell.More people probably shouldhaveevacuated than did, and mo"re would have but did not believe that officials had ordered them to do so. Relativelyfewevacuees left prior to explicit recommendations or orders from public officials, butinsome locations a substantial portion of the eventual evacuees had already left when the warningwasposted Thursday morning. Public shelter use and local refuge demand were relieved considerablybyofficials' urging evacuees to seek other alternatives. Behavioral analyses upon which evacuation studies were basedforSouth Carolina were quite accurate for most locations and most behaviors but would have been more useful had they provided numerical guidelines for planning for a greater variety of scenarios.