ML20235R684
| ML20235R684 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 10/01/1987 |
| From: | PUBLIC SERVICE CO. OF NEW HAMPSHIRE |
| To: | |
| Shared Package | |
| ML20235R605 | List: |
| References | |
| OL, NUDOCS 8710080155 | |
| Download: ML20235R684 (202) | |
Text
- - _ _ _ - - - _ _ _ _
i d
'l Date: October 1, 1987 UNITED STATES OF AMERIC7.
NUCLEAR REGULATORY COMMISSION before the ATOMIC SAFETY AND LICENSING BOARD
)
In the Matter of )
)
PUBLIC SERVICE COMPANY OF ) Docket Nos. 50-443-OL NEW HAMPSHIRE, et al. ) 50-444-OL
) Off-site Emergency (Seabrook Station, Units 1 ) Planning Issues and 2) )
)
l
! APPLICANTS' DIRECT TESTIMONY NO. 7 (EVACUATION TIME ESTIMATE AND HUMAN BEHAVIOR IN EMERGENCIES)
{
Panel Members: Anthony M. Callendrello, Manager, Emergency Planning, New Hampshire Yankee Gordon Derman, President, Avis Airmap Company Paul R. Freenette, Jr., Senior Emergency Planner, New Hampshire Yankee i Edward B. Lieberman, KLD Associates !
Dennis S. Mileti, Professor of Sociology and Director of the Hazards Assessment Laboratory, Colorado State University l
I. CONTENTIONS ADDRESSED l This testimony addresses contentions regarding the i
Evacuation Time Estimate (ETE) of the New Hampshire 8710080155 871001 PDR ADDCK 05000443.
y PDR j
.)
Radiological-Emergency Response Plan-(NHRERP) Revision'2,and' '
Human Behavior in Emergencies, including Seacoast Anti-Pollution League (SAPL)' Contention No.34, Town of Hampton (TOH) Contention No. III, SAPL Contention No.'31, and
-elements of TOH Contention No. IV, TOH Contention No.fVI, SAPL Contention No. 8A, 18 and 37.
SAPL Contention No. 34 reads as follows:
l "The New. Hampshire State and local pla L do not meet the requirement that there be maps showing the population distribution around the facility as required at NUREG-0654 J.10.b and. Appendix 4.
Therefore, there is no reasonable assurance that-adequate protective measures can and will be taken pursuant to 10 ' CFR 50.47 (a) (1) and 50.47 (b) (10) . "
As its basis for this contention, SAPL asserted that NHRERP Revision 2 erroneously reduced the peak populations in the.EPZ and that the population maps are not accurate.
The revised TOH Contention III, filed on 10/31/86, asserted:
"The Evacuation Time Estimate Study (ETE) prepared by KLD Associates, Inc., Revision 2, Volume 6, is based upon inaccurate biased factual data and unreasonable or misleading assumptions, fails to corply with NRC regulations, and fails to provide
.- reasonable assurance that adequate protective measures can and will be taken, or that adequate facilities, equipment, or personnel will be provided to the Town of Hampton, in the event of radiological emergency. 10 CFR 50.47 (b) (1) (10) ; NUREG-0654, Appendix 4. "1(a) (1) ,
1TOH revised Contention III to Revision 2 was admitted on February 18, 1987, limited only to the bases expressed on 10/31/86 and 11/19/86. The Memorandum and Order of May 18, 1987, however, admitted TOH III limited to the bases offered on 10/31/86. Asubsequent ruling on July 16, 1987 admitted the bases to TOH III offered on its pleading dated May 23, 1986 as limited to matters preserved in NHRERP Revision 2.
i As bases for this contention, TOH asserted:
(a) ETE lacks adequate data to compute permanent and transient population estimates.
(b) ETE. inadequately compensates for adverse irclement weathe'r conditions.
.(c) ETE employs unsupported assumptions when computing road network capacity'cencerning' impediments to traffic, insufficient traffic control, through traffic, highway accessible capabilities, spontaneous evacuations, local weekend work schedules, disabled vehicles.
(d) ETE employs unsupported assumptions for ETE preparation time (notification times, beach to car times, staging ares preparation time, understatement of impact of workers returning home).
(e) Growth of EPZ population and vehicles, population distribution will soon be outdated.
(f) ETE assumes evacuees will go to designated host communities.
The revised TOH Contention III, filed on May 23, 1986, and admitted on May 18, 1987, reads essentially as the revised TOH Contention III filed on October 31, 1986. The bases for this version of this contention were limited by the admitting order to matters preserved in NHRERP Revision 2.
As applicable bases for this contention, TOH asserted:
l
l
- 1. ETE unreasonably estimates vehiclo counts using limited data, and therefore, ETE projections are unreasonably low.
- 2. ETE unreasonably relies upon an inadequate factual base, 1.g., telephone survey to estimate time required for notification, trip commencement, a'd n total population I evacuation.
- 3. ETE fails to account for traffic traveling through beach areas during the summer.
- 4. ETE fails to provide reasonable assurance adequate personnel are available to implement ETE; local area traffic problems and bottlenecks may not be apparent to State
', personnel.
- 5. ETE unreasonably assumes traffic guides will be available to implement traffic control procedures.
- 6. ETE unreasonably assumes adequate equipment and 1
personnel will be available to assure evacuation routes remain passable during a snow storm.
- 7. ETE unreasonably assumes buses will encounter "little impedance" when entering the EPZ; fails to account for impact of entering emergency vehicles in delaying overall evacuation of vehicles; assumes buses may travel at 40 to 50 mph; bus loading times of 40 minutes are unrealistic.
- 8. ETE relies on inadequate data to compute population estimates; no determination if telephone survey, used to compute residents and transients without transportation, is l
-4 -
representative; ETE concedes it has no computation regarding special facilities or private citizens with medical needs; ETE fails to include through traffic.
NOTE: The basis that a " compromise estimate" of 2.8 persons per vehicle was not preserved in Revision 2, and is not therefore, applicable.
SAPL Contention No. 31 as revised and dated November 26, 1986 reads as follows:
"The evacuation time estimate report, as described in Volume 6 of NHRERP Rev. 2 does not meet the requirements of 10_ CFR 50.47 (a) (1) , 50. 47 (b) (10) and NUREG-0654 II.J.2, II.J.10 i, 10 h and 10 1, and Appendix 4 because it fails to account properly for the number of vehicles that would be evacuating the EPZ; relies in part upon unsupported assumptions; relies in part upon potentially biased input data; does not rely upon extensive enough empirical base; relies upon traffic control personnel not shown to be available; does not appropriately account for travel impediments such as flooding, snow, fog and icing of roadways; does-not account for the effect of driver disobedience on evacuation time estimates (ETEs); does not appropriately deal with topographical features; does not deal realistically with the transport of transit dependent persons; in some instances overestimates roadway capacity and, for all of these reasons, underestimates the amounts of time it would take to evacuate the EPZ and its subparts
(" Regions") under the various scenarios analyzed."
As bases for its contention, SAPL asserted:
- 1. KLD estimates of 3,000 "through" vehicles on highway inadequate, uncicar if 3,000 vehicles factored into time estimate calculations; when and by whom aerial photos of Hampton Beach showing 300 vehicles on roadway were taken is unstated.
- 2. Use of figures from Kaltman, 1981 report unreliable due to area growth (vehicle / dwelling for seasonal units, l-vehicles at campgrounds, vehicles for overnight accommodations).
- 3. Resident and employee population growth not provided for over the term of the plant's life span.
- 4. Unrealistic to assume traffic control measures will i be in effect during evacuation; scate resources would not be mobilized quickly enough; assumptions used by KLD that ETE would be extended only 20-30 minutes with immediate General Emergency "are not carefully elucidated."
- 5. KLD estimate fails to account empirically for 25%
spontaneous evacuation rate, or account for traffic outside of EPZ spontaneously evacuating.
- 6. Due to low telephone survey call completion rate, no valid basis for notification times, commencement of evacuation trips, average vehicle occupancy.
- 7. Assumes workers can return home in normal time frames.
- 8. Total of 146 traffic guides are needed for TCPs and ACPs; NHRERP does not support adequate numbers.
- 9. KLD estimate assumes no effect of stalled vehicles, recommendation for placement of tow trucks not indicated in NHRERP, only two of the tow truck locations are within EPZ, no indication given for how long a tow truck would take to respond.
- 11. No account for topographical features.
- 12. Mobilization times for buses are not reasonable, j l
particularly during off-business hours. Survey and replies not in appendices.
- 13. Number of transit-dependents underestimated.
- 14. Loading times for boarding special facilities-too short; loading nonambulatory persons omitted, no estimate of )
noninstitutionalized, nonambulatory persons.
- 15. Collected data re: road geometrics should confirm roadway widths equal to or greater than assumed.
- 16. Calculation of transit dependent based directly on percent of time average vehicle is inoperable incorrect. j
- 17. Unclear how traffic control information is handled in simulation; loading procedures not described in sufficient detail; substantial amount of passing assumed; unclear how )
light traffic patterns are treated.
- 18. Estimate of 2.6 persons / vehicle for residents unrealistic, not supported by August, 1985, and July 4, 1986 actual vehicle counts.
- 19. Insufficient empirical base of transient population inferring from beach blankets and parking spaces.
The original SAPL Contention No. 31 was filed on May 15, 1986 as " SEACOAST ANTI-POLLUTION LEAGUE'S FOURTH SUPPLEMENTAL PETITION FOR LEAVE TO INTERVENE." The Board's Memorandum and Order of February 18, 1987 did not consider further that May 15, 1986 filing with the admission of November 26, 1986 revision, detailed immediately above. The November 26, 1986
l revision was admitted, excepting bases 10 and 16. The Board's Memorandum and Order of July 16, 1987, however, reconsidered and admitted that portion of the first paragraph of page 11 of the May 15, 1986 filing which alleged
" overestimated capacities of certain roads and inter-l sections." Specifically, this basis cites Route 1A N/S as being overestimated as a " Medium" design road (Volume 6, p.
3-7); SAPL contends that this road should be treated as a low design road, as it is very narrow in places and "has at at least two points almost right angle turns."
TOH Contention No. IV as revised reads as follows:
" Revision 2 fails to provide for adequate emergency equipment, fails to demonstrate that adequate protective responses can be implemented in the event of radiological emergency, and fails to correct deficiencies in emergency response capabilities apparent from the emergency exercise."
10 CFR 50.47 (1) (8) (10) (14) .
TOH IV, Basis E., dated April 14, 1987 asserts that the Hampton RERP fails to assure prompt access for emergency vehicles to the EPZ. Included under Basis 2 of its contention, TOH asserted:
"(B)us-and-driver availability would be substantially less than as specified in the Letter Agreements (due to) drivers . . . who may become imbedded in outgoing evacuation traffic thereby substantially delaying or prohibiting a driver from timely reaching the EPZ. RAC Review, August, 1986,Section VI, p. 12."
TOH Contention No. VI reads as follows:
"The Hampton RERP fails to demonstrate that local personnel are available to respond and to augment their initial response on a continuous basis in the
event of radiological emergency. 10 CFR 50.47 (b) (1) . "
Basis A as submitted on February 21, 1986 TOH filing holds in applicable part:
The Hampton RERP' relies on a peak season population computed using secondhand information prepared by nonlocal sources. .
Estimate of Hampton peak population of 110,000 figure from earlier NHRERP revisions) is less than one-half of estimate Hampton Chamber of Commerce, using parking spaces, business receipts and shifts in municipal services.
Basic C.2 of this contention maintains in part that the Hampton RERP fails to account for " driver disobedience, panic," therefore, fails to provide assurance available Public Works Department personnel can ensure all evacuation routes are serviceable.
Basis G of this contention maintains in part that it is unreasonable to assume that officials will be available to evacuate Hampton school children.
Basis H of this contention maintains that the " State acknowledges that Hampton may have an ' upper peak seasonal population of 110,000' revised Hampton Plan I-II."
Contention SAPL-8A reads as follows:
"The New Hampshire Compensatory Plan fails to meet the requirements that there be adequate manpower and 24-hour per day emergency response including 24-hour per day manning of communications links, as required by 10 CFR 50.47 (a) (1) , 50. 47 (b) (1) ,
NUREG-0654, II.A.1.e, II.A.4, and II.F.1.a.
Additional bases under this contention and filed by SAPL holds in part that individual drivers of the Teamsters Local i
w---. - _ _ - - -
No. 633 "have in no way shown their willingness to drive into the EPZ."
SAPL Contention No. 18 as revised reads as follows:
"The NHRERP Rev. 2 significantly miscalculates the numbers of non-auto owning population for the 17 New Hampshire local communities. No buses are provided in the plans for the individuals who are not accounted for due to these miscalculations. Therefore, these plans fails (sic) to meet the requirements of 10 CFR
- 50. 47 (a) (1) , 50. 47 (b) (8) , NUREG-0654 II.J.10.g and NUREG-0654 Appendix 4, p. 4-3."
In relevant part, the basis to this contention holds that:
- 1. The NHRERP Rev. 2 erroneously assumes an EPZ-wide nonauto owning population percentage of 2.5 percent; it is based on this percentage that bus needs are assessed, and
- 2. The population on which the percentage calculation is made has been underestimated by a very significant margin.
SAPL Contention No. 37 reads as follows:
"The NHRERP Rev. 2 fails to provide reasonable assurance of adequate public protection because an adequate number of emergency vehicles are not provided for in the plans and further there is no assurance that effective use of these vehicles vill be possible in view of a potential outgoing flow of evacuating traffic and a significant lack of drivers therefore, these plans do not meet the requirement of 10 CFR 50.47 (a) (1) , 50. 4 7 (b) (3) ,
50.47(b)(10) and NUREG-0654 II.J.10.g and II.J.10.k."
The basis to this contention, in relevant part, holds that:
- 1. The KLD Study shows traffic congestion in the City of Portsmouth throughout the evacuation process; transportation resources must be requested from the State, there could be significant degrees of difficulty in getting buses through traffic.
l
- 2. Portsmouth is the location of one of the '
transportation staging areas (at OMNE Mall). The Towns of Seabrook and Hampton Falls are among the towns to be served by buses from that staging area (Vol. 4, p. 1-10).
Portsmouth is in the northernmost section of the New Hampshire portion of the EPZ. Hampton Falls and Seabrook are at the southernmost section of the New Hampshire portion of the EPZ. It is therefore held by SAPL that buses will need to travel through congested traffic conditions over considerable distances, that adequate traffic control measures will not likely be in effect due to local and State Police personnel and resource constraints, and that nothing will constrain evacuees from using those lanes needed for inbound emergency vehicles.
II. PLANNING BASLS A. Evacuation Time Estimates
- 1. Introduction Volume 6 of the New Hampshire Radiological Emergency Response Plan, "Seabrook Station Evacuation Time Study," was developed in accordance with the guidance provided by NUREG-0654, FEMA-REP-1, Rev. 1, " Criteria for Preparation and Evaluation of Radiological Emergency Response Plans and Preparedness in Support of Nuclear Power Plants." Appendix 4 of this document, " Evacuation Time Estimates Within the Plume 1
Exposure Pathway fcr Emergency Planning Zone," provides l specific guidance for the development of the Evacuation Time l
Estimate for Seabrook Station. In addition, NUREG CR-1745 entitled " Analysis of Techniques for Estimating Evacuation Time for Emergency Planning Zones" was used as a basic refere. ice.
The section of this testimony which presents the population estimates discusses thase elements of Volume 6 which have come under scrutiny as a result of filed and admitted contentions. Following this discussion, we will review activities which have been pursued since the Revision 2 publication in an attempt to gather and maintain accurate, current data relating to the ETE.
All page references in Section A of this testimony entitled " Evacuation Time Estimates (ETE)" refer to State of New Hampshire Radiological Emergency Response Plan (NHRERP),
volume 6, "Seabrook Station Evacu~ation Time Study," Revision 2, unless otherwise indicated. As documented by Appendix 4 of NUREG-0654, "the evacuation time estimate will be used by those emergency response personnel charged with recommending and deciding on protective actions during an emergency."
This ETE provides information regarding the anticipated time required to evacuate specific regions of the Emergency Planning Zone (EPZ), under various scenarios. This information, in turn, is to be utilized as one criterion by .
i j
4 I
)
emergency response' decision-makers in reaching protective action decisions.
- 2. Population Estimates: Volume 6 )
- a. General Volume 6 of the NHRERP details the sources of population j l
and vehicle estimates utilized within the'ETE. All of the sources are empirical, arising from direct observation, from )
extensive survey results, or from the files of the 23 town clerks. These estimates of current resident population, as found in Table 1 of Volumes 16 through 32 of the NHRERP, and as updated within Section A.3, " Surveys and Analyses !
Subsequent to Publication of Revision 2, NHRERP," of this testimony, are independent of prior ETE estimates. ;
- b. Permanent and Peak Population Estimates Estimates of permanent (resident) population utilized by the ETE were derived from information from Town Clerk estimates in early 1985 based on local census activities.
Compounded annual rates of growth for each EPZ town were l calculated using State census data for the year 1980 and for I the year 1985; these estimated rates of growth, and the resultant projected 1986 population and vehicle estimates are listed at p. 2-9, Volume 6. To confirm the validity of those l
population estimates provided by local sources, the 1984 population estimates of the New Hampshire towns as provided by the New Hampshire Office of State Planning were also projected to 1986 using the most recent annual growth rate
- -______ _- ___________a
l I
for each town. These estimates, listed on p. E-11, I demonstrate very close comparison of the local and state derived estimates.
Peak population estimates utilized in the ETE represent the total peak population which may reasonably be within a l l
particular town at any one time, for each of 10 " scenarios." J
.These scenarios, defined in Table 10-1, consider different 4 times of the year and of the day, as well as different weather conditions. These population estimates which are relevant to the calculation of the ETE differ considerably from cumulative (g.g., 24-hour) traffic counts and suggested unsubstantiated estimates inferred from local business figures. It is noted that KLD sought information regarding any local " traffic counts and local business figures of peak population or vehicles" from the Town of Hampton and Chamber of Commerce of Hampton Beach; KLD was informed that no such traffic counts or person counts existed.
Table 1 in each of the local community RERPs is entitled
" Population of Municipalities Wholly or Partially Within 10 Miles of Seabrook Station." It provides a listing of estimates of resident and peak summer weekend and midweek populations by municipality. These estimates were prepared using information contained in Sections 2, 5 and 6 of Volume 6 of the NHRERP. The table estimates the peak summer weekend l l
population for the Town of Hampton as 36,635. The table estimates the peak summer weekend population for all of the 1
l C_______________.______
i New Hampshire towns at 142,569, the peak summer midweek population for the New Hampshire towns at 139,929, the peak !
summer weekend for the entire EPZ population at 228,292, and i 1
the p_ak summer midweek EPZ population at 225,726. (Note l l
that peak weekend population estimates within the seacoast EPZ towns have been reevaluated following analysis of aerial photographs taken on July 18, 1987. This analysis is discussed in Sections A.3 and A.4 of this testimony.)
(The listed figure within Table 1, Rev. 2, indicating the estimated peak summer midweek population for the Town of Hampton was mistyped in Volume 18. This estimate of population should read "34,337" not "31,337.")
- c. Surveys and Analyses Sucoortina Pooulation Estimates In arriving at these population estimates used in the ETE, a multitude of detailed and varying surveys and analyses were conducted. These activities include, but are not limited to, the following surveys, counts, and observations.
- 1. On page E-6, Volume 6, it is shown that detailed examination of large-scale photos taken on July 4, 1983 consisted of a " count of people" (emphasis added) on the most crowded portion of the beach which is opposite the Hampton
" Casino," close to the sanitary facilities maintained by the State of New Hampshire. This count included the entire beach area extending into the water. It was noted that virtually I
all blankets were on dry sand above the high tide line. This dry beach area was measured and used to calculate person
~1 density, not capacity. (Note that if the beach area had been I measured to the low tide area, the value of density would have been commensurately lower.)
- 2. To estimate the maximum number of vehicles (and people) that could occupy the beach areas, the ETE " relied on empirical examination of the number of vehicles which can )
physically be accommodated within the beach area," Volume 6,
- p. 2-1. Also, parked vehicles (see pp. E-4, E-5 and 2-10) and vehicles in transit (see p. 10-16) were counted. The phrase "which can be physically accommodated" (Volume 6, p.
2-1) was stated in the context of reasonable expectation.
The intent of the phrase was to relate the observed unused
, parking space to what would be reasonably used, based on the prevalent usage in each area as viewed on the aerial film.
An extensive number of photographs of specified date and time were used within the ETE in estimating reasonable parking capacity along the beaches. Nine sets of color slides of the entire coastal area were examined with each set consisting of about 55 slides (p. E-4). The data representing the most crowded situation was that "on Sunday, Angust 11, 1985 (taken) in the early afternoon" (p. E-5, Volume 6).
- 3. Thirty-six surveys to determine vehicle occupancy rates in the beach areas were conducted, as listed on p. 4-7, Volume 6. These counts were taken of travellers in the beach ;
area and do not represent household populations. The i
L
].
l 1
estimate of 2.4 persons per vehicle was based upon the 6 surveys totalling over 5,000 vehicles taken during fair weather and on crowded beaches during July 4th and the morning of the 5th, 1986 (p. 4-8,. Volume 6). The counts
~
taken during rainy weather in late 1985 which led to an average occupancy rate of about 2.2 persons (p. 4-6) were discarded.
Furthermore, during 1983, which may be considered one of the best tourist seasons in recent years for the New Hampshire coastal area (meteorological data gathered at the Seabrook Station meteorological tower indicate that 1983 had the second highest average summer weekend midday temperatures over the past seven years), traffic counts were taken with automatic traffic recorders (ATRs) along all major beach access roads. These data were collected 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> a day, 7 days a week throughout the entire summer season. Table 2-3 on p. 2-13 documents counts of traffic entering and exiting the Seabrook and Hampton beach areas on the most crowded day of this 1983 season. As indicated there, the peak net July 16th influx of about 5,600 vehicles took place at 2:00 p.m.
This count is consistent with the data for Hampton and Seabrook beaches which were obtained using the 1985 aerial l
film. Refer to the discussion on pp. 2-8 through 2-13, i
Volume 6.
l 4. The ETE empirically observes that beach population l
varies widely from day to day, depends most strongly on l
l l
I .
weather conditions and also varies with time of day (p. 2-10, Volume 6). It is noted in this section that, "on a sunny day it (beach population) generally peaks about 2 p.m.," which is about the time the 1985 aerial photos used in the Revision 2 parking analysis were taken.
- 5. The analysis which develops the estimate of average person occupancy of evacuating vehicles used by permanent residents is detailed in Exhibit 2-1, p. 2-5, Volume 6.
Information employed for this purpose was obtained from a telephone survey of EPZ residents. Supporting data for this survey are presented in Figures 2-2 and 2-3 and Appendix G.
This estimate equals 2.6 persons per evacuating vehicle and is applied only to the permanent resident population; it does not apply to those people returning home to pick up family members.
- d. Estimated Number of Transportation Dependent The ETE estimates that 4,495 people within New Hampshire will require transportation assistance (see NHRERP, Volume 6, i Table 11-6). This estimate of 4,495 people with no car available constitutes approximately 5% of the total resident i
population within the New Hampshire portion of the EPZ. This estimate was determined by applying a statistical analysis to the empirical base obtained by the KLD telephone survey.
In response to concerns expressed by FEMA that the plan should take into account the high probability that ride sharing would provide transportation for most of these people l
l
___ _______________-_-_ - ___________a
l.
(RAC Review:of State and' local Radiological Emergency Response Plans,Section VI, A.1.), an estimate of 50% has been applied to indicate those transportation dependent l
l= evacuees who'will ride-share. This ektimate of 50% was based on the Mississauga experience (see p.'11-8, Volume 6) and on the suggestion of Mr. Edward Thomas, Region 1, FEMA, that for planning purposes, 70% of transit dependent persons would rideshare. Volume 6 utilized a lower figure of 50% to
- estimate the number of buses needed so as to ensure adequate bus resources. Thus, the estimate of 2,249 people requiring transit are those persons who do not have access to a vehicle and who also do not have access to ride-sharing. Volume 6, pp. 11-9, 11-10 and Table 11-6. This estimate of 2,249 agrees favorably with the results of the NHCDA Special Needs survey which identified a total of 2,106 people who do not have access to a private vehicle for evacuation.
The survey conducted by KLD in August 1985 which is
(_ described in Volume 6, pp. 11-1 through 11-11, and Section 5.i of this testimony was based on a sample of the EPZ population. One of its objectives was to estimate the number of persons within the EPZ requiring transit assistance (exclusive of those residing in special facilities) so that the number of buses required to accommodate them could be calculated. The March 1986 NHCDA survey, on the other hand, was conducted to identify those persons with special needs as well as those requiring transit assistance.
Finally, conclusions about the behavior of people in emergencies are supportive of a greater percentage estimate of ride-sharing. A closer inspection of this work provides a more detailed reflection of general behavior in emergency situations. (See Section B.1.b entitled " Ride Sharing".)
- e. Out-of-Service Vehicles l To estimate the number of vehicles which could be out of service (i.e. inoperable) at any given time due to mechanical problems, a telephone survey was undertaken of fleet operators on Long Island, New York. The respondent was asked to estimate how many days per year the average fleet vehicle l
l was inoperable, or alternatively, the number of miles driven before a vehicle became inoperable. The answers varied widely. One owner maintained that vehicles remained in service, except for preventative maintenance, for 100,000 miles before breakdown; another indicated a figure of 50,000 miles. We adopted the most pessimistic response which was an estimate of four days per year; or 1.1 percent of the time.
Thus, the probability that a household which owns a single car, will have that car out-of-service at any time, is 0.011. For a household with two cars, the probability that both cars are out-of-service is 0.00012. It is now possible, using the survey data of Figures 2-2 and 2-3, to estimate the number of people with out-of-service vehicles:
Percent of households with one car:
(0.14 x 0.79 + 0.35 x 0.35 + 0.19 x 0.19 + 0.19 x 0.19 + 0.08
l i
x 0.14 + 0.05 x 0.21) x 100 = 32.75% I percent'of households with two cars:
i (0.14 x 0.79 + 0.35 x 0.55 + 0.19 x 0.54
- 0.19 x 0.52 +-0.08 I x 0.51 + 0.05 x 0.28) x 100 = 45.85% 1 I
Total number of households: 142,194/2.87 = 49,545, j i
The av,erage size of households that own one car is: j l
(1 x 0.79 x 0.14 + 2 x 0.35 x 0.35 + 3 x 0.19 + 4.x 0.19 x l 0.19 + 5 x 0.08 x 0.14 + 7.2 x 0.05 x 0.22) /0.3275 = 2.27 ]
The average size of households that own two cars is:
(1 x 0.14 x 0.07 + 2 x 0.35 x 0.35 + 3 x 0.19 x 0.54 + 4 x .
0.19 x 0.52 + 5 x 0.08 x 0.51 + 7.2 x 0.05 x 0.28) /0.4585 = !
2.75 l 1
Estimated number of people whose cars are out-of-service:
(0.011 x 0.3275 x 2.27 + 0.00012 x 0.4585 x 2.75) 49,545 si 413 people We wished to avoid the documentation of this arithnetic 1
in Volume 6. The procedure outlined on p. 11-9 calculated the number of transit-dependent. persons. Within the context I of this procedure, vs determined that the multiplicative factor of 1.06, applied to this estiNate of transit dependent )
I persons, would yield a'value approximating the 413 calculated above. That is (see Table 11-6) , 74 58/1.06 x 0. 06 = 422, '
l which approximates 413. J
- f. Estimates of Enclove1._P_gight.ign ,
Section 5 of Volume 6 is devoted to' the Estiraation of Employee Population." Within the ':.'own, of Hamp":on, the number of employees who work at the behch arn during the summer .;
season was estirated to' be 25 percen :i M the total number of employees in H eptan. -
T
) '
)
h !
i I This percentage is based on the seasonal employment i l
l- picture shown in Table 5-2 which documents the significant I summer employment in Hampton. As shown in Table 5-2, there are some 2,000 more employees within Hampton in the summer :
than in the off-season. If all of these summer employees I worksd at the beach then about 38 percent of the total employee population in Hampton would work at the beach.
Clearly, this estimate represents an upper bottnd. There are many inland establishments serving seasonal traffic l 1
(restaurants, motels, retail storos) that either close j l
off-season or which depend primarily on seasonal trade.
l These establishments, by definition, provide a basis for
. sesaonal employment. The ETE provides for these seasonal !
i employees within Hampton, but away from the beach. The i
estimate of 25 percent for beach employment is equivalent to the assumption that about one of three summer employees works fL away from the beach which we have considered to be a '
reasonable expectation.
The following sensitivity study demonstrates that any uncertainty in this estimate has a miniscule effect on the ETE. Suppose that instead of 25 percent, only 15 percent of summer employees are on the beach -- this would increase the
,. number of evacuation vehicles by about 100. That is:
(706/O.75) (0.25-0.15)=94. If, on the other hand, 35 percent g are en the beach, (close to the upper bound of 38%), a I
decrease of some 100 evacuating vehicles would result. I
[
L I
mm u _____.__
Note that this consideration is limited to the summer scenarios. In the first and last columns of Table 5-4 the off-beach (i.e., inland) employees are estimated for the summer scenarios. These estimates are computed taking into-account that an estimated 25% of employees work at the beach for Hampton and Rye, and 10% for Seabrook, Salisbury and NewbKry. The vehicles observed in the aerial photographs, which belong to employees at the beach, are indistinguishable from those belonging to tourists there, who will also be evacuating. In the off-season (next-to last column), all ,
evacuating employees are treated as such, since there are few tourists at that time.
- g. Population Growth
- 1. Population croiections.
There exists no current regulatory requirement to project into'the future when formulating ETE. In accordance with NUREG-0654, Appendix 4 which states that the " evacuation time estimates should be updated as local conditions change",
the ETE will be subject to the continuous planning process.
- 2. Emolovment arowth.
It has been contended that the ETE has failed to account for the " reasonably" anticipated and substantial growth in population and motor vehicles within the EPZ due to employment growth. This, however, is not the case. NHRERP, Volume 6, Table 5-1 projects the historical growth of total i
i
)
{ cmployment from 1980 to 1984, as determined by the NH Labor I
l Services and Employment Bureau, forward to 1986.
- h. Seasonal bausina and overnicht accommodations.
l Pages 2-14 and 2-17, Volume 6 provide discussion regarding the estimates of Seasonal Housing Residents and
~
Overnight Accommodations. NRC figures as prepared by Kaltman in 1981, were used as baseline information sources in the preparation of these two population estimates, as they were the most up-to-date figures available at the time of ETE development. However, it is necessary to estimate the number of vehicles in the EPZ belonging to transients at overnight accommodations which are not at the beach when the beaches are most crowded. The figures provided by the NRC via the Kaltman Report are therefore used with the following exceptions:
- 1. In order to avoid double counting, those vehicles at the beach which originated at sea ~sonal housing away from the beach were excluded from the count of inland tourists, as these vehicles are included in the count of beach vehicles.
- 2. Discussions with managers of tourist facilities
{
indicated an estimated 74% of visitors at off-beach I facilities travel to the beach and park their vehicles there during mid-day weekend conditions. A factor of 50% was adopted by the ETE, i.e., that half of the tourists lodging at inland facilities will drive to the beach on a sunny day 4
1 4
i l
with the other half remaining within the EPZ but not at the l beach.
The aerial photographs provided the ETE with an estimate of vehicles along the beaches. The NRC data in the Kaltman Report, together with a survey which established the 74%
figure above, provides the basis for estimating the number of tourists lodged at inland facilities that are not at the beach. (Note where Kaltman estimates weekend vehicle demand for beach area seasonal residents at 10,449 vehicles (Figure 2-4), Volume 6 estimates total beach parking capacity at i
25,470 (p. E-5).)
Volume 6 does not depend solely on the NRC estimate of 2.5 vehicles per dwelling at seasonal housing units. On p.
E-10 it is stated: "KLD's on-foot survey recorded an average value of 2.6 vehicles per dwelling."
Estimates of vehicles associated with overnight accommodations, i.e., hotels, motels and guest houses, considered the same exceptions as noted for seasonal housing residents, as well as the following:
- 1. Arrival times at many overnight accommodations will be after 2:00 p.m. after the beach population has dropped below peak;
- 2. Departure times are usually before peak population conditions at the beach;
- 3. Many, if not most, patrons remaining for several I
I
- 2s -
l
-l days will leave the facility to go to the beach, or shopping
~
or go to sono recreational attraction;-
- 4. The number of vehicles'per' housing unit may be less q
than one as: (a) family / friends arriving in one car may
. occupy more than one unit, and'(b) bus travelers will; occupy many units. i Several larger motels indicated they set aside blocks of rooms on weekends for tour buses, at a rate of 20 units per bus. Estimates ranged from 5 to 40% with regard.to those guests utilizing more than.one unit per. car; thus, an estimate of 0.85 vehicles per unit was adopted.
- i. Camoarounds.
Page.2-20, Volume 6 discusses the manner in which vehicles at campgrounds within the EPZ were estimated. Again, .
the NRC base data was considered, with the exception that double-counting was avoided-as with overnight accommodations (see Section A.2.h'above). Also, campground operators estimated that 75% of campground sites are unoccupied during peak beach hours.
It is noted that NRC's figures estimated a total of 3,147 vehicles at EPZ campgrounds (Figure 2-8, Volume 6).
Current campground managers' estimates of maximum numbers of vehicles (see " Campground Maximum Capacities," an NHCDA on-going survey), including capacity overflow areas total to' l I
2,938 vehicles, close'to the number of vehicles estimated by
.the NRC's 1981 report. l 1
~ _ . _ _ _ _ _ _ _ _
- 3. Surveys and Analyses Performed Subseauent to Publication of Revision 2. NHRERP Since Revision 2 of the NHRERP was submitted, additional surveys and studies were conducted for the purpose of obtaining a broader data base and an updating of estimatas.
These included:
)
- a. Aerial Survevs. On July 18, 1987 Avis Airmap of Braintree, Massachusetts took aerial photographs of the Massachusetts and New Hampshire seccoast region extending from the southern tip of Plum Island, Massachusetts northward to Odiorne Point, New Hampshire. This photographed section j of coastline extends from approximately 15 miles south to 12
- miles north of Seabrook Station and approximately 1 mile inland from the shoreline. The weather conditions on the day of the flight were ideal for beachgoers: sunny, warm and in the mid-80s. This weather attracted an attendance at the beach which was comparable to that on the peak day in the summer of 1983, July 16th.
From en altitude of approximately 3,000' an RC10-A camera with a 6" lens was utilized to take each photograph, producing contact prints with a scale of 1" = 600' or 1:7200.
The fly-over began promptly at noon and a complete set of 187 black and white 9" x 9" contact prints were provided for
{
review and for the selection of individual enlargements. A series of 58 enlargements with an enlargement ratio of 12.5 (1" = 50') were provided covering a continuous strip from j 1
)
i i
l Plum Island, Massachusetts to Odiorne's Point, New Hampshire.
An enlargement ratio of 12.5 was selected because it represents the largest magnification consistent with visual clarity. Because the photographs were taken stereoscopically (with the exception of Flight 10 which is monoscopic) there is approximately a Got overlap on adjacent prints.
Therefore, only half of the contact prints were needed to produce the photo enlargements as there was sufficient overlap to ensure continuous coverage of the study area.
The data reduction activities using the enlarged photographs to estimate the vehicular and beach-going population were also performed by Avis Airmap. The following
., step-by-step procedure was undertaken:
Enlargements were labeled to correspond to the contact printnumbering system.
Mylar acetate overlays were placed over the enlargements. Register marks were drawn on the photographs and on the mylar so that if the overlay were detached from the photo, it could be repositioned exactly as when the data were recorded.
Municipal boundaries were inked onto the photo enlargements so that counts were totaled on a town-by-town basis.
A grid comprised of 2" x 2" squares was drawn on the mylar over the beach area. Grid squares were assigned a unique code by using numbers on the horizontal axis and the Arabic alphabet on the vertical axis (A1, A2, etc.).
For vehicular parking areas, parking lots and residential areas were demarcated and assigned specific numbers.
l Applicants' Exhibit 3 is a group of six 9" x 9" color contact prints. labeled 6-3, 6-5, 6-7, 6-9, 6-11 and 9-1 of the area photographed.
- 1. Beach Pooulation Two categories of people were counted:
People on the beach (on the sand and in the water);
and People occupying both sidewalks and parking lots abutting Route 1A or the beaches.
The beach population within each cell of the grid was cou'nted by connecting each individual observed on the photographs with a continuous line drawn with a permanent ink marking pen. The lines were drawn from top to bottom of each square and are not of a consistent length. Rather than having one long line, several shorter, non-intersecting lines are contained in each square. The number of these lines connecting observed persons varied according to the density of the persons in each grid. This method of connecting observed persons with continuous lines was selected as it ensures that all " outlier" persons who have not been counted will be detected during visual inspection and therefore not excluded in the final count.
Tallies for each grid square were then placed on a tally sheet and totaled at the bottom of each sheet. In many instances, several tally sheets were used for each photo enlargement.
r . - . _ _ - - - _ _ . _ .
l
- 2. Vehicular Counts Four categories of vehicles / vehicle-spaces were counted l and include: 1 i
filled parking spaces, i.e. all parked vehicles not I along curbs l
unfilled delineated parking spaces -
filled curbside spaces vehicles in transit on the roadways )
A grid pattern forming sub-areas was used to count vehicles. This grid pattern differs from that used to count t' people in that the vehicular grid cells demarcate separate parking lots and areas which may be irregularly shaped. This method was selected so that land use, e.g., residential, beach-front, motels, would be grouped together.
Vehicle counts were recorded by category on tally sheets and sums recorded by each photo enlargement. As was the case with the beach population counts,.one enlargement may have several pages of data depending on the number of parking areas. All counts were ultimately totaled by town.
Parked cars in four areac west of the beach (and west of Route 1A as well) were counted by using the color contact prints rather than the photo enlargements. A stereoscope was used to count the cars in these four areas. Subsequently, enlargements of these areas were obtained and the counts were re-checked for numerical accuracy. Tallies for these areas j i
were recorded in the same manner described above. l The beach population for the seacoast had previously been determined to peak at 2:00 p.m. on summer weekends (see l l
{
_ - _ _ _ i
the Beach Area Traffic Count Program, performed in 1983 by HMM Associates, pp. 2-12, 13 of Volume 6). The 1987 aerial photographs were taken by Avis between 12:00 noon and 1:20 p.m. To obtain a reasonable estimate of peak vehicle population on the beaches, it is necessary to project the observed estimates of the number of vehicles counted on the photo enlargements forward to 2:00 p.m.
The procedure used to project these counts to 2:00 p.m.
applied the data in NHRERP Volume 6, Table 2-3 as follows:
Determine the total number of vehicles observed on the photo enlargements for Seabrook and Hamptonbeaches (south of Route 51). This total count at 12:15 p.m. on July 18, 1987 was 11,889 vehicles. Determine the net influx of vehicles (i.e., entering minus exiting) onto these beaches between noon and 2:00 p.m. on the day exhibiting the peak population in 1983, on July 16th, as shown in NHRERP Volume 6, Table 2-3:
Net Influx Time (Vehicles) 12:00-1:00 p.m. 881 1:00-2:00 p.m. 426 Using the net influx data above, the percent increase in vehicle population over the time frame in which the photos were taken was projected to 2:00 p.m. The results of these calculations are presented below.
Photo Projected Net Projected Percent Set / Influx to Increase in Flight Time Photos 2:00 p.m. Vehicle Population Number ,
Taken (EDT) (Vehicles) to 2:00 o.m.
1 12:50 573 4.62 2 12:00 1,307 11.20 l 3 12:45 646 5.24 4 12:05 1,234 10.51 l
l
5 12:40 720 5.88 6 12:15 1,087 9.14 7 ~12:30 867 7.16 8 12:25 940 7.81 9 12:20 1,013 8.48 The data in the first two columns of the above table were provided by Avis. The data in the third column were obtained by interpolation from the data in Table 2-3, Volume
- 6. The final column is obtained by dividing the respective
. data in the third column by the imputed count of vehicles on Seabrook and Hampton beaches at the indicated times.
To illustrate the procedure, take for example'Hampton and Seabrook beaches (Flight Number 6) which were photographed at 12:15 p.m. We expect an additional 1,087' vehicles would park on these beaches between the time the I'
photos were taken, and when the peak condition is attained at 2:00_p.m. This'is equivalent to stating that the number of vehicles on these beaches will increase by 9.14% (3087/11889 x 100) from 11,889 to 12,976 if all photos covering this area were taken on flight number 6. The vehicle counts from all photos in Set 6 would therefore be increased by 9.14%, on <
this basis.
To calculate the percent increases in vehicle population for the other times, it is necessary to estimate the vehicle I population on these beaches at these other times, using the data of NHRERP, Volume 6, Table 2-3. For example, the estimate'of vehicle population at 12:45 is:
[ 11,889 (at 12:15) + 1/2 x 881 = 12,330 1 1.
p-l-
I l
E 1 Thus, the percent increase in population between 12:45 and 2:00 p.m. is:
(12,976 - 12,330)/12,330 x 100 = 5,24 This approach implies the assumption that the rate of net influx of vehicles onto the beaches on July 18, 1987 is comparable to that recorded by HMM on July 16, 1983. While the data gathered on July 18, 1987 does not permit a direct comparison of net influx between noon and 2:00 p.m., we can compare the number of entering vehicles on these days, on two approaches to the beaches for which data exist.
VOR Survey HMM ATR Counts 11:30-12:30 11:30-12:30 Enterina Vehicles July 18, 1987 July 16, 1983 Route 286 975 856 Route 51 111 22R TOTAL: 1,844 1,854 The close agreement in the number of entering vehicles for the indicated days lends support for this procedure which estimates the increase in vehicle populations from the time the photos were taken, to the peak condition at 2:00 p.m.
The following numbers represent the total number of observed parked vehicles, projected to 2:00 p.m.
L !
. J 1
eip th th ce0 ce0 j
eV0 1767823311 9 eV0 619 6
- 3212 2871 9 j : 008 9 od2 286 8 od2 425 1 re ,
re ,
Pkt 1 Pkt 1 ra a
ra
a
P P E E es es se se al al ec. ec.
rchi m. rchi m.
nep nep IV IV 0 0 dd0 dd0 996 ee: 3323102738 ee: 315 tk2 285 tk2 cr cr eao jPt eao jPt o o rf rf Po Po
)
v r d d u de de b ev ev w kr kr e re re N as as Pb Pb O O 723 2 v f f 683 8 e os 8444721683 7 os 315 0 l e 2212 2585 1 e ,
w rl 275 7 ) rl 1 o ec ,
t ec R bi 1 r bi =
mh o mh ue = o ue E h
c NV v NV E r i u w b s w p e I
(
N
(
h i h #
t t u t r t o n o n S e N e SN1SN m S m d e 2357715578 d e W0 n g
- - - - - - - - - - n a 991
- a r 2222224444 a r - - -
l a l a 444 s s I
l n
E I
l n
E m m u u l l P P
eip 4 eic th 150844999 9 th -
ce0 943053842 1 ce0 7698 0 j
eV0 3129 ,5354 1 eV0 4548 4
- , 1 , j : 6345 0 od2 1 6 od2 . ,
re re 1 3 Pkt = Pkt ra ra =
a E a P P E es es se se al _ al ec. ec.
r i m.
ch rchi m.
nep nep IV IV 0 792671766 0 4083 dd0 302845343 dd0 5333 ee: 1 1 ee: 1 tk2 tk2 cr cr eao jPt eao jPt o o rf rf Po Po d d de de ev ev kr kr re re as as Pb Pb O 468273233 8 O f 530298509 4 f 3615 5 os 302834353 5 os 9215 8 e , , ,
e 5344 7, rl 1 1 5 rl .
ec ec 1 2 bi = bi mh mh =
ue E ue NV NV E i
t #
n e WESNSN t m 2244668SN n e 111111133 e a - - - - - - - - - m NSNS r 444444466 e 35S7 a a - - - -
y l r 6666 r n k a u E o l b o n s
r E i b l a a e S S
lm lm dc. dc.
eip 4 ein s
I th j
ce0 eV0 :
150844999 943053842 3129 ,5354
, 1 9
1 1
th j
ce0 eV0 7698 4548 6345 0
4 0
od2 1 6 od2 . ,
re re 1 3 Pkt = Pkt ra ra
=
a E a P P E es es se se al al ec. ec.
rchi m. r i m.
ch nep nep IV IV 0 792671766 0 4083 dd0 302845343 dd0 5333 ee: 1 1 ee: 1 tk2 tk2 cr cr eao jPt eao jPt o o rf rf Po Po d d de de ev kr ev kr re re as as Pb Pb O 468273233 8 O f 530298509 4 f 3615 5 os 302834353 5 os 9215 8 e , , ,
e 5344 7 rl 1 1 5 rl . ,
ec ec 1 2 bi = bi mh mh =
ue E ue NV NV E t #
n e WESNSN t m 2244668SN n e 111111133 e a - - - - - - - - - m NSNS r 444444466 e 3S57 a a - - - -
y l r 6666 r n k a u E o l b o n s r E i b l a a e S S
\ll
11l l
s s
- e. e.
lm lm _
dc. dc. _
eip ei D -
th th 208750 -
2 .
ce0 066472 5 ce0 719619 5 -
eV0 : 559487 0 eV0 : 876315 3 _
j 1811 4 9 j ,
od2 , , , . ,
od2 3 re 123 2 9 re Pkt Pkt =
ra = ra a a E P E P es es se se al al ec. ec.
rchi m. rchi m.
neD nep IV IV 0 698279 0 311786 dd0 9361 7 dd0 6552 4 ee: 22 1 ee:
tk2 tk2 cr cr eao jPt eao jPt o o rf rf Po Po 6
3 d d de de ev ev kr kr re re as as Pb Pb O -
O f 478203 4 f 997074 6
) os 512389 0 ) os 054404 0 1 e 0691 2 1 1 e 866315 1 5 rl , , , . , 5 rl ,
ec 122 2 9 ec 3 e bi e bi t mh = t mh = -
u ue u ue -
o NV E o NV E R R -
f f o o h h t t u r o # o #
s n
( t ( t n n h
c e
m NSNSN WW h c
e a e 799992 m SNSN a e 466882 e a - e a - - - - - -
B r 666668 B r 888889 a a n l n l
_ o n o n E
t t E p p H
m a m a
H ll l il! l
ll 11 1llIIllI ll l!
lllI' ll.ll lljjlllI l if!I ll1
)
- j dc.
eip th ce0 eV0 od2 re Pkt s
e.
lm ra a
4473 4041 2
8 0
3
=
dc.
j eip th ce0 eV0 od2 re Pkt s
e.
lm ra a
94695962476601 56443595981290 451211 372 1 7
4 3
4
=
P E P E es es se se al al ec. ec.
r i m.
ch r i m.
ch nen nep IV IV 0 3531 0 45200941782677 dd0 1 dd0 341111 251 ee: ee:
tk2 tk2 cr cr eao jPt eao jPt o o rf rf Po Po 7 _
3 d d de de _
ev kr ev -
kr re re as as Pb Pb f
o O 1942 6 f 59495021794034 2 os 4841 8 os 55102484856189 2 e 1 2 e 451211 362 2 rl rl ,
ec ec 3 bi bi mh = mh =
ue ue NV E NV E t t n n n e SNS e NSNSNSNSNSNSN o m N002 m 24466880022446 t e 8111 e 11111112222222 p c - - - - c - - - - - - - - - - - - - -
m r 8888 r 88888888888888 a a a H l l n n h E E t
r e o y N R I lll l
l 1
3-
SUMMARY
Observed Projected Parked Parked .,
Town Vehicles Vehicles i Plum Island South 1,717 1,899-Plum. Island North 1,082 1,196
. Salisbury . 5,548 6,119 Seabrook Beach- 2,785 3,040 Hampton Beach South 9,104 9,905 Hampton Beach North 3,106 3,352 North Hampton 286 308 Rye -3,222 3,474 E = 29,293 l These estimates of peak beach area vehicles which are 1
used as inputs to the IDYNEV model for calculating ETE, may be compared with the listing of " Estimate of Parking g Capacity" on page E-5 of Volume 6. As indicated, the sum of new estimates are approximately 15%-higher than the estimated l . total, and amounts to a 4% increase, relative to.the entire vehicle population of the EPZ.
- b. 1987 Vehicle Occuoancy Rate Surveys comprehensive vehicle occupancy rate surveys taken at selected EPZ beach area roadway points were conducted on July 11 and 18, 1987. Over 24,000 vehicles were counted, I containing a total of approximately 56,000 individually counted occupants. The mean occupancy rate of the vehicles surveyed was 2.274 occupants per vehicle, about the same as the 2.335 estimate obtained in July-1986. An estimate of 2.4
. persons per vehicle was utilized in the ETE in estimating this population segment. (See Applicants' Exhibit 4: j I
)
l
" Report on the Vehicle Occupancy Rate (VOR) Survey Process.")
- c. New Hamcshire~ Office of State P(gnnino Populations Efd;1* nates The'New Hampshire Office of State Planning has released its current New Hampshire population estimates in a publi. cation dated August 1987, entitled, "1986 Population ,
Estimates of New Hampshire Cities and Towns." These most recent estimates compare well, overall, with the 1985 estimated permanent population figures utilized in the ETE.
In fact, comparing these NHOSP estimates (89,619) with the resident population estimates found in Table 1 of the. local RERPs (92,601) indicates that the ETE differs by a total of 1,982 persons in excess of NHOSP estimates. The largest numerical differences between the estimates are found in the listings for Hampton and Seabrook which are, respectively, 1,157 and 1,463 persons in excess of the NHOSP estimates.
See Attachment 1 hereto, selected pages from "1986 Population Estimates of New Hampshire Cities and Towns",
- d. Ramo Capacities Upon further survey of highway ramos and review of the Highway Capacity Manual below, it was determined that the initial ramp capacity was underestimated. (The Highway Capacity Manual (HCM), Special Report 209, Transportation ;
l Research Board, National Research Council, Washington, D.C.
1985. This is the most recent authoritative manual for
}-
estimating highway capacity, as organized through the Transportation Research Board, with contributions by the
Federal Highway Administration, National Cooperative Highway Research Program, individual states through the American Association of State Highway and Transportation officials, as well as many other National and International supporting projects and individuals).
Consequently, the nominal capacities of the ramps for undersaturated flow conditions is now estimated at approximately 1,330 vehicles per hour; under congested conditions,-the estimated ramp capacity is approximately 1130 vehicles per hour. These ramp capacities may be further reduced due to the frictional effects of ramp traffic merging into the traffic on the main lines. This potential reduction in capacity is handled internally by the IDYNEV simulation model.
- e. Voluntary Public Evacuation As a result of recent evaluations by Dr. Mileti, updated representations of voluntary evacuations were considered. i Section B.3. of this testimony discusses voluntary public evacuation in detail. f
- f. Throuch Traffic Within the EEI A more accurate representation of through traffic (that is, external-external trips) along the Interstate Highways within the EPZ has been introduced into the input streams for the IDYNEV Model. A detailed discussion of this topic appears in Section 4.K. of this testimony.
I i
t
I l
I i
f ,
l' I
- g. Additional Analysis Recions I*
It was decided to add three keyhole configurations shich are defined by a central area, approximately 5 miles in radius, with separate quadrant sectors, extending to the EPZ boundary. Specifically, the following additional EPZ Regions )
i are considered: j 1
Recion Soatial Extant ERPA Classification 11 To EPZ B' dry. A, B, C, D, G, Outer North Region 12 To EPZ B' dry. A, B, C, D, F, Outer West Region 13 To EPZ B' dry. A-E Outer South Region See Pages 10-3 through 10-5 of Volume 6.
- h. UDdated ETE Studies As a result of the considerations noted under sections ,
a,d,e,f, and g above, additional ETE calculations were performed using the IDYNEV model, applied to the summer I weekend scenarios 1 and 2.
The results of these ETE studies are presented below.
All ETEs are referenced to the Order to Evacuate.
I l
1 : : : :
L 6 7 6 8 5 5 5 5 2 1 3 3 0 1 : : : :
6 7 6 8 5 5 5 5
1 2 3 4 0 1 : : : :
6 7 . 6 8 0 5 0 5 1 -
0 2
2 4
6 7 6 7 0 5 0 5 1 0 2
2 4
6 7 1 6 7
s s .
e e l 0 5 l 0 5 i 1 0 2 i 2 4 M .
- : M 8 : :
6 7 6 7 2 5 n 0 5 n 5 5 i 1 1 3 i 2 3 0 h : : h : :
t 6 7 t 6 8 i i w n n o w o i 0 5 i 1 5 5 m c 1 1 3 m a 3 0 o e : : o e : :
r R 6 7 r 6 8 f f R
) 0 0 1
) 5 0 s 1 1 2 s 3 4 e : : e : :
t 6 7 t 6 7 u u 4 .
n n i 0 0 i 5 5 M 2 1 2
M 3 4
- 6 7 : 3 6 7 r r u u o 5 5 o 5 5 H 1 2 3 H 4 0
( :
6
- (
2 : :
7 6 8 e e m m i 5 0 i 5 5 T 1 2 4 T 1 4 0 n 6 7 n 6 8 o o i i t t a a u u c c a o a o v i r i E r a E r a
n 1 2 n 1 2
. e . e 1 c 2 c S S
- H4 :
6 8 6 8
0 0 5 0 1 4 1 4 2 1 : : : :
6 8 6 8 0 0 5 5 n5 :
6 5
9 0
7 0
0 1
5 5 5 5 a 3 0
3 0
_ 6 8 6 8 5 5 5 5 s
1 3 0
2 3 0
e 6 8 Z 6 8 l P i E M 5 5 1 5 5 2 3 0 e 3 0 0 : : h : :
1 6 8 t 6 8 n n 2 i 0 0 i 0 5 h
t 1 4 1 h 4 1
- : t : :
i 6 8 i 6 8 y
w n w r n 1 m o 0 0 m a o 0 5 i d i o a 1 4 1 o n c 4 1 r e : : r u e : :
f 6 8 f 5 6 8 R o R
) ) B s 0 5 s 0 5 e 1 4 0 e 4 0 t : : t 1 : :
u 6 8 u 6 8 n n i i M 0 5 M 0 5 3 4 0 3 4 0 s 6 8 s 6 8 r r u u o 0 0 o 5 5 H 2 5 5 H 0 0
( : : ( 2 : :
6 9 7 0 e e 1 m m i 5 5 i 5 5 T 1 0 5
T 1 0 0 n 7 9 n 7 0 o o 1 i i t t a a u u c c o a o a i v i v r E r E a a t. 1 2 n 1 2 6
. e . c 3 c 4 S S
As noted on p. 10-20 and Figure 10-29 in Volume 6 of NHRERP, the population of Seabrook Beach was the last to clear from the EPZ. On the basis of updated estimates obtained from.the July 1987 photographs, the population of Hampton Beach will clear at a later time. As a result, the evacuation of the entire EPZ for the summer weekend scenario has increased approximately thirteen percent. The new.
regions 11, 12 and 13 exhibit ETE which in general parallel
-- and exceed somewhat those of Regions 2, 3 End 4, respectively. For these three regions, voluntary evacuation is specified as 50 percent of the population within the EPZ, but external to the region.
- i. Sensitivity Runs
- 1. Manning of Control Points.
Several sensitivity runs were executed, using vehicle populations which include the new peak estimates of beach area vehicles. Certain of these runs were conducted to quantify the results of different arrival times of personnel at control points. The results of these runs are described below:
a) Sensitivity Run #1 The purpose of this sensitivity run was to quantify the effect on ETE of the late arrival of some traffic guides at control points during the Summer Season. Based on l information provided by the New Hampshire State Police, it is expected that four State Police will report to the assigned 1
1.
control points within 15 minutes of the Beach closure. Three more will report to their respective control points within the following 45 minutes; and six additional police will arrive within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of Beach closure. This assumed manning schedule will have the following impacts on highway capacity relative to that used for the Planning Basis:
The Route 51 overpass of 195 will service only one lane of westbound flow for a period of one hour following the beach closure. Thereafter, two lanes will be established to service westbound evacuating flow.
The intersection of Routes 1 and 101C will not be manned until two hours after beach closing. It is I
assumed that evacuees will respond to existing signal control even in the presence of no competing traffic flow.
During the first two hours, evacuees from Hampton ,
l Beach will not be discouraged from travelling south over the Hampton Harbor Bridge into Seabrook.
Subsequently, all Hampton Beach evacuees will travel north and west, only.
b) Sensitivity Run #2 The purpose of this run was to examine the sensitivity of ETE with respect to any possible further delay in manning the Route 51 overpass of I-95. Specifically, it is assumed
1 1
I I
for this run that the State Police established the TCP there two hours after the beach closing.
c) Sensitivity Run #3 The purpose of this run was to examine the impact on ETE l l
if none of the capacity-enhancing TCPs in Massachusetts were i established.
I The intersections of Routes 1 and 286, and of Routes 1 and 1A, both in Salisbury, were assumed to service evacuating flow with the existing signal control policy. It is assumed that evacuees would not violate this policy even in the absence of competing traffic flow.
Only one ramp will service westbound evacuating traffic along Route 110, onto Southbound I-95.
d) Analysis The ETE for these three sensitivity tests are given below, followed by a brief discussion of these results:
ETE (Hours: Minutes) for Vehicles Within the Indicated Distances from Seabrook Station EMD 2-mile 5-mile 10-mile EEI
- 1 6:20 6:35 6:40 6:45
- 2 6:35 7:05 7:05 7:10
- 3 8:40 8:50 8:55 9:05 As noted in Sensit'vity i Run 1 above, evacuees from Hampton Beach will travel south across the bridge into I
c
/
1 Seabrook, to the extent that traffic congestion will permit, until'the TCP which routes traffic toward the north is established. It is assumed that up to 900 vehicles will travel south by the time TCP A-HB-01 is manned; thereafter all traffic moves north.
Since Hampton Beach traffic now (using the Avis data) constitutes the critical path under the Planning Basis, any movement south over the bridge could expedite evacuation.
(of course, those travelling south from Hampton Beach move somewhat " closer to Seabrook Station, which may be undesirable if a release has taken place.) The results of Run #1 confirm this expectation; the ETE of 7:05 corresponding to the Planning Basin is reduced somewhat to 6:45.
The ETE expands somewhat if the Hampton Interchange is manned after two hours instead of one hour. In this case, Run #2 indicates that the ETE lengthens by 25 minutes, relative to Run #1.
For Run #3, we take a pessimistic view of the consequences of no TCPs being manned in Massachusetts. In f prior runs of " uncontrolled" scenarios, we assumed full (or nearly full) utilization of intersection capacity. Here, we assume that evacuees will heed the signals even if there is no competing traffic flow, thus utilizing only a portion of available capacity.
Under these circumstances, the evacuation paths from Seabrook and Salisbury beaches become critical, replacing i
l l
L____ __ _
' u ' -A j , .
I ?
j)? ,h.
e {3l/ lr ,
a ,
_,, .V . <
p .?
'" $ ( ; ) y
, , 1 Hampton Beach in this respect, with a foncon,gitant incnass '
u,. 1 in'ETE. The results of Ru W3 r;how #
thut the ETE will _')
q' 'g .i increase by two hours relatitre to the Planning Basis for t 3 Region 1, Scenario 1, if the TCPs 'in Massachusetts rer.ain '
.)
unmanned throughout the evacuation. 402 h:curre, if these TCPs are established, eveniwith a delay, the ETEs will bcf{, r' 1
greatly reduced.- Thes>'ET2swillsapproact/tboseforthe Planning Basis if thii delay in establishing theue TCPs !s /
moderate). , . }
. \
- 2. ext 91LPd_JJ1adow Evacuation T11jg,_ESt}]ngt,e,jg
., .? ,) J \ t i Additional sensitivity tests were conducteo'to assess the impact onlETE,'if any, of anticipated voluntary
~
evacuation by persons outside the EPZ, up t'o a distance of 20 miles from Seabrook St$ 1on.' To conduct t$ esp runs, it was necessarytocreateadanalysisnetworkrhpresentingthe is highway system extending fiom'-just within V,hd EPZ to a distance of 20 miles or so from *2he Sthtion. j Estimates of
(
population during the summehwere ,oi:,tained from data obtained e s 6
l l
from State and Federal Agenciec.# i,. _,
h,.
The IDYNEV model was dxecuted as follows* '
.{ $,
Evacuation traffic' leaving th;p EPZ entry,ed this y network within the' regkon ohIdide the EPh, %le { ,
s rates at which the EPZ traffic enter this outer , ;
\'
l < i network were computed by/ %DYNEV during the Region l*
1, Scenario 1 Study. These rates of traf:f'ic' '
y ,
,\
entering the outer ne%o,rk am, specified as inputs i ,. ;,
s n
,f f h *' 1 1
,0
/ /
s '
1 i ;
, ,/ )
,, ;; i l1- /,
4
- _ - _ - _ - - i !
..g
- q
., ;i< -
JS ud' ir '. r oA !/,' i ' t; to IDYNEV for these sensitivity runs. Note that -
.tr i f, p this outer network overlaps the EPZ network at the I
!! latter's outer extent. !
Evacuation traffic originating within the outer '
y, network originates at centroids established for
'[ .
- j. that purpose, as was done for the EPZ network.
y -
The output was exanined to determine the ETE for
- 1. .
traffic evacuating the EPZ.
Two sensitivity runs were conducted, each assuming a g n uniform rate of voluntary evacuation within the area outside l 4
[L the EPZ. In the first, it is assumed that 10 percent of this population voluntarily evacuates; in the second, a 20 percent tigure is assumed. The results are presented below, together with a comparison with the Planning Basis which does not consider any external shadow effect.
ETE for Evacuees From Percent.of External Within EPZ Pooulation that Evacuates (Rec. 1. Scen. 1) e 0 (Planning Basis) 7:05 7 Q
10 7:05
?
20 7:05 As indicated, voluntary evacuation from the area outside the EPZ, j a at the indicated luvels, does not influence the ETE for those s evacuating from within the EPZ.
1 1 i 1
1; 4
i . 5 4' '
l r#' .1
)
.j. Effect of Slower Accident Escalation Rates This subject was discussed on pages 10-17 and'10-18 of Volume 6, l
-l NHRERP. These sensitivity runs were updated with the recent input resulting from the 1987 Aerial photos. The associated results follow: l l
Elapsed Time from ETE for Evacuation from within the l
.the Alert Level to ' Indicated Areas around Seabrook Station I the Order to Evacuat,g Referenced to the Order to Evacuate 2 mi. 5 mi. 10 mi. EPZ b' dry
^
0:25 (Plan Basis) 6:23 6:45 7:05 7:05 1:05 5:49 6:00 6:20 6:25 2:05 4:10 4:50 4:55 4:55 l
2:50 3:30 3:50 4:20 4:45 l The data, like those on page 10-18, also display the sensitivity of ETE with the rate of accident escalation.
- 4. Hiahway Canacity
- a. General As put forth on Page 3-1, Volume 6, NHRERP, the ability of the road network to accommodate demand is a major factor in determining how rapidly an evacuation can be completed. Highway capacity, in general, may be defined as the " maximum hourly rate at which persons or vehicles can reasonably be expected to traverse a point or uniform section of a lane or roadway during a given time period under prevailing roadway, traffic, and control conditions."
(Highway capacity Manual, p. 1-3.) Section 3, Volume 6, I NHRERP, presents a discupsion of some of the major factors which influence capacity. Factors which contribute to
capacity include 1) on the approach to intersections:
geometry, channelization of traffic, composition of traffic, ,
competing traffic streams, turning movements, type of control device'and signal timing; 2) along sections of roadway: )
I roadway geometrics, traffic composition and motorist 1
behavior; and 3) general considerations: weather and pavement conditions.
- b. Link capacities Appendix N, Volume 6, NURERP provides a detailed listing of estimated network link capacities and turn movements. It has been suggested that the data in Appendix N implies that 1500 cars can enter node number 1 from "each of 3 directions." No such implication is intended. The actual capacity of a link (as compared with its theoretical capacity) depends on the role it fulfills within the structure of the highway network. For example, the actual discharge rate through nods 1, aggregated over all apprcaches, as calculated by the IDYNEV simulator, ranges from approximately 800-900 vehicles per hour, depending on the scenario,
- c. Loadina Procedures It has been stated that loading procedures are not described in much detail within Volume 6. In fact, the trip generation rates are given in great detail in Appendix M.
For the representative summer scenarios and off-season scenarios considered and for all centroids, those data !
l
_--_-__m______ ------
indicate the variation of loading rates over time. These nominal loading rates are upper bounds which can be reduced ,
by congestion on'the evacuation network. The actual loading )
rates are computed internally by the IDYNEV simulation model l
and are not provided as nutput. l
- d. Traffic Directional Distribution In response to a contention regarding traffic k
directional distribution, it must be noted that the value of '
fd as indicated on p. 3-9 of Volume 6 and as is also stated in the HCM in Table 8-4, is 0.75, not 0.675 as has been alleged. It is also noted that the value of fd is independent of passing rate. Page 3-7, Volume 6, NHRERP defines the term fd as an adjustment factor for directional distribution of traffic, obtained from HCM, Table 8-4.
Volume 6, p. 3-9 assumes a directional split of traffic (averaged over the evacuation time frame) of 0.9 on all two-way road sections. That is, 90% of all traffic is outbound and 10% is inbound; the outbound traffic is evacuating while the inbound traffic is the " light" counterblow traffic. Columns A and B, Table 4-2 of Volume 6, demonstrate that beyond the first hour of the evacuation, very little counterblow traffic exists. Thus, the overall estimate of 90% directional split, and use of 0.75 as recommended by the HCM as the value ofd f , is supported by tnese statistics. (While the directional split of traffic
! over the first hour will be somewhat more balanced than I
E_____________
90:10, the outbound capacity will remain at about the same level, according to HCM procedures.) +
- e. Licht Traffic Patterns A contention has been raised which questions how light j traffic patterns which are indicated in Appendix I had been j treated in the simulation model. Appendix I, Volume 6, NHRERP, Rev. 2, provides detailed descriptions of traffic management and control tactics for traffic control posts, developed in cooperation with most Police Chiefs. Light traffic patterns included in Appendix I represent traffic which is moving in directions which are generally counterblow to those of evacuating traffic. Therefore, while the simulation model does not explicitly consider this inbound traffic (and, to our knowledge, neither do simulation studies conducted for any other ETE studies), the effects of the inbound flow on the ETE are included in the ETE calculations through the capacity calculation of two-lane highways. That is, the presence of inbound flow tends to reduce the capacity of the outbound highway lane, for 2-way roads.
- f. Weather Weather is one of the many major factors which influence l highway capacity. Each type of influencing weather condition is addressed separately within the ETE. Highway capacity reductions of 20 and 25 percent for rain and snow respectively were utilized in the ETE. These figures are responsive to the guidelines established by the 1985 Highway l
l l
I Capacity Manual. See p. 3-11, Volume 6, NHRERP.
Corresponding with a reduction in capacity, the calculations i
of ETE reveal that rain and snow increase ETE relative to )
l clear weather (compare the ETE for Scenario 2 with those of 1 Scenario 1, and the ETE of Scenarios 6 and 7 with those of Scenario 5).
{
- 1. Snow j l
It has been contended that the reduction in capacity used with Volume 6 to account for snow is inadequate.
Necessary to the understanding of capacity reduction due to snow is the correct understanding of the term
" water-equivalent snowfall". The ratio of " water-equivalent" to " depth of snow" is commonly termed " snow-density".
Average snow densities generally range from 0.05 to 0.3 with the higher figure applicable for late-winter snow (with a high moisture content). The average figure for freshly fallen snow is 0.1, which we will use. (See Appendix H of National Cooperative Highway Research Program (NCHRP) Report No. 127).
Applying this figure, a fall of 6 inches over 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> (as suggested by contention) is equivalent to 0.075 in./hr. I water equivalent snowfall. On this basis, the associated percent capacity reduction using the 2.8 percent model (again i
as cited by contention) is: )
l 8 + 0.075 x 2.8/0.01 = 29 percent.
l
{
\
t_____ - - - _ _ - - - . - J
l This estimate of capacity reduction for the indicated severe snowfall, which only applies to conditions at the end of the 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> snowfall, compares with the representative figure of 25 percent in the ETE study, which was applied throughout the 8-hour evacuation period.
Other studies offer evidence that capacity reduction due to snow is less pronounced. Using data from Appendix E of NCHRP Report 127, capacity is reduced by 25 percent whenever
" snowstorm speed factor" is about 0.55. This value corresponds to a condition of 1 inch per hour of snowfall over 5 or more hours.
Note that the EPZ is subject to an average of 12-16 inches of snowfall per winter month, as shown in Table 1-1 of Volume 6. Thus, the suggested example of 6 inches of snow within only eight hours given above represents a little less than half of a month's average snowfall. Even so, the 25 percent capacity reduction figure is applicable.
- 2. E22 Volume 6, page 3-11 addresses the issues of both ocean and inland fog as they relate to highway capacity.
Discussions held with public officials indicate that low-lying ocean fog which can affect travel conditions is an unusual occurrence during the summer months, and when it does appear, generally dissipates by 9:00 to 10:00 a.m. or may 1 appear after sunset, both times of the day when beach population is significantly below peak levels. Thus, the
conditions of Scenario 2 (sudden rain occurring with beach' population at capacity) are'more severe than a scenario with early morning or' late evening fog. Furthermore, the ETE
'obtained for the inclement weather scenario for rain can be used.for widespread foggy conditions.
-3. Floodina Consultation with-FEMA, Region 1, indicates that flooding is a hazard which may impact EPZ land areas in one of two manners: tidal flooding or riverine flooding.
Tidal flooding is generally associated with high winds which-raise the level of the tide along the coastal areas.
(Of course, the level of the tide still responds to the lunar cycle.) It was estimated by FEMA that for the 50-year storm (a storm of particular severity with a one in.50 chance of occurring in any one year), it was possible for parts of Salisbury beach to remain under water for as long as 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> at high tide, at a depth which would make vehicle passage impossible and with no alternative roads available.
Such an event, if synchronous with'the order to Evacuate, could delay some permanent residents up to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />.
(Note that such severe storms do not, in general, occur during the' tourist season.) Since the ETE substantially exceeds 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, this extension in trip generation time does l not necessarily imply an extension in ETE -- certainly not an ;
extension of as much as two hours, in any case. (It must )
I also be considered that foreknowledge of the approach of such I 1
l l
'a severe storm may well lead toothe beach area being evacuated prior to the accident.) ' Riverine flooding is associated with rainwater and/or melting snow run-off, and generally occurs in the. Spring.
, Riverine flooding could also occur during heavy rains produced by hurricanes in Autumn. (As tabulated on p. 1-9, Volume 6,'the maximum rainfall in the months of July or'
' August over a 30 year period was less than 7 inches.) The
. Town of Exeter is most vulnerable to such flooding in that one,;and possibly.two,: evacuation routes for Exeter evacuees could be severed. Specifically, Route 108 south of the village is a potential flood area,as well as Route 150 in -
northern Kensington. There are five other evacuation routes
.from Exeter to the north and west. Calculations indicate i that the ETE for residents of the Town of Exeter are less than those for residents in the towns in the coastal region; therefore, any incremental delays experienced in Exeter due to the loss of a flooded road should not extend the overall travel time for the residents in the EPZ beyond the ETE.
- 4. Igg The effect of ice storms on the ETE can vary widely i depending on the extent and physical cor.dition of ice on the pavement, and on the temperature. The friction factor offered by an icy surface varies with temperature, increasing as temperature decreases below freezing. In the temperature range between 28 and 32 degrees F, the heated tire surface F
[.
E can ride on a thin film of water and traction is at a minimum. Under these weather circumstances, travel would be most affected.
An exhaustive literature search has revealed no estimate of the effect of ice on highway capacity. In the absence of such data, application of the ETE for snow conditions (which includes a 25 percent reduction in capacity) appears acceptable for the following reasons:
Highway capacity during an ice storm may be less than.that during a snowfall, thus tending to increase travel time relative to snow. Note, however, that sanding operations would restore capacity of icy pavements to a significant extent.
In general, there is no need to shovel a driveway in an ice storm, as is assumed within the ETE to be required for a snowstorm, thus tending to' reduce trip generation time, relative to snow. A reduction in trip generation time tends to reduce ETE.
While we are aware of no data to quantify these opposing trade-offs, it is reasonable to expect that the net effect is limited and that applying the ETE for the snow scenario 7 or 10 (see page 10-2) is a proper response. Note, however, that under severe ice conditions, in the absence of sanding, some highway sections with extended upgrades may become virtually impassable. Thus, sanding may be necessary to assure adequate traction on such highway sections. Sanding with sand, salt, or a combination of the two, is a prevalent procedure in the area.
k
- g. Disabled Vehicles
.It has been suggested that cars could break down on bridges'and "other choke points" within the EPZ which could lead to one vehicle totally obstructing the' road.-
It must be pointed out that all bridges of significant
' length within'the EPZ are. designed so that there is some shoulder space available to store disabled vehicles. -Narrow, short bridges do exist on some two-lane roads. If a car
~
should stall on such a bridge (which is a low probability event), it can be pushed along the bridge onto the shoulder I immediately ahead. Furthermore, even if a vehicle is stalled at a point on a highway where there is little or no shoulder. ;
1 room, there is still sufficient room for the traffic to move around the disabled vehicle even if it means encroaching, somewhat, into the incoming lane of travel on a two-lane road. In situations like this, the dominant flow of evacuating traffic which would be in the outbound direction would effectively "take over" that " choke point" and traffic would move past the obstruction. The " minor" traffic flow in the inbound direction would take advantage of any gaps in the evacuation traffic. In any event, tow vehicles will be dispatched to remove impediments, as discussed later.
The suggestion that the Highway Capacity Manual (HCM) estimates'that capacity is reduced by 1/3 "because the roadway's perceived width is reduced" is simply incorrect.
The referenced passage on p. 6-10 in the HCM cites one study 1
I (by Goolsby, M.)- which describes an " incident removad to the shoulders" which reduced capacity by one-third on a three-lane (in one' direction) freeway. The Goolsby paper specifies that this incident was actually an accident and that capacity reduction was caused.by the " gapers-block" phenomenon. That is, the drivers along the. freeway slowed their cars to observe the activities associated with the processing of the vehicle involved in the accident.- There is no mention:of a perceived reduction in width. Goolsby then indicates that capacity returns to normal when the -
accident- investigation is complete.
Thus, a car parP.ed on a shoulder, with no associated activity, would have a negligible effect on~ traffic flow.
For example, the narrowing of a lane, due to maintenance or construction purposes, to 10 or 11 feet widths, provides a capacity per-lane of about 1800 vph on a highway section as described on page 6-13 of the HCM. This figure compares with the value of 1728 vph estimated within Volume 6 (see p.
3-10).
In'the'HCM chapter on Intersection capacity, the impact of a lane of parked, motionless cars can reduce capacity by up to 10 percent (Table 9-8 in HCM); a single parked car should produce a lesser effect.
The estimates of highway capacity contained in Volume 6 (see discussion in Section 3), take into account uncertainty in driver responses. The ETE has reasonable expectations that under emergency conditions, there could arise somewhat uncertain responses on the part of the evacuating public. For example, it must be anticipated that some vehicles will exhaust their fuel supply and will have to be pushed to a shoulder or driveway. Such short-term disruptions also serve to reduce capacity for short periods of time, and justify the conservative posture we have adopted. We believe that this posture is prudent and responsive to the intent of NUREG-0654.
As noted on p. 3-4, the calculation of ETE asserts that capacity of a highway section (or link) is reduced by 15 percent when congestion (i.e., Level of Service F) prevails there. The following appears in Chapter 6, Freeway Systems, of the Highway Capacity Manual (HCM):
In many cases, vehicles are unable to depart a standing queue at the normal capacity rate of 2,000 pephpl. In their studies of uninterrupted flow characteristics, Edie and others (30) have noted that the relationships among speed, density, and flow may be discontinuous at the point of capacity, and that the maximum rate of flow of vehicles departing a queue may be less than capacity under stable flow. Various observations of a freeway j queue departure rates range from as low as 1,500 pcphpl to as high as 2,000 pcphpl. Local driving ]
characteristics have a major influence on this effect, which ranges from a significant reduction in capacity (compared to 2,000 pcphpl) of up to 25 percent to cases in which there is virtually no reduction.
The above citation applies to an " uninterrupted traffic stream" on freeways. We have applied a capacity reduction on all roads. The justification for such reduction on i
non-freeway segments is based on the reasonable expectation i I
I
that driver uncertainty during an evacuation and short term disruptions can reduce capacity.
Since capacity reduction during congestion can range from zero to 25 percent, our adoption of a 15 percent reduction overall is reasonable.
1 In the event that vehicles are incapacitated on the j 1
highways, the NHCDA has entered into Letters of Agreement ,
I with multiple tow truck operators located, for the most part, along the periphery of the 10-mile EPZ. A number of these j tow truck operators are located within approximately 1-3 miles of the " pre-staged" locations suggested in Table 12-1, Volume 6. (See Attachment 2 hereto, " Planning Memorandum Re:
Tow Truck Assignments".) In lieu of tow truck " pre-staging,"
the NHCDA has elected to utilize routine, existing procedures for activating these vehicles, i.e., through the NH State Police Dispatcher, to facilitate mobilization of these services. Note also that once activated, tow vehicles i equipped with radio dispatch will enable the drivers to move from one incident directly to the next.
A series of sensitivity tests were undertaken to l
quantify the effect on ETE of highway impediments which extend over substantial periods of time. Four runs were i i
executed for each series; each run simulated the effect of 10 )
I accidents of varying durations occuring on different l l
randomly-selected high volume evacuation roadway links within I the EPZ. All cases assumed that these accidents impede evacuation traffic and considered the case study, Region 1, Scenario 1, as reported in the ETE (the evacuation of the entire EPZ). Results of these sensitivity tests, which were conducted prior to the incorporation of the August, 1987 l beach data, are shown below:
Range of Increase Series Scenario Duration of Imoedances in ETE over 4 runs 1 1 1 to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 0-10 minutes 2 5 1 to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 15-20 minutes 3 1 2 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 30-60 minutes 4 l
4 5 2 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 15-40 minutes (
Ten accidents were simulated in each run since this number represents the anticipated number of accidents from defined statistics, based on the vehicle miles of travel I I
expended during the evacuation. These data indicate that if j impedances are removed by tow vehicles in a timely manner, one can expect a small effect on ETE. If the responses are less timely, then more pronounced extension of ETE will occur. Note that the ETE is not extended the same amount as the longest impediment since traffic is free to redistribute on alternative outbound routes, if one route in the area is experiencing some loss of capacity.
Even in the event that tow vehicles are not immediately available, the public is not helpless in emergencies as would 1
be suggested by the thesis underlying " myths" about emergency L _ ___ _ --_ _- _
behavior. In fact, the public is an actual resource in emergencies. .The public often gives rise.to " emergent work.
~
groups" in emergencies which form temporarily and on the spot to do what needs to be done. A detailed 1
discussion of public behavior in emergencies is found at
'Section B . 1. ,d " Evacuation. Impediments," of this direct )
i testimony. J l
- h. Tocoarachical features i l
The evacuation time estimate also accounts for
-topographical features. As indicated by the discussion on page 3-8 all two-lane roads within the EPZ are classified as
" rolling terrain." Note that the vast majority of congested highways are east _of I-95. As a result, the estimates of i capacity for these roads are roughly 10 percent lower than if level terrain: were represented. The terrain in this area is relatively flat with elevations rarely exceeding 100 feet above. sea level.
- i. Road Tvoes It is suggested that the roads were classified into "only 4 crude groups" and that the detailed data collected should confirm that the minimum width of each roadway section should be greater or equal to those values which are defined in these groups. The measurements that were taken in the field were-taken at representative sections. It is neither practical nor feasible to take width measurements at closely spaced intervals along the highway nor is it necessary. The '
- 64 -
L 1
i discussion on p. 8-8 of the Highway Capacity Manual does not stipulate that such detailed measurements need to be taken and furthermore indicates that "the impact of narrow lanes restricted shoulder widths is less deleterious if vehicles are already traveling at reduced speeds which prevail under
. capacity operations." The classifications used in the ETE are consistent with those used in Table 8-5 of the Highway capacity Manual.
The classification of roads within the EPZ was accomplished by way of capacity estimations based on physical surveys. The examples of the four general types of rural roads found in page 3-7, Volume 6, were intended to give a.
" general impression of road type within each category; it was not-intended for example, to-imply that the entire length of Route 1A is classified as a " medium" design road. Route 1A, north of node 22 in the Town of Rye, where the road is winding and exhibits sharp turns, is classified as a " low" design road for the purpose of the ETE (see Figure 1-3, p.
N-1).
- j. Traffic Control Measures ]
The ETE assumes as one planning basis that certain 1
traffic control measures are in effect. This assumption is based on the high probability that there will be sufficient time between official notification to man Traffic Control Points (TCPs) and the public Order to Evacuate to allow for i
I
t L
traffic control personnel'to be mobilized and positioned.
- The NHRERP provides for the notification to emergency l
.i response groups and to supplemental State resources under the- l Alert classification,-providing ample lead time for resource notification and dispatch prior to the recommendation of
. protective actions.
The. staffing of traffic control positions associated with an offsite response to an accident-at Seabrook Station are outlined in the ETE and Traffic Management Manual (Draft). The State of New Hampshire has sufficient personnel resources to-perform its assigned responsibilities of. fully staffing all evacuation-related traffic management pointe.
See Applicants' Exhibit-1 to the testimony of the Personnel Resources Panel.
Current State procedures call for NH State Police Troop A to man access control points with back-up personnel available from Troops from other points of the State; additional assistance is available through other State resources. The National Guard, for example, is available to supplement access control 2r traffic control points within the communities. The extensive manpower of the National i
-i Guard could essentially provide the bulk of second shift personnel requirements in this area. Also, once local jurisdictions are evacuated these local police may be asked to assist with perimeter access control.
l
I l
l In addition, each traffic control point (TCP) has been assigned a priority, while priority 1 TCPs have also been i
assigned a sequence in which they are to be manned. The most important TCPs are considered to be those which have the most )
i potential for expediting the movement of traffic. Those TCPs which are assigned lower priority are considered less ;
important, although they are helpful in expediting i traffic movements and in reassuring the public that the .
evacuation is under control. Even if lower priority TCPs were not manned, the effect on ETE overall would be negligible, under the assumption that there be no significant inflow of traffic into the EPZ from outlying areas.
Sensitivity runs have been performed to quantify the effects on ETE of most traffic control measures not being in effect. To understand the rationale underlying the sensitivity runs relating ETE to the manning of TCPs, however, it is necessary to delineate the various functions of traffic control. These functions, in decreasing importance, include:
Enhance roadway capacity; Guide evacuees along recommeded routes; Expedite traffic movement; Provide assurance to the public; Surveillance.
In general, a TCP performs more than one of these <
functions. Some TCPs, however, are designed primarily to l
l q
'I perform one of the two most important functions. The first I l
function, enhancing roadway capacity, can have a pronounced ' '
Linfluence on ETE. Such tactics are applied sparingly during an evacuation since they can be resource intensive. These tactics, however, can be effective in reducing ETE when effectively and_ strategically applied. The most recent l analysis _ performed with respect to this matter reveals that f
three TCP locations within the New Hampshire portion of the EPZ have been designed primarily'as capacity enhancing:
A-HB-03,'A-HB-04 and D-MA-02. These TCPs will serve to expedite the traffic movement from within the town of Hampton, including Hampton Beach.
The public information program and the EBS messages encourage evacueas to travel along the recommended routes in directions away from Seabrook Station. Nevertheless, some TCPs in New Hampshire are designed expressly to restrict travel in directions which are commonly used during normal times in order to expedite movement along the recommended routes. of these, TCP A-HB-05 and A-HB-01 serve to direct traffic on Hampton Beach northward, away from Seabrook Station, and toward the recommended evacuation routes. There H are, of course, other TCPs which fulfill this function; the ones specified here reflect the need for State police under the conjecture that Hampton town police will not participate in the evacuation. This function, to a lesser extent than i capacity enhancement, could influence evacuation time.
l
By definition, all police personnel will act to expedite traffic movement so-that.the available roadway capacity willL j l
be fully utilized.: .The ETE, however, did not depend on full utilization of available capacity; as noted previously, j i
. capacity was reduced by 15 percent under congested conditions l
to represent driver uncertainty under emergency conditions. '
Thus,-no capacity. adjustment was made for the TCP i
" expediting" function in the ETE calculations. Thus, with respect to-this function, and to the " provide assurance" J function,'there would be no change in ETE.due to the absence l l
of guides'at most TCPs. i
'The surveillance function is discussed on pp. 12-1 and-12-3, Volume 6.
In summary, then, it is seen that the higher priority TCPs are those which enhance capacity. Discussions with representatives from the New Hampshire State Police indicate that the manning of State-assigned TCPs will be a function of trooper availability and will occur in a consecutive fashion as statewide troopers are called to respond. Estimated time frames for manning these posts range from an approximate 15 minutes for the first 4 troopers to a span of 3 to 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> for a complement of 100 troopers.- See Section A.3.1 to this testimony for a detailed description of sensitivity runs conducted to determine the effect of this partial non-compliance with traffic control tactics. It has been suggested that emergency workers, traffic control per,sonnel i
1
_ _ _-___=_-_. -
l u
included, may. abandon their assigned emergency duties.
Section B.3 to this testimony, " Emergency Worker Role Abandonment," provides an examination of the lack of role abandonment in historical record.
- k. Throuah vehicles NHRERP, Volume 6, p. 2-27, estimates "through" vehicles i to number 3,000 additional cars. These vehicles are assumed to be traveling through the EPZ (external-external trips) and are assumed to be already on the highways at the time of the accident. Page 10-3-notes that'these vehicles are not otherwise counted. Clarification is hereby provided concerning the distinction between " peak hourly flow", and the estimated number of through vehicles on the network at the time of the order to evacuate. (Note that this estimate is not "at the time of notification of an emergency.")
Peak hourly flow is the maximum number of vehicles that can reasonably be expected to traverse a point (or uniform section of) a lane or roadway, over an hour, under prevailing roadway traffic and control conditions, and has the units vehicles per hour. On the other hand, the estimate of 3000 vehicles as used within the ETE is the total number of through vehicles on the highway system at a specified point in time, and has the units vehicles. There is no relation between " peak hourly flow", or for that matter, " peak daily flow", and this estimate of through vehicles at a certain point in time.
l j l
I I.
h_______m_ __ __. _
The primary routes servicing-external-external trips through the Seabrook Station EPZ are the Interstate Routes 95 and 495. -All other routes within the EPZ, which could be
. interpreted as through routes (e.g., Route 1, 107, 108), are two-lane roads which are not relatively attractive to through travelers. About 230 lane-miles are provided by these express routes:
i - I-495: 2 miles 9 4 lanes; 5 miles 9 6 lants
- I-95: 21 miles 9 8 lanes; 4 miles 9 6 lanes. j The calculations of ETE for the summer scenarios assume that the accident takes place when the beaches are at maximum usage, at 2:00 p.m. When developing the inputs to the IDYNEV model, it was estimated that about 3000 vehicles, not otherwise counted, would be on the network at this time.
This estimate was based on observations made while traveling the network.
'Specifically, it appeared that traffic on I-95 and on I-495 was traveling at Levels of Service (LOS) that did not exceed LOS B or C. The associated range of density is 13-30 passenger cars per mile per lane. Thus, the total number of j y
vehicles on these highways is between 2990 and 6900, many of l which are not through vehicles.
(The concept of " Levels of Service" is a qualitative measure which describes operational conditions within a traffic stream, and generally implies perceived condition in l such terms as speed, freedom to maneuver, comfort, I
convenience, and safety. Six levels of service are given letter designations, with LOS A representing the best operating conditions, and LOS F the worst. More detailed definitions are found at pages E-2 and E-3, Volume 6, and on pages 1-3, 1-4 of the 1985 HCM.)
l The NHRERP calls for access control points to be established at the periphery of the EPZ which will discourage traffic from entering the EPZ from points outside except, of course, those vehicles which will participate in the evacuation (Vol. 1, 2.6-16).
This estimate of 3000 through vehicles represents those who entered the EPZ prior to the implementation of access control and have not as yet completed their travel through the EPZ by the time the Order to Evacuate (OTE) is given. At the OTE, the number of through vehicles within the network could be substantially less than the number which occupy the network at the time the access controls are applied. This reduction in vehicles reflects the fact that many of these vehicles will have exited the EPZ between the time the access control was applied and the time the OTE is announced. Thus, it is seen that the number of external-external vehicles on the EPZ network at the OTE sNould not exceed the 3000 estimate, and may be substantially less.
Subsequent to the calculation of the ETE study, direct observations of traffic flow on these major routes provided by aerial films taken on July 4 and 5, 1986 at approximately
2:00 p.m. indicated that traffic volume is very low - Los A and B. The associated range of traffic density is less than 20 passenger cars per mile in each lane.
Thus, based on these later observations, the total number of vehicles on these highways at this time is less than or equal to approximately 4600 passenger cars (20 x 230). Many of these cars, of course, belong to EPZ residents, tourists and employees who have already been counted. It is seen, therefore, that the estimate of 3,000 through vehicles -- not already counted -- is realistic.
Traffic on these highways during the off-season generally does not exceed the volume associated with LOS C, regardless of time of day. Thus, the estimate of 3,000 through vehicles is applicable throughout the year.
The 3000 vehicles estimated as through travelers are included in the ETE calculation. For example, the computer analysis for Region 1, Scenario 1 indicated a total of 99,398 trips. This figure compares with the total number of 96,524 in Figure 10-11c.
- 1. Licht Traffic on Hampton Beach The ETE estimates 300 cars on the highway in Hampton l
Beach from vehicle counts of aerial films, Volume 6, p.
10-16. Page 10-16 indicates the source of this count when referencing Appendix E, Item 7, which states that aerial photos used to determine estimates of parking capacity and Vehicle counts were taken on Sunday, August 11, 1985. Page !
- f. i i
10-16 stater the following: "at the time of maximum parking ccupancy (about 2 p.m.), however, moving vehicles are few in number relative to those parked. For example: a total of 300 moving cars were counted in Hampton Beach from the aerial films..."
It must be emphasized that few, if any, of these 300 cars are "through" vehicles. It is highly unlikely that any vehicles moving through the area would leave an Interstate Highway such as I-95 to travel along the coastal roads which ,
1 would significantly increase their travel time. Thus, these 300 vehicles are most likely either in the process of leaving or are destined for parking locations along the coast. Since our estimate of vehicle population along the coast is based on reasonably available parking capacity, these vehicles have already been included in that count. This approach is in consonance with NUREG-0654, p. 4-2 which cautions that " Care should be taken to avoid double counting."
- m. Evacuation Canacity Der Hour: Tables 10-10 Volume 6, p. 10-70 includes a description of how capacity values as listed in Table 10-10 were computed.
These tables, which include Column 6 of Tables 10-10,
" Evacuation Capacity per Hour", were developed in the format recommended in Appendix 4 of NUREG-0654. The entries are largely of general interest only, and are not used in the calculation of the ETE. Specifically, these entries of l
l capacity in Tables 10-10 indicate the potential aggregate t
]
i l
l l
i
) -____________________J
e
,capaelties of all outward-bound highways which are intersected by arcs of 2, 5, and 10-mile radii, respectively, centered at Seabrook Station. This potential capacity can
, only be partially exploited by evacuating traffic, since ,
network topological constraints limit the access of traffic to some of these routes. It is primarily for this reason that these entries are of limited interest. (Note that the statement that " estimates of available capacity may overstate the actual accessible capacity" refers only to the entries in Column 6 of Tables 10-10.)
, in contrast, the calculations of ETE are based upon a rigorous analysis of traffic demand interacting with the estimate.d link specific capacity constraints, using the IDYNEV model developed for the Federal Emergency Management Agency. These calculations do not utilize the aggregate potential capacity values entered in Tables 10-10.
- 5. ETE Preparation / Trio Generation Times
- a. Notification Time Appendix 3, ITUREG-0654 provides, as the initial minimum acceptable design objective for coverage by the public notification system the capability for providing an alert signal "throughout the 10-mile EPZ within 15 minutes." This .
I design criteria provides basis for the 90 percent plume EPZ '
population notification within 15 minutes. i 6-
m, tc j
.% =
.. , /t .. ;
y t
,. j y 1 N -/' t i .
' 1 j-N
?;.::y L A
n ( v.' s q P 'q t
/ / }-.f_
- b. 'Rgjach to vehigle i Time' Estimate j
i .' , M ' If
<< t r i Page.4'-12; Volume 6, NHNERP, estimates that e21
, vf ' . y [4 ,
',i e
day-trippers in the bea ::h ari,th ware wi$hin 3') minutes of '
f-
/
( v 1
'{
their cars. imate of 2 d/4 mph Asserting the rossipable (* e.%p ,
3 walking speed implies 'a walkfys pistance in excess off or.e .
g
/p , * $, '?> '- '
mile for beach-goers. Since tae width of the' beach patking I
q s.*
areas are under one-half mile, %is vin!. mate of 30 minut,es is clearly reasonable. Volume 6, p. 4-lL reads in partina j '-
J
,~
part,-"While we have no empirical data to,.h support this3' si s distribution, we do know the physical domain of the beach ) '#
L , ye a) 1$
area and the activities involved." .Fugther, this section f f ., [
the ETE reads that "since we kr:.Uwm3that corgestion will o,; cur. , ?,/
>3 : '
. on the beach areas,during.the sum $ r'an)d thtt dvacuationl t,im,, e -
g ,
{
will exceed Trip Generation time, any inaccuracies in - #q
- q 0 "I'
- e. \' n i f d' distribution will not influence the ETE. Thus, an h' k' N >
- ]
4 -
>,> a approximate, re'asionabledistribution"vdllsatisfyourr(dedh.+
~
1 ,
! I
- c. Itansoortation Stagina Area Peg),paration ,Tiggg 0 A delay of 15 minutes preparation!tinie at ti.vf
)
f
(
Transportation staging area was assumed for the movement of I transit. vehicles in and out of the areas. (Volume 6, p. .M
,' i 11-20.) The ETE allows for,a one hour preparation time for s 5
/ ,
thetransportationstagingjareastobereadytore'ciive o-4 vehicles. !
1
- d. Commuter Retur1VEEeroen'cv Vehicle Entrance Time s I
J Volume 6, p. 4-9,.providesgif );
iatribution entitledy' I r.
" Distribution No. 2, Time to' fr,epare to Leave Work. " This e >
t, ' - 9 *
.. s j
5 (
I i !
- .1
r
)
1 l
distribution has been suggested to indicate a percentage-of workers returning home following notice of a " radiological emergency." .The horizontal axis of Figure 11-1 on p. 11-17 entitled " Time Distribution of Arrival Home" is improperly t i t labeled. Where it now reads " Elapsed Time From Order to 4- ~
?j. . i 3 Evacuate," it should read: " Elapsed Time from Start of notification" (Table 4-2 on p. 4-16 is referenced from the y,
start of notification. Figure 11-1 is a plot of Distribution B from Table 4-2.) Figure 11-1 and Table 4-2, Distribution B, indicate that about 88% of commuters would arrive home i
within one hour of the svacuation order (which is assumed to 6
occur 25 minutes after the start of notification). That is,
'the 88% figure corresponds to 1:25 after the start of
})>
'g' ,
notification. (Because of the improper labeling of Figure i
- 11-1, the value of 70%, which corresponds to one hour after l j.
the start of notification, was previously inadvertently l
attributed to one hour after the order to evacuate.)
l e }
s' L The rate of returning vehicles is about one-third the j ; figure cited by contention, i.e., 95% of workers within 30 6-minutos. Commuters returning home against the flow of
,jh' I
f evacuating traffic will experience little traffic impedence
! t
- j. ii on the major highways since it is reasonable to expect that l
most, if not all, unnecessary trips toward the EPZ will be deferred or cancelled. Some commuters will encounter delays l s w'ithin the EPZ which will extend, somewhat, their trip 7l; f( ci /
generation time. However, the net effect on the ETE would be e
i i .h >
4j f G . ': ' ' l
l minimal, since the ETE for most scenarios are insensitive to moderate changes in trip generation times.
Pages 11-19 and 11-20 of Volume 6 of the NHRERP provide a detailed description of inbound travel time. This section indicates that.little impedence to vehicles in general entering the EPZ following an order to evacuate would be anticipated due to the following reasons:
- 1. The first buses will not depart their respective points of origin until approximately 30 minutes following the evacuation order;
- 2. Those same 88% of returning commuters as noted above will be reaching their homes within the EPZ one hour following the evacuation order, and therefo?;e, will not be contributing to inbound traffic flow over the time frame that authorized incoming vehicles (for example, buses, ambulances, police vehicles) would be traveling to the EPZ;
- 3. Inbound traffic would b'e limited due to discouraged entry via EBS and other media messages, and by people's concern over the potential risk to their health and safety; and 4, Inbound traffic would also be limited by access control points established at the perimeter of the evacuated area which would discourage non-essential inbound traffic.
The use of estimated return times for commuters under
" normal circumstances" represents peak hour conditions when other commuters are occupying the road'.iays over the same time
frame as those who are returning to their homes within the EPZ. The very existence of a peak period of traffic during normal circumstances in the late afternoon reflects the fact that the majority of workers leave their places of work and enter the highway system within a narrow time frame. If an order to evacuate occurs prior to the normal p.m. peak period then those returning home to the EPZ will be occupying the roadways while workers who live elsewhere remain on their jobs. Consequently, under those circumstances, the trip home should take somewhat less time than during the " normal" peak traffic period. Furthermore, it is highly unlikely that other travelers, who might normally be traveling toward the EPZ, would actually make their trips under emergency conditions. This reduction in other trip-making also tends to shorten work-to-home travel time relative to the normal afternoon peak period. Travel against the direction of evacuating traffic will not have an effect on major highways such as I-95 and I-495 since the two directions of traffic are physically separated and traffic in one direction has no impact on the other. On surface roads inbound traffic could encounter frictional effects but these would be counterbalanced to some extent by the anticipated lighter inbound traffic volumes due to the factors discussed in the preceding paragraphs. Traffic cones at TCPs would not physically block desired turning movements in any direction, thereby impeding counter flow traffic. As noted on page 7-1;
)
I I
"There are always legitimate reasons for a driver to prefer a direction other than that indicated. For example: He/she msy be traveling home from work or from another location, to join other-family members preliminary to evacuating." Later, the plan states, "The implementation of a plan must provide room for the application of sound judgment. The traffic cones and barriers are deployed as indicated in the sketches L of Appendix I, so that there remains room for vehicles to maneuever through these guides. That is, cones and barriers will not physically obstruct passage." Thus, it is seen that the plan is designed to accommodate commuters who will be returning home and incoming emergency. vehicles during the initial stages of an evacuation. For these reasons, it is reasonable to anticipate that incoming vehicles, including buses traveling along at-grade primary highways (e.g., Route
- 1) may average 40 mph and those that travel along access-controlled highways may average 50 mph.
(It has been argued that buses will travel at high speeds, and that an accident at these speeds could disable a car. The higher speed vehicles, however, are traveling counterblow to the evacuating traffic and are on Interstate highways, or the major highways traveling toward the EPZ.
These major highways are designed to safely accomodate much higher speeds than the average speed cited above. The majority of vehicles on the road within the EPZ are evacuating from the EPZ at low speed.)
- e. Bus Mobilization Times A telephone survey was conducted in early 1987 by the NHCDA in order to verify the mobilization times for bus l
l company vehicles under letter of agreement. The mobilization times requested were based on unscheduled activation of the buses. The times included how long the bus companies estimated it would take to notify drivers, have drivers go to their buses, fuel the buses and be ready to drive as well as the travel times to the state staging areas. This survey indicated that approximately 25 to 30% of needed vehicles could report to State staging areas within one hour, and that the remaining needed buses could report within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of notification. (See " Estimate of Bus Mobilization and Travel Times for Bus Companies to State Transportation to Staging Areas".) These data compare well with the results of the mobilization survey conducted in preparation of the ETE, which lead to an estimated 80 percent mobilization within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of notification (p. 11-18, 19, Volume 6). This small difference will not impact the ETE when the ETE exceeds transit times since all evacuating buses will merge with the evacuating traffic. Some small beneficial effect on ETE associated with this later survey data may be experienced only for some off-season scenarios where evacuation time for evacuating private vehicles is somewhat less than that for buses.
l It has been contended that individual drivers of Teamsters Local No. 633 "have in no way shown their willingness to drive into the EPZ." However, the Teamsters Local No. 633 represents only a small portion of drivers within the overallLdriving pool, as discussed in the direct testimony of the Special Needs/ Transportation Panel. I
- f. Bus Loadina Times The ETE has estimated the time for loading a bus with members of the population from special facilities using a 15 second mean boarding headway. In Table 12-33 in the Highway capacity Manual, empirical data has led to the estimate of boarding time per passenger equal to 2.6 seconds which includes the payment of a fare with a single coin. 'Other figures are 3.0 seconds for exact fare and 3.5 seconds for passengers who are standees on a bus. It is seen that the ETE estimate for boarding people from special facilities onto buses is approximately 5 times that documented in the Highway Capacity Manual. Also, the figure of 15 seconds is reasonable as an average boarding time, balancing shorter boarding times for ambulatory persons with potentially longer boarding times required by non-ambulatory persons.
While the total time to load passengers, which consists of travel from the local transportation center to the special facility plus the time for passengers to board the bus, was estimated at 40 minutes (0.67 hour7.75463e-4 days <br />0.0186 hours <br />1.107804e-4 weeks <br />2.54935e-5 months <br />) in Progress Report No. 7,
l.
the current estimate is 45 minutes (0.75 hour8.680556e-4 days <br />0.0208 hours <br />1.240079e-4 weeks <br />2.85375e-5 months <br />), as documented on page 11-21.
The number of non-ambulatory persons outside of special facilities who are in need of special transportation assistance has been determined by way of the NHCDA special needs survey, and has been included in the local RERPs. (See testimony of the Special Needs Transportation Panel.)
- g. Translt Dependent ETE Pages 11-16 through 11-27 discuss evacuation time estimates for transit dependent and special needs' persons. A table is provided on page 11-26 which estimates the " total elapsed time, at worst, from notification to the arrival of an EMS vehicle within the EPZ" to be 3.30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />. In calculating ETE for transit dependent persons it is noted that in-coming emergency vehicles can be expected to enter the EPZ prior to the complete evacuation of an area; as such, these vehicles after picking up evacuees will become an element of the same outbound traffic stream and will not be distinguished from the other evacuating traffic.
- h. Imoact to ETE of "Immediate" General Emeroency.
Argument has been presented that an immediate general emergency would extend the ETE by more than 20 or 30 minutes.
However, analysis of this type of scenario demonstrates no support for this argument. For example, a reduction in the elapsed time of 25 minutes (assumed as the Planning Basis) between the Alert stage and the Order to Evacuate, could i
occur under the presumed scenario of an immediate general emergency. If this 25 minutes were reduced to zero, it could increase the ETE by no more than that amount of 25 minutes.
In this scenario, the Order to Evacuate, which is the starting point which defines the ETE, is advanced by 25 minutes. Since the highway system in the beach areas and their service roads are operating at capacity very quickly after the Alert stage due to precautionary beach closings (for the summer scenarios), it therefore follows the total elapsed time to clear the area from the start of the alert is relatively insensitive to the extent of the elapsed time between the Alert and the Order to Evacuate. Thus, it is reasonable to expect that the ETE would increase by the amount of time that the starting (i.e., the Order to Evacuate) point is moved forward (namely, i.e., by 25 minutes). Sensitivity studies have confirmed this expectation. See Volume 6, pages 10-16 and 10-17 for a detailed dj.scussion of the results of sensitivity tests for an "immediate" General Emergency compared to the results of the Planning Basis.
- i. Telephone Survey of EPZ Residents A portion of the information used within the ETE in support of trip generation times was gathered by way of a telephone survey of residents of the Seabrook EPZ. The First Market Research Corporation of Boston conducted the study q based on a survey instrument drafted by KLD and subsequently l
1
refined by First Market Research Corp. This organization has over,10 years experience in conducting this type of telephone survey; their interviewers are carefully selected and supervised during the implementation of the survey. Each survey interviewer was instructed to adhere to the final form of the survey instrument as developed by First Market Research. The sampling procedure used automatic, random telephone dialing equipment. The number of calls is documented below.
A total of 10,587 random dialings were made using the random dialing equipment. Most of these dialings generated numbers which were not assigned to any telephone. A total of 3,582 connections were made of which 562 were ineligible since the households were located outside the EPZ. Of the remaining 3,020 calls:
- 769 respondents refused to participate,
- 51 calls were terminated before the survey was completed,
- 818 calls were answered by persons other than head of household (mostly children), and
- 1,382 interviews were completed.
Thus, only 2,202 were connected to adults in households within the EPZ, of which 1,382 calls, however, led to completed interviews, representing a 63 percent completion rate. This sample of 1,382 responses is extremely robust for this population size. For example, a somewhat similar survey
1 was conducted throughout New York State in 1983. This random sample telephone survey interviewed 1,503 State residents out of a total population of approximately 17.8 million, compared with our'1,300 responses for a resident population of under 150,000. (Ref.: Byunso, J. and Hartgen, D., "An Update on Household-Reported Trip-Generation Rates" in Transportation Research Record 987, 1984.)
In assessing the validity of a survey, there are two forms of errors: sampling and response. Sampling errors arise from the fact that only a portion of the underlying population is being surveyed. Response errors may be due to j many factors including inaccuracies or outright guesses on the part of the respondant, ambiguous questions, improper interviewing protocol, or deliberate falsehood by the respondents. In either case, error may be random with (approximate) zero mean, or systematic.
Bias, or systematic error can result from one of two sources: sampling errors or response errors.
The former deals with " external validity" -- the ability i I
to generalize from the sample to the underlying population. I
{
The latter deals with " internal validity", which means !
I getting unbiased answers to the questions. If either or both forms of this bias were systematic then the survey results could not be generalized to the population.
Surveys must be designed to consider all forms of error with a view toward minimizing the total error. In 1
____________________a
l particular, the systematic error is of far greater concern since it is'this bias which can compromise'the efficacy of a survey.
In designing the survey, it was decided to inform.the ]
respondent of the purpose of the survey in the interest of candor. The introductory statement concluded with "the information contained will be used . . . in connection with i
preparedness plans for Seabrook Station."
's one factor that was considered was the prospect that the
-strong feelings on this subject by some members of the sample .!
could provoke responses which were deliberately-designed to bias the survey results. Specifically, we were concerned of an organized effort by opponents of Seabrook Station to-orchestrate responses which furthered their point of view. I (Mr. Lieberman was informed by an assistant Mass. A.G.
attorney that a few respondents did provide false information; thus, our concern was justified). We decided that the best way to contend with such a possibility was to complete the survey as rapidly as possible. This decision precluded call-backs to non-responding phone calls. There is a potential for bias (which is indigenous to all surveys) associated with nonrespondants. The extent of any such bias depends on the context of the survey, the sampling procedure and the sampling frame.
In assessing this potential for bias, it is instructive to examine the available statistics:
I
i of the unanswered random dialings: )
i Some 16 percent (562/3582) were outside the EPZ l About 26 percent of all phone lines within the EPZ ,
are leased to business firms, according to a '
response to an inquiry from New England Telephone.
Since,the survey telephone calls were made in the j evening (on a weekday)'and on the weekend, in the off-season, it is certainly reasonable to expect that a higher percentage (than 26) of the unanswered calls were business lines.
Business personnel are excluded from the sampling frame; such unanswered calls are of no concern. It is therefore reasonable to expect that more than half of all unanswered calls were either outside the EPZ or were business firms, and thus did not contribute to the potential for bias.
Another approach to assess whether the potential for bias is actually manifested as real, significant bias, is to compare the data obtained by the survey, with the same, or similar, data obtained from other sources. If there is essential agreement between elements of the survey data set and those of other sources, then the prospect of systematic bias of significant extent can be dismissed.
- 1. Mean Household Size According to the N.H. State Planning Office, the population of Rockingham County in 1985 was 220,466 persons in 77,629 households, in 1985. These data yield a mean of 2.84 persons per household. According to our survey, the mean persons per household in the New Hampshire portion of f
l i
__ _____________j
l l
1 the EPZ was 2.85, in very close agreement with the county !
data. This close agreement implies that the household
! demographics of the nonrespondents are no different, j statistically, than those sampled.
- 2. Nean Travel Time It is of. interest to compare the mean travel times !
obtained from the survey conducted in 1985, with those obtained in the 1980 census (see page H-2, Vol. 6 of the NHRERP). As noted on page 5-1 and documented in Table 5-1, employment within the EPZ has grown significantly in this time frame. Consequently, with increased employee traffic volume, one must expect increased travel times to work. The survey data confirms this expectation; all towns display a consistent trend of increased travel time relative to that in 1980, as shown below:
Mean Travel Times to Work (min.)
1980 1985 Community (Pace H-2, Vol. 6) Survey Amesbury 20.6 24.1 Merrimac 22.5 25.1 Newbury 25.6 28.3 ,
Newburyport 22.4 26.7 '
Salisbury 20.7 19.5 W. Newbury 26.8 35.3 Exeter 20.2 23.2 Hampton 22.5 26.0 Kingston 24.5 30.9 Portsmouth 14.5 16.5 Rye 21.0 22.0 The consistent relationship between these two columns of data, in the context of increasing population and traffic l
density over the intervening time, again suggests that the survey data replicates the travel experience of the underlying population.
- j. . Snow Removal Times The discussion on page 4-20 of Vol. 6 of.the NHRERP addressed the need to estimate the time associated with the
~
need for evacuees to clear their driveways of snow to the extent necessary for their respective vehicles to' gain access to the street system. It is assumed, as a conservative approach, that snow has already fallen and will continue to full after the Order to Evacuate (OTE).
Consistent with current practice in the area, plowing of the public street system and of driveways begins as soon as sufficient accumulation makes the effort productive, and continues, as needed, to maintain passage. This planning basis assumes that the roads will remain passable throughout the evacuation time frame subsequent to the OTE, but at lower levels of free-flow speed and capacity. This assumption is consistent with the protective action decision criteria of NHRERP, which consider off-site constraints such as roadways not being passable due to snow. (Nowhere is it assumed that equipment will continue to plow the roads or driveways after the OTE).
Subsequent to the OTE, it is reasonable to assume that driveways will have to be cleared. The estimate of this clearance time, given in the distribution on page 4-20, is I
based on direct observation by.Mr. Lieberman and on discussions with persons living in the EPZ. It is our view that these estimates are reasonable and properly account for the conditions described above.
- 6. Host Locations It has been suggested that the ETE assumption that evacuees will choose to evacuate to assigned host I communities, is unreasonable and particularly insupportable due to the large numbers of beachgoers and transients during the summer months who may be unfamiliar with host communities. Several elements within the ETE and NHRERP in general must be raised in response to this claim:
The precautionary early closing of the beaches does not consider the immediate opening of reception and mass care centers for this population segment. The EBS message for beach closing directs these persons to leave the area immediately (Vol. 4, p. G-13).
There is no reason for this population segment to report to host communities until and unless the order to evacuate is given. At that time, EBS messages will provide direction for all evacuees to travel to the appropriate reception centers; and TCPs throughout the EPZ will be manned to provide route guidance.
Public Information flyers and supplemental f materials are to be provided to restaurants, f
_ _ _ _ _ _ _ _ _ _ _ _ l
hotels, motels, rooming houses, schools, camps, health care facilities, parks and state forests (Vol. 1, p. 2.3-2). This material will contain i
information regarding the location of reception i centers and evacuation route descriptions.
The NH public information program also includes posters prepared by NHCDA for display in public places and state lands (Vol. 1, p. 2.3-2). )
l Emergency public information will be contained on !
dedicated pages in telephone books which will be distributed in the EPZ (Vol. 1, p. 2.3-2). These pages will provide general instructions on what to do following notification of an emergency at Seabrook Station, and specifically list the Reception Center assigned to each EPZ community.
Tourists that enter an area with which they are particularly unfamiliar will, in all likelihood, have in their possession maps of the area to avoid getting lost during their vacation.
The traffic control points are established to
"(f)acilitate evacuating traffic movements which serve to expedite travel out of the EPZ along routes, which the analysis has found to be most ,
effective." (Vol. 6, p. 7-1). Hence, direction is .
provided throughout the EPZ along designated routes.
I t
1 i
There is no requirement that an individual report to a host community. Reception centers are' l
established for the evacuating public to provide !
" services for any evacuated population in need of public assistance" (Vol. 1, p. 1.6-6). The EBS messages provide information as to what services-are.available at the reception centers (Vol.-4, p.
G-31.and p. G-37). However, it is an option for l l
evacuating persons to refer to reception centers-f for assistance. l The NHRERP calls for strip maps to be available to l l
all transportation drivers in reaching their designated destinations.
It is reasonable, based on the foregoing ETE and other NHRERP elements,.to assume that both permanent resident and transient populations will have adequate access to information regarding appropriate evacuation routes.
B. Human Behavior in Emergencies 1
- 1. General Character of Public Behavior'in Emergencies
- a. General Public_Pesponse Princioles Human response to emergencies has been a topic of investigation by social scientists for approximately three decades. Many studies have been performed in a variety of emergencies stemming from geological, climatological, and technological phenomena, and has been documented in a vast body of emergency literature. This research history has l
l i
i I
I i
covered many aspects of human behavior in emergencies; and q there have been several attempts to summarize the findings of I the hundreds of existing studies (see, for example, Thomas E.
Drabek, 1986. Human System Responses to Disaster: An Inventory of Sociological Findings. New York:
Springer-Verlag; Dennis S. Mileti, Thomas E. Drabek and J.
Eugene Haas, 1975. Human Systems in Extreme Environments.
Boulder: University of Colorado). An important component of 1
this work has been to investigate the behavior of people within the community at risk in response to an emergency.
Although additional research will continue, the record is clear with respect to the response of members of the public at risk in times of emergency. The principles of public response discussed below -- well established through decades of research and investigation -- would be applicable in the l event of an emergency at a nuclear facility such as the Seabrook Nuclear Power Plant. ;
l Emergencies are by definition, and in reference to human
]
i behavior, unique situations. Emergencies analogous to one at J Seabrook which pose a collective threat to an entire community are, behaviorally, in a class by themselves. Mass emergencies such as these transform communities behaviorally at both the group and individual levels. Priorities of ongoing social life shift, goals and objectives are transformed, and identifications change. The first priority for virtually all people who find themselves in such a
j d
l 1
i collective threat situation becomes the collective safety of l
-l people and the community at large. People Gbandon personal forms of identification'and personal interests, and they identify with the entire human collectiv? or community that is threatened. This " shift" in the human character has come l to be known by many names, for example, the " therapeutic community" (see Alan H. Barton, 1969. Communities in Disaster: A Sociological Analysis of Collective Stress Situations. Garden City, New York: Doubleday).
The change or " shift" in the social psychological complexion of social life and human behavior results in a variety of principles that emerge to document the character of emergency behavior. This includes, for example, a dramatic decline in activities and behavior that run counter to the good of the collective and those that are based in individual or personal interests, and a dramatic increase in acts and behavior that bring people together and help one another. This "chift" would undoubtedly occur in an emergency at the Seabrook Plant; it has occurred in every mass emergency of this sort studied by social scientists where it has been a topic of investigation (and has been evidenced even in emergencies where it was not formally a topic of investigation). This general shift in human behavior in emergencies should be considered in planning for emergencies since it will characterize the basis for human behavior in future emergencies.
l
l l
i The results of actual empirical research on human behavior in mass emergencies provides clear guidelines for planning for future emergencies. Public behavior is rational ~, and the emergency goals of helping themselves as 1
well as others take precedence over almost all else; the j character of human spirit is strong when faced with mass emergencies and most people rise to the occasion. In simple terms, the " thin-veneer of civilization" is not stripped from humanity when mass emergencies are experienced (as one would conclude from observing disaster movies); it is in fact strengthened.
- b. Ride Sharina People in emergencies become altruistic and concerned about the safety of others. consequently, people check on the safety of others; communicate with friends, neighbors and intimates; and offer help and provide assistance to each other. Further, evacuation can be generally characterized as a group behavior, rather than a behavior in which people engage as individuals. People generally exhibit -- as a consequence of becoming altruistic and concerned about the safety of others -- a tendency to form into groups prior to and for the purpose of evacuation.
Historically, in emergencies where evacuation has been recommended, people without transportation have obtained evacuation transportation from friends, neighbors and I
relatives. This is the case because of the natural tendency i 1'
l
E h
L l-L of people in emergencies to' check on the safety of others, to become. altruistic and offer help to those who need it, and to
. form groups for evacuation. Therefore, in the event of an emergency.at Seabrook necessitating evacuation, regardless of the reason forLlack of personal private transportation, the vast majority of evacuees without their own transportation would receive transportation from other evacuees. This would likely be particularly true for people whom others (e.g.,
friends, neighbors and relatives) . know do not own a car or have a car which is out of the EPZ - for example, because of l a working spouse. Thus,.the number of persons who would need l
I official transportation assistance during an emergency can be expected to be well below the number of persons without transportation in non-emergency Times. The assumption employed-for the Seabrook emergency planning zone for y planning purposes (i.e., that approximately 50% of those who 1
do not have personal private transportation would receive-
! transportation from other evacuees) is very conservative and reasonable, based on principles of behavior established through the study of emergencies.
This nation, on the average, experiences at least one evacuation a day. Some of these involve major events in which a large number of people evacuate. Despite the volume of the research record, we know of no emergency evacuation case in which the lack of transportation has ever been a problem in constraining the act of evacuation for persons who l
wish to evacuate. This, of course, should not suggest that transportation needs for evacuation should be excluded from emergency planning for nuclear power plants. It does, however, suggest that the planning assumption that 50% of those without their own transportation would receive rides from other evacuees is a very conservative assumption. In fact, unless there were something very unique about the population in question (for example, if many belonged to a religious group whose beliefs kept them from owning motor vehicles), 70% or 80% would be a conservative planning assumption.
In summary, an extensive history of human behavior in emergencies leads one to be confident that -- in the event of an evacuation due to an emergency at the Seabrook Plant --
the vast majority of evacuees without their own transportation would receive transportation from other evacuees.
- c. Spontaneous Shelterina The general principles of human behavior in emergencies which have already been discussed are applicable to all types of analogous mass emergencies, and have been observed in climatological, geological and technological emergencies including the Three Mile Island accident. Interestingly, another frequent research finding is that many people are surprised, as they reflect back on their emergency experience, that they did not observe what they might have l
l
- _ _ - - - - _ \
i i
i expected to observe, for example, panic, hysteria, selfish acts,' conflicts over scarce resources, and so on.
Expectations about human behavior in emergencies by most Americans are quite consistent and even the antitheses of actual human emergency behavior. The " myths" which permeate our society about emergency behavior are strong and likely explain why most people who experience an emergency are ,
surprised to learn that "when the chips were down in our ;
community people really pitched in to do their part and help one another." More often than not, locals attribute this inconsistency between what they would have expected and what they experienced to the " unique" character of their community's citizens. In fact, it is the general universal character of social life which leads to what is observed.
One consequence of the existence of strong myths in American culture about emergency behavior is that these myths can manifest themselves in a variety of specific concerns as planning occurs for future emergencies. The underlying and incorrect thesis on which these specific concerns rests is i
typically that emergencies bring out the " worst" in people !
who become more irrational, selfish, aberrant, helpless and so on. The empirical facts are that emergencies bring out the "best" in people who become more rational, altruistic, compassionate, helpful and so on.
The notion that strangers wenld be left outside in an I emergency at Seabrook in which sheltering were the advised l
__-_- __ -_ __-- _ - t
protective action is a concern that illustrates how particular " problems" are imagined in future emergencies because those raising the concern are basing their view on American " myths" about behavior in mass emergencies instead of empirical fact.
We do not fault Americans for subscribing to the incorrect thesis about emergency behavior. Hollywood movies and script writers have done little to dispel myths about such behavior, and, in fact, are a likely cause of their reinforcement. Nevertheless, emergency planning should be based on fact and not fiction.
It is inconceivable that people (be they shopkeepers, hotel operators, theater managers or merely citizens in their homes) would lock people out were a sheltering advisory l issued at Seabrook and people were " stranded" outside. The reverse would occur as people would encourage those outside to seek shelter in the buildings which they occupied. It is equally inconceivable that persons outside would engage in aberrant behavior in an attempt to get inside. The only l
behavior that could be reasonably hypothesized in such a circumstance would be on the part of those inside encouraging those outside to share their shelter.
Unfortunately, we know of no specific studies to cite that would empirically document this phenomenon. The reason no studies on this specific issue exist, however, is that it has never been identified as a problem in the hundreds of
- 100 -
1 i ' emergencies researched. There are, however, dozens upon dozens of studies which illustrate the generic principle of how people help one another in emergencies and share the resources available with those who need them. The prototypical conclusions are illustrated as follows:
"as a specialized aspect of the overall community response pattern, heightened levels of solidarity have been reported repeatedly" (cf. Thomas E. Drabek, 1986, Human System Responses to Disasters: An Inventory of Sociological Findings. New York: Springer-Verlag).
l
" Previous intergroup differences are lessened, cooperation and social solidarity are heightened..."
(cf. Nicholas J. Demerath and Anthony J.C. Wallace, 1957, " Human Adaptation to Disaster" Human Organization 16, Summer: 1-2).
" ... disasters create unity rather than i 7
disorganization...a set of norms which encourages and I' l reinforces community members to act in an altruistic I fashion develops; also, a disaster minimizes I conflict....(Ressell R. Dynes, 1970 a. Organized Behavior in Disaster. Lexington, Massachusetts: Heath l
j Lexington Books, page 98). 1 These conclusions about altruism are prototypical of many others. Individually and collectively they suggest that I l
it is beyond the realm of human possibility to conceive of an emergency at Seabrook in which anyone would be " locked out" a
when sheltering was ~dvised. I
- d. Evacuation Impediments We have already addressed a general description of human response to emergencies as well as provided some thoughts about how American myths about emergency behavior are quite at odds with empirical scientific fact. We have also illustrated how the " facts" of behavior in emergencies would
- 101 -
likely manifest themselves in terms of ride sharing and spontaneous sheltering were there'an emergency at'Seabrook Station. In this'section of testimony, we will address the concern that evacuees would not be able to deal with evacuation impediments, such as stalled cars on the highways or accidents.
We shall not be repetitive. The same theory about-human response in emergencies which we have already elaborated upon is quite applicable to this concern.
The public is not helpless'in emergencies as'would be suggested by the thesis underlying " myths" about emergency behavior. In fact, the public is an actual resource in' emergencies. It is the case in most emergencies, for example, that most search and rescue activities to help victims are actually carried out by other victims.
... studies have consistently shown that initial search and rescue work is carried out by persons who are in the impact area and that formal rescue organizations become involved at a later point..." (cf. Dennis S. Mileti, 1985, Disaster Relief and Rehabilitation in the United States. A Research Assessment. Boulder: University of Colorado, page 9).
The point is a simple one. The public in an emergency is not helpless, awaiting for a formal organization to solve the problems which are presented by the emergency. In fact, I
most emergency problems are solved by the public. Search and 1
rescue is an example of this principle. pushing ;
l
- 102 -
l l
1 L______________-
r incapacitated vehicles off the road would be another. Dozens of other examples could be offered.
The public is a resource in emergencies, it is not the problem. 'The public often gives rise to " emergent work groups a in emergencies which form temporarily and on the sp:,c to do what needs to be done. The importance of this pheroironon has already been put well in terms of emergency planning application. "
...too few emergency officials really grasp or understand the range of important contributions made
,, by such groups" (Thomas E. Drabek, 1986, page 154).
Emergency p14ns for Seabrook Station provide for the removal of traffic impediments along EPZ evacuation routes through the dispatch of tow vehicles as identified through established I<etters of Agreement. If, however, emergency pl 9 ns for Seabrook also assume that an evacuating public would push vehicle impediments out of>the way, then the plans are based on a very accurate impression of public behavlor in emergencies; for this is precisely what would occur.
We should also add one more thought to evacuation and road impediments caused by accidents during emergencies. The evidence regarding accidents during emergency evacuation suggests that they were no more frequent than during non-emergency tie s. For example, in : study dcn'e by Dennis S. Mileti, donald M. Hartsough, Patti Nadson and Rick Hufnagel entitled "The Three Mile Isinnd Incident: A Study of Behaviora) Indicators of Human Stress," 1984 (Mass
- 103 -
'P 1
/ pu _, i
- v j
3;
. i Emergencies and Disasters, 2,1:89-114) data were examined I that included the frequency of accident rates before, during
.and after'the TMI accident. There was no evidence to suggest that accidents increased during the evacuation, s.
I
) despite.the fact that commutation in the area was likely up j- because od' evacuation. J' The conclusion abex5 the 'IMI accident is supported-by other evacuation studier. In an extensive review of .
, evacuation behavior, E.L. Quarantelli notes "A strong theme' e- ; J is that withdrawal movement is almost always, orderly in -
.getting people away from an achual or potentially dangerpus y location. This runs counter to widely held views among some b disaster planners and emergency organization personnel that there is e,need to be concerned about evacuation turning into
.# r e disorderly flight if not wild panic" (Evacuation Behavior and
,y
- Problems: Findings and Implications from the Research Literature.
4 .
Columbus: Ohio State University, Disaster ,
Research Center, 109). I
! ), -
It has alao been noted that accidents and traffic jams are not problems in vehicular evacuations. (H.E. Moore, et
,r al., Before the Wind. Washington, D.C.: National Academy of Sciences, National Research Council, 1963) specifically 4 investigated accidents and auto breakdowns in the evacuation of over one-half million people because of Hurricane Carla.
Only .6% of the sampled evacuees reported they had even seen 1
l l
- 104 - i l ,l ' ^ '
,, i e
Y
? )
.- --_--.---__-----9
y _ _ _ _ - _ _ -
L I
or were~in an accident or breakdown (note that all .6% may have witnessed the same accident). l
- e. ', conclusions i
.An elaborate body of empirical research accumulated over the last three decades exists regarding public emergency l behavior in the United States in reference to geological, climatological and technological emergencies. This research
-base also includes many studies about public response to the 1979 emergency and evacuation at Three Mile Island. This
['recordprovides'aclearbasisfortheconclusionthatmass 1 conmunity-wide emergencies (the sort of emergency of concern in these hearings regarding the Seabrook Station) elicit altruistic public behavior. In such a circumstance the public would clearly share rides vith other evacuees without i transportation, share shelter with those in need of it, and help themselves and others to solve " problems" (such as pushing evacuation impediments like disabled cars out of the i
way) encountered during the emergency.
{
)
- 2. Emercency Worker Role, Abandonment
- a. Introduction.
The research history of human response to emergencies has covered many aspects of human behavior; an important component of this work has been to investigate the behavior of people with roles of responsibility in emergency response, y s as well as the behavior of emergency organizations, and both "s informed emergency planning efforts in the nation. The
- 105 - l
research record has grown over time. Although additional research will continue, the record is clear with respect to.
the response of emergency organizations and emergency workers in times of emergency. We are confident that the principles of emergency. organization and emergency worker response discussed below -- well'-established through decades of research and investigation -- would be applicable in the event of an emergency at a nuclear power plant such as the Seabrook Station,
- b. Engroency Role Abandonment by Workers i The notion of " role conflict" is a concept in the social sciences based on the "ollowing ideas. Individuals in society play many different roles, and each role has certain rights and obligations in particular social relationships.
Because each person plays many different roles, the rights and obligations of one role may be consistent with those of another role, or irrelevant to or in " conflict" with another role. The concept of role " conflict" is generally used uncritically, as an either/or matter in which a person is forced to choose between two or more roles (cf., Lewis Killian. 1952. "The Significance Jf Multigroup Membership in Disaster" American Journal of Sociology January: 309-314).
Conflict implies equally weighted contradictory alternatives, requiring a person to choose one role to play while abandoning another. This condition is rarely, if
- 106 -
I~
l' ever, found in actual social life. A more accurate term is l
role " strain," which denotes the difficulty of role l obligations at the same time. Role " strain" is preferable 1
because it describes more accurately the actual conditions
! 'that people experience in all of social life, not just those of emergencies. Role " strain" is something with which people cope in most social situations and is a permanent feature of social life.
It is important to distinguish between role strain, which is a mental state (a feeling of concern and unease),
and role abandonment, which is a type of behavior. Thus, while it is to be expected that emergency workers would experience some role strain during an emergency at Seabrook, this does not mean that they would abandon their emergency roles because of it.
Research conducted on the actual behavior of people with defined organizational responsibilities in emergencies has established that emergency workers who have a clear idea of their emergency roles do their emergency jobs. This research record stretches back over the last three decades.
- 1. The Prevailina Imace It was in the specialized research literature of disasters and emergencies that the concept of role conflict was first conceptualized in sociological terms. This was l
accomplished by Lewis Killian (1952) who sought to develop a I
typology for role conflict types. That is, Killian did not
, - 107 -
l
)
1 4
seek to document the occurrence of role conflict in a set of )
I particular disaster situations, but rather to describe a concept he thought important to the general sociological l enterprise. Killian described the concept well, and his
)
original article has become a classic but abused wori The ]
classic " cell" in Killian's typology has become role conflict that is experienced by emergency workers who are torn between conflicting loyalties to family and work groups when disaster strikes. Disasters provided the discipline of sociology with a good example situation in which to illustrate to students role conflict of this type. It is easy to be able to imagine role conflict by thinking about having a family at home in a disaster-striken area when an emergency begins while being at work and having an emergency job that must be performed.
Killian's original work, however, seemingly went beyond defining a concept. He also reported on what appears to be a conclusion from empirical evidence about how emergency workers resolved their family-work role conflict. He states:
"The great majority of persons interviewed who were involved in such dilemmas resolved them in favor of the family, or, in some cases, to friendship groups. Much of the initial confusion, disorder and seemingly complete disorganization reported in disaster communities was the result of families trying to find and rejoin their families" (1952:311). Other early disaster studies offered similar conclusions, for example, a study of the 1953 Holland flood disaster (cf.,
l - 108 -
1
Instituut Voor Sociaal onderzock Van Het Nederlandse Volk Amsterdam. 1955. Studies in Holland Floor Disaster 1953.
Washington, D.C.: National Academy of Sciences-National Research Council, Volume IV, p. 11) concluded that "The choice of most men was to be with their families rather than fulfilling the tasks which they should have done in their positions as members of another group;" and "The number of men in rescue increased, if they knew that their wives and children were safe."
Other 7tudies, apparently, followed which provided ,
additional " evidence" that role conflict in disasters was resolved by the abandonment of emergency worker roles in favor of performing family roles. Moore (cf. Harry C. Moore 1958. Tornadoes Over Texas. Austin: University of Texas Press, p. 245), in a study of a disastrous tornado, concluded that " Efforts to reunite the family were the first thing done in many cases. Until this was done everything else was postponed and reported to have been insignificant." Moore also provided three case examples to illustrate the general conclusion. First, the mayor, on becoming aware of the ;
l disaster while driving home went home first, picked up his wife, and then started out to see what had happened. Second,
, the civil defense director was at the scene when the impact occurred, and after making calls to alert several disaster l agencies, he then telephoned his family and went home. 4 l
Third, the city manager, who was also close to the scene of
- 109 -
l 1
the disaster, checked with his wife in the course of his official activities to see if she was injured; he was able to do this because he was at police headquarters which had special lines.
Accounts of findings of role abandonment by emergency workers due to role conflict in disasters continued to mount.
For example, Fogleman (cf., Charles W. Fogleman, 1958.
" Family and Community Disaster Upon Individuals and Groups."
Unpublished Ph.D. Thesis. Baton Rouge: Louisiana State University, p. 343) concluded that the roles of husband and father were primarily in structuring activity in the period of disaster impact. Then, in 1958, Form and Nosow (cf.,
William H. Form and Sigmund Nosow. 1958. Community in Disaster. New York: Harper and Brothers) provided several generalizations based on accumulated evidence: " Help for family members, close friends, and neighbors comes first; then, but apparently only then, other victims can be looked after" (p. 66): "Only half of those persons (belonging to disaster oriented organizations)... sought to relate their activities to occupational organizational statuses" (pp.
114-115); " ... greater conflict exists for those who have injured family members than for those who do not" (p. 102);
and "...those who experienced... conflict showed a greater concern for family, neighbors and friends" (p. 109).
Finally, Moore et al. (cf. , Harry E. Moore et- al. 1983.
l Before the Wind: A Study of Response to Hurricane Carla.
l I
- 110 -
1 l
- - _ _ _ _ _ _ - _ _ _ _ . l
Washington, D.C.: National Academy of Sciences - National Research Council) conducted a major hurricane disaster study and concluded that the presence of conflicting role obligations was an important factor in directing the behavior of some persons in disasters. These studies, and perhaps a few others, provided " evidence" that role conflict in disasters is resolved in favor of primary family roles at the expense of emergency work roles.
The portrait of role conflict and abandoment in disasters and emergencies provided by these studies has been summarized in several texts on disaster (cf. Allen H. Barton.
1969. Communities in Disaster: A Sociological Analysis of Collective Stress Situations. Garden City, New York:
Doubleday; Dennis S. Mileti, Thomas E. Drabek and J. Eugene Haas. 1975. Human Systems in Extreme Environments.
Boulder, Colorado: Institute of Behavioral Science,
, University of Colorado) and is typically and unquestioning 1y l
referenced in most freshman textbooks. The implied social process is: (1) disasters produce role conflict in emergency workers, (2) role conflict elicits abandonment of emergency worker roles, and (3) abandonment of emergency worker roles contributes, at least initially, to the social disorganization imposed on human communities by disasters.
This is, indeed, the conventional " wisdom" or image of the determinants and consequences of role conflict in disasters and emergencies.
4
- 111 -
l L _ - -- _ . _ _ _
i i
l 1
- 2. Additional Research Evidence i Several questions must be formulated and then answered before we conclude that a tornado, earthquake, chemical spill, nuclear power plant accident and other forms of natural and technological calamities mean the breakdown in I
normative structures for emergency response that society has erected to combat disaster when it strikes. Who, for example, " abandons" emergency roles in favor of family obligations during disaster? Are they people who know that they have an emergency job, or are they people who do not perceive themselves to be part of emergency response, but who outside observers judge should have thought one up during the emergency?
~
outside observers could be sociologists or anyone else expecting to find " social disorganization". Is role conflict something that results in playing one versus another role in an emergency, or can it and is it resolved by doing both? If there are cases where people who knew they had an emergency job left that job, did they leave without regard for whether or not the emergency work could get done, or was it clear before they left that others staying on the job would be able to handle it? The answers to questions like q l
these would help us understand if role conflict results in I role abandonment in emergencies (and if early accounts of role abandonment were, for example, merely based on some researchers observing a person who did not know he had an l emergency job not doing that emergency job); why it occurs if
- 112 -
l 1
it does; the theoretical character of how role conflict is resolved or played out in emergencies; and if it is a problem, how it can be managed by those who would use sociological knowledge to prepare their communities for disasters.
It is interesting that the scholar who'gave us the concept of role conflict in disaster was the first to step forward to begin a clarification. Killian (cf., Lewis M.
Killian. 1954. "Some Accomplishments and Some Needs in Disaster Study." The Journal of Social Issues x: 66-72) provided the field with some important but too often overlooked insights only two years after his initial publication. He suggested (1954:60) that "The possession by the individual of a clear conception of a role which he can or should play seems to be conducive to organized adaptive behavior." Killian was suggesting that disasters are not social disorganization, and that role conflict does not result in role abandonment in disasters, if emergency workers are provided -- before an emergency -- with a clear idea of what would be their emergency role. If Killian were correct, it would be quite inappropriate to observe the behavior of a "would-be" emergency worker without a pre-emergency idea of l
his/her emergency role not doing emergency work in a disaster, and then conclude that role conflict led to role abandonment and social disorganization. Such observations would not be role abandonment since it is difficult to see
- 113 -
how a person could " abandon" a role if they did not know that they had it. Killian (1954) went on to also suggest that role conflict can contribute to " personal disorganization" or
" anxiety" while emergency workers do their work. On the basis of this view, therefore, one would suspect that role conflict is a mental phenomenon that describes an emergency worker's need to know that loved ones are safe rather than behavioral phenomenon that could result in emergency role abandonment and consequent social disorganization.
This alternative (psychological rather than behavioral) view of role conflict (for emergency workers who know they have an emergency role rather than those who are unsure er do not know about an emergency role for themselves) rests on more than Killian's (1954) second article. For example, Fritz (cf., Charles E. Fritz. 1961. " Disaster." pp. 651-694 in Merton and Nisbet (eds.), Contemporary Social Problems.
New York: Harcourt) attempted to provide an overview of how families, people with emergency roles, and people without emergency roles all relate when disaster strikes. In essence, an informed reading of Fritz would lead a reader to hypothesize the following: (1) all people are concerned for loved ones in disaster, (2) most people do not have emergency roles, and these folks are torn between loyalty to intimates versus the community when disaster strikes -- most of these people tend to their intimates first and only then volunteer for emergency work, (3) people with clearly defined emergency
- 114 -
jobs perform.their emergency roles, but under personal stress until they are assured of the safety of intimates, (4) I provisions can be improvised, or planned for, to assure -
emergency workers of family and intimate safety and remove j this stress for them during the initial aftermath or f
disaster, and (5) finally, that role conflict for emergency j workers who have a clear pre-emergency notion of their emergency role is a mental and not behavioral phenomenon.
Another study (cf., Fred L. Bates et al. 1963. The Social and Psychological Consequences of a Natural Disaster.
1 Washington, D.C.: National Academy of Sciences - National !
l Research Council) investigating the impact of a hurricane on impacted communities reported -no role conflict in emergency workers; everyone opted for family instead of work roles.
However, there was no emergency plan in the community studied
-- no one had a pre-emergency notion about their emergency role. The picture presented thus far did offer some insights about role conflict and for whom it might mean psychological stress while doing emergency work versus doing emergency work only after intimates were attended. However, these conclusions were still somewhat hypotheses, and no one had as yet offered a clear theoretical explanation as to why.
An answer to this question was provided by White (cf.
White. 1962. " Role Conflict in Disasters: Not Family but Familiarity First." Unpublished MA Thesis. Chicago:
University of Chicago Press) is an unpublished MA thesis at
- 115 -
c-__---------------------- -
l the University of Chicago. White investigated data on emergency role performance in three different emergencies.
The findings were that 82% of the emergency work force contributed to disaster response activity first -- over 1
family roles; and that this figure rose to 89% after the j
first four hours from disaster impact had elapsed. She l concluded the " role-certainty" was the key factor that 4
explained what emergency workers do in reference to emergency '
roles when disasters strike. That is, if an emergency work role is " uncertain" to a worker when an emergency begins, that worker will opt for other roles -- like family roles --
that are more clearly known and "certain." However, for emergency workers who have clear and certain notions about their emergency werker role, emergency work role performances take precedence over other roles including family roles. ;
White (1982), therefore, provided a theoretical rationale and explanation for the performance of roles in disasters; role confict exists in emergencies, role certainty in emergency work roles yields performing emergency roles over family roles and family roles over emergency roles when emergency work roles are not clear and certain. This explanation could i account for why research had sometimes found emergency worker
" role abandonment" and sometimes not.
Finally, perhaps the most all-inclusive work on role l
conflict / abandonment in emergency workers performed to date was by E.L. Quarantelli (cf., E.L. Quarantelli. No date.
t
- 116 -
l l
l
i i
" Structured Factors in the Minimization of Role Conflict: A Re-Examination of the Significance of Multiple Group Membership in Disasters." Working Paper. Columbus, Ohio:
The Disaster Research Center at the Ohio State University). ,
1 This work looked at 150 different disaster events and the l behavior of over 6000 emergency workers. The conclusion was i
(cf., page 3) that role conflict was not a serious problem in l i
the loss of manpower in emergency situations. Six emergency ]
events were examined in more detail. No one at work when the i
( emergency began abandoned their emergency job. A dozen people, however, left their job temporarily; they were not needed for the organization to do its work. The general conclusion of this work was that role abandoment is absent in emergencies because of social changes that occur in communities when emergencies begin; for example, other family members pick up internal delegations to the family, kin, neighbors and so on which allows the family member to participate in emergency roles.
- 3. Conclusions from the Research Base The conclusions that can be reached regarding the role conflict and abandonment on the part of emergency workers are straightforward. First, when emergency work roles are "certain" -- perhaps through training and planning -- for emergency workers, role conflict in emergencies does not result in the abandonment of emergency work roles. Second, when emergency work roles are not clear or "certain" .
9
- 117 -
l 1
1
--perhaps through a lack of training or planning -- for emergency workers, role conflict in emergencies can result in seeing would-be workers play more certain roles toward intimates before attending to emergency work. Third, role conflict for emergency workers on-the-job during an emergency can elicit psychological stress or at least concern about the safety of intimates; and workers can improvise or emergency plans can formalize ways that emergency workers can check on the safety of intimates. Conclusions such as these suggest that if emergency workers -- before disaster strikes -- have a clear (Killian, 1954:60; Fritz, 1961; Bates et al., 1963) and certain (White, 1962) image of their emergency role which
, can be achieved through planning and training (cf., Russell Dynes. 1970. Organized Behavior in Disaster. Lexington, Mass.: .D.C. Health,.p. 155), that emergency workers resolve role conflict in emergencies in favor of emergency work roles while improvising ways to check on the safety of intimates Quarantelli, no date) unless formalized ways to accomplish this same objective are drafted into emergency plans. This conclusion would certainly explain why researchers (cf.,
Killian, 1952; Instituut Voor Sociaal Onderzock Van Het Nederlandse Volk Amsterdam, 1955:IV-11; Moore, 1958:66; Moore et al., 1963, and others) reported role conflict in emergencies resulting in role abandonment during times when emergency planning and training -- and consequent low levels of emergency worker role certainty -- were slight in American
- 118 -
communities. It would also explain why other researchers (for example, Quarantelli, no date) found so little role l abandonment resulting from role conflict in more i
contemporary emergencies where emergency workers could have had more certain.(cf. White, 1962), pre-emergency notions j l
about their emergency roles. It seems, therefore, that the abandonment of emergency work roles by emergency workers is not a problem in disasters if emergencies are prefaced by emergency worker training. It is, however, one very real reason among others why emergency planning and emergency worker training is essential.
- 4. Discussion The virtual lack of evidence of persons abandoning known emergency roles over a wide range of emergency events in the past illustrates that there are certain structural changes in the community during emergencies that reduce role strain.
Role obligations are based on values and, during non-emergency " normal" times, people expend efforts on achieving many different values, some potentially contradictory. An emergency changes this process dramatically. Some values become clearly more important than others, and people experiencing the emergency generally agree on the few that take precedence over all others. This phenomenon is the development of what has been called the
" emergency consensus." In this consensus, people become l I
altruistic, and protection of the community becomes the
- 119 -
l L______________.
[
- highest priority. Other values become drastically less important. This temporary shift in values simplifies the role obligations of people, and greatly reduces the potential for role strain. People are " released" to concentrate on the critical tasks of the emergency. The value priorities in an emergency thus simplify, rather than aggravate, role strain for community members. Those with identified emergency roles are therefore able to fulfill them, rather than ignore the emergency in order to tend to the obligations of other roles.
People who know in advance of an emergency that they have emergency roles to play (should one ever occur) are able to make informal family contingency plans in advance of the emergency. For example, families can make plans in advance of an emergency to ensure that -- in an emergency -- the non-emergency-worker spouse (or other appropriate person) will take the appropriate measures to protect the family unit in the absence of the emergency worker. Similar pre-emergency plans can be made to provide for the continuation throughout an emergency of child care arrangements in place at the time any emergency occurs. Such pre-emergency planning can also provide for the assumption by relatives, friends or neighbors of other usual responsibilities of the emergency worker. Indeed, in past emergencies, even in the absence of such contingency plans, the role obligations of emergency workers toward intimates (including family) have generally been shifted and assumed by
- 120 -
l other non-emergency-worker members of the intimate group, thereby freeing the emergency worker'to fulfill assigned emergency roles. As discussed above, role strain in.
emergency workers during emergencies between emergency role obligations and family (or-intimate' group) obligations is not an either/or situation. That is, emergency workers are not required to choose between mutually exclusive roles. Role
~
strain between family and organizational responsibilities would be a mental process in an emergency; it would not result in role abandonment or interfere with the ability of emergency organizations to do their emergency work.
Emergency workers would need to feel assured that their families can tend to their own safety. However, historically, in past emergencies, emergency workers have performed-their emergency roles at the same time they have sought assurance of the safety of their families through ad hoc, informal channels.
There is little difference between trained full-time emergency workers and trained part-time or volunteer emergency workers, for example volunteer fire departments, on this issue. The reason is that the number of hours typically worked and the rate of pay, if any, are not relevant factors )
in determining the performance of emergency roles by emergency workers.
One relevant behavioral difference between classes of emergency workers is whether they are at work or somewhere
~ 121 -
else, for example at home, when emergency mobilization begins. In this way, volunteer emergency workers could be different from other kinds of emergency workers only because ,
they are more likely to not be at their emergency job when the emergency begins.
Emergency workers who are at home when an emergency begins will likely continue to play the roles they are performing at the time (for example, father, husband, wife, mother, and so on) for a brief period of time. These roles are played at the same time that decisions are made that enable them to feel free about separating family members so that emergency roles can be performed. What might seem to be a potential for delay in reporting for duty is not, in operation, a real problem because (a) most emergencies do not begin with the need for immediate and dramatic actions like evacuation; emergency organizations can be mobilized in stages while families complete decision-making, with other family members or intimates assuming the potential family l role obligations of the emergency worker who has reported for work; (b) most families include at least one other member who is competent (i.e., able to drive, listen to emergency information, and make decisions) and able to assume the
- potential family role responsibilities of the absent emergency worker; and (c) workers are typically anxious to i
resolve other responsibilities and report to their emergency work stations. Workers away from home when an emergency
- 122 -
l
-begins typically improvise ways to assure themselves of the safety of intimates while tending to their emergency job duties at the'same time.
' Finally, emergency workers fulfill their role obligations in emergencies because of the cohesiveness of the work group. People with knowledge of their role do not wish to let their co-workers down. In emergencies,.'ha-nmed to have emergency work performed is obvious, and tPa emergency roles take on added importance. The sense of obligation to perform that role is strengthened, and the sense of obligation to the community as a whole is strengthened. And, since part of the traditional image of epouse and parent includes the obligation to perform adequately the necessary work-related emergency roles, the traditional role reinforces (rather than conflicts with) the emergency roles of the emergency worker.
The potential of role strain for reducing the effectiveness of emergency response is minimized by the organizational and occupational assignment of emergency responsibility. Assigning clear responsibility creates clear direction for behavior and ensures that the work of emergency response organizations is fulfilled. In addition, advance knowledge of one's emergency role allows the emergency worker to arrange to avoid role " conflict" and indeed to fill the role of family member more effectively than if he or she were not an emergency worker. The worker can preplan for family ,
- 123 -
safety in.an emergency. Also, because of organizational ties, the emergency worker is more likely to receive information about the nature of the emergency and'the nature of the risk, which allows the worker to fulfill family obligations more effectively.
Emergency worker. performance levels are ensured by training. Training provides the context.for those individuals to preplan activities that allow them to implement their emergency roles, such as providing back-un and supplementary options for continuing family l
responsibility. In particular:
(1) . Training gives the emergency worker a clear
., understanding of what is expected of him or her -- of what i his or her emergency role is.
(2) Training makes the worker aware of the advantages of making family contingency plans in advance of an emergency.
(3) Training makes the worker aware that the community and the emergency worker's co-workers depend on him or her.
Failing to report for work would increase the burden on others.
(4) Training informs the worker about the nature of the radiation risk and about the procedures to be used in dealing with that risk, such as dosimetry and radioprotective drugs.
In sum, emergency workers who know of their emergency roles in advance of an emergency perform their emergency i I
- 124 -
H functions in the event of an emergency. We do not know of even a single case in the history of emergencies in this nation where the ability of an emergency response organization to function was impaired by a failure of personnel to report for duty. There are, however, many cases where people who were not advised of their roles in advance of an emergency carried out other duties before volunteering for emergency work. Thic simply underscores the importance of advising emergency workers of their roles in advance of an emergency.
In conclusion, a large body of historical evidence shows the functioning of emergency organizations is not hampered by failure of emergency workers to perform their jobs. In spite of role " strain," such workers perform effectively.
Moreover, training for emergency work can reduce role strain and enhance the effective performance of emergency workers.
While role abandonment may be theoretically powsible, it is certainly extremely rare, and consequently it does not reduce organizational effectiveness. Indeed, the typical problem in emergencies is not that too few workers report for duty, but rather that too many persons volunteer. (In some cases --
most notably those involving response organizations which do ;
1 not require multiple shifts -- the management of excess j i
personnel has had to be addressed.) )
I
)
- 125 -
I l
- - _ - - - - - - - - - - - - i
- c. Emeraency Role Abandonment by School Worlarg Regarding their behavior in an emergency, school-l organizations and teachers are not unique. Their behavior in l
an emergency would not differ in any way from the behavior of.
others with known. emergency roles. Indeed,'the research record strongly supports the ascription to teachers and other school personnel of responsibility for supervision of g
l students in an emergency. Emergency-response is facilitated when plans are predicated on pre-existing roles, rather than on artificial' social structures constructed solely for emergency response use. (This does not mean that non-school
. personnel could not care for students in an emergency; good planning and some training in the supervision of students would ensure the efficacy of the use of such personnel in l
such a role.) Because school personnel who would accompany students on buses, for example, would be exposed to no greater radiological risk than the evacuating general public,
, and because school personnel are accustomed to' supervising students and dealing with students in the routine of their day-to-day work, school personnel need not receive the detailed training to be provided to other emergency worksrs.
It is necessary only that such personnel be informed of their emergency duties in advance of an actual emergency -- for example, in the orientation program for school personnel at
- i. the beginning of each academic year.
I.
Like other workers, the key to the response of teachers and other school personnel in any emergency (nuclear or
- 126 -
non-nuclear) is their awareness of their role with respect to the students in that emergency. We have no doubt that some school teachers (and other school personnel) might be found vno would swear in advance that they will desert their students in a radiological emergency. (This would also be the case in the vicinity of any nuclear plant, including Seabrook, we are sure.) We also have no doubt that in a real radiological emergency, teachers and other school personnel would generally remain with their students as long as :
necessary -- at least long enough to see them safely onto evacuation buses, particularly if those charges were small children. Similarly, we are confident that sufficient numbers of school personnel will be available to supervise students on buses. It is not necessary that each class be accompanied by its teacher. I l
- d. Emercency Role Abandonment Durina Three Mile Island l
As far as we know, there is no rigorous study of the phenomenon of role strain or role conflict in emergency workers during the Three Mile Island accident. Remember that I these concepts refer to mental " feelings" about concern for doing one's job and concern for intimates (for example, family members). No one knows what was in emergency workers' i
minds during the accident.
Information is available about how emergency workers actually behaved during the emergency (that is, information about role abandonment), but even this information is sparse,
[ - 127 -
I L___----------
I because apparently no one conducted a full-scale study of role abandonment per'se. There is a short passage in the Report of the President's Commission on Three Mile Island (the'Kemeny Report) that apparently reflects claims that some hospital workers left their jobs. This passage reads:
"A substantial number of other persons that is, other i than pregnant women and preschool children, including I health professionals, voluntarily left the area around
- the plant during the weekend of March 30 through April 1." (J.G. Kemeny et al., Report of the President's Commission on the Accident at Three Mile Island, 41 17 (Oct. 1979)).
This passage likely rests on what has been written about hospital workers in two medical publications. This information is discussed in a following section of this testimony.
The only other information about role abandonment at Three Mile Island that we are aware of is the result of an assessment Dr. Dennis Mileti of this panel conducted in 1983.
Dr. Mileti realized that very little information had ever been gathered about emergency workers and role abandonment during the Three Mile Island accident. He also recalled that on several of his trips to Three Mile Island in 1979 and 1980 he had had informal conversations with all sorts of people about a variety of topics. On occasion, he asked people if emergency workers had showed up for work during the accident.
As Dr. Mileti recalls these conversations, all the people queried said that as far as they knew, emergency workers had done their jobs. These conversations were hardly a " study."
- 128 -
1 Because of their limited usefulness, Dr. Mileti decided that i
it would be more helpful if someone simply phoned a variety I of emergency organizations in Pennsylvania that would have had emergency responsibilities during the Three Mile Island accident and ask whether role abandonment occurred or not.
Luckily, a Ph.D. candidate in the Department of Sociology at l i
Colorado State University was interested enough in the idea I to call and talk to several emergency organizations in Pennsylvania. This student had helped Dr. Mileti do other work on Three Mile Island and was familiar with the accident, its timing, and the Three Mile Island research that has been done.
Dr. Mileti's instructions to this Ph.D. candidate were simple: (1) call a variety or organizations that would have had emergency responsibilities of some sort during the Three Mile Island accident in the O- to 5-mile ring around Three Mile Island; (2) be sure to include some local schools to find out about the behavior of teachers, bus drivers, and-other school personnel before the schools were closed; (3) speak to someone high enough up in the organization to know about all the parts of the organization, but not so high up that he vould not know about workers; (4) make sure that the organizational respondent had actually been present during the accident; and (5) ask if " role abandonment" had occurred i
in the organization, how many people it had involved, the total number of emergency workers in the organization (so we
- 129 -
I
.____________D
l could get an idea about what proportion of each organization i
abandoned their roles) and why individuals abandoned their emergency jobs.
This student phoned and questioned about a dozen organizations of his own choosing. These included five police departments, a spokesperson for r. 100-plus-man volunteer fire force, a representative of the Middletewn inclusive school district (who was asked about both teachers and bus drivers), Civil Defense, the State Troopers, the National Guard, the Pennsylvania Emergency Management Agency, and the Personnel Department for the Commonwealth of Pennsylvania. Not one case of role abandonment was reported by any organizational respondent in reference to teachers, bus drivers, the police, Civil Defense, the State Troopers, the National Guard, or the Pennsylvania Emergency Management Agency. The Commonwealth of Pennsylvania did report a 12-13%
absentee rate on the Friday and Monday of the "IMI weekend",
however, they also reported that a 10% absentee rate is normal before and after any weekend. (In'any event, general I
state employees are hardly " emergency workers".) Also, the j spokesperson for the volunteer fire department reported that (
"a few (firemen) who wanted to leave the area did." These !
few evacuating volunteer firemen cannot be characterized as having " abandoned their roles"; there is no reason to believe they were called upon to fight fire during the Three Mile l
l
- 130 -
L_______________
7 7 ,
' , )
' 'l
(
t 1
p/ ,
i f, Island accident or that their presence was crucial to the funct! toning of local fire units.
tven thi's assessment cannot Ise called an elaborate one.
e However, it does provide some information about role I
abandonment -- or rather the lack of it -- among trained
/)
- emergency l' workers, teach rs, and school bus drivers during the Three Mile Island accident.
Two papers, both written by hospital personnet, exist in which the lntjthors catalogue their perceptions about staffing in hospitals in the area around Three Mile Island. One was written by Christopher Maxwell, " Hospital Organization Response to the Nuclear Accident at Three Mile Island:
Implications of Future Orie/ntated Disaster Planning,"
American Journal of Public Health, Volume 73, 1982, pg.
275-279. The other was by J. Stanley Smith and J. H.
Fischer, "Special Conanunications:
TMI: The Silent ,
Disaster,"j7ournal of the American Medical Association, Volume 245, April 24, 1983, pp. 1656-59. These two articles are personal accounts by administrators who worked at hospitalsinthearea$roupdThreeMileIslandandwereat work 'during the Three Mile Island accident. These articles describe the effects of the confm,ed and often conflicting
/ c 1 informationgiventothepublic[followingtheThreeHils / I
( Island accident, but they do not support the proposition that role abandonment was a prob 3pm among emergency workers during the Threm Mile Island accident.
b 131 **
i i 1 '
1
]
j t
i 4
~ Several cautions should be kept in mind before reading i these articles. 'First, in all emergencies -- including the 1
/ accident at Three Mile Ipiand -- large numbers of people )
\
d abandon their jobs. Butf it is important not to confuse the j i
behavior of non-emergoney workers with the behqvior of emergency workers.
" Routine" jobs are typically performed.
The assumption should not be made thak the behavior of workers described in these publications was the behavior of emergency workers. At go' time was there'a medical emergency !
at Three Mile Island. Second, the aftermath of an emergency is always filled with stories that often become incorporated f
into folklore.
That these stories exist does not mean th96' the events or interpretations Mf the events that they
. portray are true. Stories such as these are' based on the indhrpretationsbyuntrainedlayobserversofphenomenathat t
mIyormaynotbeconsistentwiththewaytheyarelabeled.
For example, the abandonment of factoryjwork roles by <vorkers in an emergency could be seen as " role conflict" and INandonmentbysome,butnotbeseeninthesamewayby appropriately trained researchers seeking to examine the behavior of emergency workers in an emergency.
Stories about other unreal emergency behavior -- for example, panic, looting, and so forth -- often are told by the untrained observer after an emergency. The more informed observer, however, .:bs ' rarely as biased in interpreting the behavior that underlines these myths. For example, a mar
- 132 -
l
i arrested for " looting" while sifting through the rubble of j
his own house for personal objects may be seized because the police expect to find looters. Onlookers who observe the arrests are quick to spread the myth that looting is going on. ,Their interpretations of the arrest -- erroneous on many fronts -- could well end up in the written record about what happened during the emergency. The written record and oral stories about looting, then, perpetuate the myth, alter the expectations of others about future emergencies, and result (in the example just cited) in still more erroneous expectations.and arrests in subsequent emergencies.
" Third, the published stories about " role abandonment" of medical personnel that have been provided us give little if any background with which to interpret their descriptions of the behavior of medical personnel. Suppose, for example, that a flood were-coming, that there were a hospital situated in the flood plain, and that evacuation advisements were issued for people in the flood plain. Under these circumstances, would evacuating medical personnel be
" abandoning roles" or would they be engaging in perfectly reasonable and appropriate behavior? We can answer this ,
question only if we first know what circumstances prevailed when they left. " Role abandonment" would correctly characterize their leaving if they did so while knowing that their assigned role was to help evacuate patients and if they
( also knew that co-workers left behind would be incapable of l
- 133 -
i
performing the emergency duties that had to get done. But leaving could hardly be considered role abandonment if it were done without knowledge that they had an emergency role.
Fourth, the published stories about " role abandonment" of medical personnel in hospitals should not be interpreted in the absence of knowledge about the role of emergency information in determining the behavior of people in emergencies. The importance of information is addressed within this written testimony on " voluntary evacuation," but it is relevant to " role abandonment" as well. For example, the uninformed may wonder why people who were not pregnant or young (the two groups advised to leave during Three Mile Island) also evacuated. The answer is simply that the public is not the military and cannot be expected to obey instructions blindly if a different course of action appears more rationti in light of their situational perceptions of i
the risk. The advisement that some selected subpopulations should leave a definitive geographical area was also news to others in that area that they could be exposed to some risk.
Those who would have us use a military paradigm with which to evaluate public responses at Three Mile Island would conclude that there was " overresponse" because more than the young and the pregnant left. Those who understand that people respond in emergencies on the basis of situational perceptions of risk, on the other hand, would understand that evacuation advisements were a source of risk information for everyone in l
- 134 -
._____-_________--______m
l the area that was to be evacuated and not only for workers doing routine hospital work as.it would be'for anyone else.
Fifth, one should also keep in mind, before turning to
]
these published comments, that descriptive evidence of any sort is not a good predictor of future events unless it. takes into account the determinants of these events. For example, the fact that a particular airliner has crashed in the past does not mean that others will crash. Rather,'one has to investigate why the airliner crashed -- and why others do not
-- and generalize to the future only on the basis of the determinants of successful versus unsuccessful flights. That is, one needs to determine that when x, y, and z exist, then an airliner can crash. One should never speculate about future events without once considering the established, valid predictors of those events, because the determinants of what one is speculating about could well be different from one event to another.
With these thoughts in mind, we interpret the two articles on role abandonment by medical workers during Three Mile Island as follows. The two articles are personal accounts by people who worked in local hospitals in the area around Three Mile Island and who were there during the accident. These accounts have been read carefully and also a research assistant of Dr. Mileti's has called Dr. Smith and >
Mr. Maxwell to ask some questions. Neither of these articles suggest that trained emergency workers did not do their
- 135 -
t i emergency jobs during the Three Mile Island accident. There l are several reasons why we hold this opinion.
I These hospitals were not in the 0-5 mile Three Mile Island planning zone. For that reason, none of them had a radiological emergency plan that clearly specified emergency roles for workers. At the same time, the public information at the time suggested that " risk" did exist at the location of the hospitals. Dr. Smith reported that about 50% of the staff left who were specialists and not in a role of responsibility for activities that would include, for example, patient evacuation. Also, division chiefs and other higher authorities stayed to oversee evacuation work that was likewise clearly part of their job. Put simply, the people who left did so because they had no emergency role. In addition, weekend staffing is only one-third of regular staffing anyway. Dr. Smith reported that the evacuation weekend coincided with a national physicians' meeting in another state, and many pecole were away attending the meeting. Mr. Maxwell reported that his hospital had no radiological emergency plan, that few workers knew if they had a role or not, and that the normal reduction in staffing for weekends largely affected how many people were at work in these hospitals.
What we conclude from these publications and conversations with their authors is that hospitals can have a L hard time performing an emergency act like evacuation if they
- 136 -
l
~J r
1 l
do not have an emergency plan assigning emergency roles.to j would-be emergency l response hospital staff' persons; the incident at TMI occurred when normal. staffing was down to:
one-third and another two-thirds of the sta'ff were somewhere elsel(a convention, at home, or evacuating themselves) and
]
did not perceive themselves as having an emergency role because no plan'ever gave them one. Since they knew that they had no emergency roles and perceived their routine place of routine work to be an "evacaution area," they cannot be faulted for not going to their routine jobs on Monday morning. These cases showed no incidents of role abandonment; they do, however, document a lack of planning I
and its consequences.
- e. The "Uniaueness" of Radiation j 1
The ability to transfer the principles of human behavior i in emergencies is questioned by some as not being applicable 1
in emergencies that involve nuclear power plants. There are two forms of the argument. The first (more extreme) position j is simply that radiation is so unique that it produces its own unique set of laws of human behavior. The claim is that radiation is different enough from the risks posed in other types of emergencies that human behavior would also be different.
The determinants of human behavior in emergencies in contemporary America are, however, transferable across l
emergency types. This is not to say that the descriptive J l
- 137 -
i r
1 I
i accounts of how people respond to emergencies will be the same from one emergency to another but the reasons vny people respond in different ways to emergencies -- the determinants of behavior -- are consistent across emergency types. The position that radiological events are unique argues against the basic premise on which the social sciences rest: that there are knowable reasons and patterns in human behavior that are discoverable through systematic scientific inquiry.
The position is also contradicted by empirical evidence.
Another, equally untenable, position taken by some is that they agree that there are determinants of human behavior in emergencies and that these are knowable. But they argue
. that people are fearful enough of radiation because of its
" unique" characteristics that, while ordinary principles that l explain human behavior will still apply, the behavior itself will be different from behavior in non-nuclear emergencies.
l This proposition has some elements of truth. It is true, for example, that radiation is invisible, unlike tornadoes and hurricanes. It also may well be that public understanding of the risks associated with emergencies that involve nuclear power plants may differ from the risks associated with, for )
example, floods. The potential list of " differences" that could be developed is quite long. However long the list, it is not true that the " qualities" of the determinants of l
behavior in the human response equation are different for nuclear plant emergencies as opposed to other types. For
- 138 -
l l
a i
example, the " unique invisible character of the radiation hazard" is not unique in terms of explaining human response i
i in radiological emergencies. The " character of the hazard" has long been included in the catalog of factors taken into account to understand human behavior in emergencies. It is a well-established determinant of human response. On this factor, nuclear reactor accidents are something like a flash flood that occurs on a sunny day. If radiological emergencies are unique, it is only because the " values" or
" quantities" that the determinants of human response can take are unique -- or further from the mean -- and not because the list of relevant determinants is unique.
What is important for preparedness to ensure public i safety in emergencies of all sorts -- including anticipating l
the response of emergency workers -- is that the factors that determine behavior are known and transferable across i emergency types. Put simply, the "what," "how," "when,"
"where," and "why" of emergency preparedness are known, and planners can take into account variation in the " quantity" of each determinant across different types of emergencies to draft sound plans for response in nuclear power plant emergencies. All types of emergency workers -- including those who have unique jobs and those who do not, those who work in shifts and those who do not, etc. -- will be able to get the necessary emergency jobs done in a nuclear power
- 139 -
i l
k i
)
plant emergency if plans adequately address the important factors.
Once again, the key factors of determinants that, if adequately addressed in a plan, will lead to emergency workers performing their emergency jobs in a nuclear power plant emergency are (1) adequate training so that the workers clearly understand that they have an emergency job and what that job requires them to do and so that they develop an understanding of the importance of their jobs for overall community safety and to their work group; (2) enabling workers to understand what risks they are and are not taking while doing their work (for example, providing pre-emergency education about radiation, the use of dosimeters, or other ways to achieve this same objective); and (3) recognizing that workers will be concerned about the safety of their families and will want to be able to check on them and, thereby, feel comforted as they do their jobs.
- f. Workers' Fear for Their Own Safety The supposition that emergency workers would not do their emergency jobs because they would rather "save" themselves is unfounded. This is not to say that emergency workers ordinarily will remain in buildings that are, for example, under flood waters or demolished by an earthquake.
At the same time, the history of emergencies catalogues many i
acts of bravery where both trained and volunteer emergency workers have put themselves in extreme jeopardy in order to
- 140 -
help others. It would be foolhardy to plan for emergency workers to do work that would result in their death or other extreme personal consequences; but it is a fact that in an emergency some few would almost certainly step forward to volunteer for such duty.
Particular concern has been expressed that trained volunteer emergency workers who live or work outside the Emergency Planning Zone (EPZ) whose emergency job would require that they go into the EPZ could be less likely to perform their emergency jobs than those already in the EPZ.
The obvious thesis underlying this concern is that this subset of emergency workers would be reluctant to leave a
" safe" area to enter an " unsafe" area.
In general, emergency planning can take steps to prevent circumstances from arising in which emergency workers would fail to perform out of fear for their own safety. For example, they can take into account the factors that ensure that emergency workers do their emergency jobs, and they can take care to understand the circumstances under which it would be imprudent to assure that emergency workers would perform (even though some workers would undoubtedly perform even then).
The record is clear that emergency workers do their jobs when they understand that they have an emergency job to do, when they understand what that job requires of them, and when they have a sense of the importance of their job for overall
- 141 -
I t_--__--
community safety and to their work group. The understandings ,
i can be produced in different ways. For example, people who hold jobs that are in the routine of everyday life comparable to their emergency roles -- for example, policemen -- bring ;
these understandings to the emergency setting. But these understandings can also be developed, for those whose 1
emergency roles are different from their daily ones, for example, volunteer emergency workers who live and/or work outside an EPZ, through training. Adequate preparedness with emergency workers of the latter type is not impossible, but it is more demanding of those who would develop plans and be prepared.
Also, the behavior of trained emergency workers (be they full-timers or volunteers, inside or outside an area of risk when the emergency begins), like that of people generally, depends on their situational perception of " risk" or
" hazard." It is quite possible that an emergency worker would not perform a certain act if he believed that performing it would result in his own death. It is also the case that radiation is invisible (emergency workers cannot see the hazard) unlike, for example, fires, where people can see the location and intensity of the flames. This difference between radiation and fire does not, however, mean that emergency workers would assume the worst about the l 1
hazard, come to believe that it would result in their own death, and flee their posts or not enter an EPZ to protect
- 142 -
themselves. The reason is that emergency workers of all sorts can be trained and equipped to accurately assess the risk. Emergency workers can be trained about radiation (they need not be kept ignorant of it); emergency workers can be provided dosimeters (a way can be provided for them to "see" the hazard); and emergency workers are able to hear emergency information that will keep them informed about the situation.
An emergency at Seabrook Station, therefore, would not provide the context in which imaginations about an unknown risk could run wild, leading emergency workers to conclude that doomsday was upon them. The hazard would be both knowable and known. Emergency workers would therefore behave in rational and orderly ways (as they do in other emergencies) and perform their jobs not without regard for personal safety, but rather with information about it,
- g. Workers' Intentions Versus Actual Behavior No doubt emergency workers could be found who might profess that they would likely be more concerned about other roles (for example, family member) in a future emergency than they would their work role; and that they might envision I evacuating with their family before reporting for emergency work. In the prior section of this testimony, we explained why such role abandonment does not occur in actual emergencies. In this section, we will explain why such speculation about future behavior bears little weight.
l l Social scientists have long questioned the validity of
- 143 -
inferring behavior in the future from notions about how people think they would behave in a future circumstance, for example, in an emergency.
Since the time the first work was done on how well what people think or say their behavior will be and actual behavior related to one another (see, R.T. LaPierre,
" Attitudes Versus Actions," Social Forces, 13:230-237, 1934),
the great majority of investigators who have looked at the question have concluded that there is only a weak relationship, if any, between how people think they might behave and actual human behavior (see, A.W. Wicker,
" Attitudes versus Actions: The Relationship to Attitude
. Objects," Journal of Social Issues 25:44-78, 1969). A few scholars recently have suggested that the failure of this research to illustrate that anticipated behavior and behavior covary is not because there is not a theoretical relationship between the two, but rather because researchers have measured i the anticipated behavior and the behavior which they seek to compare in an inappropriate way. (See C.A. Kiesler and P.
Munson, " Attitudes and Opinions," Annual Review of Psychology 28:887-891, 1973; C.A. Kiesler, B. Collins and N. Miller, Attitude Change: A Critical Analysis of Theoretical Approaches, New York: Wiley, 1969). These more recent papers suggest that there is a relationship between anticipated behavior and behavior, but that it does not appear in the studies because researchers have not done quality research l
1
- 144 -
l
l I
and have not measured reliability, category width, change in attitude object, item difficulty, factor structure of attitude and behavioral syndromes, the effect of non-additional contributions to answers to questions about attitudes, and so on.
But while most social scientists are of the opinion that anticipated behavior (or intentions to behave) do play some sort of role in shaping behavior because it makes
" theoretical" sense, they do not think that anticipated behavior and behavior co-vary, based upon 50 years of empirical evidence to the contrary.
The minority of scholars who profess that behavioral intentions and behavior would eventually be shown to empirically co-vary if only researchers would stop doing bad research, still await empirical proof of their belief.
Interestingly, even these scholars begin with the admission that "a host of 'other factors' can influence a person's behavior to make it inconsistent with a reported attitude" (see, M. Rokeach and P. Kliejunas, " Behavior as a Function of Attitude Toward Object and Object Toward Attitude," Journal of Personality and Social Psychology, 1972; A.W. Wicker, "An Examination of the "Others Variables' Explanation of Attitude-Behavior Inconsistency," Journal of Personality and Social Psychology, 19:18-30, 1971). They claim that one might be able to increase the chances that a behavioral intention or an attitude will correspond in some empirical 1
- 145 -
l i
.way to behavior if,.perhaps, verbal measures were taken that tried to take these other factors into account. The latter approach, it is claimed, allows-the respondent to be able to "take the other non-attitudinal things into account" when coming up with the answer since "other factors" and the attitude or intended behavior are considered at once. Some who advocate this method of "let the respondent do the measuring of sets of factors" hope that this method might show co-variation of answers to questions with actual real-world behavior.
This method, and several varieties of it, go by many names. One name is " intentions-to-behave" (see, M. Fishbein and I. Ajzen, Belief, Attitudes and Behavior. Reading, Mass.: Addison-Wesley, 1975). The " behavioral intentions" method of attempting to find attitudinal / behavioral co-variation is straight forward in logic: "Ask someone if they would do something, and if they do not change their mind, then you will find that they may do what they said they would do." The method may work, but only for familiar behaviors l J'
(like buying peanut butter). Underlying the method is the supposition that the respondent can take all "other factors" l into account when coming up with a guessed-at future behavior. When the respondent does not know what the "other 1
factors" are that he might take into account, the result is a I wrong guess -- the result that has been found in studies over l
- 146 - l
o t .
I the last 50 years'-- and the answers to questions do not f(
correspond to actual behavior.
I Another set of researchers (M. Snyder and W.B. Swann, "When Actions Reflect Attitudes: Politics of Impression t Management," Journal of Personality and Social Psychology, 1976) go one step further that Fishbein and Ajzen. They suggest -- perhaps because the respondent may not know all the "other factors" to consider -- that the researcher provide the respondent with what are all the "other factors."
Yet another pair-of researchers (see, J.M. Piccolo and J.
Louvier, "Information Integration Theory Applied to Real-World Choice Behavior," Great Plains-Rocky Mountain Geographical Journal 6:49-63, 1977) go even further and suggest that respondents should be asked to engage in information-integration tasks that more closely resemble the real world. None of this minority of scholars claim that attitudes (regardless of what label -- including behavioral intentions -- is currently being used) and behavior covary.
This minority of scholars who maintain that answers to questions will someday covary with actual behavior persist in thinking that one can provide a way to allow a respondent to take all "other factors" into account before he guesses at his future behavior. "Other factors" can be taken into account by respondents on familiar behaviors for which the .
1 respondent knows what are all the other factors. This is why voting polls sometimes are used to accurately predict future
- 147 -
l
behavior. The "other. factors" typically have'their effect on I voting long before a pollster. calls or a' voter goes into a voting booth. Last minute changes-of-mind usually are random enough to cancel themselves out, and the prediction by-the pollster prevails. . But where the respondent is inexperienced with the behavior being asked about,.and has no notion of how relatively important all those "other factors" are (or even what they are), he cannot take them into account in predicting his future behavior. As a result, his prediction will be inaccurate.
The point'of this review of the research record regarding the relationship between attitudes and/or behavioral intentions and actual behavior was to illustrate that pre-emergency notions about what a person thinks they would do or not do in a future emergency (for example, evacuate with one's family versus do their emergency job) have little predictive value for actual future emergencies.
L
. Emergency situations, by definition, are quite different l-from the routines of everyday life. In addition, they are socially complex situations in which human behavior results from a social process. For example, the public behaves in
-ways determined by the social character of the communications of emergency information which they hear (it must be confirmed, for example). Emergency workers' own behavior is j i
affected by, for example, group cohesion (who they are with), )
and so on. Factors such as these -- which are the actual '
- 148 -
r.
- determinants of human emergency behavior. -- are not '51kely
' known to people before they enter an emergency situation.
They certainly cannot be part of people's own verbal predictions of future emergency behavior and are not based in the social reality of an actual emergency.
Behavioral-intentions by emergency workers during routine times will bear little weight in determining their actual behavior in an actual emergency. The factors reviewed in the prior section of this' testimony will be the actual determinants of the behavior of emergency workers in actual future emergencies; and these factors operate in emergencies in ways that lead emergency workers to.do their emergency jobs.
- 3. Voluntary Public Evacuation
- a. Introduction In recent years, an issue has been raised that emergency
! planning efforts for nuclear power plants take insufficient account of the number of evacuees that may leave the area despite the fact they are not " ordered" to evacuate. This issue seems to have surfaced in light of the evacuatien phenomena witnessed during the days following the 1979 Three Mile Island (TMI) accident. Available research indicates ,
l that some voluntary evacuation from among a population j 1
outside an official evacuation zone can be expected in emergencies. This circumstance would be especially-
[
anticipated in cases where emergency information syste.ms are
- 149 -
i
absent or poor, or the public risk.information itself is of low quality (for example, inconsistent, imprecise, untimely, etc.). The expectation for some voluntary evacuation should be considered in emergency planning for nuclear power plants.
In our judgment, a reasonable planning assumption would be to expect that up to 25% of the population living 10-20 miles from the boundary of an official evacuation zone would voluntarily evacuate if evacuation of the population within the official zone is advised.
The basis for why, in our judgment, this 0 to 25%
estimate of expected voluntary evacuation is cautious and conservative is described, in some detail, in the sections of this testimony which follow. Additionally, we also offer judgment about voluntary evacuation in an emergency planning zone (EPZ) that is only partially evacuated.
I
- b. Determinants of Public Behavior in Emergencies Empirical research on the behavior of publics in emergencies has been performed by social scientists for over three decades, addressing public behavior in response to climatological, geological and technological hazards. The l
conclusion resulting from well over one hundred empirical studies of actual public behavior in actual emergencies is consistent and clear regarding why members of the public behave as they do in emergencies. The following text provides a brief summary and overview of what has been learned.
- 150 -
I a
l Most members of the public in emergencies behave in ways relatively consistent with their situational perceptions of risk and what to do about it. These situational perceptions are largely formed during the emergency as it is being experienced. Hearing that an emergency is underway rarely leads to an immediate protective action response. Instead, hearing initiates a sequential process in most members of the public which is: hear-understand-believe-personalize-decide-respond. In simple terms, people do not have a knee-jerk response in emergencies. Response comes after people have formed an understanding of the situation, formed belief based on alternative information to which they are exposed, personalized or not personalized the risk, and made decisions about what to do, if anything. These situationally determined perceptions (what people come to understand, what people end up believing and their belief in emergency information, how they do or do not personalize risk, what they decide to do) are the result of two sets of factors and how they interact as the emergency is being experienced.
Research findings accumulated over the last 30 years from approximately 100 empirical studies when synthesized 3 clearly illustrate that two sets of factors influence public response to emergency warnings and/or information. These are receiver characteristics (which profile characteristics of people who receive emergency information and warnings and
- 151 -
l 1
)
i I
I include factors such as sex, age, pre-emergency fear of a 1
particular emergency), and sender characteristics (which {
)
include factors such as information clarity, consistency, frequency and many others). Additionally, emergency information and warnings, if well-configured regarding factors such as frequency, consistency and so on can largely overcome many of the constraints to forming sound situational perceptions of risk and then behavior that can be imposed by receiver characteristics. For example, holding information or sender characteristics constant, and at a less-than-desirable level, women are more likely to perceive risk to be higher than men and, consequently, are more likely to engage in protective actions than men; however, with
" good" emergency information both men and women are equally likely to form more accurate situational perceptions and, therefore, both sexes are equally likely to make more accurate decisions about how to behave during an emergency.
Interestingly, however, pre-emergency judgements about what behavior of the public is thought to be likely in a future unexperienced emergency (for example, like answers to questions about intended future emergency behavior ascertained through a poll or survey) would undoubtedly detect different answers from men and women. Were a future emergency to occur that was characterized by " poor" emergency information, actual response differences between men and women would be observed. Were a future emergency to occur
{
- 152 -
)
i l )
I i
i
that was characterized by " good" emergency information, little or no response differences would be observed between men and women. " Good" emergency information, therefore, can help overcome the constraints to sound emergency public response imposed by the heterogenous differences between members of any public as people in that public seek to form ;
situational perceptions of risk and then engage in a behavioral response to the situation being experienced.
- c. The Mix of Public Information in Emergencies A fact of emergency warning settings in an open society like the United States is that public information flows from more than the " official" warning system. Sources are varied; variation exists in viewpoints about the risk and what people should do about it; local as well as sometimes national media can become involved; warnings and information often reach the public from friends, neighbors and relatives as well as from
" official" sources; and so on. It is almost inevitable, ,
therefore, that the public will be exposed to different interpretations and inconsistent emergency information considering all the potential information sources involved (these range from " official" warnings to advice from intimates). This inevitability is further compounded by the tendency of the American news media to report "both sides" of a story, and the fact that different media organizations have different " styles" that can als'o carry into how they report on emergency situations (see, for example, Dan Nimmo, 1984.
- 153 -
l l
_ _ _ _ _ _ _ - - - _ _ _ _ - l
"TV Network: News Coverage of Three Mile Island: Reporting Disasters as Technological Fables," Mass Emergencies and Disasters 2:1:115-146; and Don M. Hartsough and Dennis S.
Mileti, 1985. "The Media in Disaster." 1Np. 282-294 in Jerri Laube and Shirley A. Murphy (Eds.) Perspectives on Disaster Recovery. Norwalk, Conn.: Appleton-Century-Crafts.
This means that the public in any emergency warning situation could and is often immersed in a sea of conflicting and contradictory information. This tendency does not mean that an emergency warning system cannot provide a public with consistent information on which that public can form its situational risk perception. Instead, it means that good public emergency warning system preparedness should presume
.that conflicting information will exist in the emergency warning setting, and then take steps to assure that
" official" warnings (presumed to be the best warnings) stand out and become the focus of the public rather than to have i the public focus on information of lesser quality with content somewhat randomly determined.
The elaborate body of public warning response research which has already been mentioned suggests how to help a public focus on " official" warnings in an emergency setting, as opposed to other information in the information " soup" in which they could likely be immersed. This is accomplished, i l
for the most part, by seeking to achieve the same attributes
- 154 - )
I i
e
1 l
i of the emergency information checklist presented within the next section of this testimony.
What all this means, therefore, is that the most important aspect of an emergency warning system for a nuclear power plant becomes the Emergency Broadcast System (EBS) messages because it is here that warning system preparedness can seek to achieve the type of emergency information characteristics for the " official" public warning is needed in order to have these warnings and this information " stand out" for the public.
This reduces to a status of lesser importance the role of secondary and tertiary public information sources, for example, press releases, emergency news center (ENC)'
l presentations, media interviews, rumor control, and other l
vehicles for public information. This does not suggest that these secondary and tertiary public information sources are not relatively important, nor that they should not be part 1
of emergency planning and operations. Instead, we only mean to suggest that they are of lesser importance when " good" emergency planning has taken the necessary steps to focus public attention on the EBS system on the somewhat accurate presumption that other public information sources (for example, friends, relatives, editorials, parking lot interviews with " experts" and so on) could likely contain misinformation or at least alternative contradictory information.
- 155 -
We have provided all this background in order'to explain why it is the case that public response to a future emergency (should there ever be one) at Seabrook would largely be a function of the information characteristics of the EBS system. Criteria for evaluating the information characteristics of EBS messages are provided in the next section followed by our evaluation of EBS messages designed for the Seabrook emergency' plan.
- c. Evaluation of Seabrook EBS Messages.
- 1. Evaluation Criteria The research record, when synthesized, provides a checklist of.what constitutes the type of emergency public information that helps most members of a public perceive risk more accurately in an emergency. That is, the following ideal emergency information characteristics can help men and women,-young and old, people with high pre-emergency fear of a particular hazard, as well as low pre-emergency fear perceptions, and so on, all form more realistic situational risk perceptions and make response decisisons consistent with those situational perceptions. This checklist can be briefly summarized as follows. Whenever possible, emergency information should be: (1) message attributes: specific, consistent, accurate, certain and clear in reference to the risk, its location, what people should do and how much time they have before they should do it; (2) channel attributes:
distributed over diverse channels of communication; (3)
- 156 -
)
i frequency attributes: repeated frequently and in a clear l pattern; and (4) source attributes: the source of the '
information is best if it can be a set of people with organizational affiliations to enhance officialness, credibility (scientists and engineers usually have the most),
and familiarity.
- 2. Evaluation Results ;
The messages, channel, frequency and source attributes of emergency public information and warnings listed in the last section of this testimony constitute an " ideal type."
That is, these emergency information characteristics are those which represent, collectively, all that has been empirically documented to enhance accurate public situational perceptions and subsequent behavior in emergencies. We have read each of the Seabrook EBS messages contained in Appendix G, Volume 4, Rev. 2 (8/86) with these criteria in mind, and find that these EBS messages in many aspects advocate the
" ideal type" characteristics already presented.
A few examples will serve to illustrate how these EBS messages adhere in many ways to the characteristics of the
" ideal type." Risk location is specific, certain and clear, for example in Message G, since the specific towns for which evacuation is advised are names; additionally, people are I informed that there is no reason to evacuate or take any action if they do not live in those named towns.
t Additionally, guidance is, in several ways, certain, plear I I i
- 157 -
1
r .,
l i
I and specific concerning evacuation, and the time factor is as well: " evacuate as soon as possible." The frequency elements )
1 of an emergency public warning message is outstanding. The i
EBS system will repeat messages every fifteen minutes. This i l
is frequent and in a predictable pattern. This pattern, l l
"until new information is available," would also help members of the public focus upon the EBS system and its messages, as opposed to other random "non-official" communications.
Also, the messages are, overall, internally consistent.
For example, people at risk are not told: a release of radiation has occurred, your town is at risk but don't worry.
Instead people are told what to do about the risk The criteria we would use to judge emergency public information like these EBS messages amount to about two dozen factors,to consider. These messages are outstanding on many but not all of these factors. For example, the source of emergency public emergency warnings is important to helping the public form sound perceptions during an emergency and, in turn, making sound decisions about what to do or not do i,n response to the situation. The source of the information contained in these messages is not clear. Warning sources are best when they can be attributed to multiple persons and/or organizations that would include, for example, government officials, utility officials, scientists and/or engineers.
Also, for example, the language within the EBS message j might provide the public with more active and directed
- 158 -
s
'A l
c ; n i
a, s
courses of action. Rather than passively suggesting that l 4 more information is available in pre-distributed public I
information brochures, the messages '
wight more actively '
a, .\
direbtpersonstoget,theirbrochursk,andexplainwhythey7,j-V J l need them. ;
The'EBS mesedles found within the NHRERP, Revision 2 are m ,
currently undergoing review in l'ignt of the above ' ,'
q .
characteristics,vhich may be viewed as indicative of the t
" ideal" message type. This review is being undertaken in
( ,
ordertohighligdg.andunderncoretheseparticularNttributes j ,
i t 1 which the EBS messages now visflect to some degree. Once ' '
honed to more closely follow these ideal, prototypic o
attributes, the EBS messrdes for Seabrook Station will providethelis{.hninglpubicamostsolidfootingfromwhich to base sound decis4.*>n-making.
- d. Volun,tAfv Evacuation in E AISLRD!difE F
e .,
i Voluntary evacuation by persons not. advised to evacuato 1 has not been explicity studied as a phenomenon se'parate from evacuation in general. .Volvntary Evacuation as a phenomenon which may or may .,not be applicable to emergency planning at a nuclear power plant could take two forms.
The first of these we refer to as concentric voluntary evacuation. This form of voluntary evacuation is theoretically possible in the concentric ring which surrounds
'/
l l/. -
the area in which kvacuation has been advised in cases in l \f '
which the area is a circle; for example, the i
- 159 - >
.t
\
f 1
i l
- J
H
.s A
9,,
concentric ' ring outside. sthe 10-mile Emergency Planning Zone
, s (EPZ) in a scenario in which evacuation of the entire EPZ iq advised (see-Figure 1). ,
The second form of voluntary evacuation we refer to as keyhola voluntary evacuation. This form of voluntary
. evacuation is theoretically possible in. scenario in which evacuation is advised for only a sub-population of an EPZ.
In such a scenario voluntary evacuation could occur in those
,i.
portions of the EPZ not advised to engage in protective actions. We refer'to'this as keyhole voluntary evacuation only since, in its most prototypical form, it would resemble the following scenariod: advised evacuation would occur for everyone within two miles of the plant plus a subsegment of the EPZ beyond that two. nile ring'in the downwind direction
?
(see Figure 1 of thin ' testimony) .
Two evacuati n cases exist which provide the most detailed data regarding concentric and keyhole voluntary evacuation. Both of these cases involved advised evacuation and provide evidence of voluntary evacuation. These evacuations occurred in 1979 as a result of accidents at the Three Mile Island (IMI) nuclear power plant, and a hazardoun materials release and explosion resulting from a train derailment at Mississauga, Ontario.
These specific cases, along with general knowledge regarding the determinants of public evacuation behavior, were used as a basis to inform our estimates of the level of i
- 160 -
4 l
l 1
. voluntary evacuation (concentric and keyhole) in future nuclear power plant emergencies in the United States. Also j 4
taken into account was the current regulatory context for b
nuclear power plant emergency p'.anning in the U.S., which requires rigorous planning for emergencies.
]
l'. TMI and Concentric Voluntary Evacuation. l l
)
In order to examine the issue of concentric voluntary evacuation to estimate its likelihood and-expected incidence i'
at Seabrook, the extensive body of research and information regarding evacuation behavior in the aftermath of the l
accident at TMI were reviewed. This experience is a l I
particularly useful event for this purpose for several j j
reasons. First, the event occurred in response to a nuclear i power plant accident. Second, there is an extensive body of available literature, including several studies of evacuation behavior. Third (and to a lesser extent analogous, as will !
be described later in this section), there was never issued
)
an order or official advisory that the general population should evacuate. Consequently, many researchers interpret t
all evacuation that occurred at TMI (except for pregnant I women and pre-school children) as voluntary evacuation.
During the emergency period, no public evacuation was ordared. However, the following advisories were issued by s
the Governor. At a mid-morning news briefing on Friday, a prepared statement from the Governor advised all persons i living within a 10-mile radius of TMI to remain indoors with 1 .
L
-. - 161 -
l ,
e r
i
[. '
'their-windows closed. Later that day the Governor issued another advisory that pregnant women and pre-school children within a 5-mile radius of the plant should. leave the area..
On Saturday the Governor announced that his advisory regarding' pregnant women and pre-school children would be in effect for_at least another night.
Aside from the Governor's advisory for' pregnant women and pre-school. children to leave, two other public announcements related to evacuation were made. During a radio broadcast on Friday a local emergency preparedness director _ mentioned an evacuation order may be forthcoming shortly. . There were several reports of the detailed
. evacuation planning that was taking place and official recommendations for individuals to make preparations for a possible evacuation.
a) . Estimates of Evacuation at TMI.
It has been estimated that approximately 150,000 persons left the area _ surrounding the Three Mile Island power plant (cf. Robert A. Stallings. 1984 " Evacuation Behavior at Three Mile Island" International Journal of Mass Emergencies and Disasters 2:1:11-26). Of those who evacuated, all but four percent were located within 15 miles of the plant; within 15
- miles;of the plant, 39% of the population or 144,000 people were estimated to have evacuated (cf. C.B. Flynn and J.A.
Chalmers, 1980, The Social and Economic Effects of the Accident at Three Mile Island. Washington, D.C.: U.S.
l
- 162 - i l
l- .
- s. T1- -
A
__m.._om... __m_, __,, . .
l Nuclear Regulatory Commission, p. 22). In an effort to address voluntary evacuation and the reasons for it, it is useful to focus on the distribution of evacuated population by distance from the power plant, itself. -Table 1 provides such data for evacuees from within 15 miles of the
. plant.
TABLE 1 PERCENT EVACUATION BY DISTANCE FROM TMI
' Distance Data Source From TMI (miles) Al A2 B C 0-5 60 66 NA 48 0-6 NA NA 55 NA 0-12 NA NA 53 NA 0-15 39 41 NA 35a 5-10 44 49 NA 45 10-15 32 33 NA 27 5-15 37a 39a NA 34a Al Flynn and Chalmers, 190, p. 22 (population)
A2 Flynn and Chalmers, 1980, p. 47 (household)
B Brunn et at, 1979, p. 47 (household)
C Cutter and Barnes, 1982, p. 120 (household) a Flynn and Chalmers, 1980, p. 10 household estimates used for reanalysis of data NA Not Available Evacuation estimates from all studies were provided for geographic areas that represented concentric rings centered on the TMI plant site. Estimates were given for varying numbers of miles distant from the plant. In the area closest to the plant (0-5 miles) household evacuation estimates ranges from 48 to 66 percent. From the area 0-6 miles, 55%
were estimated to have evacuated.
- 163 -
{
i i
l
)
The overriding finding from these data is that voluntary evacuation coincided with distance from the TMI plant. site.
In all, 35-39% of the. population within 15 miles of TMI i
evacuated (cf. Susan Cutter and Kent Barnes, 1982,
" Evacuation Behavior and Three Mile Island," Disasters 6:2,
- p. 120; and Flynn and Chalmers, p.22). Of the total number of households that'left, 28% contained pre-school children (Cutter and Barnes, 1982, p. 118). The particular concern for children is a.1so revealed by the study of Brunn et al (cf. Stanley D. bronn, James Johnson Jr., and Donald J.
Zeigler, 1979, Final Report on a Social Survey of the Three Mile Island Area Residents, East Lansing, Michigan:
Department of Geography, Michigan State University, p. 54) that found those with children under five remained out of the area.for a longer period of time. For evacuees from within 3 miles of TMI 58% stayed away 3-6 days and none of these households contained children under five years. Forty-two percent of those evacuated from this area stayed away 9-13 days. Of those who stayed away the longest time, all but one family had children under five years (Brunn et al 1979, p.
54),
b) Data Equivalency: TMI and Seabrook.
l For purposes of estimating incidence of voluntary 1 evacuation at Seabrook based on estimated evacuation rates at J
TMI, it is necessary to lay some groundwork in order to focus the examination on the TMI data set that would be most
- 164 -
equivalent to the area of interest for voluntary concentric 1
evacuation incidence estimates at Seabrook. j i
Current regulations guiding emergency planning efforts 1 at Seabrook call for planning attention to the area that has l
a 10 mile radius from the plant site. Known as the Emergency Planning Zone (EPZ), this area considers the population that is located within a 10-mile circle around the plant.
It is in the ring around this circle (from miles 10-20 from the plant site) that is the study area of interest for voluntary concentric evacuation. For planning purposes, the EPZ would be subject of an official evacuation advisement, with 100% population evacuation advisory possible.
Population in the area beyond 20 niles from the plant site is not expected to voluntarily evacuate in appreciable numbers, so for emergency evacuation purposes the area can be ignored, This expectation is readily supported by the data from TMI studies that estimated evacuation beyond 15 miles as l
4 and 9 percent (Flynn and Chalmers, 1980, p. 22 and Brunn, et al, 1979, p. 47, respectively).
i In the case of the TMI experience, we have conservatively assumed that the evacuation area that is 0-5 miles from the plant is equivalent to Seabrook's EPZ of 0-10 miles. Similarly, we have concentrated our data examination .
on the evacuation rates provided in TMI studies for the ring shaped area that is 10 miles beyond the 0-5 area (in that case, from miles 5-15).
- 165 -
l i
l i
I Hence,'we have examined the TMI data as equivalent to '
Seabrook by using the following construct.
l Study Area for Concentric Voluntary Evacuation TMI 5-15 miles J Seabrook 10-20 miles )
l This is a conservative construct, mainly because of the ?
expected decay effects of voluntary evacuation with distance from the plant site. The Seabrook study area is literally 5 miles further away from the potential source of threat.
Furthermore, and more importantly, the equivalency construct is conservative because at TMI the Governor issued an advisory that persons within 0-10 miles of TMI should shelter (i.e., stay indoors with windows closed). Consequently, we expect that the TMI evacuation rate for 5-10 miles is larger than would be expected for a more strictly voluntary evacuation effect at Seabrook.
One further note about equivalency of the TMI data as a basis for estimating the incidence of concentric voluntary evacuation at Seabrook. The data obtained from survey instruments in the two mailed questionnaire surveys resulted from survey response rates that were quite low, as shown:
Mailed Survey Response Number of Data Source Rate Respondents Brunn et al, 1979,
- p. 23 56 150 Cutter and Barnes, 1982,
- p. 116 39 359
- 166 -
In this case, we believe the consequence of these low l l
response rates is that evacuation rates are overreported, I l
1.e., persons who took an active role in the emergency by f
actually evacuating should be expected to have been more likely to have responded to a mailed questionnaire that asked about evacuation in the TMI accident. Unfortunately, we have I
no way to specifically quantify this effect on reported '
evacuation rates. We expect the effect would be that {
evacuation rate data overestimate actual evacuation.
The following section of this testimony describes the bases on which we. estimate concentric voluntary evacuation at Seabrook from the date reporting evacuation rates in the ring-shaped area that resided 5-15 miles from TMI. From Table 1, the percent of households that evacuated are reported to have been 39% and 34%. We will examine the conditions that influenced these evacuation rates as a basis for estimating whether or not and to what extent voluntary evacuation can be expected in the ring-shaped zone that is 10-20 miles from the Seabrook plant site.
c) Estimating Concentric Voluntary Evacuation in a Future Accident at the Seabrook Station.
Two reasons loom forth to suggest that the amount of 10 to 20 mile concentric voluntary evacuation to be expected in an accident at Seabrook could not likely begin to approximate the 34 to 39% equivalent concentric voluntary evacuation to observed during the accident at Three Mile Island. Both
- 167 -
- {..
K ,
' factors focus on the character of emergency public g information during TMI versus the emergency public information that' would characterize an emergency at Seabrook.
The first of these reasons is that voluntary evacuation during TMI has and continues to'be unquestioningly and unfairly computed on the' basis of only one public advisory instead of all public information during the accident. The
.second reason is the contradictory and confusing nature of
' emergency information that characterized the TMI emergency versus what would occur in an emergency at Seabrook.
- 1) The Basis-for Defining Concentric Voluntary Evacuation During TMI. Voluntary' evacuation in the
. concentric ring surrounding TMI has always been couputed on the basis of the Governor's advisory. That advisory defined the evacuation area as the 5-mile circle surrounding the plant; thereby setting the standard that evacuees living greater than 5 miles from the plant would automatically be categorized as concentric voluntary evacuees. However, this approach'to defining concentric voluntary evacuation is grossly inaccurate and unfair. There are many reasons why, but the most significant reason is a simple one. The evacuees during the TMI accident were not responding to only that information which was contained in the Governor's advisory, and other significant information was the likely cause of the observed 34-39% concentric voluntary evacuation.
Cutter and Barnes (cf. Susan Cutter and Kent Barnes, 1982,
- 168 -
" Evacuation Bahavior and the Three Mile Island" Disaster 6, 2:116-124) have summarized this fact well on page 19 of their article.
Extensions of the emergency evacuation zone from 5 to 20 miles also contributed to the confusion. Emergency management officials had been assured by the NRC that their 5-mile plans were sufficient, but on Friday morning, 30th March (day 3), the NRC told state and local officials to extend the radius to 10 miles; later they ordered the zone extended to 20 miles from the plant ... Within hours the number of potential evacuees rose from 27,000 to 700,000, and this caught emergency management officials without satisfactory emergency response plans.
It is unfair, therefore, to judge voluntary evacuation during TMI in light of one public announcement alone. In fact, many of the evacuees during the TMI accident perceived the zone of likely evacuation advisement to larger than the 5 miles referenced in the Governor's advisory. For example, among all evacuees approximately 75 percent cited "to avoid forced evacuation" as one reason for their leaving the area (cf. Peter S. Houts et al, 1984, "The Protective Action Decision Model Applied to Evacuation During the Three Mile Island Crisis" Mass Emergencies and Disasters 2:1, at p. 36).
The emergency information during TMI was not at all consistent in terms of the area at risks. On one hand it was the 5-mile zone, on another it was the 10-mile zone, and on
- 169 -
)
1 yet another it was the 20-mile zone. In our opinion, it is methodologically unsound to focus solely on one piece of 4 l
information as a standard by which to measure concentric voluntary evacuation since more than one standard operated during the accident.
ii) The Contradictory and Confusing Character of Emergency Public Information During TMI. Most if not all researchers who studied evacuation at TMI have concluded that one of the most profound reasons why people evacuated (including the 34-39% concentric voluntary evacuees) was because of the confused emergency information made public during the accident. For example, Houts et al, op cit, p.
. 36) concluded that 80 percent of those who evacuated said that confusing information was one factor in their decision to leave. Brunn, Johnson Jr. and Zeigler (cf. Stanley D.
Brunn, James Johnson Jr. and Donald J. Zeigler, 1979, Final Report on a Social Survey of Three Mile Island Area Residents, Michigan State University: Department of Geography, p. 49) concluded that "Everyone who evacuated indicated one of these two reasons (safety or conflicting reports) and 61% indicated both." In fact, Dr. Zeigler along with his two colleagues concludes that both conflicting information and talk by government of other evacuation alternatives were major reasons for the voluntary evacuation observed during TMI.
The shadow cast by the evacuation order was intense j and widespread in the case of TMI for two basic reasons:
- 170 -
8 l
First,'the peculiar nature of a nuclear disaster makes I it impossible for people to evaulate the danger to which they are exposed which, when combined with conflicting information from all sources, leads them to add a geographical margin of safety to the government's l
proclaimed radius of-danger. Second, serious talk of a full scale evacuation continued ever after a limited evacuation had been ordered. This discussion prompted more families to leave than if the government had acted confidently about the outcome of the accident, dropped talk of any further evacuation measures, and offered a strong rationale for the extent of the partial evacuation which was ordered (Brunn, Johnson Jr., and 1 Zeigler, op cit., p. 49). (emphasis added)
The conclusion that the confused character of public emergency information was a profound determinant of evacuation observed during TMI was even reached by the Nuclear Regulatory Commission's (NRC) special inquiry into the accident. The Regovin Special Inquiry was instituted by the NEC within weeks of the March 28, 1979 accident to determine what happened and why, to assess the actions of utility and NRC personnel before and during the accident, and to identify deficiencies in the system and areas where further investigation might be warranted. The inquiry resulted in a two volume report Three Mile Island, A Report to the Commissioners and to the Public, and it was issued by the Rogovin Special Inquiry Group in January of 1980.
(NUREG/CR-1250). The Rogovin Inquiry (Vol. I at ix) concluded that the character of the information available to the public on the accident was a major determinant of evacuation behavior and the public's perception of danger from TMI-2.
- 171 -
1 l
To a considerable extent, information available through the media was confusing and frequently conflicting. Met
'Ed, the basic source of information on the status of the
. . plant, quickly lost' credibility. The NRC was a source
( of contradictory and, upon several occasions, alarming information. Whereas the local-media tended to be restrained.and'non-speculative in its coverage of the accident, the national media and the media outside'the area-tended to be more speculative. (Rogovin inquiry,.
Vo..II, Part 2 at p. 620; see also Rogovin Inquiry, Vol.
)
l II, Part 3 at 1073). i
- Consequently then, the TMI accident may be evaluated as one of the worst public emergency information events in our nation's history. Appended to this testimony as Appendix One is a brief thumb-nail sketch of the contradictory and conflicting character of some of the emergency information made public during the accident.
Without reviewing the complete' chronology of the accident, a summary of the key statements made to the public during'the emergency period follows. It makes apparent the causes of the confusion and uncertainty and resultant evacuation by members of the public during TMI.
The first public awareness of the TMI-2 accident was at I
8:25 a.m., Wednesday, March 28, at which time a local radio station announced an accident at TMI. At 9 a.m., an AP national news bulletin was issued. A survey to the TMI-area l population.(25 mile radius) by the Michigan State University Department of Geography shortly after the accident (Brunn et al., op cit.) indicated that 35% of the respondents first heard of the accident on Wednesday morning; 62% had heard by i
g
- 172 -
)
)
Wednesday night, and all had heard by Friday, March 30 (including 17% who had not heard until the 30th).
During the morning of March 28, because of contradictory ,
statements about radiation releases attributable to the delay in inf.ormation received by Met Ed officials who ware issuing i statements from various locales, and because of lace of an explanation of the general emergency status, the Kemeny Commission concluded that the utility's credibility, at least in the eyes of several reporters and public officials, was damaged beyond repair (Kemeny, Report of the Public's Right to Information Task Force, p. 97). The confluence of the perceived lack of utility credibility with the seeming inconsistency between information reports issued by the Governor of Pennsylvania's Office and Met Ed immediately produced uncertainty if not apprehension in the public's mind. This situation culminated in Lieutenant Governor Scranton's statement, at a 4:30 p.m. press conference on March 28, based on a misunderstanding of what was going on at TMI, that " Metropolitan Edison has given you and us conflicting information . . . There has been a release of radioactivity into the environment. The magnitude of the release is still being determined, but there is no evidence yet that it has resulted in dangerous levels." (Kemeny, Report of the Public's Right to Information Task Force, p.
100.) Further distrust, confusion and apprehension in the public's mind was created by inconsistent reports from the
- 173 -
l l
NRC and from the Governor's office during the evening of March 28 concerning the source of low levels of radiation measured offsite.
l l On March 29, the public was informed of discharges by l Metropolitan Edison of industrial waste water into the 1
Susquehanna River. The' releases were necessitated by the buildup of waste water on site ordinarily routinely released.
Although not harmful, the releases did contain small concentrations of noble gases (below maximum permissible concentrations). Nevertheless, public announcement of the discharges probably misled and alarmed some individuals, who mistakenly believed highly contaminated water was being
. dumped into the river. On this day local farmers and the Department of Agriculture began to test milk for radioactivity and outdoor feeding of cows stopped. This measure, while perfectly reasonable, was probably anxiety-provoking for the public. As a result of these actions, milk sales dropped. Also occurring on March 29 was Dr. Ernest Sternglass' alarming radio interview during which Dr. Sternglass recommended evacuation of pregnant women and preschoolers, a recommendation that the radio announcer mistakenly called an order to evacuate. (Kemeny, Report of the Public's Right to Information Task Force, p. 223.) While this statement was denied and corrected by a representative of the Pennsylvania Department of Health during the same
- 174 -
cr.-
1 afternoon, minimally, the misinformation created confusion and' uncertainty among the local citizenry.
On the morning.of Friday, March 30, a helicopter hovering over the TMI-2 stack measured a.1,200 millirem release spike as a result of the transfer of noncondensible gases between primary coolant system makeup tanks. This measurement was reported to the NRC, the State and the press, dramatically-altering perception of the seriousness of the accident. The director of emergency preparedness for Dauphin County went on the radio announcing the possibility of protective evacuation. NRC headquarters misinterpreted the information about the release, initially believing it represented a radiation level monitored at the site boundary, and consequently advised the Governor of Pennsylvania that people within five miles of the plant in the downwind direction should be told to stay indoors. The Governor made this announcement.in a live broadcast at 10:25 a.m., altering Chairman Hendrie's recommendation and suggesting that people within a 10-mile radius stay indoors and keep their windows closed. The stress produced by this information was probably exacerbated by a Met Ed press conference half an hour later at which, in response to a question, a Met Ed representative stated that he had not heard about, although he did not dispute, the 1,200 millirem reading.
- 175 -
y
.]
At about'12:30 p.m., Governor Thornburgh, after conferring again with Chairman Hendrie, held a press conference at which he announced that he had been in contact i
.with the President of the United States and with the Chairman of the NRC. He stated that the President concurred in the ,
Governor's' views that there was no reason for panic or implementation of emergency measures, and that Harold Denton would be dispatched to the site as.the President's personal representative. The Governor also announced that, based on the advice of the NRC Chairman and in the interest of taking every precaution, he was advising that pregnant women knd preschool-age children within five miles of the plant leave .!
the area. He also ordered the closing of schools within the 5-mile radius.
Contemporaneous with the evacuation advisory, an.NRC staff member, in response to a reporter's question, suggested that, although there was no imminent possibility of a meltdown, it could happen if conditions worsened. This highly alarming information was carried as the lead in many national stories, often neither qualified nor put in its proper context. Not until 6:30 p.m. did.the NRC publicly state ~that there was no imminent danger of a core melt --
reassuring but, probably, confusing information. In an 8:30 p.m. press conference, Governor Thornburgh and Harold Denton indicated that an evacuation order was not necessary; however, the earlier advisory would remain in effect at least
- 176 -
until Saturday -- also reassuring, but not anxiety-preventing information. (Rogovin Inquiry, Vol. II, Part 3,
- p. 1064-65.)
The issue of core melt was precipitated largely by the fact that a steam bubble inside the reactor' led some NRC officials, erroneously, to believe that the bubble could l explode. A news release was issued by UPI at 4 p.m. on Friday, March 30, reporting a senior NRC staff official's concern about the bubble and "the ultimate risk of a meltdown." (Kemeny, Report of the Public's Right to Information Task Force, p. 145-146.) As a result of continued confusion among NRC staff as to the significance of the reactor vessel hydrogen bubble, an NRC-confirmed AP story was released at 9 p.m. on Saturday which stated the " Federal officials said Saturday nisht that the gas bubble inside the crippled nuclear reactor at Thrwa Mile Island is showing signs of becoming potentially explosive . . . One NRC source, who asked not to be identified, said that critical point could be reached within two days." (Kemeny, Report of the Public's Right to Information Task Force, p. 153.) According to the Rogovin Special Inquiry, the AP story created a "near panic" in the Harrisburg area. State Police officers were deluged with calls, as was the NRC. (Rogovin Inquiry, Vol.
II, Part 3, p. 1067-68.) Later that evening, Met Ed, the Governor's office and Harold Denton publicly disagreed with the views of the NRC in Washington, describing the explosion
- 177 -
reports as erroneous or distorted. Denton stated that there l
l was no near term possiblity of a hydrogen explosion in either the containment or the reactor vessel. Nevertheless, one poignant indicator of the resultant level of public anxiety was the local churches' grant of general absolution of sins to the citizenry. (Kent Barnes, et al., Response of Impacted Populations to the Three Mile Island Nuclear Reactor Accident: An Initial Assessment, October 1979, Rutgers University Department of Geography, Discussion Paper, p. 6.)
On Sunday, April 1, Governor Thornburgh issued a press release which was reassuring in that it directed State offices to open as usual on Monday, April 2; however, the evacuation advisory was not rescinded, and it was recommended that schools within five miles of the plant remain closed until further notice. By Monday, April 2, with the news conference by Denton announcing that the bubble had reduced in size sufficiently to no longer be a problem, the crisis at TMI was over. However, the public was not told that the NRC had been wrong in its assumptions about the hydrogen bubble.
(Denton simply said the Staff had been very conservative in ,
their calculations.) (Rogovin Inquiry, Vol. II, Part 3, p.
1068.)
In summary, then, it is apparent from a simple chronology of the THI-2 accident that misinformation, inconsistencies, and confusion in information flowing to the public not only directly precipitated alarm and associated
- 178 -
distress, particularly for those who were identified as being especially vulnerable to radiation, but also produced the circumstances which were most distressful and contributed to evacuation. Specifically, the misinformation about a high level of radiation at the site boundary, the possibility that highly contaminated water was released into the Susquehanna River, and faulty information about an explosive hydrogen bubble were the most alarming pieces of information communicated to the public. The other primary stress-producing factor was the evacuation advisory, which was the result of misinformation.
One TMI evacuation study (cf. Peter C. Houts, Robert W.
Miller, George K. Tokuhata and Kim Skik Ham, 1900, Health
~
Related Behavioral Impact of the Three Mile Island Nuclear Incident, Parts I and II. Report submitted to the TMI Advisory Panel on Health Related Studies of the Pennsylvania Department of Health: The Pennsylvania State University College of Medicine, Part I, pp. 4-5) reports on the findings of in-depth interviews and emphasizes the importance of the role that incorrect and confusing information had on determining public evacuation. One respondent, for example, stated that she left because of "that bubble." Another individual stated, "Everything was so conflicting in the news reports. You'd hear one (local) report (saying) there's nothing to worry about; you hear the national (news), the i
! place is blowing up." (Houts, Part I, p. 4.) The Houts et
)
- 179 -
L__ ____ ___ _
al., Report reached the following conclusions about ;
evacuation:
By and large, these -evacuation- decisions seem to have been made in a'gonerally reasonable way. The absence of
' reliable information did not lead to mass panic, nor to a widespread denial that any threat existed. Those who left did so primarily because of a desire to protect family members who seemed to be unusually vulnerable (e.g., pregnant women and small children), and because of a general feeling that,.given the lack of information, a conservative approach would probably be best. Those who stayed were aware of the possibility of danger, but seem to have concluded that the possiblity of genuine harm was outweighed by the costs of leaving and the benefits of staying (Part I, Chapter 4, p. 10).
'Houts et al. (Part II, Chapter 7, p. 2) also found the
-following generalizations relevant for public policy purposes:
- 1. Information conveyed by the media was the main basis for decision to evacuate. Two aspects of this information were especially important in the decision'to evacuate:
- a. information about possible danger
- b. inconsistencies and/or contradictions in information.
- 2. There was considerable variability among residents in how information was interpreted resulting in:
- a. evacuation and returning being spread over a considerable period of time which contributed to the efficiency with which the large number of people evacuated the area
- b. a substantial percentage of the population electing not to evacuate at all.
In summary, one can conclude with reasonable certainty
.that the evacuation performed by TMI residents during the TMI-2 accident was based largely on the nature and quality of
- 180 -
L
the information relayed to the public during the accident.
Much of this information was faulty. In fact, the most alarming reports were based on misinformation. Exacerbating the already stressful situation was the fact that the public frequently heard conflicting reports from the same or l
l different sources that ordinarily would be relied upon for accurate and consistent information.
iii) conclusion concerning concentric voluntary Evacuation. Concentric voluntary evacuation during TMI has been estimated at 34 to 39% by different researchers in reference to the logically equivalent 10-20 mile concentric ring around the EPZ for Seabrook Station. It is inconceivable that such voluntary evacuation rates could occur were there to be an evacuation of the Seabrook EPZ.
First, the hypothesized Seabrook evacuation would not likely contain different information regarding larger evacuation zones than the 10-mile EPZ. This was the case during TMI.
Second, the existence of plans to facilitate the existence of consistent and nonconfusing public emergency information is the hypothesized Seabrook evacuation (including, for example, an EBS system, appropriate EBS message, an emergency news center, and so on) insure that the character of public emergency information would be sound and not even remotely resemble the poor quality information made public during the TMI accident.
- 181 -
On-these grounds it is inconceivable that concentric voluntary evacuation during an emergecy at the Seabrook Station would come close to that which occurred during TMI.
However,:it is equally. inconceivable that.it would be zero.
In our judgment, and with the quality of emergency'public
' information promised by the emergency planning-at the Seabrook Station, estimating a 0-25% concentric voluntary evacuation in the-10-20 mile zone, we would expect concentric-voluntary evacuation to vary inversely with distance from the plant.
- 2. Mississauga and Keyhole. Voluntary Evacuation.
One of the largest public evacuations in the history of North. America was successfully conducted in 1979 in-4 Mississauga, Ontario, Canada.
Estimates indicate that approximately 226,000 residents were systematically evacuated in the aftermath of a train derailment at about midnight on N'ovember 10, 1979. Of the 106 train cars derailed, 38 contained hazardous materials, including chlorine, liquid petroleum products and caustic soda.
Within minutes of the derailment an initial "BLEVE" (Boiling Liquid Expanding Vapor Explosion) occurred, starting
- a series of explosions. The area in the immediate vicinity of the explosions was evacuated starting less'bhan two hours after the derailment. During the ensuing 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br />, changing risk assessments by officials resulted in a series of staged 1
- 182 -
evacuations that eventually encompassed the entire city of Mississauga. An estimated 98% of the population of the official evacuation zone left the area.
Evacuation orders were given and evacuation occurred in phases that were sequenced by 15 official evacuation zones over a 21-hour period. The differential timing of evacuation orders coincided with changes in riak assessments made by the officials-in-charge that were made in response to events onsite (new leaks, new explosions, etc.) and ever-variable wind direction. Evacuation orders were issued and implemented in accordance with the following timing:
Zotte Egriod of Day 1-7 early morning 8-11 late morning to early afternoon 12-15 late afternoon to early evening The majority of residents left their homes after 9 a.m.
on Sunday.
a) Estimates of Evacuation at Mississauga.
A comprehensive behavioral study of this evacuation was performed within the months following this event (21. Ian Burton et al., 1981, The Mississauga Evacuation, Toronto:
University of Toronto, the Institute for Environmental Studies). The following observations regarding this evacuation are derived from an analysis of data collected for the Burton study. Two data sets detail the evacuation
- 183 -
i i
experiences of persons living in the official evacuation
. z o n e .- A third data set provides information from residents of three areas outside the official evacuation zone. This latter data set provides a particularly good basis for examining the influences on keyhole voluntary evacuation in the Mississauga emergency.
i In the case of Mississauga, a specific evacuation zone j was defined. The evacuation boundary ran down the center of a roadway. In addition, evacuation areas were identified by officials'in a serial fashion over a 24-hour period. The direction of evacuation progressed steadily toward what' ultimately became1the final perimeter area boundary.
Throughout the evacuation experience, some persons who lived across the road from one of the ultimate boundaries of the official evacuation zone were, in effect, put on alert that they may have to evacuate soon.
Persons in Etobicoche and Oaleville, just outside one area of the evacuation zone boundaries were not particularly responsive to the evacuation in that relatively few of them voluntarily evacuated (Burton 1981, p. 6-23). However, residents in one of the areas outside the official evacuation zone (Burnhamthorpe, to the north of the accident site), were particularly responsive to the emergency situation. In this area, 60% of the families voluntarily evacuated their full household and another 4% of households partially evacuated.
Apparently these departures occurred after it was known that
- 184 -
2 chlorine gas was involved and after the announcement was made <
that a nearby shopp,ing center, Square One (being used as an J evacuation center) had been closed and evacuated.
Respondents in the Burnhamthorpe area reported they left Sunday and expected to return by that night. The researchers -
asked a sample of these persons why they left. Their responses were as follows (Burton, et al., 1981, p. 6-25):
- 1. evacuated due to danger or family health concerns 48%);
- 2. evacuated because they believed they were told to evacuate (14 %) ;
- 3. evacuated because they lived near the evacuation zone (11%);
- 4. evacuated because they saw others leave (9%)
In addition, 18% indicated reasons other than those above and likely included people that were leaving for the day or found it convenient to do so. The researchers pointed out that many of these evacuees attempted to return that evening but were not allowed because of police roadblocks.
b) Data Equivalency: Mississauga and Seabrook.
There are two primary reasons that the available data from the Mississauga evacuation shed light on the idea of keyhole voluntary evacuation. First, Burton and his colleagues conducted telephone interviews with a sample of the population for which an official evacuation order was never issued. The sample population all resided within 4-8
- 185 -
~
l '
\. >.
')an of the derailment site. These distances from the source of threat are within the distances inherent in the definition of keyhole voluntary evacuation. That is, consideration'of f.
keyholevoluntaryevacuationinvolvesevacuationofthe[L t population within the 0-10 mile EPZ not advised to evacuate, l at the same time that other portions of the population in the EPZ (at the same distance to the plant site) are advised to evacuate.
It is likely that numerous voluntary " keyhole" evacuees during Mississauga were aware that their distance from the ,
derailment site was the same as (and in some cases less than),
some zones that were officially evacuated. This is i
, particularly true for those who were located across the road from zone 8 and within a short distance from the Square One Center. Unfortunately, there are no available data in the Burton report to quantify the number of evacuees this information reached or its influence on evacuation rates.
Sixty percent of this " keyhole" population evacuated voluntarily at Mississauga.
Nevertheless, this data set does provide some basis for making a judgmcnt regarding the expected incidence of keyhole evacuation at Seabrook should a partial official evacuation of the EPZ ever be issued. l c) Estimating Keyhole Voluntary Evacuation in a Future Accident at the Seabrook Station.
l
- 186 -
a
yj dv _
u y- ,
- p ,
W (N
g ig )
~~
Theamountofkeyhb'levoluntaryevacuationat I Mississauga data base likely*over-estimates k.eyhole
& 1 evacuation since a major'phblic issue in the period in which )
am 3 1 this study was conductkiS,was potentially reactive with: survey s
questions about evacuation. That issue was.one of receiving t
compensation for the costs associated with having evacuated (cf. Burton et al., 1981', p. 6-26). This issue could have q
<w v
{G] entised some study respondent to report evacuation not e at thally performed.
c Coincidentally, the equivalent of keyhole voluntary evacuation observed at Three Mile Island was also sixty y 3 percent.! Itwascbbervedtbatsixtypercent (d. Flynn and Cbihyr.s, 1980, p. 22) of people in the 0-5 mile area who L q p- ,
"^
wer'e not advised to evacuate did voluntarily leave. These evacuees are equivalent to keyhole voluntary evacuees since they were at a like distance from the plant as persons j actually advised to evacuate. The TMI data base likely over-estimates keyhole evacuation since there was other significant information in the public domain at the time supporting that everyone in the 0-5 mile area could soon be asked to evacuate (d. Brunn, Johnson Jr. , and Zeigler, op.
cit., p. 49).
The data on keyhole evacuation during TMI and Mississauga lead to the followAng conclusions. First, it is very likely that keyhole voluntary evacuation rates would be higher than concentric voluntary evacuation rates in an 187 -
a d;
i emergency at~Seabrook Station. This conclusion,is consistent-with the TMI and Mississauga. cases as well as accumulated
.k knowledge about how' the public responds to emergency risk
.information. Sechnd, it seems clear that voluntary keyhole evacuation would likely be less than those rates observed ,
during TMI (public information in an accident at Seabrock //
would be better than that which occurred during d'MI), and
.less than those rates observed at Mississauga (the straight i boundary used to define the evacuation versus non-evacuation zones during Missicsauga and therefore, creating a mixed set of relationships between distance from the risk and ,
-evacuate /not-evacuate population.~would not exist at i ,
Seabrook).
d) ConclusionshtegardingVoluntaryEvacuation.
Emergency plannir.g for an emergency at the Seabrook Station, i:t our judgment, would be very conservative were it to be based on the assumptior of the O to 25% concentric voluntary evacuation, and a higher probability for keyhole voluntary evacuation.
Several major factors that are known from the etaergency evacuation literature to create evacuation-prone behavior were operating at both TMI and Mississauga. Thecz; included the following: l
, i Evacuation occur; red on the weekend. This meant that familiesweremorelikeky~tobetogether(averyimportant ;
influence on evacuation behavior). Evacuees did not {
138 - j l
l
a
.y f
D anticipate losing work or missing school. At TMI, most evacuation occurred on a Friday afternoon after area schools had been closed. Employees at schools, at the TMI plant and State employees had been allowed to leave work early.
A second phenomenon operating at both locations was that numerous evacuees reported they left in anticipation of being officially advised to evacuate. At Mississauga, the ordered evacuation occurred in sequence throughout a 21-hour period.
The ultimate evacuation boundary was a roadway that bordered the population study area for voluntary evacuation. At TMI, there were numerous reports of possible evacuation,
^
evacuation plans being readied and in seme cases evacuees had been advised by the Governor to take protective actions.
H,. '4 4
That is pregnant women and pre-school aged children were advised to evacuate from the,0-5 mile zone and the entire population in the 0-10 mile zone were advised by the Governor to' stay indoors with the windows closed.
The coincidence of the operation of these evacuation prone conditions at TMI and Mississauga further support our judgment that these cases offer reasonable planning bases for considering voluntary evacuation in an emergency at Seabrook.
r'
- 189 -
= - _ _ _ _ _ _ _ _ - _ _ _ - -
i ill1l1l e t d h n e e d t a s h e l u t T
N r
a d e r e o e s o e g i E l n i h a c h n r D c y e s t e t i o
- I n e n e n ; e r e t C o d a y b i g h le o o b c e r w r C
A it y mp a d y in t t c n S e n a .
i n
D ia cn co n a t d e e a i C r w d N
A d e i h m
i v i in it er d EC e d o l
o i e t
L a
r r g ; .
e i
e r
i t
u mr sr e s ; e h ic My .
s e le b l
t t S
o m b, l
b e e c r h a t e r m a g u d m I t o o e n o h a b a E
n e a a p t io n e s l i t s r e l l e d n e d d t L ; l a a a i i d
e t d a a rn m
o t a ; r e
l r n r et I l i e e o g M M r v v e
v h i n d n u t e c n
-E T
n
- a. an i t r
a e a e b a b s e i e t s eb m e w t t r E N t e t u c no o n t b t d n e ot R O a g o o n g n e a u d H
T I
T t :
n n a
e i r
t es c em da in da me d h i o
m n n i
.E P n e s m s d a a r e h e h n v r a e o I
e s l a m e ot r b r ita rop cot e m H R d a i e h r l u e l b T C S
i e a c l t k
o C o
f e c s r a a e r er et n a l
a d s
G E c e e p R e e h e a o o e m a 'd _
N I
D A r d s N in a r m t w w c t r S h E R
U _
D _
N O e I
m T o A s M
R r e O o ly m F f o N l s I t a n i r Y e t o E C d n f N N i e
O E c t t G c o n g 3 X R N a p e n -
I E T O d i A
D M C f f o icc m I
N E A T o r E P A a P .C MM s s a l P I I R s s a A L / O e e s B E F n n u ;
U R N e r
e o e P U I r i s T a a r i F
O A r MO A w w e A s r a
R E
T e
s a m
C n i A o l R p l
- A t s i n re Wy M n C N e O n m o G I i e t - b I
N T A
t e
c n
n a
- n T M l u C R l o cr ce io I O u n S n t n
re se L F b T
N N
I l a ]
r o ef u q O a c 7 n C F n i a r n O o d e o a D i lb s v c i N E t u e o d A C a p h G s e R n t s m Y U C e . e R
O O
S P
A R B N [ L t r p
ot T
C I
D A
R
_ T N , . .
O . y C a 8 m. m.
- m. d 2 a a E a s
- H T E M
I 2
0 d
- e a e h n cr 0
3 0
5 5
0
- T 9 W M 1 1 l!(!llrlI lli t { 1i
illl1 i!ilIl l l !
, =
n l s
e d
e o f b y t e n o u t
io' v t c M it b t e c n i r a , l e o l e ) . a ,
i r o p i b h ? h m r c i d e n t x t u t e x t ht a t r n n e a a u p
h e ur le l
a f o l b
i w
e : c e c e w T p , e n u p _
d C d i r a h l m y l h i t ..
n ! f e v t t u
o l l e
u n o i w e o o s n o o c n l i c g e m e a t n h g
r i t e ig n n e o t b a v e
e r be i n s i
l i t t s
b d k a s s o n ; . e i o g s e u a o u c
a rv ea y d y t m d i l r p e t J b e l n d n e t d a m
t n a c I a r h o a eh o si l
s e a n a d e s i o M i
c t t dn ad _
e b r r
a l p e a i T e e i m a h v
b o
v O a a er . g t t
s s . so nio hec n a
w n e t
r a e r c
d t
a S
t a e t i l
f f
o h t
t r
a eg en e tr n v a tn r o _
t i r n s n e et h
g d
e p n ia e b i
l io t h e
e u a w io h e b a e d d a ig S e v
u t d d d d u n p s s t t o t v o d e o r a i p e h
o t
c E e
t t o r s c t e m
s u i m o a h a m ce e a
u n t S ie ns ah ht G t a d t l .
l e e e r u h o t f t i h u e a d o t A d M r o o t N a b o s S t le b b w E t L y
t i
l
_ i y h b t t ;
d i i i t e d l w c b
i r e i f e r r c b i
t c
o rr o c d i n o y
_ s d e l o c t e E r f i d
t c n o a it i n l 4
l
- e M
d E
c n a a k iib A a n l e d n g t i p l e o n e x r i i M y e n c a uc o t l d s ; i d m d e s d t e d r e c e E a c E f
o r u r i u d p t d d t
_ n s e x e a e e I a R E M r r M
)
r P o R B t e
h u ( n y t n
_ a e b o e
r n m o i e g t t u n B iotc t a i g v a s s
e et t
s f n r e r -
e i e t i t s p -
a or t
c r n b i
b e O r S P l v o
, n b a t n e u i f a r c e o p d o e n r i e n v e i
s a t d E
m g i r n e e o r -
n i s G e r d t w ip br nt . n f
o a e e p e e t o D R M N o H C L c t
s e
h t
. f
- m. . o .
a m. n m.
p i p 5
4 e
5 b 0
- 1 r 3 1
- e :
1 1 H 4 iIj1!!l
s n t e i
e o n s h y 0 n o t i e a t n 3 a t a . t t m a : l t e d c c n le e t 1 p e y y S s e a o e e a p u n h r a
t a e
t y n y s e d t s r ita re h t
h l t l e
t r e e i eh m iv l l a a c a s r h i
d n e ot w en w ae e g t t f i
. e h t o h d a b n o r it e r d o c t ep le u le
. p c s s o g e r n a t e t in d t s a l e a l r t c a h u h e m o c a i r s e o a h w eb e m a ni d b e h o s h h .
- m. h ap c e h ti F t t n o l t r l d w d [ t e S a o , t b f a d
n u l
n a o
, s r n r a f i e y i e a m i io o r d e n s t t d t o n r a o 0 ly a p i e e e h a e a i e u r a e
- m. e l
s w
e i t 0 a : etm 1 a n a g k d
e a r im w so w e a t s p t
t e ut s
t i s v b d
a f
f e i i r f
g h
t t
a m
a c d i n a 1 a e i e o t w s a e e r h l
o a u i a
r o l t n r m t l l l t e t i e n c e R r
m i s o l h v p t c s f d b a m h v iff o x o a a e n t e e n o e o t i e r o g c h e l r t b o r e ly c a t o n s f . f r n y d a
t e r n
o u c s us r r ca p i y l b a b d l e l n
y c o n d p t c t h o e o g e er e a w a o g io io e a n n n t t n i n ,
n . v e p i n n
a s s
, v e eg re e mi nn o t i s r u h e m s i t i d p s s t a r
l a t l e s t l l c a d a
. a t i i s e c o l n e n i r m e e le m. s e l
p s t d t d i v n p a w e v i d i o rp u e a m p y e a n a x o x e o eb e r ap r u f c q r a w e A s g h u R e le c e h L l b o n
o i
5 t -
a A n
a l
p x
e e
s l
a F
_ t n
e
_ m
_ e c
n
_ u o
n n
a ic l
b u
p C
R N
m.
_ p 0
0 5
i
> W
- @ e so C 3 U **
e @
O O C
- x w - e V . 2 C C *a e 6 e 9 - -
O C 6 E @ E -
ed e O > C O - D **
m 6 - e c -C e au. w .e -.
> . A e 6- x C -
- @ A ** e @ D = 0 0
@ C @ = 3 ** e - C a U e e . es > e e. Q) 6 - X a a) E @ t C E x C C D n e C -+4 6 & W m 0
M E =
E
@ U e C C
+3 m % C -> l se e @ O es en l
== 0 C C ** ** C ** 6 * @ ) C e U 6 @ ~ - CJ L C @ E - M C e b > e g g 6 g C w .a - C C C e = ** e C C ) 6 - A e - O e -O
- a v 6 C 6 e 6 6 m & = m - I e .
C @ U C C @ @ c 6 6 a es @ e 6 es a U e a C 6 > b @ *e C e 3 6 0) ed O
6 e
) @
-e6 4a G) C e
C C @
6 m C 6
- a e >
U m 3
3 --
O E 6 @
V e
O 64 @
-C V @ m V
E 3
6 g
m t C 6 C
g; e 6
-6
- 6 U L - a O en C @ L U - @ . U E O
> m C e
@ @ E O
a k
C@
=D -m** C
- D D e >
C
@ O E D C C e @
L e
- M C O
- a e O &
)
C U
- e 6
a C -
w C
.C b
a
@ @ C
.e 6 C 6 6 2 - V
@ sa E. 6 6 E
6
-3 D@
m b @ O e % -
- w - .U e O =
@ 6 @ m 6 C p g C. . e - -= 4 ,g 4A @ b e- m -
@ @ 6 es ==
C @ C E
+8 C.a **
- - OE -D6 a so O
e
) *8 ) Ee.n = C ** a (
C e E @ X A
- a Ip ea 6 E E - C E U en @ O se - en
-U @ = 6 +a U 6 E e
= ** = 0 6 & a b e b a e D a o e C *8 C @ 3 m @ m )
Om O 6 & Q @ O U U % M so o O O e 6 g C E C E @ e C 0 h
" E - E 3 : e ** O
.e C- m e @
@ > U
@ 6 C D 3 , 3 m @ m a m O m 6 e e g a
@ 3 0 - @ C @
6 D U C @ g
& C C
- O A 6 @ @ => 6 -
6 O 4 0 6 m b o eu C
@ m e C C e 6 10 6 +s C b U C - O h C @
@ C 6 O @ w 6 g; g
> @ C L 6 C O 6 C .h
- .aa .E4 O O @ 6 e m - qp 6 6 a b m mO
+8
- C 6
U a -> wg m
O O 6 6 h C 8 U O C A & M C w o,
1 i
E . .
h@ E.
- A e N E. e C
V a mC O 6 U in C.
3 6 o..
O C e ..
- e- E c .N.
l l
)
r
- n e o l d
i o r t o . o
- a h u c r n o e d n c s e v a a - r d r ,
i o v e e s s ly g t t
_ e rp cn o s i o a b s t n e u n l n n i n a h d o a g io a s o h s t n n n r s i w r
a n t i t o t o t 't a i. e a a p c d i o n t u t 0 h e n f e o s S c n 0 t d r a t m i l a e : l a
n o d w aR l
a o ve y t
. m e
8 ]
?
c u iD i o o h g t io c r r t g l c
t t e. e t a u a n b . o n a n y t s a s s o u n d t o d a n ng ne en cu
. l a n h o t s
e a
e e W m r p io e r e s e p c m r [ s t e r i m er dr k a v a s s e e et n e a m a d e s m r u r h a o p l t e r k 0 .
o c o w h c i s a t i 0 f a p e f h i o n t o p 2 m. n v e n m R o c m t I s n S 1 a I e b o E y
s . c e y n i t e e t s n t s i n i s r a l
. r o a i e e a o t s t r u h its r e
n o
a c s t l n u u e , c c i a u o n n h e q d u t s 7 l
i i i h t s -
a h o d d w t e A c t t n n a s i o i a a t .
c e t l
w t c st m i a n n i e i r o e h io o l n n t o t s w a s i
s f
n m e
h o
x h
t s e t u u o e p i u l r u f f c t e s a l
c o n n a l a n b o o n t e ot o C I a c C i s T t n
r e e m
g m ) h t io r ) e er - o m t s t u ? s n rf e to y a b a d a e o n s l t s i E e h n ( s a i s . t n l p y r v r t f a t e s g s s o e o
i D i P
o v e r o in e t f i C r l
M m s t s o d h e y [ e t r a n k n y t a t r r v r s h a r g a n a t c r i o o i o n r t e l a rt n e w t n e o f e o t u r g s y a t e t g t s l o o i r w s t t n e n P p n n n b c p d C D u e a i o ir i ar i s y e ua n b i l s e mt ,
c . n e s v g r s h i o s t n r k d i s i r n e e r t l t a e a a , e h n r e i v r
r ap a l
e i g n d m m k k p e t e i p h n e n i n a u f a n t n e e e o f h o rp in a a e H i S U R D H A m o w c t le D D g g n n i 0 i n 0 n n n r : r y 9 o y 9 o 0 o 8 o a 2 o a 2 o , 3 M M d n d n y r s h r s h r a h y e )
ly r c e r c e d c t .
u r t u r t i r lr f r h a f h a f r a a a m. a T M A T M A F M E ( a E
e d s d e a a i h h e a t y
l s s e e a n l r n r t t a i m u r s p d
n s d e
0 ie 's i e e p o u e e i o 0 b c h b 0 s n l l 3 r r i t r 0 i r r y e a i e e
h w t e et r e 2 d e o c h n d k t t - o n I I r t i t 1 v n n t o .
a i a n d r e r o a t o ic e o s r r d n d r h r w o o o l
u e o h e o C g t o r e e a n t d w . p t t ,
h n n r n t e v e r e d t e t r e n d t h a e e p e l
c e u or a s e d t p m D d w i s H o i hte e o b f w ro re ta che s't o H l l l l i i m r [ n b r h o p a l o d
- o re m p r t i o 0 o n
d o r t ic r d t d f o h t n
. 1 d . v 5 o
l g e d c t r u e t a e .
n i ohu n d n e p n h r n a b e n n e
'n I e G s t i s t a m o
p t r o e r o o d p d s i v i
h y o d i
h n s n e e s s o c h n t
, g a a s i n i s i t a d e t t a l e t a e l r e n e tu t n e a r a i t s i l n ]
v er m r o o e t r , e P w s c s v i w po s.
i a n n m e e
. n t
n s b t
n d w e e itc r d 0 t o n t t r n g e o e e s e le o a d
h o e e 5 n et S e d id n m r d a s n h a e u er l l 3 b u i l p u d p t e o n s o o n d o t
n d h s d t l s u e r d e p a oir d d l d a a a t p e h i a e r i n la o n i o o a e e e r h m e h e P s w h p o d i P t a d c t h h m r P t i m s s
- s l r y o ;
de ir y n c
n iom um e t t r n e a e r r s d n o e h m r e g r u p n a j
- t h
t
- t e c t e u n a n a n u i u e s e g m r is da a ig e o l
r d
e e m n u s r sr e r ;
- t b a r i
a e d f e c e l e p s r u s ly s e y i h s
- e e c e h l n ie bs l o h c ;
c a s r e
lp no va ug t a o p e ra l t
_ i t i e m , t s o s w i po s e es e
_ e p r y i e ra s t
m a o y t t p n s n t f l d o b r g
g a
m d o ic o h no ;
c n l a c de i oep n
- l i l i g f n n o t f ; t n u t s b s r
_ i , h a n t a a n r t
- mr ibe se es uc o r c op u e l s i
m r
h a l
o ut n n a m u a n p o e a r o o v n e o pac p s a e m h s r m
_ l A o h h e i r l
A t r a c i a r s d p T o
h g
r , ,
u e e b c c n t n n r s e e o a r r h c e e T d f f h a n n g r
o r o o o r n b c n i c ub r s e s nr e e s b s v v e r e o o i r e r h G l P H P T m.
a m.. m.
a p 5 0 0 3
2: 0:
0 1 2:
1 1 1
1llI
)
a y i r l n o
- i a s t v d n l a e iv r
- d y u y h h d o l e at " s t a r
- o s s n y y e l r o e o S d n c o t t g a o h l l " e e n t s a i n n f c i c d P e r da iot l ,
s e
m s n g e i r i n
d s
. e e A y
p f o t b ,
o e r 5 l t h o n i t i s t d p o u x T t t e n y s a s n d t n n n o i e ( n e h t n l s e d r e e d i h d e l t s i d o p o n A e Dp m n h c a y m et a l ! i p i
". eg m a t s r r g w i u t t e m x d o i o n b o i n e t m . e c
e e n w o
, s s e a a i e c n e
r is c dno e h s o i t l
o r mo ne s i t e s r t a m l l r l
l v i a t M n t f i h o i t
c r
u ,
w er o e a a n m c n w dr A im ad o n y s e l t p c s n s e in H n etr ow
, r y io t l 5 c s h t n i d ly a r h b t n i . "s r n y e g n w e o d e w oc ec d u e o n n u
c g n c M inh o a a h I T
r e e d o h t
g a r e n it ho n im dt e p s e d e
m e a e r t
l m
e e r a t
d e
a r v e e g r it d r e r u t l p r s r m t e C a e p a o w n u m l o l m e a o a a x h r R e I f E a c A i m h w B p a e T o M h -
d g
s n C t r R n n o n e n e ih a n e N ; o w o nke i g d a b t r h o t o b s
r e
n i t a t e u u m i t I isu rm s e z c l e rp f I a a C e l e r f o e n m v l p l a n f e o a e sk b o ; e o n u b e t g c i c o e e m r E l
u p e n y i h
g v a n b t t 9 o e o e i u n e t -
s w t a
s a
t e b s w b i t t h
g A
- s g g l t t , c s u n r y o e t d ; in e u o i c t g
n i
l o o m h i h r e g n g b l n i f t t
a o t a p
l in t n a f e r n d u
l e e i y n o e m u o e ;
c t m h h s c i ev t
a a d c r n r ig ty r r c t s a a a
r n o a a e c a v h l n t e h in e e t c c A f a w i s R t l r ,
. r t r a n a r t e s a s m t , t e t s n n s r c -
a s e o C e n e m s R l t u C r pe p N a n o R t m c e n N
a a o e i c n t h h n a t d s T t h s a s e - y f c
e w
e ic ie g h - e o t n l r n t l b d i
e r d r r d u n f B e u e o E p e ie t D b i H s n m n t C r t e r e e e R y B i c o m s M M b 0
0 1 .
y - m.
a p d 0 .
- 0 . 0 -
d : 3 i
2 m. : _
M 1 p 6 -
n i
a y m . r n e y e ly o r a ; v d i g d l l e e t
in d r e n a h c t n
, a l u n s w n T u a r o r d d y u t a s o o d h e i e n n a o a e d s e t v t n v e a t w S f v t r . r o a o e m s o l e r t . u i w m n s l r e p e n . n t l c
t y o m v e y a h o o me e n iu c
i t t t e o e a e o d m n l e a u v u . e o r d n i c e h b d e o f e e c ; r w d a u l a r s i t ; c t a y l r i e o t i d r a r ; ;
i s
v r r t i t b r c a s a F t e a n. e l y
. e a e n h e s i h i h s v o l o
s i a c l a s e a t r e o n I b o e a o s l n i e s h c e cn n i t a c h d N e r g e m- l o t f d l i s i t t c a e g p o e e b i e a r d s ot t e e r a 5 t n t h b s b w ot no d a : n p a o i y a t u r s e h h s l e N t t b d e s c c a g h g s s r w n c
t r a 'r g o h n o t t
c n i o l t
a n i H i a s e n a r e
u w o i h e :
n is b i . n h e h d i r w r iu ie b n d c f f e i t g t e r o o n r n re re r as o-r f
e o o i t e o r
t c i l
dt t i l t l s o b t a u c t d r o d v g e t n p s m r l d f e t u a n a a h r o e r u n e x o e e e y y e h n h e o h e T o C r p o i D e p r H f h b R t u s R c t
c t e r
a n o d r e
r oi d e o c c h t o u d n t a t g d E t i n u n e e e e c e i r r t M y p m a p c l t e e l p o v g y M d l a S e o i n t e a h e l a i h c i t t p f d l t u t d u u n i i d y r l
b o e o t b a u
o b
, a o r c u o
i w er it p d t l c g m d b e ; i ;
/ n ; i a r ic t b g s i d d d c l n i n a r e e s f e d i w u n t s t d n m er s s i p a r E o e u t s a m o c t c f a a m o , p t a n h
R e e r r pos e r e n t d o w M i s E C e
e c c n n e , e e r e i r e t h c r e f a g n d e f n u e n c n o r or r e n d o c e e H e n c t o f r a s n B n y e s s e o b f
n h
g w e c n c e r w ,
o r n p s o s a c u n w t w d b n d i n e e e s s n o E e o n l n e s r t b t d I h
e o n t r n C d i C t r h e e e e R i a R e P T D M H D N M H N B
- m. y 1 m. m. .
p a 3 a a m.
d p r h 0 0 0 u c 0 0 5 3 t r : : 14 a a 1 2 :
8: S M 1 1 2 il
o 0 t t 2 d
_ e h s e t ,
e h g n s s a e )
v 'n a h h 'n
_ t i o t t h
a g
m it a ie re t
t t c e
n a
er e
- n n a u b c d t r n d t . v d i o u o r e i u r l e y i a o o e g r r r hi -
l r i c e d s u e t a r H s b c e a o o d a v af d a n e p c 6 d s l w i u t e t a e i h f i s n p e s c n u h h g [ t r eg W ec e x d t n a e ; s o n o t g r e o a o v d i r n c e o o e a b , f a l h c e s r a le d k - e . n r p t
i r u d r o i t y r t n d s ly x s b i a ca e l l
_ n ; p s e d b r o r h u y n l e y y i n t e a e e a r e u h k b t t
e b l
o o e s h ig e h g ia a i
d d r e s d k m d s t b i c n nr re t i
l i i n at e b y is a t a
e o n n e b u
d v s e i h d
n e
l u r o o
_ w t
_ b i m r n n h o h r r
. a rp n e e t c e o a o e t s t i h s t w op e s n t n c h y a e e n a t n a o o i~ bo l
s o W t b n h z o d m o o h h le p w i n
t u
i e t i i n e n i t s g b u p n no ru a r a n t t t b t t b o a e s p o u a d
. e ze m
s e i e i e l a s a
u n d t a
u d
e l e i m b u c k p n c l e t s b b s c e l s c n b o a s e r e u i l a r b t e a m u a a r b c e a e t u r o t n e t u r v o o e v a u e h n w b s t p w m s i H b b p E w w l e W b b T i t
n s a
r i n
g s
. s r e o s h
c n e r o s s u i h t a 1 c u m 1
- lo l A
y s a t b c i
n a o l g g a l n n t
r a e i i r h m m e e t r r c n a a n eg o l l U t A A g
n i d h d c e g
r e
r n m y u u a a o n b b e n n n l n u e r c a y
s o a h n n e T i C i bt e l
r t R t cr e e N e n r o l 1 e u n l e 1 m o s r u h u et s s e b t b e v a C P
r o P y P t R C A b A s N r
e
_ f t
a .
m.
O p
0 m.
p ly t
r 0 m.
p
- O 2 o 2
- h :
'i: 8 S 8 l
e e e d h l
s a
cr g s
hp le o n c b t m u b
u w uo n i n o a s
r io o i
d h e rt e e s e t d b e r
s r e e p t t o h - n
. u g b s p r n xria t l t
- e r g d y a r n a a o r p
- - h o n b o i d t p s a ,
e t n i d n t r a a h r er ' e e o n s l
C d e d e s u h t c g o n n e e d b R 'e n C n t e i s r o n t e ic E b M s o R t a l' d e t u d s t o e n b u u i M p e p
%' l r
- h y
n b e 'y n N g e r t b h e
a c a t x e e e n
. d a o n n e nr t e o m d r t h a c a D r n a o M e i e r t r c ot h d o i o a d : d y ia h e n s e o n n s uc w b i e t m h e eg n w l t n
e r
n e r o u e hn m T r e t o h t e ot p d t b e e h t n h t
e e n d r e o d b e u m e e D f h e e n
t D io e a
- d t ly n a o e c ; n g
s m l n e b e n d T o g h n l h o l D f l i o e p e r e o n r n t b t u o e e h e e x e r c a eae who edr t g . g t f h s s r r e y n x o a r g : e r ri u iw a e . y t s a t
g n
g a t o i i s
t a
s s h w g h
g r d s h
t s t s
a t t
i r v e W n i s o t d n i e r r o g e e c c n l e e e n u u s n s a e e k i n r i ro ei ue o o e n m i h d d r r h m a d e n y m b b r t b t o t e t n
d h Ig l n
e s r n n r
e r r f h r e e i i r
t t l t t i e r b t t t s e c h a s e u i e a h t o o m p i s s e s a s a t t a p o e r p t T a h h u e r s a o v i t o p ni er iw S w o c n o O s l l T T n r b a w t o l s C C R R N N h h t t 2 i i 1
w w -
t t A
c c o I
I i
l i n
r f r n n u o o s c c s a
n e I In R s
s l
e r
, l a p r e c e n i n
t n n o n o h t e s c n c r e e e t D s p /
w e h e e l t l g c n a i e r n d lc s n n n u er b
E a o o n ef h s r t c C r o n e e R e h o M t N p T c
] ]
7 ? .
[ [
g m.
g p n n i i 0
n n 0 e e :
v v 1 E E 1
~
1 l'
,l Ii o g . r . .
e t n n e e h s e e h l t r e
w i H o su io s .
h t
t r
t u b b n e -
n t h s l t u o u e o ro r3 . a lo n r d b b b t p w
o i
y nt ic p x l a inh l e e c H a
r d n s e n b i s e , p t i u d n o ah e e r k o y n r e i t d a oi m
e rs t t t
p p h p e n e e p v d
n e h g t h e w nu e r me
. h t
s s r e u a e e g o e y a a t s C p e r r o h 'n c n o o w r - r l d y l R m n at g c t r d x o n" rd e o o n o r o e o o s l
b N u e o s v s l
a e d h
o i e n
ia mi edm , ly s h y b h; v i d
s e o h m olo t u s c h
l o
t n r e s i e p h a s e a t n a l 2 d w c o e s c i t o h i s g
h le o w a u o
a t
s e y l
t n c a t s
t l
p t
t d iff l i n n e n ; b m e f w i r h e n r ep m a i a h a e a i x b h e a t
e a h i s yo no rr ro e t s s p t n s 5 t m u t n h s s ; a p f h " e A g h e s ot i e t e ut n e p io l t t s t y t i c d
, e r i e t o ni t de r s
y a g b l
y s a p r r y d p w a a t s n a ic a h d t t i a t o ev d h n a l r a n n t n e w on o is s w s t e h i o d e ot s e i e r . t r s n o a o r e et m t b e o x n t r r w mm a e e s n o w o dp i e t i e d ic l
n r n e t v d s t a o e b a d h
e a i
t c e o r n lo t a i s a op a
t M e e d ht i e c m n e a o h r e a o l
t b u o di ee het e t lp a w c s s s S o h m m c a R s n B s d b H t
- l n n r e i i e f g t g e s e n d b i h n iv 3 lo s i 1
- m e m at i sy l -
h A y o ; t e a y t h n r d t n o s e n i s i a t e w c
t s
i a
t 'e a n e a t r l u u l l r e p c o p e c o a n o e c n e v n e h n U p e a p t U e
s a e e c l n e e r
r e _
s f _
s n _
e o r c p
s h w g e r n u _
n b _
n o r t o n h e T D m.
, 1 m. , 2 a y p y a l a l 0 d i 0 d 0 n r 0 n ir :
u p : o p 2 S A 7 M A 1
(:If'
4 4
J l
l 1
L i =
l U l C l
l
.c.
C e e O
C G O m -
C &
Q =W U
D
>6 .U W W
> U i
l l
l 3
t 4
.