Regulatory Guide 1.59

Design Basis Floods for Nuclear Power Plants

ML13350A359
Person / Time
Issue date:	08/31/1973
From:	US Atomic Energy Commission (AEC)
To:
References
RG-1.059
Download: ML13350A359 (16)

August 1973at.August 1973U.S. ATOMIC ENERGY COMMISSIONREGULATORY GUIDEDIRE"W"TORATE OF REGULATORY STANDARDSREGULATORY GUIDE 1.59DESIGN BASIS FLOODS FOR NUCLEAR POWER PLANTS

A. INTRODUCTION

General Design Criterion 2. "-Design Bases forProtection Against Natural Phenomentia." of Appendix Ato 10 CFR Part 50. **General Design Criteria for NuclearPower Plants." requires. in part. that structures. systems.and components important to safety be designed towithstand the effects of natural phenomena such asfloods, tsunami. and seiches without loss of capability toperform their safety functions. Criterion 2 also requiresthat the design bases for these structures, systems. andcomponents reflect: (I) appropriate consideration of themost severe of tihe natural phenomena that have beenhistorically reported for the site and surrounding region.with sufficient margin for the limited accuracy andquantity of the historical data and the period of time illwhich the data have been accumulated. (2) appropriatecombinations of the effects of normal and accidentconditions with the effects of the natural plhenonlena.and (3) the importance of the safety functions to beperformed.Paragraph 100.10 (c) of 10 CFR Part 100,"ReactorSite Criteria," requires that physical characteristics ofthe site, including seismology. meteorology, geology.and hydrology, be taken into account in determining theacceptability of a site for a nuclear power reactor.Appendix A. "Seismic arid Geologic Siting Criteriafor Nuclear Power Plants." was published in the FederalRegister on November 25, 1971 (36 FR 22601) as aproposed amendment to 10 CFR Part 100. Theproposed appendix would specify investigations requiredfor a detailed study of seismically induced floods andwater waves. Proposed Appendix A to 10 CFR Part 100would also require that (lie determination of designbases for seismically induced floods and water waves bebased on the results of the required geologic and seismicinvestigations and that these design bases be taken intoaccount in the design of tile nuclear power plant.TlThis guide describes a1n acceplahl' ntl lhod (itdeterminirng fOr siles strealis tit riveis ilie designbasis floods that nuclear power plants maust lie designedto withstand without loss of saltety-related functions. Itfurther discusses tlie phenomlena producing design basis floods for coastal. estuary; and Gieat Lakessites. It does not discuss the design requirements forflood protection. The Advisory Committee on ReactorSafeguards has been consulted concerning this guide andhas concurred in the regulatory position.

B. DISCUSSION

Nuclear poower plants must be designed itf preventthe loss of safety-relat ed functions resulltig front themost severe flood conditions thai call reasonably bepredicted to occur at a site as a result of sevelehydrometenrological conditions, seismic activity. orboth.The Corps of Engineers for many years has studiedconditions arid circumstances relating to floods andflood control. As a result of these studies, it hasdeveloped a definition for a probable niaxinmui 'lood(PM F)' and attendant analytical techniques forestimating with an acceptable degree oft conservattsmflood levels on streatis or rivers resulting fromihydromLeteorological conditions. For estimatingseismtiically induced flood levels. an acceptable degree of'Corps ot tEngincecr Pribahltc Ma',intsni ItIodt definlililnappears in many publication, of thait :g00ncy sch 1is IEngineeringCircular EC-I 110-2-27, Change I. 'T"ngincering :sndDesign -Policies and Procedures Perlaining 10 t)eerminaition ofSpillway Capalities and Frecboard Allowances fir t)jn<,. dated19 Feb. 1968. Ttie probahble niamimuni fhlood is atso direcllyanalogous to ftte Corps (if 1'ngineers "Spillway Design Itlod" asused for darns whose failures would result in a significant toss oflire and property.USAEC REGULATORY GUIDES Copies of published guides may be obtained by request indicating the divietoatdesired to the US. Atomic Energy Commrstiori, Washington. D.C. 20545,Regulatory Guides e issued to describe and make available to the public Attention: Director of Regulatory Standards. Comments and stuggetions fotmethods aeceptsble to the AEC Regulatory staff of implementing specific parts of Irtroovements In these guides are encouraged and should be sent to the Secrets'ythe Commission's regulations. to delineate techniques used by the stafl in of the Commission, U.S. Atomic Energy Commission. Washington, D.C. 20545.evaluating ecilfic problems or posttulatd accidents, or to provide guidane to Attention: Chief, Public ProctedingtStlff.eaplicants. RegAnftory Guides are not substitutes for regulationt and compliancewith thern is not required. Methods and solutions different from those set out in The guides are issued In the following ten broad divisions:the guides will be acceptable if they provide a basis for the findings requisite tothe itauence or continuance of a permit or license by the Commitsion. 2. Research and Test Reactors 6. Tranportation3. Fuels ard Materials racilitien 8. Occupational HealthPublished guides will be revised periodically, as appropriate, to accommodate 4. Environmentall and Siting 9. Antitrust Reviewcomments end to reflect new information or experlence. 5. Materialt and Plant Protection 10. General conservatism for evaluating the effects of lte initiatingevent is provided by the proposed Appendix A to 10CFR Part 100.The resulting I'rom the worst site-relatedflood precHble at the nuclear power plant (e.g.. PMF,seismically induced flood, seiche. surge. severe localprecipitation) with attendant wind-generatcd waveactivily constitute the design basis flood conditions thatsafety-related structures. systems. and componentsidentified in Regulatory Guide 1.292 must he designedito withstand and remain functional.For sites along streams or rivers, a hypotheticalprobable maximum iflood of the severity defined by theCorps of Engineers generally provides the design basisflood. Ior sites alone lakes or seashores, a floodCondition of cotinparahle severity could be produced bythe most severe combination of hydrometeorologicalparameters reasonably possible, such as may beprotduced by a probable maxinmum hurricane" .or by aprobable matximum seiche. On estuaries. a probableinaxinitun rivet c lood. a probable maximum surge. aprobable tuaximnuni seiche. or a reasonable combinationof less severe phenomenologically caused flooding eventsshould all he considered in arriving at design basis floodconditions comparable in frequency of occurrence witha probable ;naximum flood on streams and rivers.Ini addition to floods produced by severeIh y d rometeorological conditions. Ihe most severeseismically induced floods reasonably possible should beconsidered for each site. Along streams. rivers, andestuaries, seisinically induced floods may be producedby dam failures or landslides. Along lakeshores,coastlines, and estuaries. seismically induced ortst, namit-ype flooding should be considered.Consideration of seismically induced floods shouldinclude the same range of seismic events as is postulated2 Regulatory Guide 1L29 (Safety Guide 29), "Seismic DesignClassification," identifies waler.cooled nuclear power plantstructures. system,. and components that should be designed towithstand the effects of the Safe Shutdown Earthquake andremain funetionalt These structures. systems. and componentsare those necessary to assure (I) the integrity of the reactorcoolant pressure boundary, (2) the capability to shut down thereactor and maintain it in a ,.afe shutdown condition, or (3) thecapability to prevent or mitigate the consequences of accidentswhich could result in potential offsite exposures comparable tothe guideline exposures, of I1t CFR Part tI0O. These samestructure%, systems, and components should also be designed towithstand conditions resulting from the design basis flood andremain functional.If is expected that safety-related structures, systemns. andcomponents of other types of nuclear power plants will beidentified in future Regulatory guides. In the interim. RegulatoryGuide 1.29 should be used as guidance when identifyingrafety-related structures, systems, and components of othertypes of nuclear power plants.'See Corps of Engineers Coastal Engineering ResearchCenter "Technical Report No. 4, Shore Protection, Planning andDesign." third edition. 1966.for the design of the nuclear plant. For instance, theanalysis of floods caused by darn failures, landslides, ortsunami requires consideration of seismic events of theseverity of the Safe Shutdown Earthquake occurring atthe location that would produce the worst such flood atthe nuclear power plant site. In the case of seismicallyinduced floods along rivers, lakes, and estuaries whichmay be produced by events less severe than a SafeShutdown Earthquake, consideration should be given tothe coincident occurrence of floods due to severehydrometeorological conditions, but only where theeffects on the plant are worse, and the probability ofsuch combined events may be greater, than the effectson the plant of an individual occurrence of the mostsevere event of either type. For example. a seismicallyinduced flood produced by an earthquake ofapproximately one-hal f the Safe Shutdown severitycoincident with a runoff-type flood produced by tiheworst regional storm of record may be considered tohave approximately the same severity as an earthquakeof Safe Shutdown severity coincident with about a25-year flood. For the specific case of seismicallyinduced floods due it) dam failures, an evaluat ion shouldbe made of flood wave! which may be caused bydomino-type darn failures triggered by a seismicallyinduced failure of a critically located dam and of floodwaves which may be caused by multiple darn failur':s in aregion where dams may be located close enough togetherthat a single seismic event can cause multiple failutes.Each of the severe flood types discussed aboveshould represent the upper limit of allphenomenologically caused flood potential combi-nations considered reasonably possible, and analyticaltechniques are available and should generally be used fortheir prediction for individual sites. Those techniquesapplicable to PMF and seismically induced floodestimates on streams and rivers are presented inAppendix A to this guide. Similar apperdices for coastal,estuary. and Great Lakes sites, reflecting comparablelevels of risk. will be issued as they become available.Analyses of only the most severe flood conditionsmay not indicate potential threats to safety-relatedsystems that might result from combinations of floodconditions thought to be less severe. Therefore.reasonable combinations of less-severe flood conditionsshould also be considered to the extent needed for aconsistent level of conservatism. Such combinationsshould be evaluated in cases where the probability oftheir existing at the same time and having significantconsequences is at least comparable to that associatedwith the most severe hydrometeorological or seismicallyinduced flood. For example, a failure of relatively highlevees adjacent to a plant could occur during floods lesssevere than the worst site-related flood, but wouldproduce conditions more severe than would result duringa greater flood (where a levee failure elsewhere wouldproduce less severe conditions a[ the plant site).1.59-2 Wind-generated wave activity may produce severeflood-induced static and dynamic conditions eitherindependent of or coincident with severehydromelcorological or scisnmic flood-producingmechanisms. For example, along a lake. reservoir. river,or seashore, reasonably severe wave action should heconsidered coincident with the probable maximumwater level conditions.4 The coincidence of waveactivily with probable maximum water level conditionsshould take into account the fact that sufficient timecan elapse between the occurrence of the assumedmeteorological mechanism and the maximum water levelto allow subsequent meteorological activity to producesubstantial wind-generated waves coincident with thehigh water level produced by the initial event. Inaddition, the most severe wave activity at the site thatcan be generated by distant hydrometeorological activityshould be considered. For instance, coastal locationsmay be subjected to severe wave action caused by adistant storm that, although not as severe as a localstorm (e.g., a probable maximum hurricane), mayproduce more severe wave action because of a very longwave-generating fetch. The most severe wave activity attile site that may be generated by conditions at adistance from the site should be considered in suchcases. In addition, assurance should be provided thatsafety systems necessary for cold shutdown andmaintenance thereof are designed to withstand the staticand dynamic effects resulting from frequent flood levelscoincident with the waves that would be produced bythe maximum gradient wind for the site (based on astudy of historical regional meteorology).

C. REGULATORY POSITION

I. The conditions resulting from the worst site-relatedflood probable at a nuclear power plant (e.g., PNIF.seismically induced flood, hurricane. seiche, surge. heavylocal precipitation) with attendant wind-generated waveactivity constitute the design basis flood conditions thatsafety-related structures. systems, and compor.Entsidentified in Regulatory Guide 1.292 must be designedto withstand and remain functional.a. On streams and rivers, the Corps of Engineersdefinition of a probable maximum flood (PMF) withattendant analytical techniques (summarized inAppendix A of this guide) provides an acceptable levelof conservatism for estimating flood levels caused bysevere hydrometeorological conditions.4 Probable Maximum Water Level Is deflined by the Corps ofEngineers as "the maximum still water level (i.e.. exclusive oflocal coincident wave runup) which can be produced by themost severe combination or hydrometeorological and/or seismicparameters reasonably possible for a particular location. Suchphenomena are hurricanes, moving squall lines, other cyclonicmeteorological events. tsunami, etc., which, when combinedwith the physical response of a body of water and severeambient hydrological conditions, would produce a still waterlevel that has virtually no risk of being exceeded." (SecAppendix A to this guide)b. Along lakeshores. coastlines, and estuaries.eslimales of flood levels resulting frorn severe surges.seiches. and wave action caused by hydronteteorologicalactivity should he based on criteria cOl uparahle inconservatism to those used for probable maximumIhoods. Criteria and analytical techniques providing thislevel of conservatism for the analysis of these events willhe summai'zed in subsequent appendices to ilbis guide.c. Flood Aronditions Ihat could be caused byearthquakes of the severity used in thie design of thenuclear facility should also be considered in establishingthe design hasis flood. A simplified analytical techniquefor evaluating the hydrologic effects of seismicallyinduced dam failures disctrssed herein is presented inAppendix A of this guide. Techniques for evaluating theeffects of tsunami will be presented in futureappendices.d. In addition to the analyses of the most severefloods I hat may be induced by eitherhydrometeorological or seismic mechanisms. reasonablecombinations of less-severe flood conditions should alsobe considered to the extent needed for a consistent levelof conservatism, Such combinations should be evaluatedin cases where the probability of their existing at thesame time and having significant consequenceL is at leastcomparable to that associated with the most severehydrometeorological or seismically induced flood.e. To the water levels associated with the worstsite-related flood possible (as determined fromparagraphs a.. b.. c.. or d. above) should be added theeffects of coincident wind-generated wave activity togenerally define the upper limit of flood potential. Anacceptable analytical basis for wind-generated waveactivity coincident with probable maximum water levelsis the assumption of a 40-mph overland wind from themost critical wind-wave-producing direction, unlesshistorical windstorm data can be used to substantiatethat such an event (i.e., wind direction and/or speed) ismore extreme than has occurred regionally. However. ifthe mechanism producing the maximum water level.such as a hurricane, would itself produce higher waves,then these higher waves should be used as the designbasis.2. As an alternative to designing "hardened"protection- for all safety-related structures. systems. andcomponents as specified in regulatory position I .above,it is permissible to not provide hardened protection forsome of these features if:a. Sufficient warning time is shown to be availableto shut the plant down and implement adequateemergency procedures"b. All safety-related structures. systems. andcomponents identified in Regulatory Guide 1.29) areI tardened portection means structural provisionsincorporated in the plant design that will protect %afcty-relatedstructures, systems, and components from the static anddynamic effects of floods. Examples of the types of floodprotection to be provided for nuclear power plants will le thesubject of a separate regulatory guide.1.59-3 designed to withstand the flood conditions resultingfrom a severe slorm such as tie worst regional storm ofrecord"' with attendant wind-generated wave activityIhl1 mw. lie produced by the worst winds of record andreiain functional:c. In addition to the analyses required byparagraph 2.b. above, reasonable combinations ofFor sites along streams and rivers thik event is characterizedby the Corps of. Engineer! definition of a Standard ProjcctFlood. Such floods have been found to produce tlow ratesgenerally 40 wo fill percenrtl tihte P.SIF. For sites along seahorc,this event le ch;taracterized b% the Corp, oi t" :ineinctrsdefiNition of j Standard Projecl Ilurricane. For other 'ijC acomparable level olf risk should le assumed.less-severe flood conditions are also considered to theextent needed for the consistent level of conservatism:andd. In addition it) paragraph 2.b. above, at leastthose structutres, systems, and components necessary forcoldl shutdown and maintenance thereof are designedwith "hardened" protective fealtures to withstand tlieentire range of flo0d conditions up to and including theworst site-related flood probable (e.g., PM F. seismicallyinduced flood. hutricane, surge, seiclhe, heavy localiercipitalion) with coincident wind-generated waveact ion a s discussed in regulatory positiotn I. above andremain funictiolnal.i1.59-4

• a0APPENDIX ATABLE OF CONTENTSA.IA.2A.3A.4A.5A.6A.?AS8A.9A.10A.1 IIntroduction ..........................Probable Maxinmum Flood (PMF) ..........Hydrologic Characieristics ................Hlood Hydrograph Analyses ..............Precipitation Losses and Base Flow .........Runoff M odel .........................Probable Maximum Precipitation Estimates ..Channel and Reservoir Routing ............PNI F llydrograph Estimates ...............Seismically Induced Floods ..............Water Level Detei minations ................................... .5(1.5.......................................................... I .q...... .................5' .6I..,. I................... ...................... 1.59-7...................... 59 -8.. .... ....... ... ....... 1.5 -... ..................... 1.59-1 I.................... 1.5 .i 1 2..................... 1.59 -12.................... 1.59-)13A.1 2 Coincident Wind-Wave Activity ................................. 1.59-13References ....................................... ........ 1.59-15PROBAELE MAXIMUM AND SEISMICALLY INDUCED FLOODSON STREAMS AND RIVERSA.1 INTRODUCTIONThis appendix has been prepared to provideguidance for flood analyses required in support ofapplications for licenses for nuclear power plants to belocated on streams and rivers. Because of the depth anddiversity of presently available techniques. this appendixsummarizes acceptable methods for estimating probablemaximum precipitation, for developing rainfall-runoffmodels, for analyzing seismically induced dam failures.and for estimating the resulting water levels.The probable maximum flood may be thought of asone generated by precipitation, and a seismicallyinduced flood as one caused by dam failure. For.manysites, however, these two types do not constitute theworst potential flood danger to the safety of the nuclearpower plant. Analyses of other flood types (e.g.,tsunami, seiches, surges) will be discussed in subsequentappendices.The probable maximum flood (PMF) on streamsand rivers is compared with the upper limit of floodpotential that may be caused by other phenomena todevelop a basis for the design of safety-related structuresand systems required to initiate and maintain safeshu.tdown of a nuclear pow'er plant. This appendix.outlines the nature and scope of detailed hydrologicengineering activities involved in determining estimatesfor the PMF and for seismically induced floods resultingfrom dam failures, and describes the situations for whichless extensive analyses are acceptable.Estimation of a probable maximum flood (PMF)requires the determination of the hydrologic response(losses, base flow, routing, and runoff model) ofwatersheds to intense rainfall, verification based onhistorical storm and runoff data (fhood hydrographanalysis). the most severe precipitation reasonablypossible (probable maximurn precipitation-.lPIriinimum losses. tnaximum base flow. channel andreservoir routing, the adequacy of existing and propetsedriver control structures to safely pass a PMF. water leveldeterminations, and the superposition of potentialwind-generated wave activity. Seismically induced Ihoodssuch as may be produced by dam failures or landslides.may be analytically evaluated using many PMFestimating components (e.g.. routing techniques. waterlevel determinations) after conservative assumptions offlood wave initiation (such as dam failures) have beenmade. Each potential flood component requires anin-depth analysis. and the basic data and results shouldbe evaluated to assure that the PMF estimate isconservative. In addition. the flood potential fromseismically induced causes must be compared with thePMF to provideappropriate flood design bases. but theseismically induced flood potential may be evaluated bysimplified methods when conservatively determinedresults provide acceptable design bases.Three exceptions to use of the above-descrihedanalyses are considered acceptable as follows:a. No flood analysis is required for nuclear powerplant sites where it is obvious that a PMF or sismicallyinduced flooding has no bearing. Examples of such sitesare coastal locations (where it is obvious that surges.wave action, or tsunami would produce controllingwater levels and flood conditions) and hilltop or "dry"sites.b. Where PNIF or seismically induced floodestimates of a quality comparable to that indicatedherein exist for locations near the site of the nuclearpower planw, they may be extrapolated directly to thesite, if such extrapolations do not introduce potential1.59-5 errors of more than about a foot in PMF water levelestimates.c. It is recognized that an in-depth PNF estimatemay not le warranted because of the inherent capabilityof lihe design of some nuclear power plants to functionsofely with little or no special provisions or because thetime and costs of making such an estinate ate notconinmensurate with the cost of providing protection. Insuch cases, other nieans of estimating design basis flnoisare acceptable if it can he demonstrated that thetechnique utiliied or the estimate itself' is conservative.Similarly. conservative estimates of seisinically inducedflood potenti:al may provide adequate denmonstration ofnuclear power plant safety.A.2. PROBABLE MAXIMUM FLOOD (PMF)Probable maxir'inn Ilood sttid:,- should becoiripatible with the specific definitions and criteriasummnnarized as follows:a. The Corp; of Engineers defines the PMF as "thehyp.,thetical I1(x)d characteristics (peak discharge.Volmnc. arid hydroge? ih shape) that are considered to hethe most severe reasonrabl\ possible at a particularlocation. haised on relatIively comprehensivehvdr ometeoro logic:' I analysis o f criticalrt niill-producing precip tation (and snowmell. ifpertinent) and hydroltgic factors favorable forfltiod ruinoff." Detailed PM F determinationsare usuially prepared by estimating the areal distributionof *'prohbahe maximurn" precipitation (PNIP) over fliesubject drainage basin in critical periods of time. andcomputing the residual runoff hydrograph likely toresult with critical coincident conditions of groundwetness and related factors. PMF estimates are usuallybased un the observed and deduced characteristics ofhi St ori:al flood-producing storms anid associatedhy d ro log ic factors modified on the basis ofhydronietecorological analyses to represent the mostsevere runoff conditions considered to be "reasonablypossible" in the particular drainage basin under study. Inaddition to determining the PMF for adjacent large riversand strearims. a local PMF should be estimated for eachlocal drainae coUrSe that can influence safety-relatedfacilities, including lie roofs of safety-related buildings.to assure that local intense precipitation cannotconstitule a threat to tile safety of tlie nuclear powerplant.b. Probable maxinium precipitation is defined bytile Corps of Engineers and the National Oceanic andAtnmospheric Administrat ion (NOAA) as "thie t liheret icallygreatest depth of precipitation for a given duration thatis nieleorologically possible over the applicable drainagearea that would produce flood flows of which there isvirtually no risk of being exceeded. These estimatesusually involve detailed analyses of historicalflood-producing storms in the general region of thedrainage basin under study. arid certain nmodificalionsand extrapolations of historical data and reflect moresevere rainfall-runoff relations than actually recorded.insofar as these are deemed reasonably possible ofoccurrence on the basis of hydrometeorologicalreasoning." The PMP should represent the depth, time,and space distribution of precipitation that approachestile upper limit of what the atmosphere and regionaltopography can i Iroduce. The critical PMPmeteorological conditions are based on an analysis ofair-mass properties (e.g., effective precipitable water,depth of inflow layer, temperatures, winds), synopticsituations prevailing during recorded storms in tileregion, topographical features, season of occurrence, andlocation oh the respective areas involved. The values thusderived are designated as the PMP, since they aredeterinited wit thin I lie limitations of currentmeteorological theory and available data and are basedon the most effective combinalion of critical factorscon Iollinrg.A.3 HYDROLOGIC CHARACTERISTICSHydrologic characteristics of the watershed andsireani channels relative to the plant site should beduierniniied fromt the Iollowing:a. A topographic map of the drainage basinshowing watershed boundaries for the entire basin andprincipal tributaries and other subbasins that arepertinent. The mnap should include ; location ofprincipal stream gaging stations and other hydrologicallyrelated record collection stations (e.g., streamflow,precipitation) and the locations of existing and proposedreseroirs.b. The drainage areas in each of the pertinentwatersheds or subbasins above gaging stations, reservoirs,any river control structures, and any unusual terrainfeatures that could affect flood runoff. All majorreservoirs and channel improvements that will have amajor influence on streamfnow during flood periodsshould be considered. In addition, the age of existingstructures and information concerning proposed projectsaffecting runoff characteristics or streamflow is neededto adjust streamflow records to "pre-project(s)" and"with project(s)" conditions as follows:(1) The term "pre-project(s) conditions" refersto all characteristics of watershed features anddevelopments that affect runoff characteristics. Existingconditions are assumed to exist in the fiture if projectsare to be operated in a similar manner during the life ofthe proposed nuclear power plant and watershed runoffcharacteristics are not expected to change due todevelopment.(2) The term "with project(s)" refers to thefuture effects of projects being analyzed, assuming theywill exist in the future and operate as specified. Ifexisting projects were not operational during historicalfloods and may be expected to be effective during thelifetime of the nuce.r, power plant. their effects onhistorical floods should be determined as part of theanalyses out lined in Sections A.5. A.6. and A.8.c. Surface and subsurface characteristics thataffecl runoff and streamiflow to a major degree, (e.g..1.59-6 large swamp areas, noncontributing drainage areas,groundwater flow, and other watershed features of anunusual nature to the extent needed to explain unusualcharacteristics of streamflow).d. Topographic features of the watershed andhi-!orical flood profiles or high water marks. particularlyin the vicinity of the nuclear power plant.e. Stream channel distances hetween river controlstructures, major tributaries, and the plant site.f. Data on major storms and resulting floods ofrecord in the drainage basin. Primary at tcntion should begiven to those events having a major bearing onhydrologic computations. It is usually necessary toanalyze a few major floods of record in order to developsuch things as unit hydrograph relations, infiltrationindices, base flow relationships, information on floodrouting relationships, and flood profiles. lxcept inunusual cases, climatological data available from theDepartment of Commerce. The U.S. Army Corps ofEngineers. National Oceanic and AtmosphericAdministration and other public sources are adequate tomeet the data requirements for storm precipitationhistories. The data should include:(I) Hydrographs of major historical floods forpertinent locations in the basin, where available, fromthe U.S. Geological Survey or other sources.(2) St o rmi precipitation records,depth-area-duration data, and any available isohyetalmaps for the most severe local historical storms or floodsthat will be used to estimate basin hydrologicalcharacteristics.A.4 FLOOD HYDROGRAPH ANALYSESFlood hydrograph analyses and relatedcomputations should be used to derive and verify thefundamental hydrologic factors of precipitation losses(see Section A.5) and the runoff model (see SectionA.6). The analyses of observed flood hydrographs' ofstreamflow and related storm precipitation (Ref. I) usebasic data and information referred to in Section A.3above. The sizes and topographic freatures of thesubbasin drainage areas upstream of the location ofinterest should be used to estimate runoff response foreach individual hydrologically similar subbasin utilizedin the total basin runoff model. Subbasin runof'response characteristics are estimated from historicalstorm precipitation and streamflow records where suchiare available, and by synthetic means where nostreamflow records are available. The analysis of floodhydrographs (Ref. 2) should include the following:a. Estimates of the intensity, depth, and arealdistribution of precipitation causing the runoff for eachhistorical storm (and rate of snowmelt. where this issignificant). Time distributions of storm precipitationare generally based on recording rainfall gages. Total'Strcamflow hydrographs (of major floods) are available inpublications by the US. Geological Survey. National WeatherService, State agencies, and other public Sources.precipitation measurements are usua~ly distributed, intime, using precipitation recorders. Areal distributions ofprecipitation. for each time increment, are generallybased on a weighting procedure in which tihe incrementalprecipitation over a particular drainage area is computedas tile sum of tihe corresponding incrementalprecipitation for each precipitation gage where cacchvalue is separately weighted by the percL1ntage of thedrainage area considered to be represented by the raingage.b. The determination of base flow as the timedistribution( of the difference between gross runoff arndnet runoff.c. Computation of distributed (in time)differences between precipitation and net direct runoff.the difference being considered herein as initial andinflitrafion losses.d. The determination of the combined effect ofdrainage area. channel characteristics, and reservoirs onthe runoff regimen, herein referred to as the "'runoffmodel." (Channel and reservoir effects are discussedseparately in Section A.8.)A.5 PRECIPITATION LOSSES AND BASE FLOWDetermination of the absorption capability of thebasin should consider antecedent and initial conditionsand infiltration during each storm considered.Antecedent precipitation conditions affect precipitationlosses and base flow. These assumptions should beverified by studies in the region or by detailedstorm-runoff studies. Tile fundamental hydrologicfactors should be derived by analyzing observedhydrographs of streamflow and related stormis. Athorough study is essential to determine basincharacteristics and meteorological influences affectingrunoff from a specific basin. Additional discussion andprocedures for analyses are contained in variouspublications such as Reference 2. The followingdiscussion briefly describes the considerations to betaken into account in determining the minimum lossesapplicable to the PMF:a. Experience indicates the capacity of a given soiland its cover to absorb rainfall applied continuously atan excessive rate may rapidly decrease until a fairlydefinite minimum rate of infiltration is rcached. usuallywithin a period of a few hours. Infiltration relationshipsare defined as direct precipitation losses such that theaccumulated difference between incrementalprecipitation and incremental infiltration equals thevolume of net direct runoff. The infiltration lossrelationships may include initial conditions directly, ormay require separate determinations of initial losses. Theorder of decrease in infiltration capacity and theminimum rate attained are primarily dependent uponthe vegetative or other cover, the size of soil poreswithin the zone of aeration, and the conditions alfectingthe rate of removal f" capillary water from the zone ofaeration. The infiltration theory, with certainapproximations, offers a practical means of estimating1.59.7 the volume of surface runoll fronm intense rainlfall.However. in applying tile method to natural drainagebasins, tile following factors must be considered:(I) Since the infiltration capacity of a givensoil at the beginning of a storm is related to antecedentfield moisture and the physical condition ofthe soil. theinfiltration capacity for the same soil may varyappreciably from storm to storm.(.2) The infiltration capacity of' a soil isnormally highest at the beginning of rainfall, and sincerainfall frequently begins at relatively moderate rates, asubstantial period of time may elapse before the rainfallintensity exceeds the prevailing infiltralion capacily. It isgnerally accepted that a fairly definite quantity ofwaler loss is required to satisfv initial soil moislturedeficiencies before nnoff will occur, the amount ofinitial loss depending upon antecedent conditions.(3) Rainfall does not normally cover the entiredrainage basin during all periods of* precipitation withintensities exceeding infillration capacities. Futhermore.soils and infiltration capacities vary throughout adrainage basin. Therefore, a rational application of anyloss.rate technique must consider varying rainfallintensities in various portions of the basin in order tode te rmine tile area covered by effectiverunolf-producing rainfall.b. Initial loss is defined as thie maximnum amountof precipitation that can occur without producingrunoff. Initial loss values may range from a minimumvalue of a few tenths of an inch during relatively wetseasons to several inches during dry summer and fallmonths. Tile initial loss conditions conducive to majorfloods usually range from about 0.2 to 0.5 inch and arerelatively small in comparison with the flood runoffvolume. Consequently. in estimating loss rates from datafor major floods, allowances for initial losses may beestimated approximately without introducing importanterrors in the results.c. Base flow is defined herein as that portion of aflood hydrograph which represents antecedent runoffcondition and that portion of the storm precipitationwhich infiltrates the ground surface and moves eitherlaterally toward stream channels, or which percolatesinto the ground, becomes groundwater, and is dischargedinto stream channels (sometimes referred to as bankflow). The storm precipitation, reduced by surfacelosses, is then resolved into the two runoff components:direct runoff and base flow. Many techniques exist forestimating thie base flow component. It is generallyassumed that base flow conditions which could existduring a PMF are conservatively high. the rationale beingthat a storm producing relatively high runoff couldmeteorologically occur over most watersheds about aweek earlier than that capable of producing a PMF. Oneassumption sometimes made for relatively large basins isthat a flood about half as severe as a PMF can occurthree to five days earlier. Another method for evaluatingbase flow relates historical floods to their correspondingbase flow. The base flow analyies of historical floods.there" fore, may he readily utilized in PMFdeterminations.A.6 RUNOFF MODELThe hydrologic response characteristics of thewatershed to precipitation (such as unit hydrographs)should be determined and verified from historical floodsor by conservative synthetic procedures. The modelshould include consideration of nonlinear runoffresponse due to high rainfall intensities or unexplainablefactors. In conjunction with data and analyses discussedabove, a runoff model should be developed, where dataare available, by analytically "reconstituting" historicalfloods to substantiate its use for estimating a PMF. Theraiitfall-runofft lime-areal distribution of historical floodsshould be used to verify that tile "reconstituted"hydrographs correspond reasonably well with floodhydrographs actually recorded at selected gaging stationskRef. 2). In most cases. reconstil ut ion studies should hemade with respect to two or more floods and possibly attwo or more key locations, particularly where possibleerrors in the determinations could have a serious impacton decisions required in the use of* the runoff model forthe PMF. In some cases, the lack of sufficient time andareal precipitation definition, or unexplained causes.have not allowed development of' reliable predictiverunoff models, and a conservative PMF model should beassured by other means such as conservatively developedsynthetic unit hydrographs. Basin runoff' models for aPMF determination should provide a conservativeestimate of the runoff that could be expected during thelife of the nuclear power plant. The basic analyses usedin deriving thie runoff model are not rigorous, but maybe conservatively undertaken by considering the rate ofrunoff from a unit rainfall (and snowmelt. if pertincnt)of some unit duration and specific time-ae.raldistribution (called a unit hydrograph). The applicabilityof a unit hydrograph. or other technique, for use incomputing the runoff from an e..'uiiated probablemaximum rainfall over a basin may be partially verifiedby reproducing observed major flood hydrographs. Anestimated unit hydrograph is first applied to estimatedhistorical rainfall-excess values to obtain a hypotheticalrunoff hydrograph for comparison with the observedrunoff hydrograph (exclusive of base flow-net ninoff),and the loss rate, the unit hydrograph. or both. aresubsequently adjusted to provide accurate verification. Astudy of the runoff response of a large number of basinsfor several historical floods in which a variety of valleystorage characteristics, basin configurations,topographical features, and meteorological conditionsare represented provides the basis for estimating therelative effects of predominating influenm-i for use inPMF analyses. In detailed hydrological studies, each ofthe following procedures may be used to advantage:a. Analysis of rainfall-runoff records for majorstorms;b. Computation of synthetic runoff responsemodels by (I) direct analogy with basins of similarcharacteristics and/or (2) indirect analogy with a largenumber of other basins through the application ofempirical relationships. In basins for which historicalstreamflow and/or storm data are unavailable, synthetici .59.9

4 techniques are the only known means for estimatinghydrologic response characteristics. However, care mustbe taken ito assure that a synthetic model conse.rvativelyreflects tile runoff response expected froin precipitationas severe as thie estimated PMP.Detailed flood hydrograph analysis techniques andstudies fkor specific basins are available from manyagencies. Published studies such as those by tile Corps ofEngineers, Bureau of Reclamation. and SoilConservation Service may be utilized directly where itcan be demonstrated that they are of a level of' qualitycomparable with that indicated herein. In particular, theCorps of Engineers have developed analysis techniques(Rfs. 2, 3) and have accomplished a large number ofstudies in connection with their water resourcesdevelopment activities.Computerized runoff models (Ref. 3) offer anextremely efficient tool for estimating PMF runoff ratesand for evaluating tihe sensitivity of PMF estimates topossible variations in parameters. Such techniques havebeen used successfully in making detailed floodestimates.Snowmelt may be a substantial runoff componentfor both historical floods and the PMF. In cases where itis necessary to provide for snowmelt in the runoffmodel, additional hydrometeorological parameters must.be incorporated. The primary parameters are the depthof assumed existing snowpack. the areal distribution ofassumed existing snowpack ( and in basins with distinctchanges in elevation, the areal distribution of snowpackwith respect to elevation), the snowpack temperatureand density distributions, the moisture content of thesnowpack. the type of soil or rock surface and cover ofthe snowpack, the type of soil or rock surface and coverin different portions of the basin, and the time andelevation distribution of air temperatures and heat inputduring the storm and subsequent runoff period.Techniques that have been developed to reconstitutehistorical snowmelt floods may be used in bothhistorical flood hydrograph analysis and PMF (Ref. 4)determinations.A.7 PROBABLE MAXIMUMPRECIPITATION ESTIMATESProbable maximum precipitation (PMP) estimatesare the time and areal precipitation distributionscompatible with the definition of Section A.2 and arebased on detailed comprehensive meteorological analysesof severe storms of record. The analysis usesprecipitation data and synoptic situations of majorstorms of record in a region surrounding the basin understudy in order to determine characteristic combinationsof meteorological conditions that result in various.rainfall patterns and depth-area-duration relations. Onthe basis of an analysis of airmass properties andsynoptic situations prevailing during the record storms,estimates are made of tile amount of increase in rainfallquantities that would have resulted if condilions duringthe actual storm had been as critical as those consideredprobable of occurrence in tile region. Consideralion isgiven to the modifications in meteorological conditionsthat would have been required IOr each of" the recordstorms to have occurred over the drainage haisin understudy. considering topographical features and locationsof the respective areas involved.The physical linimiations in meteorologicalmechanisms the maximum depth. time. and spacedistribution of precipitation over a basin are I )humidity (precipitable water) in tile air flow over thewatershed. (2) the rate at which wind may carty lhiehumid air into tile basin. :ind (3) tile fraction of tileinflowing atmospheric water vapor that can beprecipitated. Each of these limitations is handleddifferently to estimate tile probable miaximumprecipitation over a basin, and is modified further forregions where topography causes marked orographiccontrol (designated as the orographic model) as opposedto the general model (with little topographic effect}) 0precipitation. Further details on the models andacceptable procedures ate contained in References 5and 6.a. The PNIP in regions of limited t opographicinfluence (mostly convergence precipitation) may heestimated by maximizing observed intense stormpatterns in thie site region for various durations.intensities, and depth-area relations and transposingthem to basins of interest. The increase in rainfallquantities that might have resulte! from maximizingmeteorological conditions during the rtcord storm andtile adjustments necessary to transpose the respectivestorms to the basin under study should be taken intoaccount. The maximum storm should represent tli.. mostcritical rainfall depth-area-duration relation for theparticular drainage area during various seasons o" itheyear (Refs. 7. 8. 9, 10). In practice. the parametersconsidered are (I) the representative storm dewpointadjusted to inflow moisture producing the maximumdewpoint (precipitable water), (2) seasonal variations inparameters. (3) the temperature contrast. (4) thiegeographical relocation, and (5) thie depth-areadistribution. Examples of these analyses are explainedand utilized in a number of published reports (Refs. 7.8.9. 10).This procedure, supported with an appropriateanalysis. is usually satisfactory where a sufficientnumber of historical intense storms have beenmaximized and transported to the basin and where atleast one of them contains a convergent wind"mechanism" very near the maximum that nature can beexpected to produce in the region (which is generally thecase in the United States east of the Rocky Mountains).A general principle for PMP estimates is: The numherand seperily of JnaximiyathiV steps must balance iheadequacy of the storm sample, additional inaximizatioun1.59-9

• .. .steps are required in regions of more limiteid stormsanmples.b. PMI1 determinations in regions of orograplhitinfluences generally are for hlie high mountain regionsthat lie in the path of Ithe prevailing moist wind.Additional maximization steps front paragraph A.77.a.above are required in the use of the orographic model(Refs. 5, 6). The orographic moxlel is developed for theorographic component of precipitation where severeprecipitation is expected it) be caused largely by tirelifting imparted to fie ait by' mounwains. This orographicinfluence gives a basis for a wind model with maximizedinflow. Assuming laminar %low of air over any particularmountain cross section. one can calctlate Ihe liife" ofthe air. the levels at which raindrops and snowflakes areformed. and their drift with the air before they strikelhe ground. Such mnodels are verified by reproducing theprecipitation'in observed storms and are then used forestimating PIMP by introducing maximum values ofmtoisture and wind as inllow at thie foot of thiemountains. Maximum moisture is evaluated just as innonorogiaphic regions. In mnotntainous regions, wherestorms cannot readily be transposed (paragraph A.7.a.above) because of !heir intimate relation to theimmnediate tuderlying topography. historical stornits areresolved into their convective and orographiccompnecnts and maximnized as follows: (I) mraximuimmoisture is assunied. (2) maxinmum winds are assumed.and finally (3) maximum values of tIle orographicconsponent and convective component (convective as innonorographic areas'l of precipitation are considered tooccur simultanretously. Some of the published reportsthat ill ustr:ute the combination of orographic andconvective components. including seasonal variation, areReferences II. 12, and 13.In somne large watersheds. major floods ate often theresult of melting snowpack or of snownilt combinedwith rain. Acco:dingly. the probable maxinmumprecipitation (rainfall) and maximunt associatedrunoff-producing snowpacks are both estimated on aseasonal and elevation basis. The probable maximumseasonal snowpack water equivalent should bedetermined by study of accumulations on localwatersheds from historical records of the region.Several methods of estimating the upper limit ofultimnate snowpack and rueling are summarized inReferences 4 and 5. The methods have been applied inthe Columbia River basin, the Yukon basin in Alaska.the tipper Missouri River basin, and the upper Mississippiin Minnesota and are described in a number of reports ofthe Corps of Engineers. In many internmediate-latitudebasins, the greatest flood will likely result from acombination of critical snowpack (water equivalent) andPMP. Thie seasonal variation in both optimum snowdepth (i.e., the greatest water equivalent inl thesnowpack) and the associated PMP combination shouldbe meteorologically compatible. Temperature and windsassociated with PMP are two important snowmelt factorsamenable to generalization for snowinell computations(Ref. 14). The meteorological (e.g., wind, temperature,dewpoints) sequences prior to, during, and after thepostulated PMP-producing storm should be compatiblewith the sequential occurrence of the PMIP, The usershould place the PNIP over the basin and adjust thesequence of olher parameters to give the most criticalrunof flor t(ie season considered.The meteorological parameters for snownielcomIpu tations associated with PNIP are discussed in moredetail in References II 12, and 14.Other items that need to be considered indetermining basin melh are optimntum depth. areal extent.and type of snowpack. and other snowmuell factors (seeSection A.8). all of which must he compatible with themost critical arrangement of the PMP and associatednueiiorological paramneters.Critical piobable maxiniuni storm estimates for verylarge drainage areas are determined as above, but maydiffer somewhat in flood-producing storm rainfall fromthose encountered in preparing similar estimates forsmall basins. As a general rule. the critical PMP in a smallbasin results primarily from extremely intense small-areastorms; whereas in large basins the PMP usually resultsfrom a series of less intense, large-area storms. In verylarge river basins (about 100,000 square miles or larger)si.:h as the Ohio and Mississippi River basins, it may benecessary to develop hypothetical PMP storm sequences(one storm period followed by another) and stormtracks with an appropriate limte interval between storms.The type of meteorological analyses required and typicalexamples thereof are contained in References 9, 15, and1 6.The position of probable maximum rainfall centers.identified by "isolyetal patterns" (lines of constantrainfall depth), may have a very great effect on theregimen of runoff from a given volume of rainfall excess.particularly in large drainage basins in which a widerange of basin hydrologic runoff characteristics exist.Several trials may be necessary to determine the criticalposition of the hypothetical PMP storm pattern (Refs. 8.17) or the selected record storm pattern (Refs. 9, 16) todetermine the critical isohyetal pattern that producesthe inaxiumtm rate of runoff at thie designated site. Thismay be accomplished by superimposing an outline ofthe drainage basin (above the site) on the total-stormPMP isohyetal contour map in such a manner as to placethe largest rainfall quantities in a position that wouldresult in the maximum flood runoff (see Section A.8 onprobable maximuni flood runoff). Thi isohyetal patternshould be reasonably consistent with the assumptionsregarding the meteorological causes of the storm. A -considerable range in assumptions regarding rainfallpatterns (Ref. 11) and intensity variations can be madein developing PMP storm criteria for relatively smallbasins, without being inconsistent with meteorological1.59-10

L,1 0.0causes. Drainage basins less than a tew thousand squaremiles in area (particularly if only one unit hydrograph isavailable) may be expressed as average depth over tiledrainage area. However. in deoerntining the BilP patternfor large drainage basins (with varing basin hydrologiccharacteristics, including reservoir etfects). runoffestimates are required for different storm patternlocations and orientations to ohtain the final PMF.Where historical rainfall patterns are not used for PMP,two other methods are generally employed as follows:a. Average depth over the entire basin is based onlthe maximized areal distribution of Ihe PMP.h. A hypothetical isohyclal pattern is assumed.Studies of areal rainfall distribution from intense stormsindicate elliptical patterns may be assumed asrepresentative of such events. Examples are the typicalpatterns presented in References 8. 14. 17. and 18.To compute a flood hydrograph from the probablemaximum storm, it is necessary to specify the timesequence of precipitalion in a feasible and criticalmeteorological time sequence. Two meteorologicalfactors must be considered in devising the timesequences: ( I ) the time sequence in observed storms and(2) the manner of deriving the PMP estimates. The firstimposes little limitations: the lhetographs (rainfall timesequences) for observed storms are quite varied. There issome tendency for the two or three time incrementswith thie highest rainfall in a storm to bunch together. assonie time is rcouired for the influence of a severeprecipitation-producing weather situation to pass a givenregion. The second consideration uses meteorologicalparameters developed from PMP estimates.An example of 6-hour increments for obtaining acritical 24-hour PMP sequence would be that the mostsevere 6-hour increments should be adjacent to eachother in time (Ref. 17). In this arrangement the secondhighest increment should bc adjacent to the highest. thethird highest should be immediately before or after this12-hour sequence. and the fourth highest should bebefore or after the 18-hour sequence. This proceduremay also be used in the distribution of the lesser second(24-48 hours) and third (48-72 hours) 24-hour periods.These arrangements are permissible because separatebursts of precipitation could have occurred within each24-hour period (Reference 7). The three 24-hourprecipitation periods are interchangeable. Otherarrangements that fulfill the sequential requirementswould be equally reasonable. The hyclograph. orprecipitation time sequence. selected should be the mostsevere reasonably possible that would produce criticalrunoff at the project location based on tihe generalappraisal of the hydrometeorologic conditions in theproject basin. Examples of PMP time sequences fulfillingthe sequential requirements are illustrated in ReferencesI1, 12. and 17. For small areas. maximized local recordsshould be considered to assure that the PMP timesequence selected is severe.The Corps of Engineers arnd theHydrometeorological Branch of NOAA (under acooperative arrane tientI since 19)39)) have madecor n prchlenrsive inet corological studies of extremnoflood-producing storms ( Ref. I ) and have developed antuimbe r o(f estimates of "probahle maximunmprecipilation." The PMP estimates arc presented invarious unpublished mnemoranda and published reports.The series of' published reports is listed on the lyv sheetof referenced Hydronietcorological Reports such asReference I8. The published memoranda reports mtay heobtained from thi e Corps of i Engineers orHyJrometeorological Branch. NOAA. These reports andmemoranda present pgneral techniques: included amongthe reports are several that contain "generalized"estimates of PM I' for different river basins. Thegeneralized studies (Refs. 7. 12) usually assure reliableand consistent estimates for various locatlions in theregion for which they have been developed inasniuch asthey 'are based on coordinated studies of all availabledata. supplemented by thorough meteorologicalanalyses. In sonic cases. however, additional detailedanalyses are needed for specific river basins (Refs. 7. 8)to take into account unusually large areas. storm series,topography, or orientation of drainage basins not fullyreflected in the generalized estimates. In many riverbasins available studies may be utilized to obtain thePMP without the in-depth analysis herein or in tihereferenced reports.A.8 CHANNEL AND RESERVOIR ROUTINGChannel and reservoir routing of floods is generallyan integral part of the runoff model for subdividedbasins, and care should be taken to assure not only thatthe characteristics determined represent historicalconditions (which may be verified by reconstitutinghistorical floods) but ;dso that they would conservativelyrepresent conditions to be expected during a PMF.Channel and reservoir routing methods of manytypes have been developed to model the progressivedownstream translation of flood waves. Tihe sametheoretical relationships hold for both channel andreservoir routing. However, in the case of flood wavetranslation through reservoirs, simplified procedureshave been developed that are generally not used forchannel routing because of the inability of suchsimplified methods to model frictional effects. Thesimplified channel routing procedures that have beendeveloped have been found useful in modeling historicalfloods, but particular care must be exercised in usingsuch models for severe hypothetical floods such as thePMF because the coefficients developed from analysis ofhistorical floods may not conservatively rellect floodwave translation for more severe events.Most of tihe older procedures were basicallyattempts to model unsteady-flow phenomena usingsimplifying approximations. The evolutiorn of computer1.59-1 I

use has allowed development ,,ofI analysis techniques thatpermit direct solution tit' basic 'Instead% flow equationsmlilizinig ntimerical analysis teclinitques adaptable to thedigital comptuter (Ref. I19). In addition. most of' theolder techniques have been adapted for computer use(Ref. 3).In all rout ing techniques. care must be ,:xercised inassurinig hat1 ijmiramet ers selectLed Jor model verificationare based on several hislorical floods (whenever possible)and that their applicationl Ith1 PMF will restilt inconserva.liVe est mates 1 l'h\ ata Cles. water levels.velocities, and ilIpacM torceI .Theoretical discussions of1the many methods availahle for such analyses arecontained in Refelences 2. 19). 20.- I .mnd 22.A.9 PMF HYDROGRAPH ESTIMATESPM F net runolf hydrograph estimates are made bhsequentially applying critically located and distributedPM P estinmt tes using the runoff timodel. conservativelylow%, estimates of prcipitalioti losses, and conservativelyhilh estimates (1' base Ilow z'nd antecedent reservoirlevels.lit PlMF determinationis it is cenerall v assumed thatshort-lerin reservoir flood control storage would bedepleted by possible antecedent floods. An exceptionwould be whet it cat be demonstrated that tileoccurrence oif a measonably seveie flood I say aboolu;one-h:alf ofl a P1I\) less than a week (usually a tinitnrtnioit' 3 to- 5 days prior :ii a lIFM c:nli be evacialetl frotilthe reservoir helfre tile artival otf a PMVF. However, it isunusual to use all antecedent storage level less thanone-halftile flood control storage available'Time applicatiomn (i P\MP in bhasins whose hydrologicfeatures vat fron llcation to location requires thedetenriiimatit, that thie estimated PM F hydrographrepresents the most critical centering of the PIMP stormwith respect to the site. ('are must be taken in basinswitlhi substantial headwater flood control storage toassure that maoire highly concentrated PMP over asmaller area dowistireant of' the reservoirs would notproduce a greater PNIF tIan a total basin storm that ispartially controlled. In siich cases more than oCe P['NIPrunoff analysis mayl he required. Usually. only a fewtrials oft a total basin l.NI' are required to determine themost critical centering.The antecedent snowpack and its contribution tothe PNIF are included when it is determined thatsnowrnell coilrihntions to thie flood Would produce aPNIF (see Section A.7). However. these typcs ofhypothetical floods are generally the controlling eventsonly in the far west and northern United States.Runoff hydrogruphs should be prepared at keyhydrologic hlcations (e.g.. strcanigages and dams) as wellas at the site of mnclear facilities. For all reservoirsitnv olvedt. in flvw. out hllow, and pool elevat ionhydrographs should be prepared.Many existing and proposed dams and oilier rivercontrol structures may niot be capaible of safely passingfloods as severe as a PMF. Tile capability of river controlstructures to safely pass a PMF and local coincidentwind.generated wave activity must be determined as partof' the PM F atnalysis. Where it is poissible that suchstructures imay nitot safely survive Iloods as severe as aPM F. tile \vtwrst such conidition withi resipect todownstream nuclear lpower plants is assuimied (hut shouldbe suhtsltanlialed hr analysis ohl lpsl eamn PNIF poi':litiallto be their failuore during a PMF. and the PM Fdetertminatiion should include the resuiltant effects. Thisanalysis: also requires that tihe consequncces otf lupsreamiidam failures on downtstreanm damis ( domtino effects) heconsidered.A.10 SEISMICALLY INDUCED FLOODSS.isinically induced bloods on streams and riversmay be caused hr landslides or dain failures. Where riverCoitrol structures are widely spaced, their arbitrarilyas.suilied indiciduwil total.l instantaneous failure andresul tinig downsttreailmi flotodl wave atltenuation (routing)mliar be showII to coTIns6lcite lbi) threat to nuclearfacilities. Where the relative size. location, and proximityof' dams !o ptentiial seismic generators indicate a threatto nuclear power plants. tite capability of suIch structures(cither singly or in combination) Ito resist severeearthquakes (critically located) shimald he considered. Iliriver basins where the flood a unoff season mayconstitute a significant portion of' the year (such as theMississippi. Columbia. or Ohio River basins). f'ull floodcontrol reservoirs willi ai 25-year flood is assuniedcoincident with the Safe Shutdown t..artliquake. Also.cotnsideration should he given to the occurrence of' aflood of approximately one-half the severity of a PM Fwith frill flood control reservoirs coincident wi\h themaximumi earthquake determined on the basis of'historic seismicity ito mainlain a consistent level ofanalysis I'or Other combinations of such events. As withfailures dime to inadequiate flood control capacity,domino and essentially simultaneous multiple f'ailuresmay also require consideration. If the arbitrarilyassumed total failure of the most critically located (froma hydrolh.:,ic standpoint ) struct ures indicates flood risks atthe nuclear power plant site more severe than a PMF, aprogessively more detailed analysis of the seismiccapability of the dam is warranted. Without benefit ofdetailed geologic and seisunic investigations. the floodpotential at the nuclear power plant site is next generallyevaluated assuming the most probable mechanistic-typefailure of' the quest ioned struci tires. IfI tile results of eachstep of the above analysis cannot be safelyacconmnodated at the nuclear power plant site in anacceptable manner, the seismic potential at tile site ofeach questioned structure is then evaluated in detail, thestructural capability is evaluated in the same depth as for-I1.59. 12

° nuclear power plant sites, and the resulting seismicallyinduced flood is routed to the site of the nuclear powerplant. This last detailed analysis is not generally requiredsince intermediate investigalions usually providesufficient conscrvalive inflormiation to allowdeterminalion of an adequate design basis flood.A.11 WATER LEVEL DETERMINATIONSAll the preceding discussion has been concernedprimarily with determinations of flow rates. The Ilowrate or discharge must be converted to water levelelevation for use in design. This may involvedetermination of' elevation-discharge relations Ifor naturalstream valleys or reservoir conditions. The reservoirelevation estimates involv,: the spillway dischargecapacity and peak reservoir level likely to be attaiiiedduring the PMF as governed by the inflow hydrograph.the reservoir level at the beginning of the 'M[:. and thereservoir regulation plan with respect to total releaseswhile the reservoir is rising to peak stage. Most riverwater level deterininations involve the assumption ofsteady, or nonvarying, flow for which standard methodsare used to estimate flood levels. Where little floodplaingeometry definition exists, a technique called"slope-area" may be employed wherein the assumptionsare made that the water surface is parallel to the averagebed slope, any available floodplain geometryinformation is typical of the river reach under study, andno upstream or downstream hydraulic controls affectthe river reach fronting the site under study. Where suchcomputations can be shown to indicate conservativelyhigh flood levels, they may be used. However, the usualmethod of estimating water surface profiles for floodconditions that may be characterized as involvingessentially steady flow is a technique called theItstandard-step method." This technique utilizes thlei- .grated differential equation of steady fluid motioncommonly referred to as the Bernoulli equation(References 22. 23, 24, and 25) where, depending onwhether supercritical or subcritical Rlow is tinder study,water levels in the direction of flow computation aredetermined by the trial and error balance of upstreamand downstream energy, respectively. Frictional andother types of head losses arc usually estimated in detailwith the use of characteristic loss equations whosecoefficients have been estimated from computationalreconstitution of historical floods, and from detailedfloodplain geometry information. Application of the"standard-step method" has been developed into verysophisticated computerized models such as the onedescribed in Reference 23. Theoretical discussions of thetechniques involved are presented in References 22, 24,and 25.Unsteady-flow models may also be used to estimatewater levels. Since steady flow may be consider,:d a class.of unsteady flow, such models may also be used for thesteady-flow water level estimaLion, Compnterizedunsteady-flow models require generally the samefloodplain georrit tv definition as steady-fiowv models.and thelrefore hit li use may allowv more accurate watersurface level t"'caini;ws whiiere approxinmatlions are inlle. ()n.e such iilwloidV-Iwcoriputier 1t1odel is dicused ill e 11).All ieas.omahly i,'cnr:ile wvacr h'ct, nlrdels reqmuire 11;1,lpl:1 &lfiminitiori l :11c.ts that cat1inatetialklv affect ticl levels. I.ood wa%( t .l;:iriom .and c:litihratlini lv by rnr:henirl~ical iecii.,-iwii ofhislorical (tit mte ,hcclioit of- c.1iblat:ioicocttficiellts based (it l the cil 'itsa,;li'c liallnIerl ofinformation derived torll SAilr 'lildies -I' oilier iv,.rreaches). Particular c:are s hould he cxercis-d it, asstiiethat corntrolling tlfomd lc.el est iniates tic tilwvayvsconservatively high.A.12 COINCIDENT WIND-WAVE ACTIVITYThe superposition tlt \n'd-wave :activitv on I'MF tirseismically induced wael! level dcte rnin ltions isrequired to assure that. in 11le event Cilt hr coildit ito didoccur, ambient nieteorological activityv would Inot causea loss of safe ty-related tun t iotn due to wav, act ion.The selection of' wind spejeds andtI critical winddirections assu.med coincident with mnxiiniini I'MI: orseismically i.'duced water levels should provide :t,,n; i rinccof virtually no risk to safety-reialed equipmientr icces.arnVto plant shutdowvn. The ('orps of' ngineecrs .uqiests(Refs. 26. 27) that average rmaximum %%-itnd siced% of'approximately 40 to (10 inph have occurred in miajorwindstorms in most regions of the United States. Forapplication to the safety analysis of nuclear facilities, theworst regional winds of record should le :ssnmnedcoincident with the PMF. However. the postuhlted windsshould be meteorologically compatible with theconditions that induced tire PMF or with tlie floodconditions assunred coincident with seismically induceddam failures) such as the season of tfie year. the ntiterequired for the PMP storon to 11r0%'e our of the area andbe replaced by meteorological conditions that couldproduce the postulated winds, ard the restrictions onwind speed and direction produced by topography. Asan alternative to a detailed study of hitorical regionalwinds, a sustained 40-inph overland wind speed t'romrany. critical direction is an acceptable positulation.Wind-generated set up (or wind tide) atd waveaction (runup and impact torces) may be estimated usingthe techniques described in References 26 and 28. Tiremethod for estimating wave action is based on stutisticalanalyses of a wave spectrum. For nuclear power planrts.protection against the maximuin wave, defincd inRefernce 28 as tire average of tire upper one percent ofl"the waves in the anticipated wave spectrumI , should bIeassumed. Where depths of water ill tronit r0'safety-related structures are sufficient (Cusually aboutseven-tenths the wave height), the wave-induiced forceswill be equal to the hydrostatic forces estimated frort1.59-13 the maxilunm rurup level. Where the waves can be-tripped' and caused to break both before reaching andon safeiy.related structures, dynamic Irces may. beestimated from Reference 28. Where waves may inducesurging in intake structure sumps. pressures on walls andthe underside of' exposed floors should be considered,particularly where such sumps are not vented and airColmpression call greatly increase dynamic forces..In addition, assurance should be provided thatsafety systems ncessary for cold shutdown andmaintenance thereof are designed to withstand the staticand dynamic effects resulting from frequent flood levelscoincident with the waves that would be produced bythe nmaximumn gradient wind for the site (based on astudy of historical regional meteorology).1.59.14I

V64 REFERENCESI. Precipitation station data and unpublished recordsof Federal, State, municipal, and other agencies maybe obtained from the U.S. Weather Bureau (nowcalled National Weather Service). In addition,studies of some large storms are available in the"Storm Rainfall in the Un it ed States.Depth.Area-Duration Data." summaries publishedby Corps of Engineers, U.S. Army.2. Corps of Engineers publications, such as EM1110-2-1405 dated 31 August 1959 and entitled,"Engineering and Design-Flood HydrographAnalyses and Computations." provide excellentcriteria for the necessary flood hydrograph analyses.(Copies are for sale by Superintendent ofDocuments. U.S. Government Printing Office,Washington, D.C. 20402.) Isohyetal patterns andrelated precipitation data are in the files of theChief of Engineering, Corps of Engineers.3. Two computerized models arc "Flood HydrographPackage. HEC-I Generalized Computer Program,"available from the Corps of Engineers HydrologicEngineering Center, Sacramento, California, datedOctober 1970 and "Hydrocomp SimulationProgramming-HSP," Hydrocomp Intl.. Stanford,Calif.4. One technique for the analysis of snowmelt iscontained in Corps of Engineers EM 1100-2.406,"Engineering and Design-Runoff From Snowmelt,"January 5, 1960. Included in this reference is alsoan explanation of the derivation of probablemaximum and standard project snowmelt floods.5. "Technical Note No. 98-Estimation of MaximumFloods," WMO-No. 233.TP.126, WorldMeteorological Organization, United Nations, 1969and "Manual for Depth-Area-Duration Analysis ofStorm Precipitation," WMO-No. 237.TP.129, WorldMeteorological Organization, United Nations, 1969.6. "Meteorological Estimation of ExtremePrecipitation for Spillway Design Floods", Tech.Memo WBTM HYDRO-5. U.S. Weather Bureau(now NOAA) Office of Hydrology. 1967.7. "Seasonal Variation of the Probable MaximumPrecipitation East of the 105th Meridian for Areasfrom 10 to 1,000 Square Miles and Durations of 6,12, 24, and 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br />," Hydromneteorological ReportNo. 33, U.S. Weather Bureau (now NOAA), 1956.8. "Probable Maximum Precipitation. SusquehannaRiver Drainage Above Harrisburg, Pa.,"Hydrometeorological Report No. 40. U.S. WeatherBureau (now NOAA), 1965.9. "Meteorology of Flood Producing Storms in theOhio River Basin," Hydronieteorological ReportNo. 38. U.S. Weather Bureau (now NOAA). 196L.10. "Probable Maximum and TVA Precipitation Overthe Tennessee River Basin Above Chltllanooea."Hydrometeorological Report No. 43, U.S. WeatherBureau (now NOAA), 1965.11. "Interim Report- -Probable Maximum Precipitationin California." Hydrometeorological Report No. 36.U.S. Weather Bureau (now NOAA). 1961.12. "Probable Maximuni Precipitation, NorthwestStates," Hydrometeorological Report No. 43. U.S.Weather Bureau (now NOAA), 1966.13. "Probable Maximum Precipitation in the HawaiianIslands," Hydrometeorological Report No. 39. U.S.Weather Bureau (now NOAA). 19)63.14. "Meteorological Conditions for the ProbableMaximum Flood on the Yukon River AboveRampart, Alaska," Hydronieteorological Report No.42, U.S. Weather Bureau (now NOAA), 1966.15. "Meteorology of Flood-Producing Storms in theMississippi River Basin." HydrometeorologicalReport No. 34, U.S. Weather Bureau (now NOAA).1965.16. "Meteorology of Hypothetical Flood Sequences inthe Mississippi River Basin," HydrometeorologicalReport No. 35, U.S. Weather Bureau (now NOAA),1959.17. "Engineering and Design-Standard Project FloodDeterminations," Corps of Engineers EM1110.2-1411, March 1965, originally published asCivil Engineer Bulletin No. 52-8.26 March 1952.18. "Probable Maximum Precipitation Over SouthPlatte River, Colorado. and Minnesota River.Minnesota," Hydrometeorological Report No. 44.U.S. Weather Bureau (now NOAA). 1961).19. "Unsteady Flow Simulation in Rivers andReservoirs," by J. M. Garrison. J. P. Granju and J.T. Price. pp 1559-1576, Vol. 95. No. IIYS,(September 1969), Journal of the Ilyt'draulicsDivision. ASCE. (paper 6771).20. "Handbook of Applied Hydrology." edited by VenTe Chou, McGraw.Hill. 9)64. Chapter 25.21. "Routing of Floods Through River Channels." EMH 10-2-1408. U.S. Army Corps of Engineers. IMarch 1960.1.59-15

.2. "'l~nLiti .'riig 1 yvdiauilics". e.'dited hy Hlu tier Rouse.John WViley & Sons. l1tc. 19Q50... 1 eW c Sil face Plroilies. HI.I-2 GenraliuedCo nipmiaUt Program.' available from( tie Corps of1:-ni neers Hydrologic Engineering Center.Sacrameilnito. C:ail._'4. "()pen Chalnel Ilydratlic'" by Ven Te Choli;-j "lack%:%tlctr (Cirv es in River (Channels." EMI I 1 40-).I4. U.S. Ariny Corps of Elpgineeis.Dc),. a',:. cr "7. 2o. "Compiitation of Freeboard Allowances ,fr Wavesin Reservoirs." I-ngineca Technic;al Leiter lTLI1 10-2-). U.S. Army Corps of lingineers. I Augist27. "Policies a nd Proceedures PerIaining toD)etermination of Spillway ('apaci ties anid FrecehoardAllowances for D)ams.'" lingincer Circular 1-C1110-2-27. LU.S. Arwy Corps or Engineers. I August28. "iShore Protect iot. and I)esign, TedhnicilRelp)rt No. 4. U.S. Arauy "Coastal ElngineeringResearch Cenler. 3rd edition. I906.1.59-16