ML20082C573

Testimony of Mc Cordaro,Ja Weismantle & Eb Lieberman on Phase II Emergency Planning Contention 65 Re Evacuation Time Estimates

ML20082C573
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	11/18/1983
From:	Cordaro M, Lieberman E, Weismantle J LONG ISLAND LIGHTING CO.
To:
Shared Package
ML20082C575	List: ML20082C573 ML20082C585
References
ISSUANCES-OL-3, NUDOCS 8311220141
Download: ML20082C573 (234)

Text

q t#

LILCO,0gy&ber18, 1983

1 1

53 tsy 21 Ali:07 i

d UNITED STATES OF AMERICA!CE Cf SECWA

i NUCLEAR REGULATORY COMMI'SSIONl? l."["

~

Before the Atomic Safety and Licensing Board In the Matter of

)

)

lj LONG ISLAND LIGHTING COMPANY

) Docket No. 50-322,-OL-3

) (Emergency Planning Proceeding)

(Shoreham Nuclear Power Station, )

Unit 1)

)

--)

TESTIMONY OF MATTHEW C. CORDARO, f;

JOHN A. WEISMANTLE AND EDWARD B.

LIEBERMAN ON BEHALF OF LONG ISLAND LIGHTING COMPANY

{j ON PHASE II EMERGENCY PLANNING CONTENTION 65

i j

a

.J Hunton & Williams P.O. Box 1535 707 East Main Street m

Richmond, Virginia 23212 (804) 788-8200 3J hI s

', 1 r

i 8311220141 831118 h

PDR ADOCK 05000322

}s T

PDR o

LILCO, Novgmbar 18,.1983 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION Before the Atomic Safety and Licensing Board

' In the Matter of

)

)-

LONG ISLAND LIGHTING COMPANY

) Docket No. 50-322-OL-3

) (Emergency Planning Proceeding)

(Shoreham Nuclear Power Station, )

Unit 1)

)

TESTIMONY OF MATTHEW-C. CORDARO,

JOHN A. WEISMANTLE AND EDWARD B. LIEBERMAN ON BEHALF OF LONG ISLAND LIGHTING COMPANY ON PHASE II EMERGENCY PLANNING CONTENTION 65 PURPOSE Suffolk County Contention 65 states numerous allegations concerning the accuracy of the evacuation time estimates pre-pared by LILCO for a 10-mile radius Emergency Planning Zone at Shoreham.

The purpose of this testimony is to summarize the nature ar.d results of the multiple series of evaluations of evacuation times for the 10-mile EPZ, starting with a descrip-tion of the DhNEV analytical modeling system.

The testimony i

shows that LILCO has properly accounted for the time necessary l

l for persons to be notified of an emergency at'Shoreham and to l

make preparations for evacuation (Contention 65.A.), and that-traffic in the EPZ during the mobilization period will not L

materially hinder the commencement of evacuation (Contention 65.B.).

It'also shows that LILCO's proposed use of traffic control strategies will materially assist traffic flow

(Contention 65.C.), and that accidents, breakdowns and road characteristics and conditions have been properly accounted for (Contention 65.D.).

In these analyses, LILCO has accounted for the potential effects on evacuation times of use of ordinary traffic controls rather than special ones (" uncontrolled" cases), and of use by motorists of evacuation routes other than those assigned to them ("non-compliance" cases).

The analyses have accounted for the number of ordinary passenger automobiles used in an evacuation.

They have also accounted for the additional vehicles used to transport school children, hospital patients, the handicapped, and others requiring special care (Contention 65.E.), as well as for the additional time necessary to transport school children and per-sons with special needs (Contention 65.G.).

Finally, it de-scribes the functions and anticipated effectiveness of route spotters, including use of a spotter in a helicopter (Conten-tion 65.H.).

The results, which are consistent with those of earlier, less sophisticated analyses of evacuations of the Shoreham EPZ, and with results for other licensed nuclear plants, are summa-rized in a series of tables attached to the testimony, princi-pally Attachment 6.

This testimony does not consider the potential effects of psychological stress on evacuation times (Contention 65.C.2.

.~..

. and F.); this matter is treated in-the testimony of Drs. Dynes, Mileti, et al., being filed simultaneously with this testimony.

I I

,,.7-

-,-.w--p n.

y

-e,

-,,---.,,,y e,,, -.,

r-

,m.,

,.,.,,, _ -.,--,,,,,.v,,,,e e

LILCO, November 18, 1983 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION Before the Atomic Safety and Licensing Board

-In the Matter of

)

)

LONG ISLAND LIGHTING COMPANY

) Docket No. 50-322-OL-3

) (Emergency Planning Proceeding) s (Shoreham Nuclear Power Station, )

Unit 1)

)

TESTIMONY OF MATTHEW C. CORDARO, JOHN A. WEISMANTLE AND ETWARD B.

LIEBERMAN ON BEHALF OF LONG ISLAND LIGHTING COMPANY ON PHASE II EMERGENCY PLANNING CONTENTION 65 TESTIMONY' 1.

Q. Please state your name and business address.

A. [Cordaro]

My name is Matthew C. Cordaro.

My business address is Long' Island Lighting Company, 175 Old Coun-try Road, Hicksville, New York, 11801.

[Weismantle]

My name is John A. Weismantle.

My busi-ness address is Long Island Lighting Company, 100 Old Country Road, Hicksville, New York, 11801.

[Lieberman]

My name is Edward B.

Lieberman.

My busi-ness address is KLD Associates, Incorporated, 300 Broadway, Huntington Station, New York, 11746.

i

__ 2'.

Q. Please summarize your professional qualifications and your role in emergency planning for the Shoreham Nuclear Power Station.

A.

[Cerdaro]

I am Vice President, Engineering, for LILCO.

My professional qualifications are attached to this testimony as Attachment 1.

I am participating on this panel to provide the LILCO management perspective on emergency planning, and to answer any questions pertinent to management.

My role in emergency plan-ning for Shoreham is to ensure that the needs and requirements of emergency planning are being met, and that the technical direction and content of emergency planning are being conveyed to corporate management.

[Weismantle]

I am Manager of the Local Emergency Response Implementing Organization for LILCO.

My pro-feasional qualifications are attached to this testi-mony as Attachment 2.

My familiarity with the issues surrounding this contention stems from work in devel-oping and implementing the Local Offsite Emergency Recponse Plan for Shoreham (referred to interchange-ably as the LERO Plan or the LILCO Transition Plan).

[Lieberman]

I am Vice President of KLD Associates, Incorporated.

My professional qualifications are attached to this testimony as Attachment 3.

My l

l

_ familiarity with the issues raised by Contention 65 results from work KLD Associates has performed for LILCO on evacuation time estimates for the Shoreham EPZ.

This work has extended beyond the preparation of evacuation time estimates and has included some involvement in virtually all of the material that appears in Appendix A to the LILCO Transition Plan.

3.

Q. Would you please summarize the issues raised by SC Contention 65?

A.

[Cordaro, Weismantle, Lieberman]

Suffolk County Con-tention 65 questions whether the evacuation time esti-mates contained in Appendix A of the LILCO Transition Plan are accurate and reliable.

Specifically, the contention raises concerns that:

Contention 65:

Evacuation Time Estimates Further Preamble to Contention 65.

Section IV of Appendix E to 10 CFR Part 50 requires that license applicants " provide an analysis of the time required to evacuate and for tak-ing other protective actions for various sec-tors and distances within the plume exposure pathway EPZ for transient and permanent popu-lations."

(See also, NUREG 0654,Section II.J.8 and Appendix 4).

Accurate estimates of the time necessary to evacuate the Shoreham EPZ (or portions thereof) are essential to evaluating the evacuation route system.

In particular, such estimates must be accurate and reliable so that command and control per-sonnel who are considering what protective actions might be ordered for particular per-sons can estimate whether, given projected release and dispersion of health-threatening

fission products from the Shoreham plant, evacuation can be accomplished before such dispersion takes place.

(See 10 CFR Section 50.47(b)(lO); NUREG 0654 Section II.J.lO.m).

A1 decision to order evacuation, if based on inaccurate evacuation time estimates,'could result in evacuees' being trapped in queues or slow moving traffic inside or outside the EPZ, thus exposing them to a release-of fission products from the Shoreham plant.

LILCO has' submitted evacuation time esti-mates for the 10-mile EPZ, which estimates are contained in Appendix A, at V-3, and OPIP 3.6.1, Attachment 4.16/

LILCO estimates that the time for evacuation will vary from about two to two-and-one-half hours for only the

-inner EPZ sectors,.to a maximum of approxi-mately six hours for evacuation of the entire EPZ under adverse weather conditions.

Contention 65:

Intervenors contend that LILCO's evacuation time estimates are inaccu-rate, unreliable and, in fact, should be far longer.

LILCO's evacuation time estimates are so underestimated that under the LILCO Plan an evacuation may be ordered which realistically cannot be completed prior to release and dis-persion of fission products from the Shoreham plant., Evacuees will be caught in queues or delayed in heavily congested traffic within the EPZ.

Under many accident conditions, there will be a dispersal of radioactive mate-rials while such traffic conditions still exist, resulting in unacceptable health-threatening exposure to the evacuees.

The automobiles of the e,acuees will offer essen-tially no protection from the plume.

,1_6/

The FEMA Report at 11-12 notes that the time estimates are inadequate in part because the estimates in OPIP 3.6.1 are incomparable to those in Appendix A.

,. The specific deficiencies in LILCO's esti-mates and further bases for this contention are set forth in paragraphs A-H below.

Contention 65.A.

The LILCO evacuation time estimates ignore or underestimate the time required for people to mobilize and ready themselves for evacuation.

The LILCO esti-matesMin Appendix A include only the time

~

involved in the actual evacuation trip out of tte EPZ.

(Appendix A, Table XIV).

LILCO assumes in OPIP 3.6.1 that complete mobiliza-tion of'the public will take about 20 minutes after receiving notification, which grossly underestimates the time it will take for mobilization, especially during working hours.

In fact, it will likely take at least from one i-to more than three hours for people to mobi-lize before they can begin to evacuate.

This mobilization time will be required because:

l.=

Following activation of the prompt notification system, it will take time for people to become aware of the-emer-sgency, to become informed of the recom-L mended protective actions and to deter-mine their own course of action.

2.

Where possible, most families will seek to evacuate as a unit.

Specifically, working parents will leave work and drive to schools and/or.home to pick up l

their children prior to evacuating.

There will also be travel to and from various locations as family groups are assembled from work locations, rela-tives' homes, day care centers, and the like.

Mobilization time must include time for the travel necessary to assem-ble family groups.

In addition, fami-lies with school children who do not pick up their children themselves, will delay the start of their evacuation until all their children have returned home.

Given the length of time neces-sary to implement early dismissals (see Contention 69), mobilization times could be increased significantly'by this fact.

~,-

-4,-

,*e.,---

-r..-

ww,--rec,w yr,----e--,.

w.--,,-.,.,.-,..--.--,yer--er

,..-%v

-,w--,r-,r,,-,ece--

,,y

.~ - _

t

<. i I

3.

It will take time for the evacuees to gather necessary provisions before evacuating..(See " Emergency Proce-l dures:. Shoreham Nuclear Power Sta-tion," at 8).

In addition, some per-sons will seek to go to banks, stores and other such facilities for money and provisions.

i 4.

Travel within the EPZ during the mobi-lization perfod (work /home, home/

4 school, to banks and stores, etc.~)

i prior to commencing evacuation will i

result.in heavy traffic congestion which will lengthen the-time.necessary L,.

to complete mobilization travel.

f Contention 65.B.

Heavy traffic congestion from mobilization traffic, due to both high I

demand and conflicting traffic flow (i.e.,

some traffic flow in directions different than prescribed evacuation directions), will lengthen evacuation times.

LILCOfs evacuation time estimates do not appear-to take this causa of congestion and resulting evacuation

~

delay into consideration.

Thus, the LILCO estimates are inaccurate for this additional reason.

i Contention 65.C.

The LILCO traffic control-plan, as described.in Appendix A, even if l

assumed to be lawful and capable of being

[

t implemented, will, in fact, constitute an additional source of congestion which has been ignored in LILCO's evacuation time estimates.

If.such congestion were taken into account, the LILCO estimates would increase substan-tially.

The Plan will cause additional con-l L

gestion for the following reasons:

1.

LILCO's estimates assume that its traf-

.fic guides will screen all motorists moving in a direction contrary to its prescribed traffic flow to determine

[

whether each person has " good reason" for going in that direction.

(Appendix A,

at IV-83; see also, IV-8).

Thus, a traffic guide presumably would stop or otherwise delay all such motorists, i

s

_._,.,___,.,_-.._,,._-,,...,,.-__.__...r,.

question them, and attempt to persuade or order them not to go in their in-tended directions if their reasons for doing so were judged not to be suffi-cient.

This screening process will im-pede traffic flow, resulting in conges-tion and further increasing the evacua tion time estimates.

It will also re-quire more traffic guides than LILCO has designated for each traffic post.

2.

LILCO's attempted use of traffic con-trols may cause aggressive behavior on the part of those attempting to take protective actions.

This aggressive behavior will stem in part from fear of a radiological emergency (which is per-ceived by the population to be differ-ent from other emergencies) and in part from confrontations that will result when motorists wish to travel contrary to the directions of the LILCO traffic guide, or are stopped by guides for screening.

Conflicts between motorists and traffic guides will result in traf-fic blockages, confusion, accidents and possibly injuries, all of which will increase congestion.

3.

Because under the LILCO Plan neither LILCO's traffic guides nor any other LERO personnel will alter traffic sig-nal lights, traffic guides may attempt to implement a control strategy counter to the direction given by the signals.

(See EEMA Report at 10 citing non-compliance with NUREG 0654,Section II.J.lO.j).

Such simultaneous and po-tentially contradictory instructions to motorists will cause confusion and con-gestion, thus further delaying traffic movement.

(Id.)

4.

In some cases, LILCO's prescribed routes direct motorists to travel con-trary to their perceptions of the most expeditious way out of the EPZ.

(See, e.g.,

Post #19 described in Appendix A, at IV-56).

This will cause confusion

i,

and anxiety on the part of the motor-ists and confrontations with traffic guides.

Contention 65.D.

The LILCO time estimates

- assume that "[njo major vehicle breakdown or other types of incidents [will] occur which block major routes for an extended time."

' (Appendix A, at V-2).

This assumption is unrealistic and leads to an underestimation of the time required for evacuation.

Examples of 4

factors which increase congestion and thus in-crease time estimates, and which should have been included in LILCO's estimates, include:

1.

Anticipated traffic accidents and auto-mobile breakdowns, including running out of gas (for example, the Suffolk i

County police responded in 1982 to.

10,000 incidents such as accidents and breakdowns on the-Suffolk. County por-tion of the Long Island Expressway, thus indicating the potential for this factor to influence severely evacuation times);

2.

The absence of shoulders on some pri-mary or secondary routes which will be used during an evacuation; 3.

Road construction / repair work which can be assumed to be ongoing at any time; and 4.

Abandonment of vehicles under emergency i

conditions.

Contention 65.E.

The LILCO evacuation-time i

l estimates do not take into account the addi-tional congestion to be encountered by evacua-t ting motorists that will result from the evac-untion and early dismissals of schools and the evacuation of those in special facilities and f

the handicapped.

Such evacuations and dis-missals will involve the use of large numbers of buses, ambulances and trains which will be traveling in all directions through the EPZ, on prescribed evacuation routes and other roads,' making frequent stops.

If the impact i

r...

w.

,m,-,,.~---.

-,-.-,-_..-,,,,_-,..,.-~c,-,_.,,.--..m,,--,m,,,.w.

., i i

i I

of speciai evacuations were taken into.

account, the LILCO time estimates for evacua-ting motorists would increase substantially.

Contention 65.F.

Behavior research demon-strates that stress and= anxiety induced by a radiological emergency at Shorehem will dimin-l ish driving skills and awareness, and impede the processing of information necessary for a driver.to make decisions and drive properly.

The geography of Long Island, with its narrow, limited land area, may create a feeling of being " closed-in," which may increase the likelihood of poor driver behavior.

Decreased driving skills and driver awareness will cause confusion, congestion and accidents and, if properly taken into account, would increase LILCO's evacuation times.

LILCO, however, has failed to take these factors into account in its evacuation time estimates.

Contention 65.G. -The LILCO Plan does not include evacuation time estimates for evacua-tion of those with special needs who cannot rely on private transportation, such as school children, persons without access to cars, per-sons in health care or other special facili-ties, and the handicapped.

(See FEMA Report at 11, citing noncompliance with NUREG 0654,

- Section II.J.10.1 and Appendix 4, at 4-9 to 10).

The individuals in. charge of making protective action recommendations must know how long it will take' to evacuate these por-i.

tions of the population.

The Plan thus fails to comply with 10 CFR Appendix E,Section IV, and NUREG 0654,Section II.J.8 and Appendix 4.

Contention 65.H.

The LILCO Plan (OPIP 3.6.3) provides for two evacuation route spot-ters.to report information'to the EOC regard-ing traffic congestion on evacuation routes.

(Contrary to the requirement of NUREG 0654 L

Section II.A.2.a.,

the LILCO employees' expect-ed to fill these positions are not identified-by job title in the Plan.

(See OPIP 2.1.1, at 32.)

Without the ability to spot congested areas effectively, LILCO will be unable to im-plement appropriate measures for evacuees to 1

avoid such congestion, resulting in 1

1

+,e

+,. -..,.

6,,,

,,,,,.-,-,-,-n,w-n.,-,,.,

n.- m ~,-v

.,-r

.--,.,,e-m.n.,

,,,.,_n.w,

,a,,,

,n.,,,+,,-n,..,

,.~c,

m

.,. 1 increased evacuation times.

LILCO's-route spotters will be ineffective because:

1.

LILCO has not provided enough route spotters to cover the evacuation routes.

(See FEMA Report at 11).

2..

The LILCO route spotters will be unable to move expeditiously through heavily i

congested traffic, especially since the evacuating motorists will not defer to LERO vehicles operating without police sirens'or flashers.

(Id.)

4.

Q. Could you briefly describe the scope of this'testi-mony?

A.

[Cordaro, Weismantle, Lieberman)

This testimony will address all parts of Contention 65, except Contentions 65.C.2 and 65.F.

Those two subsections of Contention 65 raise issues about human behavior which are more _

appropriately addressed by behavioral scientists.

Accordingly, those two subsections of Contention 65 are addressed in the testimony of Matthew C. Cordaro, Russell R. Dynes, William G. Johnson, Dennis S.

l

.Mileti, David N.. Richardson, John H.

Sorenson, and 1

John A. Weismantle that is also being filed today.

This testimony will begin with overview explana-tions of the traffic model used to calculate evacua-tion times and of the cases studied by LILCO and the results of those studies.

The testimony will then l

focus on the specific subparts of Contention 65, l

l i '.

bearing in mind this Board's Order of August 19, 1983 which admitted Contention 65 but limited the discus-sion of alleged deficiencies to those contained in the

~

subparts of Contention 65.

It should be noted that there is a group of topics referred-to in various subparts of Contention 65, whose basic substance is centered in other conten-tions.

Included in this group is the treatment of accidents (Contention 65.D.; Contention 66), insofar as they relate to the creation of obstacles which must be removed from evacuation routes.

Also included are the tennbility of LILCO's proposed means of evacuation and estimates for evacuation times for persons without access to automobiles (Contention 65.G.; Contention 67), school children (Contention 65.A.2, E.,

G.; Con-tention 68-71), persons in special facilities (Conten-tion 65.E., G.; Contention 72), and handicapped per-sons at home (Contentions 65.E., G.; contention 73).

The assumptions used by LILCO to model evacuation means and times for these populations are discussed in Contention 65 to the extent necessary to address con-cerns within the scope of that contention, i.e.,

as they related to cverall evacuation times.

The bases for and merits of those assumptions will be presented

4 in detail in.the context of the more specifically fo-cused contentions, each of which is a Group II issue.

5.

Q. Where do the concerns raised in Contention 65 fit into the overall concept of emergency planning?

'A.

[Cordaro,-Weismantle, Lieberman]

Contention 65 raises 4

concerns about traffic and other logistic aspects of an evacuation of all'or part of the Emergency Planning Zone around the Shoreham Nuclear Power Station, should 1

such an evacuation be ordered as a protective action following an accident at the plant.

Emergency plan-ning for Shoreham, as for all commercial nuclear power plants, is-structured against a background of a series of core technical / regulatory documents prepared by the NRC and other agencies, including the Environmental j

Protection Agency and the Federal Emergency Management Agency.

Of these, the principal ones are NUREG-i 0654/ FEMA-REP-1~(Rev.1), Criteria For Preparation and Evaluation of Radiological Emergency Response Plans and Preparedness in Support of Nuclear Power Plants 1

(November 1980) and NUREG-0396/ EPA-529/1-78-Ol6, Planning Basis for the Development of State and Local l

Government Radiological Emergency Response Plans in l-I Support of Light Water Nuclear Power Plants (December 1978), as well as the NRC's regulations, 10 CFR i

l' 4

., 5 50.47 and Part 50 Appendix E.

Though these docu-ments form a backdrop, this and other testimony will address them only to the extent they bear on issues in contention here.

Similarly, this and other testimony will not at-tempt to summarize, in any one place, the contents of LILCO's entire four-volume Local Offsite Radiological Emergency Response Plan for Shoreham, but simply to make use of those portions which are particularly rel-evant to those issues actually in contention.

With respect to evacuation planning, the principal portion of the Emergency Plan of interest is Appendix A, the Evacuation Plan, along with two Offsite Emergency Planning Implementing Procedures, particularly OPIP 3.6.1 (Plume Exposure Pathway Protective Action Recom-mendations) and OPIP 3.6.3 (Traffic Control).

In focusing on the detailed but narrow area of evacuation planning, and particularly time estimates, f

several precepts from the general emergency planning 1

literature referred to above should be remembered:

r Firat, the basic purpose of radiological emergency planning is to achieve dose savings to the general I

public.

We regard the first line of defense to the I

public as a safely constructed and operated nuclear l

l i

plant, but, in the event that ndtwithstanding all our precautions there is a serious accident, like the one at Three Mile Island, for example, an emergency plan should be in place to reduce the public health conse-quences of that accident as much as can reasonably be done.

Second, it is not possible to guarantee, and the law does not require, absolute protection for every member of the public against radiation no matter how serious the accident.

No matter how safe the plant, and no matter how good the emergency plan, there is always the possibility, however remote, of an accident so severe, coupled with meteorological conditions so adverse, that some people would receive significant doses of radiation.

This is true of every commercial nuclear plant in the country.

The purpose of the emergency plan is to reduce those worst-case doses as much as possible to as many people as possible.

l l

Moreover, it is inappropriate to focus exclusively or even principally on any single accident sequence, whether one calls it the worst possible or not.

NUREG-0396 tells us not to concentrate on a single accident but rather on a spectrum of accidents.

The point is that the emergency plan has to set up a l

l

~ -. - - -.

.,n

~ -.- --

_ __ 4

^

. resource base that provides the necessary flexibility l

to respond to any kind of a wide range of potential

. accidents, whose severity, timing and other character-istics will suggest the appropriate protective action or actions (sheltering"and evacuation-being the prin-cipal ones), along with the area.over which they should be applied.

i Third, as a result, there are few, if any, abso-lutes in emergency planning.

There is, for example, no requirement that emergency planners be.able to evacuate the public in any specific minimum amount of time.

Nor is there any dose level which the emergency planners must be able to guarantee that people will never receive.

Th'e EPA Protective Action Guides are not such criteria.

The EPA Manual of Protective Action Guides (EPA-520/1-75-001) points out (on page l

1.1) that a Protective Action Guide "under no circum-stances implies an acceptable dose."

Instead, PAG's are trigger levels, decisional tools to help the deci-i l

sionmakers decide whether to evacuate, whether to l

shelter, or whether to do something else.

t l-To determine the' adequacy of emergency planning at l

l any site, one relies primarily on the guidance of i

NUREG-0654.

In addition, one must understand that any l

i*

==-im ww--"-

-g--g---*ur ee'-9wex--

m-

-s

-T-e-P==9yv' tF 7-tt'-

P-t "M-' maw *J'*rF*Tf 7 *W t""

T*t""TmMP-S*"W'-*

P""*T-"N-V""'MD "T

$'"W"D'

_-. --- _- emergency plan is restricted to some extent by the available resources.

To oversimplify somewhat, the emergency planner's' job is to marshall existing re-sources in an effective manner.

He is not expected to commit massive additional resources to the effort.

For example, NUREG-0396 expressly does not recommend that blankets be stockpiled, hospitals be built, and so forth.

The principle that evacuation planning is done not to guarantee some minimum evacuation time but J

rather to make evacuation efficient is a similar idea.

These ideas strongly suggest that major additional in-

. vestments of resources and fundamental community I

changes are not required by.the NRC regulations.

LILCO has abided by these principles, except_that we have gone beyond them in the following sense.

We t

have made a major commitment of resources that is not really necessary.

This is because Suffolk County has refused to allow its resources to be used for radio-logical emergency planning.

Consequently we have had to duplicate in some respects resources that are al-ready there to be used but are being withheld by the County.

For' example, the training of over 1,500 traf-l' fic guides and other emergency workers, when the i-Suffolk County police and other County officials are

~

-ren,

n e,

m - -.

-eav

.w-

,.m,m--,..

,n.wv,, - - -

.c

,.-,-----,,,w w

.c n---

--w m

-p-r-

-*-wa~*-m o,~\\-p~

not only available to help but (our experts tell us) i probably would help in reality, is by any reasonable standard a misallocation of resources.

It was justi-fled only to prevent an even larger squandering of resources, the abandonment of a completed nuclear power plant.

1.

6.

Q. What kinds of background information are reflected in Appendix A to the Emergency Plan?

f A.

[Weismantle, Lieberman)

Appendix A, subtitled "Evacu-ation Plan,"'contains a great deal of detailed infor-mation about the specifics of evacuation planning, including automobile and bus routes, placement of traffic guides, traffic control strategies, and evacu-ation times.

It also contains in Sections I and II much background information about emergency planning in the 10-mile EPZ around Shoreham.

Some of the salient background facts are these:

Half of a 10-mile circle around the plant (roughly the northern semicir-j cle) lies in the open water of Long Island Sound.

The projected 1985 EPZ population of 138,500 (winter).and 160,000 (summer) is concentrated primarily to the west and southwest of Shoreham.

Significant portions of the EPZ, particularly to the south and east,. consist of po:k or large, lightly populated scientific and I-i l

industrial establishments (Brookhaven National Labora' tory, Grumman).

Prevailing winds are toward the northern semicircle (i.e., offshore) over 30% of the time, and toward the lightly populated parts of the EPZ approximately another 35% of the time.

The topog-raphy of'the island is generally flat, with only small hills and bluffs along the northern shore.

The high-way system is good, containing the Long Island Expressway (three lanes plus shoulder in each direc-tion) and severallother east-west roads..This depic-tion, set forth in Appendix A, of a context not inhos-pitable for emergency planning is. presumed as background for purposes of this and other pieces of testimony addressing specific contentions.

Appendix A also details the location of schools, hospitals, government facilities, major employers and roads, the layout of emergency planning zones, and other important items of background information for i

emergency planning.

Again, this information is' avail-able simply as background for the more specific issues j

raised in specific contentions.

7.

Q. As background for later discussions, could you briefly describe the traffic model used in calculating the evacuation times presented in Appendix A to the LILCO Transition Plan?

. =-

4 i l

A. [Lieberman). The model used in all calculations of Shoreham-related evacuation times being presented in

- this testimony is actually a three-model system known as DYNEV, and is discussed in detail in Appendices B and C of Appendix A to the LILCO Transition Plan.

DYNEV consists.of three major components:

1) an. equilibrium traffic assignment model, 2) a capacity model, and 3) a traffic simulation model.

In order to utilize the DYNEV system, an analyst must first represent the physical roadway system.

This is done by use of a network comprised of links and nodes.

' The network for the Shoreham EPZ is displayed in Exhibit 1 of Appendix A to the LILCO Transition Plan, which is' attached to this testimony as Attachment 4.

In general, a link represents a section of roadway, and is expressed in terms of points, or nodes, that exist at either end.

These nodes, which are referred l

L to in Attachment 4 as " internal nodes," generally rep-i

(

resent some' type of physical discontinuity.

This dis-L continuity can be an intersection; a point where the j.

' geometry changes (for example,. a lane-drop); or a l

change in the topography of a road such as a hill or a L-sharp curve.

Nodes have also been introduced in the l.

,,r...,

.n.,..

,,.,,.,,,.,.,,,., ~ _,,..

-,,.~n..---,-,,

.-.--..,_r--,

...----,,.,,.,-,..n....

,na

. Shoreham EPZ network to represent individual ramps connecting arterial highways with expressways and to satisfy a constraint which limits link lengths to less than 10,000 feet even in the absence of any physical discontinuity.

In the last case, the node serves as a

" dummy" that merely shortens the lineal extent of the links on either side.

In addition to the links and nodes that constitute the, network representation of the roadway system, the DYNEV system also requires the specification of

" source" or " origin" nodes that represent areas where evacuation traffic originates, before it actually enters the evacuation network.

For example, looking at Attachment 4, an origin ncde could represent a major employment area, such as Brookhaven National Laboratory (Origin Node 24), or a residential area consisting of a number of individual homes (e.g., Ori-gin Node 16).

These nodes are not an integral part of the roadway system, but actually serve as conduits that feed traffic from the associated areas onto the modeled roadway system.

Consequently, associated with each origin node is a roadway link that receives the traffic generated at that node.

The final set of nodes depicted in Attachment 4 are destination nodes.

These nodes are the points where evacuation routes cross the EPZ boundary.

At these nodes, traffic leaves the area of interest.and is assumed to move on to its ultimate destination.

1.

Traffic Assignment Model The traffic assignment model is designed to iden-tify the best evacuation routes between each pair of origin and destination nodes.

As an input to the traffic assignment model, the analyst specifies the traffic volume traveling to one or more selected des-tinations which " attract" the people whose trips are generated at each origin node.

The traffic assignment model then determines the best paths of travel from each origin node to each such destination node; the best path is defined as one which minimizes the travel i

time from each origin to each associated destination along the network.

This form of optimization is called the User-optimization procedure.

2.

Capacity Model The capacity model, which is referenced as a sub-model by both the traffic assignment and the traffic simulation models, is designed to estimate the permis-i sible service volumes -- defined in vehicles served

i,

per hour -- for each traffic movement on each link.

This model is needed since vehicles executing turning 9

movements i.e., making a turn -- are in general r

serviced at lowerorates than vehicles executing through movements -- i.e.,~ continuing to travel straight -- upon entering intersections.

The model i

reflects this lower service volume by longer discharge-headways for turning vehicles.

A " discharge headway" i

is defined as the elapsed time between the instant a vehicle enters an intersection --

i.e.,

crosses the stop line, and~the instant that the following vehicle enters the intersection.

This model also accounts for conflicts between left-turning vehicles and oncoming 4

m.

vehicles it, tempting to enter the intersection at the same time'.

The model computes the impact, if any, of t

pedestrian traffic on turning vehicles.

Finally, the capacity model also computes the effects on service volume of any control device that is installed at an intersection.

3.

Traffic Simulation Model The traffic simulation model describes the dynamic process.of traffic operations on the evacuation net-work.

The model divides time into small increments,

- called'" time intervals," of 3 3/4 minutes for these l

' ~

studies.

During each time interval, the traffic on l

each link is moved some distance as a function of the following factors, each of which is accounted for in

.the calculation:

the capacity of the' link on which the traf-fic moves; the impedance imposed by the traffic con-trol device, if any, at the downstream node of-the link; the density of traffic on the link and the relationship between speed and density on the link; the presence or absence of any blockage on the link or on the receiving links due either to a blockage of some kind or to congestion.

Specifically, if traffic is unable to move onto a link because of the presence of a standing queue on that link, then it will be forced to slow or stop; the turn movements of traffic at all inter-sections; the rate at which traffic enters the link either at the upstream node or from the associated source node, :Lf one exists; and any storage limitations on the link.

1.

1 The traffic simulation model also moves traffic from link to link using these same factors.

i 8.

Q. In your description of the traffic assignment model, you stated that a User-optimization procedure was used.

Are there other means of determining trip paths?

'A.

[Lieberman)

There are, basically, two forms of l

- traffic assignment optimization:

user-optimization and system optimization.

In user optimization, the mathematical programming algorithm is formulated to assign traffic to origin-destination paths in a manner that minimizes the travel time for all paths.

Stated-another way, if any motorist departs from his assigned path, he cannot reach his desired destination in less travel time, assuming all other vehicles adhere to their assigned paths.

In system optimization, the algorithm is formulated to assign traffic so as to minimize aggregate network-wide travel time.

This objective can be accomplished if some motorists are consciously willing to accept potentially longer individual travel times so that others can benefit as a result.

Under system optimization, if a motorist departs from his assigned path, he personally may be able to reduce his travel time, but the aggregate travel time over the network would increase.

1 The user optimization approach was chosen for evac-i uation planning purposes because it is responsive to people's desire to minimize their evacuation travel time.

=~

l l i 9.

Q. What information is required as input to the DYNEV system?

A.

[Lieberman]

Generally, the information required to execute the traffic assignment model is the same as i

that required to execute the traffic simulation model.

For both models, the following information is required:

1) the specification of all link attri-butes, including link length, number of lanes, the existence of any turn bays, the channelization of lanes, the topology of the network for each link (i.e.,

the identification of all links which receive traffic exiting from the subject link), the time lost in responding to a change in signal con-trol from no-go to go, the mean dis-charge headway, the mean free-flow speed on that link, whether a right

~

turn on red is permitted, and an indi-cator of pedestrian volume, if any; 2) a description of the control device, if any, at each node of.the network and the extent to which it provides access time to the vehicles on the various approaches to that intersection; and 3) the identification of origin nodes within the area serviced by the roadway network, the link on which traffic ori-ginating-from each origin node accesses the evacuation. network, and the time-varying rates at which trips are gener-ated at each origin node.

In addition to these data, the traffic assignment model also requires as input a matrix of information called a trip table.

The trip table specifies the

t number of vehicles leaving each origin node and trav-eling towards each associated destination node.

The traffic simulation model requires the data enu-merated above, but does not use a trip table.

Instead, it requires specification of the turn move-ments of traffic exiting each link, expressed in terms of a percent turning left, through, right and diago-nal.

These data are input from the traffic assignment model.

Thus, the traffic simulation model moves vehi-

-cles along those paths identified by the traffic as-signment model as the most effective means for traffic to evacuate the network.

In addition, the user may specify a blockage factor, for some specified time, to represent the effect of a disabled vehicle or other factor which is reducing the capacity of that link.

10.

Q. How are human responses represented in the DYNEV sys-tem?

A.

[Lieberman)

In order for the DYNEV system to approxi-mate accurately real world traffic performance, its logic must contain a proper representation of human i

l The traffic simulation motorist response mechanisms.

model is classified as a " macroscopic" model since l

individual vehicles and individual motorist responses are not explicitly modeled.

Instead, the responses of l

l

l motorists in the traffic stream are recognized and represented as follows:

1)

As noted earlier, the capacity model is designed and calibrated to take into account the responses of motorists to the presence of other vehicles in the traffic stream while attempting to exe-cute turning movements at intersec-tions.

For example, right turning vehicles execute their maneuver at lower service rates -- i.e.,

at longer headways -- than do the through moving vehicles, since the relatively short turning radius of their curved trajec-tories restricts their speed.

Like-wise, motorists attempting to turn left must seek gaps in oncoming through traffic in order to complete their turn safely.

Such impedances translate into lower discharge rates which are com-l puted by the capacity model according to the specific elements that exist at each intersection.

The capacity model l

also computes any decrease in discharge rates which may result from " lane im-balance."

Lane imbalance results when longer queues develop in one lane rela-tive to'another due to the presence of turning movements on that link.

2)

The traffic simulation model accepts, as an input parameter, a description of motorists' delayed responses to a change in signal indication from no-go to go.

The associated time lag is called " start-up lost-time" which acts to reduce the effective capacity of a

. link.

3)- The specified value of mean queue dis-charge headway reflects motorict i

responses to the physical conditions of the roadway system and is specified as an input by the analyst for each net-work link.

These inputs were obtained for the Shoreham EPZ by conducting a j

h thorough survey of the actual EPZ road-way network.

Based on this survey, the specified values of discharge headway i

f l

for all studies reflected the most con-servative values observed in the field.

4)

The value of free-flow speed which is specified as input reflects the behav-ior of traffic on each link of the net-work as observed by analysts in their survey of the network.

This value will vary with the type of roadway facility (for example, expressway versus a two-lane arterial), motorist behavior, and weather conditions.

s 4

5)

It is well known through empirical observation that as traffic density increases the speed of traffic can de-crease.

This relationship reflects the l

recognition by drivers that as traffic density increases, they may need to re-duce their speed in order to maintain proper, safe spacing between vehicles.

j This speed-versus-density relationship is developed internally by the model using the specified link input data and is applied during the course of the

.I '

simulation to provide rea.listic de-scriptions of traffic on each link.

6)

Finally, it was recognized that any attempt to predict and model driving I

behavior during an evacuation was sub-ject to uncertainties, such as reac-tions to a stressful situation and to potentially conflicting traffic direc-tion.

Rather than attempt to quantify l

these possibilities individually, or to determine here whether in fact they will occur (a matter at least as much within the province of behavioral sci-entists as traffic engineers), we chose t

-to reflect them collectively by reduc-l ing the estimated " nominal" values of l

capacity for links. approaching inter-sections, which are based on conserva-tive empirical observations, by 15 per-cent, whenever congested conditions occur.

These values for " congested capacity," which are.specified in Table IV of Appendix A, were then used in all modeling studies.

i.. - -.

i 11.

Q. To'what extent has the DYNEV. system been validated?

A.

[Lieberman)

The DYNEV system was based upon prior software development efforts which1were undertaken under the sponsorship of the Federal Highway Adminis-tration (FHWA) and Transport Canada..These efforts resulte'd,in a software system named TRAFLO.

The TRAFLO system consists of several distinct models.

[

JOne of these~is a traffic assignment model which has been incorp. orated in DYNEV with several enhancements.

The traffic. assignment model that is used in DYNEV was developed at the University of Montreal for Trans-port Canada.

That equilibrium traffic. assignment t

model has been extensively validated in several loca-tions, most notably in Winnipeg, Canada.

The capacity model that.is used in DYNEV is imbedded in the TRAFLO model.

Thus, the validation of l

the' entire TRAFLO model implicitly validated the capacity model.

In addition, one exercise has been conducted explicitly to validate the capacity model.

That study was conducted using a-congested roadway i

system with a high level of left turns.

The results showed that predicted capacities were within 5% of i

observed. capacities.

~.

.=

I' 4

Another model in the TRAFLO system, called Level II, is a macroscopic traffic simulation model which is an enhancement of a component of a British signal-timing optimization model named TRANSYT.

The TRANSYT i

model is probably the best known signal optimization model in existence; it has been used extensively worldwide.

Both it and the Level II model have been i

widely and carefully validated by comparing their com-puted results against empirical observations.

The i

traffic simulation model in the DYNEV system contains the same computational algorithm for describing traf-fic' operations as does the Level II model.

However, since the Level II model was limited to describing traffic on surface streets, it was necessary for DYNEV to extend that model to accommodate traffic flew along expressways and rural roads.

Consequently, that por-tion of the DYNEV system was not validated previous to this project.

Recently, the DYNEV traffic simulation model'was executed to describe traffic flow along the Whitehurst Freeway in Washington, D.C.

Comparison of

.these model results with empirical data along that facility indicates close agreement.

Although a rig-orous statistically based validation has not been un-dertaken, there is no reason to believe, on the basis 4

of the observation to date, that statistically vali-dated correlations would be materially different from those observed by the comparison to date.

In summary, the major components of the DYNEV sys-tem have been validated using rigorous statistical testing methodology.

The portion of the traffic simu-lation model that has not been statistically validated has been compared at a high level of detail with empirical results from the Whitehurst Freeway, and that comparison has shown good agreement between com-puted and observed traffic performance under congested conditions.

12.

Q. How were the input data obtained for studying the Shoreham EPZ?

A.

[Lieberman)

As indicated earlier, a significant amount of data is necessary in order to apply the DYNEV system.

KLD's first activities therefore focused on gathering and coding the information needed as input.

A large portion of the necessary informa-tion was initially obtained from the Suffolk County Planning Department and from examining ~1arge scale maps.

Further information was obtained from a detailed field survey of the entire EPZ roadway net-work.

During the survey, notes were taken on the

~ geometric features of each roadway-section, any lane channelizations, turn restrictions, other1 restrictions O

to vehicle movement, posted speeds, and the free-flow speed of other' vehicles.

Notes were also taken on a number of~other factors including adjoining land use,-

pavement condition, and the placement, number and-types of control devices.

1 Another study _was undertaken to calibrate the queue i

f j

discharge behavior of vehicles at major intersections throughout the EPZ.

The study was undertaken during peak hours, when the numbers of vehicles in the queues were sufficiently.large'-- generally more than four or five -- to permit the average queue discharge headways s

to be estimated.

The final task in developing inputs to the DYNEV

~

system was to draw the-network representation of the

~go physical roadway system and'to define all internal,

' N origin and destination nodes.

The conversion of this information into the DYNii:V input stream was an extremely labor-intensive activity t

i r, that was subjected to numerous checks.

When the input 6,

stream was introduced into the model, it activated a s

L 7 wide' range of diagnostic testing routines which detect remaining errors in the input stream.

With these 9,

s.

P

..'r

,_.,_m,,

,.,-,,,ym m

.,,wwr_,....

,,em.

y.p,-

. errors identified and corrected, the model was ready for use.

13.

Q.-How was the DYNEV system then used to calculate evacu-ation times for the Shoreham EPZ?

A.

(Lieberman]

In analyzing the Shoreham EPZ, it was necessary to develop an iterative procedure which employed both the traffic assignment and traffic simu-lation models of the DYNEV system.

The rationale of this procedure is explained in the NRC publication en-titled " Analysis of Techniques for Estimating Evacua-tion Times for Emergency Planning Zones,"

NUREG/CR-1745, November 1980:

Although traffic volume can be assigned in a relatively straightforward (although not trivial) process, travel time on a network link is usually a function of the volume assigned and the link capacity.

The basic relationship is that as traffic flow in-creases, traffic speed decreases.

If vol-umes exceed capacity, speeds can be reduced to near zero in a short period.

This unstable condition, called forced flow, re-sults in reduced flow rates as well as re-duced speed.

After the initial assignment is completed, volumes assigned to some links may result n longer travel time dhan initially assumed.

The increased travel time on certain links may then make alter-nate routes the minimum time path.

An iterative process is used to balance the mathematical model of this system.

This process would be especially useful in nuclear evacuation planning, as it would identify potential bottlenecks and alternative routings.

NUREG/CR-1745 at 11-12 (Emphasis supplied).

The iterative procedure employed in studying the Shoreham EPZ, which is explained in more detail in Appendix D of Appendix A to the LILCO Transition Plan, consisted of several steps.

First, a trip table was developed by examining a ldrge scale map of the network, and by determining through judgment which of the limited number of avail-able destinations were suitable for traffic generated at each origin within the EPZ.

The suitability of a given destination was determined using a series of criteria that included selecting destinations that would minimize the distance that.had to be-traveled,

' consistent with the constraint of locating destina-tions in a manner that avoided the necessity for traf-fic to move closer to the Shoreham plant.

The set of candidate destinations selected on the basis of these criteria was then reviewed to ensure that the collection of destinations for various ori-gins would not produce a " focusing" of traffic on a few limited routes.

This additional review was responsive to another goal of the planning effort --

that of expediting overall evacuation times -- by dispersing traffic so that the relationship between i

I h

,._..,r-y,.,.. _

....m.-7

- _. traffic demand and available roadway capacity would be as uniform as practicable over the network.

Since the steps just described all involve degrees of judgment, sensitivity studies were-conducted to determine the effects of varying the trip' table design from what was perceived originally as "best," to other possible con-figurations to arrive at the best overall design.

Second, for each specified trip table, the traffic assignment model was executed to produce a coarse estimate of travel time for each movement ou each net-work link.

By examining this' output, it was possible to identify bottlenecks where considerable delays could result.

Third, having identified these bottlenecks, engi-neering judgment was then applied either to modify the trip table or to develop control tactics to expedite the movement of vehicles through the network.

Alter-natively, in some instances, the configuration of the evacuation network was reviewed with an eye toward including more of the available roadways in the evacu-ation network.

Again, these judgmental decisions were tested by repeatedly executing the traffic assignment model and then comparing the new results to the previ-ous ones to_ determine whether the changes had improved traffic performance.

)

\\

Fourth, after performing several such iterations

'with the traffic assignment model, it became clear that little additional progress could be made by restricting the study to the traffic assignment model.

1he iterative process then turned to the traffic simu-lation model, which can produce more accurate repre-sentations of actual traffic operations since it more rigorously accounts for changes in traffic conditions with-time.

The traffic simulation model was executed using the turning percentages generated by the traffic i

assignment run that had produced most expeditious

~

evacuation of the EPZ.to that point.

These results were then analyzed to identify the problems associated with the formation of bottlenecks and to quantify the extent of the resultant congestion.

Once again, efforts were made to alter the design of the traffic control strategies to reduce the impact of such bottlenecks and to expedite the movement of traffic.

As before, all such changes in control tactics were t

tested'by introducing those chan.as into the traffic l

simulation model, executing that model, and examining the results.

This iterative procedure was a time-consuming one which required the repeated execution of both the r

r i

traffic assignment and traffic simulation models to develop an understanding of the complex interactions among demand, capacity, the spatial configuration of the network and the spatial and temporal distribution of traffic.on that network.

During the course of this initial activity, we assumed a total evacuation of the entire EPZ.

We felt that any route st ructure which was optimal or near-optimal for these conditions would also'be suitable for partial evacuation of the EPZ.

During -thic-iterative process to develop the " base case" evacuation plan scenario for a 10 mile radius EPZ, the-two models were executed approximately 20 to 25 times.

14.

Q. Could you briefly descrice the information provided by the DYNEV system?

A.

[Lieberman]

The DYNEV system provides a variety of

~

l information describing traffic operations on each l

. link, and for the network as a whole, at specified l

periods of time during the course of evacuation.

Link-specific information includes:

vehicle-miles, j

number of vehicles discharged from the link since the beginning of the. evacuation, and two measures of the travel time of vehicles, including total time in l~

vehicle-minutes and average travel gime in seconds per l

- 4

-vehicle.

Both measures of-travel time are further subdivided into unimpeded travel'and delay time.

In addition, the model provides the average speed of Jtraxal.on the link aggregated since the beginning of

the evacuation, a measure of the extent of queue for-mation'on that link, the percentage of vehicles on that link that are forced to stop, and the current number of vehicles occupying that link.

The measures aof queue formation and the number of vehicles on each link are particularly helpful, since the DYNEV system recognizes that each link has limited vehicle storage capability.

This characteristic of DYNEV is more use-ful than the technique used by various other models, t

which permit cars to " pile on top of one another."

Since this information.is printed at specified periods over time through the evacuat on, it is possi-i I

ble to' describe a history of traffic operations on each link of the network.

The. level of detail _thus afforded by the model permits the analyst to identify the time-dependent traffic flow processes over the en-tire network, the location of each bottleneck, the conditions which led to the formation of the bottle-l-

neck, the time at which the-bottleneck formed, and the consequences of each bottleneck in terms of delays l

l:

l

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ experienced by the motorist and the length of queue formed from the head of the bottleneck.

These data provide the analyst with insights into important cause and effect relationships which permit him to identify management techniques which will expedite the flow of traffic.

15.

Q. How many different scenarios were modeled during the preparation of the LILCO Transition Plan?

A.

[Lieberman)

Initially, 21 scenarios were investi-gated.

These cases are listed in Table II on page II-8 of the Appendix A of the LILCO Transition Plan, which is attached as Attachment 5 hereto.

Subsequent to these original 21 scenarios, we have examined approximately-15 additional scenarios.

16.

Q. Why were this many scenarios studied?

A.

[Lieberman]

The first 20 cases were necessary in order to be responsive to the FEMA /NRC Guidelines out-lined in NUREG-0654, Rev.1, Appendix 4, page 4-4.

These cases are reprinted as Cases 1-20 in Table 1:

" Evacuation Time Estimates," attached as Attachment 6 hereto.

The first three cases addressed the three regions making up the 90* quadrant east of the EPZ.

These regions within the east quadrant are, respec-tively, the region within approximately two miles of the plant, the regjon within approximately five miles, and the full ten mile region.

Similarly,_ Cases 4 to 6 provide estimates of evacu-ation times for the various. regions of the 90 quad-rant west of the plant, and Cases 7 to 9 provide evac-uation time estimates for the various regions of the central 90* quadrant, which, as it happens, overlaps significantly both the east and west quadrants.

Case 10 provides time estimates under the condition that the entire two mile region is evacuated; Case 11 is for the entire five mile region; and Case 12 is for the case where the entire 10 mile radius EPZ is evacu-ated simultaneously.

Cases 13 through 20 repeated selected analyses of the first. set of runs, under the conditions of inclement weather.

Two types of inclem-ent weather were considered, namely, summer and win-ter.

For summer inclement weather, which consists of rain, we have applied a reduction in capacity, and a reduction in free-flow speed for each link in the net-work, of 20% relative to those estimated values which l

l were specified for normal weather conditions.

Thus, Cases 14, 16, 18 and 20 provide evacuation time esti-i mates for summertime populations under inclement sum-L mer weather conditions for the entire east quadrant, I

l 1

_ _ _ _ _ _ _ - _ _ _ _ - _ _ _ _ west quadrant, central quadrant and entire EPZ, respectively.

Cases 13, 15, 17 and 19 perform the same kind of analysis under the assumptions of win-tertime population and~ inclement winter weather, which could consist of compressed snow with patches of ice.

For these runs, we have reduced capacity and free flow speeds by 30% from normal weather conditions.

Case 21 was undertaken to determine the sensitivity of evacuation travel times to change in assumptions about the trip generation period.

This case was run assuming a three hour trip generation period rather than the two hour period assumed in Cases 1 to 20.

This case is discussed ~in greater detail in the answers to Questions 25 and 26 below.

In additien to these cases which are presented in Appendix A of the LILCO Transition Plan, a series of additional cases were studied in response to concerns that were raised by Suffolk County in its Phase I emergency planning testimony and its later contentions on Phase II issues.

Specifically, these cases, which are detailed later in this testimony or the testimony on Contention 23, examined the effect on evacuation times of a variety of differing assumptions including:

e.

1) the-voluntary evacuation from areas outside the EPZ -- the so-called " shad-

.ow phenomenon" (Cases 22, 23, 26, 27 4

and 28);

i 2) the presence of accidents during an evacuation (Cases 29 and 30);

3) the absence of traffic guides and evacuation route signing -- an

" uncontrolled" evacuation (Cases 24 and 25);-

4) the deviation of evacuees from their recommended paths -

"non-compliance" cases (Cases 31,-32, 33 and 34); and 5) the creation of an additional evacua-tion route along the LILCO right-of-way extending parallel to Route 25A (Cases 35 and 36).

17.

Q.1Could you briefly summarize the results from these case studies?-

A.

[Lieberman]

Yes. to this testimony sum-

~

marizes the results from these studies.

The results are presented in terms of the elapsed time from the first notice to evacuate and the time when 50%, 90%

and 100% of the people ordered to evacuate cross the 10-mile ~EPZ boundary..More detailed discussions of these results are presented later in this testimony and this panel's testimony on Contention 23.

18.

Q. Have odner evacuation time estimates been performed

'for the Shoreham facility?

A.

[Lieberman]- Yes.

To the best of my knowledge, three

~ _.. -.. _. - ~ _. _ _.. _ _ _.. _, _., _.. _ _.. _,. -,. _ -. - _ -., _

organizations have conducted studies of evacuation times for the area surrounding the Shoreham facility:

Wilbur Smith and Associates (WSA) as reported in FEMA-REP-3, dated February 1981; Suffolk County Planning Department (SCPD) in late 1981 or early 1982 (this time esti-mate was part of the work performed by SCPD during the period when LILCO and Suffolk County were cooperatively preparing an emergency plan for Shoreham.

This time es-timate was never presented in a formal doc-ument.); and PRC Voorhees (PRCV) as reported in a memo-randum dated June 1982 and in a report.en-titled, " Preliminary Evacuation Time Esti-mates for the Shoreham EPZ," November 1982.

The WSA and SCPD evacuation time estimates were calcu-lated for a lO-mile planning area that was identified or virtually identical to the EPZ shown in Figure 3 of Appendix A.

The PRCV memorandum of June 1982 also contains an evacuation time estimate for a ten-mile EPZ.

The later PRCV estimate, which was prepared for the Suffolk County legislature's hearings of early 1983, studied a 20-mile EPZ.

The results of the " base case" analyses which studied a 10-mile planning zone are tabulated below, in terms of the total time required.to evacuate the entire summertime population of the 10-mile EPZ to the west, assuming normal weath-er and no " shadow phenomenon" beyond the 10-mile boundary, following the first evacuation advisory:

l

=

i Agency Eut. Time (Hrs. - Min.)

WSA 3 - 05 SCPD 5 - 35 PRCV [10-mile EPZ) 5 - 45 19.

Q. How do-these results of these compare with those you summarized in response to-Question 15?

A.

[Lieberman]

The. comparable estimates obtained by KLD i

for the summertime population are 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes for a planned evacuation, and 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> 10 minutes for uncontrolled evacuation.

These estimates of evacua-tion times agree closely with the SCPD and PRCV esti-mates and, in fact, bracket those estimates.

20.

Q. How do the evacuation time estimates for the Shoreham EPZ compare with the evacuation time estimates for other nuclear facilities?

A.

[Lieberman]

A convenient source of information about evacuation time estimates for other nuclear power j

plants is an NRC publication entitled "An Analysis of l

L Evacuation Time Estimates Around 52 Nuclear Power Plant Sites," NUREG/CR-1856, May 1981.

Table 8 of that document presents information of the median evac-uation times, by 90 sector, as a function er the sec-to-population being evacuated.

These data were com-pared to the sector evacuation time estimates-for the Shoreham facility -- sector evacuation times were selected for comparison, because they avoid the l -

,,. ~, - -,

--.,----.n

-..-.-.,...,,,,,n.

w

,-,.-.--n-,-.n

,..n,.-,..

i

- problems of comparing EPZs with different geographical configurations, for example, plants that require a 360* area to be evacuated, and plants, like Shoreham, that require far less than a 360* evacuation.

The results of this comparison are displayed in Attachment 7 hereto -- results are presented for the Shoreham Eastern, Central and Western sectors (SNPS E, SNPS C and SNPS W, respectively).

A review of this Attach-ment shows that the sector evacuation times for Shoreham are comparable to the median evacuation times presented in NUREG/CR-1856.

Contention 65.A.

21.

Q. How is " mobilization" defined in the LILCO Transition Plan?

A.

[Lieberman]

Mobilization time within the context-of the LILCO Transition-Plan is defined as the elapsed time between the issuance of the notice to evacuate and the time that the first person within the EPZ begins an evacuation journey from a home or place of business to a location outside the EPZ.

Thus no evac-uation activity is assumed to take place during the mobilization period.

The twenty-minute mobilization period assumed in LILCO's model runs reflects the fact i

, that some individuals will' receive, and understand, e

their notification virtually instantaneously upon issuance of the evacuation notice and will be at or near home at the time, but that they will need some time to prepare to evacuate.

The LILCO modeling runs clearly recognize that a majority of people within the EPZ will require more than 20 minutes to begin their evacuation journey.

For example, as indicated in Table X of Appendix A which is attached as Attachment 8, the number of peo-ple who have begun their evacuation trips within the first 35 minutes following the advisory to evacuate constitute only about 3% of the total population with-in the EPZ.

22.

Q. Does this definition of " mobilization period" vary from that used in SC Contention 65.A.?

If so, how?

A.

(Lieberman]

Yes.

Contention 65.A. apparently assumes, erroneously, that the 20-minute mobilization period used by LILCO implies that the entire public begins to evacuate simultaneously at the end of the 20 minute period.

As just described, this simply is not the case.

The evacuation time estimates assume, in all but one case, that trip generation and loading of the network will take place over a period of two hours

following the end of the mobilization period.

Thus, everyone is assumed.to have begun their evacuation journey by 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> and 20 minutes after the issuance of the advisory to evacuate -- not 20 minutes, as the contention appears to assume.

Contrary to the suggestion of Contention 65.A.,

this loading period implicitly accounts for a variety of necessary or optional activities, recited in Con-tention 65.A.,

that are preparatory to evacuation.

These activities include receiving notification of the advisory to evacuate, preparing to leave work, travel-ing home to unite the family unit, and any necessary preparations that must be made before departing home.

-23.

Q. How are people loaded onto the evacuation traffic net-work during the two-hour loading period that follows the mobilization period?

l A.

[Lieberman]

The source of the data describing this two hour trip generation period was the Suffolk County Planning Department's work in 1981 and early 1982.

This work produced a series of loading histo-l

. grams, such as the one depfcted on page III-9 of l

Appendix A, for each origin node.

These histograms indicated the number of cars that would leave each origin point during each half hour of the two hour period.

Unfortunately, the analytical basis for these

+

_ _ histograms could not be directly established by KLD and LILCO personnel, since communication with person-nel from the Suffolk County Planning Department was terminated by.Suffolk County at approximately the time these histograms were first produced.

Since the most direct means of-validating these histograms was not available, KLD personnel compared this two-hour load-ing time with loading times used in other evacuation plans and determined that a two hour trip generation period was in fact typical, or " reasonable".

As a re-sult, the loading histograms produced by the Suffolk County Planning Department were employed for the studies that KLD conducted in the spring and summer of 1982, and that are documented in Appendix'A as Cases 1 through 20.

To account for.any uncertainty associated I

with these loading histograms, an additional run --

i Case 21 -- was performed using a three-hour trip gen -

eration period to test the sensitivity of evacuation l

time estimates to a change in the trip generation l

l period.

l' In the fall of 1982, KLD commissioned a survey by i

j.

the National Center for Telephone Research to obtain l

L detailed information describing the travel patterns, household structure and automobile ownership ef the l

..... _. ~ _. _. _,...

. population. residing within the EPZ.

One purpose of this survey was to gather empirical data which could be'used to generate more accurate trip-generation time distribution.

That organization surveyed about 1,000 members of the public living within the Shoreham EPZ

-to obtain a variety of information, including informa-tion about the amount of time it takes for people to return home from work during the normal commuting period.

Some pertinent results from this study are attached as Attachment 9 hereto.

A review of the information collected during this survey added cre-dence to the earlier conclusion that the two-hour loading period was reasonable.

24.

Q. Have you performed any detailed analyses of the survey results, or have you attempted to account explicitly for the four items listed in Contention 65.A.?

A.

[Lieberman]

Based upon the results of the National l

l-Center for Telephone Research survey and additional information on daily transportation schedules for school children, we were able to produce a detailed, independent analysis of the time distributions for the pre-evacuation preparatory events identified in Figure l

4 of Appendix 4 of NUREG-0654.

These preparatory I

events include the receipt of emergency information, i

workers traveling home to reunite with their families,

and families preparing their households prior to departure.

These events correspond with items 1, 2 and 3 of Contention 65.A.

Item 4 of Contention 65.A.

does not represent a separate preparatory event; rather, it raises a concern about the time required to complete some of these events.

As such, it is addressed in the time distributions that have been prepared for each preparatory event.

This analysis, and the results obtained are documented in a report entitled " Development of Time Distributions for Evacu-ation Events and Activities," KLD TM-139, which is appended to this testimony as Attachment 10.

The results of this detailed analysis were compared to.the loading histograms prepared by the Suffolk County Planning Department.

This comparison, which is displayed in Figure 1 of Attachment 10, shows close agreement, with differences of less than 15 minutes in the time distributions.

The results in KLD TM-139 were also compared with time distributions that were prepared for Suffolk County by PRC Voorhees, which are presented on page 30 cf a report entitled " Preliminary Evacuation Time Estimates for the Shoreham EPZ," PRC Voorhees, November 1982.

This comparison is displayed in Figure

. 3 of Attachment 10.

While this figure shows differ-ences in the time distributions for individual prepa-ratory events, the distributions for the auto-owning population leaving home -- which is the critical dis-tribution for modeling purposes -- are virtually iden-tical.

Thus, the loading histograms, that were pre-pared by Suffolk County and were used in all but one of KLD's modeling runs, have been verified by indepen-dent analyses conducted by KLD and PRC Voorhees.

25.

Q. Please describe the sensitivity study that was con-ducted using a three hour trip generation period.

A.

[Lieberman]

To resolve any remaining uncertainties about the loading histograms prepared by the Suffolk County Planning Department, it was decided to conduct a sensitivity study of the estimates of evacuation

-travel time with respect to trip generation distribu-l tions.

Consequently, Case 21 was conducted essential-l ly to repeat Case 12 -- the base case evacuation of the entire 10-mile EPZ -- using a 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> trip genera-

~

. tion period instead of the 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> period used in Case 12.

These trip generation distributions are depicted in' Figure 4 of Attachment 10.

The results of this study indicate that increasing the trip generation l

l time by one hour has no effect on the total time l

l required to evacuate the EPZ (compare Cases 12, 21 on ).

Careful examination of the computer output reveals some differences in the internal dis-tribution of evacuation travel time, but that these differences are small.

26.

Q. How do you explain this lack of sensitivity of evacua-tion travel time to a one-hour lengthening of the trip generation time?

A.

[Lieberman]

In order to understand this result, one must first appreciate that the effect of trip genera-tion time on evacuation travel time depends on whether a network is operating at an " undersaturated" or " sat-urated" condition.

A network is undersaturated when virtually all-the links --

i.e.,

sections of roadway -- in the system are servicing traffic demand levels that~are below the available roadway capacity.

During saturated conditions, several and perhaps many, links in a network may experience traffic demands that exceed roadway capacity.

For an undersaturated condi-tion, the total time required to service the total de-mand is equivalent to the time ov6r which that demand is generated -- the trip generation time -- plus the travel time through the EPZ.

For the saturated case, however, extending the trip generation time has no little or influence on the total time required to t

f'

_ _ _. - _ - service the total demand, as long as the trip genera-tion time is less than the time required to service the total demand.

To illustrate these relationships, let us consider a single link a mile in length having a capacity of 2,000 vehicles per hour.

For the case of under-saturated flow, let us assume that the total demand for service on this link is 1,000 vehicles.

If these 1,000 vehicles demand service over a' period of 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, then the total time to move this traffic through the link is approximately 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> (if we ignore the travel time along the link).

If the demand for service is assumed to extend over 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, then the total time required to service these 1,000 vehicles would be 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

For the saturated condition, let us assume a total demand of 8,000 vehicles.

If these 8,000 vehicles leave their respective homes and demand service over a period of 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, then at the end of 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, 2,000 vehicles will have been serviced and 6,000 would be in a queue awaiting service.

At the end of two hours, a total of 4,000 vehicles would have been serviced and another 4,000 would be still in queue awaiting ser-vice.

Only at the end of 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> will all vehicles

~.

.. -. - -. have been serviced given the. constraining link capac--

ity -- 2,000 vehicles per hour.

If we assume these same 8,000 vehicles demanded service more or less uniformly-over~a two hour period then at the end of the first hour, 4,000 vehicles would not have left home, 2,000 would have been ser-viced, and another'2,000 vehicles would be in-a queue awaiting service.

At the end of 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, all vehicles

.will have left home, a total of 4,000 vehicles would have been serviced and another 4,000 vehicles would be l

in queue.

At the end of four hours, all 8,000 vehi-cles will have been serviced.

3 It must be recognized that this example is a micro-cosm of an' evacuation network.

Nevertheless, the principles which apply to a' single link apply as well for a network of roads.

RSo long as the trip genera-tion period is less than the time it takes to service l

the demand in a saturated network, then any change in the trip generation time will not, as a general rule, influence the time required to service the demand.

Only when the trip generation-time approaches the overall evacuation travel time will there be noticea-L ble lengthening of the total travel time.

Of course, if the trip generation time is longer than the

4,

- evacuation travel time associated with saturated flow conditions, then-the system reverts to an undersaturated condition and any further lengthening in trip generation time translates into a equivalent lengthening of. evacuation-travel time.

Traffic demand during a 10-mile EPZ evacuation cre-ates a saturated flow condition on parts of the Shoreham evacuation network during much of the evacua-tion period.

Thus, lengthening the trip generation time from 2 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> should not, and does not, affect the overall evacuation travel time, which is approximately 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes for both cases (com-pare Cases 12, 21 on Attachment 6).

Though not modeled, I would expect a similar insensitivity to a one-hour extension in the trip gen-eration period in all butLfour of the cases presented in Appendix A.

In those cases (Cases 1, 4,

7 and-10),

i which simulate an evacuation of all or part of the

. two-mile region, the total evacuation times range from 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 15 minutes to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 35 minutes.

For the 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 20 minute loading period that was assumed in the sensitivity study, I would expect that the total evac-untion time for these cases would increase slightly to between 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> 40 minutes and 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

w y.

,r---

-e e-

---

r--,

=

w

+v w---m--

-w---

1-

--s>

-r e

.---e--

w-se-----e-~ * - - - - - -. -

i.

( l Contention 65.B.

4 27.

Q.'Has.KLD considered how traffic created by people tak-ing preparatory steps for evacuation will affect evac-untion travel and evacuation time estimates?

A.

(Lieberman]

.Yes.

Preparatory trips are likely to fall into two-general categories:

1.

Trips taken to unite the family unit at home; and 2.

other trips taken by members of the family unit prior to initiating the

}.

evacuation trip.

The work-to-horae trip constitutes the largest portion of.these preparatory trips.

A potential exists for these trips to restrict the service volume for evacua-tion trips whenever traffic demand approaches capacity and the associated turning movements at intersections act to reduce the effective capacity of some of the approaches.

2 According to the results from Case 12 --

i.e.,-

evacuation of the entire EPZ, some congested condi-tions begin to occur approximately 40 minutes after the first evacuees depart from their homes.

Reference to Figure 3 of Attachment 10 indicates that based on the'KLD analysis of normal rush hour travel times, 87%

of the work-to-home trips are completed within 40 minutes after the start of evacuation; based on the

A-.

PRC Voorhees analysis, 93% of the work-to-home trips are completed at this time.

-Thus, both the KLD and PRC Voorhees analyses indicate that work-to-home

. travel is nearly complete by the time the evacuation travel demand begins to reach roadway capacity.

It can therefore be concluded that any interaction be-tween the work-to-home travel and the evacuation trav-el-should be limited in extent.

Other trips undertaken by members of the family unit that are " preparatory" in nature will probably take place outside the EPZ.

This conclusion is prem-ised'on:

1.

.The likelihood that most retail estab-lishments within the EPZ would close shortly.after the sirens are activated.

and well before evacuation travel be-comes' heavy; 2.

A recognition by the public that'equiv-alent establishments are plentiful in Western Suffolk and Nassau Counties, and a coupling of this perception with a desire to avoid unnecessary delays within the EPZ; and 3.

Radio broadcasts which will urge evacuees to proceed promptly out of the EPZ.

Thus, this category of preparatory trips is unlikely to affect evacuation time estimates.

Contention 65.C.

28.

Q. Why was a traffic control plan included in the LILCO Transition Plan?

A.

[Lieberman)

A traffic control plan was included in Appendix A to the LILCO Transition Plan primarily to minimize overall evacuation times, and hence, to reduce the potential radiation exposure of persons living within the EPZ.

The traffic control plan, l

which is detailed in Section IV of Appendix A, is de-signed to accomplish the goal of minimizing evacuation times in two ways.

First, the plan utilizes cpecial traffic control tactics to limit many potential bottlenecks.

The special tactics are beneficial dur-ing evacuation because they replace less efficient control tactics designed to serve normal traffic con-

)

ditions.

The most important evacuation control tac-tics are displayed on pages IV-10 to -13 of Appendix A.

Second, the plan specifies the establishment of up to 138 traffic control posts, depending on the zones to be evacuated (see Figures 8 and 8.3 in Appendix A).

Each post will be manned by the number of traffic guides specified in Figure 8 of Appendix A.

The pri-mary purpose of the traffic guides will be to

,.. facilitate traffic. flow and to assure compliance with routing assignments, to the extent possible.

Route compliance will further be aided by the placement of

' trailblazer signs throughout the EPZ.

These signs will. identify evacuation routes by the use of direc-tional arrows and-zone letters.

29.

Q. Has LILCO. evaluated the effect on Appendix A evacua-tion time estimates if these traffic control measures were not used?

A.-[Lieberman]

Yes.

As indicated in the answer to Ques-tion ~16, KLD has conducted studies to determine the effects of " uncontrolled". evacuations during both nor-mal and inclement winter weather.

For the purposes of these studies, an uncontrolled evacuation was defined as an. evacuation.during which existing traffic signals were operated as normal and people followed the routes assigned by'the traffic' assignment model.

These assigned routes were not.necessarily the same as the recommended routes that appear in Appendix A, since the' traffic assignment model was rerun to produce com-parable levels of optimization in the " uncontrolled" 1 cases.

These studies are explained in more detail in LC a KLD' report numbered KLD TM-77, which is attached as

' Attachment 11 hereto.

1

\\

s i

\\

)

Q.

Lt -

'k.

30.

Q.'What!were t results of these studies?

z.

A.

[Lieberman).The results of these studies, which appear as Cases :24 and 25 in Attachment 6, indicate that for^an uncontrolled evacuation under normal con-

. )3 ditions the total evacuation time would increase-by 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 35 minutes, from 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes to 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> 30

. minutes.

For an uncontrolled evacuation under inclem-e kJpj ent winter weather-conditions the total evacuation-f

'tiae would increase by 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 55 minutes, from 6 hour6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />'s I ;yb

.to 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> 55 minutes.

y 31.~

Q. Turning to the specific concerns of Contention 65.0.,

will LERO traffic guides " screen" people entering the EPZ or-people proceeding in other than prescribed

{y directions?

C'.A.

[Weismantle, Lieberman]

No.

As documented on page IV-8 of Appendix A, "no motorist will be denied access to any location within the EPZ."

The LERO traffic guides, through the use of hand and arm movements and the deployment of traffic cones, will discourage traf-fic from entering the EPZ, but will not " screen" or

,l};

prevent any-vehicle from entering the EPZ or from pro-w

'if.

2 -'

ceeding in other than prescribed ~ movements.

Further, te

'the function of discouraging traffic from entering the s.

t EPZ is secondary to that of facilitating its depar-ture.

~;

?

r.

'hs O -

4 uu =

_ _ _ _ _ These control measures, and the rationalia for their selection, are discussed in more detail in LILCO's testimony on SC Contention 23.H.

32.

Q. In subparts 2 and 3 of Contention 65.C., Suffolk County suggests that aggressive driver behavior and conflicting traffic signals will increase congestion and delay traffic movement; were these effects consid-ered in KLD's modeling work?

A.

[Lieberman]

As was noted in an earlier answer, the actual presence of aggressive driving behavior during an evacuation, mentioned in subpart 2.,

is a matter outside the expertise of this panel, and is being dis-cussed in the testimony' filed today by Drs. Dynes, Mileti, et al.

As a traffic engineer, I am aware of many instances where temporary traffic control mea-sures may conflict with existing traffic control mea-sures, for example, construction zones, malfunctioning traffic signals where police are directing traffic and

(

utility workers routing traffic around a work area.

i To account for potential lengthening of headways that l

l may result from driver uncertainty about conflicting signals, all roadway capacities were conservatively reduced by 15% during congested conditions in all evacuation time estimates.

l l

I

~.

_. _. _. _ t i

L 33.

Q. How were the various evacuation routes, which are depicted on pages IV-75 to -165 of Appendix A, deter-

. mined?

A.

(Lieberman]

As is described in more detail in my answer to Question 13, an iterative process was

. applied to identify the best routes for facilitating i

the movement of vehicles and people from the EPZ in the event of an accident.

Briefly, the process began with the development of a trip table that associated each origin node with one of more destinations at the periphery of the EPZ.

This information was then input to the traffic assignment model of the DYNEV system.

This model identified the " minimum-time" trk<.s-routes for traffic moving from each origin node to eacn cor-responding destination node.

With this information, the more' detailed traffic simulation model within DYNEV was executed to produce detailed estimates of evacuation travel time.

i l

From this information, an analyst can determine ways to improve the plan and thus expedite the move-ment of traffic over the network.

Such improvements can take the form of a revised trip table, revised or l

added control tactics, or an awareness that additional existing roadway sections should be included in the evacuation network.

These changes are introduced into I-t

the DYNEV system, and the iterative process is con-tinued, as has been done for Shoreham, until a near i

optimal solution is obtained.

The routes associated with this near optimal solution were then identified 4

and documented on the cited pages of Appendix A.

34.

Q. What assumptions were made in the KLD model runs about evacuees' compliance with the routes depicted on pages IV-75 to -165?

A.

[Lieberman]

Consistent with the usage of the traffic assignment model and the subsequent use of the traffic simulation model, it was assumed that vehicles would follow these routes during a " planned" evacuation.

Similarly in " uncontrolled" evacuations, it was assumed that people would follow the routes assigned by the traffic assignment model.

35.

Q. Is it likely that people's perceptions of the most expeditious way out of the EPZ will vary from these designated routes?

A.

[Lieberman]

The answer to this question is based in l

l part on common sense supported by sociological litera-l tura as reflected in the testimony of Drs. Dynes, Mileti, et al.,

and in part on application of traffic I

engineering tools.

It is likely;that during an emer-gency situation people will be strongly motivated to leave the EPZ by the most expeditious route.

This i

4 i motivation, reinforced and guided by a public informa-tion program that informs the public as to the best path for meeting that motivational goal, and by imple-mentation by traffic guides and trailblazer signs con-sistently with that information, should result in a high level of compliance.

The trip tables developed by KLD were designed to be responsive to these desires.

Destination nodes were chosen to evacuate people to the west by the shortest path, subject to two constraints.

First, in some cases alternative paths nearly as direct as the shortest path were chosen in the interest, and only in the interest, of dispersing traffic and thus reducing travel time.

Second, the NUREG-0654 guideline that evacuation routes not require evacuees to move closer to the power planc was observed.

l l

l 36.

Q. Has KLD tested how non-compliance with these recom-i-

mended routes, if it occurred, wc;uld affect evacuation f

time estimates?

l A.

[Lieberman]

Yes.

KLD has analyzed the effect on evacuation time estimates of non-compliance with rec-t j

ommended routes for both a " planned" and an

" uncontrolled" evacuation.

Two studies were conducted for each case.

These studies are explained in detail l

in a report entitled " Determination of Varying Route Compliance Levels and of a Proposed New Roadway on Evacuation Travel Times within the Shoreham EPZ," KLD TM-140, which is attached as Attachment 12 hereto.

The first study assumed that 25% of the population

-having access to cars would not comply with the recom-mended routes, while the second study assumed 50% non-compliance.

For each study, it was assumed that the vehicles not complying with the recommended routes would, in most cases, be equally distributed between two alternative destinations.

The first step in these studies was to revise the trip tables used in " base case" runs to reflect the fact that some vehicles would be traveling on paths other than those recommended.

This was accomplished by identifying reasonable, alternative destinations for these vehicles.

The reasonableness of these al-ternative destinations was judged by the same criteria that were used to select the original trip table.

Specifically, the destinations were selected to com-port with the public's likely perceptions of the closest and most expeditious means of evacuating the EPZ, but the alternative destinations were not located so as to take people closer to the plant.

Depending upon the location of each origin node, either one or

, two alternate destinations were selected.

For origin nodes relatively distant from the western EPZ bound-ary, two alternative destinations were generally specified.

For origin nodes located relatively near the western EPZ boundary, the previously discussed criteria greatly constrain the analyst's flexibility in choosing alternative destinations.

Many of the or-igins near the western EPZ boundary have only one via-ble alternative destination, while a few have none other than the originally chosen route.

For'the non-compliance cases, a model variation'was introduced that permitted traffic to deviate from the assigned paths if traffic congestion blocked that path and a viable, alternate path was available.

Following revision of the trip table, the traffic assignment program was again applied in order to de-tennine the traffic pattern within the EPZ.

Then, the traffic simulation model was applied to compute the evacuation times for the total evacuation of the EPZ.

37.

Q. What were the results of these studies?

A.

[Lieberman]

For the case of a controlled evacuation i.e.,

traffic guides, trailblazer signs, traffic control tactics in place, it was found that if 25% of the population diverted from the recommended paths,

the-estimated evacuation time would not be affected (compare Cases 12, 31 on Attachment 6).

If the level of non-compliance increased to 50%, then there would be an increase of 35 minutes in the total time re-quired to evacuate the entire EPZ,'from 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> 30 minutes (compare cases 12, 32 on ).

For the case of an uncontrolled evacu-ation, it was found that no increase in the total

. evacuation time occurred for either 25% or 50% non-compliance (compare Cases 24, 33 and 34 on Attachment 6).

Careful examination of the oute?t did, in fact, show that detailed traffic patterns over the network were different in each "non-compliance" case from those occurring in the originally assumed case of 100%

route compliance.

The reason why these detailed dif-ferences did not significantly affect overall evacua-tion times is because trade-offs occurred among alter-native routes.

For example, if an alternative route for population originating at one source node shifted a portion of that population from destination A to destination B, then that shift tended to be balanced by another population group originating at a different source node shifting from its recommended destination

. -B to destination A.

Such balancing shifts in popula-tion routing did not, however, result in symmetrical trade-offs.

There were some net' differences in the distribution of traffic volumes at points where traf-fic exited the EPZ.

Nevertheless, such net differ-ences in traffic volumes were not sufficient to cause material differences in the overall estimates of evac-untion time.

Probably the primary reason for the insensitivity of results is due to the fact that the limiting area for overall time of evacuation of the EPZ is that area that lies directly west of the plant and north of Route 25A.

This area is the most densely populated f

region within the EPZ, and all other source nodes and paths generate evacuation travel times which are.less than those experienced by this segment of the EPZ pop-ulation'.

Nor does deviation from recommended paths by these groups ever increase travel time to an extent that any of these other paths takes longer than evacu-ation of the area north of Route 25A and west of the plant.

The set of criteria that were used to identify the destination points for the population in this area l

north of Route 25A and west of the plant was based I

' primarily on considerations of proximity between the origin nodes in this area and their destination pointo.

For the vast majority of the population liv-ing in this area, the originally assigned destination points represented the most direct route out of the EPZ.

For a relatively small portion of the population in this area, who live closer to the plant, they were originally recommended to take routes that moved them first in a southerly direction so that they would not encounter the congestion to the immediate west (see, for example, Zones A and F at pages IV-77 and -101 of Appendix A).

Consequently, when non-compliance was considered, a portion of this population was directed to the west.

This additional westward traffic, and the consequent increment in congestion, were responsi-ble for the 30 minute increase in the evacuation time for the case of a planned evacuation and 50% non-compliance.

The major portion of the population in this~ area, however, had few viable routes other than those already recommended in Appendix A.

Accordingly, the impact of non-compliance on the population resid-ing in this area was limited.

Overall evacuation travel time results for the uncontrolled case are relatively insensitive to

-. _. I non-compliance with predetermined recommended routes, for readily explicable reasons.

First, for the planned case, the control tactics were developed expressly to service expected traffic volumes at a high level of efficiency, under the assumption of 100%

route compliance.

Hence, any substantive departure of the traffic routing pattern from that which was used to design the control tactics could adversely impact evacuation travel times.

For the uncontrolled case, by contrast, no traffic control tactics are in place and the existing control system is assumed to be active.

Since this control system is not " customized" relative to a specific routing pattern of traffic, it is reasonable to expect that a change in routing would have a more moderate effect, if any, on evacuation travel times.

Secondly, the planned evacuation is "near-optimal," and provides for a more efficient evacuation than the uncontrolled case as verified by

~

documented study results.

Hence, any departure from the factors which produce this near-optimal condi--

tion in.either control or routing has the potential to compromise the efficiency of the evacuation perfor-mance.

The uncontrolled evacuation is, in contrast, far from optimal in efficiency.

Hence, since the

d 1 efficiency of the evacuation process is already rela-tively low, there is a smaller opportunity for any change in traffic routing, short-of a radical change, j

i to' degrade performance further.

38.

.Q. Have any other studies been conducted which provide information about effectiveness of the traffic control pian contained-in Appendix A?

A.

[Lieberman)

Yes.

As a supplement to the traffic con-trol plan, we were asked to study the effects of the potential construction of a new access road along an existing LILCO right-of-way on overall evacuation time

. estimates.

This proposed access road would be located just north of Route 25A and would run in an east-west direction.

For a major portion of the route, two

~

lanes of. service would be provided.

Two case studies were performed for this proposed roadway, one assuming a controlled evacuation, the other an uncontrolled one (Cases 35 and 36, respec-i tively).

The results of these studies, which are detailed in Attachment 12 and are summarized in At-i tachment 6, indicated that if this roadway were con-structed and used during an evacuation, that the resultant evacuation times for controlled and uncon-

' trolled evacuation would be identical at 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 30 minutes, and that both of these evacuation times would

_,.-.,_.r.,_~--

.., _....,..... _ _., _, - ~ _ _., _.. -. _...., _. _.. _,.. _. -.

i.

l be lower than for a controlled evacuation without the l

roadway (compare Cases 12, 35, 36 in Attachment 6).

l The reason for the apparent insensitivity, of evac-untion times under these conditions to the presence or absence of control mechanisms is that, as mentioned earlier, there are a number of means of reducing evac-uation times, such as revising origin-destination assignments, adding special control tactics and i

increasing evacuation roadway capacity.

In this case, the proposed access road is so effective in providing additional roadway capacity in the critical population area north of Route 25A, that the.use of special con-trol' tactics is not necessary to allocate more effi-l ciently limited roadway capacity.

It.should be noted I

f that there are no active plans to construct this hypo--

thetical roadway, which would require various County, and perhaps Stato, permits.

39.

Q. Could you briefly summarize the importance of a traf-fic-control plan to evacuation planning?

l A.

[Weismantle, Lieberman)

The primary importance of a traffic control plan is that it provides an oppor-tunity for reducing evacuation times and thereby low-l ering the' prospects of exposure of the public to radiation.

In addition, the presence of LERO guides' l

. -, -. ~. - -. -., - - - -., -, -,, -,, - - - -.,.. - - _, -, -... -,..,.

_ and the training of these guides helps to provide

~ assurance to the evacuating public that they are under the protection of a planned evacuation which has been carefully developed in their best interest.

The model runs performed for the Shoreham EFJ have shown that the plan provides important benefits in terms of re-duced evacuation travel times relative to a situation in which no operational plan is in effect.

The assignment of people to recommended evacuation routes offers a guide to people who recognize that perceptions based upon normal traffic patterns may not be applicable during an emergency situation.

While the effect of non-compliance has been shown to be lim-ited, as expressed in terms of lengthening evacuation time, the recommendation of a route should serve to relieve the public of uncertainty in that respect.

Certainly, the results have shown that non-compliance will not reduce travel times and therefore, the pref-erable course would be to comply with the recommended routes.

Finally, a traffic control plan is not the only means of reducing evacuation times.

Adding sufficient roadway capacity in critical evacuation areas has been shown to provide comparable time savings.

-- 40.

Q. In summary, are the allegations raised in Contention 65.C. meritorious?

A.

[Weismantle, Lieberman]

No.

Taking allegation 65.C.l. first, the assertion that LERO traffic guides will screen all motorists moving in a direction con-trary to prescribed traffic flow to determine whether each such person has " good reason" for going in that direction is inaccurate.

The LILCO Transition Plan has never contemplated that traffic guides would actu-ally prohibit motorists from going in any direction at any. time.

Earlier revisions of the Plan contemplated that traffic guides-would inquire, as appropriate, of t

motorists to determine where they were going if it was not in accordance with the prescribed traffic flow, in order to ensure that motorists were aware of the pos-sible consequences of their intended course.

Since that time, it has been concluded that if motorists are adequately aware of the existence of an evacuation from a predesignated area and that area is adequately marked, then there is no need for traffic guides to discourage trips into the evacuation area.

Accord-ingly, a combination of cones and signs will be used in order to define the boundary of the area to be evacuated and traffic guides will no longer attempt to discourage traffic from proceeding in directions

. other than those encouraged by the Plan.

Thus traffic guides will not be diverted from their primary duty of facilitating evacuating traffic.

The number of traf-fic guides assigned to traffic control posts, depicted in Figure 8 of Appendix A, has been confirmed by recent re-evaluation to be adequate to facilitate the flow of traffic; if any areas of further need are dem-onstrated, additional guides can easily be designated.

Thus this part of Contention 65.C.,

if it ever had any validity, no longer has any.

The second part of Contention 65.C.,

asserting that traffic controls may cause aggressive behavior on the part of potential evacuees, is addressed in the testi-

- nony of Drs. Dynes, Mileti, et al., being filed here-aith.

The assertion in Contention 65.C.3, that traffic gu3 des will be attempting to implement control strat-egies counter to the directions given by ordinary traffic signals, does not p' rove anything.

As is indi-cated in the discussion above and in the testimony of Drs. Dynes, Mileti, et al.,

filed herewith, common sense, experience and behavioral literature all tend to indicate that the directions given by qualified traffic guides to facilitate evacuation flow will tend

~ _

i to be obeyed by all or virtually all motorists.

Thus the confusion and congestion alleged by the contention should not occur.

However, even if motorists disre-garded the suggestions of traffic guides and proceeded to follow the instructions of normal traffic signals, the outcome would merely tend to approach the times listed-in the " uncontrolled" case.

In short, the traffic analyses have bounded the potential degrees of observance and nonobservance of traffic guides' direc-tions.

The testimony of Drs. Dynes, Mileti, et al.,

suggests that the " compliance" case is the more likely one.

However,.the effects on evacuation times of all

. extremes of compliance and non-compliance are readily discernible.

Thus the argument raised by Contention 65.C.3-has been accounted for.

4 Contention 65.C.4 asserts generally that some pre-scribed routes direct motorists to " travel contrary to their perceptions of the most expeditious way out of the EPZ."

As indicated above, in most cases this l

allegation simply is not the case, even where the con-

- straint of not allowing motorists to approach signifi-cantly closer to the plant at any time between the 1

j origin node and the destination. node is observed.

In those relatively rare cases where there is some I

i l

I I

m,

-..,-.~,m-

--,-----mmy,-,,,t,

--r m.g

.r mm,.

..,r-e--

y,--

,v,-+-,er-4 departure from the route which is quickest in fact, it is a logical alternative route and the " confusion and anxiety" forecast by the contention should be readily reduced by advance notice and information to motorists and implementation of the Plan in accordance with that advance information.

Thus the contention is not pro-bative.

Contention 65.D.

41.

Q. What information exists about the expected frequency of automobile accidents and breakdowns during an evac-untion?

A.

[Lieberman]

The existing documentation describing large scale evacuations of the. general public from areas at risk is very limited.

There are however, two documents which provide important information on the subject:

U.S.

Environmental Protection Agency Report EPA-520/6-74-002, entitled " Evacuation Risks - An Evaluation" by Joseph Hans, Jr.

and Thomas Sell published in June 1974 (hereinafter the "Hans & Sell Report").

" Hurricane Carla" prepared by M.E.

Treadwell and published by the Department of Defense, Office of Civil Defense in 1962.

The Hans & Sell Report describes a study which sur-veyed evacuattons over a 13-year period involving a

_ _ _ _ _ _ _ - _ _ _ _ _ - _ _ _ _ L total of over 1.1 million people.

The results of this survey indicate that few vehicle accidents of any con-sequence have occurred during evacuations and that none of these accidents has resulted in a fatality.

The report also concludes that for incidents where the total evacuation-vehicle-miles are large, predicting risk using National Motor Vehicle Accident Death and Injury rates may lead to an overestimation of expected injuries and deaths.

These lower accident rates dur-ing evacuations were presumably due to the low vehicle velocity and greater traffic control during these evacuations.

The Hurricane Carla report describes the evacuation of approximately half a million people from eastern Texas in anticipation of the approach of a major hur-ricane in 1961.

Over 300,000 of these persons are reported as having evacuated from a two-county area surrounding Galveston within a span of six hours-following' issuance of the evacuation order.

While the evacuating population had several days' notice of a potential need for evacuation, the episode illustrates the capacity of large numbers of persons to evacuate substantial areas in relatively short periods cf time.

The " Traffic Safety" section of that report reveals

- _ _. that during the entire evacuation only two minor traf-fic accidents occurred, neither of which resulted in injury or death.

42.

Q. If national accident statistics were used, what would be the predicted number of accidents and fatalities during an evacuation of the entire 10-mile EPZ?

A.

[Lieberman]

National accident statistics for 1980 indicate that accidents occur at a rate of approxi-mately one per every 77,000 vehicle-miles and that a fatality occurs every 26 million vehicle-miles (see Transportation and Traffic Engineering Handbook, Chap.

18).

An evacuation of the entire EPZ would produce approximately 304,000 vehicle-miles of travel.

Ac-cordingly, using national accident statistics, approx-imately 4 accidents would be predicted to occur, and there is a low probability that any of these accidents would result in a fatality.

Accident statictics from 1980 show that the rate of fatal accidents in the State of New York, measured in terms of fatal accidents per registered vehicle, was within 5% of the national average, and that the rate of fatal accidents in Suffolk County was slightly below that in the State of New York.

(Sources:

" Highway Statistics, 1980," U.S. Dept. of Transporta-tion, Federal Highway Administration, HHP-41/10-81; 1

_ 1 1

" Auto Vehicle Statistics, 1980," Annual Report of the New York State Dept. of Motor Vehicles).

d 43.

Q. How do these four predicted incidents compare with in-t formation presented in Contention 65.D. on accidents and breakdowns on the LIE in 1982 and the data relied on by Philip Herr in the overview report prepared by him for the Suffolk County Plan?

A..[Lieberman]

A direct comparison with either value is difficult to impossible, given the scarcity of infor-mation about the values.

However, some general obser-vations are possible.

If the 10,000 incidents on the entire Suffolk County structure of the LIE in 1982, a

cited in Contention 65.D.,

are converted into an hourly incident rate, then the result is 1.14'inci-a dents per hour.

If we take the evacuation time esti-mate for the entire EPZ under normal weather conditions -- 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes -- then 5 to 6 inci-dents would be predicted.

It must be observed that this comparison is of the grossest nature, since a i

majority of the Suffolk County portion of the LIE lies outside.the EPZ, and the incident rate reported in I

Contention 65.D. is not presented in terms, such as accidents or breakdowns per vehicle-mile, which can be readily translated into units more useful in pre-dicting the number of accidents during an evacuation.

a 4 -,

---r

,c, e----

,,, -..- - +

--,--,---e.w--,--m-

.,e, cy

-.y

,v-,y.

ww ww.,y-~+v+-=,-

-p- - --,-.--

---wr

In the. report entitled " Discussion Overview of the Suffolk County Emergency Plan," which was prepared by the Suffolk County Radiological Emergency Response Plan Steering Committee (November 1982), Mr. Herr

. states:

rough calculations indicate that an ordered 10-mile evacuation, with its spon-taneous voluntary expansion, would likely result'in one auto fatality and about 100 3

injuries at " usual" accident rates. 2g/

21/ Herr Associates calculations, rates from J. Hans and T.

Sell', Evacuation Risks,

'USEPA, Office of Radiation Programs, Washington, DC, 1974.

3 Discussion Overview at 16.

-Using Mr. Herr's handwritten calculations, that were obtained during discovery, I have reached the following conclusions about Mr.-Herr's statement:

1) the number of vehicle-miles used by Mr.

Herr was.70 times greater than that as-sociated with a total evacuation of the Shoreham EPZ; 2) the accident rate used by Mr. Herr is approximately half the rate obtained for national accident statistics; t

3) thus, had Mr. Herr used the vehicle-miles traveled during an evacuation of the entire Shoreham EPZ, he would have predicted that approximately 2 acci-t dents would occur during an evacuation.

y

.2.--,

-we--.-,--,.-,....-m.

,-_,,,.m-.

From these cursory reviews, neither the statistics cited in Contention 65.D. nor Mr. Herr's interpreta-tion of the Hans & Sell Report suggest that the use of national accident statistics is an inappropriate means of predicting the accidents likely to occur during an evacuation.

44.

Q. How did LILCO use these accident data to study the effect of accidents and breakdowns on evacuation time estimates?

A.

[Lieberman)

A study that consisted of two model runs (Cases 29 and 30 on Attachment 6 to this testimony) was conducted to determine the effect of accidents which block, at least to some extent, the passage of traffic along a stretch of roadway.

For each of these runs, it was assumed that four accidents had taken place simultaneously on different links of the net-work.

The accident locations were selected in a ran-dom manner, limited only by the constraint that each accident location must service a high volume of traf-

'fic.

It was further assumed that one of these four accidents created a blockage for a period of 30 minutes while the other three created blockages of 15 minutes each.

By causing these accidents to occur concurrently, it was felt that a worst-than-typical, if not worst, case condition was being simulated,

\\

_______ ___ _ _ _. since interaction among the effects of these accidents was permitted to occur.

45.

Q. What were the results of these studies?

A..[Lieberman)

These studies indicate that accident blockage influenced the movement of traffic in and about the area where the blockage occurred and, in one case, lowered the average speed of traffic movement over the network.

However, the effect-upon evacuation time estimates was negligible.

Specifically, the av-erage speed over the network for one run was reduced from about 6.8 miles / hour to approximately 6.5 miles / hour, while the total evacuation time remained at about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 55 minutes (compare Cases 12, 29-and 30 on Attachment 6).

46.

Q. Does the presence or absence of shoulders on the road affect the time estimates contained in Appendix A?

A.:[Lieberman]

Our physical survey of the roadway net-

. ork reveals that, in general, all two-way, two-lane w

roads on the evacuation network are sufficiently wide from one shoulder edge to the other, to permit at least one lane of moving traffic in the outbound

' direction and one lane in the inbound direction even if a disabled vehicle is parked on one should<2r or the other.

The only exception to this general observation

l is Shore Road, which is located in the extreme north-west of the EPZ.

The capacity assigned for this stretch of roadway in the KLD's evacuation time esti-mates is only 650 vehicles per hour.

This is about half the capacity estimated by PRC Voorhees.

Consequently, it can be reasonably assumed that if a blockage were to occur on a two-way, two-lane road, then there would be sufficient room on the shoulders to permit evacuating traffic to move around the block-age albeit at a lower rate of service.

A conservative estimate of the effect of such blockage is that capac-ity is ceduced by 50%.

If a vehicle is moved to a shoulder, then this assumed reduction in capacity is extremely conservative.

A 50% capacity reduction was applied in the studies when an accident occurred en a section of roadway which was specified as having only one available lane for evacuating traffic.

When an accident occurred on a section of roadway that had two or more lanes in each direction, then it was assumed that an entire lane was removed from service.

This assumption is also conservative, since all multi-lane roadways have adequate paved shoulders, making it likely that a disabled vehicle would be pushed onto the shoulder and the roadway would not lose an entire lane of service.

- 4 47.

Q. Were road construction and possible abandonment of vehicles considered in KLD's modeling runs?

A.

[Lieberman]

No.

Within the context of.an evacuation plan as specified by the guidelines in NUREG-0654 and

)

as implemented by many agencies for other nuclear power plants, the possibility of road construction was not explicitly considered.

While road construction or repair does occur with some regularity in Suffolk County, its location, effects and frequency are only speculative.

The possible abandonment of vehicles was likewise not considered to influence the movement of' traffic during the evacuation, on the bases that such abanconed vehicles would be pushed out of the way of evacuating traff c and their presence would not unduly influence the service rate of slow-moving traffic.

As noted above, there are adequate shoulders on all major roadways comprising the EPZ network.

Contention 65.E.

48.

Q. Are the "early dismissal of schools and the evacuation of [ people] in special facilities and the handicapped" considered in the evacuation time estimates contained in the LILCO Transition Plan?

A.

[Lieberman]

Yes.

49.

Q. How are they considered?

A.

[Lieberman)

Contention 65.D. arguably raises two issues about the dismissal of school children:

first, whether the transportation of school children to their

.,.m O. : L 0-fl homes will delcy the departure of families during an q-evacuation, and if it does, how will it affect evacua-p.1, '-,'

+ tW : "'.

tion time estimatec, and second, whether the presence fp'.:+a' (S f f-d of school buses on roadways will cause congestion and

%?.

..+.;',.,

thus lengthen evacuation times.

The LERO Plan assumes a l?g %.

that school children will normally be dismissed and J '.. '

'iksi; :

bussed home from school at the Alert stage (see pages

'i[( h' [

->.. : i IV-167 to -172 of Appendix A).

Given this assumption, dk['y. $

"n;.y p

' = -

three periods of time must be considered in determin-ing whether the return of school children will delay 9 @.',?

family departures.

The first period of time extends

.j :.4 from roughly 7:30 a.m. to 9:30 a.m. when children are

';{' #.

..9

-I

?y

,o y. <-

in the process of being taken from home to school.

c5 M e ?

. 4,+

a

- n gi.,,;

The second period, that extends from 9:30 a.m.

to 1:30

(

a ji p.m.,

is the time when children are in school, the 1

UgN53

, k'?,g,'~ -

?

buses are back at the depot and the bus drivers have g[hghgg left.

The third period of time, from 1:30 p.m. to

g.-

3:30 p.m.,

is when children are in the process of p,j ;, ;.

oq/- f '

d

. 1 being taken from school and transported home by bus.

y,";; -

7.

4 jpf +*.

,, g

[ i.:-

f.b.'.f 'I '

. m. ~.

"I i f,{ s.

.....y

.' r;-,., ;. y..

.h m,.'d b. %.

.% J... 5

' N N.% ', k.2 % 'V.j l.4.:,^

'l,N_.... j.;f %,,' l p,

].

L ! G.;'.* g;c.,;. ;.2. 3

,.< 7._... g y

.x

. w-

~

l..'

. 3, /. ?

?. :

.n.

..n,

i g. ~,.

y.,.A M'.

90.,:, r,.

M W,.2,def.

'. y?

..:l ; A c-

...v

' {;

Ql. - [,-

g.

u..

5 -

Of these three time periods, the second poses the

'.>.:wc-a 5 i

, %t,,..~ '

m,:

most serious questions about possible departure delays 4ttf;;g.

e

.a <.

since buses are at remote locations relative to the

~

^7i

. g* 1' g

. s..g schools and the bus drivers are not with their buses.

f. j,'. W

..i' r,

.i

. 5 & :-

., ".,J 7, "l :

Under these conditions, discussions with the various school districts have indicated that anywhere from 45 d r, e.

minutes to 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 30 minutes is required to summon bus

. X-4 +.w,, a,,

a

}f ;f "../.

?"

drivers and to get the buses to the schools.

It will

':s_ ;

y.3 3 then take between 30 minutes and 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 30 minutes to

.' c >

i t.

p. 4.... c; transport the school children home Consequently, it

.Y.k.

l

.,m-g

[

could take up to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> to return school children ft > k '.-

A,.

y.,

e g

.{,

home following the initial tone alert.

By comparison, p' ;..

1$

if an accident occurred during the first or third time g.

9.

s

..,y g.

+q= ~i _

b periods, it would only take 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> 30 minutes, at W. s.)..

s.:.

' '9 ~ 1. f ;

il most, to return all children home.

f ;r$9 -

r. :.. : ;..

• %,3,f,f f

M.[

These school-to-home travel time data were incorpo-

.,s:.

....y :.

rated explicitly in the statistical analysis detailed

.A d

.? s -

u

.+-e. u g' ;p"p,. ].., t.

A.c:.

in KLD Report TM-139, which is attached as Attachment 1., w T.,.

=.:

y, g.:,3;*/ -

t,

'O.

The first step in this process was to convert the

,M h-.'.

a. r,,9.

3. '.

~

school-to-home travel time data for the second t., V.P ' ' '

.y

....,_y J? ? ?y;,1 e..

40 period -- 9:30 a.m.

to 1:30 p.m.

-- into a time dis-4 t,,'

4 9

4 tribution (see pages 7 and 8 of Attachment 10).

Then,

.'. j % ;1., -

s,. : -

to account for those cases where a returning school 6..-,=.:.%.?.

Tu.' Q. a 3

W child was the last family member to return home, the c.C, 1 -

. ",ff.g't,

a. m

..+

k' v.

..,,,.4

_b.

f aj

.y.

,,(. f

n..m

..z

.,<..,.4

,,c7 k' %p?f, y 1 (.

', ":. y,y r

r w

hkA

-f.k%s.

c

3) b. _ k, h

r.;

[f

.,{,.,, i

. f,.,

e..

34 ~....

.b 7

l-M

/.'

7 ^'

time distribution for returning school children was combined with the time distribution for workers re-turning home, to produce a joint distribution of da-parture, times for families with school children.

This joint distribution.was then combined with the distri-bution of departure time for families without school children to produce the trip generation distribution for the entire auto-owning public.

This overall dis-tribution is displayed in Figures 1 to 3 of Attachment 10.

A review of these figures indicates that the re-turn of' school children, even during the period when buses are not readily available, yields a trip genera-tion. time distribution for the auto-owning public, which closely approximates time distributions produced by the Suffolk County Planning Department and PRC Voorhees.

The second possible concern about.the dismissal of school ~ children raised by Contention 65.D. is whether the number of buses transporting children home will create congestion that will extend evacuation times.

The simple answer to this concern is that it will not.

'The LILCO Transition Plan recommends the dismissal of school children at the Alert level.

Thus, for all but the' fastest breaking accidents,- the dismissal and f

l i

- I transporting of school children should have been com-pleted, or nearly completed, at the time the notice to

-evacuate is given, and thus, school bus traffic will be of no concern.

For fast breaking accidents, returning school buses could' interact with evacuating vehicles.

However, this interaction is unlikely to make a significant contribution to traffic volumes and thus to have any_effect on evacuation times, since school buses will comprise a fraction of one percent of the number of evacuating vehicles on the roadway network, and a large percentage of their travel will be on residential streets that are not part of the evacuation network.

For special facilities and thc handicapped, Conten-tion 65.D. questions whether the ambulances, ambu-lettes and buses needed to evacuate these people will create congestion that will lengthen evacuation times.

The total ~ number of vehicles-involved in evacuating these groups is less than 1 percent of the total of evacuating vehicles.

Accordingly, there is no basis to assume that any meaningful increase in evacuation time will result from the presence of such vehicles.

50.

-Q.

Contention 65.D. also mentions that trains "will be traveling in all directions through the EPZ."

Could trains affect the evacuation time estimates?

.. ~.

l -

A.

[Lieberman]

No.

All major. evacuation routes, except Edwards Avenue, which crocs the-Long Island Rail Road I

(LIRR) right-of-way, do so in a manner which does not 1

require automobiles to cross the railroad tracks --

these crossings are commonly referred'to as " separate i

grade" crossings.

Thus, these soparated grade cross-ings have no effect on traffic movement.

The Edwards Avenue. crossing is an "at grade" cross-ing,.which requires evacuating vehicles to cross LIRR tracks.

However, it is highly unlikely that the LIRR will disprtch trains through the EPZ if an evacuation is ordsred.

Even if it did, conflicts between evacua-ting' vehicles and passing trains should generally be i

less-than one minute in duration.

Contention 65.G.

51.

Q. Does 'idue LILCO Transition Plan include evacuation time estimates for people without access to private trans-portation?

l A.

[Weismantle, Lieberman]

Yes.

In response to a FEMA.

comment that time estimates were not expressly shown for the permanent population, transient population and special facilities population,-Table XV (Appendix A,

p. V-8) was added to Appendix A'by Revision 1 of the e

e-

...--r,.--v mw...,,.-

.-%v,e.--,-w,

,---y r,m.,,,-rtp.,,,-m-.+,.e---ymre-,,w,

-v-,-

--ir---g-,-ve,v-ywry-ww-e-,,-e,*.

v e

LILCO Transition Plan.

This table, which is attached as Attachment 13 hereto, was designed to comport with

'tdue example format-table cantained on page 4-16 of NUREG-0634,-Rev.

1.

In' addition to the breakdown of evacuation times contained in Table XV, Appendix A also contains detailed bus schedules for people _with-f out access to cars (Appendix A, pp. IV-74e to -74x),

and evacuation times for individual special facilities (Appendix A, p..IV-180).

52.

Q. What are the estimated evacuation times for the four groups listed in Contention 65.G.,

namely, " school children, persons without access to cars, persons in health care or other special facilities, and the hand-

-icapped?

A.

[Lieberman] Ecr school children, it is assumed that they will be evacuated as members of a family unit.

Therefore, their evacuation time will be the same as f

'either the automobile owning public or the public l

without access to cars.

l The_ evacuation of the public without access to_ cars I

will take from approximately 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> 30 minutes to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> 30 minutes following the notice to evacuate, I

under normal weather conditions (see Appendix A, p.

l' IV-76b).

A range of possible evacuation times is presented in Appendix A to reflect the impact of the speed of the accident on bus evacuation times.

As noted in the previous answer, evacuation time estimates for special facilities are presented in Table XV of Appendix A.

The evacuation time estimates

'for special facilities and the handicapped living at home but without access to private vehicles are in the process of being revised to reflect more precise in-formation on ambulance and ambulette availability and on the number of people needing t2 - service of these vehicles. - Revised evacuation time estimates indicate that evacuation of these groups will be completed within 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> following the notice to evacuate.

Contention 65.H.

53. ' Q.'What-is the purpose of. route spotters in the LILCO Transition ?lan?

A.

[Weismantle]

As stated in the LILCO Transition Plan Implementing Procedure OPIP 2.1.1 - Organization Implementation, the responsibility of the Evacuation

-Route Spotters-involves, "[t] raveling through the areas being evacuated to verify and report on evacua-tion traffic flow as directed by the Evacuation Route Coordinator."

,_m

...,.a.

_,,5

.~

,,,,._,,y.y,_..,,,,,,..,,.,,, -

m,,.,

,,y._

,,,y,,%y.,.,..,-m,__,.,7.yy.,_.,,,,

,_m,y,

,.-,~ym,

"95-i

'54.

'Q.

How many route spotters are specified by the Plan?

+

A.

[Weismantle]

There are six Evacuation Route Spotters identified and specified by the Plan (see Figure 2.1.1 of the LILCO Transition Plan, and OPIP 2.1.1, Attach-ment 3 ).

These route spotters will primarily patrol-the important evacuation routes that are listed in Figure 8.1 of Appendix A.

In addition, unless weather

-prohibits, LERO intends to adopt FEMA's recommendation and. utilize helicopters for aerial surveillance of evacuation routes.

The personnel, who will man these helleopters are in addition to the six route spotters listed in the Plan.

-It is tentatively planned that these personnel will be supplied by the helicopter companies.-

55.

Q. Will congested traffic conditions interfere with'these i

route spotters' ability to carry out their assigned tasks?

A.

[Weismantle, Liebermanj Evacuation route spotters are directed by OPIP 3.6.3 to travel routes opposite to j

l evacuating traffic to avoid congestion.

Of course, this will not always be possible.

Therefore, aerial surveillance by helicopter will constitute the primary means of route spotting because of its speed and insensitivity to traffic congestion.

_ - - 56.

Q. What is the importance of route spotters to the over-all LERO Plan?

3 A.

[Weismantle]

The role of route spotters in the LERO h

Plan is' limited because evacuation routes are pre-planned ~and not selected on an ad hoc basis.

Route spotters' primary duty will-be to monitor traffic flow and to report disabling-accidents back to the EOC.

Since these accidents are likely to be small in num-ber,.and the number of. people available to spot them are plentiful -- route spotters, helicopters and traf-fic guides, if'the accident were to occur at or near a-manned intertection -- the route spotter is not as basic to the successful implementation of the LERO Plan as are various other roles.

i:

_ -... -.-. -...-,., ~...._ -.

-