ML19248D253
| ML19248D253 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 07/26/1979 |
| From: | Harold Denton Office of Nuclear Reactor Regulation |
| To: | Wallis J AMERICAN GEOPHYSICAL UNION |
| Shared Package | |
| ML19248D254 | List: |
| References | |
| NUDOCS 7908150449 | |
| Download: ML19248D253 (8) | |
Text
.
y n%
f
\\
L. i T. c S i.4 T E S 5 % _Qf 'e
L'CLE AR H ECUL.L iGRY CGT.U.;iSSION 2
i a m.o c m ss e
e JUL 2 61979 Mr. J. R. Wallis President-Elect Section of Hydrology American Gecphysical Union c/o P.O. Box 218 Yorktown Heights, 'Y 10598 Cear Mr. Wallis:
Your letter of May 18, 1979 to Chairman Hendrie has been referred to me for response.
During the licensing reviews of the site the staff shared your concern over site flooding.
As I interpret your letter, your principal concerns are (1) inappropriate use of statistical estimates to assess flood risks at nuclear power facilities, (2) the " flexibility" of the MPF estimate at Three Mile Island, (3) whether we are aware of how the flood estimates are developed, and (4) why an alternate site was not chosen.
I appreciate the opportunity to consider and respond to your comments, and I'm providing the following background information and explanations.
While we do not concur with your characterization of the Water Resources Council's flood frequency estimating technique, we agree that contemporary statistical techniques are not appropriate for estimating rare flood events because (1) available streamflow records are short-term and may not be repre-sentative of the extreme flood producing potential of a watershed; and (2) we can establish no rational basis for the selection of confidence levels to minimize the residual error in estimates of severe flood event magnitude or likelihood.
These shortcomings preclude the quantification of flood risks frcm extreme events in any meaningful way.
The deterministic approach which we use also prohibits useful quantification of risk.
(Our approach is discussed below.)
Because we view the quantification of risk as a desirable goal, NRC has initiated a research program towards this end.
To date, the results are not promising.
All this does not inply, however, that the risk of flooding is high.
It simply means that the probability of occurrence while small is, in your words, unknowable.
g The Probable Maximum Flood (PMF) estimate, which you term the MPF, did not O
change as a result of Tropical Storm Agnes.
The applicant based the original Ci peak flood discharge estimate of 1.1 million cubic feet per second (cfs) on a C
preliminary and unapproved Corps of Engineers analysis.
The Island flood e
protection levee system was designed for that discharge.
The NRC staff indepen-dently reviewed the PMF estimate and questioned its adequacy.
At about this
% time, the Chief of Engineers, Corps of Engineers, was completing a review of
% the original Baltimore District 1.1 million cfs estimate.
The Corps of Engineers
'~ c;ncluded that the Susquehanna River PMF estimate (transferred to TMI) would be about 1.6 million cfs.
The NRC 5taff independently reviewed the estimate and concurred.
b)b Ob Vi 09504%
7 Tropical Storm Agnes occurred after these events.
A subsequent comparison of the storms generated by Agnes and the resulting Susquehanna River flood reaffirmed the appropriateness of the PMF estimate.
We have concluded that the risks at Three Mile Island associated with the occurrence of a PRF are sman because, in part, of the conservatism exercised by the Corps of Engineers, the utility, and the NRC staff in the development of PMF estimate and in the ficod protection provided.
You have provided no information which, in our view, alters that conclusion.
The PMF is derived by maximizing various flood producing factors, in conjunction with an occurrence of the Probable Maximum Precipitation (PMP).
The PMP is defined as the greatest depth of precipitation for a given duration, drainage area, and time of year for which there is virtually no risk of exceedance.
The PMP approaches the maximum possible precipitation within the limits of current hydrometecrological knowledge.
The PMP estimates used by the NRC staff were prepared by the National Oceanic and Atmospheric Administration (NOAA) and reported in Hydrometeorological Report No. 40, " Probable Maximum Precipitation, Susquehanna River Drainage above Harrisburg, Pa."
The approach used by NOAA for developing PMP estimates consists of using extreme record storm rainfalls as an indirect measurement of parameters and mechanisms causing extreme rainfalls.
A large sample of storms that include extreme record rainfalls are analyzed such that the optimum rainfall producing mechanisms and efficiencies are considered to have been experienced.
After transposing the storms to the area in question, they are further adjusted to produce maximum moisture.
Moisture maximization is accomplished by adjusting the historical storms to conditions represented by the maximum moisture that could have been contained in the storm, based on the maximum observed 12-hr. persisting dew points in the region of interest.
After the maximum rainfall depths for the many transposed storms are determined, the points are envek ped according to duration and area.
In addition to this maximum precipitation there are many additional conservatisms in PMF estimates.
These conservatisms include:
(1) the occurrence of antecedent rainfall, (2) critical storm centering, (3) critical rainfall sequence, (4) critical runoff parameters, (5) conservative computation of water surface profiles, (6) ceincident wind-wave activity, and (7) freeboard provided at TMI above the PMF level.
Each of these important conservatisms is discussed below:
1.
Antecedent Rainfall - About 3 to 5 days prior to the occurrence of the PMP, it is assumed that another rainfall event with approximately 50% of the PMP occurs.
This assumption is made to saturate the ground so that the postulated PMP rainfall will run off more readily, and also to produce a high base flow.
2.
Storm Cente ing - The rainfall is assumed to occur in a multiple elliptical pattern based on recorded historical storms; this pattern is adjusted at d centered over the drainage basin to produce the worst flood.
4 g f, n /, {
\\
.' 3.
_ Critical Rainfall Sequence - The rainfall amounts in the total PMP are distributed with time in such a manner as to produce the greatest flood peak.
For example, very large amounts of rainfall are assumed to occur later in the storm rather than at the beginning.
This means that the cound would become further saturated and the streams would have more inicial flow in them prior to the occurrence of larger amounts of rain.
The chance of rain falling exactly in this critical sequence is very small cased on observations on many intense storms.
4.
Critical Runoff Parameters - The more critical runoff considerations include infiltration rates, unit hydrograph peaking and stream routing and combining.
Each consideration is conservatively assessed.
5.
Water Surface Profiles - After the PMF discharge is computed, standard-step backwater models are used to compute the water surface elevation for the peak flood pr Mile.
The models are calibrated based on historical floods and river flows to determine friction coefficients (commonly known as Manning's 'n' values).
For TMI, these coef t icients were determined by the applicant during the safety review for Unit 1.
The determination was based on the 1936 and 196? floods, which had discharges of 750,000 and 464,000 cfs, respectively, at a Harrisburg gaging station.
The coefficients were then checked against the 1972 Agnes flood, which had a discharge of 1,020,000 cfs.
It was found that the applicant's model, using friction coefficients calibrated with the 1936 and 1964 floods, overpredicted the river stages which actually occurred in 1972.
This is in accordance with theory, in that the fridion coefficient generally decreases with increases in stage. k _ have concluc'ed that the friction coefficients used are very conservative and that the computed river stage at TMI is conservative.
6.
Freeboard Above PMF Level for Coincident Wind-Wave Activity - An extra margin of about four feet above the PMF level is provided at tne TMI, Unit 2 reactor to accommodate the wave effects produced by a 40 mile per hour custained wind blowing from the most critical uirection, coincident with the peak PMF level.
More freeboard is provided at cther TMI plant locations.
A significant increase in the discharge would be required to produce a flood whis.n exceeds the freeboard provided.
The levees at TMI were designed for the original PMP estimate of 1.1 million cfs.
The levees would be overtopped by a flood of 1.6 million cfs.
- However, the levees have been designed to be overtopped, beginning at the downstream ad, where the freeboard provided for the levees is less than that provided upstream.
The basis for this design was to preclude a " dam failure wave" that could impinge on plant structures if upstream portions failed first.
At any rate, it can be seen that failure of the levees is anticipated, but resulting water levels would not endanger plant facilities because of the extra freet.'ard margin providea at plant structures and flood protection provided at vital structures.
1, P '/b U/)
q/A
_4
!Ie have no basis for concluding why sites oi.. r than Three ' tile Island were considered less viable since our oractice is not to require alternative site considerations from a safety viewpoint.
Thank you for the opportunity to consider and respond to your concerns about the flood vulnerability at Three flile Island.
Sincerely, P
arold R. Denton, irector Office of lluclear Reactor Regulation DISTRIBUTION Chairman Hendrie Commissioners H.
lenton E. Case D. Cunch R. Bovd
~
V. Stello R. 71attson R. DeYoung D.
- uller R. ?ollmer R. Denise W. Kreger L. Hulman W. Bivins Johnson Response has been cleared w/Ccmmission 0G'-
and apcroved fcr discatch.
SECY : tail Facility (3) (70-1501) f per SECY 7/~c '70
"' ~
G. Ertter (EDO =C6338)
NRR Reading L
roff (EDO =06338)
Hf18 Reading
^
Central ;iles NRC PDR w/incomina SEE PREVIOUS YELLOW FOR OTHER CONCURRENCES
. ? S E..
JRR.. (
. Sc...
o, ic =
- LGHulman::RESS HRE
- ton O
ears >
k.[. [..
/
.. l..
\\
NgC PO8tM 313 (9 76) NRCM 0240 D 4.5 so v e a am m a? ****** o* **c 8 5 '-8***
hr f,' / )
.)
4 s
- t
GLL4 d a!,
/
/
We have no basis for concluding why sites other than Three Mil VIsland were considered less viable since our practice is not to require fternative site considerations from a safety viewpoint.
In the foregoing, I have responded to your concerns an ass. tions regarding the flood vulnerability at Three Mile Island, and have' presented the bases for our flooding criteria and evaluation.
Sincereig',
arold R. Denton, Director Office of Nuclear Reactor Regulation DISTRIBUTION H. Denton E. Case D. Bunch R. Boyd V. Stello R. Mattson R. DeYoung D. Muller G. Ertter (EDO #0 338)
M. Groff (ED0 #06338)
R. Vollmer R. Denise W. Kreger L. Hulman T. Johnson W. Bivins Central / File NRR Reading HMB Rsading NRC/PDR j
'd _,,.
c.o m
)
1
- d i-1 s\\ !
s CRESS:mim SE 454J DSE DSE ELDi t/
NRR EDO PH2014-A WSBivins LG dlinan Kreger D$
r i
HDenton 6/6/79 6/ /79 6/}/79 6/ ' /79 6/10 /79 6/ II /79 6/ /79 6/ /79
- e.,:
/*'
t' C
y
n 1.
.c..
FROM.
ACTION CONTROL DATES CONTROL NO J.R. Allts cOMnL DEAOLiNE 6/b/7, 06338 j
latsrnat10eaal adjDd33 Had61&c.5 AC KNOW L E DG ME N T DATr OF DOCUMENT lNTERIM REPLY h][9 T O:
PREPARE FOR SIGNATURE OF-
O CHAIRMAN L
FILE LOCATION C EXECUTIVE OIRECTOR OTHERr D$CRIPTION Z LETTER UMEMO C REPORT [ OTHER SPECIAL INSTRUCTIONS OR REMARKS
@csticas concaming the flood rist at Answer to be cleared with Ccenission. EDO i
tna M site will handle before dispatching.
I f
CEASSIFIED OATA COCU V ENT, C O PV NC.
'C LA SS IF I C A TIC N NUMBER QF DAGES l
CATEGORY POS T A L R E 3pT RY N 7.j
[
N5'
] 90 2 FRO ASSIG'/3 TO I
DA :
INFORMATION ROUTING LEG AL REV1EW FIN A L
- COPY M
S Q /79 ASSIGNEO TO-OATE J NO LtJ AL C8 JECT!ONS NOTf' Y.
[
[ ECO ADMIN i COR RES B A l
E x T.
CO\\1MENTS. NOTIFY.
EXT.
JCAE NOTIFICATION RECOMMENDEC O vEs O NO NRC ov 232 EXECUTI'.'E CIPECTOR FOR CPERATIONS CO tv 0 T R E tf 0 VE Td/S COP Y PRINCIPAL CORRESPONDENCE CONTROL l
i uun u=~
,. r, I%
5
s 6 6##
~~
79 1 R O 1 Logging Oate No.
/
NRC SECRETARI AT Oate TO.
O Commissioner Gen. Counset
,5M mec. Dir Oper solicitor O Cong L.aison O secretary O eubhc Affairs 0
J.R. Wallis, Pres.
, nc om,,,,
American Geophys1 Cal Union g,om.
5/18/79
~ Chairman Hendrie o,,
ro Provides info. on TMI flood plain sumect O erenare reoiv to, signature oi:
O Cnairman C Commissioner O eoo. oC. Ct. sot. ea. sECv 0 signature bioca omittee O a turn oriainai ot mcomin9.itn resoonse XHXXFor $ rect reply *
$Uspense: May 31
._/
O Foraoo,ooriateact,on s
~ <
l
' j > ;., t 0 For iniormation For recommendation Cpys to:
Chm, Cmrs, PE, OGC Remark s Combs aor en. Commiss.on send inre-i3) coa es of reci, to secy Mao F ety ACTIGN SL P N6 c o2
- }%
,r u;%
t EDO CONTROL NO.
DUE:
THIS EDO CONTROL SHOULD RECEIVE HIGHEST PRIORITY IN HANDLING AND PREPARATION OF RESPONSE.
PERSON RESPONSIBLE FOR PREPARING REPLY SHOULD CALL M. GROFF, EXT. 27750, IMMEDIATELY UPON RECEIPT.
/
- k. t -
,-f DATE:
TIME:
'/ '
.)
-~
[.,-
ASSIGNED TO:
1 TIME AND DATE ASSIGNED Received Dispatched Division Director km 5/43 - Af a r f/:L3 -.7.:g y Assistant Director
'/l a. 8 ' #
hl,4 - I M hM Branch Chief
- WWN4 f 24.- $
k 5)24 - F'13 bid'4C ~
5/24 - $lTr0
[
Other OELD TIME AND DATE FORWARDED FOR CONCURRENCE Received Dispatched OELD Other Branch Chief G /7 9.@
/,/7 9~?o _
Assistant Director i}4 U
Divisior) Director THIS SLIP SHOULD STAY WITH THE PACKAGE AT ALL TIMES AND BE RETUPNED WHEN COMPLETED.
Harold R. Denton, Director Office of Nuclear Reactor Regulation
\\
'e' y a
,