ML19345C216
| ML19345C216 | |
| Person / Time | |
|---|---|
| Site: | Beaver Valley, Millstone, Fermi, Limerick, Maine Yankee, Midland, Shoreham, Bailly, Crane  |
| Issue date: | 06/30/1980 |
| From: | Cosby J, Sheppard W WILBUR SMITH ASSOCIATES |
| To: | |
| Shared Package | |
| ML19345C217 | List: |
| References | |
| NUDOCS 8012040155 | |
| Download: ML19345C216 (43) | |
Text
---
, I,, o '
y_u AN
.EV AC J A-O N V E ASS.ESSV'E N T OF NINE NUCLEA9 ? OWE 9 3LAN' S EME 9GENCY PLANN..N G ZONES.
VOLUME I
PROGRA\\A RE3 ORT fQMLmyscg737nm,z9,
2!E#EE#atig=
Prepared for FEDERAL EMERGENCY MANAGEMENT AGENCY 9fidut Smi/4 and dsuce;2/es
^
8 01 e 040 l55 f
ACKNCW!iDGE.M T The independent assessment of the evacuation times centained in this report was performed under the technical directicn of John C. Cosby.
Mr. William V. Sheppard, Vice President, was the Principal-in-Charge of the Project.
The principal centributors to the individual velumes of the report were:
Volume I
- Program Re= ort
- John C. Ccsby Volume II
- Bailly
- James R. Bancroft Volume III
- Beaver Valley
- Richard A.
Day Volume IV
- Enrico Fermi
- Elbert L. Waters Volume V
- Limerick
- George S. Coulter, Jr.
Volume VI
- Maine Yankee
- Robert P. Jurasin Volume VII
- Midland
- James R. Bancroft and Elbert L. Waters Volume VIII - Millstone
- Frank LaMagna Volume IX
- Shoreham
- H.
Cean Browner Volume X
- Three Mile Island - Welbourne E. Thempscn All reports were revised and edited by John C. Cosby and H. Cean Browner.
All of the above personnel are per:nanent employees of Wilbur Smith and Associates.
l
i TABLE OF CCNTENTS PAGE INTRODUCTION 1
Evacuation Time Assessment Versus Evacuation Plan 2
General Assumptions 3
GENERAL APPROACE AND METHODOLOGY 7
Data Collection 7
Evacuation Route Cevelopment 8
Evacuation Time Assessment Program 12 Evacuation Route Refinement 19 Special Problem Area Evacuation 19 Evaluation of Scenarios 22 Sensitivity Analysis 25 METHODOLOGY FOR CCNFIRMING EVACUATION 26 Visual Confirmation by Helicopter 26 Ground ?atrol 27 Telephone Confirmation 23 Computer Cards 28 RECOMMENDATIONS FOR IMPROVING EVACUATICN 30 Public Warning System 30 Security During Evacuation 30 COMPARISON OF EVACUATION TIME ASSESSMENT AND EVACUATION PLANS 31
ILLUSTRATIONS FOLLCWS FIGURE PAGE 1
Evacuation Route Cevelopment Block Diagram 8
2 Evacuation Time Assessment Basic Flow Diagram 12 3
Combination of Four Separate Reaction Distributions 13 4
Ccmbination of Two Separate Reaction Distributions 15 5
Combination of Four Separate Reaction Distributions 15 6
Time Distribution of Input Traffic Volume 16 7
Formation of Queues and Delay Resulting From Limited Capacity 16 8
Merging of Restrain 2d and Unrestrained Traffic 16 9
Queue Formation and Additional Celay Resulting From Limited Capacity 16 10 Evacuation Route Refinement 19 TABULATICNS TABLE PAGE 1
Variables Employed in scenario Assessments 23 g
p, m
~
INTRODUCTION An independent assessment of evacuation times around nine nuclear pcwer plant sites was made for the Federal Emergency Management Agency. The results of this three-conth study are cen-tained in ten volumes, as folicws :
Volume I
- Program Paport - Evacuaticn Time Assessment of Nine Nuclear Pcwer Plant Emergency Planning Zenes (EP Z ' s)
Volume II
- Bailly Nuclear Pcwer Plant Evacuation Time Assessment Volume III
- Beaver Valley Nuclear Pcwer Plant Evacuatien Time Assessment Volume IV
- Enrico Fermi Nuclear Pcwer Plant Evacuation Tire Assessment Volume V
- Limerick Nuclear Pcwer Plant Evacuation Time Assessment Volume VI
- Maine Yan'cee Nuclear Power Plant Evacuation Time Assessment Vclume VII
- Midland Nuclear Pcwer Planu Evacuation Tira Assess =ent Volume VIII - Millstone Nuclear Pcwer Plant Evacuation Time Assessment Volume IX
- Shoreham Nuclear Pcwer Plant Evacuation Time Assessment Volume X
- Three Mile Island Nuclear Pcwer Plant Evacuation Time Assessment In addition, an Executive Summary is also available.
a This volume contains a technical discussicn of the generai apprcach and methodology empicyed in making the independent evacuation time assessments for these nine nuclear pcwer plant EP
's,.the evaluation of four scenarios, the discussien of l
f
evacuation of special problem areas and the methodologies for confirming evacuation of the areas.
This report cencludes with recommendations for i= proving evacuatien times.
The scenarios evaluated are those exepcteci when evacuation takes place at night (the optimum time frem the standpoint of evacuation time), during a normal workday, during bad weather (the worst case condition),
and, where applicable, the evacuation with sum =ertime resident and transient pcpulation.
Evacuation Time Assessment Versus Evacuation Plan The assessment employs available demographic data and trans-portation facility information to predict the public respense time to an evacuation warning on the assumptien that such a warn-ing is made within 15 minutes of an en-site nuclear incident warranting such emergency acticn.
The assess =ent must provide for estimates of public respense time to these warnings, assembly of family and other groups, preparatien f::r departure, travel time en the network including consideration of capacity limitations en the network possibly forming queues which add to delays, and clearance of the 10-mile radius around the site.
It must consider the evacuation of special problem areas and groups.
These would include schools, nurseries, nursing and re tirement homes, hospitals, penal f acilities, beaches and recreational areas, and other activities which may provide periodic cr seasonal concentrations of people.
Population groups without access to their cwn transportatien or unable to provide the special transportation facilities required for evacuation must be included in the evacuation time assessment.
Evacuation time assessment methodology combines selected techniques of traffic management and planning, land use planning and cperaticnal analysis.
Because some condtions prevailing during an evacuaticn are not well documented, =cdificatiens ec
-.2-
scme established principles may be required to meet evacuation requirements.
Assumptions may be required in lieu of well formu-lated relationships because of the highly speciali:ed prcblems being addressed.
These assumptions must be founded en best pro-fessional judgement and/or extrapolation from existing kncwledge.
The assumptiens must be specifically identified.
The bases upon which the assumptions are founded shculd be appropriately dis-cussed.
Evacuation time assessments contain basic =ethodology commen to evacuation plan development.
Ecwever, the assessment is not an evacuation plan.
The major distinction between the as sessment and a plan is the extent to which the elements have been coordi-nated with all participant agencies and jurisdictions.
For example, the assessment may assume that a specific traffic management element is established to cptimize traffic operations at a specific location alcng an evacuation network.
The feasi-bility of such an element in the assessment would be based upcn established technical principles.
Ecwever, the element would not be coordinated with specific law enforcement agencies to establish what agency would exercise the element control and management nor identify the type and number of persennel to be required.
The study time allotted makes such coordination is-possible.
The assessment must identifv what is required for the evacuation time to be realized, and assume that such an element would be implemented.
General Assumotions In the assessment of evacuation times, certain general assumptiens were mandatory.
More important of these are summa-rized as follows:
1.
Emergency evacuatien of the general public frem the EP:
will be performed largely from the home by the family as a united
~
I group.
This assu=ption is prefaced by the follcwing quote:
. people will not evacuate an are a, regardless of the danger, if their family group is separated, unless they knew that members of their family are safe, accounted for, and that arrangements have been made for them to evacuate."
It was felt that this psychological pressure is so prevalent and streng that the above assumption appears to be justified.
In additien, to assure that segments of the family are safe and accounted for would have required the establishment of shelter locations and the develep-ment of a shelter support plan.
In view of the next assumption and due to the short time period of the study, this was not dene.
2.
Public use of shelteis in previcus mass evacuation exper-ience related to natural disasters appears to be a very small percentage of total evacuees.
Examples cited in literature include:(2)
"In a California fleed, only 9,260 out of 50,000 persons evacuated registered in the 38 Red Cross shelters ; during Hurricane Carla, 75 percent of the evacuees went to cther than public shelters; and during Eurricane 3etsy, only 20 percent requested assistance.
Generally, shelter centers are used cnly if nothing else is' available or if one cannot financially care for himself. "
In this evacuation time assessment study, it was assumed that the predcminant traffic, after leaving the 10-mile EPI, went diverse routes rather than to a shelter destinaticn.
Therefore, the evacuation time asaessment ended at the EPI boundary.
An analysis of route capacities and service levels of highway f acilities beyond that boundary was made to assure that delays or problems were unlikely to cecur.
(1)
EVACUATION RISKS - AN EVAI.UATICN, U.S. Envircnmental Pro-taction Agency, Office of Radiation Programs, EPA-520/6 00 2, June, 1974, p. 49.
(2)
Ibid., p. 52. 1
i 3.
Experience gained in a large range of evacuations indi-I3) cates that private vehicles
. were the predcminant mode for evacuation (more than 99 percent).
Pcpulation density ranged frem approximately 15 persons per square mile to 20,000 persens per square mile. "
It was assumed that this was applicable to this time assessment study.
It was further assumed that persens without private vehicle transportation would be provided, at their telephone request, adequate transponauien in high occu-pancy vehicles (HOV's).
The additional vehicle volumes en the network would therefore be small, could be affected during the general public evacuation time, and would not affecu the ccmputed evacuation times of the general pcpulation.
4.
It has been cbserved that not all persons will avacuate the EPI.
"In many cases, even when presented with a grave threat, pecple refuse to evacuate."I4)
This scurce centinues, "Results of this study indicate that approximately six percent of the total pcpulation refused to evacuate.
Other reports indicate this figure can run as high as 50 percent.
There is no reascn to believe that because the disaster agent is radiatien rather than some other agent
. will provide sufficient motivation to leave.
Rather the cpposite viewpoint should be taken--people will hesitate to leave. (5)
It is believed that a majority of this hesitance is based en fear of exposing their. property to looting and vandalism.
Nctwithstanding this evidence, this time assessment study assumed that all persens evacuate.
5.
It has been assumed that the traffic network within the EPI has beea isolated so that no through traffic is permitted to enter it within 15 minutes after the evacuaticn warning has been iss ue d.
(3)
Ibid., p. 52.
(4)
Ibid., p.
48.
(5)
Lec. cit.
)
i i
)
6.
Traffic management by appropriate law enforcement officers will be performed at selected intersections where evacuation traffic ficw is given priority.
7.
All persons in the IPZ have been provided, in advance, sufficient information regarding the assigned evacuation route from their place of residence (referred to as the " centroid" in the report).
8.
It was assumed that the public response to an evacuation order can be defined as a combination of up to four categcries of statistically distributed respenses :
Teceive warnine, le ave work, travel home, and evacuate home.
. : was assumed that these responses are time-distributed folicwing a nor=al distributien curve.
The details and applications of this assu=pticn are =cre fully discussed later in this report.
Additional assumptions were made which are discussed in the body of each report.
M N
N
I GENEPAL APPROACH AND METHOCOLOGY The evacuation time assessment for each of the nine nuclear power plant EPI's were all made, follcwing as near as feasible identical procedures, depending upon local circu= stances.
Thereby, the resulting evacuation time assessments reflect a commen measure for all sites.
The methodology can be subdivided into major tasks:
DATA COLLECTICN, EVACUATICN ROUTE DEVELOPMENT, EVACUATION TIME ASSESSMENT, EVACUATICN ROUTE REFINEMENT, EVALU-A'2ICN OF VARICUS 'SCINARICS, AND SPECIAL PRCBLEM A'4EA EVACUATIONS.
In addition, a sensitivity analysis was perfor=ed on the evacu-ation time assessment mcdel to test the infit.ence of some of the input assumptions on the evacuation time.
In separate sections of this report, discussions of the rethods for confirming evacuations and recemmendatiens for i= proving evacuations are included.
The assessrent was conducted by four ceams comprising traffic engineers and transportation planners, working to ccm-men guidelines and =ethcdologies.
Assign =ent of nuclear pcwer plant assessments was =ade en the basis of geographic groupings.
Data Collecticn The initial activity was to collect planning data and gen-erate a brief familiarity with the facilities and local problems and objectives which might influence the evacuatien time assess-ment.
Visitations were made by each team to the cperating utility or utility holding the constructicn permit, local and regional planning agencies, county and state highway depa_%-
ments, local and state Civil Defense organi::ations and other such groups which could be of assistance in cbtaining these.
=aterials and infor=ation.
Ccpies of prior evacuatica plans and O
O
studies were solicited.
Cn',.y four areas had prepared evacuation plans which were supplied to the study.
The remaining areas either had no develcped plans or had plans which the sponsors thought were proprietary and therefore not available.
The lack of plans for the five sites did not significantly inhibit the perfo=.ance of this independent evacuatien time assess =ent be-cause suf ficient planning data were cbtained to fulfill the re-quirements.
Possibly the only influence on the study was that of local issues which may not have surf aced during the visita-tions but which might be expected to have been addressed in a detailed evacuatien plan.
Time for this study did not permit a complete assurance that all such issues were ccepletely de-veloped, althcugh every reasonable effort to this end was made.
An irportant data input was the United States Geological Survey Quadrangle maps (7.5 minute series) o f each EPO 's.
Cur-rent editiens of these maps were e=plcyed to develcp standard base maps of the zone to the same scale and fermat for applica-tien in the assessment.
They also provide information of route structure and intensity of land use.
Evacuatien Route Develcpment Figure 1 illustrates the general work ficw of this Task.
Development of EP 3 Base. Maps - Base maps were created for the report using the USGS Quandrangle. Maps for inclusion in the report.
Each quandrangle maps was reproduced and combined to obtain working maps for each area.
Cetermination of Planninc Ocne Soundaries - The available demcgraphic data essential to the evacuation ti=e assessment was related cc varicus planning enes which varied in format among the nine areas.
In scme cases, these data related to
_g_
-- M l'
CETAIN ETERMINE PLN4N I NG PLANN!!4G 3 3 741 C7 OEVELOP E8Z 20hE POPULATICNS DEMcGR AP'il C 3ASE *AP BCUNCARIE3
'ATA OEFINE
- AaCR I
PCPULATICN ZCNES v
OEVELCP EV AC'J ATI CN PCPULATION CENTRO!05 R
v lETERMIME ICENTIPICATICM
- p 3pge:At EV ACUATICN PRCBLEM RCUTES PCR 2agAs F AC:4 CENTRotD OETER. INE LINK C4 ARACTER! STICS CEVELCP LINK /NCCE LENGTH, AVERAGE DESCRIPTIONS SPEED, EVACUATICN CAPACITY v
CENTROLD SCPULATiON AND ANERAGE CAR CCOUPANCY T
T l
)
v INPUT TO EV ACUATION ASSESSMENT PROGRAM
}
l EV ACUATION ROUTE DEVELOPMENT l
l BLOCK DIAGRAM FIGURE 1
- 44. f A ou 4 r.am
i I
tcwnships, incorporated tcwns and similar political jurisdicticns.
In other cases, notably SECREHAM, these data were related to arbitrary enes defined by roads and _ other geographical referen-ces because Suffolk County is one township and no political subdivisions are made.
The information provided by Suffolk County Department of Transportation was used.
The same sub-divisions employed in their preparation of the Evacuation Plan for Lcng Island Lighting Company were incorporated into the SHORIHAM assessment.
Definition of Major Population Centroids - The major popu-lation zones were assigned the related demographic data.
Where the nuclear pcwer plant was in operation, these data were 1980 data er adjusted by standard projection means to that date.
In those cases where the plant was under construction, data was adjusted to reflect 1985 anticipated population.
In all cases, local data, generated by local planning areas was used.
In most instances, these areas were tco large to be related to reascnable evacuation route development.
It was necessary to subdivide most of the planning =ene into a nu=ber of population centroids.
The populauien assigned to each centroid was based upon relative intensity of land use as indicated on the USGS maps to the total land use within the planning zene.
The number of centroids was governed by a rule of thurb related to the number of available lanes of potential evacuate route facilities.
This rule assumed that at least one lane of route facilities be available for a maximum of approximately 9,000 centroid popu-lation or less if the route was particularly suitable for evacu-ation of a centroid.
As a result, the number of centroids in each EP" varied from as icw as 31 in Maine Yankee to 95 in Limsrick.
-g_
m p
s e
-,---v
---nr om-en"
,,-a-e r,
s v
Cevelocment of Evacuation Routes for Each Centroid - It was observed in some available evacuation plans that there is a great temptation to use high-capacity interstate and similar type route facilities.
These prove not to be as advantageous and efficient when the interchange on-ra=p capacity was included in the analysis.
Such facilities may have one-way capacities of 4,000 to 8,000 vehicles per hour, but the on-ramp may ha*m a capacity of 1,500 vehicles per hour or less.
Therefore, the through capacity can seldom be realized within the EPZ with the limited number of interchanges provided.
In this evacuation time assessment, the development of evacuation rcutes was based upcn the premise that a facility, even a two-lane rural road of good quality was more effective in evacuation than attempting to force large volu=es onto interstates.
Selection of evacuation routes for each centroid were made of routes preferrably oriented in a radial direction out-ward from the power plant site to the 10-mile boundary of the EP Z.
Evacuation routes for a centroid, in most cases, co=bined at certain modal intersections with other centroids, where capacities of the remaining route structure permitted.
There-fore, route " trees" were formed by traffic engineering judge-ment.
In the development of the nodal patterns of the network, a node was assigned wherever the route:
e changed operational characteristics; i.e., capacity, or operating speed; e
Route departed from the planning :ene; and, o
Route left the 10-mile radius of the EPZ.
. 1 1
)
l Development of Link / Node Descriptions - Each centroid and its asscciated nodes were assigned numerical designations.
Listings of the evacuation route for each centroid related the j
nodes assigned to the route to the external nede, defined as the point on the route where it passed out of the 10-mile radius of the EPZ.
Input of these were then generated to be used in the EVACUATION TIME ASSESSMENT PRCGRAM.
Determination of Link characteristics - The traffic charac-teristics of each link in the evacuation network were determined by traffic engineering analyses.
The link was described by the two nede numbers; the "A" nede at the beginning and the "3"
node at the end of the link.
A listing of the link character-istics was preparad relating the two nede nu=bers, the link distance, operating speed, and link capacity (the number of lanes times the assigned lane capacity).
The operating speeds and lane capacities, assigned by engineering judgement, were basically higher than that which would be exployed for other traffic engineering studies.
While these assignments appear high, they were conservative in view of other evacua",1cn studies.
"In cbservations dur:.ng evacuations, 1,100 to 4,080 cars per lane per hour were cbserved.
The average of actual cbservations was approximately 2,600 cars per lane per hour.
The average vehicle occupancy was four persons.
--- Vehicle speeds observed ranged f:cm 25 to 45 mph (with an average of 35 mph) during evacuatien. "(6)
(6) I3ID, p.
- 42.,
e
~..
Identificatien of Special Problem Areas - The input plann-ing data was searched for special prcblem areas.
These were identified and 1ccated en the respective EP:: base maps.
These problem areas included schools, nursery schools, hospitals,
nursing homes, j ails, recreational areas, beaches and the like which represented large concentrations of population within the EPZ.
s Centroid Pcpulation and Averace Car Occupancv - Another input to the EVACUATION TIME ASSESSMENT PROGRAM ccnsisted of the centroid population and average car occupancy to be empicyed in estimating the volume of vehicles to be evacuated from each of the EPZ centroids.
Although it was anticipated that this latter figure might vary from site to site, analysis of average household occupancy, from current census data, indicated that this statistic varied only slightly, in most cases from about 2.7 to 3. 3 persons per household.
Based upon this small variation, the figure of 3 persens per vehicle was used.
Evacuation Time Assessment Program The evacuation time assessment was perforrad in two separate packages of computer programs.
The general flow of these packages is illustrated in Figure 2.
Build Evacuation Network - This portion of the program is accomplished, utili:ing the link /nede descriptiens from Evac-uatien Route Development task.
The computer program, utilizing well-established principles, was extracted from an in-house program, TRANSIT, which was basically created for transit route development.
It organizes the input data and it asse=bles link data such as distance, speed and capacity.._
e' F' Bilt 8t hrea>&
Pem4 04 estruesht l4n'e El63Bibulleest BInt titialbullon 6844848800 tima khte8Pitu#3 Avsm A4 4 II #I #88 38 Vt Ml(t a 8 I m& #seubt 8 0 telsIN. Avt 4 &et
&V8Cuait #Ndd ttAwa touht Nth Al8h uClaar AmC f Listaltglom 6ptag, (ApA(g gl N
N ewea o (cetusa runiat etsetesa Punt et me srun>&
t wat uA s s ue wantons paa siset tessiasmusson sent sistessostoej M 3medA tACM tih8 holt htCalvt teASN! ass 18A7ff90822W
'88 '"",jawa ct a s c,.,.,,.,,,,ia....
,,,,,,u.a u_.. t
......... a.t..e a
.t
..t vahvat isng vetamtn er enesta Aeeon as tsma tent encatments isnt I
reneones ( Ar A(a st tattaf AhA4V586 camiaaan se taitanAt m444 1'
id it hniht DetAs s tem ton E Ata ats=4ea asus a4 stana nt sotAt s4tAv Inne e Aos canteole to L a tt ahAt esLE4 14 f thmswt 1984 twatualltHe ting 0
4ALH ((Mikel$ 30 e mit an At muse BASIC FLOW DIAGRAM EV ACUATION TIMt; ASSESSMENT FIGUHE 2 C.&. L./J../.4. - /..
Assicnment of Vehicles to Each Network - Using centroid population and car occupancy inputs, the program commutes centroid vehicle volumes to be evacuated and assigns these volumes to each link of the centroid evacuatien network.
Re-maining steps are also computed in the evacuation time assess-ment routine.
Public Rescense Time Distributions - Since the predicted volume of vehicles entering the network from each centroid is a function of various public response times to the evacuation waming, it is necessary to establish quantification of these respenses by certain assumed conditiens.
Up to four public respense time distributions are ecmbined to assess evacuation l
times under the four scenaries - normal work, nighttime, summer peak and adverse weather conditions.
In quantification of these public response times, it was assumed that all responses are normally distributed in time.
That is, the percent of the pcpulation respending follcws an accumulated statistical nor=al distributic.n curve.
There are four public responses:
Receive Warnine, Leave Work, Travel Time-Work-to-Ecme, and Evacuate Home.
The combined distributien then determines the distribution of vehicles leaving any centroid.
The ccmbination process is illustrated in Figura 3 for distributions employed to simulate an average workday, when household workers are at work and children are at school.
Curve 1 in this illustration shews a linear approximatien of the accumulated distribution of population versus time af ter warning for the Receive Warning respcnse.
Fifty percent of the population are assumed to receive the warning 10 minutes after the original warning event.
Assuming a nor=al.
O 4
l h
4 i
1 M
S.
w
=
~
a o.
s a
Es.
3 I3 C
r r:
- 5 2
3 33* 3333
!2 e
i 3a 3
(
42 9
2222
~
a 22$*
1 q
\\
o z
w
\\
s
=
\\
w3 0
$5
~
a aw.
w
- g }- =
- a 3
a n
g 4
Y 3::TTj d
- :v 3 35 s
s*
- ? ? ?,
s g
-=
a e
g M
K.g
~
=
s'fe.
-E g
- ,O h
t f
g
- 2 A
\\_
4 O
9 g
g a
% g l
R
%s%
i s
i s%'%
C g U
's's
/
g 2
tf
\\
- 'n g
/
g g
%,,I@ #47,w,%,,%
t s
s s's
% g 4
% en, 9
N 3
s
's,
%A O
C of.7.=,
s
- 7Eq' %,
s
- iss, s
s
,., s g %
i i
i
- 3 i m.u:= 2.
aa m ann::v 5
m m
dis tribution, about 16 percent of the pcpulatien are assu=ed to receive the warning 5 minutes after issuance (or ene-sig=a-en the normal distribution curve)'.
This also implies that in 15 minutes, 84 percent of the pcpulation will receive the warning after 15 minutes.
The accumulated normal distribution between these three points can be approximated by a straight line with insignificant error, as far as its application and other variable variations are concerned.
The distributien empicyed is truncated at a time qual to two-sigma to zero and 100 percent at zero time and 20 minutes respectively.
Curie 2 in the illustration, represents the approximate time distribution of pecple leaving work.
The distribution assumes that within 10 minutes af ter receiving the warning, 50 percent of the workers leave work.
Curve "I and 2" therefore represents the combinaticn of distributions 1 and 2, and represents the time distribution of people receiving the warning and leaving work.
1 Curve 3 describes the distribution of workers ' time to travel frcm work to home.
When combined with the first two dis tributions, the response time becomes the curve shewn as Curve "1,
2 and 3".
This ccmbination represents the distri-bution of people who arrive at heme.frem work after the initial 4
warning.
Curve 4 indicates the distribution of time the family will evacuate home.
This curve, unlike the first three in the illustration, assumes that the average family will evacuate the hcme in 20 minutes after the evacuatien period starts, here initially assumed to be the arrival at hcme of the worker.
In the evacuate-home distributien, 16 percent would evacuate home within 10 minutes and 84 percent in 30 minutes.
The curve truncation to zero and 100 percent at 15 minutes and 75 minutes respectively.
The distributien is represented by four straight line segments as -illustrated. _
y y
,y--
y
-g
--gy-
-w
-g,,,+-w-
Note that the linear approximation of the four combined response distributions is zero at 15 minutes.
prior to this time. it is therefore assumed that no vehicles have departed from any centroid during the period.
In other scenarios, two groups of distributions were employed.
Nighttime conditiens were assumed to be based upon everycne being at heme.
Therefore, distribution curves two and three in Figure 3 are omitted in Figure 4.
Only the distribution for Receive Warning and Evacuate Home are used.
The resultant combination and its linear approximation show that 15 minutes after initial warning, 16 percent of the population are assumed to evacuate, 50 percent by about 28 minutes, 84 percent by 45 minutes and 100 percent by 60 minutes.
The third group of distributions employed are shown in Figure 5.
These responses represent the public response distributions maintaining during a period of bad weather.
Three respcnse times, Receive Warning, Leave Work, and Evacuate Home are assumed identics1 to the normal workday. conditions since 4
these responses should be unaffected by weather conditions.
Ecwever, the work-to-home distribution is twice that used for normal workday conditions to reflect an assumed halfing of travel speed due to weather.
Later in the assessment program, operational speeds on the network are dropped to one-half of the value assumed for daytime evacuations to account for the lower speed in returning home.
The combined distribution for bad weather shows that the same 15 minute. mobilization time is experienced, but 50 percent of the population did not evacuate the home until about 52 minutes had elapsed af ter initial warning.
It required 35 minutes for 13 percent of the pcpulation to leave home, and about 75 minutes for 92 percent of the pcpulation to depart.
Ninety-five minutes elapsed before 100 percent of the pcpulatien had evacuated home. i
<~
e 10 0
,s j
,s
/
/
j
,s'
,/
/
l j
f'r/
a l
r i
/
s s
j c' /
1
&/
l y
l 4/
l
/
$/
2 1
1 0 60
-l s'
5
/'
l a
r l
/
l V
fr a
o
~ g'.
- s-j 49 80 fs'
/
u El dl
/
~
accumutaito nacan
/
as tre or sortuvat j'
4 e
x/
/
t, Ng attl0N 4
2 3
4 3
6 F
4
((
/
[
/
/
/
-atctivt wanutus is so as soo 20 l
I
/-
. 2-tvacuait tnw uout a is ss so si se sz soo I
/
/
/
s
/
s'
/s s s' 0
I I
I I
I O
S 10 IS 20 25 50 35 40 45 SO SS 60 65 TO MINUTES AFIEN ALEHI COMBINATION OF TWO SEPARATE REACTION DISTRIBUTIONS FIGURE 4
%!4-e %4. /.4,
.at.*
60 0 f
,r 7
s s
s J
,s
/
s' s'
- /
s'
,/
,s
,s s
j y,
,L-g
..___g 4-
\\'
l
/
s' l
/
s' s'
0 J
l l
4 /
a 2
l
( 4'l f,
E, g
/
r
,g
,,p..
/g' 3
Es g
f i
f/
ll0 5
ci
/
=es s
~,
?g
$ 4Y N
4 4
0 s
$l-fi
,s 40 S,
.#/
7i o
s 4
E8 a#
s a(Ct*wut alte Pt acant 4[
at tea op ealtavat g
,s h( ACl40as t
2 3 4
3 6
i e g'
j 8 W(Cliv( E Ah864NG 16 50 e4 00 0 1
s 20 s'
/
j#
I'"II " "
W ** "
/
3 IhavtL escast e
46 33 30 6F 64 92 80 0 f
[
f' 4 tv4CUAll ikdasIsbett 8
16 33 30 C7 84 92 10 0 s
s'
/
s o
,s l'
,s' 4'
I-L#,.
I i
i l
i I
I g
o a
10 as 20 23 30 35 40 4$
so SS 60 63 70 To 80 es 90 93 MINUIES AFTEN ALERT COMBIN ATIONS OF FOUR SEPARATE REACTION DISTRIBUTIONS C A.. L.4 s.4u 4.
FIGUf1E 5
Determination of Mcbilizatien Time - Mcbilization time is defined as that period between the issuance cf the evacuation warning and the time taken for the last vehicle to leave any centroid under the specified scenario conditions.
In scenarios using the group of public response distributions shown in Figure 3, it can be seen that mobilization time is 75 minutes.
For these represented in Figure 4, the time is 60 minutes, while these in Figure 5, it is 95 minutes.
Time Distribution of Traffic Volumes en the Network - The traffic volumes previously assigned on each link of the evacuation network are then distributed over time in four incremental periods, as indicated in the ecmbined respense distributions of Figures 3, 4, and 5.
Fo r exa=ple, in Figure 3, no vehicles entered the centroid evacuatien route until 15 minutes had elapsed.
There then followed four 15 minute intervals, each representing a certain percentage of total centroid vehicles entering the network.
These latter four-interval distributions were used to determine the network input volume end link volume distributions in time.
Capacity Celay Analysis - The capacity delay analysis is performed in the assessment program by the four time increments determined in the above step.
It is based upon the rudimental principle of queuing -- which is, if the input to a network element during a specified time period exceeds the service capacity of that element, a queue of input vehicles is for=ed and a delay is generated to those vehicles in the queue.
These vehicles must be added to the input of the next interval and ccmpared to the service capacity to dete_- nine if another queue is formed at the end of that interval.
The process is con-tinued until all vehicles to be serviced have past through the element.
The process is illustrated in Figures 6 through 9.
i 1 i
)
.,-._.m.
=
v, P V, 3
w 5
s>
u E P, V, 5
TIME DISTRIBUTION OF INPUT TRAFFIC VCLUME p, y, T,
T, T,
r.
TIME Flouma e I ro l V
,/ 0 f 1/
/
P,V,
/
/
Os w
h/
/
/
A
" P,V,
,;h, Y,8 I
/ "t a:
/
/
FORMATICN OF QUEUES
/
ANO DELAY RESULT 1NG P V, FROM LIMITED CAPACITY T,
7; T,
Te il M E h J as ar >--
eleumE 7
v, + v, V, + V' V
j v
/
/,
s i
/
+#s oi
\\
g,/
/
c
,,,s
,,.o...
g WERGING CF RESTRAINED 50 AND UNRESTR AINED TRAFFIC I
I, I,
I47pgl f
TlWE M TDl!
TD 2 l
v,,v, c,
O' /
j f-i
/
o, e-g
/
8 r #,e e
3 3
n.u...
1 CUEUE FORMAT 1CN g
dL AND AcclTICMAL oELAY REsutrinG FRCW UWITED OAPACITY ewle w '-
9 Figure 6 indicates the distribution of a traffic volume, over 4 time increments in which P1 is the percent of the V,
P2 the second and total assigned to the first time incrament, so forth.
Figure 7 indicates the condition where the input of the first and fourth intervals dc: not exceed the link capacity.
After the first interval, all vehicles in the input are served Hcwever, the input with no delay.and no queue is formed.
during the second interval exceeds the capacity, shewn as a At the end of rate of vehicles served per time increment.
second time interval, the inr:ut has resulted in a queue of vehicles shewn as Q2 (Q1 being zero).
This volume is therefore At the end of this added to the input of the third interval. These vehicles are period a queue of Q3 vehicles is formed.
added to the input of vehicles during the fourth interval because that volume exceeds the capacity, a number of vehicles Q4 remain at the end of the period.
This queue must then be A
dissipated at a rate equivalent to the link's capacity.
delay of TD is required for this discharge and must be added to the four time increments to obtain the total time for the volume to pass through the link.
It is important to note that this process has " metered" As the traffic proceeds the input to equal the link capacity.
to the next link of the evacuation route, its input is at the If no rate commensurate to the previous link's capacity.
additional volume has been assigned to the subsequent link and that link has the same capacity as the upstream li.d, no
~
~
additional delay is experienced.
If either the capacity or volume of the dcwnstream link the analysis procedure must be repeated, using is different, the respective input volumes and capacity of that 11.2.
4 i
i Figure 8 illustrates the circumstance where two volumes t
are merged at the input node of a link (the "A" node).
One volume, represented by Curve 1, has been limited upstream by a limiting capacity.
The vehicles have been delayed by a period TDl.
The other volume, shewn as curve 2, has not been capacity-restrained it therefore follcws the original time distribution.
The volumes of the two inputs at the end of each incremental period are su=ced to generate Curve 3.
The resultant total input volume by increments is sub-jected to the same analysis procedure as shown in Figure 7.
The analysis is illustrated in Figure 9.
The input volvme distribution is shewn as Curve 1.
During the first interval that volume exceeds the service capacity, resulting in a queue of Q1 vehicles.
At the end of the basic fcur-interval period, Q4 vehicles remain the the queue.
At the end of the input period, considered as the original period plus the previous delay of TDl, 05 vehicles remain to be served.
These vehicles must be dissipated at the rate equivalent to the link capacity, and results in an additional time penalty of TD2.
Ceterminatien of Delay Times for Each Link - The evacuation routes for each centroid are analyzed using the delay analysis technique described above.
The delays, if any, are assigned to each of the links.
Previous delays for any link resulting frem the analysis of another centroid for the same link are ecmpared in the program.
Appropriate adjustments to each link delay are made by the program and the proper delay assigned.
Cetermination of Link Travel Times - Travel times for each link are computed by the assessment program using the link distance and the anticipated link operating speed inputs.
These travel times assume no capacity delays.
Therefore, when the travel time for a link is added to the. prgper delay time, the actual speed for the link is represented. 4 w
r
-v
Cetermination of Total Delav Times - Total delays due to queuing. delays and travel time are dete_ ined by the pregrams by tracing the network from centroid to external node (leaving the 10-mile EPZ) and summing all delays in the evacuation route.
Determination of Total Evacuation Ti=e av Centroid - The total evacuation time, or that time for the last vehicle from a centroid to cross the 10-mile radius of the IPZ, is the sum of the total delay times (travel time plus delay time) and the mobilization time.
This time is computed by the program and provides the evacuation time for each centroid.
These results are then compared to determine the time for all vehicles to evacuate the IPZ.
Evacuation Route Refinement An initial evacuation time assessment program was run.
The resultant evacuation times for each centroid using the initial network were reviewed for reasonableness.
In every area assessed, some i=provement in evacuation time was indicated.
Rafinement of the networks and further subdivision of pcpulation centroids were made.to lessen the evacuation time for selected centroids.
These refinements were then confirmed by a re-run of the assessment program.
Not all high evacuation times could be improved however, because of the limited in-place highway j
and street facilities.
In most instances, these became the limiting cases in the assessment.
The refinement process is shcwn in Figure 10.
Scecial Problem Area Evacuations Special evacuatien problem areas within each IP: were iden-tifie'd and their locations shcwn on the appropriate base maps.._.
EVACUATION ROUTE 4-DEVELOPMENT REFINEMENT v
EVACUATI ON TI ME 4_,__
ASSESSMENT PROGRAM I
l v
I I
EVACUATICN TIME BY 3
CENTROID i
l EV ACU ATION ROUTE REFINEMENT i
- i/h L,74 and du
/,.
F1G U R E 10
The number of persons assigned in each were identified in the reports for each EP2.
Certain considerations and assumptions were made in their regard in this assessment.
School Evacuation - Based upon Asserption 1 en page 3 --
that public evacuation would criginate frem the home as a unified group -- it was likewise assumed that schcols would be notified directly.
It was f urther assu=ed that schcol childr,n would be dispersed to centrali:ed pick-up points near their residence such as fire staticns, police stations, or other public places where, supervision might take place.
The parents could either pick the children up during the 15 minute period between initial evacuation warning and first evacuation of a hcme, or on the way cut of the cent =cid frc= home to external nede.
No additional delay in evacuaticn time =f the general resident public would be entailed due to school pcpulaticn.
Seaches, Parks and Seascnal P.ecreation Areas - The distri-bution of public respcnse employed for this scenario assess =ent was the same as that for the nighttime evacuation.
No worker had to leave work, no children were in school.
The exception-to this generalization occurred in the assessment of evacuation times for Maine Yankee.
Two su=mertime scenarios were censidered here.
One used the distribution for the weekday conditions.
A second scenario, that of bad weather, cc=bined the su=me.~.ime peak pepulation widi the distribution of public respense assc=ed for bad weather.
Further discussion of thic is made in the section of scenario evaluations.
In theory, the evacuation of beaches would precede that of the general public.
This was due to the absence cf the Leave Work and Travel Time-Werk-to-Home response distributions.
Ecwever, technical difficulties were encountered when atta= pts were rade to co=bine the two -
1 i
i network volumes.
It was therefore decided to assign a centroid for these locations in the network, assign a pcpulation of "1" during other scenario conditions, and assign a population commensurate with the transient population at the centroids for the summertime evacuation scenario.
In three areas, Millstone, Maine Yankee and Shoreham, there were significant changes in resident summertime pcpulation involving certain centroids.
This required the creation of a new centroid population input for +he evacuation time assess-unt program for this category of scenario evaluation, which produced higher than expected time distributions of volur.es to the network.
The approach was considered acceptable since the results were conservative, but proved not to be worst case scenario.
Evacuation of Hoscitals - These special problem areas were carefully documented.
However, to perform precise evacuation time assessrents would have required the creation of detailed evacuation plans.
The question then resolved to what effect would evacuation of these facilities have on the limiting evacuation time of the EPZ?
The subject of who would require special evacuatien was also considered.
There appear to be three general categories of patients:
ambulatory, confined and intensive care patients.
Some of the first categorf could be released to their families for evacuation.
The remainder of these patients could be evacuated to receiving hospitals outside the EP3 in high occupancy vehicles.
Patients confined to bed would require transportation by ambulance.
Intensive care patients were censidered.
It j
was concluded that intensive care patients would be exposed to a greater risk in an evacuation than letting them remain in the hospital in areas of the building =eeting radiation ahelter, _-
require =ents.for plume fallout.
Ecspitals generally have such areas.
Therefore, the third category of patients would not be evacuated.
The conclusien was that the evacuatien of hespitals could occur simultaneously with the general public evacuation.
Such evacuations could be cc=pleted within the limiting ?vacuation -
time of any EP: under any scenario condition.
The relatively small nu=ber of additional vehicles en the network would have insignificant effect upcn the general public evacuations.
Jails - Only a few EP 's enccmcassed this special problem area.
The evacuation of these facilities was assumed to be accc=plished by special buses.
Again, the evacuation by these means could be accenplished during the limiting evacuation time for any scenario.
Only a small nt=ber of additional vehicles would be on any evacuation rcutes.
Their presence would not significantly affect the evacuation time en any evacuation route.
Evaluation of Scenarios A series of scenarios were examined to test each EP: for its most optimistic evacuation times and to develop the worst case evacuation times.
Four categories of scenarios were con-sidered in the assessment:
normal work day, nighttime, bad weather, and su=merti=a population, both resident and transient population at entertainment areas such as beaches, parks, and other = ass - entertainment centers.
These scenario evacuation ti=a assessments were conducted by the above procedures using the conditions as shcwn in Table 1.
1 -
Table 1 VARIABLES EMPLOYED IN SCENARIO ASSESSMENTS PUBLIC RESPONSE DISTRIBUTIOd MEAN TIMES (b0% RESPONSE)
Travel EVACUATION Receive Leave Work-Evacuate SCENARIO Warning Work to-Eome Home OTHER Normal Workday 10 min.
10 min.
10 min.
20 min.
Nightti:ne 10 min. Not Used Not Used 20 min.
Summer Peak-Resident 10 min..10 min.
10 min.
20 min.
Bad Weather 10 min.
10 min.
20 min.
20 min.
Reduced Network Speed By 50%
D ^
4 l
It will be noted that in the had weather scenario, cen-ditions expected over an average year were examined.
It was the concensus of this review that icy roads constitute the most hazardous weather condition to be encountered annually in eight of the nine EPZ's assessed.
These weather conditions were modeled in the evacuation time assessment program using the same centroid populaclon and ne twork as that employed in the normal workday or scenario.
However, the speeds on the network links were reduced by 50 percent to account for weather impacts on traffic operations.. The public reaction distribution representing the Travel Time - Work-To-Home was doubled for the same reason.
The exceptional EPZ not assessed in this manner was Maine Yankee.
In this EPZ, the above criteria was used, but the had weather condition proved to be summertire fog conditions.
People may be at beaches when it occurs, so resident summertime population was combined with beach population.
This centroid population was then used to assess evacuation tire.
Heavy snowfall occurs in most areas.
In some of these EPZ's, disabling weather conditions cccur every five to six years when traffic may not move for several days.
These occurrences were considered as exceptions to the evacuation time assessment, because of the very small and almost insignificant probability of the simultaneous occurrence of both an evacuation warning and a snowfall occurring during the same days.
This probability of simultaneous occurrence could be best addressed in the development of other plant operational strategies to be effected in these rare cas.es than attempts in any meaningful way to address this problem in an evacuation time assessment.
Flcod hazards were examined for all pcwer plant areas.
Bea' Valley, Three Mile Island and Limerick IPZ's appear w be the most endangered.
The remaining pcwer plant areas border.
---,w.,-.--
r
.-y-y 9
-w.
I w--mi.-.-~w e
re-+.
f on large bodies of water with little exposure to large volure
- rivers.
The floed level histories for the three areas were reviewed.
No link interruptions were discovered, with the exception of a catastrophic bridge failure, which would disrupt evacuation.
These failures to a network would be handled by re-design of the affected evacuation route.
The incident was not included in this evacuation time assessment.
1 Sensitivity Analysis The computer program provides means of rapidly evaluating changes in the input values, including quantificatiens of the assumptions of public response times, speeds, capacities, volumes, and other variables.
Figures 3, 4 and 5 illustrate the sensitivity of the mobilization time, and, censequently, the sensitivity in evacuatio.1 time to variations in the various public response distributions employed.
The pregram can receive any distributions at any interval of minutes desired and apply them in the assessment.
J Evacuation time is a function of travel time and capacity i
delay times.
In the limit, capacity delay time was significantly larger than travel time.
This implies directly that operating speed; a dependent variable of travel time, had relatively litule influence on evacuation time.
Changes in capacity influence the delay time directly.
Therefore, the value chcsen for the assess-ment was an important, direct result of this input variable.
1 1
METHOCOLOGY FOR CONFIPJ4.ING EVACUATION Considerable thought was invested under this study topic.
The results were such that no specific methodology for confirming evacuation evolved for any specific site.
It was decided *w incorporate some of the concepts in a general manner in this volume to be applied to all areas.
In light of the assumption that not all people will choose to evacuate (Assumption 4 in the INTRODUCTICN), the important interpretation of this topic should be rephrased to include the i
deter dnation of persens receiving the warning and either choosing to evacuate or not, as their individual preferance dictates.
Thus, the topic erbraces not only the confir=ation of i
receipt of warning, and the public response thereto, but includes many interlacing consideratiens of the degree of security and protection of personal property during the evacu-ation, the nature, extent and operational considerations given to its patrol, and the nature of advanced public information provided to the general public residing in the EPZ and those transient populatien groups temporarily in the EPI.
A number of methode, were postulated during the study.
These were tested against feasibility and efficiency.
The salient concepts tested are discussed in the following secticns along l
with the pros _and cons of the concept.
t 4
Visual Confirmation bv Helicopter At least one operating utility has developed a concept for public information on evacuation procedures which led to this concept.
Long Island Lighting Company, cwners of the SHOFIEAM l
f acility under construction, is considering an intreging public information program.
Each household would be supplied a packet l
i.
- - - ~
e-
of infor: nation which includes a windshield sticker to be placed i
on the windshield when evacuation is under way.
This sticker would expose to those. outside the vehicle the identification of the centroid assigned to the vehicle.
This would assist in traffic direction and control.
On the side exposed to the car occupant would be the evacuation route description to be used by residents of that centroid.
Similarly, packets would be available for transients at beaches, etc., to be distributed there as part of the warning system for that area.
If such a packet also contained a folded panel of suitable dimensions, the occupant of the household could indicate that he has received the evacuation warning by displaying the un-folded panel in his yard or on the sidewalk.
One side of the panel would have a contrasting color, say, orange.
Displaying this side would indicate that he has received the message and has evacuated.
By displaying the obverse side whose color is, say, red, he would indicate that he has decided not to evacuate.
Fly-over by helicopter could determine the status of the household.
These households without a displayed panel could then be checked by ground patrol.
Such a concept has its weaknesses.
It vould not be effective at night and peor visibility due to weather.
It requires back-up from a ground patrol.
It also has a weakness that it advertizes those houses which are vacant to would-be looters.
This would discourage cooperative use of the panel and reduce the advantage of aerial reconnaissancs, and increase the burden of the surface patrol.
Ground Patrol A system of confirmation of evacuation could be develeped employing ground patrol units.
This could be combined with a
4 security patrol of the area.
It is understcod that a form of this procedure is planned by Civil Defense.
Rather than re-quiring the patrol to ring doorbells, the occupant could be instructed to tie a-white cloth on the doorkncb to indicate he has evacuated.
This is again a possible invitation to looters.
Te1ephone'Confirmatien This concept would require a special telephone equip =ent to be effective.
The idea is to computerize the dialing of numbers of residents within the evacuation area.
Those who fail to answer would be assumed to have evacuated.
Those answering could be questioned as to their status.
The concept has negative attributes.
Not all residents have phones.
Its use could tie up the telephcne system at a critical ti=e when it is required for emergency purposes.
It is based upon the assumptien that no answer signifies evacuation, which may not be the case.
Ccmeuter Cards This concept would require the issuance of computer card to all residents.
These cards would be collected after evacu-ation, either at the shelters or when leaving the area.
The cards would then be processed by ccmputer to indicate evacuation status.
The weakness in this concept is the collection of cards.
It has been assu=ed that only a small percentage of evacuees use shelters.
The:Jefore, collection there appears impractical.
Collection of cards on the evacuation route would seriously hamper the network ficw and cause delays and congestion.
Suitable parallel collection points as are employed en toll facilities would be required to relieve congestien.
The capacity of such a collection system is 650 vehicles per hour per gate.
l Sis figure assumes easy, diracu approach and cnly handing card i
i j.. - - -.
to the collector at that point.
Implementation of this concept would require detailed analysis of the gecmetry of the roadway at each external node and the careful design of collection lanes at each of those points.
Processing by computer could be accomplished at a central point.
This processing would add insignificant time to the operation.
The major delay would be the transport of cards from collection points to this central location.
Of all concepts, this appears most promising.
The evaluation of confirmation time, hcwever, and the collection point design requires detailed study which was beyond the scope of this study.
W.
i RECOMMENCATICNS FOR IMPROVING EVACUATICN l
This independent evacuation time assessment indicates that I
a detailed traffic engineering approach to the design of effective evacuation routes can produce a reasonably satisfactory evacuation of an area.
Avoidance of the temptation to use high volume expressway facilities should be exercised and limited by ramp capacities.
The use of lower grade, radially criented routes should be encouraged.
This apprcach improves evacuation time at the expense of a more ccmplex evacuation network.
The success of such a plan depends heavily upon an adequate public information system where all residents are knowledgeable of their assigned routes and other details of the evacuation process.
Public Warning System The study disclosed that no EPZ had in place an adequate public warning system.
Such'a system should cccbine a general area acoustical warning system of sirens or horns to alert the public to tune into designated public broadcast radio and TV stations for instructions.
In remote areas, the use of special, automatically operated radio receivers might be necessary to 4
acenemical cover those areas.
The radio would present both an aural tone and a visual signal to alert the occupants.
Such a system would reduce the RECIIVE WARNING public response time.
Security During Evacuation It has been pointed cut in a number of places in this re-port that public concern over security of household has a drastic effect upon their decision to evacuate.
Adequate planning and
)
implementatien of a security patrol and the public discicsure of such a system should increase the effectiveness of the evacuatien. - -.
e.,
COMPARISONS OF EVACUATICN TIME ASSESSMENT AND EVACUATION PLANS In the independent Evacuation Time Assessment Study, a Certain assumptiens standard uniform methodology was employed.
Care must were necessary and were identified in the reports.
be exercised in attempting to compare the results of these In the four plans
]
assessments and existing Evacuation Plans.
obtained during this study, each employed a different methodology There fore, and were based upon diversely different assumptions.
comparisons among plans would be misleading without complete It follcws that comprehension of their differences in approach.
the same comprehension must be present when ecmparing their results with this assessment.
Y t.
i -
_