Flood Hydrology Manual
A Water Technical Resources Publication

Arthur G. Cudworth, Jr. Surface Water Branch Earth Sciences Division FIRST EDITION 1989

United States Department
Bureau Denver of Reclamation Office

of the Interior

As the Nation’s principal conservation agency, the Department of the Interior has responsibility for most of our nationally owned public lands and natural resources. This includes fostering the wisest use of our land and water resources, protecting our fish and wildlife, preserving the environmental and cultural values of our national parks and historical places, and providing for the enjoyment of life through outdoor recreation. The Department assesses our energy and mineral resources and works to assure that their development is in the best interests of all our people. The Department also has a major responsibility for American Indian reservation communities and for people who live in Island Territories under U.S. Administration.

UNITED STATES GOVERNMENT PRINTING OFFICE DENVER: 1992 For saleby the Superintendent Documents,U.S. Government Printing Office, of Washington,DC 20402, and the Bureau of Reclamation, Denver Offie, Denver FederalCenter, P.O. Box 25007, Denver, Colorado 80225, Attention: D-7923A 024-003-00170-8.

The need for a comprehensive Flood Hydrology Manual for the Bureau of Reclamation has become apparent as increasing numbers of civil engineers are called upon to conduct flood hydrology studies for new and existing Bureau dams. This is particularly important because of the increasing emphasis that has been placed on dam safety nationwide. In general, these engineers possess varying backgrounds and levels of experience in the specialty area of flood hydrology. Accordingly, the primary purpose of this manual is to provide the necessary background, relationships, criteria, and procedures to allow the engineer to conduct satisfactory flood hydrology studies. As a result, these studies should reflect greater consistency and reliability of results for most of the drainage basins encountered in Bureau projects as well as those for other water resource construction agencies. These relationships, criteria, and procedures are based on detailed analyses of hydrologic and meteorologic data and studies of observed flood and severe rainfall events that have accumulated over the years. The information contained in this manual reflects the methodologies currently used by the Bureau in performing flood hydrology studies. These methodologies have been proven to provide satisfactory results for use in the planning, design, construction, and operation of the Bureau’s water control facilities. However, it would be inappropriate to infer that these methods are the final solutions to hydrologic problems because other methods are constantly evolving. It can reasonably be expected that, as additional data are collected and as new and more advanced techniques emerge, these methodologies will be improved upon and modified with the passage of time. It is also expected that such modifications and improvements will be fully incorporated into future editions of this manual. The Bureau of Reclamation is primarily responsible for the development of water resources in the 17 Western States to meet agriculture, municipal and industrial, power, recreation, and environmental water supply requirements. In addition, with the recent emphasis on dam safety, the Bureau has been assigned the responsibility for safety of dams studies for Department of the Interior dams throughout the nation. The Bureau has no authority, as legislated by Congress, in the area of flood control except for a few unique, specifically authorized projects. Flood control is, and has been for the past 50 years, within the purview of the United States Army Corps of Engineers. Therefore, the flood hydrology discipline, as applied to Bureau activities, deals with two prin-

cipal areas: (1) determination of the upper limit or probable maximum



flood potential at a damsite so that dams whose failure would result in loss of human life or widespread property damage can be designed to safely accommodate this flood without failure, and (2) determination of more commonly occurring floods for use in the design of diversion dams; very low hazard storage dams; cross drainage facilities for the extensive canals, aqueducts, and roads associated with the delivery of water to users; and for diversion of flood waters that may occur during the construction of dams. It should be noted that the technical hydrologic procedures used in these two principal areas are essentially the same as those used in flood hydrology studies that support the planning, design, construction, and operation of flood control projects. Recognizing the Congressionally authorized function of the Bureau, this manual concentrates on three major technical aspects of flood hydrology: (1) hydrometeorology related to probable maximum precipitation determinations, (2) probable maximum flood hydrograph determinations, and (3) statistics and probabilities relating to the magnitude and frequency of flood flows. Other important, but essentially nontechnical, aspects related to flood hydrology studies are also treated. Chapter 1, “Background and Historical Perspective,” provides a brief historical perspective of the flood hydrology discipline. Included in this discussion is a general overview of procedures and philosophies that have evolved and been applied over the years, from those used in the development of early water control works to the present. Chapter 2, “Basic Hydrologic and Meteorologic Data,” provides a discussion of the basic hydrologic and meteorologic data that are available for analysis in the development of criteria for upper limit or probable maximum precipitation estimates, methods of determining rainfall-runoff relationships used in deriving probable maximum flood hydrographs, and in generating flood probability estimates. The methods and equipment currently used for collecting these data are briefly discussed to provide the reader with a general background on the mechanics involved. The important consideration of the reliability and accuracy of these data is included in subsequent chapters dealing with the specific uses of these data to provide the reader with an appreciation of the resulting accuracies of flood magnitude and frequency estimates. This chapter concludes with a more detailed description on assessing a drainage basin’s physical features as they impact the rainfall-runoff phenomena through field reconnaissances. Chapter 3, “Hydrometeorology,” begins with a general discussion on the theories underlying basic atmospheric processes. Emphasis is placed on those processes that are important in the development of regionalized or generalized criteria used in estimating probable maximum precipitation amounts that are, as shown in subsequent chapters, necessary in determining probable maximum flood hydrographs. This chapter then describes the procedures used in the analysis of observed events that lead


and the Soil Conservation Service. Corps of Engineers.” provides a comprehensive presentation of an extremely important area of flood hydrology. These procedures are needed when studying large basins that have been divided into subbasins or include one or more reservoirs. and discusses the importance of report preparation in documenting the basis and results of a flood hydrology study . and topography so that curves represent homegeneous hydrologic and meteorologic conditions are discussed in detail. .PREFACE to the generation of criteria used for developing estimates of probable maximum precipitation and storm levels throughout the conterminous United States. Chapter 5. season.” is the final chapter of the manual. Chapter 4. The chapter concludes with an example where a hydrograph representing the probable maximum flood event is generated for a relatively complex basin.” provides a discussion on the purpose and procedures used to develop envelope curves of experienced flood discharges. storm type.. The hydrologic cycle is discussed as it pertains to extreme runoff phenomena up to and including the probable maximum flood. Procedures for routing flood hydrographs through reservoirs are also included.” provides a discussion on procedures used by the Bureau in accomplishing the channel routing and combining of flood hydrographs. These curves provide a check on the adequacy of probable maximum flood hydrograph estimates. Examples of these distributions are provided and their use in flood probability analyses is discussed. Proper procedures for segregating the data by region. Discussion on both natural and developed basins is included because development has been shown to considerably modify the hydrograph resulting from a given amount of precipitation. Chapter 6. “Envelope Curves of Recorded Flood Discharges. Chapter 8. National Weather Service. both peak and volume. Considerable attention is focused on the uncertainties and relative reliabilities inherent in such relationships. Chapter 7. 111 . The process of infiltration losses into the soil and their effect on the runoff hydrograph is presented in more detail. “Flood Routings Through Reservoirs and River Channels. “Flood Hydrograph Determinations.” presents the theory and procedures currently used by the Bureau in converting rainfall over a drainage basin into a hydrograph of flood runoff. “Statistics and Probabilities. The chapter concludes with remarks on continuing Federal interagency activities leading to the development of hydrometeorological criteria by the Bureau of Reclamation. “Flood Study and Field Reconnaissance Reports. The chapter includes a discussion of the basic concepts encountered in the collection and statistical analysis of streamflow data with emphasis on theoretical frequency distributions. The chapter then discusses the current standard methodologies used by Federal water resources development agencies in determining discharge-probability relationships for basins in the United States.

The recommended report content prepared during of such a report the initial phase is also discussed. E. Manager of Planning Services. In addition. Preparation of the manual for publication was supervised Ronald D. under the general direction of Ronald Head. and John Dooley. Assistant Commissioner for Resources Management. This manual was written by Arthur G. by the Editor. Galvan for their patience.. in bringing the preparation of this manual to a successful conclusion. Leon Hyatt. The authors are indebted to David L. There are occasional references to proprietary materials or products in this publication. Earth Sciences Division. perseverance. for their critical review of the manuscript and their valuable comments and suggestions. Lane who were responsible for writing chapters 3 and 7. iv . McGregor. (retired) former Head. These references are not to be construed in any way as an endorsement because the Bureau does not endorse proprietary products or processes of manufacturers or the services of commercial firms. respectively. and untiring effort in typing the many draft versions and revisions. with considerable assistance from Louis C. Bullard. Lanky. and Terry P. Jr. The authors are also grateful for the administrative support and direction of M. Kenneth L. Mohr. we would be remiss if we did not extend our sincere appreciation to Diane C. The Bureau of Reclamation expresses appreciation to the organizations who have permitted the use of their material in this text. Chief. all of the Denver Office. at the Bureau’s Denver Office. Lynott. Nielsen and Monica A. Robert K. Surface Water Branch. Schreiner and William L. Cudworth. Sveum. Flood Section. Planning and Editing Section.FLOOD HYDROLOGY MANUAL as well as the field reconnaissance of the study.

.............. . (a) Recorded streamflow data . .. ... . .. .. Derivation of PMS ... . . ........... ... . ...... ... .. ... ........... ...... ................. ............ .......... .. .. . ....... .... . .......2 3.... BASIC HYDROLOGIC DATA AND METEOROLOGIC 2. .............. .... . ... ................ ..... .... ........... ... .... (b) Supplemental meteorologic data acquisition or bucket surveys ... ........ ... . ..... . ....... .... ... ..... .... ........... .5 General considerations. . ...... ..... ..... . .... ... ............ .. ... .. ... ...... .... .. .. ....... ..... ................. ........ .... . ..... ... ........ ... ........ . ....... .... .... .. ..................... .. .. ...... . ... ..... . . ..... . . . . .. ... ...... ...... .. ... . .... ..CONTENTS Section Page 1 .. ... ....... .... ................. ....3 General considerations and background .. .... . .. .. . ......... ............ .. .. ... . ............... . (b) Pecos River . .... . ....... 1..... ...... . . . ... .. ....... ...... ............ ... .. (a) General . .... .... ... ............... ..... ...3 2......... Examining nearby basins that have experienced significant recorded floods. ...... .... . ... . ...... .... ..... .. ..... ........ ... .... . .. .. (c) Condensation ..... . ..... .. . Hydrologic data ..... .. (b) Participation in field reconnaissance...... .. . ... ..... CHAPTER 2... ...... ... . .1 2..... ..... . ....... ...... . .. .. ... .. . Meteorologic data ... ........... .. ...... . (b) Lifting mechanism . . .... . ... .. . (e) Ware River ........ ......... . . .. . . ..... ................. . ...... ....... .. ....... .... ........ . . . .. ...... .............. . ... .. ....... ...... .. ............. (d) Water droplet-ice crystal growth... ....... .. ..... 25 27 27 27 29 30 30 31 40 V ................ ... . ... ................ .... ........ ... . .. . . (a) Storm maximization approach . ... . ...... ... .. . . .. .. HYDROMETEOROLOGY 11 11 11 14 16 16 16 20 21 21 21 24 3............... . . ......... ... 7 8 CHAPTER 1. .. BACKGROUND AND HISTORICAL PERSPECTIVE 1.. ......... .. ... ... . ... .............. ... ......2 Historical procedures.... . . .. (c) Field observations ... .. . .. . . (c) Plum Creek . .... . .. .. ...... ... .. ....... . ............... . ....... CHAPTER 3.... .. ... . . (a) Systematic data acquisition.4 2.... . ...... ... ..... ..... ... ..... .... .. .......... .. . . . . .. ............. ...... . .. ..... .... ........ .. .. . ........ . 1.... .... . . ... . 7 . .. . . Atmospheric processes . . . ... ... ........ .. .. ... . .... (d) Sun River ... ... .. ... .1 3. ........ (b) Model technique ........... ....... . .. . . .. . ..... ... ..... ... . .. . (a) Atmospheric moisture. ..... . . ..... ... ... .. .. .... .... . .... ...... .1 Background.... . . (b) Indirect discharge measurements....... ......... . . ....... ........ .... .... ..... ..... ...... . . ... .. ...... ............ . . ... ..... .......... ..... ........ . ... .... .. . zieLeeFissance of drainage basins for flood hydrology . ......... .. ....... . ...2 2.. .. ........ ... . (a) Arkansas River . .. ... .... . . ........ .. .. ... ...3 Hydrometeorologic approach. ... ...... .

. .. .... ...... . .. .... . .. ... ..... ...... . .. .. .... . .... ...... (b) Routing constants and routing procedures.... Foss Dam example 133 136 140 140 142 143 144 150 155 158 163 167 171 vi ...... . .... . ... . .. .. ... . .... . ...... . .... .... ....... . ..... . . .... ......... . .. .... ... ...... . . ... .. .... (a) Determination of K constant ... ..... .. . . (g) Computing synthetic unit hydrograph ordinates .. .... ... . .... .. ........ ... ....... . .. ..... .... ... . ... .. . ... .... .. . ...... . ... .. .. ..... .. . . ......... ... ...... . .. . .. Western Mountain snowmelt equation....... . . .3 5......... .. . . ... ........ .. ... .... . ... (i) Summary . . .. ... .. . . . ...... .. .. ... . . . ... .. .... Distributions of PMS . .....2 4..... ...... ..... .. . ...... ..... .. . . . .. . ... . .... .. .. .. . .... .. ...... .. (h) Observed flood hydrograph reconstitutions. (f) HMR 36 ..... . ... .... ..3 4.. . Foss Dam example 5. .. .... . . ........ . .. .... ..... . . . . (g) HMR 43 ......... .... and 53 ... ....... .. .... .. . ..... . ... ...... . .. .......... . . .... Antecedent floods... . .. . .. ... (e) Development of synthetic unit hydrographs ... ... (c) Unit hydrograph lag time..5 5.......... .... .. .. . .. .... FLOOD ROUTING RIVER THROUGH CHANNELS RESERVOIRS CHAPTER 5.. Reservoir flood routings Techniques for routing floods through river channels .... ... . . . (k) Base flow and interflow Probable maximum flood hydrograph .. . ..4 5. ..... .. .. .. .. .. . . . .... . .... .. ... .. .. . ... .. . ...... ... ... . . ... ....... .. .. .. .... . ... .... ... . . (b) Basic unit hydrograph concept . ........ ... .. ...... ........... ... ...... .. (d) HMR 51............ ....... (a) Temporal ... (c) Selection of reach lengths... ........ ... ..... ... . . .... . . .. .. .. .... .. ... . . .... . .... .... .. .. . ... .... . . .. .... ... ..... ...2 5... . .... . . . Muskingum Routing Method. ..... .. ............ .. . .. .4 4......... . .. .. .... .. ... . .. . .. .....4 4...... .. .... .. ............ .. . .. . . ... . . .... ... .................. .. ... (h) HMR 49 .. ..... ... (b) Determination of X constant. .. . ... ........... ....FLOOD HYDROLOGY MANUAL CONTENTS-Continued Section (c) Individual drainage estimates and regionalized studies.... . . .. .. . ..... .. .. . . .. Successive Average Lag Method .. ... . .... . ...... .. . .. (e) HMR 55A ... . .. . . . . .. ..... . . . ..... ...... . . ... .. ... ........ . .... .. (d) Temporal distribution of unit runoff......... .. ...... .. ...... . ......... . . ..... .... .. . (a) General ...... ...... ... .. ... . . . .. ... .... ... .. . .. . .... . ..8 Genera1 considerations. ..... ...... ... ... ............ .... ..... . ... ........ .... . . (i) Hydrograph reconstitution procedure.. .. . CHAPTER 4.. ... . ..... . ... .. . . . . .. .. ..7 5. (i) Infiltration and other losses. . . . .......... ...... ... ........... .. .... .. .. .. ... . .. .. ..... ...... ...... .... .. ......... FLOOD HYDROGRAPH DETERMINATIONS 63 63 63 67 69 71 87 87 104 105 109 112 114 115 116 118 AND Page 41 44 46 49 53 55 57 58 58 59 3......... .. . ... .1 5. . ... ..... .. . . .. .. . ... ..... . . . .... . .. . ..... .... . . . .. ... .. . .. . . . . ..... .......... .. . . . ... (a) Basis of method .. .. ..6 5.. 52. ... .. . ... ... ..... . ... ... .. . . .... . .... ...... . ..... .... .... ....... ... ... . .. ....... . . ... .. . .... .... ... .5 Flood runoff from rainfall .. ........ .. ..... .. . . .. .. . .. .. . .. .... .. . ... ... . ....... . ... . ..... ... ...... ..... .. .... .. .... ....... .. ... Modified Puls Method Modified Wilson Method .... . .. ............ .. ........... . . ....... .. . (f) Dimensionless unit hydrograph or S-graph selection ..... .. . .. .. ..... .. ... .... ... . ......... . .. ... ........1 4...... ... .. .... . .... .. .... ..... . . ... .. ...... ... .. . ........ .... .. .. . ..... . . . (b) Spatial . .. .. . ... .. .

............................ (d) Combined frequency curves........... (a) Transposition of frequency data ................................................................................... (a) Causative meteorological (b) Types of mixed populations.. ..... ...................................... 7....................3 Frequency distributions.................................................. ................................................ STATISTICS AND PROBABILITIES 7........................8 Volume analysis...................................5 Example application using Bulletin 17B............................ ............... factors...................................... ........ 183 185 185 186 187 188 189 189 191 194 196 197 198 198 199 200 201 201 202 203 203 204 204 205 205 206 206 206 207 209 210 210 210 211 212 212 vii .... 7................................................ 7..... (a) Concepts and philosophy ............................................................ (b) Regional flood frequency. CHAPTER 7.......9 Probability relationships for ungauged areas ................ ..................................................................10 General trends in frequency relationships.................................... Hydrologic data samples ........... ............ ................... ........... ....... ........................... (b) Data sources ............................................................................CONTENTS CONTENTS-Continued Section CHAPTER 6......... (d) Parameter estimation ............................................... ................................... (i) Adjustments................................. (c) Probability distributions........................................... (c) Separation of data.............. Procedure ........ ........ 7................................................................. (c) Shape ... ..................... (a) Balanced flood hydrograph....................................... (h) Outlier criteria........ .. .... ..1 6............. ..........................1 7.................................................. (a) Type of data ......................................................................................... 7..... .................................................................................................. ................. (c) Data analysis .................................................2 Introduction ........................ 7...3 General considerations ..................... (c) Frequency analysis from synthetic data...... (a) Adoption by Federal agencies...................................................2 6..................... (e) Disadvantages................. (b) Forms of presentation ............ (d) Accuracy of data.... ENVELOPE CURVES OF RECORDED DISCHARGES FLOOD 177 179 180 Page 6.......................... (g) Generalized skew ................................................................................................ 7................................................. (b) Fixed interval analysis.............................. (d) Advantages............... (f) Fitting procedure ...... .............................7 Mixed populations.. Sources of data ...................................................................... (b) Form .................. 7.......4 Log-Pearson Type III distribution....................................................................................................6 Limitations on frequency curve extrapolation ........................................

.... ... ... ........... ........ Specific contents of a flood study report .......... .... (b) Summary of results............. .. Vlll . . .......... 7....... ....... (e) Historical floods...... Utah............. ....... ....... Oregon. . ..... ..... (k) Antecedent flood ... ...... . ......... Unit hydrograph lag data for the Rocky Mountains... ............. .. ..... .... ... ... .. .... ... .. and Oregon .. ................. .......... .. . ... .. . ...... (j) Snowmelt runoff ..... .. . .. ......... .... ..................... .. .. (h) Infiltration loss rates........ .... ................................ .......... ... ... . . .... . ...... .... (a) Authority... . ..2 8... .. ....... New Mexico.... ..... ........ BIBLIOGRAPHY INDEX... 8... . .... ..... ......... Colorado. .... .. ....... .. ... Unit hydrograph lag data for the Coast and Cascade ranges in California......... .. ............ (g) Unit hydrograph......... .... ........ Unit hydrograph lag data for the Sierra Nevada in California .. ... . ........ ....... .... ......1 8............ ..... .’ ...................... .... ... .. . ..... ... ..... . . ... .. . ....... CHAPTER 8.MFlHpydrograph nve 0 e curves.............. ... .... FLOOD STUDY AND FIELD REPORTS RECONNAISSANCE Page 213 2 19 General .... (d) Description of drainage basin ..... .. .....FLOOD HYDROLOGY MANUAL CONTENTS-Continued Section 7.......... ....... ......... .................. .... .......... ... .......................... ...... ...... ..................... .......... .... ...... 228 228 231 237 TABLES Table 4-l 4-2 Unit hydrograph lag data for the Great Plains . ........ ......... (f) Hydrometeorological analysis. ........ .. .... ....... I”.................. . . ... .... .............. ............. ....... ........ ............. ..... (z) Discharge-probability analyses... ... ..3 Field reconnaissance report ............... . ............11 Foss Dam example ......... ... ... ..... (c) General background .... ....... ...... .. ............... ............ . .... ....... ..... ............... ....... . .. .................. (m) Reservoir routing criteria ..... .. .. Montana...... ..... . ........ . Great Basin.... Idaho........... . . ...... . ........ . Unit hydrograph lag data for urban basins .. ... ..... ..... . ............... ...... ............ .............. ............ Page 79 81 82 84 85 86 4-3 4-4 4-5 4-6 .. .......... and Washington ... (i) Base flow and interflow .. .. .... ........... ...... .. ...... ..... Wyoming....... 223 223 223 223 224 224 225 225 225 226 226 226 227 227 227 ................ .... .... ... ... ............ .......... and Colorado Plateau ... .. ...................... ......... .... .. . .. . .. ....... ... .......... . .... . .. . .. . .......... ..... ............ . Unit hydrograph lag data for the Southwest Desert... ....... ........ ...... . .... .. .... ...... .........12 Example on mixed population analyses. ........ . ..... ... ......... . . ...... ............................................ ..................... .. ... ‘ (1) Flood hydrograph routings..... .... ............ . . . .... ...... .. ................ .... ....

.......... 104 4-19 Values for synthetic unit hydrograph................................. Coast................................................ ........................... and Colorado Plateau ........... 4-9 General storm dimensionless unit hydgrograph data for the Rocky Mountains ...... 172 5-11 Total PMF hydrograph for Foss Dam............ .................................... 5-3 Flood routing constants by Successive Average Lag 145 Method ............ and Colorado Plateau ....................... .................. 5-4 Example on routing computations by Modified Puls Method ........ 5-l Tabular procedure for routing and combining subbasin flood hydrographs to obtain total basin hydrograph.............. 96 Coast.................. and Cascade ranges........................................................................................................... of coefficients K and X for the Muskingum 5-8 Determination 162 Routing Method......... 90 4-10 General storm dimensionless S-graph data for the Rocky Mountains ............................................................. 5-10 Routing constants by Muskingum Method for reaches with 168 different flood-wave travel times.............................................................. . 98 4-17 Dimensionless unit hydrograph data for urban basins .. 120 4-20 Properties of subbasins in Washita River Basin...................... ....................... 4-15 Dimensionless unit hydrograph data for the Sierra Nevada......... ....................................................................... 123 125 4-24 Flood hydrograph development for subbasin 2......... 88 4-7 Dimensionless unit hydrograph data for the Great Plains 89 4-8 Dimensionless S-graph data for the Great Plains . ................. 97 and Cascade ranges..... 127 4-25 Flood hydrograph development for subbasin 3.. Great 95 Basin.... 91 4-11 Thunderstorm dimensionless unit hydrograph data for the Rocky Mountains ............... ............................ ................. ................................. 130 4-26 Flood hydrograph development for subbasin 4.................... ............................................ 4-21 Unit hydrograph lag time for each subbasin in Washita River Basin......................................................... .................... 92 4-12 Thunderstorm dimensionless S-graph data for the Rocky Mountains ......... ix .. 4-14 Dimensionless S-graph data for the Southwest Desert................................................. 99 4-18 Dimensionless S-graph data for urban basins .... 121 122 4-22 Unit hydrograph development for subbasin 1 ...CONTENTS TABLES-Continued Page Table ........................................... 148 151 5-5 Stream routing data by Modified Puls Method....... 5-9 Routing of Columbia River discharges using routing 166 constants derived by Muskingum Method..................... ........... 4-16 Dimensionless S-graph data for the Sierra Nevada.................. 5-6 Stream routing computations by Modified Puls Method ....... 153 157 5-7 Flood routing by Muskingum Method............ ................... ......................... . 4-23 Flood hydrograph development for subbasin l....................................................................... Great Basin............................. 135 141 5-2 Flood routing computations........ 93 4-13 Dimensionless unit hydrograph data for the Southwest 94 Desert...................

.... 1946-63.... Dimensionless unit hydrograph lag time.. ............. .......... ......... ........... ... . ........... .... .... ... .. ........... Sequential arrangement of PMP increments for a 24-hour storm.....and 95-percent confidence band calculations for the Washita River near Cheyenne. ... ...... .. ..... 215 7-2 Development of data for construction of dischargeprobability relationship ............... Discharge-probability relationship.....000 to 700 millibars..... . ... Pecos River near Comstock.... ... ......... Oklahoma ............. .. .......000 square miles........... . . .. ................ ..... .. ... .. .. ......... .... Montana... .......... .. ....... Discharge-probability relationship...... . .. ......... ........ .......... ........ Page 3 4 5 6 19 33 34 35 39 39 40 44 52 59 65 66 67 69 70 X ........ ... . .. ......... ....... .... Storm precipitation survey form ....... .:: 3-3 3-4 3-5 3-6 3-7 3-8 3-9 4-l 4-2 4-3 4-4 4-5 Discharge-probability relationship. Unit hydrograph principles .......... ... .............. Colorado . Colorado.... ...... .. ........................ . Typical flood hydrograph reconstitution.... ..... .... ............... .. Pseudo-adiabatic chart for dewpoint-elevation adjustment ... ...... ....................... .... ............ ........... Nork Fork of Sun River near Augusta... 1901-53 . .................... ............... ......... .. Discharge-probability relationship........... ... Diagram for precipitable water determination from 800 to 200 millibars .. . .. . . .......... S-Graph lag time.. ... . ... .... Plum Creek near Louviers................ .... . Texas... ..... 1948-64 . ... 221 7-6 Flood frequency parameters for North Fork of Big Thompson River near Drake. ........... Orographic laminar flow precipitation model....... ... . .... ... ........... .. .. ......... ........... Colorado. ...... ... ...... . ........ ......... ................. ........... ............ ..... . ..... ..... .. ...... ...... .............. ..... Depth-area envelope of transposed maximum 24-hour precipitation . ........... ... .. Arkansas River near Pueblo.. . .. ... . ......... ... ......... ... . ....... . 218 7-4 Discharge-probability relationships at Foss Dam .. ..... ... .FLOOD HYDROLOGY MANUAL TABLES-Continued Page Table 7-l Annual peak discharges for each year of record for the drainage area above Foss Dam . Colorado ..... ...... 2 18 7-3 The 5........ Unit hydrograph application.......... Diagram for precipitable water determination from 1... Depth-area-duration envelope......... .. ......... .... ............ . ....... .. .. ... ... .... . .......... ...... .............. 222 FIGURES Figure l-l 1-2 l-3 l-4 2-l .. 1895-1920 ... Hydrometeorological report series coverage of conterminous United States ....... . .... ........... . .. ........... . ...... .. 220 7-5 Annual rain peak and snow peak discharges for North Fork of Big Thompson River near Drake........ .... ............... Depth-duration envelope of transposed maximum storm values for 1.. ............... .. .... .... ...... .. .............. ...... ...... .......... ............ . ........ ..

...................................................... 78 4-l 2 Dimensionless unit hydrograph and sample computations...... 158 5-12 Typical valley storage curves..... 149 5-8 Storage indication and percolation loss curves.................. .............. ........... 137 5-3 Typical spillway discharge curve..................... .... 163 5-14 Illustration of wedge and prism storage................. 73 4-7 Unit hydrograph lag relationships for the Rocky 74 Mountains .......... 108 4-15 Basin outline for Foss Dam ................ 4-8 Unit hydrograph lag relationships for the Southwest Desert.............................. and Colorado Plateau.. 134 5-2 Typical inflow and outflow hydrographs...... 154 5-10 Muskingum routing principles ......................................................... 164 5-15 Effects of a varying K constant and number of subreaches on a routed hydrograph............................................. 178 7-l Probability density function curve .... 152 5-9 Example of a routed flood hydrograph by Modified Puls Method .... ................................... .................................... ...... ............... 75 4-9 Unit hydrograph lag relationships for the Sierra Nevada in California ......... 208 xi ................................ 192 7-2 Examples of skewed probability density function curves ..................................................... Great Basin..... ......................... 192 7-3 A typical cumulative distribution function curve .... ......................... ..... 10 1 4-l 3 Typical dimensionless S-graph......................................................... 103 4-14 Typical components of total flood runoff hydrograph ....... 143 5-6 Storage indication curve .......... 161 5-13 Effects of a varying X constant on routed hydrographs ................................... 193 7-4 Mixed population frequency curve .................... ........................... Oregon.............................................. ........ ....... 119 5-l Hypothetical drainage basin for flood hydrograph routing and combining considerations. 165 6-1 Typical peak discharge envelope curve .................... ........................................... 138 5-5 Hypothetical flood hydrograph routed by Tatum Method... ... 156 5-l 1 Application of channel storage equation...... 76 4-10 Unit hydrograph lag relationships for the Coast and Cascade ranges in California....................................... 165 5-16 Effect of subreach length counterbalanced by a varying X constant........................................................ 137 5-4 Typical reservoir capacity curve................................. 77 4-l 1 Unit hydrograph lag relationships for urban basins......................................................CONTENTS FIGURES-Continued Page Figure 4-6 Unit hydrograph lag relationships for the Great Plains...... and Washington.................................... 146 5-7 Hypothetical flood hydrograph routed by Modified Puls Method .......................................................................... ........................................... ...........


Jansen. evidence that this land may have been the birthplace of the art.BACKGROUND 1. India.” During this period. many noted philosophers speculated. In the ruins at Sialak. In China. suggesting that irrigation was practiced there from very early times. Menes.D.1 Background Chapter 1 AND HISTORICAL PERSPECTIVE Archeological evidence indicates that man has been attempting to regulate the flow of water in streams and rivers for beneficial purposes for about 6. this knowledge and the lack of scientific rainfall and runoff measurements was insufficient to enable the designers of that time to estimate. construction of impressive dams was accomplished on the Min River for flood control and diversion of water to nearby farmlands. Robert B. and provided lifelines on which their civilizations depended. The Persians of ancient times recognized the importance of irrigation to the sustenance of civilization. near Kashan. Study of ancient China. the first Pharaoh of Egypt. about the nature of the hydrologic cycle. This evidence generally consists of the remnants of dams and irrigation systems.” that the science of hydrology commenced. the flood potential at their damsites. former Assistant Commissioner of the Bureau of Reclamation. Iran. and wells. they accomplished projects which rank among the greatest in history. The sacred books of India cite the very early operation of dams. in quantified terms. 1400 as the “Period of Speculations. this led to underdesign and ultimate failure of their constructed works. are to be seen traces of irrigation channels which are considered to be as much as 6. in his excellent publication Dams and Public Safety [l]‘. The first measurements of rainfall. channels. ‘Numbers in brackets refer to entries in the Bibliography. By excavating underground water tunnel and gallery systems (quanats) and by constructing many dams. Ven Te Chow [2]. even before the arrival of the Aryans in the land now known as Iran. evaporation. these philosophers gained a certain measure of practical knowledge. Generally. often erroneously.000 years. “History does not record exactly when irrigation systems were first constructed. in his 1962 publication Handbook of Applied Hydrology.000 years old.” Jansen goes on to recount a number of dam failures associated with these ancient works. termed the period dating from these ancient times up until about A. Unfortunately. due to the number of water control works constructed. . ordered irrigation works to draw from the River Nile. However. and Egypt does reveal that such work in these lands was begun thousands of years ago. It was not until the 17th century. many of which were apparently the result of an inability to safely pass or accommodate floods that their tributary watersheds had produced. states that. the period that Chow termed the “Period of Measurements.

In other cases. using the Embudo gauge. On 2 .000 feet in elevation. Four examples of occurrences where floods were much greater than previous maximum recorded floods are graphically presented on figures l-l through l-4.” as it is currently known. designers based their dam designs on the historic flood of record raised to a higher value by applying a factor. the annual maximum peak discharge recorded was 30. From sometime in 1885 to June 3. Another approach used was the application of a number of empirical formulas generally relating peak discharge to drainage area. Geological Survey) [3]. statistical procedures were used by some in evaluating the flood potential of basins where dam construction was planned.2 Historical Procedures In the late 1800’s and early 1900’s. and flows easterly across a high plains region passing through Pueblo. 1. These four examples and one other are discussed in the following paragraphs. or “discharge data acquisition.-The Arkansas River at Pueblo. Colorado rises near the Continental Divide and drains areas exceeding 14. 1921. which remains in use today. Each annual maximum flow was the result of the spring snowmelt runoff from the higher elevations in the basin. established on the Rio Grande River at Embudo. These factors were at best quite arbitrary. This station. served as the forerunner to systematic stream gauging. The establishment of a systematic stream gauging program by the USGS provided the needed data base for later (20th century) development of both physical and statistical flood hydrology procedures presently in use by the major Federal water resource construction agencies. Systematic stream gauging on the Arkansas near Pueblo was initiated with the establishment of the Pueblo gauge in 1885.000 cubic feet per second on May 21. with continuous daily records starting in 1895. These examples illustrate the extreme variability of flood flows and the potential consequences of relying on historic peak discharges as a basis for design of high hazard water control structures. Colorado. Each of these three approaches was used with apparently satisfactory results until a number of flood events occurred that could not have been reasonably predicted by application of any of these approaches.S. (a) Arkansas River. Measurements of river flow were accelerated in the 19th century with the development of measuring instruments such as the Price current meter. The 19th century marked the installation of the first stream-gauging station by the USGS (U. a method proposed during the 19th century. New Mexico by John Wesley Powell in December 1888. 1901.FLOOD HYDROLOGY MANUAL and river discharges were made in Central and South Central Europe. served as a training facility for instruction in the use of instruments and the methods used in the relatively undeveloped art of stream gauging. The river leaves the mountains near Canon City. The Embudo Camp.

Arkansas River near Pueblo. Colorado.. 103-D-1901.-Discharge-probability relationship. 1895-1920.000ff% PROBABILITY OF DISCHARGE BEING EOUALED OR EXCEEDED IN ANY GIVEN YEAR~PERCENTI Figure l-I. .I921 FCOOD:l03.

wo 300.71954 900.000ft3h. Pecos River near Comstock.000 200. 1901-53.000 6w.-Discharge-probability relationship.030 -wJo 400. 103-D-1902.053 7w.000 100. I III 0 E 5 = K! L E 0 5 s z iii % 2 9 0 4’ L 800.ooo I FLOOD!94B.000 II 0’ 9999 999999 99 99 995 WOBABILITY OF III 95 90 DISCHARGE BEING EOUALED OR EXCEEDED IN ANY GIVEN YEARIPERCENT) Figure l-Z. . Texas.


1946-63. . Nork Fork of Sun River near Augusta.8 0.-Discharge-probability 103-D-1904. relationship. Montana.1964 50.93 99993999599 ” ’ ’ 98 OF ’ I 95 III 90 80 BEING Illlf 10 60 50 40 OR 30 III 20 IN 00 ANY I 5 GIVEN I 2 I I I as III 0.100f& o-J 99.2 0.05 I 0.00 PROBABILITY DISCHARGE EOUALED EXCEEDED YEARIPERCENT) Figure 1-4.000 FLO0D:51.

In March 1936. 1965 flood on Plum Creek. this event was almost nine times higher than any recorded event in the previous 53 years.960 cubic feet per second in 1933. In the mountainous region of Montana in early June 1964.000 cubic feet per second. The resulting discharge from this storm was measured at 154. l-4) had recorded an annual maximum discharge of about 4. Texas. a large Great Plains type thunderstorm centered over the high plains region immediately upstream from Pueblo. the maximum recorded peak flow was measured at 2. (d) Sun River. four times larger than 7 . This discharge represents a threefold increase over the previously experienced recorded maximum shown on figure l-l. yielded a maximum peak discharge of 7. Streamflow records accumulated at the USGS stream gauge on the North Fork of the Sun River near Augusta. (e) Ware River.200 cubic feet per second. the highest annual peak flow recorded was 116. During the first 53 years the gauge was in operation. This severe rainfall event produced a measured peak discharge of 948.000 cubic feet per second in 1932. A stream gauge was established at this location in August 1912. in 1901.700 cubic feet per second. This extreme event. a flood occurred that had a recorded peak of 11. produced a peak discharge at the Pueblo gauge of 103.162square mile area of the Pecos River Drainage Basin.800 cubic feet per second for the period from the time the gauge was installed in 1946 through 1963. As shown on figure 1.June 4.100 cubic feet per second.-Such phenomena are not limited to the high plain regions. Montana (fig.000 cubic feet per second. over 10 times the previous maximum peak recorded discharge.500 square miles of the basin immediately above the Comstock gauge. (c) Plum Creek-One of the most significant examples in terms of exceedance of previously experienced discharges was the June 16. a severe storm on the Sun River produced a flood event of such extreme magnitude that it overtopped the Bureau’s concrete-arch Gibson Dam. Colorado.-A stream gauge was installed on the Pecos River near Comstock. the 1965 event was about 22 times larger than what was experienced in the previous 23 years that the gauge was in operation. Systematic measurements of streamflow at the Plum Creek near Louviers gauge since 1942. with rainfall amounts of over 10 inches within a 6-hour period.000 cubic feet per second. consider the Ware River at Gibbs Crossing. Massachusetts. This event produced a flood with a peak discharge of 51. and there are many more.-As a final example. (b) Pecos River. a major storm event centered over the lower 2. and measured runoff from a 35. As shown on figure l-3. In June 1954. Colorado centered over this 302-square mile drainage area in the foothills of the Colorado Rocky Mountains. From 1912 to March 1936. 1921. A severe storm similar in type to the June 1921 event in Pueblo.2.

It can readily be seen from these few examples that engineers. Leroy K. this has proved to be the case. falling on a drainage basin. This branch has published a series of reports. A year later in 1933. for converting rainfall excess on a drainage basin to a hydrograph representing surface runoff. In 1932. This group of meteorologists became what is presently known as the Hydrometeorological Branch in the Office of Hydrology at NWS. This series of events probably led the dam engineering community. A. which provided the analyst with the capability to account for the amount of rainfall. Failures. 1. to the mid 1930’s. catalogs 55 failures of private and public dams in the United States due to overtopping during the period from 1889. as a result of extreme rainfall that took place during a hurricane.3 Hydrometeorologic Approach The 1930’s saw the basic hydrologic and meteorologic tools become available for what became known as the “hydrometeorologic approach” to upper limit or PMF (probable maximum flood) hydrograph development. or model. a peak discharge of 22. that provide definitive criteria for developing PMP estimates.700 cubic feet per second was recorded. This latter flood event had a peak twice as large the preceding record event and over seven times the next largest event. 0. and Accidents [4]. the only significant part of the procedure missing was a technique for estimating the upper limit rainfall amounts that could be used to compute the PMF hydrograph. lost by infiltration into the ground and therefore not available for surface runoff. Merrill Bernard [7] of the NWS (National Weather Service) and Gail Hathaway of COE (Corps of Engineers) reached a cooperative agreement whereby NWS would establish and COE would fund a specialized group of professional meteorologists to study severe meteorologic phenomena and develop criteria that could be applied by the hydrologic engineer to determine upper level or PMP (probable maximum precipitation) values. generally referred to as the HMR (Hydrometeorological Report) series. with modifications. when the well known Johnstown. discussed in more detail in chapter 4. Babb. to seek ways to more reliably estimate the upper limit flood potential of a drainage basin. provided the basis. in his 1968 Bureau publication Catalog of Dam Disasters. could greatly underestimate the flood potential of a drainage basin and develop an inadequate hydraulic design. basing their dam and spillway design on either factoring up a recorded maximum flow or by applying statistical procedures to estimate a given flood probability. in the mid 1930’s. Only 2 years later in September 1938. Horton [6] proposed his infiltration theory. In 1933. Robert E. 8 . Pennsylvania disaster occurred. In many instances. is still being used by the Bureau and other major Federal water resource development agencies. Sherman [5] proposed the unit hydrograph theory which. This theory. In 1936.FLOOD HYDROLOGY MANUAL the previous maximum.

and did not always reflect the most efficient meteorologic processes. The document Unit-Graph Procedures [9] was published by the Bureau in November 1952. As will be discussed in detail in chapter 4. and the National Weather Service [8]. as shown with their regions of coverage. Review of these studies under the Bureau’s safety of dams program indicated that these past practices yielded results that were frequently inconsistent within meteorologically homogeneous regions. and embodies the basic principles advanced by Sherman 20 years earlier. are shown on figure 3-7 in chapter 3. This objective has been achieved with the publication of HMR 55A. developing PMF hydrographs.ssed in general terms in chapter 3.observed storms suitable for analysis. did not systematically account for all available historically . the principal modification to the document was that of introduction of a factor into the general unit hydrograph lag time equation that varies as a function of a drainage basin’s hydraulic efficiency characteristics. The derivation of these criteria is discu. Soil Conservation service. Statistical analyses are generally required to estimate flood discharges used in assessing cross drainage requirements. reliable. and reasonable PMP criteria are available for use by the dam engineering community for the conterminous United States. The techniques outlined in this document were modified to incorporate work conducted by hydrologic engineers in the Bureau and in the Los Angeles District and South Pacific Division offices of COE in the 1960’s and early 1970’s [lo]. considerable work in flood hydrology relies on data developed using the former methods. it became Bureau policy to adopt the HMR series for estimating PMP and. The unit hydrograph technique and Horton’s infiltration theory have been the basic tools used by the Bureau for many years to convert rainfall to flood runoff.BACKGROUND Beginning in 1980. The present practice and policy of the Bureau is to use these reports for determining PMP values for virtually all PMF studies. The primary objective of this committee has been to assure that consistent. hydraulic and structural design of diversion 9 . Exceptions are encountered when the drainage basins being studied are larger in area1 extent than that covered by the reports. in turn. Corps of Engineers. This practice departs from past Bureau policy where individual hydrometeorological studies were conducted for individual basins tributary to dams. The most up-to-date publications in the report series. To overcome these deficiencies. the activities of the Hydrometeorological Branch have predominately been steered by an informal committee comprised of representatives of the Bureau of Reclamation. and a detailed derivation is presented in each individual report. The introduction of this factor was generally considered to improve the reliability of synthetic unit hydrographs derived for ungauged drainage basins. Even though statistical or probabilistic methods for determining upper limit flood hydrographs have been discarded in favor of the deterministic or hydrometeorological method.

The goal of this effort was to achieve the greatest degree of consistency possible in probabilistic studies conducted by these agencies. two or more entities conducted discharge probability studies for the same basin. On several occasions. the determination of flood control benefits and the associated statistical hydrologic analysis lie within the province of COE because of their legislated authority. and 17B [ 141 were successively published. which was placed in use by all agencies. These criteria and their development are discussed in chapter 7. computing PMF hydrographs. Bulletin 17B provides the basic probabilistic criteria being used by the Federal establishment. that would be applied by all agencies in the development of discharge-probability relationships. 10 . this situation led the National Water Resources Council to implement a study group comprised of representatives of each Federal water resource development agency to develop guidelines. Currently. establishing crest elevations of auxiliary and fuse plug spillways. and for determining economic benefits for flood control projects. Statistical analyses associated with cross drainage design and river diversions are a problem common to all water resources development entities. The following chapters of this manual will provide definitive information on current procedures used by the Bureau of Reclamation in estimating PMP values. deriving unit hydrographs. river diversion requirements during dam construction. as problems with Bulletin 15 criteria became apparent through application. this chapter has provided the reader with some general baclcground information as to where the specialty area of flood hydrology has been and where it is today. 17A [ 131. assigning appropriate infiltration loss rates. each being essentially an improved revision of its predessor. At that time. As previously mentioned. Hopefully. Subsequently. routing flood hydrographs through reservoirs and river channels.FLOOD HYDROLOGY MANUAL dams. Bulletins 17 [ 121. these entities used several different technical approaches to determine discharge probability relationships. using a single technique. and conducting statistical analyses of streamflow data. The result of this 1967 effort was the publication of Water Resource Council Bulletin 15 [ 111. with results that often varied considerably. Prior to 1967.

snow surveys conducted by Federal. timing. The well is connected ‘Numbers in brackets refer to entries in the Bibliography.These data are collected primarily by the USGS at continuous recording streamflow gauging stations. temperature.2 Hydrologic Data (a) Recorded Streamflow Data. these data are required in the development of criteria by hydrometeorologists for making PMP estimates. when used with data collected at the official NWS stations. commonly referred to as “bucket surveys”. and reservoir operation records from which inflow hydrographs may be determined based on outflow and change of storage relationships. State. 2. and degree of development. provide fairly definitive information on the area1 extent. and local agencies in areas susceptible to significant snowmelt runoffs. StreamCaging Procedure [15]l. Briefly. geologic setting. As will be seen in later chapters. 11 . and data from supplemental precipitation surveys. and for preparing discharge-probability relationships. The first type is actuated by a float installed in a well located below the gauge house. The purpose of the latter is to provide additional data which. amount and type of vegetation. indirect peak discharge measurements based on flood marks at locations where there are no stream gauges. A comprehensive discussion on the installation equipment and operation of these stations is presented in USGS Water Supply Paper 888. the stations are equipped with devices that sense and record the variation of the river stage above an arbitrary’ datum. conducted immediately after the occurrence of severe storm and flood events. and wind records collected at “official” NWS first and second order climatological stations.1 General Considerations Cha ter 2 A J D METEOROLOGIC DATA The most fundamental part of any flood hydrology analysis is the compilation and analysis of hydrologic and meteorologic data accumulated during and after severe flood events. crest stage streamflow gauges. drainage network development. Meteorologic data include precipitation. and intensity of a storm. dewpoints.BASIC HYDROLOGIC 2. There are two primary types of sensing and recording devices currently in use at these stations. development of unit hydrograph and infiltration parameters necessary to determine the rainfall-runoff relationships for both gauged and ungauged basins. Other vitally important data include data related to watershed characteristics such as topography.. Hydrologic data include records of flood runoff measured at continuous recording streamflow gauges.

and is driven by a clock-actuated device. usually filled with waterproof ink. that causes a pen in the device to move back and forth with the changing river stage. As a result. As the river stage rises or falls. or thalweg of the river channel. are usually made using a current meter that measures the velocity of flow. causing the counter-weighted cable to move upward or downward. as discussed in chapter 4. The recorded river stage data. The pen. The primary advantage of the punch-tape unit is that the data recorded on the paper tape can be extracted by automated means. it is possible to miss recording the peak of a runoff event. whereas the continuous recording device requires manual reduction of the data. The A-35 device produces a continuous trace of the river stage whereas the punch-tape unit only records the stage at predetermined time intervals. The clock drive advances the paper chart at a fixed rate so that. The variation in pressure is then sensed by a mercury servomanometer which. over the maximum range possible.FLOOD HYDROLOGY MANUAL to the river using pipe intakes that allow the water level in the well to maintain the same elevation as the water level or stage in the river. which would render the record unusable in observed flood hydrograph analyses. To overcome this problem. If the time interval between readings is too long. drives a recording device. Discharge measurements. in most cases. the float in the well does likewise. particularly for smaller drainage basins. or stage. in turn. the water-level recording device produces a continuous trace on the chart of the variation of river stage over time.” Rather than sensing the water surface level by using a well and float. in the river as the pressure in the tube corresponds to the head on the orifice. and either staffs or calibrated cables that are used to measure depths of flow across the cross section of the channel at the measuring 12 . usually either a Stevens A-35 or a Fisher-Porter punch-tape unit. The float is connected to a recording device by a counter-weighted cable. whether accumulated by a continuous recording device or a punch-tape unit. The cable is connected to a drive wheel on a recording device. Because pressure must be exerted on the gas to overcome the water pressure or head at the orifice. The second type of device is called a “bubbler gauge. This is accomplished by developing a rating curve for the stream gauging site that relates discharge to river stage. The time interval set for a punch-tape unit is of importance in flood hydrograph analyses. generally a Stevens type A-35 recorder. it is possible to measure the variations in pressure and relate them to variations in the water surface level. 1 inch on the recorder chart equals 12 hours. must be converted to discharge values for use in the hydrologic analyses. many bubbler-gauge installations include both a punch-tape unit set at a relatively long time interval and a continuous recording device such as the Stevens A-35. records the stage on a continuously moving recorder chart that moves at right angles to the movement of the pen. the bubbler gauge feeds nitrogen gas under pressure through a tube to an orifice located at or near the bottom.

This is generally due to the inability to obtain accurate velocity measurements at higher flows. The total discharge of the section at the particular stage is then found by adding the discharge determined for each part of the total section. In general principle. and a detailed discussion on the procedure is beyond the scope of this chapter. Average daily flow values are developed from recorder charts that provide a continuous record of river stage versus time at each gauging site. This general procedure is repeated for various river stages.BASIC DATA section at or near the gauging station. the procedure is somewhat more complicated than described here. Stream-Gaging Procedure-A Manual Describing Methods and Practices of the Geological Survey. As a result. a statement as to whether the records are excellent. Horizontal measurements are taken between staff readings. If interested. the records will be rated at one level up to a given discharge and then rated lower at higher discharges. a rating of good implies an accuracy of the velocity measurement within 5 percent. however. or fair. The streamflow records published by the USGS will provide.and volume-probability relationships as discussed in chapter 7. A rating of excellent implies a velocity measurement within 2 percent accuracy. the data are rarely used in hydrograph analyses. The data are suitable for developing both peak. The streamflow records are compiled and published by the USGS in a series of “Water Supply Papers” that are generally available in each Bureau office. dated 1962 [151The hydrologic engineer should. Accordingly. 888. however. local USGS offices. and a fair rating implies an accuracy within 8 percent. In actual practice. the velocity of flow is measured at several locations across the channel resulting in an average velocity for the section. for each station. the accuracies of the resulting discharge data may be less. the rating curves have to be extrapolated beyond the measured data. the data are of limited value for observed flood hydrograph analyses (ch. If the upper end of the rating curve is based on indirect measurements. In many instances. or to having a hydrographer present to measure the flow when the peak discharge occurs. The depth reading times the sum of one-half the distance between adjacent staff readings yields the cross-sectional area that may be applied to the velocity to obtain the discharge for that part of the total cross section.2(b). good. 4) in all but very large drainage basins. The staff or cable readings for depth of flow are taken at each point that the velocity is measured. be aware of the relative accuracies inherent in these data. and a plot of discharge versus river stage is then constructed which is the rating curve for the station. this is discussed later in section 2. River stage is shown on the recorder chart as an elevation. and in major city libraries. in 13 . These charts can provide valuable information for the flood hydrograph analyses. These publications present streamflow in terms of average daily flow for the period of time that the stream gauge has been in operation. the reader is encouraged to study the USGS Water Supply Paper No.

and then converting those values to discharge using the rating curve for that station. To supplement the indirect discharge measurements are data acquired at these stations. the system also stores data for about 20. In view of these high costs. As the procedures for accessing the system are modified from time to time. taken at other locations following flood events.-The costs associated with installing. and the annual peak discharge-frequency relationships discussed in chapter 7. and expensive in terms of manpower requirements. The Bureau currently has access to this system in all of its project and regional offices. Copies of these charts can be obtained from the Water Resource Division in the local offices of the USGS along with the rating curve for each gauging station.000 stations that are no longer operating. The time interval selected should be sufficiently short to properly define the recorded hydrograph. there are relatively few continuous recording stream gauges in the United States considering the number and length of all the rivers and streams involved. The previously mentioned Water Supply Papers also present the instantaneous peak discharge for each gauging station for each year that the station has been in operation. at the Denver Office for current accessing procedures. and in the Denver Office. the USGS has provided for storage and retrieval of these data in a system known as WATSTORE (National Water Data Storage and Retrieval System). (b) Zndirect Discharge Measurements. The process of abstracting the required data from the many volumes comprising the Water Supply Papers for use in the statistical analyses discussed in chapter 7 is tedious. A hydrograph representing the discharge in cubic feet per second for a particular location can then be developed by reading river stage values from the recorder chart at selected time intervals. Proper hydrograph definition is important to gain an understanding of the runoff characteristics of these gauged basins. 14 . and maintaining continuous recording streamflow gauging stations and compiling and publishing the resulting data are rather high. the new user should contact personnel of the Surface Water Branch. operating. These data form the basis for developing the discharge envelope curves discussed in chapter 6. some peak flow values less than the maximum annual event but greater than a base flood flow level are presented. With the advent of current highspeed electronic computers and their large data storage capabilities. above some arbitrary datum. laborious.FLOOD HYDROLOGY MANUAL feet.1 of chapter 4.000 streamgauging stations presently in operation. In addition. There is also a high potential for human error when performing this task manually. especially in the case of hydrographs with rapid rise and recession limbs. In addition to storing streamflow records for over 11. Flood Section. This point is treated in more detail in section 4. which are of use in developing partial duration discharge-probability relationships.

Indirect discharge measurements are made as soon as practicable after flood events. This equation is used to determine the 15 . Chow [16] cites live categories of this type of measurement including (1) measurements taken along channel reaches. 6). This type of gauge is currently installed and operated principally by State highway departments in many regions of the country. as described in USGS Water Supply Paper 888 [15]. There are a few instances where an observer has marked flood heights in a permanent fashion at recorded time intervals during both the rise and fall of the water surface during a flood event. Also important to an adequate indirect measurement is the establishment of the high water marks attained during the flood event. (4) where flow takes place over what may be considered a broad-crested wier such as a highway embankment. A graduated rod is inside the pipe and granulated cork is placed at the bottom.” where high water stages are recorded. or may be actually recorded as in the case of flow over a dam’s spillway by the reservoir’s water-level recorder. One of the best methods of recording the maximum stage is by using a crest stage gauge. or the discharge is determined from the maximum stage using the “slope-area” method of indirect peak discharge measurements. This maximum stage is then related to discharge using a “rating curve. at bridges along the stream. This provided a stage versus time relationship that was then converted to a flood discharge hydrograph using either an existing rating curve or a curve developed from data derived using slope-area procedures. The maximum stage may be in the form of physical evidence such as debris or scour along the banks or overflow area of a watercourse. usually through a short section of the river or stream. 7). (2) those taken at contracted openings such as bridges. as discussed in chapter 4. as described in section 5.16] that uses Manning’s equation. They are simple devices consisting of a length of 2-inch-diameter pipe mounted vertically on a post or bridge pier. Another category where indirect discharge measurements are used is called the “crest stage gauge.BASIC DATA Indirect discharge measurements are of considerable importance in flood hydrology in the development of envelope curves of experienced peak discharges (ch. Entry of water during flood flows causes the cork to rise and adhere to the rod at the maximum stage achieved during the event. as it can be ascertained. In most cases. and (5) where flow passes through a hydraulic structure such as a spillway or outlet works at a dam. and in the development of peak discharge-probability relationships (ch. (3) readings taken where flow passes through culverts.” if one exists. The measurements are based on hydraulic computations using the flood water’s maximum stage.4. The pipe is capped at each end with the lower part of the pipe perforated in the direction of flow to permit entry of water. The majority of indirect discharge measurements are made along channel reaches using the slope-area method [15. these indirect measurements determine only peak discharge and are therefore of only limited value in observed flood hydrograph analyses.

this average velocity is multiplied by the average cross-sectional areas to arrive at the discharge estimate. radio and TV stations.-Systematic acquisition of precipitation data is accomplished primarily through the efforts of the NWS.FLOOD HYDROLOGY MANUAL average velocity in the reach and then. and “fair” for those within 15 percent. A second order station is operated by meteorologists during the day. As Bureau personnel may initiate or be called upon to participate with other interested entities in these important supplemental precipitation surveys. the Flood Section maintains records for “second order” NWS stations and for data collected by NWS cooperative observers. to supplement the data obtained at “official” rain gauges with “unofficial” observations. collecting the same data as a first order station. as compared with the past. Naturally. who maintains a network of what are termed “first order” weather stations that are operated on a 24-hour basis by meteorologists. having measured the cross-sectional areas along the reach. (b) Supplemental Meteorologic Data Acquisition or Bucket Surveys. Care must be taken in selecting the reach to minimize the variation in crosssectional areas and to be reasonably sure that steady flow occurred at the maximum stage reached during the flood event. A complete set of microfiche containing these systematic data is maintained by hydrometeorologists in the Flood Section at the Bureau’s Denver Office for use in PMP and storm studies for which they are responsible. following outstanding storms.3 Meteorologic Data (a) Systematic Data Acquisition. 2. Cooperative observers collect 24-hour rainfall and maximum and minimum temperature data. Also. guidance is provided relative to recognizing the need 16 . It should be noted that the accuracies of flow measurements using the indirect method are not as good as those of the systematic recording gauge. Each weather station in this network collects continuous precipitation. relatively more complete data are being obtained for current storms. requiring some experience on the part of the analyst.- Hydrometeorologists require data on historical storms for use in preparing PMF hydrographs and developing operating criteria for performing flood routings through reservoirs. and relative humidity data. The USGS classifies as “good” those measurements within 10 percent of the true value. wind. With the numerous recording rain gauges now operating. the network of precipitation stations is still far from sufficient to provide the necessary temporal and spatial data for detailed analyses of observed storm precipitation. However. rather detailed. These observations include measurements of precipitation caught in any type of open receptacle. the selection of proper Manning’s n values is critical to an adequate estimate. and city and county public works departments. temperature. the following. These unofficial observations may be made by individuals. It is therefore necessary and extremely important.

Plans should be adjusted to eliminate duplication of effort among the participants. .. (3) It is difficult and not particularly desirable to give the survey team members hard and fast instructions on the routes they are to travel in 17 :i (. In the interest of preserving and maximizing the reliability of the data. When such an event occurs. the appropriate project or regional office of the Bureau. as well as other interested agencies. and television reports of precipitation and stream or river stages and from observations available at local weather service offices. in any part of the area for any duration. efforts should be made to coordinate the decision to conduct a bucket survey with hydrometeorologists of the Flood Section at the Bureau’s Denver Office. and SCS. approaching or exceeding the maximum observed for such area and duration. and local agencies should be contacted to determine their capability in supporting the field survey and whether they desire to participate as survey team members. radio. However. and appropriate regional offices of NWS. and developing procedures for. In particular. State engineers. it is best to accomplish the necessary coordination in person or by telephone because correspondence by mail has been found to involve unacceptable delays for storm survey work. State and local agencies that might have an interest should also be approached for possible involvement during these preliminary coordination activities. It must be reemphasized that this activity is of the highest importance in the area of hydrometeorology. a supplemental precipitation survey of the storm should be made. In general. (1) Within the Bureau. The occurrence of unusual precipitation events may be indicated by newspaper. reported rainfall equivalent to that of an actual or potential major flood-producing storm should be investigated. the data must be collected as quickly as possible after the occurrence of the storm so that the decision on whether to conduct a survey can be made quickly. as well as later in the survey. USGS. NWS. The plotting of available reports on precipitation amounts on a map of the genera1 area will be helpful at this phase. . \ \‘. these local information sources does not reduce the responsibility of the Bureau offices to be independently cognizant of the occurrence of severe events in their areas of responsibility. the responsibility for recognizing particular storms as being outstanding and for collection of supplemental rainfall data lies with the regional and project offices in whose area the storm occurs. In general. (2) When the decision has been made that a particular storm merits investigation.BASIC DATA for. primarily the COE. and to ensure full coverage of the area over which the storm occurred. If the available data show the possibility of rainfall rates. conducting these surveys. detailed plans for the field survey should be coordinated with others that have indicated interest at the preliminary stage. district office of COE. Local offices of the Bureau of Reclamation in or near the storm area. this plotting will later allow the survey team to concentrate their efforts in what appear to be data deficient areas.

the itineraries of team members will be governed by the distribution of population in the storm area and the location and condition of the roads and bridges. briefly explaining the purpose of the survey. if possible. lightning. it is considered best procedure for the team to first locate the center or centers of precipitation intensity and then obtain data along lines crossing those centers. All data may be of value regardless of whether the person interviewed has actually measured the precipitation. heavy precipitation. If l l 18 . it is helpful to inform local residents of the existence and purpose of the survey and the locations where team members may be reached. The times of beginnings and endings of precipitation should be obtained from the interviews. the beginnings and endings of the separate and snow or hail should be oboccurrences. County highway maps have been found to be very useful in assigning areas to team members and in identifying the location of data points. Extensive use of news media is desirable in this effort. straight-sided. The instructions should be specific enough to preclude the omission of significant areas and the overlapping of work areas. To a large extent. Specific items for the survey team to follow are: l Bureau of Reclamation Form HD39 is intended for use in field surveys of storm precipitation. tained and entered on the form. but pleasant and tactful. Generally. the most desirable open receptacles are level. If precipitation was discontinuous or if its character varied. The team members’ approach in interviewing residents should be businesslike. and should be carefully filled out before leaving the site. The NWS cooperative observers frequently do not enter much data on their forms other than the amount of precipitation. This is also an effective check on the observation. therefore. or wind shifts should also be noted. some of the information already obtained could be mentioned during the interview. For these receptacles. In obtaining measurements of precipitation. For the purpose of interest. Emphasis must be placed on obtaining every bit of pertinent information available. Figure 2-1 shows a sample of this form that has been completed in a proper manner. etc. If container has been emptied without measurement. it is desirable to have the observer refill the container and then measure the depth in the presence of the interviewer. if the measurement was made previous to the interview. particularly near the center of the storm. In planning the survey. strong winds. The observation of thunder. flat-bottomed containers that should be left standing in an open area unaffected by buildings. The water depth is measured by inserting a thin ruler or steel tape vertically into the container. The upper part of the form is devoted to the identification of the location of the observation. a direct measurement of water depth is satisfactory. it is usually beneficial to interview them for that type of additional information.FLOOD HYDROLOGY MANUAL obtaining the rainfall data. trees.

I.1~ True Depth = g = s = 1.. Dixson T~~. 1947 - E.. the rainfall between observations may be similarly determined.tOn v = J (49+211.5 4 6 v = 33.BASIC DATA aTORN PRECIPITATIONSURVEY Be* 5/16/47 smm hiti +-ll 37.17 A = $ 8.542 4 Figure 2-I. it should be indicated whether this is a matter of observation or a reasonable assumption.. If receptacle overflowed. If several observations were made during the storm.. AT = 2 7..4+56. If receptacle is described as empty at beginning of storm. -.52 = 18.9)2. 19 . NW Vernon. -approx Colorado 3800 PRECWiT..I~~w. container was not empty at the beginning of the storm.--Storm precipitation survey form. the beginning depth should be subtracted from the total depth at the end of the storm.83 in.. Marvin Tuell 46 II sulk28 W. the time at spillover should be noted. 103-D-1905.

= = = = d = volume of water in cubic inches. A. it is necessary to know the horizontal area at the bottom. cross-sectional area at half-depth in square inches. Adequate accumulation 20 . (1) must be divided of precipitation.4 Field Studies As equally conducting ical features important as the basic hydrologic and meteorologic data in a flood study is the accumulation of data regarding the physof the drainage basin being studied. cross-sectional area at water surface in square inches. Reconnaissance of Drainage Basins for Flood Hydrology l 2. For receptacles having a rolled edge. The basic data usually required are height. the most direct method of determining the depth is to divide the volume of water (cubic inches) by the receptacle opening or catchment area (square inches) to obtain the depth (inches). the interviewers should understand that photographs cannot replace the detailed sketches with dimensions of the receptacles. In cases where volume or weight cannot be readily determined. The volume can then be computed: l (1) where: V A. An obstacle may exert effects on precipitation for a horizontal distance two or three times the height of the receptacle. at surface of water. However. Liberal use of photographs cannot be overemphasized to add to the written descriptions. For this method. and location of receptacle with respect to obstacles and other potential shelters. bottom dimension. cross-sectional area at bottom in square inches.FLOOD HYDROLOGY MANUAL l For precipitation caught in irregular-shaped receptacles.. top dimension. shape. by the and The area l volume V in equation to get the true depth catchment Presentations in the supplemental precipitation survey report should include a sketch of the receptacle showing dimensions. depth of water in inches. and depth of water. use the prismoidal formula to find the measured volume of water in containers with sloping sides. and amount of tilt. the catchment area is taken as extending to the ridge of the roll. A. direction. Each team member should carry a steel tape or rule and appropriate graduates or measures. and at catchment level. at half-depth of water.

including existing water control facilities and existing and potential development. This includes any reconnaissance conducted at the appraisal level because these estimates are often used for later studies.. When these parameter agreements are made at an early stage. pertinent physical features of the basin that may have an impact on the magnitude and timing of flood runoff. the time for review and approval of the study by the Denver Office is shortened because the necessity for technical revisions will essentially be eliminated. (a) General-The field reconnaissance should be conducted prior to the initiation of any flood hydrology study. in turn. At least one of the participants should be a senior level flood hydrologist with extensive experience in this specialty area. the time to complete the study is shortened due to the elimination of uncertainties as to acceptable parameter selection that may arise in the course of conducting the flood study. representatives from the hydrology staffs of the project and regional offices and the Denver Office’s Flood Section should participate in the field reconnaissance.-In the interest of familiarizing all levels within the Bureau regarding a basin’s hydrologic characteristics. simultaneous.-Partic(1) Drainage network or “hydraulic ular emphasis should be placed on observing and documenting the hydraulic roughness characteristics of the drainage network of the basin. meander 21 . Studies conducted solely on the basis of office use of USGS quadrangle maps without physically visiting and inspecting the basin have led to many erroneous hydrologic analyses. (c) Field Observations.-The field reconnaissance team will observe and document the following four primary characteristics of the drainage basin: system” of drainage basin. (b) Participation in Field Reconnaissance. all critical hydrologic parameters to be used in the flood analysis can be agreed upon in the field where the actual conditions are immediately apparent. The advantage of this wide representation lies in the fact that based on direct. These observations should specifically consider overbank flow because these reaches are likely to be less hydraulically efficient than the main channel. The characteristics are most readily determined by visually inspecting representative reaches of the network and assigning average Manning’s n values to each reach. in a trip report. and are not upgraded due to lack of time or funds. Also. The purpose of the reconnaissance is to identify and document. observations. These characteristics have a direct impact on the drainage basin’s hydraulic efficiency and. on the selection of unit hydrograph parameters discussed in chapter 4.BASIC DATA of these data can only be accomplished by hydrologic engineers conducting a field reconnaissance of the basin. It should be kept in mind that the n values assigned are to reflect water levels or stages that would be expected during extreme flood conditions.

and the material comprising the channel bed (boulders. The description should include a discussion on the type of channel (swale. the overbank areas. such as when analyzing basins of very large area1 extent like the Colorado River Basin. and should be appropriately included and referenced in the narrative portion of that report. A complete description of the channels and overbank areas as related to hydraulic conditions should be included in the trip report. The n values selected and the stream reaches should be called out and delineated on the maps used in the field reconnaissance. The density of the well defined channels comprising the drainage network should be observed and described in the field reconnaissance report. It also provides excellent colored photographs of the measuring sections and associated channel reaches. The description should be tied directly to available USGS quadrangle maps used in the field work. should always be included as supplementary information in each reconnaissance report. scour potential. these values will be averaged and will form the basis for selecting appropriate coefficients for the general unit hydrograph lag equation. there are cases where this size of basin will be exceeded. bare rock. This description will necessarily be somewhat subjective. Subdivision is generally required to properly apply the unit hydrograph approach when converting rainfall or snowmelt into a runoff 22 . etc. well incised. meandering. This document provides Manning’s n values. cobbles. straight. In most cases. which results in a more rapid runoff. An excellent guide for the selection of Manning’s n values is the USGS Water Supply Paper 1849. Agreement should be reached during the field reconnaissance as to the subbasin breakdown to be used in the study. the denser the drainage network. that have been determined from measurements during and after major flood events. This type of flow occurs in those portions of any basin where runoff must flow generally in “sheets” prior to reaching a point where the flow becomes concentrated in a channel or swale. Generally. The report should also include information relative to the extent of overland flow. the more limited the distance overland flow must travel. Roughness Characteristics of Natural Channels [ 171. Photographs.). This information is useful for reference in supporting the results and conclusions that will have to be made at a later time when study participants may not be available.). if appropriate. but can be enhanced by the inclusion of representative aerial photographs. etc. As will be discussed in chapter 4. It is desirable that these photographs be taken from bridge crossings or at bends in the channel to show both channel and overbank areas. etc. native soil. for a variety of natural channel and overbank conditions. and the time of year the flood is likely to occur. However. if available. bedrock.FLOOD HYDROLOGY MANUAL cut-off that shortens the length of travel of flood waters. character of overbank area (heavily wooded. preferably in color. this breakdown is required for basins larger than about 500 square miles and for those basins where significant topographic variations are present. grass covered.) and.

BASIC DATA hydrograph. Systematic observations and adequate documentation of these observations will provide the basis and support for selecting appropriate minimum or ultimate infiltration rates commonly used in the development of probable maximum and other flood hydrographs. delineated on a map. Topographic yariations usually occur when both mountainous areas and flat valley or! outwash plains are present in the basin and each has a significantly different response to a given amount of rainfall. If such features are present. (4) Land use. particularly in the channels and overbank areas. In many areas of the United States underlain with limestone beds.-Drainage basins tributary to Bureau dams are often natural or undeveloped basins. Ground observations supplemented with aerial photographs have been found to be the most satisfactory way to accomplish this task. depressions in the land surface have developed. If this\ is the case and the basin is expected to remain undeveloped. during the field inspection. (3) Vegetal cover.-Another factor important to the satisfactory selection of estimated infiltration loss rates and unit hydrograph parameters is a general knowledge of the vegetal cover of the drainage basin. The geologic setting will. the description of the geologic setting will have to be researched and extracted from the technical literature. This description should be supplemented with color photographs as appropriate. The general geologic setting. should be described in the report. have an impact on the selection of infiltration loss rates. These depressions. it should be so stated in the field reconnaissance 23 . Such literature is generally available from other related Bureau studies or USGS libraries. in many cases. Therefore. it is necessary to observe and document the types of vegetation present in the basin. and should be discussed in the trip report.in a drainage basin. Because the reconnaissance team members will probably have only limited backgrounds in geology. such features should be fully described in quantitative terms in the field reconnaissance report. with the inclusion of color photographs. The results should be delineated on the USGS quadrangle map used in the field inspection. as it relates to runoff. and the area1 extent and location of each type. Also.” usually result in areas that will not contribute to runoff from a drainage basin. called “sink holes” or “playas. they can have a significant effect on the flood runoff that can be expected because they will act as small reservoirs either detaining or storing runoff during flood events. and assessed as to their floodwater storage capability. The soils observed to be present in the drainage basin should be classified as to type using the four general SCS types discussed in chapter 4. Therefore. it is of prime importance that such areas be identified. (2) Soil and geo1 g-t conditions.-Soil o ‘c conditions should be observed and documented on a suitable map in terms of types of soil comprising the drainage basin and the area1 extent of each type.



report. However, there are an increasing number of situations where all or a portion of the basin will be used for agricultural purposes, including both crop production and livestock grazing; forestry, such as the tree harvesting that is prevalent in the Pacific Northwest region of the country; and urban development. In the case of agricultural and forestry land uses, the type, extent, and intensity of such uses should be determined during the field inspection and properly documented in the field report. Regarding the degree of urbanization, only development that presently exists can be observed and documented. Since dams will be in existence for an extended time, estimates of projected urbanization must be obtained along with an estimate of the storm drainage facilities that will be constructed in connection with the urbanization. Therefore, when conducting the field reconnaissance of a basin that includes or is near an expanding urban center, the local government entity responsible for land use planning and zoning should be contacted and a projected land use map secured for use in the flood study. Knowledge of projected urban land use is of considerable importance because the rainfall-runoff response of an urbanized drainage basin is very likely to be entirely different from that of the same basin in a nonurbanized condition. For example, in a relatively flat area in central Texas, the peak discharge from essentially the same rainfall from a particular basin was found to increase by a factor of almost 8 after being completely urbanized. 2.5 Examining corded Floods Nearby Basins That Have Experienced Significant Re-

When the route of travel to or from a specific basin being inspected passes near a basin where a significant flood event has been recorded, time should always be allowed for a reconnaissance of that basin. Observations of the types previously discussed should be made and documented in the reconnaissance report. The documentation of these observations may serve as a basis for conforming hydrologic parameters used in the flood study for the drainage basin being analyzed, or for other ungauged basins within the hydrologically homogeneous region. Particular emphasis should be placed on conducting field reconnaissances of the basins described in chapter 4 to develop the general unit hydrograph lag time relationships for the six regional categories.


3.1 General Considerations and Background

The “design storm” is the estimate of the rainfall amount and its distribution over a particular drainage basin, with respect to both time and area, that is used in the development of a PMF hydrograph or; in some instances, a flood hydrograph representing a specific frequency event. The purpose of this chapter is to provide the Bureau hydrologic engineer a basic familiarity with the atmospheric processes that produce extreme precipitation events, and the methods by which these processes are used in computing individual drainage PMS (Probable Maximum Storm) values and in developing regionalized criteria used to compute such values for particular drainage basins. In 1981, the Bureau, NWS, and COE adopted a mutually acceptable, uniform definition of the widely used term PMP (Probable Maximum Precipitation) as “theoretically, the greatest depth of precipitation for a given duration that is physically possible over a given size storm area at a particular geographical location at a certain time of the year.” Complete technical application of this definition results primarily in a storm isohyetal pattern; isohyets are lines of equal precipitation that are analogous to contour lines on a topographic map. This pattern is overlaid and critically centered on a drawing of the drainage basin outline, and then the average precipitation is computed. The average basin precipitation is then distributed over time, which results in incremental basin average precipitation values that are used to generate the PMF hydrograph discussed in chapter 4. The average basin precipitation is highly dependent on the total understanding of atmospheric processes as to the cause of severe event precipitation and the analysis of accumulated basic storm data. As hydrometeorologists expand their knowledge of severe storm meteorology, future revisions to present PMP estimates can be expected. However, at least for the conterminous United States, only minimum modification to current values of PMP is expected in the foreseeable future because knowledge of severe storm phenomena has reached a plateau. Some discussion among meteorologists has centered on the effect of general climatic variations on estimates of PMP. Climatic trends are not incorporated into estimates of PMP for the following two reasons: (1) such trends progress rather slowly or fluctuate to an extent that their effect on PMP is small, especially in relationship to the other uncertainties associated with PMP development; and (2) effect of climatic variation on




the production by members

of extreme precipitation has not of the meteorological community.





The terms PMP and PMS have often been used interchangeably. For certain studies, these terms are used synonymously; e.g., those studies concerned with small area, short duration rainfalls. However, for the majority of studies, a precise distinction is realized. By definition, PMP refers to the “maximum precipitation amount for any area size and duration at a particular location during a certain time of the year.” The PMS results from adjustments applied to the PMP that provide a total design storm event, realistically patterned (spatially-temporally) after actual storms of record, and that has been adjusted to a single PMP area size and duration deemed critical for design. An example of such adjustments is the use of within/without storm DAD (depth-area-duration) relationships as described in the PMP applications report HMR NO. 52 [IS]‘. It should be noted that these adjustments will tend to result in a lower average basin rainfall value. The adjustments are negligible for very small basins and significant for very large basins.

Values of PMP are called “estimates” because there is no direct means of computing and evaluating the accuracy of the results. Judgments based on meteorological principles and storm data are the most important factors in assessing the level of PMP. The lower limit of PMP is set by observed storms of record. In evaluating the appropriate level of PMP, it is necessary to account for such factors as the length of record available, number of severe storms that have occurred, degree of envelopment of PMP over observed data, transposition limits to severe storms of record, maximizing steps incorporated, validity of models used for assessing PMP, and care that excessive compounding of the meteorological factors associated with rare events has been avoided. The evaluation is often complicated because of the uniqueness of the problem presented. No two basins provide exactly the same problems, and the available data base of severe storms is hardly sufficient for most regions. Because of these variable conditions, a “cookbook” approach to PMP determination and its evaluation is not available. Hydrometeorologists charged with the task of determining PMP must appropriately weigh what has been previously derived for past specific studies with what their combined meteorological judgment provides. As our knowledge of the severe storm process expands, it is expected that present-day estimates of PMP will be refined. The Bureau’s current approach in determining PMP is the adoption of material obtained in various hydrometeorological reports issued by the Hydrometeorological Branch of the NWS. Although modifications to these reports are sometimes necessary, it has been determined by Bureau hydrometeorologists that these reports provide the most accurate and consistent method of PMP derivation. For these reasons, PMP derived using the HMR series is the method used by the Bureau where applicable.
‘Numbers in brackets refer to entries in the Bibliography.



3.2 Atmospheric


To more fully understand the derivation of PMP, some knowledge of the precipitation process should be useful. Basically, four conditions are necessary to form intense precipitation: (1) abundant atmospheric moisture, (2) a lifting mechanism, (3) condensation, and (4) water droplet-ice crystal growth.

(a) Atmospheric Moisture.-The term “atmospheric moisture” refers to water vapor only and not to water in either its liquid or solid state. Water vapor in the atmosphere is commonly measured in the United States in terms of inches of precipitable water. Precipitable water is defined as “the height of condensed total atmospheric water vapor contained in a vertical column of air of unit cross section located between any two levels of the atmosphere” [ 191. It should be noted that, in severe storms, precipitation often exceeds the measured precipitable water depth of the total overlying atmosphere. This situation is explained by the fact that, during the storm event, additional moisture is constantly being fed into the area due to the nature of the storm circulation. Measurements of atmospheric moisture through depth are typically made by radiosonde and/or weather satellite related observations.
Water vapor is always present in the atmosphere, with the maximum amount limited by air temperature. Air becomes saturated when it contains the maximum amount of water vapor at a particular temperature. Saturation is achieved by either the addition of moisture or the cooling of air to a lower temperature, where the maximum amount of water vapor is held and condensation begins to occur. Although both processes work simultaneously during the formation of precipitation, it is the latter that has the greatest effect in producing intense precipitation. The temperature to which a given parcel of air must be cooled at constant pressure and constant water-vapor content for saturation to occur is called the “dewpoint”. Air at a temperature near the dewpoint temperature is likely to produce condensation particles.

(b) Lifting Mechanism. -Precipitation is basically caused by the cooling of moisture-laden air to a temperature at which the atmosphere cannot retain its moisture charge (saturation). To produce intense precipitation, nature achieves the required rapid cooling by air expanding as it rises or is lifted into the atmosphere. The total process that causes this upward forcing of air and resultant cooling is termed the “storm lifting mechanism.”
Four separate storm lifting mechanisms can be distinguished that provide the necessary lift for large quantities of moist air [20]. During actual storm occurrences, from one to all four mechanisms may interact to produce resultant saturated air and eventual precipitation. These four lifting processes are briefly described as follows:




(1) Atmospheric convergence.-Lifting by this mechanism results from the interaction of atmospheric pressure and circulation to concentrate moist air. The convergence that takes place near the Earth’s surface forces the air upward, where expansion and cooling take place. Precipitation formed mainly through this process can be intense over large areas and for long durations. Such precipitation is often related with extratropical (storms that originate outside the region of tropical easterlies) and tropical storm types that are readily interpreted on a synoptic (large area) scale of analysis. (2) Orographic li@ng.-Wh en moist air flows up a range of hills or mountains, orographically induced precipitation may occur. The rate and quantity of air cooled is dependent on the steepness and orientation of the barrier encountered relative to the direction and magnitude of the inflowing moist air mass. (S)Fronts.-A front is described meteorologically as the interface between two air masses of different density. An air mass is a large body of air that is almost homogeneous in horizontal direction, vertical temperature and moisture variation are the same over its horizontal extent [19]. Air masses become established while situated over a particular region of the Earth’s surface for prolonged periods. The surface of contact between air masses is an inclined surface with the lighter warmer air above and heavier colder air below. When a colder air mass is displacing warmer air; i.e., when front is moving in direction from cold air to warm, a cold front results. A warm frontal surface is one in which warm air is displacing cold, and this surface has a gentle slope as the less dense warm air smoothly overruns the retreating cold air mass. Because of earth surface frictional forces, a cold front is steeper in the lower atmospheric levels than that established by a warm front. Because of the shallow surface presented, neither a cold nor warm front can produce significant precipitation. Frequently, a frontal wave is developed along a front when the movement of a portion of the front is retarded and a counterclockwise wind circulation (northern hemisphere) develops. This occurs most frequently when a low pressure trough in the upper atmosphere, commonly termed a “short overtakes the surface frontal system. Under favorable wave”, conditions, usually in concert with one or more of the other storm lifting mechanisms, a wave may develop rapidly and cause significant precipitation. Storms developed in this manner occur mainly in temperate climates and are known as “extratropical cyclones.” A series of waves may be generated when a front does not indicate significant motion toward either the cold- or warm-air mass. Such frontal positioning is known as a “stationary or quasi-stationary front”. A series of small waves, generated under these conditions, may successively cause precipitation over a given area and may even produce a major storm


Meteorologists predominantly refer to convection as “mixing of atmospheric properties in the vertical direction.-An extremely important concept in the production of extreme precipitation is that of convection/instability. as a major (4) Convection /Instability. Additional latent heat of condensation must be removed from a parcel of saturated air. Forced convection can be demonstrated by the previously mentioned three storm lifting processes of atmospheric convergence.e. derived precipitation can be intense over large areas for long durations. maintaining saturation through substantial depth often producing intense precipitation over small areas for short durations. resulting in showers and thunderstorms. An excellent example of convective activity is the development of local storm (thunderstorm) precipitation. A parcel of air that has reached saturation will continue to cool with increasing elevation at a rate less than unsaturated air.-When water vapor is transformed into a liquid or directly into a solid (sublimation). the parcel will continue to rise and therefore eventually cool to its saturation temperature. and condensation and/or freezing nuclei must be available in sufficient quantity before droplets of water or ice crystals are 29 . which can be defined as “a condition that if a parcel of air is less dense than its surrounding environment.” Local storms are typically generated by the heating of air near the Earth’s surface. Combined with the previously mentioned storm lifting mechanisms. the process is called “condensation. Free convection arises from atmospheric instability. Atmospheric instability is present to some degree during any severe storm event.” i.. it will become buoyant. fog.” The formation of dew. and achieving saturated air does not necessarily mean condensation will occur. If atmospheric instability is present through significant depth. A local storm is defined as “an isolated precipitation event restricted in both spatial and temporal distributions. which causes explosive cloud development and major precipitation. a parcel of saturated air in an unsaturated environment will remain warmer than its surrounding air with increasing height and continue to rise. (c) Condensation. rapid vertical cloud growth takes place. Other storm lifting mechanisms often provide the initial impetus (lift) to place moisture-laden air into a unstable environment. Therefore.” A further distinction is made between forced and free convection.HYDROMETEOROLOGY ~ even though none of the waves could be regarded individually storm system. orographic. and clouds are typical visual results. and fronts. Local storms involving widespread activity of this type are of maximum frequency and magnitude during the summer when maritime tropical air masses occur frequently and when the heating of air near the ground is sufficient to initiate convective activity. This results from the release of latent heat due to condensation.

Such airborne particles are dust. and combustion nuclei. Such growth continues until the ice crystals become large enough to fall. duration. Theories explaining such growth are found in most meteorological texts concerning cloud physics. both ice crystals and liquid water droplets exist together in a supercooled cloud. Further growth occurs as the ice crystals collide with other crystals (aggregation) and with supercooled cloud droplets (riming). 3. another process must occur to further the growth of the initially formed droplets/crystals. In summary. in deep storm clouds. and (3) the rate of growth of cloud droplets/ice crystals to form precipitation-size hydrometeors. the equilibrium vapor pressure of water vapor to ice is less than that to liquid water. and area1 extent of the PMP. Condensation nuclei are hygroscopic particles that act as a catalyst upon which water vapor condenses. (d) Water Droplet-Ice Crystal Growth. However. At low latitudes (tropics) where rainfall frequently occurs from warm clouds. only the collision-coalescence theory offers a reasonable explanation for cloud droplet growth. The two most advanced explanations are the Bergeron-Findeisen theory (icecrystal) for cold cloud (temperature less than 0°C) and the collisioncoalescence theory for warm cloud (temperature greater than O’C) particle growth. and ice crystal growth will then occur at the expense of the water droplets. however. the most critical element is the calculation of the magnitude. precipitation finally occurs. (2) rate of cooling (updraft) produced by lifting mechanisms. The collision-coalescence theory is the merging of two or more water droplets into a single larger droplet or raindrop.-Condensation of water vapor to liquid droplets or ice crystals results in particles of such size that only the slightest updraft is required to suspend them in the atmosphere.3 Derivation of PMS In the development of the PMS. sea salt from evaporated sea spray. In the Bergeron-Findeisen theory. the quantity of precipitation formed is dependent upon the following three primary factors: (1) amount and rate of inflowing moist air into the storm area. To overcome the substantial updrafts that are associated with extreme precipitation events. size and concentration of drops. the ice crystal processes are valid at low latitudes. Initial contact occurs from impact. At some subfreezing temperature.FLOOD HYDROLOGY MANUAL formed. whether the coalescence activity is completed and a single large drop is formed depends on the relative velocity of impact. It is entirely possible at‘ mid and high latitudes for one or both of the theories to form precipitation. In the 30 . and the electric charge of individual drops and surrounding field. Once sufficient raindrop mass is obtained so that gravity overcomes the storm updrafts.

HYDROMETEOROLOGY deterministic approach to evaluating PMP. the greatest precipitation potential for that particular storm event is assumed to be determined. the Canadian Atmospheric Environment Service has prepared similar analyses of 400 storms in Canada [25]. two methods are generally approach. A set of similar but unofficial data for about 300 additional storms analyzed by the Flood Section of the Bureau. For any given storm. The DAD values represent the average depth of precipitation that has occurred over an area of given size within a specified time interval. aspects from both approaches are combined to produce appropriate results. The storm maximization approach uses the information described in section 3.-Section 3.2 indicated that the amount of precipitation formed is basically controlled by the rate at which available moist air can be processed by a storm. the solution is to use the record of extreme precipitation amounts as an indirect measure of the maximum potential of these conditions. The reasoning supporting these assumptions is twofold. (1) Data base. or found in the literature [24]. For the United States. and envelopment performed upon this data base are given in the following paragraphs. Brief descriptions of the data base and operations of storm moisture maximization. The first method uses the “storm maximization” and the second uses the “storm model” approach. Second. and has been accumulated in an open-ended publication [23]. 31 .2 for the formation of precipitation except for atmospheric moisture. By adjusting observed severe precipitation amounts to maximum moisture conditions. studies have shown that the amount of precipitation obtained is directly related to the quantity of atmospheric moisture available. The lack of a sufficient storm data base at individual locations is compensated for by the introduction of storm transposition and envelopment to achieve the level of PMP. Hydrometeorological Branch of NWS. supplements the basic data in reference [23].2. applicable. First. these values represent the highest average depth for selected area sizes and durations. excluding moisture content. DAD type data has been assembled for more than 500 such storms. For drainages situated near the Canadian border and the United States. (a) Storm Maximization Approach. due to the lack of sufficient measurements during actual storm events to quantify each of the conditions causing precipitation. and (2) that the record of severe storms is sufficiently large that an optimum “storm mechanism” has been realized. Occasionally.-The basic data are the records of the largest known observed areal-duration precipitation amounts. This storm mechanism includes all the conditions described in section 3. These data are developed by the standardized DAD procedures described in [21]. and is based on the following two assumptions: (1) precipitation can be determined from the product of available moisture and storm mechanism. transposition.

The actual moisture maximization factor formed is the ratio of maximum precipitable water to water available during the storm. The moisture charge in an air mass is expressed in terms of precipitable water. The sea level dewpoint is read at the intersection of the curve and the zero-elevation line (sea level). and is represented by the solid sloping lines on figure 3-l. The standard is to adjust all surface dewpoints to a pressure of 1. plus the density of stations taking such observations is large enough to sense narrow tongues of inflowing moisture often critical to small area severe storm development [26]. i. dewpoints are chosen as the highest value persisting for 12 hours. tables of precipitable water have been prepared 32 . Tests performed with data from severe storms of record indicate adequate support for this assumption [20]. In the storm maximization process. the surface dewpoint temperature. a point corresponding to the elevation (ordinate) and observed dewpoint (abscissa) is plotted.FLOOD HYDROLOGY MANUAL (2) Moisture maximizution.000 mbar. that the rainfall was closely proportional to the moisture charge available for each model [20]. on a diverse set of storm models. a dewpoint of 58°F at 5. Such a process is called pseudoadiabatic (saturation-adiabatic). and then compressed adiabatically to 1. Initially. Surface dewpoints. Moisture maximized precipitation has the same ratio to the observed storm precipitation as the maximum moisture charge has to the moisture charge of the observed storm. instead of data collected by radiosonde. to basically normalize dewpoints for differences in station elevations.. based on associated dewpoint temperatures and a saturated atmosphere. a curve is then traced through that point parallel to the adjacent pseudoadiabats. In the determination of a storm moisture maximization factor. with moisture supplied to keep the air saturated during compression.000 feet is equivalent to a dewpoint of 70°F at sea level. Figure 3-l is used in converting an observed dewpoint to its sea level equivalent. It is believed that this time period is more representative of inflow necessary to establish severe precipitation.-Moisture maximization is a process whereby observed storm precipitation is increased to a value consistent with the maximum moisture in the atmosphere for the storm location and time of year. in conjunction with an assumed saturated atmosphere above surface level. is used as an index of available precipitable water. as well as reducing the error of instantaneous observations. For a saturated atmosphere. Dewpoints are comparable only if they represent conditions under the same atmospheric pressure.000 mbar (millibars).e. The adjusted sea level dewpoint is the temperature that a parcel of air would achieve if cooled to the dewpoint at observed pressure. For example. It has been mathematically shown. are used as a measure of available moisture because they are the only measure of moisture potential for early severe storms of record (prior to 1940). which is approximately sea level pressure.

which is a measure of the moisture inflow in an actual storm. The first is the “representative storm dewpoint”.27 inches. over the same time period.000 mbar and persisting 33 .HYDROMETEOROLOGY 2. air temperature is equivalent to dewpoint (saturation) temperature. From this intersection.OOO-mbar dewpoints [27]. It is necessary to record both distance and direction of the selected dewpoint location in relation to the precipitation center. scale along bottom of page) value of 2. scale along top of page). Representative storm dewpoints are selected in the moist air flowing into the precipitation area.000 feet) having a sea level dewpoint of 70°F. The second is the maximum dewpoint for the same location and time of year when the storm occurred.000 1. proceed vertically downward to read precipitable water (abscissa. For a particular storm. Since saturation is assumed. Locate the point on figure 3-3 at the intersection of the 208mbar surface (ordinate) and the 70°F temperature (abscissa. Details of guidelines used in the selection of representative storm dewpoints are available in several publications [29. assume it is desired to determine the total precipitable water in a column of saturated air from sea level pressure (0 feet) to the 200-mbar level (about 40. each reduced to 1. Figures 3-2 and 3-3 show the results of these tables in nomogram form [28]. 103-D-1906. As an example using these figures. Determination of two dewpoints is necessary for moisture maximization calculations. standard practice is to average dewpoints from several stations. providing such measurements based on l.000 0 50 52 54 56 56 60 62 64 66 IN 66 OF 70 72 74 76 78 80 TEMPERATURE Figure 3-l.31].-Pseudo-adiabatic chart for dewpoint-elevation adjustment.30.

S Weather eurew SurfaceTemperatures HydrOrn. 23 Rate for the indicated from “. 103-D-1907.tm01oplml Figure J-Z. water determination from 1. 34 .FLOOD HYDROLOGY MANUAL DEPTHS OF PREGIPITABLE WATER IN A COLUMN OF AIR OF GIVEN HEIGHT ABOVE 1000 MILLIBARS Assumiw A*Opl*d Saturation with o Pseudo Adiabutic Lapse Rep-x+ NO.000 to 700 millibars.-Diagram for precipitable From [28].

Diagram for precipitable From [28]. 103-D-1908..~+her with a Pseudo-Adiabatic Surface Temperature BUWJU Hydrom~twxologlcal TEWPERAWRE Lapse Report NO. 35 .S. water determination from 800 to 200 millibars.rom Saturation U. w. 23 for the lndicutrd 8( PRECIPITABLE WATER IN INCHES Figure 3-3..HYDROMETEOROLOGY 1 DEPTHS OF PRECIPITABLE WATER IN A COLUMN OF AIR OF GIVEN HEIGHT ABOVE 1000 MILLIBARS Assuming Rate ~d.a~+ed .

The exact location of the maximum dewpoint used in the moisture maximization process is the same location used for the representative storm dewpoint. Charts of the largest persisting 12hour. = observed precipitation. This procedure is applied to both representative storm and maximum dewpoint related calculations of available precipitable water before determining the storm maximization factor. from the precipitable water in the total column for the same dewpoint from sea level to the 200-mbar level. Wp (max. The basic storm maximization process can be expressed mathematically as: P. The selection of the maximum dewpoint is normally taken 15 days into the warm season (higher dewpoint) from the date of the representative storm dewpoint occurrence. therefore. subtract the precipitable water in the column of air extending from sea level to the height of the barrier or precipitation elevation. a reduction to the available precipitable water above sea level due to the barrier or elevation is necessary. to maximum Wp and representative storm Wt. Maximum dewpoints typically express the highest dewpoint observed for a particular location and time of year that occurs in the vicinity of associated rainfall potential. To determine the moisture charge available for a drainage basin at a given dewpoint under such circumstances. dependent on the region of interest. storm refers to 36 . further moisture adjustments are applied. = We (storm) (2) where: P. This method accounts for the assumption that a critical storm could reasonably have occurred 15 days before or after its storm date without appreciably affecting the storm mechanism. and Wp = precipitable water (max.OOO-mbar observed dewpoints that are regionally and seasonally smoothed to provide a maximum analysis are found in several publications [32. P. Under these conditions.33. 1 .FLOOD HYDROLOGY MANUAL for 12 hours as representative of individual storm moisture. the distance and direction of the selected representative dewpoint must be compatible with observed lower level winds and time of dewpoint selection.. Additional increases of up to 10 percent in adjusted storm precipitation are often realized from application of this procedure..) PC. A representative storm dewpoint is chosen upwind of the precipitation area. = adjusted moisture maximized precipitation.34]. whichever is greater. If a mountain barrier interrupts air flow from the moisture source to the precipitation region or the precipitation occurs at an elevation different from sea level. A distance is selected that allows the greatest moisture content available to be processed during the storm.

for the transposed location to the maximum available for the storm in place. and is applied only if meteorological or topographical explanations of potential discontinuities cannot be postulated. The moisture maximization and transposition adjustments can be expressed mathematically.35. transposition becomes an important tool for providing additional data at a particular site. and then determining what atmospheric processes were at work to produce the event.3 (a)(2). two types of storm transposition are used: (1) “explicit” transposition. As each drainage basin of interest has not likely experienced the number of severe storms necessary for PMP development. In deriving the PMP. Details of storm transposition procedures are available in several publications [29. see section 3. Adjustments based on moisture availability are normally the only modifications permitted within the explicit transposition region. -Transposition involves relocating individual storm precipitation within a region considered homogeneous relative to topographic and meteorologic characteristics deemed significant to that storm.36. Limits of transposition are then established. Storm transposition begins by clearly identifying the location of the precipitation. The typical storm transposition adjustment is determined by the ratio of maximum precipitable water. The maximum transposed dewpoint is selected using the same distance and direction from the transposed storm location as was determined for the selection of the representative storm dewpoint.31. which leads to a classification of the particular storm type. Meteorological records are then researched and surrounding terrain features examined to identify similar regions where the storm could reasonably be transposed.HYDROMETEOROLOGY (3) Storm transposition. The final step in this procedure is the application of adjustments within the explicit transposition limits for the specific relocation of the storm.” An exception to this definition would be the distance from the coast adjustment for tropical storms. so only general comments concerning these procedures will be made in this manual. Knowing that nature would not typically permit atmospheric discontinuities to occur at limiting boundaries to explicit storm transposition. and (2) “implicit” transposition. A complete meteorological and topographical analysis is made by the hydrometeorologist.37]. implicit transposition (or regional smoothing) is applied. Implicit transposition extends the explicit limits beyond what would normally be determined. and simplified to produce the moisture maximized-transposed precipitation as follows: Adjusted moisture maximized and transposed precipitation 37 . which is usually defined as “the region where a particular storm may be transposed without experiencing significant change to its storm mechanism.

transposed) = (PO) where: W. In the development of regionalized studies. Information taken from these figures are eventually combmed shown on figure 3-6. (4) Enuelo merit. transposition. additional regional smoothing of a number of individual analysis.-Envelopment involves the selection of the likely greatest value Prom a set of data. similar to that shown on figure 3-6. recipitation. respectively. and decreases in elevation or barrier considerations result in associated increases to adjusted precipitation.3 (c)(2). and envelopment performed on observed precipitation are the basis for developing the PMP using the “storm maximization 38 . Wr (max.3 (c)(l). For individual dramage analysis.35. note that the same storm does not control the enveloping curves for all the area sizes and durations indicated. pat= (PO) in place) I( Wp (max. indirect re ional smoothing Keve1 of PMP has methods are often used to ensure that the appropriate been achieved and that consistency has been maintained with adjacent basins.FLOOD HYDROLOGY MANUAL Wp (max. section 3. Also. Increases in elevation or barrier considerations result in associated decreases to adjusted precipitation. of individual moisture-maximized and transposed-adjusted precipitation. necessarily acThe adjusted counts for dif Perences in horizontal or vertical (barrier) effects due to moisture maximization and transposition of observed precipitation to other locations.. location.38]. (storm) moisture maximized-transposed precipitation.-The three processes of storm moisture maximization. In line with the defin a corn lete DAD analysis. tt should be noted from figure 3-6 that the eninition o P envelopment. Figures 3-4 and 3-5 are examples of depth-duration and depth-area smoothing. P.>(max. Inconsistencies in PMP estimates within or between regions stud nee cl to be meteorologically or topographically justified. or such anomalies smoothed. = observed We = precipitab posed recipitation. and location of interest. For additional transposition adjustments that may ap ly for special circumstances dependent on storm type. area. and storm refers to representative storm Wp). refers to maximum Wp in place or translocation. Envelopment accounts for the random occurrence and variation of individual severe storm precipitation. = adjusted P. veloping lines do not necessarily pass through every moisture-maximized and transposed-precipitation amount.. or avat P able data. see references [29. This step becomes a requirement due to the lack of a uniform storm data base for every duration. section 3. and horizontal transposition away from the moisture source causes related decreases to the adjusted precipitation. transposed) in place) Wp (storm) W. are performed for the entire study area. (5) Summary. as expressed in equation (3).31. Horizontal transposition closer to the moisture source causes related increases to the adjusted precipitation. 7 e water (max.

-Depth-duration envelope of transposed maximized square miles. envelope of transposed 24-hour precipitation.000 lO. storm values for 1. STORM STORM STORM STORM STORM A S c 0 E IO IO 0 0 0 I I IO I 20 20 PRECIPITATION 30 IN INCHES maximized 40 I xl - 24-HOUR Figure S-L.DDO 5. 39 .-Depth-area 103-D-1910. 103-D-1909.000 In 4ooo IA 1 500 I i z z _ 2 a so- 7 - lOOr 0 x A .HYDROMETEOROLOGY 0 12 24 36 DURATION IN 48 HOURS 60 72 Figure 3-4.

40 .2 is hypothesized.-A second approach to a deterministic PMP evaluation is accomplished using storm models.000k 5.39. technique. section 3. the various parameters attributed to the storm mechanism and moisture inflow are then maximized in an attempt to realize the full PMP potential. To date. (b) Model Technique. 103-D-191 1.” Elements of these processes are also incorporated in the “storm model” approach. Limited use of an orographic model is described in section 3. Calibration of the model is obtained by varying.000 c I k! -=l 2 z Q E -a 1.FLOOD HYDROLOGY MANUAL I I I I I I0.38.-Depth-area-duration envelope.40]. Having replicated precipitation from observed storms. the processes describing the storm mechanism and moisture availability until satisfactory comparisons are achieved with actual severe precipitation events of record. Individual reports previously referenced should be consulted for details.000 500 100 50 2 a O 0 I IO I 20 PRECIPITATION I 30 IN INCHES I 40 I 50 I 1 Figure 3-6. mathematically.3 (b). where a model describing the various atmospheric processes of section 3.40]. and only limited success using an orographic model has been attained [39.3(c)(2). Modifications to these processes are occasionally required. dependent on the study location and data available. satisfactory models describing PMP potential have not been fully realized. Various difficulties encountered using models are discussed in several publications [29.

(c) Individual Drainage Estimates and Regionalized Studies.HYDROMETEOROLOGY It is generally accepted. The Bureau’s alized reports a professtonal Denver Office.--The approaches to PMS development of storm maximization or modeling techniques can be applied to either an individual basin or to a large region that contains a multitude of drainages of varying size and shape. Regionalized PMS criteria [ lS. (3) consistency among individual basin estimates is obtained. and (3) the scale of analysis is such that minor refinements are not incorporated because of the smoothing involved.44.38. the term “regionalized” or “generalized” approach is used for the PMS evaluation. The final result is obtained using appropriate figures.42. and (5) regionalization serves as a base of severe storm information and criteria to further develop individual drainage study requirements for specific locations when additional information becomes available. With few exceptions. and equations for which values of the PMS are obtained for any drainage located within the study area and within the limits (durational-areal) of the regionalized report. The primary disadvantages of regionalized studies are: (1) time required to complete and document studies often take several years. that use of such models is the preferred method.31. is always conducted by hydrometeorologist in the Flood Section of the Bureau’s or through consulting meteorologists in conjunction with 41 . the regionalized approach as set forth in the HMR report series is to be used in determining PMP and PMS values for PMF development. an area of similar topographical and meteorological features is defined and the procedures of storm maximization and/or model technique are applied to portray the PMS in generic form for the entire region.29. (2) storm maximization or model techniques provide a greater degree of reliability to the PMS if analyzed on a regional basis. the “individual drainage estimate” and the results are to be exclusively applied to the single drainage under study. development of the PMS. It is expected that models will be developed that will eventually simulate the severe storms of record and be adapted to provide more reliable estimates of PMP. For the regionalized approach. For a large region.39.41] are used because they possess several distinct advantages such as: (1) greatest use of available data can be incorporated. (2) extensive manpower requirements that include several hydrometeorologists with a specialty in PMS criteria development. For the individual basin. tables. among practicing hydrometeorologists concerned with severe storm precipitation events. Usually.40.35. (4) individual estimates of PMS can be readily obtained from completed regional studies by hydrologic engineers. exceptions arise when the drainage basin being studied is larger than that for which criteria are presented in the report series.45].20. unless obtained from region[ 18.31.

For orographic regions. Storms located within a homogeneous meteorological-topographical region. Using the previous assumptions and the storm moisture-maximization and transposition techniques of sections 3. The PMS estimates developed at various Bureau regional offices using approved regionalized criteria are reviewed by personnel in the Flood Section for compliance with regionalized criteria and recent agreements. Use of the storm maximization technique for an individual drainage is based on the assumption that adequate data are available describing severe historic storms that have occurred in or near the basin of interest. Basically. the isopercental technique adjusts inplace storm pattern precipitation by the ratio of the analysis of precipitation given for some specific return period and duration at the transposed site to that evaluated at the storm inplace location. (1) Zndividual drainage estimates for PMS. maps of mean annual precipitation are helpful in this adjustment.3(a)(3). a sufficient number must be considered transposable to the study basin with only limited modifications imposed. each storm’s isohyetal pattern is transposed and fitted over the basin in the most critical position. revisions. is typically used in relocating the precipitation patterns to account for differences in complex terrain features between the inplace and transposed storm locations. Care must be directed in the selection of appropriate geographical limits of transposability. Occasionally.-To date. may be assumed to have the potential to occur over the drainage area.3(a). The technique works best for large area transpositions of highly orographically controlled storms where precipitation amounts are influenced by the same parameters that caused the variation in precipitation described by the selected base map at both the inplace and transposed locations. all individual drainage estimates for PMS development computed by the Bureau have been completed using the storm maximization technique by professional meteorologists.FLOOD HYDROLOGY MANUAL hydrometeorologists on the Bureau’s staff. It is also assumed that several of these severe storms occurred having the most efficient storm mechanism.3(a)(2) and 3. or refinements. The procedure basically involves the considerations described herein. considerations to pattern orientation may apply. the storm isopercental analysis technique. and could be brought to their full precipitation potential through application of the moisture maximization procedure. and judiciously reorientated over the subject basin. as described in section 3. and is presented as background information for the flood hydrologist. individual storm precipitation patterns are directly shifted from their inplace location. as the study basin. In nonorographic regions. Of these storms of record. Basin average precipitation precipitation that would is determined by calculating have fallen over the drainage the volume area after of the 42 . often used to develop inplace storm isohyetal patterns in data deficient orographic regions.

are rearranged after a selected storm of record. Each report 43 . Average basin precipitation is maximized using the ratio of maximum precipitable water at the transposed location to that available (measured) during the actual storm at its inplace location with appropriate considerations given to intervening barriers. Timing of the precipitation in the basin follows the temporal distribution associated with the storm occurrence inplace. maximized for moisture and transposition. Typically. read from the smooth depth-duration curve. Methodologies and criteria obtained from these publications have been examined by Bureau hydrometeorologists and are considered to provide the best estimate of PMP potential within the limits of each report. the various steps in deriving the PMS for the basin should be carefully reexamined. see the list of references at the back of this manual. -Regionalized studies leading to the eventual development of the PMS are currently available in the various HMR and Technical Paper series of publications issued by the NWS. The PMS estimates derived from individual drainage analysis should be compared for consistency with estimates for similar drainages in the same homogeneous region. The incremental precipitation values from the smooth enveloping curve of individual storm adjusted data represents the maximum precipitation for the basin. Adjusted average basin precipitation. and revision of the criteria available in several of these publications.HYDROMETEOROLOGY storm was transposed. Adopted temporal distributions of incremental average basin precipitation. More consistent and reliable results are achieved using the regionalized criteria described in the following subsection. and justified or modified. for each storm transposed to the study basin is plotted and analyzed in a smooth depth-duration plot similar to the one shown on figure 3-4. Consistency is difficult to maintain when PMS estimates are completed using the individual drainage storm maximization approach when applied by different individuals at various times. elevation changes. they should be studied carefully. it may be accomplished using one or a composite of the isohyetal patterns that have occurred in or been transposed into the drainage. review. Since 1980. as developed in reports [21. The PMS derived from these reports and used in the derivation of the PMF is reviewed by Bureau personnel at the time the flood study is reviewed. Bureau personnel have participated with NWS hydrometeorologists in the development. Regionalized PMS reports available for the entire conterminous United States are regionally depicted on figure 3-7.46] concerning storm DAD analysis. 3-4) for the study basin. When appreciable differences are identified. If spatial distribution is required. the temporal and spatial distributions are obtained from one or more of the storms controlling the depth-duration enveloping curve (fig. and distance from moisture source. (2) Regionalized PMS. For those results considered suspect.

The PMP values obtained from HMR 51 are considered to be all-season estimates.-Hydrometeorological States.e. A set of regional charts 2HMR 51 indicates estimates of PMP westward to the 105th meridian.’ Application of these storm area1 precipitation estimates for a specific drainage to obtain average basin PMS values can be obtained using the procedures in HMR 52 [18]. Each storm was maximized for moisture and transposed in accordance with section 3.3(a)(2) and 3. they are the greatest values obtainable throughout the year. and quickly determine vide a synopsis of the approach develop regionalized criteria. conthe PMS.7. 52.3(a)(l). section 3.3(a)(3). report series coverage of conterminous United provides a step-by-step procedure sistently.. A companion report. that allows the user to accurately. 103-D-1912. The derivation of PMP criteria in HMR 51 initiated with the collection of severe storm observed area1 precipitation data from storms occurring in or near the surrounding boundaries of the study region. the estimates of PMP between the 103d and 105th meridian have been revised and are shown in HMR 55A. HMR 53 [43].FLOOD HYDROLOGY MANUAL Figure 3. Since publication of HMR 51 in 1978. 44 . The following subsections proused for the various HMR reports to (d) HMR 51. and 5J. provides a seasonal variation of small area PMP for the same region based on information obtained in HMR 51.000 square miles and durations from 6 to 72 hours are provided in HMR 51 [35].-Generalized estimates of the PMP for the United States east of the 103d meridian for storm area sizes from 10 to 20. i.

lO. The temporal distribution was obtained from examination of the actual occurrence of incremental precipitation from major storms of record. While developing the spatial distribution based on storms of record. 48-. smooth regional isohyets were analyzed on each chart. section 3. and HMR 52 provides general comments concerning what is and is not permitted. 5. the spatial distribution of precipitation. The DAD values were then enveloped areally and durationally and plotted on a new set of charts from which a revised smooth regional analysis was developed. Using DAD information from critical storms of record. A grid was established for these charts from which DAD values of precipitation were read.52].000-square mile storm areas. and regional distribution of maximum persisting l. Use of 45 . and monthly observed precipitation amounts [47. and 72-hour PMP for lo-. Also. Reduction in storm PMP is provided to account for restrictions or physical preferences as to the orientation of the PMS pattern relative to the geographic orientation of the drainage basin under study. This same concept was incorporated into the spatial distribution of the PMP using within/without storm depth-area relationships to realistically distribute area1 PMP precipitation. which is that precipitation occurring outside the PMP portion of the pattern and not considered to be PMP magnitude. or the degree of precipitation concentration within the isohyetal pattern.5 to 1. and 20.HYDROMETEOROLOGY for selected storm area sizes and durations was developed. Such information is obtained from plots of mass rainfall curves determined for individual storms. weekly.OOO-. and the adjusted transposed area1 precipitation from each critical storm was plotted on the appropriate chart. Determination of the proper spatial distribution led to the concept of residual precipitation. an analysis of severe storm precipitation patterns led to the adoption of an elliptically shaped storm isohyetal pattern having a major to minor axis ratio of 2.OOO-.3(a)(3). These charts show the 6-. Any inconsistencies were noted and addressed.48. 200-. 12-.49]. The general shape and gradients of the isohyets were patterned after several rainfall indices such as minimum envelopment of greatest daily. A brief explanation on using the charts is also provided from which the user may quickly obtain required estimates of PMP for a specific drainage basin within the scope of the report. 24-.51. Final charts of PMP were then drawn covering the study region for selected area sizes and durations. The actual selection of a single temporal distribution is left to the user to test what is hydrologically critical.OOO-mbar. 100-y ear precipitation analysis [22.000-. Individual storms indicate a variety of possible temporal distributions. Procedures are provided in HMR 52 that translate PMP values from HMR 51 to a spatially and temporally distributed estimate of the maximum storm potential for a specific drainage basin. In the determination of the appropriate spatial distribution. 12-hour dewpoints [32]. was developed.50. it was noted that critical area1 precipitation only occurred over a restricted area and that lesser magnitude precipitation fell over both smaller and larger area sizes encompassed by the entire storm pattern. l.

local PMS criteria are provided for durations up to 6 hours and area sizes up to 500 square miles. This method assumes that total precipitation from individual storms occurring in an orographic setting can be divided into two separate components for evaluation.24. which is precipitation caused by all atmospheric processes.e. The combined use of HMR 5 1 and 52 permits development of the PMS for the drainage under examination. The revised report. Using the storm separation method. These components are convergence. Because of the variation in terrain features and difficulties encountered in transposing observed area1 precipitation in such complex regions.FLOOD HYDROLOGY MANUAL this important concept permits the determination of concurrent precipitation.-Regionalized estimates used in the derivation of the all-season PMS for the region of the United States between the 103d meridian and the Continental Divide are found in HMR 55A [31]. and orographic. Storm rainfall data [23. Also. and areas up to 5. These storms were classified according to the nature of the precipitationcausing mechanisms. The values obtained provide a PMF that will result from a complete hydrologic analysis.25] were examined to identify critical storms of record occurring in or near surrounding borders of the study region. it was decided to evaluate the maximum storm potential using a “storm separation” method. HMR 55A.000 square miles in the orographic influenced portions of the study area. A tentative evaluation of the maximum precipitation potential was given in a similar report published in March of 1984.. A generalized map indicating subareas of similar topography was developed to assist in transposing available storm data throughout the study region. simultaneous precipitation occurring on an adjacent drainage to that for which the PMP pattern was applied. a modified approach to the PMS determination to that used in the eastern nonorographic United States (HMR 51) was required. The percentage of total 24-hour moisture maximized. variety of storm types affecting the region. i. Because of complexity of terrain. Recent reviews and applications by the Bureau and NWS of the material in this 1984 publication have led to several modifications of the tentative evaluation. isohyetal patterns. (e) HMR 55A. provides information needed for determination of the general PMS for durations up to 72 hours. and meteorological data and analysis to develop an estimate on amount of precipitation due to FAF (free atmospheric forcing) and amount due to orography. for area sizes up to 20. transposed. storm precipitation and enveloped based due to FAF was on the principles 46 .000 square miles in basically nonorographic regions. and lack of available storm area1 precipitation data. the total 24-hour storm precipitation was evaluated based on individual storm relationships of observed precipitation. which is precipitation predominantly caused by terrain influences.

After meteorological and topographical inconsistencies were addressed. 24-hour probable maximum precipitation. The loo-year. 24-hour precipitation) were examined for minimal precipitation values in defined nonorographic areas of the study region. 24-hour probable maximum precipitation. and T/C = 24-hour orographic precipitation factor. 24-hour precipitation was created. and C represented the convergence component of i‘? To obtain C. It was assumed that the loo-year level of precipitation and the ratio of T/C represented the orographic variation of PMP for a duration of 24 hours. maps of T (loo-year. which was called T/C. which was the ratio of the most intense precipitation in a storm to the precipitation measured for the duration of interest.3(a). The next step was to determine the orographic portion of PMP and its variation over the study region. IO-square mile PMP was obtained. The calculations were determined for a dense grid covering the study region. The T simply represented the analysis of the loo-year. 24-hour precipitation frequency data [22] were used to derive this orographic factor. and a smooth analysis of the data was produced. probable maximum precipitation) representing the nonorographic (convergence) component of PMP over the entire region was derived for an area size of 10 square miles and a duration of 24 hours. 47 PMP [iW(l -T/C) + T/C] (4) . A smooth map of the FAFPMP (free atmospheric forced. A smooth analysis indicating the convergence component of the loo-year.HYDROMETEOROLOGY explained in section 3. The terms that have been discussed here can be combined in the following equation: PMP = FAFPMP where: = total combined convergence/orographic. These regions occurred in the eastern plains and broad valleys of the study region considered to contain little orographic influences. and is used as an additional adjustment of the durational variation of dynamic forces within a storm. FAFPMP = free atmospheric forced. 24-hour precipitation that provided the combined regional variation of orographic and convergence precipitation at the loo-year level. The factor T/C was further modified by the evaluation of a storm intensity factor M. It was determined that the most reasonable method to evaluate the regional orographic variation was to determine a factor by which the FAFPMP could be multiplied to achieve total PMP. The relationship shown in equation (4) produced 24-hour total PMP for specific locations assumed to represent areas of 10 square miles. a final map of a 24hour. Smooth analysis of M values were produced over the study region. M = 24-hour storm intensity factor. The M factor was obtained by evaluating such relationships in severe observed storms of record.

isolated from strong atmospheric circulations.40]. Using relationships from severe storms of record and data supplied from adjacent PMP studies.” In the Western States. The derivation of PMP previously described results from the analysis of precipitation due to genera1 storm type events.to 24-hour. two unique storm types were identified as data sources for estimates of severe storm precipitation. Because the methodology for local storm derivation is similar throughout the Western States. it will be briefly described now and referenced when other Western States HMR’s are discussed [38. 6-. Local storm precipitation is defined as “being restricted in both duration and area1 extent. because the local storm data base is severely 48 . Fortunately. and therefore not normally conspicuous on a synoptic scale of analysis. for up to 6 hours and up to 500 square miles. These storms are termed “local” in nature. Because of the sparse population in regions where local storms are effective to PMP determination and the fact that they often cover areas extremely limited in size. meteorological analyses of these events indicate that they can occur almost anywhere within the United States. Ongoing studies are being developed that will permit determination of spatial and temporal distributions of PMP for use in connection with the PMP criteria furnished in HMR 55A. and 72-hour durations. depth-area relationships were assembled for various subregions based on topographical and meteorological variations in such parameters.3(c)(3) be applied if required. Therefore. The PMS derived from each method will have to be individually evaluated in the PMF analysis and the more critical flood used for design. reports of their occurrences are rare. the PMS from such events. can often exceed the values derived from general storms for short durations and small area sizes [31.to 6-hour.” A genera1 storm precipitation is defined as “precipitation occurring from atmospheric/orographic processes readily defined on a synoptic scale of analysis.40]. a lo-square mile PMP was derived for I-. In the examination of PMP potential for the region.38.FLOOD HYDROLOGY MANUAL Using smoothed regional analyses of I. Investigations of local PMS potential begin with identifying intense precipitation events that occurred over small areas and for short durations. Nomograms were then constructed that allowed the user to determine values of PMP for various size drainages within the HMR 55A study region and within the areal and durational limitations of the report. Until these studies are complete. and 72to 24-hour precipitation ratios derived from observed storms of record.” This type of storm can cover large areas and persist for long durations. 6. the user must determine the PMS potential for both general and local storms. In HMR 55A. Meteorological charts are then examined to select those storms that do not result from distinguishing rain-producing synoptic weather features. These two distinct storm types are termed “general” and “local. it is suggested that the interim spatial and temporal distribution criteria described in section 3.

Critical assumptions used are that total PMP is derived from two separate components of precipitation (convergence and orographic). df) HMR 36. there is little need to provide estimates of local storm PMP. which resulted in an index map of a l-hour. Intense precipitation centers located in the Central Valley of California are identified as resulting from 49 . The need for local storm evaluation is basically dependent on the comparative magnitude of PMP derived from general storm types. the use of storm transposition adjusts for this deficiency. In this case. and the data were smoothed. l-square mile local storm PMP.40]. An elliptically shaped isohyetal incremental local storm PMP basic observed local storm data. The magnitude of wind and temperatures occurring in conjunction with the PMS are presented for use in computing snowmelt. PMS for use in the hydrologic pattern and the temporal distribution of were also derived from examination of From these criteria. Details of these modifications are found in reports covering local storm PMP determinations [31. the appropriate local analysis of the PMF was obtained. Data from all major storms of record are tabulated and classified as to their precipitation causing mechanisms. l. Seasonal (October through April) general storm PMP criteria for area sizes up to 5. The local storms selected are adjusted to a common area size and duration.000 square miles and durations from 6 to 72 hours are provided. and that each component can be assessed individually and later added together to produce the complete representative level of PMP.38. The methodology used in this report is called the “orographic separation” technique. Necessary durational adjustments to observed precipitation to provide a l-hour amount follow the depth-duration relationships described in HMR 49 [38]. point precipitation measurements are often assumed to represent an area size of 1 square mile if an actual l-square mile analysis is unavailable.HYDROMETEOROLOGY limited in quantity and quality. A regional grid of storm maximized values was established. usually 1 square mile and 1 hour. Analysis of observed severe local storm precipitation revealed appropriate durational and area1 relationships that were used to adjust the l-hour. For small area sizes less than 100 square miles. but modified to account for the nature and availability of associated data necessary to perform the storm maximizing procedure. l-square mile local PMP index values to obtain a local PMP for other durations and area sizes.and 3hour values of PMP are given. Where general storm type precipitation controls the level of PMP for all critical durations and area sizes required.3(a)(3). The sequential distribution of incremental PMP is also provided.-A d etailed derivation of the regionalized general storm PMP for the Pacific Coast drainage of California is given in HMR 36 [39]. Local storm precipitation representing l-hour and l-square mile values were moisture maximized and transposed using procedures similar to those described in sections 3.3(a)(2) and 3.

Two sets of P/M ratios were developed. Great care was taken so that the combination of optimum precipitation causes from the data examined did not produce unrealistically large values of total PMP. and (2) restricted convergence PMP (to be added later to an orographic component to provide the computed total convergence-orographic influenced PMP). The first dealt with the estimation of convergence precipitation associated with sole convergence PMP storms that occur in strictly nonorographic areas. The value of P is the amount of storm precipitation for a given duration.FLOOD HYDROLOGY MANUAL solely convergence (no orographic influence) storms. A regional smooth distribution of the restricted convergence component to the PMP was accomplished by taking into account the effect of changes in elevation. one based on convergence storms where instability played a minor role. From the information supplied by observed precipitation events in the region. which is the constant ratio formed by P/M values computed from convergence storms with major instability to those with minor instability. The development of the orographic separation model resulted from the meteorological combination of the causes of intense precipitation observed in both types of storms. produces unrestricted convergence PMP. Multiplying the restricted convergence PMP by a factor of 1. This information led to conclusions as to the possibility of combining maximum meteorological parameters or to the imposition of various restrictions on the merging of other precipitation indices. A single convergence restricted 6-hour. and M is the measurement of observed precipitable water associated with a storm event. barrier influences. are classified as basically an orographic event. Values of convergence PMP that were so derived are restricted in that they have to be combined with an orographic component of PMP to produce total convergence-orographic PMP. Multiplying the appropriate P/M ratio by maximum values of moisture produced the two desired levels of convergence PMP: (1) unrestricted convergence PMP (entirely nonorographic). and adjustments for various area sizes. and the other was determined from convergence storms where instability is an important factor.33. Those storms having major centers of extreme precipitation occurring on windward slopes of the Coast and Sierra Nevada Ranges. The second computation dealt with the level of convergence precipitation to be later combined with the orographic component to provide the total PMP. with minimum values of corresponding precipitation in nonorographic regions. These ratios seasonal-durational are the measure of the highest observed storm efficiency. Other monthly graphs were drawn indicating the variation of this index map with area size and duration. the computation of two levels of convergence precipitation were necessary. 50 . Evaluation of these two types of convergence precipitation was accomplished through the smooth analysis of P/M (precipitation/moisture) ratios. 200-square mile PMP index map for the month of January was constructed.

Ap.HYDROMETEOROLOGY The determination of the orographic component to the total PMP. The model was calibrated at various locations from a set of selected storms in an attempt to replicate observed precipitation. A ground profile was constructed at the test location. Figure 3-8 shows a view of the orographic model where resulting precipitation from multiple layers are summed to produce average precipitation over distance Y. q2 Y g e = = = = = = pressure difference at inflow. q. and density of water. distance from inflow to outflow. = mean inflow air speed. which is to be combined with the restricted convergence PMP component. was accomplished using a laminar flow orographic model. mean specific humidity at outflow. The following equation may be used for computational purposes: (5) where: -R = precipitation. the inflow adjusted to maximum values of wind and moisture. parameters were and then run to 51 . The region along the windward facing slopes in the Western States presents an ideal setting for application of the model because of the broad extent of unbroken barriers and the fact that many basically orographically induced precipitation events occur in this region. From this information. which is assumed to be laminar. Moist air flow. was lifted over a mountain barrier. Precipitation is considered to be the difference in moisture inflow above the base of the barrier and that measured as outflow above the barrier crest. The orographic model was tested and calibrated from data for extreme storms of record. the convergence component of observed total precipitation was removed by subtracting the observed precipitation occurring at a nearby nonorographic location from the total observed precipitation at the test site. After successful calibration of the model. acceleration of gravity. In test storms. V. air streamlines were constructed and the freezing level determined so that trajectories of falling snow and rain could be computed. This was done to ensure that the model was properly calibrated to indicate precipitation due only to orographic effects. mean specific humidity at inflow. and observed storm inflow (moisture and wind) with components of wind speeds normal to the barrier was tabulated.

both an unrestricted value of combined restricted component of convergence PMP component producing total PMP need greater precipitation value should be used in it is only necessary to dePMP. seasonal orographic precipitation was established.bostd computed snow ond raindrop on doto from observed troJectorits storms of from record their Figure 3-8. recommended by the Denver Office’s Flood Section personnel that falls within the criteria established for such distributions in HMR 36.3 (c).the study region. when combined with charts indicating the durational and seasonal variation of index orographic PMP. 103-D-1913 produce maximum orographic precipitation at various locations throughout the study region.FLOOD HYDROLOGY MANUAL NOTE:The symbols lnltlotlon * andlevel to represent the the ground. provided the appropriate orographic component of PMP within the scope of the study. section 3. In nonorographic areas of. An orographic 6-hour PMP index map for January was created for the study region. Information regarding the sequential rearrangement of PMP increments based on similar distributions found in actual storms of record is also provided in HMR 36. In orographic areas convergence PMP and a PMP and orographic to be determined. termine the unrestricted value of convergence of the study. Area1 extent of orographic precipitation was incorporated in the determination of a basin width factor. The calculation of the PMP. is that 52 . The sequential distribution.--Orographic laminar flow precipitation model. These indices. By inputing monthly variations of maximum inflow data.

000 square miles west of the Cascade Divide and for lo. it should be patterned after an extreme storm isohyetal pattern that occurred in the drainage.3 (c)(3). a summer local thunderstorm PMP is provided for area sizes up to 550 square miles and for durations up to 6 hours. Major storms of record affecting the study region are identified for three separate major subregions: (1) storms occurring west of the Cascades.3 of chapter 4 be considered before application of the criteria given in HMR 36. General storm seasonal (October through June) PMP estimates are provided for area sizes from 10 to 5. suggested sequential rearrangement of the general storm PMP is provided along with associate temperature and wind sequences for use in determination of snowmelt contributions to the flood hydrograph. (g) HMR 43. If a spatial distribution of the PMP is required. and the difficulty of transposing individual storms in the region due to the complex terrain. Columbia River Basin. It is recommended that the alternate snowflood criteria discussed in section 4. the decision was made to separately compute components (convergenceorographic) of total storm precipitation to develop the PMP potential for the region. and (3) storms located near the Continental Divide. From the meteorological analysis of these storms. The development of PMP criteria in HMR 36 was predominately based on the meteorological analyses of historical severe storms of record. For the region of the study area east of the Cascade Divide. The method chosen was a modified application of the orographic separation technique previously developed in HMR 36. derivation of local thunderstorm criteria is similar to that previously described in the discussion of HMR 55A. The alternate and recommended distribution is one of successive subtraction of subbasin volumetric precipitation. section 3. Criteria for calculation of associated winds and temperatures with the occurrence of the PMS event for snowmelt considerations can also be determined from HMR 36.to l. Also. Much of the methodology used to develop the PMP for the Northwest States is similar to the procedures given in HMR 36 and will not be reviewed in detail here.OOOsquare mile areas east of the Cascade Divide to the Continental Divide. Similar to the procedures in HMR 36. (2) storms in the central interior of the Columbia Basin. Details of these analyses are reported in HMR 37 [53]. lack of analyzed storm data.-R egionalized estimates of PMP for the Northwest United States.HYDROMETEOROLOGY which should be used unless one of the other suggested distributions in HMR 36 would produce a more critical flood hydrograph. Investigations into the determination of the convergence component of total storm PMP began with the calculation of separate P/M ratios based 53 . and Coastal Drainages of Washington and Oregon are provided in HMR 43 [40]. General storm durational PMP is provided up to 72 hours.

Seasonal values of maximum windspeeds and moisture were used in conjunction with the calibrated orographic model to develop corresponding seasonal maximum values of precipitation. Monthly charts of 6to 24-hour and 24.OOO-mbar PMP (restricted) to be combined with an orographic component of PMP. l. a smooth field of P/M ratios (storm efficiency) was created based on such ratios calculated along the west coast and distributed inland based on variations reflected on seasonal charts of maximum moisture. Using the restricted P/M result.OOO-mbar surface level.23. Storms examined were those occurring in the nonorographic areas along the west coast of the study region. displayed in figures or tables. Combination of these various adjustments. up to 5. similar to the method used in developing the model adjustment for HMR 36. lo-square mile. enveloped seasonally.OOO-mbar orographic (restricted) storm convergence PMP were prepared. Observed orographic precipitation occurring at these sites was used to calibrate the model for use in these particular subregions of the study.000 square miles east of the Cascade Divide. Unrestricted convergence PMP for locations west of the Continental Divide is obtained by multiplying restricted convergence PMP by a factor of 1. Maps of highest monthly dewpoint data (moisture index) had been tabulated. adjusted to a l.OOOmbar convergence PMP to surface elevation. lo-square mile. used as a measure of storm efficiency. 54 . allow the user to quickly adjust convergence (restricted) PMP to specific locations throughout the study region. Model tests were made at four widely separated sites in the HMR 43 study region.to 72-hour precipitation ratios were prepared to provide convergence precipitation for other durations based on ratios developed using data from observed storms of record. were calculated for the month of January from severe observed 24-hour precipitation and associated moisture at the time of the storm event.23. Monthly graphs of maximum observed precipitation data were constructed to distribute seasonally the January convergence PMP. see HMR 36 for details of model adjustments. Depth-area relationships were developed to adjust lo-square mile convergence PMP for various area sizes. The orographic component to PMP was derived using the orographic separation model developed in preparing HMR 36. and smoothed regionally. l. Index maps of seasonal 24-hour. Adjustments for terrain features (elevation-barrier/depletion-stimulation effects) associated with the individual basin location within the study region were developed to adjust l. A comparison of P/M ratios associated with maximum convergence storm conditions (unrestricted) to ratios associated with maximum orographic storm conditions (restricted) indicated a constant relationship of 1. Multiplication of restricted P/M ratios by maximum moisture produced a regionalized map of January convergence 24-hour. These ratios.000 square miles west or 1.FLOOD HYDROLOGY MANUAL on those convergence type storms indicating minimal or maximum associated instability in their storm process.

Seasonal variations of PMP values were obtained by examination of model computations and seasonal variation of observed precipitation. Observed precipitation data. (h) HMR 49. were established and model maximum precipitation values calculated for various inflow directions. generalized terrain profiles for the entire study region. General-storm PMP criteria covers area sizes up to 5. lo-square mile orographic PMP index map was developed. To compute total storm PMP.HYDROMETEOROLOGY Because optimum moisture-inflow direction and orientation of slopes varied in the region east of the Cascade Crest.OOO-mbar dewpoints.-Regionalized estimates of seasonal general-storm PMP for the Colorado River and Great Basin drainages are provided in HMR 49 [38]. it is necessary to compute a convergence storm (unrestricted) level of PMP for the more nonorographic areas located west of the Cascade Divide. is obtained by multiplying restricted convergence PMP by a factor of 1. in defined nonorographic subareas of the study region. components of orographic PMP and orographic convergence (restricted) PMP are added together. This convergence precipitation. The storm maximization process necessitated the need for updated charts of maximum 12-hour persisting 1. were moisture maximized and enveloped. all-season local-storm PMP is provided for all of California. orientated in different directions. The methodology applied in the determination of general-storm PMP is to derive separate estimates of orographic and convergence PMP. Each hydrograph should then be routed through the structure to determine which hydrograph is critical for design purposes. Envelopes. HMR 43 and HMR 36. Seasonal distributions of maximum moisture and maximum observed 55 . The two PMS values should be determined and the PMF hydrographs developed. a smooth regional 6-hour.000 square miles for durations of 6 to 72 hours.of greatest computed orographic precipitation from various combinations of slope or inflow orientations were plotted for each region. The method is somewhat similar to the method for developing the general-storm PMP for the Northwest States and California. Local-storm PMP covers area sizes between 1 and 500 square miles and for durations from 15 minutes to 6 hours. Relationships were developed to adjust envelope values to specific basin critical inflow and slope direction. which is not to be combined with the orographic component. These components are later added to find the total general-storm PMP. respectively.23. Derivation of local-storm PMP follows the same procedure previously discussed in HMR 55A. In addition to total storm component calculations of PMP. Durational variation of orographic PMP was obtained from model calculations related in time at various test locations. In addition to the above drainages. An area1 size variation in orographic PMP was established based on depth-area relationships from several observed storms occurring in the region. From the above analysis.

Examples of these modifications would be the adjustment of the first approximation to orographic precipitation by the examination of precipitation occurring along terrain profiles as analyzed from charts of observed storm precipitation and mean annual and seasonal precipitation. 24-hour precipitation was expressed as a percent of this regional loo-year. Depth-area and depth-duration relationships were developed to obtain convergence PMP for other area sizes and durations. This first approximation was modified based on several meteorological parameters. This method of orographic component evaluation assumes that the relationship of component convergence to orographic 24-hour PMP is similar to that derived using loo-year. These relationships were based on DAD relationships from observed nonorographic storms occurring in both the study region and in the plains of the Central States. In calculating the orographic component of PMP for the HMR 49 region. Based on an understanding of the greater convective activity of severe storms occurring in the region and the complexity of terrain features. and the loo-year level of precipitation would most likely represent that for the level of PMP with little additional modification. Adjustment to convergence PMP at l. The alternate procedure developed was based on observed storm precipitation and its variation due to terrain effects. Multiplication of these percentages by the convergence component of PMP resulted in the first approximation of orographic PMP for the study area. A first approximation to the estimation of an orographic component of PMP involved the use of loo-year. Total loo-year. Resulting convergence PMP values were in general agreement with similar convergence component estimates derived for adjacent regions to the study area.FLOOD HYDROLOGY MANUAL precipitation were used as guides to interpolate between locations of nonorographic moisture-maximized precipitation. 24-hour orographic PMP index map. the orographic precipitation laminar flow model developed in HMR 36 was not used. Seasonal index maps of l. HMR 36 and 43. only limited use of the orographic inflow model could be used satisfactorily. The 56 . 24-hour precipitation. It was assumed that the regional complexity of the loo-year precipitation was mainly due to orographic influences. Incorporation of these adjustments and those from other indices resulted in the adoption of a final lo-square mile. In the nonorographic regions of the study region. A variation of orographic PMP with basin size was developed using depth-area relationships from major storms occurring in the region. 24-hour convergence component.OOO-mbar convergence PMP for 24 hours and 10 square miles were established for the study region.OOO-mbar for the effects of elevation and barrier were incorporated. frequency values from NOAA Atlas 2 were assumed to represent precipitation from entirely convergence storm mechanisms after modifications for surface level and barrier effects were applied to adjust frequency values to a common surface for analysis. 24-hour precipitation data obtained from NOAA Atlas 2 [22].

(2) seasonal variation of maximum moisture and winds. Supporting data upon which regionalized PMP criteria were developed are also shown in HMR 50 [33].HYDROMETEOROLOGY durational variation of orographic PMP was adopted from examinations of the durational variations noted in maximum winds and moisture affecting the region. Temporal and spatial distributions of general-storm PMP are not provided in HMR 49. and (3) seasonal computations of precipitation expressed by the orographic precipitation mode1 described in HMR 36 and 43. Seasonal variation of orographic component PMP was accomplished by tying into seasonal variations obtained from HMR 36 and 43 near boundaries of the study region and by using various seasonal related precipitation indices within the study region. The adopted durational variation was checked and adjusted by comparison of durational precipitation associated with severe storms occurring in the region and durational ratios formed by combining maximum values of 6-. separate evaluations of convergence and orographic component PMP are established. Use of the individual drainage approach should be left to trained hydrometeorologists to develop. the resulting values should be checked for consistency. By using convergence and orographic index charts and other figures. (2) with observed storm data of record. durational. 24-. Examples of the indices examined are: (1) seasonal variation of observed maximum precipitation at various stations within study region.3(c)(3) to sequence increments of total precipitation as well as to obtain the spatial distribution of general-storm PMP. and 72-hour observed storm precipitation. The steps for determining the PMS as set forth in various HMR’s are in simple. (i) Summary-The methodologies briefly describing the derivation of the PMS from an individual drainage study or through regionalized analyses have been presented. and seasonal variations of index values of PMP. easyto-follow. step-wise procedures. and (3) with extreme values 57 . and graphs or tables to describe the areal. In the development of individual drainage estimates of PMP. These components must be added together to provide total PMP for desired individual drainages bounded by the limiting scope of HMR 49. Discussions of individual genera1 and local storms important to the development of criteria for estimating PMP in HMR 49 are presented as well as the derivation of seasonal maximum moisture charts (dewpoints). It is Bureau practice to use the procedure described in section 3. and these computer programs are available from the Flood Section at the Bureau’s Denver Office. Samples of such checks generally consist of comparisons: (1) with previous estimates derived for drainages or regions in surrounding area. Several of these procedures have been adapted for computerized analyses. and the genera1 magnitude achieved.

and methoAs new severe storms occur and are dologies are an ongoing process. each producing an array of possible distributions. 3. The combined efforts of Bureau regional and Denver Office hydrologists have resulted in some basic concepts that can be generally applied to derive appropriate PMS distributions keeping within meteorological criteria and the desire to achieve critical conditions for PMF hydrograph analysis. most regionalized reports internally address some type of evaluation as to the level of PMP obtained. revisions and refinements will be made and documented in subsequent reports. precipitation at the two- Step 4. Obtain increments of unit duration precipitation for total storm period from appropriate individual drainage or regionalized studies from the smooth PMP depth-duration curve for the basin or subbasins. Following the general comments made in sections 3. For PMP obtained using the HMR series. Place before and second and third greatest unit adjacent to the greatest increment fourth greatest incremental greatest incremental amount.FLOOD HYDROLOGY MANUAL of precipitation derived by statistical means. current procedures and reports will be evaluated for their adequacy. duration precipitation in descending order. Place greatest thirds position in the amount of unit duration total storm period. Revision and refinements of PMS procedures. evaluation of the latter two comparisons have been completed for most of the United States [34]. precipitation four greatest 58 . temporal and spatial distributions of the design precipitation are often required. duration for unit hydrograph analysis as dis- Step 2. Determine cussed in chapter unit 4.3(c)(2). Additionally. Place remaining unit duration incremental amounts in a similar fashion to the placement of the precipitation increments. a variety of methods exist. recorded. techniques. or better theories developed. (a) Temporal. Where deficiencies exist. Step 3.4 Distributions of PMS In the development of the general PMS based on individual drainage estimates or regionalized studies. the following steps were derived to provide a single PMS temporal distribution that has been adopted for use throughout the United States: Step 1. Place jacent to the precipitation Step 6.-From examination of individual drainage and regionalized storm criteria combined with various hydrological tests.3(c)(l) and 3. after and ad- Step 5.

59 .HYDROMETEOROLOGY Figure 3-9 illustrates this preferred distribution as applied to a 24-hour storm of l-hour duration. - 2 4 6 8 TIME IN HOURS FROM BEGINNING OF STORM Figure 3-9. For obtaining the spatial distribution of regionalized derived general-storm PMP for the United States. (2) For the remaining regions of the United States. arrangement of PMP increments for a 24-hour storm.--Sequential 103-D-1914. an interim technique called “successive subtraction of subbasin PMP volumes” should be applied until adequate regionalized spatial distributions of PMP are developed. and their incorporation should generally be determined by trained hydrometeorologists. Caution should be used in the application of these methods. I. (b) Spatial. Because hydrologic analysis is performed from averages of A Incremental Precipitation by Rank.-Various methods of areally distributing PMP have been developed over the years. the spatial distribution criteria shown in HMR 52 is deemed applicable. the following criteria should be applied: (1) For nonorographic regions of the Eastern States covered by HMR 51 and portions of HMR 55A. Longer duration storms would use a similar distribution.Greatest 2= 2nd Greatest 3= 3rd Greatest etc.

The basic equation for this technique is: P. = PI... 43. tests of various subbasin centerings of the PMP should be evaluated. to the summation of precipitation volumes over subareas A.. and P. = P. .A.. over area A. + . Find in subbasin concurrent 1: average precipitation P. A. is the value determined in step 1 for subbasin 1. + A. The precipitation volume. and then determining concurrent precipitation in subbasins 2 and 3. When the PMP is determined for a certain basin or combination of subbasins. The P. Find centered the concurrent in subbasin 1 with average concurrent precipitation precipitation P. times the area. over which the precipitation falls.A. Find the average precipitation P..A. A. P. Step 2. Step 1.A. The first step consists of centering the PMP in subbasin 1.. the spatial distribution of PMP by the “successive subtraction of subbasin PMP volumes” technique can be applied. when PMS is centered P. the P. Details of this technique are best described in the following example: Assume a basin with total area size A.P.. A.. + P. where the PMP is assumed centered. . and A. centered in subbasin 1.v.. Using the above criteria. as discussed chapter 4.A. . is broken into three separate subbasins of area sizes A. (6) Equation (6) equates precipitation volume over an area..A. and that required for individual or joint subbasin analysis. is the average precipitation.. Subscript IV identifies the individual T represents total area of the drainage.A. or 55A. i. it is only necessary to determine the averages of design storm precipitation calculated for the total drainage. and subscript the purpose of spatial distribution is to provide information for peak discharge calculations. for hydrologic analysis.2 A. . + P. and the duration average depth of PMP. when PMP is in subbasin 2: 60 . A. it is tantamount to saying the storm is centered over that basin or subbasin combination. is obtained directly from the appropriate HMR for the combined area size A. HMR 36. A. The result is the HMR for area size A.. Step 3. A. . + P. 49.e. is directly obtained from the appropriate of interest. 4 The value for P... It is necessary to assess the spatial distribution of the PMP determined for the entire drainage area from the appropriate orographic area report. PA.FLOOD HYDROLOGY MANUAL meteorological and hydrological parameters over basins or subbasins.. Since subbasins.A.

. The various spatial distributions are then tested hydrologically....2. 61 .2. + A.3 is obtained directly from appropriate HMR for combined area size A.2.3 Au.A. The entire procedure is evaluated for different storm centerings.3 A..P. PMP centered over subbasin 2 and then subbasin 3. A. and the distribution resulting in the critical peak discharge PMF is chosen for design.. Until such information becomes available. is the value calculated in step 2 for the PMP volume centered in subbasins 1 and 2..3 .2 A3 The value for P. Studies updating and developing regionalized temporal and spatial general PMS distributions in the United States for HMR 43 and 55A are currently underway. the temporal and spatial derivation procedures outlined in this chapter should be applied. = P I.HYDROMETEOROLOGY P. These computations are repeated for each durational increment of precipitation necessary to describe the spatial distribution for the entire storm period.. Similar information for other regions of the Western States will be addressed in future reports. and P. i. + A..e...


-The unit hydrograph may be defined as “the discharge hydrograph resulting from 1 inch of direct runoff generated uniformly over the tributary area at a uniform rate during a specified time period. retained the unit hydrograph approach because of its simplicity. This chapter concerns the processes involved with converting rainfall and/or snowmelt into a hydrograph representing the upper level or maximum flood runoff that a drainage basin might reasonably be expected to produce. 1932. As previously mentioned. In 1932. and the relatively low costs associated with its application to flood hydrology studies. The advent of high speed electronic computers has led a number of hydrologists to devise approaches using complex watershed models. has undergone considerable refinement over the years. the relative “goodness” of an approach is measured by how well that approach reproduces actual recorded flood events. as alternatives to the unit hydrograph model. it should be noted that there are a number of other techniques available for making this conversion. Comparative studies have indicated that both approaches are able to satisfactorily reproduce these events with neither one being notably superior to the other. Sherman [5]’ initially proposed the unit hydrograph approach to convert rainfall occurring over a drainage basin to flood runoff from that basin. Because these complex watershed models generally require extensive calibration to adequately represent a drainage basin’s physical properties. including highly complex computerized watershed models. the primary interest is in simulating a basin’s runoff response to extreme rainfall events. However. considerable-effort must be expended in the field and office in acquisition of data relative to these properties. issue of Engineering News Record. over the years. (b) Basic Unit Hydrograph Concept. In the final analysis. in the Bureau of Reclamation’s application.1 Flood Runoff Chapter 4 HYDROGRAPH DETERMINATIONS From Rainfall (a) General. Leroy K. the unit hydrograph approach is the basic tool or “model” to convert rainfall to runoff after abstracting suitable infiltration losses. reliability. 63 . However. Accordingly. Many of these watershed models are an appropriate basis for simulating a continuous series of runoff responses to normal precipitation events. to predict a drainage basin’s runoff response to rainfall.” The concept of the unit hydrograph theory follows five basic assumptions [2]: ‘Numbers in bracketsrefer to entries in the Bibliography. which was formally presented in the April 7. the Bureau has.FLOOD 4. Sherman’s approach.-Chapter 3 presented the criteria and procedures for the determination of PMS amounts that could be experienced over a drainage basin.

Proper application of the concept is treated in subsequent parts of this chapter. The unit hydrograph theory assumes that the l-hour hydrograph ordinates are proportional to the rainfall excess. The resulting time duration of the base from the effective rainfall runoff rainfall.” a specific period of the surface runoff hydrograph of unit duration is constant. and that this runoff flows to tributary watercourses arriving eventually at a stream-gauging station. Ordinates of surface amount of unit effective 5. which is referred to as the 3. The fact that the rainfall-runoff process is actually nonlinear is one of the acknowledged shortcomings of the concept. Assume the rate at which the rain is falling is such that 1 inch of surface runoff results. and other losses have been deducted. Effective rainfall is distributed uniformly over the entire drainage basin. hydrograph are proportional to 4. surface ponding. 64 . The basic concept of the unit hydrograph theory [9] can be explained by considering a situation where a storm of l-hour duration produces rainfall at a constant rate over that duration. and occurs uniformly over the entire drainage basin above a recording stream-gauging station. The runoff at the gauging station will be recorded as a hydrograph representing the temporal distribution of runoff resulting from 1 inch of “rainfall excess” occurring in 1 hour. the concept provides entirely satisfactory results for developing flood hydrographs. Now consider the situation where the “rainfall excess” was 2 inches in a l-hour period. The surface runoff hydrograph for a given drainage all of the unique physical characteristics of the drainage basin reflects basin. distributed over “unit duration.FLOOD HYDROLOGY MANUAL 1. if properly applied. and the unit hydrograph is said to have a “unit duration” of 1 hour. Naturally. It follows that the runoff hydrograph at the gauging station resulting from 2 inches of rainfall excess can be predicted by multiplying each of the l-hour unit hydrograph ordinates by a factor of 2. it appears that the unit hydrograph concept represents the modeling of the rainfall-runoff process as a linear system. as shown on figure 4-1. this is true for any multiple or fraction of an inch of rainfall excess. 2. The recorded hydrograph is then said to be the “l-hour unit hydrograph” for the basin contributing runoff at the gauging station. The significance of unit duration will be discussed later in this chapter. Considering the above basic assumptions. The effective rainfall is uniformly of time. However. and is the amount of rainfall available for runoff after infiltration.

Each incremental hydrograph is determined by multiplying the increments of rainfall excess by the drainage basin’s unit hydrograph ordinates. regardless of location. only the excess is shown on figure 4-2.4 to 1. as shown by the 0.HYDROGRAPH DETERMINATIONS 0 0 6 12 I6 24 30 36 4 TIME IN HOURS AFTER START OF RAINFALL EXCESS Figure 4-l. 65 .-Unit hydrograph principles. Unfortunately. The discussion to this point has basically considered an isolated rainfall event sustained for a period equal to the unit duration of the unit hydrograph. These time intervals are usually equal to the unit duration of the unit hydrograph.5inch curves at the bottom of the figure. The total runoff from the complex rainfall event. are both longer than the unit duration and more varied in intensity from one “unit” period to another. Figure 4-2 illustrates the manner in which the unit hydrograph approach takes severe storms into consideration. The severe storms that occasionally visit every drainage basin. 103-D-1848. nature does not usually behave in such a simplistic manner. can be determined by adding the ordinates of each runoff hydrograph at discrete time intervals. Figure 4-2 shows that each of the five increments of precipitation excess results in an incremental runoff hydrograph. The resulting runoff hydrograph can be drawn by graphically connecting these ordinate points in a curvilinear fashion.

and a synthetic unit hydrograph. and (2) a unit hydrograph for each basin analyzed. which are then routed and combined to form the hydrograph for the total basin.-Unit hydrograph application.FLOOD HYDROLOGY MANUAL TIME IN HOURS Figure 4-2. The surface runoff portions of flood hydrographs at these sites are developed using hypothetical rainfall amounts that could occur over the drainage basin. The total flood hydrograph is determined by adding appropriate base flow or snowmelt flow to the surface rainfall-runoff hydrograph. synthetic unit hydrographs are satisfactory when generated for drainages up to about 500 square miles. and a separate synthetic unit hydrograph generated for each subbasin. the hydrologic engineer is usually faced with the problem of providing a flood hydrograph for design purposes at a location where no streamflow data have been obtained. appropriate infiltration loss rates. In general. 103-D-1849. In actual practice. Larger basins should be divided into subbasins of about 500 square miles. Associated with each unit hydrograph are two features that are used to determine synthetic unit hydrographs for ungauged 66 . Synthetic unit hydrographs are developed from parameters representing the salient features of the rainfall-runoff phenomena found by reconstituting observed flood events on similar drainage basins. Reconstitution of observed events generally provide two significant features or items of information: (1) an indication of infiltration rates to be expected with the composite soil types present in the drainage basin.

Snyder [55] developed the following relationship for lag time based on studies of basins in the Appalachian Mountain region: Lg = c. 67 . = L = L. 103-D-1915. .-Typical flood hydrograph reconstitution. -Over the years. in hours. The latter feature is called the “unit hydrograph lag time.” In 1938.HYDROGRAPH DETERMINATIONS drainage basins: (1) a relationship representing the variation of runoff over time. . in miles. A graphical example of flood hydrograph reconstitution is shown on figure 4-3. Lag time was originally defined in 1936 by Horner and Flynt [54] as the “. time difference in phase between salient features of the rainfall and runoff rate curves. Analysis of these reconstitution results has led to the conclusion that a unit hydrograph’s lag time varied as a function of certain measurable basin parameters. numerous observed floods have ben reconstituted using the unit hydrograph approach. 0' 0 I 6 I 12 I 18 I 24 I SO I 36 I 42 I 48 I 54 I 60 I S6 I 72 I 78 J (14 Figure 4-3. = (1) lag time. and length along L to a point opposite centroid of drainage basin.” (c) Unit Hydrograph Lag Time. length of longest watercourse.2.8 to 2. in miles.)OJ where: Lg = C. (LL. and (2) the time that the rise in runoff lags the rainfall causing the rise. a coefficient that varies from 1.

. Accordingly. It should be emphasized that K.. are as follows: (1) Dimensionless center of unit unit hydrograph rainfall excess technique [9]. in hours./. = lag time. = 26K. the K.. used in the Bureau’s 1952 publication formed the basis for procedures Unitgraph Procedures [9]..33 regardless of the regional location of a particular drainage basin.FLOOD HYDROLOGY MANUAL In 1944. can be expressed as 26 times the average Manning’s n value representing the hydraulic characteristics of a drainage basin’s drainage network.5 1 where: L. This average n value is identified as K. the Los Angeles District of the Corps of Engineers introduced the slope of the longest watercourse into Snyder’s which resulted in the following general relationship between and measurable basin parameters: [56. L = length of longest watercourse from point of concentration to boundary of drainage basin.. Lag time deftnitions. This “nonlinear” “linear” basic assumptions of the unit hydrograph concept.-Lag to the time that time is the time 50 percent of the from 68 . Recent analyses of unit hydrograph data for many drainage basins throughout the United States. depending on the technique used. L. = length along L from point of concentration to a point opposite centroid of drainage basin. S = overall slope of L.57] equation. ( LL. that are somewhat different than originally proposed by Horner and Flynt [54]. therefore. = c. = a constant.33).!F’. Additional detailed analyses of these data have led these same engineers to conclude that C. in miles (point of concentration is that location on watercourse where a hydrograph is desired). so. Current Bureau practice uses two definitions of unit hydrograph lag time. have led Bureau hydrologic engineers to the conclusion that the value of the exponent N can be taken as 0. in feet per mile. and N = an exponent (the Bureau currently uses N = 0. lag time s L. LL. C. is primarily a function of the magnitude of discharge and will normally decrease with condition is inconsistent with the increasing discharge. depending on the technique for synthetic unit hydrograph development being used. value should be set to reflect hydraulic conditions that would exist in extreme flood conditions. as they have become available. C. The index of equation (2). in subsequent considerations of lag time in this manual. in miles.

This temporal distribution is accomplished by using a dimensionless form of an observed unit hydrograph for a hydrologically similar drainage basin. There are currently two methods used by the Bureau that utilize the dimensionless form of the unit hydrograph: (1) Dimensionless Unit Hydrograph Technique. This relationship is shown graphically on figure 4-5. The remaining half is the means by which the runoff from the unit effective rainfall is distributed over time. 103-D-1916. The ultimate discharge is an equilibrium rate achieved at the time when the entire drainage basin is contributing runoff at the concentration point from the continuous series of unit ranfall excess increments. and (2) S-Graph Technique. Figure 4-4. differences in drainage basin size and variations in unit hydrograph lag time and unit duration are automatically taken into consideration.HYDROGRAPH DETERMINATIONS volume of unit runoff from the drainage basin has passed the concentration point. By using a dimensionless form. (d) Temporal Distribution of Unit Runoff.-The determination of a basin’s unit hydrograph lag time is only half the information required for developing a synthetic unit hydrograph.-Dimensionless unit hydrograph lag time. (2) S-graph technique[56.-Lag time is the time from the start of a continuous series of unit rainfall excess increments to the time when the resulting runoff hydrograph reaches 50 percent of the ultimate discharge.57] . . 69 . This concept is displayed graphically on figure 4-4.

plus 50 percent D. 103-D-1917 (1) Dimensionless unit hydrograph technique [9]. Mathematically. in l-day cubic feet per second. The time base (abscissa) is expressed centage of the lag time L. this relationship is: Q (LK + 0.50 Increments drograph of time t are equal to the unit derived from the recorded flood duration event. Expressed mathematically. this may be expressed hydrographs to dimensionless as follows: in terms of time t as a perof the unit rainfall duration as loot L. + 0.-S-Graph lag time.50 divided by the unit runoff volume for the basin V’. Dimensionless discharge ordinates q are expressed in terms of the product of the recorded flood unit hydrograph ordinate Q.50) 4= V’ (3) 70 .-Unit developed from recorded flood events are converted form for use in synthetic unit hydrograph development a. and lag time L. plus 50 percent of the unit rainfall duration Lx + 0.FLOOD HYDROLOGY MANUAL Unit hydrograph lag Ultimate discharge rate Time at which discharge point equals 50 percent discharge rate at concentration of ultimate TIME Figure 4-5. of the unit hy- b. in cubic feet per second.

The drainage basin’s physical parameters K.. in hours. the length of the longest watercourse. consideration should also be given to K.. and K. and to drainage network density. and S are then entered into the general lag equation (1): where: Lg = lag time. considerable attention was given to the specific observations that should be made during a field inspection of a drainage basin.-Unit hydrographs developed recorded events are converted to dimensionless form as follows: from a. In assigning a K. L. L. in miles. and the length along the longest watercourse to a point opposite the centroid of the drainage basin. each successively out of phase by one unit period. 71 .57]. the elapsed time from the beginning of rainfall to the time when 50 percent of the ultimate discharge is reached. The lag time for this particular technique is determined by reading from the plotted summation hydrograph. L.HYDROGRAPH DETERMINATIONS (2) S-graph technique [56. L = distance of longest watercourse. in miles. A summation hydrograph is initially developed by algebraically adding the ordinates of a continuous series of identical unit hydrographs. values developed from analyses of observed flood hydrographs for basins that are similar with respect to general topography.. Once the value of K. S = overall slope of L measured from gauging station or point of interest to drainage basin divide. L. in feet per mile. (e) Development of Synthetic Unit Hydrographs. S. by stream length. value for a particular basin. are measured. average Manning’s n value for the principal watercourses in the drainage basin.. The slope of the longest watercourse. is also determined using contour data from the topographic map. L. A suitable topographic map. = distance from gauging station to a point opposite centroid of drainage basin. to channel and flood plain characteristics. has been estimated. b. The dimensionless hydrograph is then developed from the summation hydrograph by converting the time base (abscissa) to time in percent of lag time and converting the ordinate values to discharge as a percent of the ultimate discharge.. such as a USGS quadrangle map is usually used for these measurements. Observations made relative to the basin’s drainage network or hydraulic system form the primary basis for establishing an appropriate K.--ln chapter 2. = a trial value based on an estimate of the weighted. value [lo] to be used in estimating the synthetic unit hydrograph lag time.

. computed equation (2) which is equal to 26 K. which result in lag equation coefficients C. value. Those hydrographs not included were considered to represent either interflow runoff or runoff that included significant contributions from snowmelt. constant in hydrograph reconstitution..S. The relationships shown on figure 4-6 reflect K. The 162 hydrographs were then segregated on a regional and topographic basis. the overland flow occurs for fairly short distances before entering a well-defined watercourse.069 to as low as 0. 72 .77. The supporting data for these figures are listed in tables 4-l through 4-6. The for either the Graph To aid in determining an appropriate lag time. These data may be used as a guide during the field reconnaissance to establish an appropriate K. As previously stated. it is of considerable value to conduct a field reconnaissance of the basins represented in the data set to gain an understanding of the physical conditions that are indicative of a particular K. particularly with respect to the impact on runoff from various levels of development.O-~ unit hydrograph lag time determined from the flood K. as shown on figures 4-6 through 4-l 1. Many upper reach watercourses are swales. and the overbank conditions reflect relatively low Manning’s n values. These tables include the station index number. drainage area (in some cases.z value.8 and 0. 162 flood hydrographs considered representative of surface runoff from rainfall events were selected for analysis relative to regionalized trends in the lag time relationships and the time versus variation of discharge relationships. Figure 4-6 and the data in table 4-l represent conditions on the Great Plains west of the Mississippi River and east of the foothills of the Rocky Mountains.. and the C. As a result of the examination of these reconstitutions.030. station name and location. and the welldefined drainage networks are limited to the lower parts of the basins. These reconstitutions represent flood runoff from natural basins throughout the conterminous United States west of the Mississippi River and from urbanized basins for several locations throughout the States. many flood hydrograph reconstitutions have been examined. values from 0.I FLOOD HYDROLOGY MANUAL Equation (4) yields results of applying dimensionless unit (Lag Time)S-Graph the synthetic this equation hydrograph = (Lag time unit hydrograph lag time are considered adequate or S-graph approach: + Semiduration)Dimensionless in hours. Overbank flow conditions reflect relatively high Manning’s n values. basin factor LL. value for the drainage basin being studied. of 1. Data for urbanized basins are included in this manual because of the increased interest in the flood hydrology of such areas. The lower limit values generally reflect a well-defined drainage network reaching points near the basin boundary.J’. respectively. only the area contributing to flood runoff). The upper limit values generally reflect basins with considerable overland flow before reaching moderately well-defined watercourses.



_ I< 0. Great Basin. ._ 1" BASIN FACTOR." ..0 Partially urbanized desert bosins I .100. 103-D-1852. LL.1 100‘ 4000 30 Figure 4-&-Unit hydrograph lag klationships for the Southwest Desert.^ I./S”’ . and Colorado Plateau.



FLOOD HYDROLOGY MANUAL SUnOH NI ‘1 3Wll 9Vl HdVti9011aA .H 1lNf-l 78 .

3 129. at Puerto D. Dam nr.065 .04 1.0 2.063 .0 ‘68.0 3230.0 960.2 17.9 12.059 .0 40.66 1.2 27.053 . nr. OK Elm Fk.0 0. nr. nr.0 ‘6.038 . Ft. CO Jimmy Camp Cr.0 3970.38 1.064 . OK Medicine Cr.07 1.0 312.64 1. at Dawson.66 1.48 0.040 .5 7.4 20.0 25.1 10.030 . Masonville. nr.0 156.8 3.0 2645. above Clay Cr.0 time.69 0. L” ho&s Lag K 0.1 13.0 838. Diamond A Ranch. nr.0 722.0 648. OK Red Willow Cr. Station and location Black Squirrel Cr. Red R. CO Washita R. OK Beaver Cr.057 .53 1.04 1.0 1566. NM (2d reconstruction) Rio Hondo nr.Table 4-l.0 99.0 Basin factor.9 8. Lune. nr.0 131.91 1.0 ‘1050.0 299.0 83. Drainage area. OK Pond Cr.043 .0 1.98 1. CO Dry Creek nr.067 . LL. CO Smokev Hill R.59 1.041 . Red R.0 980.5 5.0 54.33 1.076 .4 3300. nr.99 1. Ellsworth.46 1.0 65. NM Buckhorn Cr.0 0.9 13.’ NM Pecos R.4 5. *Lamar.5 21./.8 17.12 0. above Marmath. nr.74 1.9 353. nr. nr.0 14. CO Willow Cr. of North Fk. Cobb.5 1.061 .0 134.78 1.036 . WY area Station) Station) 13 14 15 16 17 18 19 20 x: 23 24 25 26 27 28 ‘Contributing 3. nr. Cheyenne.040 . ND Middle Fk. nr.0 275. NE Pecos R. KS Cimmaron R. at Anton Chico.3 73.064 .0 787.5 1.0 243. nr.9 44. Granite. ‘OK North Fk.0 7.051 .+ 92.5 4.0 299. above Kaycee. Lamar.0 ‘2005. NE Little Beaver Cr.7 7.7 860.030 . Powder R. ‘OK Barnitze Cr.0 307.035 .33 1.88 3.5 7. No. NM Vermejo R.0 5.5 213. No.056 . Widefield.12 1.8 5.56 1. NE (Central Plains Exneriment 1 Washita R.059 G 0.0 2150.17 1. McCook.030 .051 .045 . Red R.78 0.5 8. 3 NE (Central Plains Experiment Beaver Cr. nr.043 .0 890. mi” 353.2 5. OK Rock Cr. Lamar. at Clinton.78 1. nr. CO Clay Cr.076 . 8.94 0.0 300. Dougherty. OK Salt Fk. Magnum.98 1. Hondo. NM Vermejo R. NM Rio Ruidoso nr.1 2. Ellicot.0 550.19 5.0 83.060 .0 794. at Dawson. Cambridge.53 No.-Unit hydrograph lag data for the Great Plains.0 10.2 306.0 920.0 797.0 73.7 11. Boise City.7 5 8 z: g 9 Fii ii! E 5 8 0-J . Magnum.0 1050. Arapaho.

which have hy. indicate lag equation coefficient C. values range from a high of 0. Unit hydrograph lag time data representing basins located at the higher elevations of these mountain ranges are generally lacking.073 to 0. values of a high of 0. The third lag curve. San Juan. and Colorado Plateau regions of southern California.I drologic characteristics similar to those of the Rocky Mountains. and Montana. Relationships shown on figure 4-7. values ranging from 1.160 (lag equation coefficient C. representing the thunderstorm phenomena in this region. Oregon. on the bed material in the principal channels. Wasatch. representing the general storm phenomena. Big Horn. and Bitteroot Ranges of New Mexico.050. Nevada. Included are the Front. respectively. Colorado. and the lower value is typical of the usual desert terrain. In addition. Arizona. Figure 4-8 and the data in table 4-3 represent conditions in the Southwest Desert. respectively. Absaroka. fairly welldefined drainage networks.042 with lag equation coefficient C. Wyoming. the low-intensity general storm and the high-intensity thunderstorm event. This limiting value is consistent with data for drainage basins located in the Sierra Nevada of California. Higher values are considered appropriate for developing flood hydrographs of more common frequency than the loo-year event. values of 1. Because most of the data reflect low-intensity storms. Selection of a value within these limits depends primarily on the character of flow retarding vegetation in the portions of the basin where overland flow will occur in the overbank flow areas. a K. the infrequency of severe rainstorms in these areas and in the Northern States precludes acquisition of a good data base representing severe event phenomena.4. of 4. of 0.8 and 3. Wind River. Great Basin. two sets of relationships are presented on figure 4-7. and also on the extent to which the drainage network has been developed by erosion. Examination of the available data has led to the conclusion that they represent two types of storm phenomena.9 and 1.130 with resulting lag equation coefficient values of 6. represents 80 .1. indicate K. reflecting relatively high hydraulic efficiencies.260 to a low of 0. Relationships shown on figure 4-7. and western Colorado and New Mexico.8 to 1.070 to a low of 0. Sangre de Christo. Utah. for example. Utah. one for each type meteorologic event.3 for K.FLOOD HYDROLOGY MANUAL Figure 4-7 and the data in table 4-2 represent conditions in the Rocky Mountains. The higher value is indicative of decreased basin hydraulic efficiency consistent with the coniferous forests found at the higher elevations in this desert region. dashed line on figure 4-8.2) or less should be used in the development of PMF hydrographs. Idaho. respectively. values ranging from 0. Basins in this arid region generally have sparse vegetation. Relationships shown on figure 4-8. indicate regional K. and terrain varying from rolling to very rugged in the more mountainous areas. Accordingly.

McPhee.260 . CO North Fk. nr. at Logan.0 469.0 194. lag L 4.5 Co c g 3 P .0 0. nr.0 11.0 1110.059 . nr.07 6. nr.5 34.0 43.0 92.I96 . and Oregon. Hermosa.9 1. at Cedaredge.0 2. CO Uintah R.058 .0 0. Colorado.241 .4 0.0 35.155 . above Craney Cr. CO Dolores R.0 28.Table 4-2.0 2. Cedar City. WY Piney Cr. MT Surface Cr.81 LL <a /.0 ho:rs 8.0 260.56 4.065 .29 5.42 6. Index No.3 10.l26 .51 8.076 .0 1795. ID Madison R.324 .0 3.0 216.0 Basin factor.0 2511. Payette R. WY San Miguel R. nr.59 8. 0. Neola. UT Sevier R.0 2060. UT Parrish Cr.7 32. WY Grey Bull R.43 1. Utah. Centerville.0 21. UT South Fk. nr.2 0.0 681. nr.3 3. Wyoming.0 123. Hatch.28 6.3 181. Meeteetse.51 1.1 59. Estes Park.339 c.259 .214 . Big Thompson R. time.8 41. Garden Valley.0 37.P 69. CO South Piney Cr.10 5. CO Dry Gulch nr.0 310. Montana.3 174. Glen Haven.0 114.73 1. Meeteetse.5 16.9 68.0 36. MT Gallatin R. Weiser.69 1. at Delta.14 6. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Station and location Purgatoire R. Three Forks.0 30. Idaho.0 30.0 1160.0 6. Drewsey.236 .0 0. 1.4 1.53 1.46 3.4 5. ID Malheur R. nr. Estes Park. OR Weiser R. at Trinidad.0 284.0 1110.6 41.056 .4 2.27 8. at Keamey.0 28.-Unit hydrograph lag data for the Rocky Mountains.30 1. nr. UT Centerville Cr. nr.2 0. nr. miP 742.0 38.3 12.98 6. CO Los Pinos R.124 . Centerville. at Naturita.19 6. UT Sevier R.1 3.0 443.0 779. WY Coal Cr.242 .2 15.9 2.324 . nr.0 0.03 5. CO Rabbit Gulch nr.209 .058 .0 1080. Kingston.195 .42 6.76 5.050 .11 1. New Mexico. Bayfield.0 11.22 3. nr.9 106. CO Uncompaghre R.5 2.235 . CO Wood R.238 .5 9.8 29.0 34. at Willow Park. CO Drainage area.4 793.0 910.0 50. nr. nr. UT Florida R.061 .0 69.5 193. nr.1 11.

8 ft.46 1.27 0. AZ 17.8 23:: Fast Fullerton Cr.0 1476.056 . CA 645. UT 15.0 99. AZ 678. AZ 4341.3 740.5 2.0 1722.050 .0 4::: Temecula Cr.3 2. 4 Damsite.8 San Diego R.6 26:3 108.051 . at Fullerton Dam.3 San Vincente Cr.9 1:2 Santa Clara R. nr. CA 4.2 5.6 Santa Anita Cr. CA 137.20 1.0 28.3 ?X San Antonio Cr.8 6. CA 3. AZ 590.Index No. Watson.94 1.0 64.3 6.38 1.2 New River at Rock River. Cameo. at Pauba Canyon.0 Moencopi Wash nr..6 2000. mi* Salt River at Roosevelt.46 1.0 Blue R.4 3.3 187.050 .0 Plateau Cr. CA 8::: 6.33 1. nr. Phoenix.053 .“i. Saugus. CA 1. CA 162.2 2:o Eaton Wash at Eaton Wash Dam.20 1.0 14.0 9:2 95.0 Puerto R.35 1. Santee.56 1. nr.0 Verde R.38 1. at Conner No. nr.30 1.0 Tonto Cr. CO 7. at Bell Road nr.5 t?::C%%?if . Hesperia. with Blue R.1 10.4 1.036 .1 Santa Margar@ R. San Gabriel R.43 1.79 1.0 12. Tuba City.75 0. New R. AZ 10. CA 3.046 .7 i:: 5.7 67. AZ 85. CA 9.35 1.0 14.8 81.3 790.20 1.59 1. AZ 20. CA 380. AZ 16.40 1.048 . at Foster.0 San Francisco R. at Santa Anita Dam. Drainage By~ayj@.052 . 16 34 z: 37 38 lag data for the Southwest Desert.4 San Gabriel R.35 1.055 . nr.38 1..058 . AZ 473.51 1.0 220.8 Tujunga Cr.047 .058 .1 0.069 . AZ 1570.5 Agua Fria R.:m?A “E 0.03 0. at Cogswell Dam.30 1.OtiU .0 16.0 18. nr.0 760. nr.9 Los Angeles R.6 0.. nr.-Unit hvdrograph 0. AZ Skunk Cr. AZ 21.6 355. Claremont. Great Basin.0 E 12. Clifton.046 . and Colorado Plateau. at Temecula.052 .2 2%:: 570. above E.48 1.0 1261.4 Table 4-3. at San Gabriel Dam. CA 24.043 . at Lees Ferry.0 1688.5 Murrieta Cr. nr.0 Clear Cr.77 1. at Sepulveda Dam. at Workman Mill Rd. at Jet.061 .5 2840.7 4020. CA 75.7 168.z White R. at San Dimas Dam. above Gun Cr. nr.5 1.38 2.046 .0 1473:o Paria R. CO 16.0 2.051 .33 1. CA 152.054 .0 24.07l . nr.64 1. CA 40.062 .83 1. Winslow.0 9. nr.85 1.: Sand Dimas Cr.056 .56 4 6 f F 0.ght:t% Station and location area.0 352. at Big Tujunga Dam.8 28. Phoenix.4 Dee Cr.2 4730.0 Gila R. AZAZ New Springs.1 Bill %illiims R at Planet.053 . AZ 3190.063 .053 .91 . Verde and below Jerome. Admana.3 Pacoima Wash at Pacoima Dam.0 63.3 West Fk.0 ‘228i.052 .61 1. CA 27. AZ 15:9 2760.9 604.:%cd. t.5 San Jose Cr.0 66. Fallbrook.032 :E . Mayor.057 .2 7.0 22-: 10. CA 16. CA 5.035 1.

Figure 4-9 and data in table 4-4 represent conditions in the Sierra Nevada of California. Oregon. considering the few points shown onfigure 4-3 at or near the lower value.88 to 0. A highdensity development combined with a good collection system is typical 83 . or a lag equation coefficient C. At the low end of the K.HYDROGRAPH DETERMINATIONS partially urbanized basins in the desert region.65.160. a value of 0.033 to 0. respectively. As shown by the relationships on figure 4-11. from 0. reflects the increased hydraulic efficiency normally found to be associated with the urbanization of a drainage basin.150. care should be taken before selecting a low K. However.34. below the two limiting curves. ranges from a high of 0. However. show that K. when attempting to assign values representative of those present during a PMF. This is indicative of very heavy coniferous growth extending into the overbank flood plain. Basins in this region normally have well-developed drainage networks and substantial coniferous growth throughout those parts of the basins above about elevation 2000. values. which lowers the hydraulic efficiencies of these basins. at 0. to ensure that the basin being studied has essentially the same hydraulic efficiency characteristics in terms of geology. River and stream channels are generally well incised into the bedrock. which is the reason for establishing the upper limit K. primarily reflects the density and type of development and extent to which engineered floodwater collection systems have been constructed. a conservative approach should be taken and values assigned toward the low end of the scale for those conditions cited in the previous paragraphs. Again. range. The hydrologic and hydraulic characteristics of the Sierra Nevada basins are mostly quite similar to those of the Rocky Mountains. . Therefore.42.013. Figure 4-10 and the accompanying data tabulated in table 4-5 represent conditions in the Coast-and Cascade Ranges of California. Relationships shown on figure 4-10 indicate at the high end of the K. intense storms. drainage network development. used for generating PMF’s for basins in the Rocky Mountain region.1) is typical of the low lying basins where considerably sparser vegetation results in a higher hydraulic efficiency.080 (C. the range in K. range. the greater amount of data available for the Sierra Nevada reflect flood hydrograph reconstitutions for flood events resulting from major.064 with associated lag equation coefficients of 4.0 to 1. of 2. it should be emphasized that the lag data accumulated for a condition may not reflect the hydraulic conditions present in a PMF event. a value of 0. and stream hydraulic roughness as those in the data set shown in table 4-4.reflecting the varying degrees of hydraulic efficiency. of 3. respectively. Relationships shown on figure 4-9. Such is not the case for the Rockies.9.150 to a low of 0. or a lag equation coefficient of 0. and Washington. The position of this curve. Associated lag equation coefficients range from 0. Figure 4-11 and data in table 4-6 represent urban conditions at several locations throughout the United States.

51 1. Calaveras R. CA North Fk.76 4. below Tamarack Cr. nr.0 64. CA Kings R.03 3. San Andreas.4 30.03 3. nr. CA Kaweah R.16 1.9 388.065 .I55 .0 1261.07 2. nr.0 990. CA Calaveritas Cr.I35 .I06 .0 164.54 2. CA Kern R.6 15.I23 .0 6.0 2075. Consumnes R. CA Cosumnes R.143 . San Andreas.46 2. Cliff Camp.0 363.1 25.44 2.67 4.9 21.4 30.P 1. CA American R. CA San Joaquin R.6 .0 290.5 13. Reservoir.77 2.93 3.I22 . Kings R.-Unit hydrograph lag data for the Sierra Nevada in California.5 10.0 53.33 3.056 . at Bullard’s Bar Dam. at Melones Dam.068 .5 17. CA area.42 1.3 Basin factor. LL ca /.0 497.0 66. at Success Dam.0 8.8 10.0 235. Cosumnes R.3 36.118 . at Terminus Dam. CA Y&a R. at Pine Flat Dam. at Folsom Dam.2 12. Drainage area. River Pines.. CA Stanislaus R.94 iz 4.2 9.72 3.0 560.0 1875.4 168.0 18. CA Tule R.7 7.I32 .2 8. CA Big Dry Cr. CA North Fk.7 31. CA South Fk. Index No. at Isabella Dam.I41 .6 10.17 3. nr.I36 .I09 .4 6.4 85.83 3.093 .083 . CA Calaveras R.0 1542.2 9.0 17.5 269. nr.8 11.Table 4-4.151 .7 time.7 395.0 20.5 13. nr.I55 . CA North Fk.7 ‘70.2 133.0 4.2 9.43 2.17 3. CA North Yuba R. at Michigan Bar. at Calaveras Reservoir.6 98. CA Cosgrove Cr.4 6.6 7. at Cosumnes Mine. at Friant Dam.0 537. Kings R.0 897. at Englebright Dam.0 5.I13 ct 3.6 15.0 481.7 8.I22 . below Rancheria.I28 .20 3. CA Calavetas R.69 3. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ‘Contributing Station and location Pitman Cr. mi2 22.4 16. at Hogan Reservoir.0 86.0 ‘116.5 7.0 143.0 8. Valley Springs.094 . CA North Fk. CA Woods Cr. Jacksonville. L ho&s Lag Kz 0.

CA Sisquoc R. CA San Lorenzo R.0 465.4 5. nr. LL.086 .86 3.6 63. CA East Fk. at St.09 2.9 3.0 8.6 5.09 3.3 37.153 .0 38.37 3.129 . mi9 577. at Boyes Hot Springs.llO . Drainage area. Helena.90 2.0 10. CA Basin factor.9 4.8 9.97 6.0 45. CA Uvas Cr.2 10.4 43.94 2.0 3.2 4.1 20. nr. nr. CA Corralitos Cr. Russian R. 0.107 . OR Arroyo Del Valle nr. Williams. Branscomb.0 288.13l .104 .0 119.P 190. CA Stony Cr.78 3.43 3.11 3.9 16.082 .-Unit hydrograph lag data for the Coast and Cascade ranges in California.0 5. CA Huasna R. CA Napa R.0 18.0 114.12 3.72 3.3 17. Winters.47 2.9 5.8 2. nr. CA Trinity R.76 3. L ho&s 17.0 8.0 14.8 4. Carey. nr. Hopland.096 G 3.0 Oregon. nr.129 . CA Powell Cr.4 3.70 2. at Palo Alto. nr.120 . CA Russian R. Hamilton City. Stanford University.98 2.4 4. nr.0 17.132 .4 6. CA Feliz Cr.5 21.143 . at Hayward. Cadzadero.2 81. and Washington.5 62. CA Redwood Cr.0 30. CA Salinas R.0 4. CA Sonoma Cr.145 .143 .6 726. Novato.5 10.24 3. at Lewiston.77 4. at Ross. CA San Francisquito Cr.0 0. nr.6 30. CA South Fk.139 .8 3.150 .7 3./.9 147.2 278.13 2.35 3.076 . nr.095 .lO6 .0 17. OR Slate Cr.09 2.113 . at Santa Cruz.0 2.0 0. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 i: 23 24 25 26 27 28 Station and location Putah Cr. CA Matadero Cr.7 4. at Big Trees. Eel R.5 157.1 1.98 2.081 . Wonder.1 3. CA Pinole Cr. at Orick. San Lorenzo Cr.9 17.5 4.3 7.04 3. nr.8 66.72 3. Santa Maria. Corralitos.61 3.8 7. Napa. lag time.119 .6 6.168 .6 10. nr.I g .0 Index No.4 76.8 4.35 1. nr. CA Dry Cr.4 4. nr. Pinole.7 6.8 17.119 .7 14.120 .8 2.117 .12 3.9 3.41 2.5 4.8 8.4 31.8 4. CA Novato Cr. at Morgan Hill..0 99. CA Austin Cr.5 1. CA Corte Madera Cr.50 CA z 5 8 B g 9 a .Table 4-5. Calpella. nr.0 170. Pozo.1 93. at Ukiah.47 2. CA Branciforte Cr.8 6.1 111.119 .0 8. nr. Livermore.0 764.

017 . VA Walker Avenue Drain at Baltimore.62 0.01 0. % time.4 0.011 .... below Hooper Ave.2 0.4 0. TX Boneyard Cr..0 81. L.4 4.1 9. at Sawtelle Blvd.030 . L ho&s 0.4 0.5 6.52 0.91 0.017 . Whittier.034 .034 .26 4. Louisville. Louisville. VA Tripps Run at Falls Church.-Unit hydrograph lag data for urban basins..36 0.31 0.3 0.6 9.2 1.75 0.8 1.0 0. CA Brays Bayou. Louisville.4 0.2 1.5 88.6 0..83 0.25 0.6 2.8 1.7 0. TX Wailer Cr.032 . CA Ballona Cr. Louisville.024 .014 . TX Beargrass Cr..2 Basin factor.024 . CA San Jose Cr.0 134.7 8.86 0. VA Four Mile Run at Alexandria.4 92. TX White Oak Bayou.8 4.04 s 9 0./P 5 4.0 4. KY Southwest Outfall. Monterey Park. at Workman Mill Rd.15 0.86 0.6 1.5 1.04 0.8 0.62 0.1 0.029 . Louisville.3 121.023 .4 2.9 0.014 .A.033 .29 0.1 3.020 . KY 17th Street Sewer.3 1.6 88.. CA Broadway Drain at Raymond Dike. Storm Drain.017 . 8 2 B 0 Index No.4 5.78 0.88 0. KY Southern Outfall.012 .3 2. Austin. KY Northwest Trunk. L.1 0.8 14. mi* 14.60 0. KY Beargrass Cr.026 . MD Drainage area. Louisville. KY Tripps Run nr.022 Cl LL.5 19.2 KZ 0. CA Compton Cr..5 4.4 1. VA Piney Branch at Vienna.033 .9 0.2 0.0 0. Houston.57 .7 0. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Station and location Alhambra Wash above Short St.1 3.0 1.8 24.2 2. Austin.Table 4-6.44 0.50 1.88 0.44 0.3 0. Falls Church.68 0.035 .A. Houston.9 6.36 0.4 7. VA Little Pimmit Run at Arlington.44 0.3 4.

for each of the five geographic regions and for the urbanized conditions. it is imperative that anticipated future developments be considered. values. Dimensionless Unit Hydrograph or S-Graph Selection. To provide adequate definition near and at the peak of the unit hydrograph as well as the eventual flood hydrograph. This process for each of the two techniques currently in use by the Bureau is discussed in the following two subsections. (1) Dimensionless unit hydrograph technique. The dimensionless unit hydrograph or S-graph selected should represent a drainage basin that is as similar to the ungauged basin as possible with respect to basin shape. the analyst must resort to relationships for nearby hydrologically homogeneous basins. do not include a flood suitable for reconstitution. vegetal cover. Most urban development eventually tends to become high-density and. Currently.5 [ 5 51. The most appropriate dimensionless unit hydrograph or S-graph to be used for a flood study would be one developed from analysis of a flood occurring in the basin under study. and each regional office has copies of those that are applicable to drainage basins within their regional boundaries.HYDROGRAPH DETERMINATIONS of fully urbanized drainage basins with the lower K. also tends to have more extensive storm water collection systems. drainage network development. Low-density or partial development with only minor floodwater collection facilities are typical of basins with the higher K. the unit duration should approximate the lag time divided by 5. values. (g) Computing Synthetic Unit Hydrograph Ordinates. However. this is usually impossible because the records. and channel hydraulic characteristics. topography.-The first item that must be determined is the unit duration of the synthetic unit hydrograph. (f. Therefore. if any. Examination of these relationships has led to the conclusion that the typical ones shown in tables 4-7 through 4-18. with continued flooding problems. will produce satisfactory results in unit hydrograph development.. values that could reasonably be expected with full future development during the functional life of the project being studied.-Data gath- ered during the field reconnaissance along with a careful inspection of the drainage basin features as outlined and presented on the topographic map are compared with similar data for basins where unit hydrographs have been developed from analysis of observed flood hydrographs.-Once the unit hydrograph lag time has been determined and the dimensionless unit hydrograph or S-graph has been selected. The hydrologic engineer must anticipate such eventualities and assign lower K. it is basically a mechanical process to determine the synthetic unit hydrograph ordinates. or no flood event may have occurred. the Bureau’s Denver Office has a catalog of over 30 of these relationships representing conditions for basins throughout the 17 Western States. As a result. However. This calculated unit duration 87 . the hydrologicengineer can improve results if sufficient knowledge of the drainage basin being studied is gained through field reconnaissances so that the most appropriate selection is made.

27 .40 10.46 . Time t.79 2.44 .52 2.41 2.26 7.50 9.22 .92 1.26 .04 13.37 16.24 2.18 .40 22.5 D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 q 0.78 .58 .08 2.27 1.23 1.05 1.25 18.65 .88 8.32 Lg + 0.72 1.59 8.18 .16 s 2 .10 14.75 .54 1.20 31 1.96 6.81 20.26 .19 21. + 0.14 1.83 7.70 . in % of Time t.66 1. in % of 5 8 0 Time t.04 3.95 5.67 L.52 12.81 .17 .32 2.25 .16 .94 .10 1.29 .91 .37 1.79 1.48 1.64 4.36 5.40 .07 20.18 11.72 .19 .66 3.53 3.32 1.38 .20 .59 18.00 1.85 1.91 24.84 .5 D 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 .23 .48 .52 .65 Time t.87 .60 .51 11.02 22.05 4.56 .5 D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 q 15.37 .43 .21 .50 .54 .78 3.59 1.20 .5 D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1.39 4.15 2.Table 4-7.24 .62 .30 .06 9.18 1.24 .5 D 5 10 :i 25 30 Lg + 0.E 300 q 3.03 16. in % of 4 0. in % of unit hydrograph data for the Great Plains.10 .02 0.18 2. in % of Time t.45 5. + 0. in % of Lg + 0.41 .98 2.34 .38 L.33 .29 .5 D 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 q 0.67 2.42 Lg + 0.35 .-Dimensionless Time t.57 6.98 .23 4.

25 90.58 97.31 5.61 89.66 Time t.19 94.86 90.05 85.-Dimensionless Time t. in % of 4 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 Discharge.85 96.70 81.11 9.29 98.20 20.19 96.22 87.95 100.89 93.19 86.42 85.39 46.15 99.15 75.93 99.23 37.71 65.34 97.52 1.54 82.04 89.33 91.55 98. in % of LA505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 Discharge.99 92.74 69.35 96.17 28.46 89.95 97. in % of 4 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Discharge.43 94.28 95.43 63.47 98.69 97.94 61.46 76.66 95.21 .81 99.91 99.62 90.00 Discharge 96 of ultimate 83.70 88.59 99. % of ultimate 98.02 96.08 99. % of ultimate 0. Time t.87 95.80 80.78 Time t.76 16.70 93.11 2.53 71.38 98.70 98.86 99.98 91.44 93. % of ultimate 96.16 88.65 99.80 97.57 33.53 99.00 90 95 100 190 195 200 .48 99.70 99. 65 70 75 80 85 Discharge.60 92.Table 4-8.08 95.95 42.29 99.01 3.07 Time t.76 84.00 Time t.65 94.08 97.28 56.02 24.73 74.91 98.22 99.63 86.81 67.21 97.66 91.20 98.25 58.01 98. % of ultimate 53.95 94.40 50.17 93.80 78.85 98.84 79.30 99.35 99.70 12.81 96.67 77.51 96.48 95. in % of L&T 5 10 15 20 25 30 35 40 45 50 2. % of ultimate 92.76 99.63 98.02 7. in % of 4 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 S-graph data for the Great Plains.06 .41 99. in % of -% 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 Discharge.46 97.02 .72 87.20 72.33 83.11 98.

80 8.80 6.44 .5 D 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 q 3.28 1.86 .-General Time t. + 0.63 .87 2.51 9.31 .20 .83 .63 1.60 .71 Lg + 0. in % of data for the Rocky Mountains.5 D 5 10 15 20 -25 30 35 45 40 50 55 :.00 5.5 D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 q 0.65 2.00 5.10 3. in % of Time in % of t.54 .77 L.81 24.97 .90 Lg + 0.26 .01 9.30 4.50 1.5 D 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 5 9 .90 2.40 6.91 15.65 .25 6.34 .91 11.40 3.00 18.71 .99 1.07 1.70 Lg + 0.40 .28 .27 .31 16.22 .Table 4-9. q 0.15 2.50 .00 4.29 .24 2.17 2 g 0 Lg + 0.21 .32 .34 1.26 .5 D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 q 11.00 6.10 3. in % of Time t.01 21.42 .25 7.75 2.05 1.40 8.80 4.45 1. in % of unit hydrograph Time t.33 .35 Lg + 0.33 2.70 7.21 19.48 .41 .21 14.23 .41 12.01 0.00 2.39 1.51 22.24 .90 .55 3.37 .74 .23 1. Time t.19 .13 1.19 .11 21.00 7.68 .57 1.56 .58 . 70 75 ii 90 95 100 q 0.00 3. in % of storm dimensionless Time t.38 .5 D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1.76 1.91 1.61 10.21 13.52 2.93 .55 4.25 .35 5.80 .83 1.09 1.46 .21 10.65 5.72 3.52 .42 2.17 .19 1.25 3.70 14.18 .23 .

57 44. % of ultimate 52.21 98.00 Time t.41 88.64 99.43 80.06 81.89 98.97 93.66 97.85 99.82 98.01 97.67 92.58 98.58 94.65 95.08 64.17 90.03 95.Table ClO.50 68.58 34.71 96.21 96.29 97. in % of L&T 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Discharge.45 95.26 83.00 98.51 70. % of ultimate 96. in % of -5 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 Discharge.18 85.19 99.55 96.55 47.40 98. in % of L* 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Discharge.95 91.68 92.46 4.26 storm dimensionless Time t.53 75.42 97.62 1.05 69.35 50.75 99. in % of 4 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 Discharge.52 99.75 90.93 84.88 72.87 89.21 61.00 50 55 60 65 70 ii !z 95 100 .42 87.11 98.96 99.89 98.61 78.08 94.39 74.17 73.47 26.05 . in % of 47 5 10 15 20 25 Discharge % of ultimate 0.04 99.85 96.78 97.26 93.32 S-graph data for the Rocky Mountains.92 88.62 76.15 97.49 98.57 90.31 98.33 12.82 94.20 2.84 21.80 99.-General Time t.86 97.11 99.66 98.83 82. % of ultimate 81.15 3.64 77.56 83. in % of 4 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 Discharge % of ultimate 91.20 63.89 87.54 97.58 99.25 95.33 99.87 57.97 6. % of ultimate 98.66 38. Time t.35 92.74 16.69 99.51 54.35 86.17 30.02 92.03 96.72 9.57 85.38 Time t.32 41.26 99.32 89.55 93.54 79.39 99.81 95.78 86.95 100.10 59.84 66.46 99.33 94.74 Time t.23 .90 99.

in % of data for the Rocky Mountains.59 .11 .48 .81 14.86 11.36 1.61 10.76 18.55 4.38 . q 0.74 6.00 0.55 1.5 D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 q 20. in % of Time in % of t.24 1.68 L. in % of Time t.15 . + 0.lO . Time t.62 3.09 .lO .18 .21 .57 .30 25.44 2. in % of dimensionless Time t.62 1.04 9.14 .65 .16 .92 .74 .16 .62 .50 .96 .13 .75 -2.42 1.76 23.31 2.99 12.5 D 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 6 2 i% .14 .32 .36 .41 7.29 .09 Lg + 0.04 1.68 .70 18.84 16.89 3.77 .69 5.12 .17 .47 4.97 4.12 .20 .5 D 205 210 215 220 225 230 235 240 245 250 255 260 261 270 275 280 285 290 295 300 q 2.14 1.33 .95 1.76 1.43 .13 2.13 .34 3.71 22.98 5.5 D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1.42 .26 .71 .19 1.27 .Table 4-l l.19 L.17 .84 1.25 3.21 .52 8.08 5 g 0 Lg + 0.93 Lg + 0.40 .30 1.17 2.55 .-Thunderstorm Time t.5 D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 q 0.5 D 5 10 15 20 25 :: 40 45 50 55 :1: 70 75 80 98: 95 100 q 0.81 .23 20.23 .61 2.35 .69 1.25 .51 .36 26.ll .84 1.33 .53 24.59 3. in % of unit hydrograph Time t.28 .26 15.38 8.52 .88 .94 13.09 .49 1.46 Lg + 0. + 0.24 .50 6.83 28.22 .15 .05 1.31 .14 .84 .

-Thunderstorm Time t.14 65.57 84.44 88.12 94.32 70.63 94.82 99.73 99.39 19.91 98.76 98.69 98.17 99.28 96.81 97.87 88.02 79.01 Time t.78 99.27 93.55 81.77 96.11 98. in % of -% 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 Discharge. in % of Lg 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 Discharge. in % of J% 5 10 15 20 iz 35 40 45 50 ii 65 70 2 85 90 95 100 Discharge. % of ultimate 99.39 90.53 78.09 95.94 99.21 98.62 99.36 Time t.41 23.45 96.30 98.20 97.48 62.86 11.62 98.96 93.38 94.64 92.22 85.92 97.60 91.95 90.69 97.52 99.58 99.57 93. in % of -% 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Discharge.03 .24 .88 99.40 99.83 72.81 91.14 .65 83.85 94. % of ultimate 0.72 95.43 58.39 2.91 99.86 76.99 45.96 92. Time t.44 85.65 99.76 99. in % of 4 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 dimensionless Discharge % of ultimate 87.Table 4-12.57 4.18 50.22 91.47 89.46 97.31 8.97 89.38 98.92 28.35 80. % of ultimate 54.06 99.44 99.64 87.31 99.00 St .43 39.82 98.52 95.92 96. % of ultimate 95. in % of 4 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Discharge % of ultimate 98.65 82.70 1.86 S-graph data for the Rocky Mountains.42 68.93 34.58 97.21 6.31 92.68 99.06 97.61 96.10 96.27 99.85 99.40 .95 86.97 100.46 98.31 95.07 .22 99.95 99.98 74.71 99.12 99.89 98.01 99.88 15.27 Time t.33 97.55 99.54 98.48 99.00 Time t.

87 3.21 1. P E E .55 Lg + 0.74 1.08 14.40 Lg+0.68 5.50 1.50.5 D 505 510 515 520 K .33 .12 .61 8.47 2. .33 2.08 1.18 21.66 .82 .63 3.43 1.5 D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 q 18.26 .63 2.59 1.97 .21 Lg + 0. Time t.12 .59 .15 1.78 1.53 .04 9.lO in % of 8 8 Lg + 0.20 7.86 .69 .07 24.22 .36 6.99 1.47 .31 4.19 .91 26.41 16.21 . in % of Time t.13 Lg + 0.93 2.38 .84 4.99 9.22 2.28 3.50 23.47 3.10 3.36 1.5 D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1.78 .75 2. Great Basin.15 4.34 .42 .45 .Table 4-13.5 D 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 q 3. + 0.17 -16 .63 .75 28.15 .74 .61 11.5 D 5 10 15 20 25 30 35 40 45 :t 60 65 70 75 80 !E 95 100 q 0.48 .27 .28 1.32 .20 .04 8. in % of Plateau.97 27.92 16.50 20. in % of Time t.20 5.36 .47 5.10 1.10 2. and Colorado Time t.19 . in % of Time t.47 12.24 -23 .-Dimensionless Time t.56 .18 .5D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 q 0.57 4.28 .78 6.88 1.91 .81 2.19 12.38 28.68 L.02 0.83 5.15 . q 0.68 1.30 .I1 . in % of unit hydrograph data for the Southwest Desert.

20 .36 88.68 89.14 93.64 40.34 73.13 99.28 80.25 98.87 94. in % of 4 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 Discharge.92 98.57 98.36 98.00 Time t.44 82.34 .03 23.69 99.77 99.88 92.21 94.93 91.56 96.85 99. % of ultimate 95.89 99. % of ultimate 86.93 99.19 57.66 Time t.12 97.47 76.75 98.29 89.33 68.57 .64 87.04 .60 97.09 85.97 100.42 91.88 98.35 96.61 99.94 97.55 18.02’ 63.65 99.22 28.81 z .40 2..84 78.83 95. in % of 4 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Discharge % of ultimate 54.74 97.90 34.89 96. and Colorado Dis- Plateau.13 96.47 98.47 99.08 4.32 92.52 99.43 83.83 66.29 97.73 99.39 Desert.06 99.12 95.57 99.37 99.42 99.74 93.25 85.9 40 45 50 2.Table 4-14.00 Time t.lO .40 81.75 96.86 90.91 1. Disin % of charge.53 70. Tmiy.99 99.52 94.f 4 505 510 515 520 525 Discharge.57 6.41 90. 65 70 ii izl 95 100 Discharge.15 45. Great Basin.10 79.37 84.53 72.32 99.04 88.79 9.45 97.86 61. in % of 4 5 10 15 20 25 .88 S-graph data for the Southwest Time t.20 99.30 50.99 75.79 13.84 98.65 95.-Dimensionless Time t.26 99. % of 4 ultimate 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 98.08 3.51 93. % of ultimate 99.14 98. in % of 4 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 charge. % of ultimate 0.01 98.

83 13.77 .22 1.24 4.75 .87 .05 20.58 3.90 .25 4.62 1. in % of data for the Sierra Nevada.28 .73 3.33 2.26 Lg + 0.30 1.13 I .63 .52 1.16 .40 .14 1.17 9.11 1.69 6.34 1.85 4.52 .95 2.63 4.47 1.99 5.36 5.69 2.42 1.65 1.59 21.43 .73 8.89 3. in % of Time t.53 11.5 D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 q 1.49 .15 6.07 4.00 0.79 2.26 1.17 15.08 2.26 2.31 .46 505 510 515 520 525 530 535 540 545 550 555 560 0.84 .92 1.82 .14 .77 16.05 1.88 2.83 12.60 3.63 18. in % of Lg + 0. and Cascade ranges.25 .36 10.19 . in % of Time t.34 12.50 2.38 .: 100 q 0.17 7.-Dimensionless Time t.58 .5 D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 q 13.73 1.33 5.61 .29 9.65 7.23 5.41 2.69 .30 1.66 .11 2.54 23.22 .33 8.98 Lg + 0.60 2.5 D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1. Coast.88 18. Time t.03 Lg + 0.18 1.93 .34 .67 1.38 1.06 Lg + 0.13 15.06 1.30 3.18 2.51 7.98 2.48 19.85 1.67 5. in % of unit hydrograph Time t.Table 4-15.80 .55 .5 D 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 q 3.5 D 5 10 15 20 25 30 35 40 45 50 2: 65 70 ii 85 .44 3.57 1.96 .72 .78 1.43 4.18 3.

25 81.04 74.23 99.90 98. in % of Lb205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 Discharge.40 89.35 84.69 94.22 98.70 a-grapn aara ror me xerra Time t.47 87.47 88.00 84.91 50.98 88.84 90.32 57.38 97.68 83.27 90.22 93.53 93.13 94.33 99.73 97.01 69.04 8.50 46.14 .22 95.60 59.66 99.00 78.89 93. % of ultimate 80.84 99.39 96. % of ultimate 99.35 65.56 68.44 2.13 75.10 78.64 21.94 87.16 76.lame 4-lo.40 86. Discharge % of ultimate 91. in % of 4 505 510 515 520 525 530 535 540 545 Discharge.51 98.24 85.89 73.47 95.56 97.00 97.06 17.08 Time t.01 66.73 99.55 92.59 Time t. Loax.66 61.21 10.17 3.78 43.--vimenwomess Time t.69 91.89 99.43 .94 89.38 6.94 96.51 99.41 94.79 99. in % of L&z 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Discharge 96 of ultimate 52.60 96.71 95.00 5 .28 35.13 4.19 97.96 99.73 26.83 94.92 99.99 100. in % of L&T 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Discharge % of ultimate 97.00 Time t.13 99.90 99.96 95.67 71.87 79.63 85.15 77.86 1.79 55.78 98.84 92.81 ana Lascaue ranges.42 99.36 98.38 70.64 98.17 96.02 99. in 96 of L&T 5 10 15 20 25 30 35 40 45 50 2: 65 70 75 if 90 95 100 Discharge % of ultimate 0.72 39.57 63.49 81.94 14. in % of -% 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 luevaaa.47 91.06 98. Time t.83 86.42 31.20 92.98 82.

31 .21 .56 2.94 5.31 10.43 . in % of Lg+0.5 D 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Time t.5D 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 600 a g s 2 q 0.67 2. in % of Lg+0.49 1.5D 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 q 1.60 1.27 9.07 1.36 .50 .25 .95 8.50 13.37 3.39 1.52 .-Dimensionless Time t.34 .42 1.72 16.48 .08 12.27 .47 2.84 q 0.69 .23 .21 zi .00 0.74 20.27 19.5D 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 unit h‘YC lrograph data for m-ban basins.5D 405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Time t.87 11.18 17.67 .78 1.52 3.24 .37 2.30 2.93 q 3.Table 4-17.33 .45 .12 2.90 1.19 15. in % of Lg + 0.97 .75 .63 4.73 .5D 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Time t.74 19. in % of Lg+0.39 .00 19.63 8.52 13.92 4. in % of Lg+0.27 5.75 6.15 1.62 .80 6. in Y0 of L.03 1.30 .46 . Time t.24 .32 19.83 1.39 4.75 2.27 7.21 1.38 9.64 1.27 17.64 1.49 .04 1.60 .81 .26 .53 2.64 .+O.24 3.93 .93 2.90 .41 q 0.78 .96 8.53 1.29 .54 .57 4.32 .18 3.04 2.37 .11 1.12 q 14.55 5.21 2.22 6.98 1.36 5.55 3.32 1.56 .22 .28 1.87 .23 1.28 .24 4.73 3.40 .58 .71 Time t.19 11.75 7.

59 71.91 98.75 84.77 95.99 100.45 98.97 Time t.67 99.45 99.42 62.66 Time t.71 34. in % of 4405 410 415 420 425 430 435 440 445 450 455 460 465 470 475 480 485 490 495 500 Discharge % of ultimate 96.67 12.91 76.64 58.13 97.13 89.71 74.22 88.15 95. in % of L&T 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 Discharge % of ultimate 82.14 .12 43.85 S-graph data for urban basins.60 86.47 30.62 99.73 98.06 83. Time t.11 5.04 1.33 96.04 79.77 91.95 39.95 94.57 95. in % of L&T 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 Discharge % of ultimate 52. % of ultimate 99.33 99.82 93.57 99.21 15.17 86.14 91.92 95.10 78.67 97.98 97.79 97.53 64.35 98.49 9.51 22.42 73.09 46.15 96.81 98.68 89.00 .64 98. in % of L.12 93.87 99.40 93.26 99.18 92.46 94.39 99.03 98.64 7.51 92.94 80. in % of Ls 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375 380 385 390 395 400 Discharge % of ultimate 92.78 81.84 4.06 72.54 98.50 66.27 97.94 55.38 90.40 85.13 60.50 96.91 99.72 50.95 99.-Dimensionless Time t.98 90.82 2.68 93.01 69.00 Time t.58 Time t.82 96.21 94.34 83.77 99.23 87.Table 4-18.14 98.82 99.69 94.50 91.19 99.02 85.89 98.54 97.48 1.05 99.72 99.51 99.7 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Discharge % of ultimate 0.33 26.25 98.96 96.12 99.04 77. in % of Lg 505 510 515 520 525 530 535 540 545 550 555 560 565 570 575 580 585 590 595 Discharge.14 18.71 87.10 79.56 89.41 97.32 68.36 95.73 88.

As previously discussed.2 hours. or 8. The first column in the table lists the time in hours. and then dividing by the volume of 1 inch of runoff for the basin under study. The synthetic unit hydrograph ordinates Q. the basin should probably be divided into subbasins and an individual unit hydrograph developed for each. or 6 hours.5 = 2.50 (percent of lag time plus semiduration of unit rainfall).067 l-day cubic feet per second. 2. or 30 minutes. The ordinates so developed represent the unit discharge at the end of the respective time period.89. The value of lag time plus semiduration of unit rainfall is equal to 12 + 2/2. This methodology is explained by the following example: Consider a 300~square mile drainage basin whose unit hydrograph lag time has been determined to be 12 hours. use 2 hours) has been selected for use in developing the unit hydrograph. If the result is greater than 6 hours. Also assume that the dimensionless unit hydrograph shown on figure 4-12 has been selected as the basis for developing the unit hydrograph for the basin in question. a table is set up as shown on figure 4-12. The runoff hydrographs resulting from the application of rainfall to each subbasin will be routed and combined at the concentration point as discussed in chapter 5. are determined by dividing the corresponding value in the first column by the lag time plus semiduration value. 100 .FLOOD HYDROLOGY MANUAL should always be rounded down to the closest of 5. resulting from multiplying values in the third column by the quotient of the l-inch runoff volume divided by the lag plus semiduration value are listed in the fourth column. The volume of 1 inch of runoff is equal to 300 times the conversion factor 26. Since the lag time is known and the volume of 1 inch of runoff can be determined using the drainage basin’s area in square miles. or 13 hours. and that a unit duration of 2 hours (12 hours/5..89 is used to convert 1 inch of rainfall excess over a l-square mile area in 24 hours to runoff expressed as l-day cubic feet per second. the selected dimensionless unit hydrograph can be used to compute the unit hydrograph. percent of Lg + 0. and the ordinate is expressed in terms of discharge multiplied by the value found by adding lag time and semiduration of unit rainfall. This is generally the case for basins exceeding 500 square miles. or 1. with each increment equal to the unit rainfall duration. The first factor of 26. the dimensionless unit hydrograph is expressed in terms of time in percent of lag plus the semiduration of unit rainfall on the abscissa. 15. After these values are determined. 10. Values in the second column. Values in the third column are obtained by reading the ordinate value from the dimensionless unit hydrograph for the corresponding percent of lag plus semiduration value in the second column. and multiplying by 100 to convert to percent. 3. resulting in the flood hydrograph for the total basin.

673 ’ Read from graph (a) *Determined by rearranging equation (3) to yield: Q=q(V’/~~+0.924 5.0 46. I 9. 0 3 s IO Q8 I6 g4 :: 4 5 0 oo" 14 12 2 5 I8 a. 0 246 1. \2hour unit hydrograph 3333 6 I2 I8 24 30 36 4248 TIME IN HOURS (b) and sample computations.0 0. 103-D-1856. hours 0 2 4 6 0 Time in percent of Lg + 0. for D Since this is a synthetic unit hydrograph derivation.16 Time.5D = 13 hours z 022 s 20 v) ii IA.SYNTHETIC UNIT HYDROGRAPH DERIVATION Drainage area = 300 square miles Lag time Lg= 12 hours Unit duration D = 2 hours Unit volume of runoff from basin.067 I-day cubic feet per second Lg+0.and substituting Q.6 Dkmensionless unit hydrograph ordmate’.5D).2 61.50 (0) 400 5 Figure 4-12.-Dimensionless unit hydrograph . Q. V’=6.5D 0.2 Synthetic unit hydragraph ordinate*.4 30.0 17. q 0. I. 0 100 PERCENT 200 300 OF Lg + 0.4 3.584 10.0 15.

and is selected for use in this example. in square miles. the unit duration is the first item to be determined. The theoretical unit duration is then 12/5. The conversion factor 645. the dimensionless S-graph is expressed in terms of time in percent of unit hydrograph lag time on the abscissa and discharge as a percentage of ultimate discharge on the ordinate.6)/2. The dimensionless S-graph shown on figure 4-13 is assumed to be appropriate for the hypothetical basin under consideration.3 and then dividing the result by the unit duration of rainfall. application of these values to the appropriate dimensionless S-graph yields a synthetic unit hydrograph. The ultimate discharge for this basin from a continuous series of rainfall increments of 1 inch in every S-hour period would be 250 (646. Each value in the third column is multiplied by the ultimate discharge 102 . is then developed as shown in table 4-19. and is found by dividing the time value in the first column by the unit hydrograph lag time. the final synthetic unit hydrograph ordinates used in developing a flood hydrograph should reflect values read from the curve rather than the computed values. truncated at the 16th hour for brevity. time is tabulated in the first column at increments equal to the unit duration. A graphical representation of the final synthetic unit hydrograph and a tabulation of the final ordinates should be included in every flood study report. Since the curve will not pass through all the points.700 2-hour cubic feet per second. The ultimate discharge is an equilibrium rate of. The ultimate discharge for a drainage basin is found by multiplying the drainage area. they should be plotted on graph paper and a smooth curve drawn through the points. In table 4-19. or 80.5. Values in the third column represent ordinate values read from the dimensionless S-graph at the corresponding time in percent of lag values shown in second column. discharge achieved at the time when the entire drainage basin is contributing runoff at the concentration point from a continuous series of unit rainfall excess increments. The synthetic unit hydrograph. With both the lag time and ultimate rate of discharge known. As previously stated.-As was the case with the dimensionless unit hydrograph technique. as described in the following example: Consider a drainage basin with an area of 250 square miles and a lag time of 12 hours. The same constraints apply to this technique in determining the unit duration as applied to the dimensionless unit hydrograph relative to the length of the unit duration and the subdivision of the drainage basin.18 hours (use 2 hours for computational purposes). by the conversion factor 645. Time is expressed as a percentage of lag time Lg in the second column. or 2. (2) Dimensionless S-graph technique.FLOOD HYDROLOGY MANUAL After the synthetic unit hydrograph ordinates f& are determined.3 represents the volume of 1 inch of runoff from 1 square mile of drainage and is expressed as l-hour cubic feet per second.


it can be seen that the results of analyses of recorded flood events in the form of hydrograph reconstitution [9] provide the primary basis for development of design flood hydrographs used for sizing hydraulic features on Bureau projects. (h) Observed Flood Hydrograph Reconstitutions.033 1.784 12. hours 0 2 4 6 8 Time in percent of lag Discharge in percent of ultimate 0 1 6 21 35 43 50 55 58 60 Unit hydrograph ordinates.646 4.331 44.364 46. 104 .232 34. The resulting synthetic unit hydrograph ordinates in the fifth column are obtained by subtracting successive values from the preceding one. whether computed manually or by computer. Except for large drainage areas whose flood hydrographs do not peak rapidly. Continuous discharge values should be obtained as outlined in chapter 2.939 28. The Bureau uses the instantaneous end of period ordinates rather than an ordinate representing the average discharge for each unit time period.453 5.293 6. 33 50 67 83 16. ftS/s 0 807 4. should be plotted on graph paper at the proper time intervals and a smooth curve drawn through the points.397 resulting in the summation hydrograph ordinates shown in fourth column. It should be noted that these values represent the ordinates of the synthetic unit hydrograph at the specific time indicated.613 10 12 14 15 16 100 117 125 133 48.099 11.033 2. The basic data required for derivation of a unit hydrograph through the reconstitution of an observed hydrograph recorded for a drainage basin tributary to a particular gauging station are as follows: l Continuous discharge hydrographs for all major flood rises occurring at the gauging station for which adequate corresponding storm precipitation data are available.685 40. mean daily discharges as published in the USGS Water Supply Papers are not satisfactory for reconstructing the flood hydrograph. The final synthetic unit hydrograph ordinates will reflect the position of the smooth curve rather than the computed values.FLOOD HYDROLOGY MANUAL Table 4-lg.-Values for synthetic unit hydrograph.840 Time. ftS/s 0 807 4.420 1.-From the preceding discussions. The synthetic unit hydrograph ordinates. Summation hydrograph ordinates.

Step 3. (3) overall slope of longest watercourse. of drainage basin. (2) length. On a topographic map (a USGS 7. particular storm with a suitable symbol. Also. and (5) distance. The stations should be sufficient in number to define the area1 distribution of rainfall. and type of station (recording or nonrecording). by planimeter and opisometer (map measurer) or computerized digitizer: (1) area. in miles. The template is then rotated about 90”) pin is again placed near the template edge. in whole numbers. of longest watercourse.-The basic “trial and error” process for reconstituting an observed flood hydrograph by the unit hydrograph procedure takes the following steps [9]: Step 1. Identify each station or location with a number that will be used as a cross reference to a table that is to be prepared listing station name. in square miles. and data from sufficient recording gauges should be available to determine the variation of rainfall intensity throughout the storm period. Isohyetal lines representing pre- cipitation values. usually a cross for stations where a continuous record has been taken and a circle for nonrecording stations or locations where total storm data have been accumulated. This is accomplished by using either an isohyetal or Thiessen Polygon approach: a. location of each station (latitude and longitude). in miles. the centroid of the basin can be found by suspending a heavy cardboard template of the drainage basin by using a pin near the template edge and drawing a vertical line from the pin across the template. are drawn using each station’s 105 . identify each location where precipitation data have been collected for a.5-minute quadrangle map for smaller basins and 1:250. Step 2. including hourly rainfall data from available recording rain gauges. Determine. l A topographic map showing the area of the drainage contributing to streamflow at the stream-gauging station. and another vertical line is drawn from the pin across the template. (4) centroid of drainage basin.HYDROGRAPH DETERMINATIONS l Precipitation data for corresponding storms are required for all rainfall stations in and adjacent to the basin. Determine the basin average precipitation for the total storm. in feet per mile. The intersection of these two lines is the centroid of the basin. (i) Hydrograph Reconstitution Procedure. from gauging station site to drainage basin boundary. If a digitizer is not available. outline the drainage basin boundary.000 maps for larger basins). and location of centroid of the basin. longest watercourse from gauging station site to drainage basin boundary. The isohyetal approach uses contour lines (isohyetal lines) representing equal rainfall amounts. from gauging station along longest watercourse to a point opposite centroid of drainage basin.

which is determined either by intimate knowledge of the drainage basin.000 map. This station. Each area has a representative rainfall amount associated with it that is determined by averaging the values of the two isohyetal lines encompassing the area within the drainage basin. The trial unit hydrograph lag time can be estimated for the purpose of a first approximation of observed hydrograph reconstruction by using equation (4) from section 4. step 4. The area within each polygon is determined by planimetering or by using a computerized digitizer. The Thiessen Polygon approach for determining average basin rainfall for the total storm consists of developing polygons formed by connecting the perpendicular bisectors of lines connecting each of the precipitation stations within and outside of the drainage basin. rainfall data that will provide incremental rainfall depths for time increments equal to 15 to 20 percent of the estimated unit hydrograph lag time are used to develop the rainfall hyetograph. manual readings of instruments at recorded time intervals are entirely satisfactory. with the exception of the K. As previously discussed. b. Sufficient polygons must be developed so that all of the drainage basin is covered.FLOOD HYDROLOGY MANUAL precipitation data as a base and a straight-line interpolation between station values. value. The average basin rainfall is then computed by dividing the product of the total rainfall recorded at the station within each polygon and the area encompassed by that polygon by the total drainage basin area.5-minute quadrangle map or from a 1:250. Once the unit hydrograph lag time has been approximated. The determination of these time increments are dependent on the required unit duration of the unit hydrograph. When the lines are drawn. the unit duration of rainfall should approximate l/5. will be used to distribute the total storm rainfall by discrete increments of time. an estimate of the appropriate unit duration is necessary. 106 . A hyetograph is a bar graph indicating the depth of rainfall occurring in a specified time. Since the unit hydrograph has yet to be developed for the case at hand. All of the values in equation (4) are readily obtainable from a USGS 7. It has been assumed that at least one of the precipitation stations within or near the drainage basin is of the recording variety. In addition to mechanically or electronically recording the data. by previous studies of nearby basins. The basin average precipitation is then determined by dividing the product of each area and its representative rainfall amount by the total drainage basin area.5 of the unit hydrograph’s lag time.1(e). the areas between each adjacent line and lying within the drainage basin boundary are determined. or by using an average value from one of the appropriate lag relationships shown on figures 4-6 through 4-l 1. or stations.

Separate the base flow and interflow components from the recorded runoff hydrograph. A trial infiltration rate decay curve is superimposed on the hyetograph such that the volume of rainfall excess equals that portion of the recorded hydrograph assumed to represent surface runoff. Because this water travels underground to the watercourse. the dashed line represents the base flow component. Many previous flood hydrograph analyses for smaller drainage basins have encountered difficulties simply because the rainfall data were on a daylight savings time base and the runoff data were on a standard time base. the easiest and most reliable 107 . ideally of 50-year or less frequency. the elapsed time between rainfall and the resulting appearance as runoff in a watercourse is considerably longer than the surface runoff. as discussed in step 6. The interflow component is one of the three factors that must be adjusted and balanced within reasonable limits to eventually reproduce the recorded flood hydrograph. excluding base and interflow. the infiltration rates toward the end of the storm should approach the minimum or ultimate values for the types of soils in the drainage basin after the soils have become saturated. Construct a trial unit hydrograph. see figure 4-14. Since the storm and resulting flood being analyzed should be a major event.” it is now necessary to assign appropriate infiltration losses to the rainfall occurring over the basin. Referring to figure 4-14. when adjusting either the infiltration losses or the base and/or interflow components. the other two factors are infiltration loss and unit hydrograph. It should be remembered that the volume of surface runoff. or visa versa. Incremental values of trial infiltration losses and the resulting rainfall excess are tabulated as shown on figure 4-14. Since the trial unit hydrograph lag time was determined in step 4. Step 6. Referring to the theory and rationale presented in more detail in the following subsection (j) on “Infiltration and Other Losses. which is primarily a function of ground-water supply of perennial streams and may show an essentially constant rate of discharge for some period prior to the flood event. Step 5. compensating adjustments must also be made to ensure a reasonable balance between rainfall excess and surface runoff. Step 7. and the analyst failed to recognize the fact. The base flow component may be entirely absent in such cases as the dry arroyos in the desert Southwest or in intermittent streams elsewhere. The interflow component. Therefore. shown at the bottom of figure 4-14.HYDROGRAPH DETERMINATIONS and is plotted on the graph showing the recorded runoff hydrograph. will always be present to some degree because it represents rapid subsurface flow of infiltrated water to an adjacent watercourse. It is imperative that both the hyetograph of basin rainfall and the recorded runoff hydrograph are on the same time base. The actual position and form of the interflow component for a given flood reconstruction is assumed. must equal the volume of rainfall excess.

FLOOD HYDROLOGY MANUAL ON033S Wd 1334 319113 NI 39tlWH3SlCl .

(j) Znfiltrution and Other Losses. this apparently tedious process becomes relatively easy to accomplish. If this is the case. This method is repeated for each successive time increment.-This subsection deals with the atmospheric processes that result when either rainfall or snowfall become separated into several parts upon reaching the ground surface. If it is found that the computed peak is lower than the observed peak. Naturally.HYDROGRAPH DETERMINATIONS method of developing the trial (or “first cut”) unit hydrograph is probably by using the S-graph approach previously discussed in section 4.” 3. and then only in relation to a stage versus discharge relationship comparable to an uncontrolled spillway on a reservoir. there are several computer programs available to accomplish this task in a much more rapid and less time consuming manner. Retention in surface depressions that act as miniature reservoirs that do not release their waters until their storage capacity is exceeded. the trial unit hydrograph should be flattened. then the trial unit hydrograph should be peaked up by concentrating more volume around the peak. Move the strip down one time increment and multiply the unit hydrograph ordinate by the adjacent rainfall excess increment. Interception by vegetation and subsequent evaporation dation when reaching the ground surface. Evaporation from ground surface during prolonged rainfall events or when accumulated in frozen form from snowfall. the lag time should be adjusted accordingly and a new trial unit hydrograph developed for the next trial at reproducing the observed flood hydrograph. or retar- 2. With experience. The flood hydrologist is primarily concerned with four of these parts: 1. [6]. Then. align the strip such that the first ordinate of the unit hydrograph is opposite the first increment of rainfall excess. Multiply theunit hydrograph ordinate by the rainfall excess increment to obtain the resulting surface runoff at that point in time. It is rare when the analyst successfully reproduces the observed flood hydrograph on the first trial. Step 9. which is generally termed “sublimation. If the computed peak is higher than observed peak.ww Step 8. The trial unit hydrograph ordinates are then listed on a strip of paper in the reverse of their actual sequential order. 109 . Apply the trial unit hydrograph to the incremental rainfall excess amounts determined in step 5 by manually arranging the rainfall excess amounts in tabular form. Occasionally. and sum the successive products to obtain the total surface runoff hydrograph ordinate at that time increment. the peak of the trial hydrograph will occur either earlier or later than the observed hydrograph.

rock. or combination thereof. and falls to the ground. organic matter. When water is ponded at the surface of a dry soil. asymptotically pacity decreases in an exponential 110 . These rates are controlled by soil profile conditions. However. the soil-water tension gradients are large. Under such conditions. the soil-water tension gradients decrease. part of the precipitation is intercepted by vegetation and. part of the rainfall falls directly on the ground surface and enters the soil. gravity and soil-water tension forces act together to move the water downward through the soil profile. therefore. some of this precipitation evaporates back into the atmosphere. To illustrate the phenomena that occurs in the soil when water is applied in the form of rain. whether it is a concrete airport parking and loading area or the most sandy of soils comprising some areas of the arid West.FLOOD HYDROLOGY MANUAL into the receiving soil. interception and evaporation losses are so small compared to the magnitude of the precipitation that they are neglected. after the vegetation has reached its capacity to retain water by surface tension. the soil wetness. and are called the “soil’s infiltration capacity. As the soil profile becomes more saturated. Also. as the soil profile becomes saturated. however. The process of water moving through the ground surface into a soil profile is referred to as “infiltration”. etc. and animal burrows can all have a marked effect on infiltration rates. Initially. shrinkage cracks. consider a condition at the onset of a rainstorm where the soil is comparatively dry as a result of no precipitation having recently occurred. compaction. In nature. and the soil is said to absorb water in the same manner as a blotter absorbs ink. in the hydrologic analysis of floods. The infiltration rate is the “volume per unit area per unit time passing through the ground surface and flowing into the profile.” and assumed to have been satisfied by antecedent rainfall occurring prior to the onset of the PMS. the first three are often grouped with part of the infiltration loss and termed “initial losses.” Tension gradients are greater than gravity gradients. 4. stems.. additional precipitation simply runs off the leaves. Znjiltration The first three of the above loss processes are usually low when compared with infiltration when rainfall intensities are sufficient to produce severe flood events such as the PMF. hydraulic conductivity increases. Any of the constituents of the Earth’s mantle have a capability to absorb water. making the infiltration rates high. conditions deeper within the soil profile become more important and eventually may control the process. and the infiltration cadecay fashion. swelling of clay soils. Although the hydraulic conductivity is small. root holes.” Infiltration rates are sensitive to local conditions at or just below the ground surface and.

Bureau hydrologic engineers consider rates resulting from reconstitution studies to be more valid because they tend to reflect the integrated infiltration rates for the various soil conditions over the entire drainage basin. At some point during a precipitation event. the infiltration capacity also changes. precipitation rates may exceed the infiltration capacity of the soil. the hydrologic engineer concerned with the magnitude offc in equation (5). base of natural logarithm. Continued rainfall will then produce the characteristic decay or decrease in infiltration rates. When the precipitation rate is less than the soil profile’s infiltration capacity. in hours. initial rate of infiltration capacity. When infiltration rates developed from the infiltrometer tests are compared with those from observed flood hydrograph reconstitutions. The four groups as generally defined by the SCS are: 111 . infiltration rates are identical to the precipitation rates and no water is available for surface runoff. time from start of rainfall. in hours. as it relates to severe flood occurrences. which results in ponding and/or surface runoff. The SCS (Soil Conservation Service) has proposed the subdivision of soils into four groups relative to their respective infiltration capacities or ultimate infiltration rates. (5) and In the development of the PMF. the test results from the infiltrometer are almost always higher. Horton [58] proposed the following mathematical relationship to represent this function: f=fc+ukfwkt where: f = fc = f0 = e = k = t = resulting infiltration rate at time t. Since the infiltrating water alters the water content in the soil profile. in inches per hour. in inches per hour. The ultimate or minimum infiltration rates of these four groups have been found by the Bureau to be in reasonably close agreement with the rates resulting from observed flood hydrograph reconstitution studies. In 1940. When more than one group of soils is present in a drainage basin. a constant dependent primarily on soil type and vegetation. can be represented by a decay curve function where the infiltration capacity rather than rainfall rates control the infiltration rates. is primarily Many attempts have been made to measure infiltration rates using devices known as infiltrometers. The ultimate or final infiltration rate is theoretically equal to the saturated hydraulic conductivity because tension gradients are no longer present and the hydraulic gradient is due solely to gravity.HYDROGRAPH DETERMINATIONS approaching the ultimate infiltration rate. This phenomenon. minimum or ultimate infiltration rate. an average value for the basin should be calculated based on weighted areas.

Ultimate infiltration rates for these soils have been found to range from 0.5 inch per hour. which would include sandy loams and shallow loess. and consisting mostly of moderately deep to deep. and consisting mostly of wellto excessively well-drained sands or gravels. (k) Base Flow and Znterjlow-The base flow and interflow components to a flood hydrograph are graphically shown on figure 4-14. These soils may also include moderate organic matter. Ultimate infiltration rates for these soils range from 0. (3) Group C soils. and consisting mostly of clay soils with high swelling potential. The base flow is generally depicted as a recession curve.15 to 0. and that infiltration rates would be at the minimum or ultimate rate at the onset of the PMS. These soils have a slow rate of water transmission. desert pavement) or clay layer at or near the surface.to well-drained soils with moderately fine to moderately coarse textures. Hydrologic analyses leading to PMF estimates should be based on the assumption that minimum or ultimate infiltration rates prevail throughout the duration of the PMS. -Soils that have high infiltration rates even when thoroughly wetted. moderately well. soils with a claypan (e. and consisting mostly of soils with a layer that impedes downward movement of water..g.3 to 0.05 to 0. (4) Group D soils (high runoff potential). (2) Groups B soils. This assumption is based on consideration of conditions that have been shown to exist prior to extreme storm events. and shallow soils over nearly impervious material. Examination of historical conditions have shown that it is entirely reasonable to expect one or more storms antecedent (prior) to the extreme event due to meteorologic persistence.30 inch per hour. and include heavy plastic clays and certain saline soils. This flow continues to decrease until the water surface in the stream is in a state of equilibrium 112 . and include many clay loams. shallow sandy loams. soils low in organic matter. The minimum or ultimate infiltration rates for these soils range from 0. The base flow component generally consists of the water reaching a basin’s watercourses after flowing a considerable distance underground as ground water.05 inch per hour. and soils usually high in clay content.-Soils having slow infiltration rates when thoroughly wetted.15 inch per hour.FLOOD HYDROLOGY MANUAL (1) Group A soils (low runoff potential). or soils with moderately fine to fine texture. These soils have a high rate of water transmission. it is reasonable to assume that any antecedent storm has satisfied any soil moisture de& ciencies and initial losses. which indicates a gradually decreasing rate of flow. These soils have a very slow rate of transmission. Minimum infiltration rates range from 0 to 0.-Soils having very slow infiltration rates when thoroughly wetted. soils with a permanent high-water table. Accordingly.-Soils having moderate infiltration rates when thoroughly wetted.

It is entirely proper to assume that the base or recession discharge rate is uniform for the duration of the PMF hydrograph. If this is the case. The magnitude of the base flow is highly dependent on antecedent storm conditions both in terms of that storm’s magnitude and the time elapsed between its occurrence and the onset of the storm that produced the flood being analyzed. The separation of the observed flood hydrograph into the three components shown on figure 4-14 requires a considerable amount of judgment because the interflow and base flow (recession flow) are considerably more indeterminate than the surface flow component. After subtracting the base or recession flow component. Providing that sufficient data are available. a higher recession flow is used when there is a l-day separation between the antecedent storm and the PMS than would be used if a 5-day separation was assumed. This may be accomplished by converting the observed component to cubic feet per second per square mile of drainage basin area. is generated by precipitation that enters the ground by infiltration but emerges as a direct contribution to the surface runoff within a relatively short period. The interflow component is determined essentially by a trial and error approach in the course of observed flood hydrograph reconstitutions previously discussed. When separating these latter two components. results of observed flood reconstitutions on hydrologically similar drainage basins relative to the base flow component are used to estimate this component for the ungauged basin. however. The result is then applied to the area of ungauged drainage basin under study to determine an appropriate rate of base flow for that basin. When the water table is at a level below the channel bed. When preparing a flood study for an ungauged watershed. depending on the particular drainage basin. The interflow component. which is rarely the case. the remaining observed flood hydrograph is composed of the surface flow and interflow components. care must be taken to ensure that neither too much nor too little flow is assigned to the interflow component.HYDROGRAPH DETERMINATIONS with the surface of the adjacent water table and the flow is maintained by inflow from the ground-water reservoir. subsurface flow may be taking place in the river gravels. For example. the recession curve will approach and finally go to zero. a complete recession curve representing the base flow component for a given drainage basin can be determined. Bureau hydrologic engineers currently believe this phenomenon is present in every severe flood event in varying degrees. occasionally referred to as the subsurface storm flow. 113 . Quantification of the base flow and interflow components in a flood study is usually based on the results of flood hydrograph reconstitutions. there will be no surface flow in the stream. A reasonable balance can be achieved by the adequate selection of infiltration loss rates previously discussed. The recession or base flow used in the development of the PMF hydrograph should represent conditions that are consistent with the antecedent storm conditions provided for in the storm study report.

since the unit hydrograph approach is used to develop the design flood hydrograph. The PMF will be based on the PMP or storm values developed using the criteria furnished in the appropriate HMR discussed in chapter 3. 3.2 Probable Maximum Flood Hydrographs The Bureau’s definition of the PMF (probable maximum flood) hydrographic is “the maximum runoff condition resulting from the most severe combination of hydrologic and meteorologic conditions that are considered reasonably possible for the drainage basin under study. interflow information from observed flood hydrograph reconstructions for hydrologically similar nearby watersheds are used to estimate the magnitude and rate of change of discharge over time. a rather broad peak. When a flood study involves an ungauged basin. the storm rainfall values will be arranged in such a manner that the maximum peak discharge and the maximum concentration of discharge around the peak is achieved. The temporal distribution of the storm rainfall will be consistent with the hydrometeorological report criteria. as discussed in chapter 3. Infiltration rates that are subtracted from the storm rainfall to obtain the excess amount available for surface runoff should be at minimum or ultimate rates consistent with the soil types and underlying geologic conditions of the drainage basin under study. and should be more conservative than values reflected by averages. These minimum or ultimate rates will be assumed to prevail throughout the duration of the PMS event. These infiltration rates should also be consistent with the PMF concept. and a long recession limb. care should be taken to ensure the unit hydrograph parameters adequately reflect conditions likely to occur in a probable maximum event. considerable judgment must be used to ensure that the 114 . 2.” Accordingly. the conversion from the observed hydrograph to that for the ungauged basin is based on a direct ratio of the respective drainage basin areas. value in the general either an upward or downward direction to reflect the decreased or increased hydraulic efficiency of the drainage basin’s network that would be expected in association with the high discharges that occur during such an extreme event. the following considerations should be followed when computing the hydrograph: 1.FLOOD HYDROLOGY MANUAL When an ungauged watershed is being studied. In the absence of such criteria. The unit hydrograph used to compute the PMF should be representative of extreme discharge conditions. The resulting interflow hydrograph should incorporate a rising limb. In situations where studies are being prepared for gauged basins for which results of observed flood hydrograph analyses are available. It is entirely appropriate to unit hydrograph lag equation in modify the K. As was the case for the base flow component. 4.

acceptable criteria for determining reasonable values for the several parameters in the snow compaction method. the drainage basin is divided into elevation bands for analysis. Proper application of this method requires either data or assumed values for air temperatures.3. depth of snow. this method can become quite capricious. and density of snow at various elevation bands. Depending on the size of the basin and the elevation variation within the basin. The rain-on-snow contribution from each elevation band is computed and averaged over the total basin. and air temperatures required to assure the largest flood has been computed.3 Western Mountain Snowmelt Equation A method that has been used by the Bureau has become known as the “Snow Compaction Method. The wind speeds and air temperatures are usually furnished by Bureau meteorologists as part of the PMS study when it is anticipated that the snowmelt runoff will contribute to the PMP to generate the PMF. the basin’s expected hydraulic efficiency as 4. Without experience and good judgment. 4. The interflow component should reflect the optimum condition that could be expected to result from a PMS event. For those basins that are relatively flat. Currently.HYDROGRAPH DETERMINATIONS K.or l.OOO-foot intervals. and then added together to yield the total area runoff. This method requires several assumptions by the hydrologic engineer because there are usually several trial arrangements of rainfall. there are no consistent. percent of forest cover. However. 5. wind speeds. The base or recession flow hydrograph component should reflect the maximum rates of discharge that are consistent with the magnitude and timing of any antecedent flood event. this component will probably not be significantly different than would be experienced in a relatively minor event because the hydraulic efficiency of the subsurface media through which this component passes is essentially fixed. a rain-on-snow PMF hydrograph is desired. This problem has been a topic 115 . wind speeds. In most cases. the elevation bands are usually selected at 500. 35 [59]. From a record search.” This method is described in detail in the Bureau’s Engineering Monograph No. the hydrologist determines snow depths and densities that are considered reasonable for initial watershed conditions at the onset of the PMS. Frequently. The basis and rationale for adding a snowmelt runoff component to the rainflood hydrograph are discussed in section 4. The initial snow depth and density may also have to be adjusted to ‘assure that a reasonable amount of snow has melted and only a moderate amount of rain has been trapped in the remaining snow of the upper elevation bands. the entire basin may be used as one elevation band. values approximate closely as possible.

will be the development of suitable In view of this lack of criteria. The occurrence of antecedent precipitation.05 inch per hour to the PMP when generating the PMF rainflood hydrograph. this period may extend to 30 or 60 days or.FLOOD HYDROLOGY MANUAL of discussion and future criteria. is the basis for assuming wet or saturated ground conditions and adopting minimum or ultimate infiltration losses in developing PMF hydrographs. For drainage basins less than about 3. The normal period of runoff selected is 15 days. Naturally. The resulting flood hydrograph. both meteorologic and hydrologic. however. is generally referred to as the PMF series. the melting snowpack may form the antecedent flood discussed in section 4. the period may extend through two yearly runoff seasons. It should be noted that since the melting snowpack tends to satisfy the infiltration losses. among action representatives of this group of the Bureau NWS. The resulting snowmelt hydrograph is generally expressed in terms of mean daily flows for the 15-day period with diurnal fluctuations being neglected. the standard practice of the Bureau is to combine the rain generated part of the PMF hydrograph with a snowmelt that could reasonably be expected to occur at the time of year that the probable maximum rainfall occurs. this practice is only used in those areas where snowpacks of sufficient magnitude to be considered do occur. it is normal practice to consider antecedent conditions. This assumption is made so that the maximum rain occurs during the warmest period. Current practice is to apply a loss of 0. The resulting combined rain-on-snow flood represents the probable maximum event.4. in the case of the Colorado River above Hoover Dam. COE. 4. Such losses apply only to the area assumed to be covered by the snowpack. The concept of an antecedent event is based primarily on the meteorological factors discussed in chapter 3. For larger basins.000 square miles. the losses to the rainfall increment are minimal. and SCS. in large drainage basins with significant areas where snowpack accumulates.4 Antecedent Floods When developing a PMF hydrograph. The loo-year snowmelt flood is then distributed over time using either the largest snowmelt flood of record as the basis for distribution or by using the balanced flood hydrograph approach as discussed in chapter 7. A frequency analysis of the maximum annual snowmelt volume is performed using the procedures discussed in chapter 7. 116 . The Bureau currently uses a loo-year snowmelt flood to account for snowpack. either in the form of rain or snow. the rainflood hydrograph is superimposed on the snowmelt hydrograph with the rain assumed to occur during the day or days of the greatest snowmelt. in combination with the PMF hydrograph.

The two exceptions are antecedent precipitation criteria adopted by the Bureau and COE for the State of Texas. most probably in an interagency effort. and by the Tennessee Valley Authority for the Tennessee River Basin.HYDROGRAPH DETERMINATIONS Currently and with two exceptions. (b) For PMF’s generated by general PMP events west of the Sierra Nevada and Cascade Ranges.000 Square Miles in Area” [60]. no antecedent event is used. the following criteria are used by the Bureau: (a) For PMF’s generated by general PMP events in areas east of the Sierra Nevada and Cascade Ranges. If the balanced flood hydrograph approach is used. excluding areas covered by the previously mentioned NWS reports. However. 117 . National Weather Service. the time between the end of antecedent rainfall and the beginning of the PMP event is assumed to be 3 days. it is reasonable to assume that a storm of sufficient magnitude has occurred to satisfy initial infiltration losses and provide for minimum or ultimate infiltration loss conditions at the onset of the probable maximum local storm event. Therefore. until such time as definitive studies are conducted. “Probable Maximum and TVA Precipitation Estimates With Area1 Distribution for Tennessee River Drainages Less than 3. the antecedent flood is estimated by either converting loo-year precipitation to a flood hydrograph or developing a balanced loo-year hydrograph using statistical analyses of runoff data. certain provisional criteria have been adopted as established by hydrometeorlogists and hydrologic engineers who have given specific preliminary consideration to the governing meteorlogical factors. Meteorological conditions are such in this region that hydrometeorologists do not consider it reasonable to assume an event of loo-year magnitude to precede the probable maximum event. a time interval of 3 days between the peak of the antecedent flood hydrograph and the beginning of the PMP event is used. and NWS Technical Memorandum “Precipitation Antecedent to the 24-Hour Probable Maximum Precipitation for Small Basins in Texas”[Gl]. Documentation of the studies and the resulting criteria are contained in HMR 56. Meteorological studies leading to the development of these criteria were performed by the Hydrometeorological and Special Studies Branch. none of the Federal water resource development agencies have criteria that reflect definitive hydrologic and hydrometeorologic antecedent storm and flood studies. If loo-year precipitation method is used. Office of Hydrology. the same criteria given in (a) apply except that the time intervals are 2 rather than 3 days. (c) For PMF’s generated by local PMP events in the entire region west of the 103rd meridan. Since no similar criteria other than the Texas criteria are currently available for the 17 Western States where the Bureau has primary interest.



(d) For PMF’s generated by PMP falling on a snowpack, the antecedent meteorologic condition is assumed to be the snowfall that accumulates and forms the snowpack. It is further assumed that runoff resulting from the melting of the snowpack will occur during and somewhat prior to the probable maximum rainfall. For the majority of drainage basins, those with areas smaller than about 3,000 square miles, current practice is to develop a loo-year balanced snowmelt hydrograph and then superimpose the probable maximum rainflood onto it in such a manner that the peak flow of each hydrograph occurs at the same point in time. For basins larger than 3,000 square miles, it has been frequently found that the snowmelt will peak earlier than it is meteorologically reasonable for the probable maximum rainstorm to occur. In such situations, an evaluation of the timing of the two events should be requested of and performed by experienced hydrometeorologists. Application of the western snowmelt equation discussed in section 4.3(a) is not currently used. However, when definitive criteria for application of this equation have been developed and generally accepted among the Federal water resource development agencies, it is expected that the equation will replace the loo-year balanced hydrograph approach.
4.5 Foss Dam Example

As an example of probable maximum flood hydrograph determinations, consider the Washita River Basin above Foss Dam in Oklahoma. As shown on figure 4-15, this dam and its 1,490-square mile tributary drainage basin are located in west-central Oklahoma, with the basin centroid located at latitude 35 “40’N. and longitude 99”45’ W. The upper reaches of this drainage basin extend about 30 miles into Texas. For hydrologic reasons, the basin was divided into four subbasins as shown on figure 4-15. This subdivision will result in the generation of four component flood hydrographs that will require routing and combining to obtain the total flood hydrograph at the dam. This routing and combing will continue this example in chapter 5. Probable maximum storm precipitation values were obtained for this basin using procedures described comprehensively in HMR 51 and 52 [35,18]. The process for obtaining these values is identical with Example No. la shown on pages 108 through 126 of HMR 52. For brevity, the actual PMP computations for the drainage basin above Foss Dam are not included in this manual; however, the results of the computations for the four subbasins are shown in table 4-20. The unit The were first step in the rainfall-to-runoff conversion process is to derive hydrographs for each of the four subbasins shown on figure 4-15. salient physical properties shown in table 4-20 for each subbasin determined using a USGS quadrangle map.






Table 4-PO.-Properties Drainage Area,

of subbasins in Washita River Basin.



L <I, P
miles 8.2 11.2 19.2 14.0

ft/mi 26.5 23.1 17.6 20.6

- so 9 41.4 62.7 189.0 99.0

1 2 3 4

429 333 389 315

26.0 26.9 41.3 32.1

As a result of a field reconnaissance of the Washita River drainage basin tributary to Foss Dam, a factor was assigned for use in the general unit hydrograph lag equation, a suitable dimensionless unit hydrograph selected, and an appropriate infiltration loss rate was determined. Because of the uniformity of the basin, a single factor of 1.8 was assigned to each of the four subbasins. This corresponds to the upper curve shown on figure 4-6. The value of 1.8 reflects a K,, value of 0.069, which is considered characteristic of the drainage network’s hydraulic characteristics at extreme flood levels. The unit hydrograph lag time for each subbasin is shown in table 4-2 1. The new Arbuckle dimensionless unit hydrograph was selected for distributing the unit runoff over time because this hydrograph was derived from a major recorded flood on a nearby similar drainage basin, and was considered to be more suitable than the Great Plains dimensionless graph for this case. It should be noted that it is entirely proper to substitute a dimensionless graph based on a flood hydrograph reconstitution for the regionalized graphs previously discussed. Using the shortest unit hydrograph 1, the unit duration is determined 6.15/5.5 = 1.12 hours. Thus, a unit though a longer unit duration might best to retain a single unit duration avoid confusion during the routing chapter 5. lag time of 6.15 hours for subbasin by dividing the lag time by 5.5, or duration of 1 hour is adopted. Even be used for subbasin 3, it is always for all subbasin computations to and combining process discussed in

All the required information is now known to derive a unit hydrograph for each subbasin. For brevity, only the unit hydrograph for subbasin 1 will be derived in detail. Table 4-22 shows the details of the unit hydrograph development for subbasin 1. Table 4-22 shows the time, in hours, in the first column, which is the interval corresponding to the unit duration, in this case, 1 hour. The second column records what percent the value in the first column is of the lag plus semiduration of unit rainfall, which is 6.15 + 0.5 = 6.65 hours for this example. The third column records the value read or




Table 4-21.-Unit


lag time for each subbasin in Washita River Basin.

1 2 3 4

G Coefficient
1.8 1.8 1.8 1.8 0.069 .069 .069 .069




3.42 3.92 5.64 4.56

6.15 7.05 10.15 8.20

interpolated from the Arbuckle dimensionless graph for values corresponding to lag plus semiduration values for each time increment. The fourth column records the unit hydrograph ordinate for each time increment, and is found by multiplying the total volume of 1 inch of runoff in l-day cubic feet per second by the ordinate shown in third column divided by the value of the lag plus semiduration. The volume of 1 inch of runoff for the 429-square mile drainage basin is found by multiplying the drainage area by 26.89, which results in 11,536 l-day cubic feet per second. If the unit hydrograph is computed manually, a check computation should be made. This can be done by totaling the unit hydrograph ordinates in table 4-22 and comparing this total with the volume of 1 inch of runoff from the basin. For this example, the total of the ordinates is 276,846 l-hour cubic feet per second. The volume of runoff from 1 inch of rainfall excess falling in 1 hour is found by multiplying the drainage area, 429 square miles, by a factor of 645.33 which yields 276,846; therefore, the computations check. After a similar procedure is followed to develop the unit hydrographs for each of the other three subbasins, the next step is to determine the effective rainfall, or the rainfall that is available for surface runoff. This is accomplished by subtracting the infiltration losses from the probable maximum rainfall. As a result of the field reconnaissance of the Washita River Basin above Foss Dam, it was determined that an ultimate infiltration rate of 0.30 inches per hour would be appropriate. At this point, tables are set up to facilitate both the computation and documentation of the flood hydrograph development, as shown by tables 4-23 through 4-26. In these tables, the first column lists the time increment, and the incremental probable maximum precipitation values in the second column. The third column lists the infiltration losses; e.g., 0.30 inch per hour. Note that the loss cannot exceed the input rainfall; i.e., when rainfall is less than 0.30 inch per hour, the loss equals the rainfall. The fourth column provides the rainfall excess amounts that are available for surface runoff; these are the values that are applied to the unit hydrograph in the manner previously described to obtain the surface runoff hydrograph. The unit hydrograph and resulting surface runoff




Table 4-22.-Unit



for subbasin 1. Unit hydrograph ordinate Q ,, l-hour ft”/s 1,615 6,896 18,426 35,298 44,387 28,786 18,860 15,900 13,859 12,078 10,525 9,182 7,998 6,988 6,081 5,298 4,618 4,027 3,521 3,070 2,670 2,321 2,027 1,768 1,542 1,349 1,177 1,019 897 778 672 589 520 467 415 362 309 275 242 214 276,846

Time, hours 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40



L, + 0.50
15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 230 255 270 285 300 315 330 345 360 375 390 405 420 435 450 465 480 495 510 525 540 555 570 585 600

Ordinate q from Arbuckle dimensionless graph 0.93 3.98 10.52 20.35 25.59 16.60 10.87 9.17 7.99 6.96 6.07 5.29 4.61 4.03 3.51 3.05 2.66 2.32 2.03 1.77 1.54 1.34 1.17 1.02 0.89 .78 .68 .59 .52 .45 .39 .34 .30 .27 .24 .21 .18 .16 .14 .12



670 2.106 .061 .298 4.274 1.768 1.077 . hydrograph development for subbasin 1.300 .121 .084 .083 .860 15.069 .lOl .073 .200 .064 .079 .027 3.300 .246 35.109 .234 .139 .lOl .200 .760 1.117 .060 .182 .121 .088 .064 .859 12.129 .300 Excess.234 .439 .106 .070 2.339 .096 .300 . inches 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.058 .090 .245 .070 . fP/s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 63 358 1.896 18.027 1.207 .090 .088 .129 .039 .601 .245 .065 .083 .300 .349 1.145 .094 .155 .065 .900 13. inches 0.660 Unitgraph.460 1.300 .019 897 778 672 589 520 467 415 .079 .300 .182 7.161 .067 .072 . inches 0.618 4.362 309 275 242 214 - Hydrograph.133 .068 .525 9.210 .HYDROGRAPH DETERMINATIONS Table 4-23.974 1.099 .182 .068 .061 .141 .901 1.507 .298 44.070 .094 .-Flood Time.521 3.175 .067 .078 10.786 18.960 Loss.055 .988 6.387 28.072 .210 . hours 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Rain.073 .615 6.133 .069 .177 1.315 3.117 .084 .300 .175 .580 123 .840 66.161 .278 .077 .321 2.607 17.081 5. ftS/s 1.096 .998 6.058 .965 35.060 .299 .673 8.355 .109 .155 .099 .278 .145 .542 1.299 .141 .

222 .137 .030 138.169 .830 228.113 .589 7.410 19.071 .690 101.086 .023 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 for subbasin 1 .Continued.600 273.592 .800 33.300 .150 . Unitgraph.260 .100 1.501 6.-Flood Time.150 .425 0.525 17.904 3.222 .100 170. inches 0.800 1.900 222.FLOOD HYDROLOGY MANUAL Table 4-23. inches 3.300 . fP/s 113.066 .300 . fP/s Hydrograph.840 29.071 .520 25.415 2.510 38.103 .425 3.730 22.190 .093 .660 170.081 .310 9.546 5.058 - hydrograph Loss.063 .075 .069 .098 .954 2.063 .473 987 343 90 24 5 0 - - - 124 .190 .030 14.300 .113 .190 58.025 4.420 77.719 5.400 270.864 8.103 .650 88.292 .069 .260 .058 - development Excess.093 .464 1.098 .059 .300 .080 67.323 .440 117.388 .081 .060 44.169 .125 . inches 2.960 11.059 .983 1. hours 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Rain.088 .590 51.075 .125 .066 .086 .857 12.137 .125 0.

135 . Hydrograph.880 2.193 .065 .115 .712 8.211 .061 1.732 Loss.100 973 864 771 675 605 535 472 421 369 338 306 275 243 211 192 170 152 24 - 0.223 .580 1.064 .099 .420 1.432 Unitgraph.298 . ftS/s 897 3.407 26.118 .095 .079 .272 2.089 .065 .900 86.300 .068 .127 .109 .573 1.632 6.106 .761 1.135 .109 .760 5.075 .166 .154 .070 .300 .149 . inches hydrograph development for subbasin 2.261 .106 .072 .261 .093 .070 .740 Excess. inches 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.370 .966 9.139 .083 . fP/s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21 143 537 1.300 .249 .784 17.273 .087 .166 3.986 5.083 .300 .243 1.552 3.066 .149 .068 .072 .223 .102 .249 .193 .099 .298 .093 .091 .969 12.186 .300 .127 .192 .300 .172 .300 .710 4.150 4.323 .166 .720 7.537 2.115 .123 30.186 .780 49.717 10.084 .080 .172 .076 .084 .571 2.139 .211 .612 7.274 2.308 4.237 1.124 .023 .070 .080 .236 18.102 .120 1.068 .787 1.391 .066 .118 .HYDROGRAPH DETERMINATIONS Table 4-24. inches 0. hours 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Rain.900 2.809 27.492 .076 .691 3.798 14.154 .990 136.300 125 .089 .075 .404 1.450 4.180 2.095 .300 .079 .300 .022 1.068 .087 .-Flood Time.474 9.064 .300 .124 .

500 37.500 197.144 .970 42.090 54.201 .260 113. Unitgraph.914 1.020 99.300 .-Flood Time.098 2.063 - hydrograph Loss.954 8.349 3.091 .673 .065 .131 .065 .800 68.351 .24.224 9.220 130. ftZ/s Hydrograph.082 .131 .067 .121 .FLOOD HYDROLOGY MANUAL Table 4.431 .190 23.104 .540 243.086 .179 .051 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 for subbasin 2 .220 20.078 .069 .570 26.495 3.164 5. ft”/s 193.082 . inches 0.201 .660 33.078 .278 .112 .500 155.069 .560 18.031 415 123 37 11 1 0 - - - - - - - - - - - 126 .091 .150 47.073 .063 - development Excess.300 .903 3.315 1.236 . hours 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 Rain.373 .160 .665 11. inches 0.144 .104 .472 4.879 4.121 .930 61. inches 0.Continued.469 1.278 .097 .719 2.820 77.131 .067 .370 29.299 12.137 14.340 87.097 .206 16.200 250.946 6.837 7.300 276.300 .236 .112 .179 .160 .819 6.086 .073 .


211 .560 51.068 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - for subbasin J-Continued.151 .248 .067 hydrograph Loss.125 .210 110.075 .102 .300 .070 . 367 340 313 291 273 254 236 217 198 180 164 153 141 132 120 Hydrograph.085 . inches 0.680 23. fP.443 7.419 .FLOOD HYDROLOGY MANUAL Table 4-25.280 78. inches 0.114 .211 .300 .900 169.500 187.290 .200 291.530 55.020 46.190 33.878 10.166 .440 36.114 .138 .366 14.290 .085 .095 .067 development Excess.070 .234 16.990 25.248 .125 .112 6.198 8.940 60.970 134.530 101.089 . Unitgraph.952 11.081 .457 .157 .900 85.187 .513 5.982 5.260 92.187 .820 42.860 18.068 .210 30.630 121.100 265.081 .068 .095 .500 278.300 232.089 . inches 0.070 .368 .075 .840 65.is 118.250 71.109 .-Flood Time.102 .006 9.560 154.980 39. ftT/s.070 .166 .300 .494 5.640 19.719 .151 .042 4.733 7.138 .490 27.590 21.894 10. hours 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 Rain.109 .657 - - - - - - - - - - - - - 128 .830 222.102 12.748 15.

408 2. inches - for suhbasin 3-Continued. 129 .657 2. Unitgraph. and interflow. A seventh column would list the combined base flow and interflow. two more columns would be added to the tables.178 1.912 2. and an eighth column would list the sum of surface runoff.620 1. and this example will then be continued.-Flood ‘1-i tne.178 2. fiZ’/s liydrograph. This subject will be addressed in chapter 5.720 3.918 1. respectively. base flow. These flood hydrographs will now require routing and combining to obtain the inflow flood hydrograph. ftY/s 4. both the base flow and interflow were neglected because they would be insignificant as compared to the surface runoff.HYDROGRAPH DETERMINATIONS Table 4-25. If these two components had been significant and taken into account. In this example. hours 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 Rain. inches - hydrograph I .311 4.265 892 281 81 28 8 0 hydrograph are listed in the fifth and sixth columns.003 3.oss. This example has reached the point where the concurrent probable maximum flood hydrographs have been developed for each of the four subbasins above Foss Dam.447 3. inches - development Excess.

inches 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.410 1.184 .145 .642 17.028 .072 .144 1.302 8.163 .131 .163 .073 .122 .-Flood Time.063 .086 . fP/s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 45 193 608 1.056 .793 6.lll .190 0.264 1.300 hydrograph development for subbasin 4.070 . inches 0.563 1.062 .290 1.260 .059 .103 .014 .314 .082 .229 .960 20.063 .157 1.300 .067 . Hydrograph.069 .062 .151 .059 .072 .086 .300 .060 .300 .060 .490 1.741 1.379 7.080 .821 1.548 6.067 .075 .053 .122 .300 .099 .643 2.099 .440 2.053 .120 5.509 4.069 .088 .744 4.053 5.080 .563 21.135 .857 1.056 .379 2.260 .184 .082 .073 .118 .217 .403 .300 .FLOOD HYDROLOGY MANUAL Table 4-26.217 .168 . hours 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Rain. inches 0.168 .293 15.145 .055 .279 .992 Unitgraph.108 .063 .438 9.094 .075 .194 .300 .092 .229 .590 Excess.I35 .956 4.521 .590 1.lll .930 1.070 .933 2.108 .165 .279 .462 4.328 .626 3.021 3.141 1.300 .028 927 833 754 678 611 544 496 445 400 358 326 293 271 249 226 204 182 162 130 .102 .092 .096 10.465 .088 .740 2.102 .118 .265 2.332 10.830 61.055 .292 Loss.194 .077 . ft”/s 593 2.094 .063 .151 .853 9.783 27.131 .300 .200 36.300 .077 .

053 - Excess.074 .490 139. Rain.543 -357 .069 9.068 .333 13.870 2.910 23.906 17.157 .164 7.378 4.064 .610 81.114 .-Flood Time.057 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - ftS/s 147 135 120 21 - fP/s 90.300 .353 2.062 8.436 1.356 6.244 .126 . hours 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 hydrograph development for subbasin 4-Continued.590 43.370 147.300 . inches 0.074 .820 119.170 102.157 .775 1.969 5.430 60. Unitgraph. inches 0.300 .430 11.610 129.000 18.274 1.330 21.114 .140 .470 31.951 1.097 .105 .260 54.184 10.358 3.090 .140 .105 .079 .842 4.183 2.054 . Hydrograph.629 5.058 .068 .804 12.206 .061 .061 .090 .604 1.024 15.084 .750 39.084 .244 .053 - Loss.390 35.000 48.097 905 691 131 .175 .097 .054 .770 25.206 .175 .079 .532 3.071 .480 68.071 .139 1.HYDROGRAPH DETERMINATIONS Table 4-26.592 2.940 28.243 .064 . inches 0.300 .917 3.126 .058 .

inches - hydrograph Loss.FLOOD HYDROLOGY MANUAL Table 4-26. fP/s Hydrograph. inches development Excess.-Flood Time. inches - for subbasin 4-Continued. Unitgraph. fP/s 472 207 58 12 1 0 - - - - 132 . hours 99 100 101 102 103 104 Rain.

This combined flood hydrograph is then chanto the reservoir at point B using one of the channel routing described in section 5. This inflow hydrograph is then routed through the reservoir using the reservoir flood routing procedure described in section 5. which has been subdivided into six subbasins for hydrologic analysis purposes. This chapter deals with the procedures used by the Bureau of Reclamation in routing these flood hydrographs through both reservoirs and channel reaches. Step 3. The channel-routed hydrograph ordinates from point A are then added to the flood hydrograph ordinates from subbasin 3. assume that concurrent flood hydrographs have been determined for each of the six subbasins. This combined flood hydrograph represents the total inflow hydrograph to the reservoir at point B. The total basin flood hydrograph at point E is required. The routing and combining of these subbasin hydrographs to obtain the total hydrograph at point E would proceed as follows. The at point A hydrograph nel routed techniques ordinates of the flood hydrographs for subbasins 1 and 2 are added directly together to find the combined flood at point A. For purposes of this discussion. Step 1. Consider the hypothetical drainage basin shown on figure 5-1.2.2(a). An existing dam and reservoir have been assumed to be located at point B in subbasin 3. Any such regulating effects must be accounted for in determining the flood hydrograph at the point of interest. The outflow routed to point C. In these situations. Step 2.FLOOD 5. Also.1 General ROUTING Chapter 5 THROUGH RESERVOIRS RIVER CHANNELS AND Considerations It is common for the hydrologic engineer to encounter situations where a drainage basin is either so large or has such significant topographic variations that it must be subdivided into smaller subbasins for hydrologic analysis. These facilities may have a regulating effect on the flood runoff prior to its reaching the point of interest. hydrograph from the reservoir is then channel 133 . the component flood hydrographs for each subbasin must be routed and combined to arrive at the final flood hydrograph for the total drainage basin. there are situations where there is one or more dams and their associated reservoirs located in the study basin.

..FLOOD HYDROLOGY MANUAL *-w-c -. 103-D-1919.-Hypothetical drainage basin for flood hydrograph considerations.BA 3 v Total Drainage Basin Boundary Boundary Subbasin Watercourses Control hint Subbasin and Letter Number Dam and Reservoir routing and combining Figure 5-l . 134 .

Also. column (12) plus (13) to point E E E. This combined flood hydrograph represents the total flow at point C. To ensure that all routing and combining factors are properly accounted for. Step 5.-Tabular procedure for routing obtain total basin hydrograph. The resulting combined flood hydrograph represents the total flood hydrograph for the entire basin. column B from C C. column (15) plus (16) 135 . and combining subbasin flood hydrographs to Column No. column B B. in hours Subbasin 1 hydrograph ordinates at point Subbasin 2 hydrograph ordinates at point Combined hydrograph ordinates at point Column (4) ordinates routed to point B Subbasin 3 hydrograph ordinates at point Combined hydrograph ordinates at point Inflow to reservoir at point B Outflow ordinates from reservoir at point Column (8) ordinates routed to point C Subbasin 4 hydrograph ordinates at point Combined hydrograph ordinates at point Column (11) ordinates routed to point D Subbasin 5 hydrograph ordinates at point Combined hydrograph ordinates at point Column (14) hydrograph ordinates routed Subbasin 6 hydrograph ordinates at point Total basin hydrograph ordinates at point A A A. The channel-routed flood hydrograph ordinates from the reservoir at point B are then added to the flood hydrograph ordinates from subbasin 4. or no such facilities may be present. While the general procedure is not complex. which yields the total flood hydrograph at point D. The number of subbasins may be more or less than the six used in the example.FLOOD ROUTING Step 4. from point D is then routed to point E and added to the flood hydrograph ordinates for subbasin 6. a table such as table 5-l should be Table 5-I. the “bookkeeping” associated with ensuring that all subbasins are added in the proper sequence and that hydrograph timing is in proper synchronization can become somewhat involved. Procedure Time. The flood hydrograph representing the total flow at point C is then channel routed downstream to point D and added to the hydrograph for subbasin 5. column (2) plus (3) (5) plus (6) reservoir routing (9) plus (10) D D. the number of dams and reservoirs may be greater. The hydrograph The previous example provides a general illustration of the type of flood routing problems encountered in most flood hydrology studies. Step 6. depending on the individual hydrologic situation encountered.

in the case of a gated spillway. the general relationship for routing flood hydrographs through reservoirs can be expressed by: AS = Q. The storage accumulated or depleted in a reservoir is a function of the inflow and outflow rates. rate of inflow during time interval At. Similar curves for the outlet works or other appurtenant release structures are used when studies indicate that these structures are reliable and will be available for accommodating the inflow flood. = average Q0 = average At . The quantity of water a spillway can discharge depends on the size and type of spillway. The rate of outflow from the reservoir is obtained from the rating curve that relates spillway discharge and reservoir water surface elevation. The change in storage is determined from the elevation capacity curve (fig.Q<. the instantaneous rate of inflow at any time t can be determined from the hydrograph representing inflow to the reservoir. Outflows through during flood routings need not necessarily be limited spillways. but may be assumed to be supplemented to releases by other 136 . However. where: AS = storage 4. the outflow can be varied with respect to reservoir stage by operation of the gates. For example. or an alternative assumption might be made that the gates will be opened at a known rate so that surcharge will accumulate before the gates are fully open. This is because the three necessary components of reservoir routing are hydrograph. the discharge will vary with the hydraulic head on the spillway crest and the surcharge storage. 5.2 Reservoir Flood Routings The routing of flood hydrographs through reservoirs is considerably more straight forward and precise than routing through river channels.FLOOD HYDROLOGY MANUAL prepared for all studies when flood hydrograph routing and combining are part of the flood hydrology study. (2) reservoir storage known with some precision: (1) inflow capacity verses elevation relationship.At accumulated or lost during time interval At. one routing assumption for the operation of a gated spillway might be that the gates will be regulated so that inflow and outflow are equal until the gates are wide open. Each flood hydrology study should present a table similar to table 5-l to facilitate full comprehension by anyone knowledgeable in the flood hydrology area. and outflow during time interval At. Table 5-l reflects the physical condition of the basin as shown on figure 5-l. Referring to figure 5-2. For a discrete interval of time At. 5-4) that relates the reservoir water surface to its capacity at a given elevation. and (3) outflow rating curves for spillways and other release works. as shown on figure 5-3. For a simple overflow crest spillway.

releases such as through river outlets.Inflow hydragraph ( Inflow-time curve) I TIME tt) IN HOURS Figure 5-2 .6 3.2 3. and powerplant turbines. the size.-Typical spillway discharge curve.-Typical inflow and outflow hydrographs.. 288-D-2399. / 302 /’ I ’ Overflow length spillway crest of 25feet --. 288-D-2401. In all such cases. and method of operation of the spillway and other release works with reference to reservoir storage and flood inflow must be predetermined to establish an outflow versus elevation relationship for the reservoir routing. 1 C: 3.0 lpoo 300 0 200 400 600 C’= CLHsn For H:I 2 3 4 5 I 600 __ (200 DISCHARGE IN CUelC FEET PER SECOND Figure 5-3. irrigation outlets.FLOOD ROUTING .7 3. 137 . These operations should follow the standard operating procedures for existing reservoirs.4 3.. type.

a solution of reservoir flood routing could be made by mathematical integration. These techniques vary from a strictly arithmetical method to an entirely graphical solution. which are the same regardless of the approach used.200 1. If basic equations could be established for the inflow hydrograph curve. outflow releases (as may be modified by operational procedures).200 RESERVOIR CAPACITY Figure s-4. The data required for reservoir flood routing. a manual “trial and error” method similar to the computerized approach is used for illustration in this chapter. In column putations. are as follows: l Inflow Outflow Reservoir hydrograph versus to reservoir. are shown computations The reservoir flood routing the stepwise procedures for Step 1. each with advantages and disadvantages. figure 5-3. and l relationships. list the time increments increments are usually equal to be used in the comto the unit time period 138 . 288-D-2400. and such a solution is not possible. elevation versus figure 5-2. and the reservoir capacity curve. capacity l elevation relationships. Many techniques for reservoir flood routing have been devised.000 2.” z 306 5 F 2 Y w Y 2 oz 2 LL ? 2 LL 0 > 5 z DI 304 302 300 1. However. This computer program was developed using a mathematical (or linear) interpolation technique. These computations making these (I).600 I. general basic equations usually cannot be written for these variables. For simplicity.000 1.-Typical reservoir capacity curve.600 IN ACRE-FEET 2.FLOOD HYDROLOGY MANUAL L . Electronic computers are routinely used to make reservoir flood routing computations. figure in table are: 5-4 5-2.400 1. A computer program for reservoir flood routings is available from the Concrete Dams Branch at the Bureau’s Denver Offtce.

Use the reservoir capacity versus elevation relationship on figure 5-4 to determine the reservoir water surface elevation. Step 2. Step 5. and then enter in column (12) the reservoir surface elevation that corresponds to the storage shown in column (11). Step 11. List in column (2) the interval between time increments in column (1). enter this average value in column (8).1 (g)(l) of during the recession of to the unit time period used near the time the level to ensure accurate reservoir inflow flood hydrograph discussed chapter 4. Column (3) is completed by reading the hydrograph ordinate in cubic feet per second at the time increment listed in column (1). Step 9. If these elevations do 139 . These individual time increments will be referred to as At in the remaining steps. Longer increments may be used the hydrograph. Step 6. Step 3. Compare the reservoir water surface elevation in column (12) with the trial reservoir elevation in column (6). Step 8. Assume a trial water surface elevation in column (6). Values in column (5) are obtained by converting the values in column (4) from cubic feet per second to acre-feet. and then determine the corresponding rate of outflow from the outflow versus elevation relationship shown on figure 5-3. Column (4) is the average inflow for time interval At in cubic feet per second. Column (10) is column (5) minus column (9). 1 cubic foot per second flowing for 12 hours can be assumed to equal a volume of 1 acre-foot. Step 4.FLOOD ROUTING used in developing the in section 4. Obtain a value for column (9) by converting column (8) values in cubic feet per second for time interval At to acre-feet. Step 10. Values are obtained by averaging successive values in column (3). For these computations. Step 12. Determine subsequent values by adding the AS values from column (10) to the previous column (11) value. Average the rate of outflow determined in step 6 with the outflow for the reservoir water surface elevation that existed at the beginning of the time period. record this value in column (7). Step 7. however. increments equal for the inflow hydrograph should always be reservoir attains its maximum water surface computation of this critical elevation. The initial value in column (11) is the reservoir storage at the beginning of the inflow hydrograph to the reservoir.

The hydraulic approach relies on consideration of the principles of conservation of both mass and energy.4 Successive Average Lag Method This method. Since the area under the reservoir inflow hydrograph (curve A) indicates the flood volume. Currently.3 Techniques for Routing Floods Through River Channels The routing of floods through river channels may be accomplished using two basic approaches that are generally called the “Hydrologic Routing Approach” and the “Hydraulic Routing Approach. only the Modified Puls Method is strictly a hydrologic routing approach. this method does provide reasonable results for project designs when the flood wave travel times are based on the analysis of actual major floods in the basin of interest or in nearby hydrologically and hydraulically similar basins. However. there are four methods in general usage among the Bureau’s various offices: (1) Successive Average Lag Method.FLOOD HYDROLOGY MANUAL not agree within a specified degree of accuracy. and (4) Muskingum Method. Use of this method is suitable for preauthorization type planning studies. Of the four techniques. Inflow to a river reach minus the outflow from that reach equals the change in storage volume contained in the reach. The surcharge storage computed for column (10) in table 5-2 can be checked by comparing with the measured area converted to acre-feet on figure 5-2.1 foot. (2) Modified Puls Method. the volume indicated by the area between the two curves will be the surcharge storage in the reservoir during passage of the inflow flood. 5. Tatum [62]. These techniques are described in the following sections. a second trial elevation in column (6) and repeat the procedure the specified degree of accuracy is achieved. If suitable data are available. 140 . developed by Fred E. The following discussion relative to the basis of Tatum’s method and procedures for its application has been substantially extracted from Corps of Engineers Manual Routing of Floods ‘Numbers in brackets refer to entries in the Bibliography. the others fall somewhere in between the two approaches. as discussed later. Therefore. about 0. the area under the outflow hydrograph (curve B) indicates the volume of outflow. for which there are many different application techniques. (3) Modified Wilson Method. 5. relies primarily on the basic storage relationship.’ is not as sophisticated as alternative methods that more rigorously account for the effect of channel storage on the attenuation of a flood hydrograph.” The hydrologic approach. methods using “storage routing” techniques should be used for studies supporting postauthorization or final designs. make until The outflow versus time hydrograph that results from the reservoir flood routing shown in table 5-2 has been plotted as curve B on figure 5-2.

for dt.7 306. at elevation at time t.415 .050 1 Inn I I”” 1:100 101.417 2.505 1.000 333 6.300 2.610 2. 377 Oh” JIJ 369 (11) Total storage. acre-feet (7) (8) (f-5) Trial Average reservoir Outflow rate of storage outflow.6 306. J 10 A 80 on0 J”” 260 *In 690 0 J 5 00 JL 45 *no J” 170 AOC 475 Outflow.1 306.Jl 236 Ohm J.630 1.0 0”” L 306:2 306.&1x I:213 1 “Ah 11449 1 “C” I ooc11826 0 anA & IJJ 21195 At.000 417 2.610 4.000 5.6 306. feet Remarks 300. Qi for At: fP/s ftS/s (3) 400 600 50 (5) routing computations.J”L 1 302:l no 0 J”J 0 303:8 ens.430 1. acre-feet 1.000 3. c-9 (10) Incremental storage AS.400 117 (2) (4) Average Inflow rate of inflow tii: t. hours 0 1 1 1 2 1 E 3 1 4 1 5 1 6 1 7 1 8 1 9 2 11 500 1.100 800 133 306. acre-feet II 0 J 4 1P I” 14 *n ” 40 c)” d 77 (12) (13) Reservoir elevation at end of At. acre-feet .7 1.350 446 4.528 2.n J” 50 11” IIT 113 0”” L.350 112 306.400 1.850 238 306.535 2.6 1. QO.1 1.630 1.000 167 306.4 1.-Flood (1) Time t.0 306.Table 5-2. fP/s time t fP/s nn L “” cl 300:3 0 .650 3. hours Inflow.1 r VW OK L OK r L”W OK OK OK OK High OK 1..400 2.520 128 136 136 251 253 110 31 -24 -118 -120 2.3 87‘3 I .700 4.6 1.600 1.000 250 on0 n J”J J 303:8 800 1.540 1.559 2.000 5.

FLOOD HYDROLOGY MANUAL Through Riuer Channels [62]. + w2. is equal to the mean inflow A for the time interval t... . Hydrograph B is hydrograph A translated without change in shape to a downstream point to which the travel time is t/2.)/2 at station the inflow at time t.. l If I. (a) Basis of Method. the routed reach outflow discharge for hydrograph C is obtained. The inflow discharges . let hydrograph A be defined by inflow its properties.. .. reflects network the cumulative effect above the measuring l The shape of the observed hydrograph of the storage conditions of the drainage point..I. In each successive subreach... + 1. the midpoint inflow discharges of 142 . Therefore. This is presumed due to the floodway having been configured by successive inflows from the tributary area over a considerable period. I. (I. It is likely that the successive hydrographs obtained by this procedure are not necessarily spaced at equal intervals along the stream. at a downstream measuring station B. .+l are : at times t. . discharges lo. . Lag l The basic premises of the Successive Average Method are as follows: The storage versus flow relationship and the associated flood hydrograph shape tend to vary uniformly along a watercourse and its associated flood plain. but that they are equally spaced in time. to t. to the routing reach at times t. the process may be repeated for as many subreaches as desired to determine the change in shape of the hydrograph as a result of routing through channel storage... -_ (~0+ 1. It is assumed that this relationship between inflows at stations A and B applies to all time periods with the same time interval as t. l It is assumed that the hydrograph at B reflects. which represents the “first-step” hydrograph of the successive average la) method. to t. the change due to storage conditions in the reach between A and B. t. at measuring station A then. t. For some subreaches. t. .. in its altered shape..-Further consideration of the method discloses On figure 5-5. t. G-1 + znv2 By drawing straight lines between these inflow discharge hydrograph. . (1. I. and ZY represent flood hydrograph ordinates at times tr and t. with time intervals At small enough so that the inflow discharge varies linearly.)/2.. the velocity of translation of the flood wave may be more or less than for others due to local variations in storage conditions.t.

the more the routed hydrograph retains the shape of the original hydrograph. and n is the number of 143 . That is.Zn+i. (b) Routing Constants and Routing Procedure. t. Consideration of the geometrical properties inherent in this method is sufficient to disclose that the crest and shape of the hydrograph. TIME flood hydrograph t5 routed t. C I \ \ t. by Tatum Method the preceding hydrographs are connected. Figure 5-5.-The ordinates of a hydrograph routed successively through n subreaches may be expressed in terms of the outflow ordinates. Thus. t7 t* [62]. + 19 4 (2) and. t. + 21. + 31. 0. which results in a flattening of the wave form in its downstream procession. if: 0.-Hypothetical 103-D-1920.FLOOD ROUTING I I Hydrograph I3 I H ydiograph. t. t.. . if there are three subreaches: I. Z2. the smaller the time increment. 1%. varies with the choice of the time increment of the successive steps. + 31. = 8 (3) If the inflow subreaches. + 2 + 4 4 2 13 1 = I.. ordinates are I. of the original hydrograph. . . t. + z4 0. =+ [ -++ 1. at some distant downstream point.

... = - (4) 2” 11 cq = 2” n(n 1) l)(n 2) 2). in the routing process.000. . . The unit time of the hydrograph is usually the same as for the unit hydrograph used to develop the hydrograph.5 Modified Puls Method The original Puls Method developed by L. which is designated the “storage-indication curve” (outflow versus S + O/2).FLOOD HYDROLOGY MANUAL 0 n+1 = c.. I. the factors are applied to the known hydrograph in the same manner as the rainfall excess was applied to the unit hydrograph to obtain the runoff hydrograph. I. where: 1 c. The modified method is similar to the original except that the routing process is simplified by using only one curve. The factors associated with the appropriate number of routing steps are obtained from table 5-3. with the routing constants being similar to the precipitation excess increments and the known hydrograph assumed to be the unit hydrograph. G. = 2”(n n! I _ 1 C n+1 2% ! Routing routing constants subreaches. + cpz* + CQ +. 2” are shown in table 5-3 for the various numbers of The number of routing steps for a particular application can be determined by dividing the travel time in hours (from the point where the hydrograph is known to the point of interest) by the unit time of the hydrograph. and then multiplying the result by 2. from rating curve at lower end of reach. as ordinates storage plus one-half of outflow (S + O/2) values as 144 .(2) l)! c3 = cq = 2”2 ! n(n n(n -1) 2”3 ! (Yr 1 6. For a particular application..+* z. 5.. Puls [69] was also known as the Method of Inflow-Storage-Discharge Curves. The storage-indication which are obtained and corresponding curve is drawn using instantaneous outflows.. This curve is also applicable to the routing of floods through reservoirs.+ c. It should be noted that the sum of the routing constants for any routing step is always 1.

0039 9 0.1172 .0020 10 0.0078 .3125 .2187 .0176 .0547 .0039 .0020 .0313 .2734 .1250 4 0.1562 .3126 .0156 7 0.0625 .0703 .0176 .0098 .3125 .0937 .0625 of Routing 5 0.Table 5-3.1095 .-Flood routing constants by Successive Average Lag Method Number 1 0.2743 .OOlO .0313 .2050 .0098 .0703 .5000 2 0.2500 .0078 8 0.0313 .5000 .1172 .1641 .2344 .3750 .2050 .0440 .1250 .2500 .2734 .0440 .0313 Steps 6 0.2187 .2460 .1641 .2500 .0156 .2460 .1641 .3750 .2344 .0010 .0937 .1741 .1562 .2560 .0547 .3750 .2500 3 0.1094 .5000 .

0. Note that the right side of equation (5) corresponds to the term on the abscissa scale of figure 5-6.-Storage indication curve. S.FLOOD HYDROLOGY MANUAL abscissas. Also. = storage in reach at end of time period t. I./2 on the left side of D 200. = inflow at start of time period I. by subtracting 0.000 $ ii % L f ” m : 5 G g ii I2 eqooo I20. = storage in reach at start of time period t. in cubic feet. = outflow at downstream end of reach at end of time period cubic feet per second. in cubic feet per second. which is a rearrangement of the traditional inflow minus outflow equals change in storage equation: (5) where: t. t. or (S. in cubic feet. yields S. in cubic feet per second. and S. . + 0. The signifkance of this curve in routing can be readily understood by referring to the following basic equation.000 i60. This resulting expression is identical to the S.O.000 (S+f)IN I-DAY CUBIC FEET PER SECOND Figure 5-6. = inflow at end of time period t. 103-D-1921 146 ./2) . = outflow at lower end of reach at start of time period t. in 0. ..0. (ordinate scale of figure 5-6) from S2 + 0.OJ2./2. in cubic feet per second.OOO 40.

Z*. Knowing the left side of equation 0. . then S./2 for the first time increment. solve for left side of equation (5): 130.000 2 l-day ft5/s The outflow from the river reach at the end of the first day may then be determined from figure 5-6 using an S.FLOOD ROUTING equation (5) except for the subscripts. Subtract the 0.000 = 17. S.000 = 202.0.000 l-day cubic feet per second. + Of/2 value./2. Since subscript 1 denotes values at the start of a time increment and subscript 2 denotes values at the end of a time increment. + 0.000 on the ordinate scale. Step 2. To determine outflow at end of second day.000 l-day fP/s 2 147 ./2.000 yields a value for S./2.000 cubic feet per second at the start of first time increment is shown in column (5) of table 5-4.0. outflow obtained from figure 5-6 from the corresponding computed abscissa.0. a value for S./2 is determined by subtracting the outflow at end of first day from corresponding S.000 + 12. for the first time increment.0. . ./2 for the second time increment is equal to S../2 value of 112.95. Subtracting the outflow of 70.0.OJ2 at the end of one time increment is numerically equal to S. Compute a numerical value for the left side of equation (5) by substituting given values of I./2 in equation (5) can be found from figure 5-6. This at the end of the Step 3. the left side of equation (5) can be computed for the second time increment and the procedure repeated. Again.0. . where the flood hydrograph shown on figure 5-7 is routed over a single river reach. which (fig. Consequently. + 0.000 and reading an outflow value of 95. + 0. + 0.000+240.000 + 17. or 112. Given this outflow./2 of 12. The storage-indication curve representative of that reach is shown on figure 5-6. and is found to be 82. The detailed routing process is as follows: Step 1. (5).000 l-day cubic feet per second. 5-6) value for the reach is equal to S2 + and read the 0. which gives a numerical value for S./2 at the start of the succeeding time increment.. A tabular example of the Modified Puls Method for routing computations is shown in table 5-4. S. . enter the storage-indication curve outflow corresponding to the computed outflow is the instantaneous outflow from first time increment. .000 from 82. S. The given O2 outflow from the river reach of 70. . The value of S.000 = 112. Solving for the left side of equation (5): 70. the value of S. and 0.000 ./2.000 l-day cubic feet per second.000 + 130.

ft3/s (2) Instantaneous Inflow I.000 80. R = hydraulic radius.000 112.000 125. in feet per second. S = slope of channel.000* 95.000 70. + 0.000 100.FLOOD HYDROLOGY MANUAL Table 5-4.000 182. cross-sectional 148 . computations (3) Average Inflow per Day I.000 185.000 l-day cubic feet per second is 176.000 237. + 47 9 2 ft3/s by Modified (4) outflow . in feet.000 95.000 163.000 cubic feet per second.-Example (1) Time./2 value of 202. days on routing -. In the preceding example. Table 5-4 shows that this process is repeated for each time increment until entire flood hydrograph has been routed through the reach. and n = Manning’s hydraulic roughness coefficient. R2/3 S”’ n (6) where: V = velocity.000 68.500 112.S+.000 209.000 211.000 time increment The outflow at end of second day can then be determined using figure 5-6.000 90.000 240.000 is either known or assumed for start of first 2 3 4 5 6 7 *This value problem. gauging stations were in operation at both ends of the reach. which illustrates the basic principles of the Modified Puls Method. gauging stations will not be available and these relationships will have to be computed using Manning’s Equation: v= 1 486 . Frequently.000 in any routing 130.000 146. which facilitated the derivation of the storage and discharge relationships. (5) Computed outflow.000 121. 188. which is found by dividing area of flow by wetted perimeter of channel.000 138. in feet per feet.000 202.000 160.000 50.500 105.000 82.000 176.500 76. l-day fts/s Puls Method. ft3/s 0 1 70.000 63. where the outflow value corresponding to an S.

With the discharge at the lower end of the reach and the volume in the reach both related to a common factor. can then be found by multiplying the velocity V. in square feet. or Q = VA. The basic data required to apply equation (6) are slope of streambed. 5-6) can be derived. the storage-indication curve (fig. from the flow velocities and average cross-section. the outflow in cubic feet per second at the lower end of the reach can be computed for various depths of water. In the preceding example. it was assumed that all of the water that entered the upper end of the reach was discharged at the lower end of 149 .000 tz a 73000 50. typical cross section for reach. the storage in the reach can be computed for various depths of water by calculating the volume in the reach corresponding to the depths of water on the average cross section.000 0 I 2 3 4 TIME 5 IN 6 DAYS 7 8 9 IO II Figure 5-7. the average flow velocity can be determined for various depths of flow and. in feet per second. Using equation (6).000 AT LOWER END 6 22 c3 &j 100. Also. depth of flow. times the cross-sectional area A.000 25. and a coefficient of roughness for the reach.000 L3 25OQOO iti 8 m 225. The discharge Q.-Hypothetical 103-D-1922. in cubic feet per second.FLOOD ROUTING 275. flood hydrograph routed by Modified Puls Method.

if the ground-water inflow is appreciable. By applying the wetted perimeter for various depths of water. This condition simplifies the routing process. in most cases. Typical cross section shown on table 5-5 Percolation loss assumed to be 0.7 miles Change in elevation of streambed over length n = 0.5 cubic feet acre Puls Method a percolation as it will be are given: of reach = 800 feet per second per wetted The computations used to derive the storage-indication curve and the percolation loss versus outflow curve are shown in table 5-5. that of no tributary inflow. The given hydrograph at the upper end of the reach and the computed hydrograph at the lower end of the reach are shown on figure 5-9.075 Coefficient of roughness. The storage-discharge relationship derived in previous examples is for only one given condition. it was assumed that the total outflow at the lower end of the reach originated at or above the upper end of the reach with no contribution to flow from the intervening tributaries. the procedure is identical to that described in the previous example. These two curves. However. Wilson’s streamflow both the translation This method is based on a modification routing procedure [63]. Percolation losses may be reduced to zero. This method 150 . but frequently the tributary inflow in the reach is relatively large and should be considered in the routing process. which is comparable to table 5-4 used in the previous example except that the percolation loss is included in table 5-6. The following basic data Length of reach = 10. this is not exactly true because some of the water will infiltrate by percolating into the streambed as the water flows downstream. Having determined or assumed a percolation loss in cubic feet per second per wetted acre. a curve for percolation loss versus outflow can be drawn. for illustration purposes. This percolation loss will vary with the geologic and so-i1 conditions of the flood plain but. In the previous example. 5. respectively. In many cases.6 Modified Wilson Method of Walter recognizes T. The following example is given to illustrate the Modified for a stream that has no gauging stations in the reach but has loss. and column (7) versus column (1 l). are determined by plotting the column (7) values in table 5-5 versus corresponding values of column (9). or actually become negative. After the storage-indication curve is derived. the wetted acres in the reach can be computed.FLOOD HYDROLOGY MANUAL the reach. The results of the routing computations are shown in table 5-6. the loss will be less than 3 cubic feet per second per wetted acre of streambed overbank area. shown on figure 5-8. a percolation loss will be assumed in the next example. Tributary inflow is not considered in this example discussed in subsequent sections.

118 78. 1.92 1.5 g ?ii loss = 0.o'.875 2.34 3. ft* (4) Hydraulic Radius R.43 1.520 67.420 15.91 3.980 29. acres Percolation Loss fF/s ft3/s 1 2 3 4 5 6 7 8 9 10 *Storage 300 350 400 450 500 550 600 650 700 750 in l-hour 275 600 975 1.358 R9.70 (43.494 34.409 26.jo1416)r/z = 2.275 5.02 5.38 7.54 6.30 6.028 10.95 1.163 7. 12 = 0.486 R2.400 2.389 25.27 5.-Stream (1) routing data by Modified (7) Puls Method.71 2.496 = 0.560)(1.13 2.44 3.67 2. feet Wetted Perimeter.400 1.120 20.983) Length of reach = 10.804 83.763 45.708 56.600 4. (8) e-9 (10) (11) Depth.496 feet Slope = 800/56.308 21.687 41.88 8.975 3.318 9.650 15.Table 5-5. feet (2) A.568 33.37 4.11 6.11 3.96 5.680 46.41 2.35 (6) Discharge Storage* s+ 9 1-ho% ftS/s Wetted Area.36 4.912 69.Percolation ::.240 56.81 2.438 37.54 2.137 0.962 99.375 389 454 519 583 648 713 778 843 908 973 194 227 259 291 324 356 389 421 454 485 cubic feet per second: S = Area times (56'4g6)(24) = Apa times 15.750 4.68 6.13 3.24 3.7 miles = 56.500 10.01416.86 7. (Not to stole) . feet (5) p/3 Average Water Velocity V.000 0. ft/s 2.431 17.I:.75 4.5 cubic feet per second per wetted acre of channel.075 Cross-sectional area 750 feet '+I .022 4.


450 16.700 2.200 3.300 25.400 11.500 7.100 :xE 151500 13.900 4.300 6.700 3.100 2.000 2.650 11.300 2.550 15.~8 6:800 6.o 71000 6.600 9.000 1.000 1.200 6.450 14.500 3.400 5.200 1.100 2.400 4.750 5.600 2.000 23.000 2.050 14.500 22.300 2.300 2.600 3.900 5.000 1.200 2. hours routing computations by Modified InstatELeous Outflow.400 2.600 2.250 :3:x 26:500 29.~~ 2.900 12.000 4.150 2.500 4.500 2.200 3.350 2.300 2.400 3.-Stream Tk!e Period.800 8.600 12.300 11.300 3.800 3.300 50.900 18.650 zz 11:800 11.200 1.450 17.000 25.250 4.250 I ~ X8 26:200 24.100 3.300 153 .000 1. ft$/s (3) Storage Indication l-hour 1. fF/s 800 800 800 1.000 1.550 13.800 8.900 2.700 4.000 Ki8 2:250 3.100 6.100 12.000 (2) 9’.600 38.750 28.000 .800 2.000 6.600 18.050 $%X8 52:250 63.800 2.~.800 4.400 5.550 18.700 Fxt 64:750 60.400 1.450 2.900 9.300 23.900 6.700 2E 24:lOO 22. InstaELeous outflow Minus Percolation.900 2.000 %o” 8:050 9.900 2.550 11.000 3. ftJ/s Puls Method.900 56.100 2.900 32.750 15.200 3.700 388 5:300 4.600 4.850 41.750 :~%~ 14:700 ::‘% 10:000 9.100 ft5/s 1.100 10.500 %.FLOOD ROUTING Table 5-6.o” 20:650 19.900 %8 2:650 _.600 4.900 3.200 2.700 2.000 2.200 3.400 2.800 45.‘fi’r Per Hour.800 11.550 2.300 4.000 7.600 3.400 3.400 2.250 5.200 2.400 ‘.500 11.000 34.650 ‘2p.000 13.500 17.250 2.2E 211700 :Kz 151200 13.800 10.

5 T. (8) of peak flow through T.000 0 0 6 I2 18 TIME 24 IN 30 HOURS 36 42 48 Figure 5-S. in hours the two components 2T + At for T < 0. and consists of 154 . 0 10. the Modified Wilson Equation takes the form: 0. and I.-Example 103-D-1924. and = 0. which is inflow minus outflow equals change in storage over some discrete time interval. of K is then found by applying At the equation: The value K= where: T = travel time. = consecutive incremental inflow discharges at upstream end of stream reach. reach.. Based on this fundamental relationship. + z* 20.) 0. channel storage component (expressed as a function of time T. where: 0. and K = a coefficient dependent on translation time of flood wave T. I. and T...FLOOD HYDROLOGY MANUAL P 9 “. and the routing time interval At. The basic mathematical equation used in this method has as its basis the standard relationship. + K(Z. of a routed flood hydrograph by Modified Puls Method. = consecutive incremental outflow discharges at downstream limit of stream reach. time of a flood wave and the effects of channel storage on peak discharge attenuation.002 5.

(0. and C2 have values of: 155 and ending of time period . and the term KX (I. for (9) the The term KO in equation (9) has been considered as representing the prism storage under the profile for steady flow 0 (fig.FLOOD ROUTING Investigators have found that this method is relatively well verified in the mountainous areas of California. 5. The following explanation of the Muskingum Method has been substantially extracted from the Corps of Engineers Routing of Floods Through River Channels [62]. reach. The Muskingum method stems from an assumption of relating storage within a routing reach to outflows at each end of the reach.0) as representing the wedge storage produced by variations from a steady-flow profile due to differences between inflow and outflow occurring during rising and falling stages. are about equal. and coefficients C. and storage time T. (Z2 ZJ where the subscripts refer to the beginning (11) t.0. + 0.)] to yield: (10) 0. In these areas. respectively. For application of flood routing.0.” was developed for application by COE in connection with studies of the Muskingum Conservancy District Flood Control Project in the mid 1930’s. . 5-lo). equation (9) is expressed in increments and combined with the continuity equation (1) in section 5. the translation time T. + Z2).2: AS = 0. This method.7 Muskingum Routing Method This method is one of the more sophisticated of the hydrologic flood routing procedures. and I and 0 = simultaneous inflow and outflow. In either mountainous or plains basins. This method has also been used successfully in the Texas high plains basins. The Muskingum or Modified Puls method should be used in flood hydrology studies to support final designs leading to project construction. and hydraulic roughness characteristics. = c. (II . channel cross section.) + c. which results in the basic equation: s=K[XI+O(1-x)]=KO+KX(I-0) where: S = storage in the reach. it is imperative that the travel time value T be determined from the analysis of an observed event on a hydraulically similar channel with respect to slope. sometimes called the “Coefficient Method of Routing.5 At [Z. K and X = constants.

10. consideration in selecting At. in operating problems requiring accurate stage forecasts throughout a flood period.) = (l/1. (I. c.0) = 0. The inflows are tabulated in column (2) of table 5-7. The coefficient method may be expanded by using a variable for K and/or X. the routing may be made with several values. and C.2) (2.2KX 2K(1 . and X = 0.) = l/6 (7. If these coefficients are assumed to vary as functions of outflow. length of routing period. The practical that its value consideration tively straight.X) + At and C.-Muskingum routing principles.2.0 . The first entry in column (3) is C. the values of C. The second entry in column (5) is the summation of columns (3).I. .5 day. each applicable to a specific range of outflow discharges. which also completes the routing. the computed outflow hydrograph may be found to have more or less volume than the inflow hydrograph. for example.8. In this example. = 2K(l 2At . The use of constant values for coefficients K and X throughout a flood routing is not adequate in all instances.0. the first entry in column (4) is C. From equation (1 l). The entries in columns (3) and (5) are determined alternately until column (5) is completed.3. A secondary is that the water-surface profile in the reach will be relaThe relationship of At to the constant K is discussed in 156 . may be plotted against the outflow for use as the routing proceeds step by step. (Z2 . or working curves using C. an inflow hydrograph for a reach of the Tuscarawas River in the Muskingum Basin is considered to be represented adequately by instantaneous values of inflow at half-day intervals. 103-D-1925.0) = 0. (4). then C.0 . Also given are the conditions that K = At = 0. are computed.FLOOD HYDROLOGY MANUAL Figure 5. is multiplied in turn by each increment of inflow to obtain column (4). Therefore. and C. it is assumed there is no tributary inflow and no loss in the reach. When using varying values ofK and X. The first entry for outflow in column (5) is assumed to be equal to the inflow at the corresponding time. A readjustment of the relationship between the coefficients and outflow or stage is one means of bringing the volumes into balance. For the example shown in table 5-7.X) + At Routing of a hydrograph using equation (11) is not an involved process. and (5) of the preceding line. is is small enough to define the hydrographs. = At .2.

600 28.400 15.5 day. K = 0. :::: 2.700 -1.400 15. For t = 0.400 5.400 P 8 g 5 is 3-04-29 .200 ftS/s 0 3.500 8. Method.400 -3.000 29. Ohio Reach from Newcomerstown to Coshocton. = At-2KX 2K(l-X) + At = 7 (3) WI-0.200 6.m. 5. a.100 11.000 5.300 11.200 8. .800 19. p.100 28.500 23. and X = 0.200 800 -100 -800 -700 -700 -700 -500 -300 -200 - (5) outflow (Coshocton) 0 ftS/s 2.=1/6 fP/s 800 800 800 1.m.m.:.800 7. Muskingum Basin.m.-Flood routing by Muskingum February 26 to March 4.400 6. = 2At 2K(l-X) + At = 1 1.300 11.= l/1.m.900 23. .) c.700 - (4) W-Z. a.m.400 23.600 -3.500 24.2 1 (1) Date lnE4 I (Newcomerstown) ftS/s t.m. p.Table 5-7.800 4.000 7. + 2-26-29 2-27-29 2-28-29 3-01-29 3-02-29 3-03-29 p.5 day. a.700 19.m.600 200 -3.m. a.900 4.2 and C.100 -3.500 3.000 2.m.400 -2.000 11.100 27. a.) C.700 16.700 16. Tuscarawas River.3: c. 1929. p. p.

it would appear that At should equal K. In actual applications. which is based in part on changes in the discharge at the upstream end of the reach during the routing period. Figures 5-1 l(a) and 5-l l(b) differ in the ordinate scale. the selected length of that a relationship should exist between routing period. figure 5-l 1 shows an application of this storage equation to looped storage curves derived from flood hydrographs for a river reach. (a) Determination of K Constant.7 (a). of time. and it can be shown that As defined above. For example. with the objective of showing that. A routed hydrograph is relatively insensitive to small changes in At when X and K are held constant.-To show the significance of the constant K in equation (9). Or these two figures. Storage. which also covers the determination of K and X and the selection of reach lengths. The relationship between K and the celerity of the elementary wave also signifies that K is very close to the time interval between the centers of mass of the inflow and outflow hydrographs. if At is less than K. 158 . K is the change of storage per unit change of discharge.-Application of channel storage equation. ( b) S 103-D-1926. four different methods: (1) The K value hydrographs are it is obvious that K can be determined by flood of the for a reach between two stations for which available may be taken as the time of travel Storage. However. From such reasoning. that is. K has the dimension the time interval represented by K is equivalent to the time required for an elemental discharge wave to traverse the routing reach. an elemental discharge wave will not have traversed the reach during a routing period and the computed hydrograph for the downstream end of the reach.FLOOD HYDROLOGY MANUAL section 5. At should not be less than 2KX to avoid negative values of C. a moderate departure from this equality is permissible because the routed hydrograph is relatively insensitive to the value of At.. the value of K for an incremental time period is the reciprocal of the slope of the mean line representing storage versus outflow. by a suitable selection of X. the storage loop of figure 5-11 (a) may be reduced to the relatively small loops shown on figure 5-11 (b).S (0) Figure 5-1 I. From the previous discussions. This indicates K and At. cannot represent the actual hydrograph.

67 1. numerous cross sections in the reach. the K value will be approximately equal to the slope of a curve that represents the volumes under computed steadyflow profiles versus corresponding outflows. 2L B dY where: V.FLOOD ROUTING center of mass of the flood wave. By evaluating dQ/dy by the Manning Equation. Obviously. (12) dQ = slope of rating curve. = celerity of elementary discharge wave.. of a selected discharge on the recession curve of the hydrographs. or of some other characteristic point on the hydrographs.5. The ratios in the previous tabulation were derived for a condition of constant slope. The value of K is then the ratio of reach length to wave celerity V. the ratios between VW and the mean velocity V for various types of channels would be: Type of Channel Wide rectangular Wide parabolic Triangular VluIV 1. and a roughness coefficient.44 1. equation (6) in section 5. from actual hydrographs This method is preferable 159 . VW. (4) The K value may also be determined using an inverse process of flood routing. and therefore as the selected At value increases. Equation (12) becomes less applicable as the wave height increases. can be approximated from the discharge rating curve of a station whose cross section is representative of the reach by using the following equation: v. (3) If the basic data consist of a discharge rating curve at the downstream end of the reach. of the midordinate discharge of the rising leg of the hydrographs. (2) The celerity of the elementary discharge wave.33 The value of the mean velocity V may be obtained from the discharge and cross-sectional area of the representative section. and therefore do not apply to the reach of a river entering a reservoir where slopes pivot on the pool level. and dY B = breadth of channel at water surface. this curve and K can also be determined if the storage is computed from sounding surveys for several steady-flow profiles.

5 day. the mean line being taken as straight throughout the range of discharges. and the ungauged tributary runoff was approximated from distribution graphs and rainfall records. + WI x (4 . and the value of K for each reach would be 0.) + (1 . and therefore determines the proper X for the reach. X would not necessarily retain the value of 0. the accumulated numerator (storage) values in column (9) of table 5-8 are plotted against the corresponding accumulated denominator (weighted discharge) values shown in columns (ll).x) (02 . The accumulated numerator values are plotted as abscissae and the accumulated denominator values as ordinates. with X as a parameter. The total inflow was adjusted to equal the outflow volume. For this method. The result is a series of curves for various values of X. equations (9) and (10) are solved for K to obtain: K = 0.FLOOD HYDROLOGY MANUAL to the three preceding methods because it provides a simultaneous determination of X and K. (13). The ratio of reach length to K is the rate of flood wave movement. To conform to a criteria that K = At = 0.01) Equation (13) is described in Fort Belvoir. . The one approaching more nearly to a straight line for the entire flood satisfies most closely the equation. Runoff values for the main stream and one tributary were obtained from USGS records.2.5 day. On figure 5-12. it is preferable to compute K and X for several floods and to adopt for routing an average of these values. (13) School in a publication as follows: [67] by the Engineer successive values of the numerator and denominator are accumulated for floods for which the inflow and outflow are known.2 and K = 1.00.1. + 1. The K value is the reciprocal of the slope of the curve. as shown by equation (11). In table 5-8. However. The best fit appears to be for X = 0. (15). assuming a constant wave celerity for the two reaches.) . the reach was subdivided into two equal reaches. and because K represents the change of total storage per unit of weighted discharge. column (2) represents the inflow into the reach. and (17) of table 5-8. “ . The best fit for the curves shown on figure 5-17 is assumed to be that for which there is the least variation of the valley storage curve from a single line passing through it. A plotted methods will result curve of K values 160 . The numerator and denominator of equation (13) were evaluated for each period using four assumed values of X. which are based on material shown in reference [62].(0. Because of the limits of accuracy of the runoff data usually available. The determination in a K value that of K by any of the four preceding usually varies with stage.” The latter method is illustrated in table 5-8 and on figure 5-12. . which can be applied to reaches of shorter or longer length unless there is a marked difference in storage characteristics between the desired reach and the reach for which the K value was derived. Virginia.5 At [(I.

000 l0. If trial values ofX are applied to the corrected 161 .000 20.000 STORAGE Figure 5-12.000 30. versus outflow may be prepared and.DAY C IJBIC FEET PER SECOND valley storage curves. if a specific routing for a single value of K is to be selected.000 0 0 10.000 20. If discharges derived from average rating curves are corrected for this rise or fall.000 20. the corrected discharges would usually result in loops of lesser size. 103-D-1927.-Typical IN I. The loop in the storage curves shown on figures 5-11(a) and 5-12 for X = 0 may result in part from the fact that discharge generally is not a single-valued function of stage but depends on the rate of rise or fall of the stage hydrograph.FLOOD ROUTING z 40#000 30. this K value may be taken as the value corresponding to the average of the initial flow and anticipated peak outflow.000 3opOO 0 10.

m.300 9. a.600 -900 -800 - 2-26-29 2-27-29 2-28-29 3-01-29 . a.300 -4.000 7.100 3.500 3.900 16.200 3. Denominator.000 29. ft’/s (7) 4 .500 23.800 20.900 8.600 5.) .600 2.000 -1.800 7.200 500 -2. is column (7) + X [column (6) .900 ‘Inflow to reach was adjusted to equal volume *0utflow is the hydrograph at Newcomerstown of outflow 3Numerator.500 28.200 -3.200 14.800 -1.000 -4.300 4.500 52.300 20.700 25.000 - 9.700 14.200 -2.700 19.700 5.600 -1.400 6.800 7.900 21.m.500 -4.700 26.000 17.100 -3. is At/2.m.400 -4.) .600 2.column (5) ‘Denominator.900 -3.600 4.800 19.400 7.200 61.600 -4.900 60.600 -1. a. N = 0.300 11.600 3.000 -1..100 -700 -4.900 3.600 11.m. +. p.200 -600 - w (11) D W (13) D (14) 6. 2.300 24.500 -4.100 -1.700 36. ft’/s (3) Outflows.100 -700 - 6.200 4.Table S-8.000 21. a.200 -1.f M 3-02-29 3-03-29 3-04-29 3-05-29 a.m.800 2.800 -1.900 -3.200 34.m.3 2 !i 1s + 1.000 2.700 28.) + (l-x)(0* .100 -3. p.200 4.600 2.700 -900 -500 -1.-Determination of coefftcients K and X for the Muskingum from Dover to Newcomerstown.300 -2.1.700 4. ft’/s (4) (5) (6) 12 .000 4.900 4.200 43.500 24.m.100 26.000 16.000 9.000 -2.100 8.300 -3.600 -3.000 1.400 14.1 for Assumed Values of X x = 0.200 3.900 -4. p.100 26. p.600 16.500 5.000 16.m.000 23.700 4.000 -1.o.700 42.m.400 -4.700 4.400 -1.600 -1.200 200 5.600 -4.000 -4.000 11.100 8.600 9.100 5.800 14.400 3.m.(0.800 -1.900 22.900 23.200 8. Tuscarawas River. column (4) .800 17.200 14.600 9.700 9.300 5.400 -2.400 16.400 5.300 -2.400 13.500 5.400 -4.100 -700 -4. p.500 ZLI (15) D ZD (17) 5 2 (12) 5. D.column (7)J Note: From plottings of column (9) versus columns (1 l).. a.600 6.5 day (2) Inflow’.700 11.200 6.m.300 6.000 27.300 8.700 18. 02 + 01.m.100 7.300 15.300 11.900 -1.800 - 12. Number&or. a.400 21. a.200 - 5.200 2.400 23.300 13.000 12. the plot giving K= __ the best fit is considered to define K and X.m.500 53.000 4. N.000 11.500 3.800 5.600 -2. p.500 4.)] D xv* .000 18.500 13. Routing Method. p.100 -4.900 11.100 -4.700 12. 1929.700 16. (15).500 15.m..800 -1.600 2.600 26.01.200 6.100 -3.900 19.100 -4.800 29.m.100 -4.400 -4.100 28. ft’/s ft’/s x = 0.700 5.600 -4.700 -3.700 1.500 19. ftS/s (8) SN (9) m x=0 Values of D x= and W 0.600 21.400 15.300 12. and (17).0.500 4.900 2.500 22.I.700 28.200 1.500 5.700 -1. Muskingum Basin.000 4. (IS).900 -3.200 21.500 55.600 -1.700 5.400 2.5Al [(I* + I.200 40.300 16.700 6.900 -3.700 25.400 31.700 8.800 -2.2 'D (10) 1.200 -600 - 5.000 5.000 45.200 7. February 26 to March 4.100 57.200 -600 - (1’3 7.100 7.900 6. Ohio Reach (1) Date At= 0.600 25.

(b) Determination of X Constant. it can be shown that: x= OAY 2Y0 (Z-0) (14) on figure 5-14. When At equals K. An X value of zero produces reservoirtype storage routing.FLOOD ROUTING discharges according to the fourth method for determining K. The symbol X has been considered to represent a dimensionless constant that is an index of the wedge storage in a routing reach.-Effects of a varying X constant on routed hydrographs. If the drawing on figure 5-14 represents a unit width of a wide rectangular channel. 163 . if the slope is as- the variables of which are depicted By differentiating the Manning sumed to be constant. 103-D-1928. graphically discharge equation. z-o -=-AY 50 3 ( Yo1 (15) % B s 0 0 0 (Routing8 5 through IO TIME IN HOURS four rubrrochra with At=K=l 15 20 in each rubrroch) Figure 5-lb.5 results in a hydrograph translated through the reach without change in shape.-The effect of the variation in X on the shape of a routed flood hydrograph is shown on figure 5-13. the loops for the best fit condition of K and X would more nearly approach a single curve. an X value of 0. the wedge storage and prism storage for this unit width can be expressed as shown on the figure. By substitution.

On figure 5-16. For a wide rectangular channel where changes in discharge are small and there is no variation in slope for such a change in discharge.FLOOD HYDROLOGY MANUAL Wedge storage Prism = KX (I-O) storage =F = KO = Ly. 103-D-1929.438 at Ay/yO = 0. A good example 164 . the effect of subreach length has been counterbalanced by varying the value of X.7. The first hydrograph.5. X = 0. On figures 5-15 and 5-16. The significant point in these evaluations is that.--Illustration ofwedge and prism storage. Under similar conditions. the X value for a triangular channel section increases uniformly from 0. Similiar evaluations for X can be made for other crosssectional shapes. is shown to be practically equivalent to a second hydrograph through eight subreaches with X = 0 and At = K = l/2. roughness coefficient. the hydrographs are shown for a reach that has been divided into subreaches for the purpose of routing. and having constant or small tributary inflows.3 in equation (14). within the limits of the noted assumptions. On both of these figures. the value of X is not independent of reach length. two hydrographs are represented within the width of line of the outflow hydrograph. This method is satisfactory on streams of relatively uniform cross section and slope. for a routing through four subreaches with X = l/4 and At = K = 1. This method may be necessary if the flood wave travel time in a reach between gauging stations exceeds At/&X. as might be inferred from an analysis such as the one described above. On figure 5-15. it is convenient to make the routings in n subreaches of travel time K = At such that nK equals the travel time of the recorded data. when using the coefficient method of section 5. On figure 5-16. Figure 5-14. A second method consists of making trial routings with different values of X until one is found that satisfactorily reproduces the outflow hydrograph. The hydrographs at the downstream end of the reach are shown to be dependent on the number and therefore the length of the subreaches. an X value of l/4 applies to all routings. and length of routing reach.375 at Ay/yO = 0 to 0. X depends primarily on the shape of the cross section and is relatively independent of river slope. In this case. The previous paragraphs have discussed how to evaluate X from actual hydrographs. the hydrographs indicate that.

of these conditions would be the Columbia River from the Grand Coulee gauge to the Trinidad gauge. For each subreach.3 resulted in a closer reproduction of the 165 .000 L L 0 2 : qooo iI ii 6 0 0 5 IO TIME IN HOURS 15 20 Figure 5-15.-Effect 103-D-1931. of subreaches on routed 0 5 IO TIME IN HOURS 15 20 Figure 5-16. of subreach length counterblanced by a varying X constant. Table 5-9 shows a trial routing through this reach for a minor rise in September 1945. An X value of 0. The time for flood wave travel through the reach for this rise was 24 hours.FLOOD ROUTING “z 15.-Effects of a varying K constant and number hydrograph. and the routing was based on six subreaches. 103-D-1930.000 s ii ! I0. the values of At and K were both 4 hours.

raizzzzrl~ pids the routed outflow. etc. *Routing constants.-. 6 :s :!I 17 16 14 11 . 21 . :. ~2658 $720 ft’/s Actual Discharge Trir%ad. taken from table 5-10 sValues of C.4 constants derived by Muskingum Method. 56: 8 z . ft’/s ! 5 8 P B 17 18 z1 il 86 2 li’ 0 2 f 8 . ft’/s 59 65 of Columbia Product River discharges’ using routing of Inflow and Routing Constants* $666 =&J3 &3 &I c&$5 &g &. have so little effect on values on routed outflow that they were omitted from this table.P September 1945 Day 16 HOW 9 13 Table 5-9.006 is offset by two lines. Discharge ft’/s ft’/s Trir%ad. 8 z :i :i z i 13 2. The summation $na$u~t~h. Routed LoCal Total ovemow.000 cubic feet and six subreaches. 8 f :z :: ii ‘Discharges are shown in 1. and c. 4 hours for each of six subreaches. E 15 16 :: 12 12 11 : 6 f 2 1 1 58 fil 6 A % :z 19 20 1 2 2.-* = 0.-Routing Discharge CAtud COUk. and the product in the next column is offset by three fines. Inflow. *The product of Q = 59 and c.

If the variable nature of X must be determined. It is occasionally necessary to derive X from data based on the amount of storage between steady and unsteady flow profiles. 167 . it generally may be done by establishing. Also. For the extreme case. -The ends of routing reaches should be selected with. the value of X must usually be changed. may be represented by a single curve showing X versus 0. curves of X and K versus 0 may be developed initially for the primary stream for several values of tributary inflow. A section should also be located at the upstream end of pooltype storage reaches. for which X = 0.reference to the storage characteristics of the river valley. A section should be located at any channel control that creates relatively large storage upstream.0). the stretch should be subdivided at convenient sections. flood wave travel time. Also. and time interval. (c) Selection of Reach Lengths.FLOOD ROUTING measured hydrograph at Trinidad than X values of 0.2 and 0. Although X is usually taken as a constant for flood routing. Flowline profiles in the reach may be determined by computations or field surveys initially for a constant flow and then for discharges increasing directly with the distance from one end of the reach. it is evident from figures 5-11 and 5-12 that X may vary during a flood for a given reach. If a long stretch of river intervenes between pooltype reaches. and X. The discussed method is sometimes the only practicable approach for developing X and K values for routing flow in a primary stream affected by tributary inflows or by other independent variables. like K. in which case X. The routing was made using the routing constants shown in table 5-10 because it was more convenient than routing through six subreaches by the more commonly used coefficient method. and then for the tributary for several values of main stream discharges. the selection of X must be based on judgment and derived from determinations on other streams. It is always desirable to evaluate X from several sets of data. it is desirable to retain the value of the time interval At that was used to determine X for any other applications of flood routing in the reach. by trial and error. If no data are available. Z. The difference in the storages under the two profiles having the same discharge and stage at the downstream end of the reach is the wedge storage KX(Z . and the reach in which the reservoir-type storage is located should not be subdivided.4. If At is changed. a reasonable curve of X versus outflow that will reduce the loops of the storage discharge curves to approximately a single line. The complete determination for a reach will result in a family of curves relating 0. and allowances made for effects of reach length. X varies more with 0 than with I. For tributary inflow. The storage computed from cross sections under the steady flow profile is the prism storage KO. the storage-outflow relationship for such a reach is the reservoir type. the relationship between X and storage or outflow may vary from one flood to another. Generally.

000 0 0 0 .033 .236 .071 .167 Reach with total flood-wave travel time equal to 2At .264 .I46 .041 .000 0 0 0 0 0 1.137 .012 .010 .125 .188 .032 . G-5 G-4 G-3 L-2 &l-l C” constants by Muskingum Method Value for reaches with different of Routing Constant x = 0.004 .029 .159 .176 .167 .lll .165 .026 .030 .004 0.247 .029 .592 .159 .058 .OOl .2 x = 0.OOl .296 .510 . - x=0 x = 0.271 .231 .091 .012 0.188 .077 .008 .009 .053 .694 .039 .049 .000 0 0 0 0 0 0 1.023 .008 0 0 0 0 1.037 E Reach with total flood-wave travel time equal to At 0.5 0.060 .ooo .4 x = 0.251 .005 .038 .095 .082 .OOi .125 .413 .042 .417 .087 .003 .016 .007 .173 .826 .183 .078 .521 .171 .025 .003 .697 .-s C. Residual Gl-6.015 .018 . C.005 .001 0.344 .273 .-3 G-2 G-1 c.007 .028 Reach with total flood-wave travel time equal to 3At .001 .I48 .059 .-Routing Routing Constant Symbol Residual cn-.002 .075 .333 .000 .296 .3 flood-wave travel times.259 .116 .Table 5-IO.231 .019 .444 .016 .-P G-1 C” Residual C”-5 GZ-4 C.1 x = 0.020 .286 .150 .017 .292 .045 .022 .148 .012 .324 .599 .

098 .2 flood-wave travel time x = 0.040 .ooo .244 .083 .211 .-Routing Routing Constant Symbol constants by Muskingum Method for reaches with different Value of Routing flood-wave travel times.002 .048 .004 .003 .3 equal to 4At x = 0.260 .I54 .047 .023 .-Continued Constant x=0 x = 0.048 .-1 G .045 .205 .013 .225 .023 .OQO .225 .180 .lOl .4 x = 0.OOl .066 .017 .048 .097 .202 .003 0.-7 G-6 G-5 G-4 G-3 C.149 .273 .075 .000 0 0 0 0 Reach with total flood-wave travel time equal to 5At Residual cl-s c.095 .032 .003 .I42 .242 .082 .007 .066 .llQ .150 .007 0.144 .I73 .024 .5 Residual 0.1 Reach with total x = 0.054 .086 .008 .065 .027 .156 .230 .029 .Table 5-IO.Oll .042 .351 .131 .229 .-2 C.009 .012 0.141 .ooo 0 0 0 0 1.039 .055 .051 .188 .204 .052 .I93 .212 .032 .ooo .034 .193 .465 .240 .060 .095 .307 .002 .002 .OQQ .180 .005 .194 .180 523 .013 .071 .I64 .OOl 0.ooo .

077 ..173 .419 .OOl .096 Reach with total flood-wave travel 0.077 .Oll .020 .070 .ooo + c.-s C”P7 C”d f&5 cn-4 C.002 .ooo .000 0 0 0 0 0 0 0 0 0 0 1.061 .-J.185 .064 .053 .187 .068 .141 .058 .013 .037 .071 .171 .038 .-~ + .-Routing constants by Muskingum Method for reaches with different flood-wave travel times.200 .104 .051 .ooo flood-wave travel .094 .033 .OOl Reach with total .OOl .Oll .063 0.066 .ooo 0.232 . .ooo .136 .I21 .-J.085 .015 .252 .-2 r”.ooo .018 .143 .166 .027 .-to C.220 .157 ..234 .143 .084 .009 .044 .005 .. .148 .023 . = c.137 .OO4 .-s C.2 of Routing Constant x = 0.-~ .134 .139 .1 x = 0.lOl .lOO .382 .201 .-.217 .043 .208 .189 .164 .058 .128 .152 .002 .I84 .205 .5 : 3 B P 4 E E Residual G9 c-7 0.OOl .061 . time equal to 6A6 0.ooo 0 0 0 0 1.I36 .194 .ooo .000 0 0 0 0 0 0 0 G--B 2 G-6 c-5 cl-4 G-3 cn-2 G-1 G Residual c.-Continued P Routing Constant Symbol Value x=0 x = 0. .OOl .091 .OOl .275 .170 .3 x = 0.003 .097 .I02 .ooo .ooo .Z.OOl .042 .+ c.ooo time equal to 7At .028 .094 .018 .006 .099 ..I.Table 5-10.-g c.113 .040 .006 .050 .159 .-.4 x = 0.009 .ooo + c.231 .014 .006 .064 .lOl .j C” Note: 0.060 .

and would reasonably have a higher value than the main stem itself. Referring to the basin map shown on figure 4-15. This process is shown in table 5-l 1. of Tatum Routing Steps 13 5 3 After computing the travel time.03 was used. 5. value of 0. wetted perimeter. The combined and routed flood hydrograph is then combined with the flood hydrograph for subbasin 2.2 No. Then. a Manning’s n value of 0. and channel routing length. the combined flood hydrograph is channel routed from control point C to control point B. 171 .1 3. The channel routing method used for this example is called the Tatum Routing Method. the flood hydrograph for subbasin 4 is channel routed from control point D to control point C. the number of routing steps is determined by dividing the travel time.6 5. the routing constants to be applied to the inflow hydrograph ordinates to determine the outflow hydrograph ordinates can be determined. The total PMF hydrograph at Foss Dam is the sum of the hydrograph routed from control point B and the flood hydrograph for subbasin 1. It should be noted that this n value is for the main stem of the Washita River only.FLOOD ROUTING including gauging stations. but they still needed to be routed and combined to obtain the probable maximum inflow flood. by the time increment of the flood hydrograph. in hours.069 used in the unit hydrograph derivation reflects the hydraulic efficiency of the entire drainage network including overland flow. where the routed hydrograph is combined with the hydrograph for subbasin 3. the 6. If a major tributary joins the stream. Using table 5-3 or equation (4). in accordance with the criteria requiring K to approximately equal At. which requires that a travel time through each routing reach be determined. The total hydrograph at control point B is then routed to control point A (Foss Dam). the following information was developed for each reach: Control Point to Control Point D to C C to B B to A Travel Time. Using topographic data to determine the main stem channel’s cross-sectional area. In this case. This time was determined using Manning’s equation (6) to determine velocities of flow. Water-surface profile charts are useful in locating channel controls and inflow sections for pool-type reaches. for the reach from control point D to C. hours 12.8 Foss Dam Example The example study on Foss Dam presented in chapter 4 had four subbasin flood hydrographs that had been developed. the downstream end of the routing reach should be at or near the confluence. yielding the total hydrograph at control point B. for this case. 1 hour.

050 9.600 (7) 0 0 20 140 540 1.100 285.990 10.300 87.300 113.500 139.600 129.600 81.600 324.400 147.280 8.800 77.670 8.400 (9) 0 0 0 0 30 110 380 1.200 133.200 (5) 0 10 30 130 380 950 2.500 295.390 15.600 273.500 (4) 0 10 30 130 380 950 2.200 62.500 243.200 36.200 28. (1) 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 (2) 0 0 0 10 50 190 610 1.600 77.900 222.700 193.300 118.700 27.770 7.800 228.600 88.100 265.330 10.800 !70.600 387.30c! 211.610 18.900 169.400 265.100 5.800 61.800 119.““” 198.000 36.900 87.050 2.300 232.800 222.200 102.700 90.800 287.000 134.300 276.700 27.200 77.800 127.500 149.000 99.900 295.900 345.100 424.200 28.700 101.200 130.400 cm+2film -vu.060 4.900 305.200 341.600 77.Table 5-l 1.800 27.700 207.000 281.500 5.500 320.100 170.800 253.400 117.600 155.800 66.400 60.400 331.400 163.700 371.570 3.600 22.500 197.600 (3) 0 0 0 0 0 0 0 0 0 0 0 0 10 40 110 290 680 1.400 270.800 3.900 169.800 14.400 173.200 163.400 425.600 257.600 90.600 20.600 253.300 100.300 118.000 290.660 20.600 51.000 138.000 136.500 278.000 48.300 54.100 37.900 222.600 113.430 4.500 68.600 121.200 291.300 309.400 313.-Total PMF bydrograph for Foss Dam.700 (‘3) 0 0 0 0 0 20 50 150 370 850 1.000 35.620 6.600 143.600 47.390 15.900 293.320 3.800 18.400 77.060 4.100 423.500 203.720 7.300 382.700 2 F F $ N .550 3.960 20.500 41.800 49.700 206.300 283.100 265.100 384.100 67.440 9.600 397.700 296.500 (10) 0 0 0 60 360 1.800 (8) 0 0 20 140 540 1.200 250.300 89.740 4.500 155.100 414.000 300.920 12.100 59.700 170.500 187.800 42.200 (11) 0 0 0 60 390 1.900 423.200 242.600 47.280 8.950 15.510 3.

500 169.470 4.500 101.000 72.900 138.400 65.600 16.300 12.200 60.000 39.000 172.950 2.900 23.800 209.300 98.100 166.300 9.900 20.590 7.360 6.400 11.800 34.700 209.200 48.600 322.000 188.800 175.400 54.000 182.800 33.700 224.500 6.500 31.000 11.900 153.720 5.-Continued (1) 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 (2) 43.500 111.020 4.380 4.500 115.100 9.100 54.700 80.600 203.100 188.000 (10) 58.800 181.100 55.300 202.-Total PMF bydrograph for Foss Dam.400 22.350 (8) 303.400 181.300 255.700 15.900 13.000 44.000 15.700 103.300 89.840 4.160 7.300 194.000 218.200 333.800 65.200 227.100 83.300 21.300 92.700 366.Table 5-1 I.820 6.430 3.400 176.600 51.700 11.700 75.400 278.400 28.630 5.200 210.500 338.900 176.000 300.600 49.800 93.500 5 7 8 8 5 is .700 80.800 43.900 80.600 189.500 38.000 42.500 25.800 153.600 59.900 89.200 176.200 20.500 40.600 26.360 3.400 181.900 138.900 e-9 332.100 54.100 110.400 (5) 169.600 205.100 44.800 75.200 100.800 29.600 271.400 36.900 85.160 5.000 18.800 384.800 124.700 213.600 18.400 29.500 220.700 171.000 65.800 103.000 14.500 17.400 112.800 170.300 38.300 78.500 28.500 37.400 110.800 83.600 125.500 129.800 39.300 71.000 46.700 15.100 123.400 144.700 43.100 210.900 3.600 19.400 59.100 14.400 212.000 49.200 155.600 (7) 68.410 2.200 243.700 240.500 55.200 183.200 9.500 210.400 160.500 53.600 168.200 184.000 104.000 223.900 61.200 16.100 196.200 33.900 31.200 177.860 8.800 33.500 72.300 13.900 191.100 79.880 4.840 7.800 73.600 21.600 207.900 18.200 30.400 19.800 232.060 8.800 12.700 22.100 273.200 10.300 139.600 248.300 213.900 307.400 99.200 48.700 33.700 23.800 230.900 91.700 210.900 17.700 88.700 139.200 88.400 35.900 73.300 25.500 36.900 98.700 210.800 63.400 30.460 (11) 390.950 8.970 5.800 25.300 153.200 23.100 66.200 16.920 (3) 58.100 (4) 110.300 94.600 51.700 110.700 184.600 18.550 5.000 25.900 28.800 109.300 123.950 6.900 60.500 (6) 234.700 72.


040 670 400 230 120 60 30 10 0 P 8 z i is * Time increments in tables 4-23 through .490 3.030 2.460 970 590 330 170 90 40 20 10 0 0 0 4-26.480 3.460 970 590 330 170 90 40 20 10 0 0 0 c-4 6.040 1.000 1.490 710 370 210 110 50 20 10 0 0 0 0 0 0 0 0 0 (6) 4.390 3.940 3.000 1.500 1.330 4.010 2.Table 5-l 1 .830 5.390 3.590 2. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (7) (8) 4.510 2.330 4.620 1.010 2.940 3.380 3.850 4.840 4.830 5.030 2.-Total PMF bydrograpb for Foss Dam.490 3.000 1.380 3.-Continued (1) 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 127 126 128 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (2) (3) 970 780 600 430 290 180 100 50 20 10 0 0 0 0 0 0 0 0 0 correspond (4) 1.530 2.510 2.530 2.370 5.480 3.850 4.260 890 280 80 30 10 0 0 0 0 0 0 0 0 0 0 0 0 to those tabulated (5) 2.840 4.370 5.040 670 400 230 120 60 30 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (10) (11) 6.500 1.000 1.940 3.940 3.


As time progresses and more flood data are collected. 5. Figure 6-l shows a typical example of the envelope curve relationship for peak discharges. the envelope curve will tend to approach the PMF as an upper limit. Envelope curves are of particular value to the hydrologic engineer in the development of PMF estimates because they provide information with which to judge the adequacy of those estimates. the unique hydrologic or meteorologic feature that provides the apparently low estimate should always be identified and fully discussed in the flood study report. volume relationships will represent durations of 1. There are four primary meteorologic causes that should be recognized and the data segregated accordingly: (1) thunderstorm type events where the resulting flood is caused by high intensity. 3. These relationships simply represent flood discharges that have been recorded or observed.ENVELOPE 6. Eventually. both the data on which the curve is based and the PMF estimate should be carefully reviewed to determine if some hydrologic or meteorologic parameter has either been neglected or improperly evaluated and selected. 10. each flood hydrology study should include information relative to flood peaks and volumes that have been experienced in the hydrologic and meterologic homogeneous region. If these parameters are judged to be adequate. 177 .1 General Chapter 6 CURVES OF RECORDED DISCHARGES FLOOD Considerations To provide the hydrologic engineer with information regarding flood potential in the area being studied. Exceptions are encountered in studies involving large drainage basins where the snowmelt flood component may extend from 1 to 2 months. if not. the hydrologic engineer must ensure that the flood values used represent similar type flood events with respect to their meteorologic causes. In most studies. each envelope curve will inevitably be altered in the upward direction. Considering the limited data base from which they are derived. When preparing envelope curves. relationships representing recorded flood events should never be construed as indicating an upper limit of the magnitude of future flood events. This information should be presented in the form of a curve that envelops the data points representing observed or recorded peak discharges or volumes for specified time durations versus the drainage area contributing to the observed flood runoff. The PMF peak and volume values should always be higher than properly developed envelope curves. as on the Colorado River at Glen Canyon Dam in Arizona. and 15 days because the PMF hydrograph duration seldom exceeds 15 days.


the records contained in annual water supply papers for stream gauges in the region under investigation should be closely examined to determine if the values contained in the aforementioned series have been exceeded and. the drainage area used in the development of the envelope curves is that of the storm area contributing to the flood runoff. For example. and (4) floods resulting from a combination of rain falling on a melting snowpack. For subsequent years. it is usually improper to include data representing flood runoff from steep mountainous drainage basins with those representing flood runoff from drainage basins in low relief plains regions. In these cases. (2) general rain type events where the resulting flood is caused by moderate intensity. These gauges. are installed to 179 . The computer program WATSTORE mentioned in section 2.2(a) of chapter 2 may also be used to obtain lists of peak discharge values and volumes for specified durations for a specific geographic region. rainfall that produces high peak discharges and relatively low volumes. 6. Many State governments. have installed networks of crest stage gauges. but will still produce an extremely high flood.” The volumes comprising this series provide summary tabulations of peak discharges at the national USGS stream gauging network up to the year the particular volume was prepared for publication. the more recent values should be included in the data set used to develop the peak discharge envelope curve. rainfall. that provide information from which peak discharges can be determined as described. the hydrologic engineer must assure that similar type drainage basins are represented from the hydrologic standpoint. long duration. A severe storm will frequently cover only part of a large basin.in chapter 2.2 Sources of Data As previously discussed in chapter 2. Data for developing volume envelope curves are contained in the annual water supply papers. the primary sources of basic data used to develop envelope curves are the water supply papers published by USGS.ENVELOPE CURVES short duration. Of particular importance in the development of these curves are those papers in the general series titled “Magnitude and Frequency of Floods in the United States. These lists are obtained by inputing the latitude and longitude of the point of interest and the radius of a circle around that point within which all available records are desired. generally through either their water resource agency or highway department. Each peak and volume envelope curve presented in the flood study report should provide information as to the causative factor involved. Equally as important as meteorologic causative factors. not the entire drainage basin area above the location at which the runoff measurement is made. (3) snowmelt floods resulting from the melting of an accumulated snow pack. The output will then provide all the available data. if so.

within the limited geographic area. Colorado. NWS..cluded in the flood study report. Where only the contributing area is used.3 Procedure The procedure for developing envelope curves is relatively simple and straight forward. USGS Water Supply Paper Column 5: Drainage area in square miles. Arkansas River at Pueblo. a tabulation of the data used to develop the envelope curves should be prepared for inclusion in the flood study report. e. The table should arrange the data in tabular form under the following headings: Column 1: List the identifying number as provided on the map. source of data. 180 . list only that area and provide a table footnote to that effect. Column Column 6: Date of flood event. Records of peak discharges at these gauges are published at various intervals by the States involved. Initially. boxes. These data have generally been found to provide a valuable supplement to the systematic data acquired and published by the USGS. This map should be in. 1537. Other data sources would be the various reports prepared by the Bureau. Column 4: Identify No. in water supply papers Column 3: Location of gauge on river or stream as listed in water supply papers or other source. 7: Peak discharge in cubic feet per second of time or volume over a specified period. and some local governments that provide considerable information on observed flood events. The hydrologic engineer should contact the appropriate agency with jurisdiction in the area of interest to determine if such data are available. If data are available. This is useful to cross reference the location of data points on the map and to identify their position on the envelope curve.g. 6. and bridges.g. These stations should be plotted on the map and properly identified with the conventional USGS station number or its name. a small scale map should be obtained for plotting the limits of geographical area reflecting similar hydrologic characteristics and meteorologic phenomena. e. The next step is to determine the location of all streamflow gauging stations. Column 2: Stream or river name as identified or other source. COE. it should be obtained for use in developing peak discharge envelope curves.. In addition:. both recording and crest stage.FLOOD HYDROLOGY MANUAL provide the basic flood data used in establishing design criteria for highway cross drainage structures such as culverts.

The l-. the final curve enveloping these points may be drawn. vegetation. soils. 5-. After verifying that all control points are suitable. The l-day values may represent either a 24-hour clock time amount or a calendar day amount. At this point. and for each duration for which volumes are examined. 3-. 181 . In recent years. the data point should be eliminated. The data for each point that controls the preliminary curve should then be rechecked to ensure the data represent flood runoff from basins of comparable topography. the drainage or contributing areas are to be plotted on the abscissa and the peak discharges or flood volumes on the ordinate. it should be apparent that only a few of the total data points control the position of the preliminary envelope curve. lo-. In all cases. and 15-day volumes in all cases are to represent volumes associated with the peak discharges used for the peak discharge envelope curves. data have been provided from which the 24-hour amounts can be determined. The data listed in columns 4 and 7 are then plotted on log-log paper with sufficient cycles to cover the range in discharges. and meteorological characteristics to the basin being studied. Data for the volume envelope curves are compiled from the annual water supply papers.ENVELOPE CURVES A separate table should be prepared for the peak discharge relationship. Peak discharges are readily obtained from various sources and may be used directly in the development of peak discharge envelope curves. and the drainage or contributing area sizes represented by the data. A smooth preliminary curve is then drawn that envelops the plotted data points on the high side. while the older papers provide only daily amounts. If inconsistencies are found in this final check. as illustrated on figure 6-l.


The scope of this chapter is necessarily limited because volumes could be devoted to the many aspects involved in flood flow frequency analysis. The purpose of this chapter is to describe and illustrate the application of statistics to frequency analysis used in flood hydrology work. 183 . The desired relationship is intended to provide an estimated flow corresponding to a given probability of occurrence. For a good textbook-type approach. see Bulletin 17B [14]‘. A complete review of all flood frequency approaches is also not presented because to do so would require a book in itself. the flow level is often equivalently identified as the n-year flood. For example.STATISTICS 7. This chapter is more of a guideline that provides reasoning and a “feel” for what is important. As such. ‘Numbers in brackets refer to entries in the Bibliography. where n is equal to one divided by the annual exceedance probability.1 Introduction Chapter 7 AND PROBABILITIES The objective of frequency analyses in flood hydrology is primarily to estimate the frequencies or probabilities of future flood events. The probabilities of concern are annual exceedance probabilities. this chapter will provide some insight into the process involved. it cannot be considered as a textbook on how to perform standard frequency analyses. which are the probabilities of a given flow level being exceeded once or more in any given year. Hopefully. An inverse but equivalent relationship is used to indicate a probability associated with a given flood flow. the loo-year flood is the flow level with a 0. peak flows are the subject of these analyses. there are many books written on these subjects for all levels of interest and background. however. The material is not intended to be a substitute for theoretical background or experience. what is analyzed. The term “peak flows. therefore. and examples are provided for probability relationships and some indication of their reliability. the main orientation of this chapter is towards analyses using this type of data. A review of basic concepts and an introduction to further topics is presented. This chapter also covers many of the situations that often require analysis of different types of flows or for different unittime periods. Primarily. Also. which are the largest flow rates that have occurred for each year of recorded events.01 annual exceedance probability. and what is important to study. this chapter is not intended to be a treatise on statistics or statistical methods.” generally means annual instantaneous peak flows. While this chapter does provide some excellent examples and guidance. These probabilities are best expressed as annual probabilities. theory will be presented only as it provides a basis for aiding in the application and interpretation of flood frequency results.

usually the PMF. a knowledge of the general behavior of flood flow data will allow practitioners to benefit more from their past experiences. Following that. (2) volume analysis. For this chapter. but statictics cannot prove a theory to be correct. generation capability. however. and do not cover present or future trends in research. Statistics can also be used to aid in indicating areas or relationships that show promise. To decide between competing relationships based “fit” is not defendable. diversion channel headworks. the limitations on frequency curve extrapolation. This is followed by three major subtopics: (1) mixed populations. the general philosophy of probability distributions and a basic background of their application to flood hydrology are presented. Statistics can indicate obviously incomplete or unreasonable theories. While flood control benefit evaluation is a most obvious application. Small differences in only upon the statistical statistical fit have frequently been used wrongly to reject one relationship over another. the orientation is directed towards Bureau practice in the design and planning of projects and in the analysis of existing dams under the Safety of Dams Program. Statistics should not be used to contradict reason. statistics can be used to verify that a hypothesis cannot be shown to be obviously unreasonable. and (3) ungauged analysis. An attempt was made to provide some practical guidance as to what is important and what to expect for results.FLOOD HYDROLOGY MANUAL The magnitude and frequency of floods are very important to the proper design and operation of water resource projects. The relationships may also aid in the design. who has primary responsibility in determining these benefits. A serious mistake is often made when the rejected relationship can be defended with reason. Hopefully. To complete the chapter. is then covered in a manner complimentary to that presented in Bulletin 1’7B [14]. and siting of equipment and facilities such as pumping equipment. and sill elevations for spillways. selection. A short section covers a vitally important topic. the next section presents some information on general relationships that are found in flood frequency analyses. which is the standard for Federal practice. Please note that statistics do not provide answers and do not prove theories. a section is devoted to the data that are being analyzed. The log-Pearson Type III distribution. The discussions are not oriented towards the researcher. design of low hazard dam spillways. Having given a somewhat complete background. and antecedent events for design floods for high hazard dams. The Bureau’s main uses for the derived relationships include diversions during construction. To further clarify the intentions of this chapter. while the accepted relationship cannot be. this topic will not be discussed because of the nature of Bureau projects and the Bureau’s working relationship with COE. 184 . reason is required to defend and support any such relationships. cross-drainage design. it is oriented towards the practicing hydrologic engineer who is actively involved in the estimation of flood frequency relationships.

2 Hydrologic Data Samples The data used in frequency analysis varies with regard to several important factors. the ratio could be estimated using a similar nearby stream. while instantaneous values are only available for a short time period. Although instantaneous peak data are the main form analyzed. Usually. October of previous calendar year through the following September. peak weekly. the instantaneous peak is not of critical importance. the accuracy and homogeneity of the data deserve consideration. In this situation. and annual values could also be applicable. Alternatively. monthly. a dam with considerable surcharge storage but a relatively small discharge capacity is often most critically stressed by a large volume flood rather than by a high-peaked flood. The data consist of a series of the instantaneous peak values that have one peak value for each year of record.-The data used in frequency analysis may be of several forms. other forms of data are often used. instantaneous peak flood values are used. The years are usually either water years. In some cases. Annual peak values are usually used with 3-. seasonal. either because of the unavailability of instantaneous data or because instantaneous data is not applicable to the study. unless otherwise stated. it may be assumed that the data involved are instantaneous. frequency estimates are often required for flood volumes over various time increments. daily valvues may be readily available. The analysis to be performed also influences the choice of data to be used. In addition to the basic form of the data. Occasionally. 90-. It should be emphasized that the recorded data set is only a limited sample of the total number of floods that have occurred.STATISTICS AND PROBABILITIES 7. Not all of the statistical techniques will be 185 . IO-. or by using several streams in the region. The data set itself may be recorded in any of several time units. (a) Pype of Data. 5-. 15-. For example. or calendar years.or loday (or longer) flood volume may be of prime importance. The relationship could be a simple ratio between the instantaneous peaks-and the peak daily values. 30-. Therefore. The longer record of daily values could then be used to fill in surrogate values for the instantaneous peaks. In some cases. the orientation is primarily towards instantaneous data and. In this chapter. the frequency distribution of the daily values could be analyzed and the results used to estimate corresponding instantaneous values by using the derived relationship. annual peak 30-day volumes will not behave in the same statistical manner as instantaneous peak values. a 5. which is the type of data that has been studied the most and the type for which most analysis techniques have been specifically developed. and 120-day volume values. Frequently in a study. and may result from either direct or indirect measurement or estimation. Obviously. A relationship or correlation may be developed between the instantaneous peaks and the peak daily values.

FLOOD HYDROLOGY MANUAL equally applicable to all volume-type must be taken when using the volume data and data. although rare. The same situation is true when dealing with peak 30-day values as compared to peak monthly values. and the time basis would vary from event to event. If a balanced loo-year hydrograph using the results of a set of frequency analyses with 3-. The peak 24-hour value is the largest volume in any 24-hour period starting at any time during the day.” The balanced flood hydrograph is a reasonably shaped hydrograph representing a given frequency that preserves an entire set of volume values for differing time periods. this type of data is less desirable. However. considerable extra care Most often. some of the data could be based on actual measurements while additional data is obtained from other sites. and lo-day volumes was developed. (b) Data Source. although this would be a rather rare application. Peak 24-hour values could also be analyzed. This lack of 186 . The peak daily value is restricted to the largest volume in a 24-hour period that starts at midnight. it is not clear what time period is most critical for a dam. however. 5-. the instantaneous peak could also be preserved. to start with a set of volume data that represent the peak events from each year. One approach that seems to overcome these problems with good results is the use of a “balanced flood hydrograph. an entire hydrograph is needed. To be reasonably accurate. this can be an involved process. Regional relationships using data from several sites and based on the hydrologic attributes of the drainages could also be used to furnish synthetic data. This approach is usually less than satisfactory for the frequency estimation of floods because of the lack of a distinct correspondence between rainfall frequency and flood frequency. In most cases. this is often not appropriate and is not the usual case. Another source for the basic data could be through the modeling of the rainfall-runoff process. this type of data source requires a continuous modeling process with continuous rainfall data. the resulting hydrograph would preserve all three of the loo-year volume values. the peak 24-hour value will always be larger than the peak calendar day value. however. but sometimes very important. Rainfall frequency data are occasionally used as input to an event modeling of the runoff process using the unit hydrograph approach discussed in chapter 4. Thus. It is also possible. either actual data or stochastically generated data. the data may also be estimated flows based on the transposition of flows from an upstream or downstream site on the same stream. It should be noted that there is a subtle. At times.-The source of the data is often the actual measurements taken at the site of interest. This could also be estimated flows based on transposition from a nearby similar stream. difference between the peak 24-hour value and the peak daily value. Depending on the exact problem existing. Each data element would be for the entire volume of the event. however.

All flow values above some arbitrary level are included in the analysis. An example would be the case where losses are suffered with each and every occurrence. This type of analysis is sometimes used where multiple flooding within a year is important.STATISTICS AND PROBABILITIES correspondence is due mainly to the effects of the initial soil conditions. the annual Cxceedance probability using annual peak flow data and the exceedance frequency using the full daily peak flow series (flow duration analysis) will be very similar. for example.-This type of analysis is used to express the fraction or percentage of time that flows exceed various levels. the flood potential must be evaluated on a seasonal basis. but the results should be carefully used only as a guide to floods having the same probability of exceedance as the rainfall being used. partial duration analysis. a flow of 200 cubic feet per second is exceeded 23 percent of the time. (2) Partial duration u&y&-This type of analysis is similar to flow duration analysis in that more than one value is used per year.-Frequently. for example. and analysis by flood type. a flow of 300 cubic feet per second is exceeded on a daily average basis five times per year.e. In some cases. Other types of analyses that require other types of data include flow duration analysis. (c) Data Analysis. Normally. (1) Flow duration analysis. which is the main orientation of this chapter. While this type of analysis is often of interest in quantifying flood control benefits. a multiple-peaked event would only be counted once. Flow duration analysis is most often used when ordinary flow levels are of importance. i. (3) Seasonal analysis. For the more rare type of flooding. Results may also be expressed in terms of the number of exceedances per year. this subject will not be discussed further. only one value is included for each flooding event.-The hydrologic data sample can vary with the type of analysis. this level is set low enough to include at least one value in each year and will generally have an average of three to five values per year. Usually. Most analyses are performed using annual maximum values. Since this is rarely the concern in flood hydrology. the operation of a reservoir may make it more susceptible to floods at certain times of year. During construction. It should be noted that this type of analysis is not intended for extraordinary peak flows (floods) but rather the more day-to-day flows. The resulting frquencies for the more rare floods are naturally very similar to normal annual peak analysis procedures. . The results may be expressed as the proportion of the year that a given flow level is exceeded. The rainfall-runoff modeling procedure will give a general indication of the variability and potential for floods. which must be estimated. it is not usually used by the Bureau.. This procedure should be used only after careful study and after all other approaches have proven infeasible. the site may only be vulnerable during a limited time span. seasonal analysis.

This subject is discussed further in section 7. quantify the seasonal distributions of those consequences. the probabilities are the frequency or chance of exceedance in any given single season of the type of flood being studied. it is probably much closer to the 50-year flood because. These types of flooding may take place at different times of the year and each may exhibit different statistical behavior. consider a situation where each year has two seasons of equal flooding potential with all other seasons having inconsequential flooding potential. (d) Accuracy the basic data is usually assumed to be free of errors. a stream may be characterized as having thunderstorm floods.FLOOD HYDROLOGY MANUAL and it may be necessary to analyze the flood frequency on a seasonal basis. and snowmelt floods. The 1 in 100 seasonal floods may be equal for both seasons. and combine those distributions into one single annual probability distribution. Conversely. one for each season in a loo-year period. the raw flood data should be divided into a separate series for each type of flooding. the data would consist of only data from the season in question. However.7. flooding on a given stream may be due to more than one type of flooding. the flow values are assumed to be accurate estimates of the actual flows. (4) Analysis by Jood type. this combining process is not just the simple addition of exceedance probabilities. although they increase the variability of the data. of Data. are generally considered to have only an insignificnt effect on the results.-1 n many instances. however. For example.-For the application of flood frequency analysis. This situation is referred to as a “mixed population condition. there is the potential for errors. A mixed population analysis is often required and does improve the flood frquency analysis. As an example. the effect is only significant for the smaller 188 . Random errors. In this case. To avoid understating the risk. If the systematic error is a constant value. In fact. Errors tend to be classified as being either random (no pattern) or systematic (having a pattern). systematic errors can be very significant. To properly evaluate the actual risks involved. However. this is not the case. tend to compensate and. on the average. being sometimes high and sometimes low.” because one population consisting of all floods is really a mixture of the floods from distinctly different subpopulations. two floods should be expected. floods from general storms. There is a possibility of understating the flood risk when using seasonal analysis. care should be taken to evaluate the total risk rather than only the individual risk due to each flood type. Protecting against the 1 in 100 seasonal flood for each season in the year would seem to provide the same protection as protecting against the loo-year flood. Rather. Naturally. it is necessary to evaluate the consequences for each season. Some cautions need to be applied in this case because the resulting probabilities are no longer “annual” probabilities. and one should be aware of the effect that these errors could have on the fitted distribution. assume 150 cubic feet per second. that is. In this case. this flow is not the loo-year flood.

This assumption implies that the basic process is also unchanged from year to year. and these risks are implicitly accepted without carefully assessing the odds involved. Peak flow is a very random variable. the flows are determined as the result of other random inputs. As a result. The first assumption is that the process be random in nature. Usually. It should be noted that the assumption the data is error free does not mean that the data sample is truly representative of the population. This is a necessary and reasonable assumption. Annual instantaneous peak flows are assumed to be identically distributed from year to year. such as precipitation and the initial conditions of the basin. it is assumed that the values are statistically independent from year to year. The assumptions of flows being independent. This is obvious because if the process was not random in some way there would be nothing to study. it will be reflected in the same percentage change in all of the frequency curves. Where possible. also random. In everyday life. the interest is in the probabilities of failure. and will only distort the value of the log mean and not the slope or shape of the curve if the log-Pearson Type III distribution is used. identically distributed random variables are basic to the frequency anal- yses discussed in this manual. but have some random behavior. should be removed from the data series to make it homogeneous. that the outcomes are not entirely predictable. Even if flows are considered to result from a deterministic (or semideterministic) process. by definition. To be able to calculate risks. Sampling variation is. not to concepts and philosophies is desirable. The distribution of possible values is not altered by previous values.-The basic premise behind all frequency analyses is that the process is in some way governed by laws of chance.STATISTICS AND PROBABILITIES floods and usually of little concern. These assumptions are easily justified for 189 . although adjustments may be necessary in the case of changing basin conditions.3 Frequency Distributions An understanding of the statistical along with the knowledge of the used with flood data. however. If the error is a constant percentage error. however. The intent the more important ideas. known changes in the basin response. more easily quantifiable using statistics. (a) Concepts and Philosophy. there are risks involved in practically all activities. traditional approaches that have been of this section is to emphasize some of provide a complete discussion. Probability and statistics. 7. A more subtle but potentially serious source of error is when the stage discharge relationship for a stream becomes less accurate with increasing flow. certain basic assumptions must be made. the systematic error may distort the frequency curve. are devoted to the study and calculation of the probabilities of incurring these risks. and that these values are not dependent upon one another. flows are. that is. Also. such as the effects of an upstream reservoir. however. Since the prime interest is in the less frequent rare flood. the errors involved may be highly compounded due to the extrapolation of the frequency curve.



instantaneous peak with long duration

flows but need careful examination when working volume values such as seasonal or annual runoff. that deals with the calculation the basic laws of chance. Statistics, known, and these processes must frequency analysis is, in fact, a have been inferred from the data the conclusions may be drawn. while probability does not re-

Probability is the branch of mathematics of risks based on known processes using however, involves processes that are not be inferred from observations. Flood statistical analysis. Once the properties and the process assumed to be known, Statistics relies heavily upon probability, quire statistics.

Basic to all flood frequency analyses is the concept that there is a population of potential values that may become flood values. This population includes all of the values that could take place, and is the set of all possible results. For flood flows, this population is made up of continuous values, generally considered to be bounded on the low end with zero flow and unbounded at the upper end. These values follow a frequency distribution. Both past and future events are assumed to come from this distribution and that the process (distribution) is unchanging with time (stationary). The order of the events is random, therefore the events are independent. The record of flood flows constitutes the data sample, which is a sample of the flood flow population. The sample is frequently called the “sample population,” while the total flood flow population is called the “target population.” The sample is made up of the actual flows that have been experienced and recorded in some manner. While drainage basin changes such as urbanization, agricultural use, or construction of water control facilities may affect the homogeneity of the record, this is usually not the case and the sample is assumed to be homogeneous. More importantly, the sample is assumed to be a representative sample of the population. From this data sample, the behavior of the entire population is inferred. The choice of the basic form of distrubtion of flood flows is based on the general behavior of numerous data samples from many streams and rivers. In a strict sense, the form of the distribution is also an assumption. Only the parameters of the assumed distribution are inferred from the sample data. The reliability with which the population can be inferred from the sample data is a function of how representative the data sample of years of record, is indicative is. The size of the sample; i.e., number of the reliance that can be placed on the sample. The larger the sample, the less the chance will be that the sample is unrepresentative. However, a larger sample is not automatically more representative than a smaller sample, but is more likely to be so. From the previous discussions, it has been shown that the flood approach relies on several fundamental assumptions:




laws of chance




Process does not change with time Sample data are representative Population characteristics can be inferred from sample identically is




Sample data are actually results that are independent, distributed random variables A general form involved or framework for the probability



This form or framework could be the commonly known “normal distribution” or, more typically for floods, the “lognormal distribution” or “log-Pearson Type III distribution.” These assumptions, while not always acknowledged, form the basis for the actual flood frequency analysis. Generally, the analysis consists only of inferring the characteristics of the population based on the sample data. By inferring the characteristics of the population, what is generally meant is the estimation of the parameters of the probability distribution. A final step in the analysis, which is of considerable importance and should not be neglected, is the assessment of the uncertainties. This usually takes the form of an evaluation of the confidence in the results. (b) Forms of Presentation.-The results of a frequency analysis are in the form of a relationship between the values and the probabilities associated with those values. Results may be in the form of charts, tables, or more commonly as curves. The results are usually in terms of the probability of an event taking place (discrete probability), probability of an event exceeding a certain value (exceedance probability), or the probability of an event not exceeding a certain value. For floods, the most common form of presentation is the exceedance probability. The terms frequency and probability are not the same because frequency has to do with the counting of occurrences of some event within some stated time period, and probability is a statement concerning the chance of an event taking place during some specified period. Note that a frequency can exceed one, while a probability must be between zero (impossible to occur) and one (definitely to occur). As events become rare, frequency and probability are nearly the same, although it can be shown that the frequency will always exceed the probability no matter how rare the event. In flood hydrology, frequency is generally used synonymously with probability with the understanding that the meaning intended is, almost without exception, that associated with probability. Terms such as “flood frequency” and “frequency distribution” are used, but the actual intent is almost always “flood probability” and “ probability distribution.” 191



There are two main forms of curves used to present flood frequency distributions: (1) probability density function and (2) cumulative distribution function. The probability density function curve has the familiar “bell” or “hat” shaped look. The ordinate scale has units of probability density, probability per unit chnage in the variable, while the abscissa is in terms of the variable; the axes are usually Cartesian. This form of presentation is shown on figure 7-l. Integration of the area under this curve between two values of the variable yields the probability of having an outcome with a value between these two values. Naturally, the total area under the curve is equal to one. For many variables in our day to day environment, the distribution is nearly normal and symmetrical. For floods, the distribution is skewed to the right. The logarithms of the flood flow values tend to be more nearly symmetrical, and the log of the values are closely approximated to the normal distribution. Examples of skewed probability density function curves are shown on figure 7-2. The probability density function is difficult to construct, and while it conveys a good deal of knowledge about the more common events, it does not give much information about the tails of the distribution that are indicative of rare events. Since it is the tails of the distribution that are of


Figure 7-l.-Probability

density function

curve. 103-D-1933.




Figure 7-2.-Examples curves. 103-D-1934.

of skewed probability





most interest in flood hydrology, often used.

the probability

density function

is not

The cumulative distribution function is one that relates the exceedance probability to the value of the variable. In fact, this function is the integration of the area under the probability density function from the value in question to the largest value possible. The exceedance probability form is the most commonly used form for presentation of flood frequency distributions. In most statistics texts, however, the cumulative distribution function is defined in terms of nonexceedance; that is, the function gives the probability of an outcome less than or equal to the given value. It is easy to keep the two forms separate because each form is simply one minus the other. A typical cumulative distribution function is shown on figure 7-3. Usually, the abscissa represents the exceedance probability and the ordinate represents the variable value. However, the two axes are frequently switched. This form of presentation, which is fairly easy to construct, does not convey the meaning concerning the more common events as well as the probability density function, but it does better for the more extreme events. This form of presentation is ideal for flood hydrology because the concern is with the more rare events. As shown on figure 7-3, values for the cumulative distribution function range from zero to one. The graph paper used to present this curve is often constructed with a highly distorted axis to arrive at a curve that is as near to a straight line as possible. Usually, the abscissa is transformed from Cartesian to a normal probability scale that is constructed such that a normal probability distribution would plot as a straight line. This is accomplished by creating a Cartesian scale in terms of standard deviations of the normal distribution and then labeling the scale in terms of the corresponding exceedance probabilities. In addition to the distortion in the abscissa, the ordinate axis is also often “distorted” by using a logarithmic scale in this direction. When both scales are distorted, lognormal paper is created and a lognormal distribution would plot as



PROBABILITY distribution function curve.


Figure 7-3.-A typical cumulative 103-D-1935.




a straight line. Lognormal paper is very useful in that it facilitates the drawing of frequency curves and, for most flood data, the curves are more nearly straight lines. There is a considerable amount of distortion in this type of graph paper and the hydrologist should resist the false security that it may convey. In some cases, other types of graph paper are used for plotting. The probability scale is always distorted and, in some cases, both scales are distorted so that a certain probability distribution will plot as a straight line. In this manual, only lognormal paper will be used. Flood probabilities will generally be stated not as probabilities at all, but rather as the recurrence interval (waiting time). This interval is the expected (average) time between the occurrence of values exceeding a given level. It can be shown that this interval is equal to one divided by the exceedance probability for that level. The value of this interval is expressed in terms of time. In hydrology, terms such as 25, 50-, and lOOyear floods are used, which refer to floods having annual exceedance probabilities of 0.04, 0.02 and 0.01, respectively. False meanings are often attached to floods that are referred to in terms of a waiting time, such as the loo-year flood. For example, a loo-year flood may be exceeded several times in 100 years, or it may not be exceeded at all during that time. The waiting time should not be confused by saying that once a loo-year flood has taken place, another loo-year flood cannot take place until 100 years have passed. Every year has an equal chance of having a loo-year flood, that chance being a probability of 0.01. If a loo-year flood occurred this year, the probability of another loo-year flood occurring next year is still the same 0.01. To rephrase, the chances of a loo-year flood are not altered by the flooding or nonflooding in other years.

(c) Probability Distributions.-Many probability distributions are in common use; however, only a few of the distributions that are used in flood hydrology will be discussed. The discussion is oriented towards the concepts and elementary theory that support the use of these distributions. The applicability of the various distributions to flood hydrology will be discussed briefly.
In virtually all approaches used for flood frequency, the basic form or family of distributions is chosen prior to the analysis of the actual data. Once the basic form has been selected, the parameters of the distribution are estimated based on the sample data. An exception to this would be the use of a nonparametric approach, although this approach is rarely used. The choice for the basic form of the distribution can be based on the natural bounds assumed for the flood process, the theoretical considerations based on the assumed process and the behavior of statistics, and the statistical behavior that has been observed in data from other sites. Zero flow purposes, is an obvious example a lower bound equal of a lower bound for floods. to some type of pseudo-base For practical flow may be


it is logical to assume the best choice for a flood distribution would also have the ability to preserve an upper bound. Other distributions can be formed as direct applications of the lognormal distribution to nonlinear transformations of the variable. If the factors are not additive but rather are multiplicative. A distribution that preserves a lower bound would seem to be indicated. The same situation would also apply to the upper bound of the PMF. this distribution uses the coefficients of skew of the logarithms. there are not enough data to extend frequency curves to anywhere near this limit. and whether the selected distribution preserves such a bound is of little concern. the lognormal distribution can be argued to apply. In many natural processes. Statistical theory demonstrates conclusively that a random variable that is calculated as the sum of a number of other random variables that are not totally dependent upon one another will tend to be distributed according to the normal probability distribution. however. In actual practice. Another distribution that is a slight variation from the lognormal is the log-Pearson Type III. In fact. The constant is a lower bound that is preserved by the distribution. Also. if the PMF is considered to be an absolute upper bound. The case for multiplicative factors can be argued for many types of flood flows because most flood data seem to behave in a lognormal manner. For a zero skew. the population mean and standard deviation. It should be noted that using this distribution preserves a zero flow lower bound. the influence of the factors is in fact additive in terms of their logarithms. For floods. the distribution is exactly the lognormal 195 . and the logarithms of the variable are normally distributed. where the logarithms of the values minus a constant are found to be normally distributed. This conclusion governs many natural processes. the normal distribution is an obvious choice for cases where it can be argued that the influencing factors are several. a three-parameter lognormal distribution is used. In addition to the mean and standard deviation of the logarithms. In some cases. each of which has an additive effect on the variable. Since the effects of this physical limit would probably not be felt until floods were quite near the limit. the random variable is the sum of several other variables or is the result of the influence of several factors. it is doubtful that anything would be gained through the use of distributions that explicitly recognize this limit. The two parameters are the population values of the mean of the logarithms and the standard deviation of the logarithms. floods near zero or the lower bound are rarely of any interest. data do not exist to either verify or deny the existence of such a bound. and are additive. are all of comparable weight or influence. are somewhat independent. The three parameters are the constant and the mean and standard deviation of the logarithms of the flows reduced by the constant. In this case. The normal distribution has only two parameters.STATISTICS AND PROBABILITIES reasonable to use.

two of which will be discussed here. the parameters of the distribution are actually population moments. This methodology for parameter estimation has the advantage of being straightforward and simple. This form of distribution has been found to fit a wide variety of flood flow data. has gained some popularity for use in flood hydrology.. that has five parameters. variance. or a straight line. while it does fit some data fairly well. the parameter estimation is simply the use of the sample moments as estimates of the parameters (i. The “Gumbel” or “extreme-value” distribution. The “Method of Moments” [68] is probably the most commonly used parameter estimation method. as such. In general. When the coefficient of skew is positive. This distribution is actually one of only three limiting extreme value distributions and. By using sample moments as estimates of the population moments and by using the indicated functions. Usually. In some cases. In this method. when plotted on lognormal paper. the parameter estimates are related to the moments of the sample data. population moments). In some cases. This distribution combines all three into one. the parameters are estimated as functions of the sample moments. The first three sample moments used in hydrology are the mean. and skew. For this normal distribution.e. has some theoretical justification for being used. the curve is concave downward. it has been found that the upper or lower tail could be fit by a distribution but that both tails could not be fit at the same time. and when the coefficient of skew is negative. it is concave upward. This form of distribution can easily be supported as a modification of the lognormal. (d) Parameter Estimation. The mathematics involved in defunctions are riving the functions is relatively basic. and the resulting 196 . the mean and variance. The tails of this distribution behave much like the tails of the normal distribution. the relationship is calculated from the mathematics of the basic distribution. has gained attention as a distribution that could handle this situation. this distribution is not widely used. the assumptions and conditions required for the extreme value distribution are not satisfied with flood data and. A distribution known as the “Wakeby”. such as the normal distribution.FLOOD HYDROLOGY MANUAL distribution. in this case.-Several methods of parameter estimation are available. For other distributions. the curve is no longer a straight line. also known as the “double exponential” distribution. A more flexible distribution distribution” has also been extreme value distributions known as the “generalized extreme value proposed. where the assumptions for that distribution are not exactly met. the parameters can be shown to be related to some function of the population moments. The additional parameters allow for the increased flexibility needed to fit two seemingly incompatible tails.

The mean is the first moment and is an indicator of central tendency or a location parameter. a complete section will be devoted to it. when presented in the probability density form. Karl Pearson [65] developed a series of distributions that have the ability to fit many different shapes of observed sample frequency distributions. often stated in terms of the kurtosis. slope. the partial derivatives are taken with respect to each of the parameters and then these derivatives are set equal to zero. with n unknowns. These equations are then solved for the parameter estimates. hold some inherent meaning for the practitioner. This section 197 . and is usually stated in terms of the coefficient of skewness. The application of LP-III is covered in detail in Bulletin 17B [ 141. The “Method of Maximum Likelihood” for parameter estimation is very appealing in that it appears to have a slight statistical advantage over the Method of Moments in the estimation of parameters for several types of distributions. The fourth moment. In addition. 7. the higher the likelihood. and shape of the plotted frequency curve. it is easily adaptable to the use of historic and censored data. This results in n equations. and has been found to be very adaptable to floods and even more so to the logarithms of floods. The third moment is the expected value of the cubes of the deviations from the mean. To maximize the function with regard to the choice of the parameters.STATISTICS AND PROBABILITIES generally elementary. can be interpreted in terms of either the peakedness of the distribution or the heaviness of the tails of the distribution. Initially. the number of parameters. The logPearson Type III distribution is an application of the Pearson Type III distribution applied to the logarithms of the data. the formulation procedure and application are more complex. Unfortunately. usually has a bell shape or. with some parameters. hereafter called LP-III. the sample moments are easily calculated and. it is characterized by being noticeably skewed to the right. Regardless of the shape of the curve. one for each partial derivative that has been set to zero. Unlike the Method of Moments. This moment relates to the symmetry of the distribution. a J-shape. The higher the value of the likelihood function. or square of the standard deviation. These four moments have a direct relationship to the location. The method estimates the parameters such that the likelihood of experiencing the observed data is maximized. The probability density function is evaluated for each data value.4 Log-Pearson Type III Distribution Because the log-Pearson Type III distribution is the principal distribution used by the Bureau. This is done in terms of unknown parameters that are still to be estimated. and then all of the probability density values are multiplied together to get the likelihood function. The log-Pearson Type III distribution. and is an indication of the scale or spread of the distribution. in themselves. this method involves evaluation of the likelihood of the sample data being observed. and the likelihood is usually expressed as a product. The variance is the second moment.

(b) Form.FLOOD HYDROLOGY MANUAL will not attempt to reproduce supplemental information. Bulletin 15 was expanded and reissued as Bulletin 17.S.13. and is fully supported by the Bureau of Reclamation.-The LP-III distribution is a modiftcation of the Pearson Type III distribution. The actual printing and distribution of this bulletin are currently handled by the Office of Water Data Collection of the USGS. function The mode is the value written in a cumulative is easier to use: The distribution can also be form that. Considerable study and debate was expended prior to the selection of LP-III as the basic distribution to be used throughout the United States.14]. but will provide (a) Adoption by Federal Agencies. no better approach has been found and proven for general usage. however. most likely distribution to occur. However. most recently in 1981. A Unzjbrm Technique for Determining Flood Flow Frequencies” [l l] in 1967.-In an attempt to achieve consistency in the Federal sector with regard to flood frequency studies. 17A. Water Resources Council issued Bulletin 15. and x = deviations c = parameters 1 + T -cx'a a I and (1) of the variable from the mode. This bulletin was created as a basis to standardize flood frequency procedures within all Federal agencies. Later. This effort came under much scrutiny and attack from both inside and outside the Federal sector. where the logarithms of the flows are taken and then fit to the Pearson Type III distribution. in most instances. log Q = M + KS (2) 198 . This bulletin is now issued jointly by the various agencies under the auspices of the Interagency Advisory Committee on Water Data because the Water Resource Council no longer is in existence. the procedures described in this manual are sufficient and reliable. estimated from sample data. the U. a. This was a pioneering effort that focused all flood frequency research for several decades. all of that information. and. The primary concern was in having a distribution that would be appropriate for use with data resulting from all types and causes of flooding. It should be noted that Bulletin 17B allows other approaches to be used when justified. The LP-III distribution has the following probability density form: f@) = p ( where: p. in actual practice. as 17B [ 12.

There is negligible theoretic support and experience to justify the use of LP-III distributions that have high positive skew. the use of values over 1.5. it will be curved in a concave upward direction. the cause may only be due to an unrepresentative sample. If the coefficient of skew is zero. Often. Generally. the flood data will have skew coefficients well below 1. can also be a reason for a high positive skew. Other reasons for a high positive skew include upstream flood control regulation. however. Values outside this range should be examined carefully and only used after thorough justification.4 to +0. if negative. The values for K in equation (2) are provided in Bulletin 17B[ 141 in an extensive table for a large range of combinations of probability and skew. Normally. and trends in basin response. large amounts of upstream natural storage with constricted or regulated release. The effects of two or more independent causes for flooding. it will curve concave downward. it is recommended that the independent causes be separated and the population treated as a mixed population. A high negative skew is also subject to suspect.0. If this is the case. such as different storm types.0. (c) Shape. A high negative skew may be justified when the assumptions that tend to support a normal distribution are closer to being true than the assumptions for a lognormal distribution. even with recorded periods over 100 years. which were called the Monte Carlo experiments. these effects need to be accounted for and the data adjusted to arrive at frequency values that can be used with confidence. as discussed later in this chapter. Again.STATISTICS AND PROBABILITIES where: ow rate in cubic feet per second. A reasonable range of skew coefficients. Although Bulletin 17B [14] gives values of K for coefficients of skew up to 9. For example. K = a factor that is unique for each pair of values of exceedance probability and skew coefficient of the logarithms of the flows. Statistical analyses of randomly generated data samples. demonstrated a high degree of unreliability in the estimation of the coefficient of skew. and S = population standard deviation of the logarithms of the flows. a high sample skew value is the result of an unrepresentative sample.-The LP-III distribution is usually plotted on lognormal paper. the normal distribution 199 .0 should be highly justified. if the floods being analyzed are caused by many factors that are additive in their effects. the high positive skew is indeed proper for the population. A high positive skew may be reason to subject the frequency analysis to further study. the distribution is lognormal and will plot as a straight line. as computed by procedures described in Bulletin 17B. is from -0. $=fl = population mean of the logarithms of the flows. If the skew is positive.

there is an upper bound. there is no bound. However. LP-III has been found to compete favorably with other distributions. (d) Advantages. if the absolute value of the coefficient of skew exceeds 1. Logical support and reasoning must be used and included in the flood study report to support the use of a highly skewed distribution. For zero skew. it would probably be best to plot the sample data on both lognormal and normal probability paper. This distribution also usually has a boundary set on the logarithm of the flow. which can be defended from a theoretical basis. or if resulting distribution is to be extrapolated beyond a probability of one divided by twice the length of record. the loo-year flood. if the LP-III is still used. In this case. and for positive skew. this is seldom found in the data due to the rare nature of any upper bound and the likelihood that boundary effects are only felt near the boundary. By using a coefficient of skew other than zero.0. for example. The distribution can also be easily applied without a computer program. the analyst should explicitly recognize the bound and justify the use of the bounded distribution. The estimation procedure and method of moments are both well accepted and are also easily understood. Unusual sample data alone cannot support acceptance of a frequency curve that defies common sense and logic. for negative skew. the LP-III distribution has several more advantages.FLOOD HYDROLOGY MANUAL may be expected. In this case. As with any distribution that is adopted as a base. the sample data may still fit reasonably well but a negative skew should be expected. Several computer programs are available to aid in the application to most situations. there is a lower bound. the distribution can be made to fit most data in a reasonably satisfying manner. However. It is also a 200 . the flows cannot be below zero. and the fact that logarithms are used automatically sets a boundary. However. Normally distributed data will plot on lognormal paper as negatively skewed. it probably would be justifiable to deviate from normal procedure and use normal distribution. It should be noted that of K = -2 divided these bounds are not ordinarily of much influence or concern unless the skew values are highly negative or highly positive. It should be noted that the LP-III distribution is usually a bounded distribution.-In addition to the advantage of offering consistency to Federal agencies. The bounds for both positive and negative skews can easily be evaluated because they correspond to a value by the coefficient of skew. Obviously. All evidence suggests that upper bounds must be much larger than. The LP-III distribution is almost the same as the lognormal distribution. Another theoretical justification for a negative skew would be the existence of an upper bound that would be reflected in the sample data as a tendency for the frequency curve to bend over as it approaches the bound. In comprehensive studies. LP-III offers a reproducible analytic method in which the parameters involved are easily understood.

then an additional error has been introduced. in reality. because of its ability to fit data from many parent distributions. If the true population distribution is not LP-III. masks this behavior. A major disadvantage of LP-III is the fact that it is highly dependent on the value of the adopted coefficient of skew. In addition. and that LP-III. the results are still only an estimate of the population distribution and are prone to errors due to errors in basic data.STATISTICS AND PROBABILITIES versatile distribution that can closely fit data that varies from normal to lognormal to even heavier-tailed distributions. basin size has only negligible correlation with standard deviation. and coefficient of skew are estimated from the sample data. however. Sample estimates of this coefficient have been shown to be highly unreliable.The current fitting procedure used in Bulletin 17B is the “method of moments” estimation. In this manner. (e) Disadvantages. This value differs from basin to basin and from one geographic area to another. random sample variation. -Although the LPI11 distribution has the advantage of closely fitting the data. The mean value is indicative of the basin size and the basic runoff regime. The coefficient of skew is assumed to be 201 .. Confidence bands are a highly recommended approach for this type of evaluation and should be presented in the flood study report. may actually be a more technically efficient approach. Statistics does offer some help in evaluating the reliability of the results. with the choice dictated by the circumstances of the application. and the effect of an unrepresentative sample including the effects of outliers. which means that the form of the distribution would restrict the results to be consistent with good overall judgment and to be consistent with the statistical behavior that has been found with previous data. One objective in selecting the basic distribution is that the behavior of the results are constrained. The mean. that may have come from many different distributions. standard deviation. As with any fitted distribution. this may also be a disadvantage. This method uses the logarithms of the flows and. The LP-III distribution offers little in the way of beneficial constraint. the results are subject to very real errors with respect to the validity of the underlying assumptions that have been made. is a fitting of the LP-III distribution to the logarithms of the flows. The use of other more restrictive distributions. I’$)Fitting Procedure. all in terms of logarithms. This problem has been recognized in Bulletin 17B and mitigating procedures are recommended. It can be argued that flood discharge values should not be expected to follow only one type of distribution. This bulletin encourages the use of a generalized skew coefficient and using a weighted average of the generalized and station (sample) values of the skew coefficient. the choice of distribution would tend to compensate for the effects of unrepresentative samples and sampling variation.

The use of maps that relate coefficient of skew to location are an indication of the philosophy that the values for this parameter should be expected to vary in a rather gradual manner with location. These maps only reflect the influence of locations on the values of skew. ensure that prediction equation is indeed statistically significant. and that the equation is meaningful and consistent with hydrologic common sense. is a weighted average of the skew estimates from several basins in a hydrologically and meteorologically similar area. slope. There are three approaches in Bulletin 17B for the estimation of a generalized skew. this assumption is not easily shown using actual data. These iso-skew line maps are based on the analysis of many individual stations. and be mostly independent of the basin characteristics. cover. (g) Generalized Skew. the term “skew” is used when the term “coefficient of skew” is the one intended.FLOOD HYDROLOGY MANUAL reasonably stable. it is very difficult to support or deny this philosophy with actual experience. and the one most often used by the Bureau. vary in a gradual manner with primarily location and secondarily with other factors. These estimated values are then related to the skew values by a fitted prediction equation. which do depend on drainage area. by using this type of smoothed value. because of the extreme sampling variation experienced with this parameter. and other factors are negligible. In this case. Factors believed to be potentially important would also be estimated for each area. Generalized skew is a skew value that reflects the basic philosophy that the true population values for skew are relatively stable and. A second approach used for estimating a generalized skew is a prediction equation approach. It should be noted that the skew maps in Bulletin 1’7B were developed without taking into account any effects of mixed population. care must be taken to avoid spurious results. elevation. The most commonly used weight is the number of years of record. skew values are estimated for many basins in the same hydrologically and meteorologically homogeneous area. but care must be taken not to cover up the effects of mixed populations. If this approach is used. One approach is the use of maps that have contour lines showing equal values of skew. Again. Users may develop equivalent maps for the area around the basin of interest. This approach seems to produce somewhat reasonable results. while not necessarily constant. This same nomenclature is used in this chapter. drainage density. and reflect an assumption that basin size.-Throughout much of the literature. Weights could also be based on 202 . however. The Bureau’s general experience with this approach is that the relationships are weak and prone to producing questionable results. A third approach. The reliability of the map values is estimated as a constant throughout the United States based on how well the contour lines match with the data points for the individual stations.

however.STATISTICS AND PROBABILITIES the statistical reliability of the individual estimates. It is recommended that the weighted average of the station skew and the generalized skew be calculated. The outlier test is based on the assumption that the data fit the lognormal distribution and the sample mean and standard deviation are used. the effects on the upper portion of the frequency curve should dictate the elimination. This adjustment can be justified when evaluating flood damages or performing other economic evaluations. but this is nearly the same as using the number of years of record. (h) Outlier Criteria -Bulletin 17B includes a procedure for the identification of outliers. or retention of them. Low outliers are generally not of concern in a flood frequency analysis. (i) Adjustments. It is this weighted average value of skew that is to be adopted as the parameter estimate. Practitioners should be cautious when throwing out data points because elimination of high values from the data set will result in a general downward bias in flood frequency results. however. It is important to note that this procedure does not call for the automatic elimination of these data. Bureau practice is that high outliers be treated as historical data rather than being eliminated. These procedures are used in Bureau flood frequency studies with the following two exceptions: Exception 1. -Bulletin 17B includes many adjustment procedures for such conditions as zero flows and incomplete records common in the Western States. The test is performed at the 1O-percent level (one sided). These procedures are not used for the mean and standard deviation parameters because their estimates are much more stable. Since economic analyses for flood control are not usually performed by the Bureau. and that some adjustment may be necessary. The procedure should definitely not be taken to mean that the data is in error. modification. this adjustment is not applied. The weights are based on the estimated reliability of the two estimates. -The historic data adjustment has been conclusively shown not to perform well with paleohydrology data or other data 203 . Exception 2.-The expected probability adjustment is generally not used in Bureau flood frequency analyses. The regionalization inherent in the generalized skew and the averaging of the generalized skew with the sample estimate reflects a general lack of confidence in the sample skew estimates. It should also be noted that the presence of a high outlier may well indicate that the data are comprised of a mixed population. it does call the analyst’s attention to the need for further study and the possibility that the outliers may not be representative.

but as events are estimated that have return periods comparable to or longer than the sample length. The reliability of the frequency curve gradually decreases for the more extreme events. This of course is not possible because sufficient data do not exist to verify the choice of the base distribution. The errors that are unavoidable in the parameter estimates become intolerable once the frequency curve is extrapolated beyond a certain point.or 300-year return interval when the basis for choosing the distribution are records rarely longer than 60 years.11 and 7. it is difficult to defend the tail behavior of a distribution for a 200.6 Limitations on Frequency Curve Extrapolation An ultimate goal would be to arrive at a frequency curve that is valid over the entire range of possible flood flows. it is recommended that this adjustment not be used. For more common events. indicates that frequency curves are reasonably reliable out to return periods of about the sample record length. Practical rule-of-thumb knowledge. In addition to the deterioration in the quality of estimates due to sampling errors. the rate of deterioration increases.5 Example Application Using Bulletin 17B It is not practical to reproduce the entire contents of Bulletin 17B in this manual even though any example would be incomplete without the necessary tables and other material contained in this bulletin. The sample length is the main controlling factor concerning the reliability of the frequency curve. brief sample problems are shown in sections 7. which is supported by statistical calculations. As the exceedance frequency becomes smaller. For example. If choices concerning human life and dam safety are involved.FLOOD HYDROLOGY MANUAL that represent a much longer time period than that of the systematic record. it initially deteriorates slowly and takes place in a gradual manner. The sample data is only sufficient to provide estimates for the distribution’s parameters. The current Bureau practice is to limit the extrapolation 204 . 7. This reliability deterioration does not take place at a particular point. more caution should be taken about using extrapolated frequency curves than if the use was for determining the size of a cross drainage structure where only a short term loss of service is the main consequence of failure. 7. the reliability becomes poor. It should be noted that the use for which the curve is to be put is also a consideration. the reliability of the basic choice of distribution becomes questionable.12 to indicate how a mixed population data set is analyzed and how to compute confidence limits. For these cases. the reliability of the frequency curve is fairly good. However. or even twice the sample length.

200 years or four times the length of record. whichever is longer. However. intensity. (a) Causative Meteorological Factors. the data are assumed to come from one basic distribution. the curve is negatively skewed and. Confidence bands are used as an excellent approach to quantifying the uncertainty in the frequency curve estimates. Due to the nature of this calculation.STATISTICS AND PROBABILITIES of the curves to twice the length of record. The lognormal distribution is a reasonably typical distribution for general storm-produced floods. the sample length is used whenever the gauged data is at the site and also for data transposed from another site. it is generally considered that improved results are obtained with an analysis that specifically treats the mixed population. In cases where catastrophic loss. an assumed length of record of 20 years is used.-The population is the existence of Different storms are typically of a snowpack (with resulting intensity and long duration. In these cases. the population follows a distribution that is a combination of two or more base distributions. however. confidence bands are constructed at the 5. The typical frequency curve for snowmelt runoff is usually very flat. The length of record is needed to calculate the confidence bands. In most locations in the mountainous 205 . whichever is longer. 7. that is. or 100 years.and 95-percent levels. is more nearly normally distributed rather than lognormally. the typical or average record length is used. When precipitation frequency data are used with rainfall-runoff modeling to generate flood data. further extrapolations can be used as justified on a caseby-case basis. or dam safety are not involved. This is especially true in the Western States where data reflecting runoff from snowmelt and rainfall are quite common. it is characterized by a low standard deviation of the logarithms. and hurricanes that usual cause of a mixed flooding due to different types of storms. it is frequently found that the floods are the result of distinctly different causative factors. the analysis is conducted as an expediency without taking this into account. This phenomenon was recognized by Allen Hazen in his 1930 publication. Within the Bureau. and are generally extrapolated twice as far as the frequency extrapolation curve. Flood Flows [64]. Usually. while the existence of a mixed population is recognized. When using regional approaches. snowstorms that lead to the accumulation snowmelt flooding). The frequency curve for floods resulting from general type storms is steeper (has a relatively higher standard deviation) than the snowmelt curve. that is. Often. with large basins. general storms of low local storms of short duration and high produce intense rainfall over large areas.7 Mixed Populations In most flood frequency analyses. loss of life.

The data resulting from each cause is then analyzed separately with the results later combined to form a single curve.I FLOOD HYDROLOGY MANUAL regions of the West. Either type of flooding may tend to overshadow the other for the entire set of data. most years will not even have a flood of this type. floods resulting from local storms are closely approximated by the lognormal distribution. These hurricane type floods are much rarer than the others. Hydrologic differences may help explain the existence of a slightly positive skewed frequency curve in the absence of mixed meteorlogical causes. (c) Separation of Data. and it is best to develop a separate annual series for each cause. As with floods resulting from general storms. An example of this would be a rainon-snow flood. The situation is more complex when different meteorological causes tend to blend together. the relative intersection of the curves also changes. As the basin location changes. as a result.-Although a mixed population condition is probably due to different meteorological factors. The probability of exceeding any given level of flooding is the sum of all of the individual exceedance probabilities minus a correction for the possibilities of having more than one of the causes producing a flood exceeding this level. The hurricane frequency curve is the steepest of the curves discussed. with the differences possibly related to time of year. the data must be segregated by meteorological cause. These factors might include differences in infiltration. Often. the time of year is taken as an expedient method to achieve almost the same results. and may tend to be unidentifiable. a hydrologic cause for mixed population behavior is not nearly as clear. The intersection tends to move towards rarer floods as the basin elevation increases. The frequency curve for floods resulting from local type storms is much steeper than the general storm curve and. local and general storms are distinctly different as are their flood frequency curves. (d) Combined Frequency Curves. rather than identifying the meteorological cause of each flood. (b) Types of Mixed Populations. Usually. reflected by a typically large standard deviation. channel roughness.-To analyze a mixed population.-The final or combined frequency curve is a simple combination of the individual frequency curves using basic statistical principles. and antecedent conditions. the general storm-produced flooding will dominate the rarer floods while the more common floods are predominately snowmelt. The magnitude of the flooding is highly variable. local storms tend to dominate a less frequent portion of the frequency curve. 206 . a mixture could also be the result of distinctly different hydrological factors. However. cover. with a single flood being the combined result of two or more causes.

Pa PC . Some of this caution is due to the fact that the 207 . where the mixture is somewhere between snowmelt floods and rainfloods.Pu( Q > x ) Pb( Q > x). Criteria provided in Bulletin 17B [ 141 are intended for peak analysis and must be used with caution when applied to volume data. The resulting total curve is typical of what might be seen in the actual data (not separated) for the site. and such a flood may be so rare that the recorded data is entirely dominated by a more common type of flooding. the rare type of flooding may become dominate. if not more critical. for this case.Pa Pb where: ’ Pt = total curve value. This behavior is typical. and Pa and Pb = individual exceedance probabilities.STATISTICS AND PROBABILITIES In the case of two causes.not be important except for rather rare floods. In this situation.Pb PC + Pa Pb PC (4) Figure 7-4 illustrates the case with two casual factors. a mixed population analysis is even more desirable. In the operation of a dam. than the peak flow. are straight lines on lognormal paper with a rather sharp break. the equation is: Pt = Pa + Pb + PC . the same approach may be repeated. 7. . a complete hydrograph is often needed. This situation has been shown to be common in the higher elevations of the foothills in Colorado. the equation is: Pt( Q > x ) =>Pu( Q > x ) + Pb( Q > X) . Equations are easily derived for any number of casual factors. this joint probability is based on the assumption that the two types of flooding are independent. Frequently.8 Volume Analysis Although most flood frequency analyses are performed using terms of instantaneous peak flows.Pa Pb . The probability of having both floods in the same year is simply equal to the product of the two individual probabilities. the probability of having a flood exceed a given level is equal to the sum of the probability of the first flood type and the probability of the second flood type. After extrapolation. or Pt = Pa + Pb . Note that the data points and confidence bands are not shown. only the individual curves and the resulting total curve. The total curve shows two distinct segments that. the volume of water is sometimes as critical. For two causes. minus the probability of having both types of floods in the same year. Naturally. one casual factor may . For three causes.When more than two causes are present.

Colorado. . North Fork of Big Thompson River near Drake.000 100 Y /’ / \ \I I b-Rein I I I I j! Nood I IO 50 40 PROBAI$ILITY OF BEING 20 IO EQUALED 4 2 I 0.5 0.05 GIVEN 0.061 (PERCENT) 0. 103-D-1936.20.K I.0001 OR EXCEEDED IN ANY Figure 7-4.-Mixed population frequency curve.01 YEAR 0.10.

especially if the values of skew 209 . the most critical water level will be very close to the most critical water level determined by using separate hydrographs that have been fit to each of the different unit time volume values. In this case. (a) Balanced Flood Hydrograph. The storm types that define the frequency curve may even vary as the unit time changes. In effect. To construct a balanced hydrograph. however. For example. the desire is to afford a certain degree of protection. While this procedure is not conceptually pleasing. approach would be to fit a single hydrograph to all of the volume data for the various unit times. Obviously. then the peak daily value.” This approach results in only one hydrograph to be tested. . This approach is known as the “balanced hydrograph technique. In the construction of’a balanced hydrograph. It is not uncommon to have inconsistencies between the volume frequency values for the different unit times. In this manner. and most used. the loo-year volumes are usually extrapolated values. to 15 days. By routing such a hydrograph through a dam. it is not uncommon. this is a physical impossibility. be aware that different flood types may actually be mixed within the hydrograph. The instantaneous peak is initially fit. to 3 days. a hydrograph is created that provides a constant level of probability regardless of the unit time selected. Generally.STATISTICS AND PROBABILITIES parent or population distribution form is expected to change as the unit time changes from instantaneous to daily. the routing process actually selects the most critical unit time. etc. it would not be expected that the loo-year peak flow and loo-year volume would be experienced in the same flood. A second. One approach would be to evaluate the volumes of floods of all unit durations and to fit reasonably shaped hydrographs to those volumes. it is desirable to provide protection against all durations of floods that have that exceedance probability. it must be remembered that this approach is an expediency and not a model of reality. the balanced hydrograph approach is an expediency that allows for a fairly complete analysis with a fraction of the effort and produces somewhat conservative (high side) results. What is not obvious is the unit time that is most critical for a given return period. Also. etc. Each of these hydrographs would then be tested to determine which was most critical. The storm expected to result in the loo-year peak flow may not be the same storm expected to produce the loo-year. a symmetrical or other reasonable shape is usually selected as a guide for the fitting of the volumes. Using the balanced hydrograph. 7-day volume. followed by the next shortest unit time volume. it may be obvious that volume is very critical. and it is possible to actually estimate frequency curves that have a higher 15-day volume than 30-day volume at the loo-year return period.For a particular flood study. and that may be the 1 in 100 or loo-year level.

(2) regional flood frequency. the peak daily flow value is usually used as being the same as the peak 24-hour value. (a) Transposition of Frequency Data. the use of fixed interval data will result in low estimates. thus the term “fixed interval. and (3) use of synthetic data.FLOOD HYDROLOGY MANUAL vary considerably between unit times. (b) Fixed Interval Analysis. Again. 7. and the site is said to be an “ungauged area.-The transfer of information from one site to another is required for any ungauged area.-One source of potential problems in the construction and use of balanced hydrographs is what is often called “fixed interval analysis. When constructing the balanced hydrograph.5 to 0. the balanced hydrograph has slightly less volume near the peak. A peak 24-hour value for 1 year will always exceed the peak daily value.” For example.8 as the unit time is increased. This type of conflict must be resolved in a reasonable manner. with the power increasing from 0. This value is determined from daily flow values. The most common approach would be to achieve some smoothing of the skew coefficients with respect to unit time. The decision as to whether the fixed interval data is adequate will depend on how critical the volume frequency data is to the study at hand. the site where the frequency analysis is needed is not a site of recorded flow data. A more serious fixed interval problem exists if monthly data are taken to be 30-day volume data. but some awareness is justified. and care should be taken to quantify the magnitude of this problem. in the construction of the balanced hydrograph.” while the peak 24-hour values start at any time of day. or: This square root relationship has been found to be fairly accurate for instantaneous peaks and for the short duration volume frequency values. a higher exponential power of the ratio of the drainage areas may be more applicable. As a result. A basic relationship found to work well is that the ratio of flows at two locations are assumed equal to the ratio of the square root of the drainage areas. Note that peak daily values are constrained to start at midnight. For longer duration volumes of about 60 days or more. the peak l-day volume is used. 210 . This type of problem is usually only found at the longer unit times. and the inconsistency usually amounts to only an insignificant volume. This problem is considered to be of minor concern.” Three basic approaches are used to estimate a frequency curve for such an ungauged site: (1) transposition of frequency data. not peak 24-hour values.9 Probability Relationships for Ungauged Areas Usually.

The depth and scope of these three approaches vary considerably. a relationship is developed that relates the values for a single return period.STATISTICS AND PROBABILITIES For annual runoff. Then. such as the lOOyear values. and cover. for an ungauged area.-There regional flood frequency: (1) average parameter approach. the drainage area (square root) might be used to scale the data. basin elevation. to the basin areas and other basin factors. the standard deviation and coefficient of skew are not altered. this power may approach 1. Also. however. this approach produces results similar to those obtained using the average parameter approach. The coefficient of skew could also be related to other factors (generalized skew) but. The index flood approach usually requires the use of concurrent data from several stations. In effect. It is preferable to transpose information from one site to another on the same stream rather than from one stream to another.O. (2) specific frequency flood approach. are three basic approaches to (b) Regional Flood Frequency. while the standard deviation and skew coefficient do not. the parameter is not really an average but rather a function of basin factors. Lacking this preference. The resulting scaled data are then combined as if all the data came from a single site. the value for the parameter is an average value for that basin’s size and characteristics. Small changes in basin size are also preferable. similar results are obtained from all three. Another more complex and most often not justifiable type of average parameter ap preach is to also relate the standard deviation to basin factors. Generally. the log mean of the flows is plotted against drainage area and a graphical relationship developed. the LP-III frequency curve being transposed is only changed in terms of the mean. The average parameter approach uses data at many similar sites in a homogeneous region to estimate an average value for a parameter. with the data from each station being scaled using the mean annual flood or some similar value. When 211 . and (3) index flood approach. In effect. The most common average parameter approach used is to assume that the mean value parameter (logarithms) varies with the drainage area. Obviously. In many applications. It should be noted that this approach does not take into account any differences in basin factors other than drainage area alone. annual precipitation. The next most common type of parameter approach would be to relate the mean not only to area but also to other factors such as location. In some cases. this refinement often cannot be justified. some attention should be given to the similarities and differences between the sources of floods within the basins. it is preferable to transfer information between basins that are hydrologically similar. The specific frequency flood approach begins by analyzing the flood frequency at many stations in the region.

experience shows that the values for the flood increase with channel length when that factor is examined alone.FLOOD HYDROLOGY MANUAL all of erage area. The model must include the ability to model soil-water retention and loss to produce reliable results. An alternative approach to using actual data would be to make use of stochastic rainfall and. the parameters of the LP-III distribution can be expected to change in a logical manner. and to decrease with increasing channel length if the basin size and all other factors are held constant. These discussions are based on what has typically been found in actual practice. the flood values would be expected to be directly related to the area while inversely related to the length. the model must run continuously. This approach is usually not justified because of the added work involved.-Rainfall runoff models have been used for developing flood frequency curves by using precipitation data. As a single basin parameter changes. stochastic soil conditions. it is not considered to be practical for most studies at this time. with other parameters held constant. If both factors are incorporated into one relationship. since area and channel length are strongly correlated. (c) Frequency Analysis From Synthetic Data. 7. The drainage area is the linear relationship is found strongest between single factor. the model must be run for a considerable time with concurrent inputs and measured flow data to calibrate the model. and the effects of changes in two parameters may be either complimentary or contradictory. it is not practical to find sites where only one basin parameter changes because many parameters change simultaneously. However. this approach is equivalent to the avparameter approach where the mean is related to the drainage The USGS has developed several guides for determining regional frequency estimates that are usually done on a State-by-State basis a State often divided into several regions. but also the complex interactions between them. Also. possibly. Care must be taken to recognize not only the isolated effects of individual basin parameters. Obviously. This relationship might be of the form of a power relationship where the basin area would have a positive power while the length would have a negative power.10 General Trends in Frequency Relationships Experience has shown several identifiable trends in frequency relationships. Some of the more important basin factors will be discussed in this section with respect as to how they typically influence the frequency distribution. the loo-year peak flow would be expected to increase with increasing basin size. To achieve reasonable initial soil moisture conditions at the time of major flood producing storms. flood with the stations are concurrent. and a strong almost the log of the area and the log mean 212 . Since this approach is still being developed. The basin parameters are often interrelated. As an example.

This relationship is true even when the basin size is held constant. not the log mean flow. the log mean usually increases while the log standard deviation and log skew decrease slightly. If the more common flood subpopulation is taken as the base. this is not a strong relationship. soil. and an even larger increase in log skew. slope. The log standard deviation is not strongly related to basin area. snowmelt becomes more important in colder climates.g. Slope is typically closely related to one divided by the square root of the drainage area. to the product of the drainage area and slope.5) of the drainage area. which results in a reduction in the log coefficient of skew. Note that basin slope and channel lengths all change with changes in basin area.11 Foss Dam Example This section continues the example on the drainage basin above Foss Dam in Oklahoma. The log coefficient of skew can be expected to change in an inverse relationship with area. however. If the effects of basin channel slope or channel length are examined. the power would be expected to increase. as unit time increases. a larger increase in log standard deviation. for this reason. climatic effects on the flood type. the distribution can be expected to change slowly from lognormal to normal distribution. the entire population will have a small increase in the log mean. With a change in location. is closely related to the square root of the drainage area. these observations are with typical concurrent changes in these factors. For an increase in the unit time. or to the drainage area divided by the length. With large increases in area. the effects are mainly related to the distance from the moisture sources. and effects on other basin factors.. Sampling variation is high in the case of skew and. 7. and cover all tend to change in a somewhat consistent pattern with a change in location. Basin factors such as elevation. the effect of basin size must be examined simultaneously or spurious relationships may be developed. With volume floods. all three parameters of the LP-III distribution decrease. and many relationships are possible. possibly to a power slightly less than one. relationships are not distinct. As the elevation increases. while channel length and basin channel slope are inversely related. The derivation of the concurrent PMF hydrographs was illustrated in section 4.STATISTICS AND PROBABILITIES of the LP-III distribution. The mean flow. and the routing and combining 213 . this relationship is on the order of the flows being directly related to the square root (a power of 0. Usually. Mixed populations result in an inflated log skew when the data is analyzed as a single population. e.5 of chapter 4. This is a direct result of the mathematics.

for the gauging station’s annual peak discharges. 47 years of record. Since Foss Dam is an existing dam for which modifications are assumed to be required.6528.87558)‘/47 47-l The skew coefficient is: G I = G.6528)5 = -0.33778. it has been determined that. a discharge probability relationship.l)(N-2)(P) (47)‘(194O.5 1 = 0. Bulletin 17B recommends weighting the skew coefficient with the generalized skew coefficient given in the bulletin.222-(156. For the station in this example. = -0. Table 7-l shows the annual peak discharges for each year of record along with other information necessary for determining the discharge-probability relationship. including confidence bands.6528 543.87558/47 = 3. which is used in both the determination of the Weibull plotting position and in the development of the analytical discharge-probability relationship.3378. The mean of the logarithms of the annual peak discharges X is: x = C X/N = 156.176 ft3/s The standard deviation s = ( EX-(XX)2/N N-l 1 S of the logarithms 0.FLOOD HYDROLOGY MANUAL of these hydrographs was discussed in section 5.2949 From the previous computations.4225)-(3)(47)(156.87558)3 (47)(47-1)(47-2)(0. The available annual peak discharge record extends from 1938 to 1984. the bulletin’s generalized skew coefficient for the logarithms 214 . A stream gauge. has recorded flows at a location on the Washita River near Cheyenne. Oklahoma since 1938.2949.8 of chapter 5. the discharge-probability relationship at the gauging station will require transposition to the point of interest. the mean of their logarithms X = 3.222)+(2)(156. standard deviation of logarithms S = 0. operated by the USGS. Since the data reflect runoff from only part of the basin above the point of interest. Table 7-l shows that the number of items of data N is 47. This gauge records flows from 792 square miles of the total 1. and skew coefficient of logarithms G.466-square mile drainage area above Foss Dam. is required to provide for the care and diversion of flows during construction of the modifications. or 2. of the logarithms of the annual peak discharges iy2(~3)-3~(~(~~)+(2)(~3 = N(N.5 = ( of the annual discharges is: 0.87558)(543.

750 6.800 1.890 3.93693 9.900 7.26635 44.45709 13.56925 9.51731 10.60413 12.96506 15.400 1.830 3.12650 31.190 1.17896 17.35793 3.510 7.70799 13.98707 41.89813 9.77072 6.94939 3.94128 3.74819 2.900 8./s 69.67302 3.990 4.29226 3.070 1.240 1.07555 3.083 .354 . ftJ/s 14.708 .313 .55299 49.854 .625 .69533 63.420 1.875 .75287 49.25527 3.375 .958 .479 .60133 61.14613 4.14082 30.62428 3.930 574 159 1.30738 20.66834 8.48714 3.710 4.792 .13541 12.49570 543.729 .104 .229 .900 11.500 .16435 4.542 .292 .042 .15229 3.03342 3.21224 7.960 4.97968 18.050 734 592 574 560 465 427 400 297 265 159 119 38 Weibull Plotting Position Logarithm of Discharge X (6) X’ (7) X3 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 0.458 .70243 3.STATISTICS AND PROBABILITIES Table 7-l.667 .66839 3.57618 23.040 4.189 .000 2.400 1.271 .60133 60.000 14.84616 4.86570 2.646 .800 5.83220 14.68473 34.21737 7 1.59768 15.76846 29.24304 3.61011 15.27359 67.360 265 592 1.470 4.10591 9.75584 18.420 1.09342 3.167 .49539 34.85126 3.080 1.125 .99563 3.61781 15.91243 27.688 .210 1.000 11.02119 2.42325 2.10805 32.75891 2.86392 3.3 .11454 5.66896 37.240 9.58995 3.100 8.61158 7.-Annual Foss Dam.960 1.42023 14.20140 2.280 734 4.49108 13.87214 4.30791 2.550 3.12759 8.12267 55.94266 1940.438 .021 .979 Totals 4.771 .890 4.360 1.34181 17.563 .420 427 119 (2) (3) (4) (5) Ranked Annual Peak Discharge.2220 113.63952 47.08028 12.896 .83898 10.250 .01919 11.99968 20.900 8.32409 31.46298 21.600 14.60170 27.080 40.917 .063 .20164 9.990 2.146 .59768 15.59678 10.39822 50.604 .66745 2.65031 3.74757 42.17898 3.280 2.14613 3.800 40.55023 3.60206 4.92686 3.86299 37.40413 41.13354 3.900 9.60639 46.53381 21.450 7.190 1.070 2.36584 48.32476 13.68785 57.84386 4.660 297 400 560 38 1.16015 12.76567 3.420 5.100 6.81907 9.333 .15854 10.710 1.60206 2.417 .55308 57.310 2.68576 7.46683 72.49736 14.396 .510 1.19039 16.4225 215 I .34044 3.900 8.19928 53.750 .583 .18027 13. (1) Year peak discharges for each year of record for the drainage area above Annual Peak Discharge.27443 35.750 1.040 465 3.470 2.210 3.57978 156. ft.208 .94939 3.77232 2.11980 14.11529 6.80754 3.07555 1.88774 12.833 .930 2.800 2.20037 17.600 3.550 69.938 .87558 23.450 5.63043 2.660 4.79047 61.900 8.91916 6.830 5.22967 10.92988 14.47276 2.55255 7.310 7.900 8.47567 3.27569 11.000 14.8 1.050 1.65139 97.521 .46687 3.

: Bul- MSEc (G) + MSE. by using equation (6).1299 + 0. (6) MSE. = 10 exp where: (A-B[log. MSE... A = -0. = = = G = G = weighted skew coefficient. = where: G. MSE.1299 is now available to fit the LP-III peak discharges using the equation: log Q = % + KS distribution (8) exceedence probability. (N/IO)]1 (7) A = -0. mean-square error of gauging station skew.251 = 0.26 ICI.(47/10)] Glz.90 + 0.150.FLOOD HYDROLOGY MANUAL of annual peak discharges is -0. that is a function of the weighted skew coefficient and selected exceedence probability. and standard deviation of annual peak discharge logarithms.150) = -0.30 IGI. = 10 exp -0. and generalized skew from Bulletin 17B[14].3064-0. in equation is found by applying the following equation: MSE.1299 0.8633 Then.302. therefore.50 = if IGI > 1.3064 into equation (7) results in: MSE. solve for [log..302(-0.8633. letin 17B are used to estimate the The following weighted skew equations coefficient from G. computed gauging station skew.08 IGI. and the value of MSE. + 0. + MSE.55. (6) is always 0..: (-0. if lG1 5 0. G.2949) Grr = 0.52 0.33 = -0. mean-square error of generalized skew.2949. if lG1 5 I.94-0.302 The required information to the recorded annual + 0.90 B = 0. (G) G. Entering found these to be values The computed gauging station -0. if lG1 > 0. The value of MSE.50 skew G was previously and B = 0. 216 . where: log Q = x = logarithm mean of discharge Q at selected K = a factor S = of logarithms of annual peak discharges.

and 95-percent confidence band levels. it remains to compute using the following equations and Kg= where: K GW' + CK2G . x = mean of logarithms of annual peak discharges. Since x and S have previously procedures: (10) Kg and K$. and . The next step is to develop the confidence limits. and S = standard deviation of annual peak discharge logarithms. and x and S = as previously defined in equation (9).ub)". K$ = value of K at probability P that defines lower confidence band curve L.STATISTICS AND PROBABILITIES The development of the data for constructing the discharge-probability relationship at the gauge is shown in table 7-2.P = LP III deviates. Oklahoma. values for KGw are shown in table 7-3 for different values of P. where Z. Bureau practice is to display confidence bands at the 5. The values for both the actual observed annual peak discharges and their respective plotting positions.5 P W’P a (11) K:! = as previously defined in equation (9). This figure depicts the discharge probability relationship for the stream gauge on the Washita River near Cheyenne.64485 for the 5. K GW. = 1. which is accomplished been computed. as shown on figure 7-4. and N = number of years of record.and 95-percent levels as an indicator of the inherent hydrologic uncertainties in discharge-probability relationships. (I$ = sz + K$ (S) where: @ = logarithm of discharge on lower confidence band curve L at annual probability P. The equations for computing points that define the upper and lower confidence band levels are: rk”=X+K:!(S) (9) where: @ = logarithm of discharge on upper confidence band curve U at annual probability P. Kg = value of K at probability P that defines upper confidence band curve U. and the analytically derived discharges and their respective exceedence probabilities are plotted on lognormal probability paper.

25170 1.37023 .75613 x 3.14117 hl w 00 * Interpolated from the values -0.36495 3.000 ft’/ s and to nearest 100 fts/s for values over 10.6528 .16340 1.91729 2. 0.58941 4. ftS/s 2.52705 0.89357 4.3378 3. these numbers would be rounded Table 7-3.6528 .97059 -0.798 50.02 .6528 .04 .5 .61844 4.3378 3. Oklahoma.25170 1.628 79.6528 .08472 1.61860 2.09343 2.3378 s 0.348 12.6528 K’ 0.5 .3 given in Bulletin 17B[14].291 9.97059 . Y-m 2 5 10 25 50 100 (3) (6) KS 0.3378 3.23647 .6528 . Exceedence Probability P 0.3378 3.97059 .and 95-percent confidence band calculations for the Washita River near Cheyenne.317 7.73556 x 3.28648 1.39260 2.286 26. Oklahoma.02 .85136 1.l *dtL K (2) (3) a (4) b (5) K2G .70349 3.97059 .6528 3.25161 1.66164 1.33055 1.645 30.223 16.66164 1.6528 .31929 .3378 3.376 117.6528 .-ab W.97059 .39776 (7) ‘og a (3) + (6) 3.97059 .827 14.083 22.3378 .Ol of data for construction of discharge-probability (4) relationship.81711 1.02715 0.-Development (1) Annual Exceedence Probability 0.04 .2 and -0.609 5.91729 2.493 .55817 1.55577 0.6528 .04159 0.85136 1.000 ftS/s. (5) Washita River near Cheyenne.42252 4.05583 .6528 .27783 .59091 .04159 .3378 3. * Discharges shown are to nearest cubic foot per second (fts/s) to facilitate reader’s understanding.851 54.20079 .6528 .Ol 0.50919 2.928 1.2 .40494 .2 .3378 3.3378 3.3378 3.65597 -0.l .510 24.3378 3.66725 1.395 5 si r (2) Return Period.43669 (‘3) (7) (8) (9) S (10) (11) GUf Kpu 0.14117 0.96066 1. Normally. to nearest 10 fts/s for values up to 10.Table 7-P. (8) Discharge*.15491 4.456 38.-The 5.

p=o.170 feet at the gauging station near Drake. ’ 0.1 bP-O. from 1947 through 1976.25170* . This is accomplished by multiplying each discharge value by the ratio of the square roots of the drainage areas: d Foss Dam Drainage d Gauged Drainage Area Area = q7Fif-iF the discharge-probability Analyses = 1 3605 * relationships are The final results representing shown in table 7-4.0.01 K2.STATISTICS AND PROBABILITIES b = K&.05756 = Table 7-3 shows the salient features of the 5.50919 = 1. P -7 Z.02 = K2 GW.70349 = 3. The record available for this analysis covers 30 years.05756 = 0. It now remains to transpose the discharge relationship from the gauging station downstream to the Foss Dam site. lying on the eastern slope of the Rocky Mountains. 219 .2 square miles.05756 = 2.97059 bP-O.52705 confidence = K2.000 feet.0.0.66164* .9172g2 .w.5 = 0.2 (Note that 6 varies as a function bility P) of annual proba- The equation for computing level is: K.and 95-percent band calculations.5 bP-O.0.0.05756 = . Colorado.04 = 1.W. and the drainage area involved is 82.01 = 2.p. = points that define the lower confidence band K 'WsP - (K9.0851362 ..0. The North Fork.14117* . 7.04159* .05756 bP-O.05756 = 1.12 Example on Mixed Population Consider the North Fork of the Big Thompson River near Drake.0.04 bP-O.61844 = 4. heads at Rowe Glacier in Rocky Mountain National Park at an elevation of over 13.66725 = = K2Gw.0.. p= P= o’.02 = l. The river is subject to flooding from two sources. -K2 GW.P-02 K2GW.2 bP-O. p a ab)'= (12) Continuing the example computations: = 0. The river descends to 6. P= 0.05583 = 0.



Table 7-4.-Discharge-probability


at Foss DamI

P Annual Exceedance Probability 0.5 .2 .l .04 .02 .Ol






3,152 10,649 19,436 35,994 52,857 74,005

4,555 17,020 33,726 68,880 107,993 160,443

2,189 7,199 12,548 21,881 30,809 41,486

‘Discharge values shown are to nearest cubic foot per second to facilitate reader’s understanding. Normally, these values would be rounded to the nearest 10 fty/s for values up to 10,000 fP/s, and to the nearest 100 ftl/s for values over 10,000 ft?/s.

snowmelt and summer thunderstorm. A paper by Elliot, et al. [66] provides a listing of the annual rain peak and snow peak discharges, see table 7-5. Using these data, a separate discharge-probability analysis can be made for both the snowmelt and rainfall runoff related data. This is accomplished in the same manner shown in the example on Foss Dam, section 7.11, for the base frequency curve. The results of the analyses are shown in table 7-6. Using the parameters shown in table 7-6 in equation (8) section 7. II, leads to the individual discharge probability curves for rain floods and snowmelt floods shown on figure 7-4. Since it is desired to have a single curve that reflects the annual probability of a given level of flow occurring, the two curves must be combined. This is accomplished by applying equation (3), section 7.7. By applying this equation at a number of flow levels yields the points necessary to define the total population curve shown on figure 7-4.



Table 7-5.-Annual rain peak and snow peak discharges for North Fork of Big Thompson River near Drake, Colorado. From [SS]. Rain Peak Water Year Date Discharge, ftS/s 158 86 820 450 211 159 206 44 114 228 850 147 52 43 131 138 107 74 1,290 584 131 156 800 194 79 134 355 156 845 8,710 Date June 21 11 06 12 19 05 13 27 25 02 08 May 24 June 20 18 03 15 16 08 10 18 21 06 17 24 19 04 11 17 16 09 Snow Peak Discharge, ftS/s 410 166 766 129 232 283 223 47 51 161 334 295 153 120 275 110 157 84 250 60 148 148 293 251 227 170 398 184 216 82

1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976

June July Aug June July Aug May July June July Sep Aug it;

04 10 03 26 I? 14 21 29 25 31 18 01 ::

July 29 June 16

Aug 12 May 07

June 17 May 06 July 15 June 18 July 31




Table 7-6.-Flood frequency Drake, Colorado.


for North

Fork of Big Thompson

River near

Logarithmic Mean Rainfloods: Generalized Skew Systematic Record Weighted Skew Final Parameters Snowmelt Floods: Generalized Skew Systematic Record Weighted Skew Final Parameters

Logarithmic Standard Deviation

Logarithmic Skew

2.3225 2.3225

0.4877 0.4877

0.22 1.3710 0.30 0.30

2.2507 2.2507

0.2738 0.2738

-0.10 -0.2522 -0.10 -0.10


8.1 General The importance of providing complete technical documentation of flood hydrology studies cannot be overemphasized. At any time after a flood study report is completed, any hydraulic engineer should be able to independently reproduce all flood values contained in the report based solely on its contents. If the documentation is complete, the reviewer will be in a better position to recommend approval of the study without having to resort to numerous contacts with the author to resolve points that are unclear. Each assumption and decision made in the course of the study should be thoroughly documented along with the rationale for making each assumption. The report should also include the field reconnaissance report as an appendix. All calculations leading to the development of a PMF hydrograph or discharge-probability relationship should either be included in the body of the report or in an appropriate appendix. In addition, all review comments and correspondence indicating approval of the study should be included as an appendix. 8.2 Specific Contents of a Flood Study Report items

To achieve the objectives discussed in section 8.1, the following should be included in all flood study reports.

(a) Authority.-The appropriate legislation, formal request, or contract for the report should be cited, including their dates. At the onset of a report, it is of considerable value to include a discussion on the objective of the study and a description of the physical features of the area involved. If the study is to support project planning efforts, a discussion on alternative physical project configurations should also be presented at this time. This will provide the reviewer the opportunity to assess whether the required hydrologic and meteorologic combinations have been properly recognized and analyzed. The level of the investigation or design that the study is conducted to support should also be presented.

(b) Summary of Results. -A summary of the study results should be presented near the beginning of the report to provide ready reference. In many cases, more than one PMF will be developed to reflect the seasonal variation of the magnitudes of the events. The peaks and volumes for each PMF derived are presented for the particular season (spring rain-on-snow, summer, fall, or winter rain-on-snow) and the type




of storm that produced the flood hydrograph (snowmelt, general storm, local thunderstorm, etc.). In some flood studies, hydrographs of specific frequency or probability will have been developed, as discussed in chapter 7. If this is the case, the peaks and volumes for each associated probability should be tabulated. The final item in the summary should be a statement regarding the criteria to be used for routing the flood through the reservoir. These criteria should include the starting reservoir water surface elevation to be assumed at the onset of the PMF; release capability of outlet works and powerplant, if any, to be assumed during the routing; any constraints on spillway operations for gated or stoplogged structures; and any transbasin diversions that may be entering the reservoir.

(c) General Background.-This section of the flood study report should begin with a brief description and purposes of previous flood studies prepared for locations at or near the site under investigation. A summary of the resulting flood values including peaks and volumes for PMF’s and floods of specific frequency, if appropriate, should also be included for these previous studies. This should be followed with a discussion that includes all formal and informal agreements reached between the various organizational elements relative to technical aspects. Reference to the field reconnaissance should be made and should include the dates it was conducted, participants, and the offices the participants represented. The field reconnaissance report, the content of which will be discussed later in this chapter, should be included as an appendix to the flood hydrology report. (d) Description of Drainage Basin.-Much of the information that should be contained in this section may be extracted from the field reconnaissance report. The information should include a narrative description of the geographic location of the drainage basin, and should always include a location map. This may should be of sufficient scale so that readily identifiable geographic features such as highways, towns, and watercourses will enable the reader to easily identify the general location of the drainage basin. The drainage basin boundary should also be clearly delineated on the location map. The discussion should then proceed to fully describe the topographic features of the basin. This should include the range of elevations and stream slopes present in the basin, drainage network development and type, general geologic setting, prevalent soil types, and vegetal cover. The degree and type of development or land use, both present and projected, should always be included. Any projected development should reflect those of the most authoritative local source available, such as city or county planning and zoning entities. This section of the report should conclude by documenting the existing and future water control facilities in the basin. The potential effect of these facilities on flood runoff should be discussed, and the pertinent structural and hydraulic capacity and sizing data for each facility should be presented in tabular form.


A graphical representation of the reproduced flood hydrograph versus the observed flood hydrograph should always be presented. This is generally encountered when performing hydrologic investigations for drainage basins larger in area1 extent than considered in the report series. fast or slow rising. The body of the flood hydrology report should present the pertinent data resulting from these applications. Precipitation amounts and their distribution in time and space should also be included. a summary of the unit hydrograph lag parameters and the dimensionless hydrographs should be presented along with a brief narrative documenting the rationale for their selection. Hydrographs. and mass curves of rainfall should be included. if available. the report should include all details of the reconstitution analysis. In these cases. However. the hydrometeorological analysis used in developing the individual drainage estimate should be included in its entirety as an appendix to the flood hydrology report. In virtually all cases. infiltration loss rate assumed.Hydrometeorological Analysis.. and average and extreme temperature and precipitation experienced in the basin should be included. Summary information relative to general air mass circulation. Also to be included on this graphical representation is a hyetograph of the basin average rainfall. Since details of their selection are presented in the field reconnaissance report.-This analysis should begin with a discussion on the climatology of the region and specifically of the basin. storm types typical of the region that impact on the drainage basin. pertinent data from the individual drainage estimate should be included in the body of the report. the latter should be tabulated by season. temporal distribution of rainfall. The nature of the events should be described in terms of their rates of rise. These details would include the estimated storm isohyetal pattern. -This section of the flood study report should provide a discussion on the historical floods that have occurred in the basin under study and in nearby hydrologically and meteorologically similar basins.REPORTS (e) Historical Floods.-If a synthetic unit hydrograph has been used in the flood study. reference should be made to this report. and the final unit hydrograph used to produce the reconstituted flood hydrograph. as discussed in chapter 3. There will be situations where the criteria contained in the HMR series reports do not apply. with the loss 225 . which is included as an appendix to the flood study report. The discussion should include information relative to the peak discharge volumes. However. In cases where the basis of the unit hydrograph used in the study is the result of an observed flood hydrograph reconstitution. cf. (g) Unit Hydrograph. the PMP or storm will be derived by application of criteria contained in the appropriate hydrometeorological report discussed in chapter 3. including their durations. the work sheets and a brief narrative of the process involved in deriving the PMP estimate for each case should be included as an appendix to the report. isohyetal maps. base and interflow assumptions.

226 . snowmelt runoff of a specific frequency or annual probability of occurrence will be adopted as the snowmelt component of the PMF rather than results obtained from the Western Snowmelt Equation. interof presentation is (h) Infiltration Loss Rates.-When the Western Snowmelt Equation (sec. density. and wind velocities and temperatures assumed for the study. The report should then include the complete water budget analysis in tabular form. a full explanation of the rationale for reaching this conclusion should be documented. and area1 extent of snow cover assumed. In this situation. In most cases. 35 [59]‘. and where the land use is predominantly urban. or an explanation of any substantial interflow that is likely to occur due to topographic. or geologic conditions that may exist in many drainage basins. (‘j) Snowmelt Runoff. Include an explanation on the recession flow typical for the season in which the PMF could occur. hydrographs of this type (base flow. 4) is used to estimate a drainage basin’s snowmelt runoff potential. 4. and its temporal placement in relationship to the probable maximum rainflood hydrograph should be fully documented. both peak and volume. statistics INumbers in brackets refer to entries in the Bibliography. forest cover percentage determined to be representative of the basin. the summary should include information relating to the density of development and the resulting assumptions regarding composite loss rates reflecting the degree of pervious and impervious areas existing and projected.FLOOD HYDROLOGY MANUAL rates superimposed. a full explanation of the discharge probability study should be presented.-This part of the flood study report basically reiterates. general geologic setting. (i) Base Flow and Znterjlow. soils. The discussion should also include consideration of snowcover and frozen ground as they impact infiltration loss rates. and the component flow. in summary fashion. In situations where the base flow and interflow have been concluded to be negligible. the rationale for determining the snowmelt runoff hydrograph. and surface runoff). In developed drainage basins. the base flow component may consist of snowmelt runoff. An example shown on figure 4-14 in chapter 4. the conclusions reached relative to infiltration loss rates determined to be appropriate in the course of the field reconnaissance or in the flood hydrograph reconstitution analyses. and consideration of vegetal cover as it impacts infiltration losses. this part of the report should provide full documentation as to the depth.-Included in this section should be the complete rationale for selection of the base flow and interflow hydrograph. In some cases. cp. and should include the streamflow records utilized. If so. as described and illustrated in Engineering Monograph No. This summary should include a brief discussion on the soil types present in the drainage basin. or where development is projected.3.

The results of such reconstitutions should be summarized in the report. Generally. The only criteria currently available for the Western States are found in NWS Technical Memorandum [61].-As discussed in chapters 4 and 5. In some cases. a routing reconstitution of an observed flood event may have been conducted to provide site specific routing parameters. power facilities. many flood hydrology studies are based on dividing basins into smaller subbasins that require channel routing and combining of the individual flood hydrographs. and results of the flood routing and combining performed. and guidelines can generally be found in the Standard Operating Procedures for existing projects. Denver Office until such time as regionalized antecedent storm and flood criteria become available. this justification should be based on consultation with hydrometeorologists in the Flood Section. at the point 227 . the basis and rationale for parameter selection. Include any constraints that are to be assumed relative to the capability of the outlet works. for assumed conditions producing the antecedent flood. The hydrologic engineer should keep in mind that assumptions of reasonable meteorologic conditions are basic to antecedent flood estimates. (m) Reservoir Routing Criteria. (I) Flood Hydrograph Routings. (k) Antecedent Flood. (n) PhfF Hydrograph. This section of the report should provide documentation on the type of flood routing technique applied. In conclusion. These constraints are generally determined by hydraulic design engineers knowledgeable in the area of hydraulic structures.-This section of the report should provide the basis and rationale for selecting the antecedent flood.-This section of the report provides a brief summary type discussion on the manner that each of the foregoing components are combined to yield the final flood hydx-ograph. Therefore.-The summary information regarding this item is included at the beginning of the report under “Summary of Study Results. the report should include justification. Transbasin diversions that may affect reservoir operations should also be discussed. and spillways to pass or assist in accommodating the PMF. and the details included as an appendix. or hydrographs in those cases where seasonal PMF’s are determined. in meteorological terms.” This section of the report should include a detailed discussion on the basis and rationale for the starting reservoir water surface assumed to occur at the onset of the PMF. and the results of the analysis.REPORTS generated based on the records. particularly in reference to its magnitude and timing as related to the PMF. Include the consideration of the season when the PMF is likely to occur and the attendant possibility of the reservoir being at or above certain pool levels during this season. the basis and rationale of the timing of the statistically determined snowmelt component with the probable maximum rainflood component should be provided.

etc. names and offices represented individuals of personnel on field by these personnel. trav- 228 . separate tables should be provided for each type of flood event. The narrative part of this section should provide a discussion on the stream gauges selected. and 15-day volumes. and for flood control requirements. the PMF in terms of peak discharge and volumes over specified durations.-The flood hydrology study will usually include discharge-probability analyses. and volume amounts are presented. thunderstorm.) represented by the curves. 50-. for concurrence. Generally.-Each flood hydrology report should contain envelope curves of experienced peak discharge. The narrative portion should provide information relative to the source of the streamflow data used. and the type (snowmelt. and routes recon- l and offices visited. and the type of flood event (general storm. cross drainage design for project roads and water conveyance facilities. reconnaissance. This report should be based on the detailed field notes taken by each of the participants during the field reconnaissance. peak flood discharge. and base flow and interflow components. A summary table should be included that lists the peak discharge and appropriate volumes for the 5-. general storm. if required. relationships between drainage area. addition of snowmelt. This section should present the basis and results of the analysis leading to the development of peak and volume discharge-probability curves. Locations eled during reconnaissance. Denver Office. contacted. The results of these analyses are used for determining diversion requirements during construction of dams.FLOOD HYDROLOGY MANUAL of interest. (0) Envelope Curues. in tabular form. representative geographic area. and loo-year events.3 Field Reconnaissance Report This report should be prepared by the Bureau regional or project office involved in the study as soon as practicable after the field reconnaissance is completed. use of a regionalized approach (if applicable). In general. and should speciflcally contain information regarding the following items: l Dates of field naissance team. Conclude by summarizing.) of flood event represented. unit hydrograph application to excess rainfall. length of record available at each of the stream gauges used. 8. These notes should address each of the points discussed in chapter 4. A graphical representation of the final flood hydrograph should be presented. (p) Discharge-Probability Analyses. Each plot of data should be accompanied by tabular information as described in chapter 6. When completed. lo-. To be included is the abstraction of infiltration losses from rainfall. the field reconnaissance report should be included as tin appendix to the flood hydrology report. snowmelt. usually 5-. the report should be transmitted to the Flood Section. 25-. etc. lo-.

correspondence that l Synopsis of field trip. including a detailed discussion on observations made relative to defining the drainage basin’s drainage network. be adopted for use in the l l 229 . determined flood study. pertinent water control facilities. Flood Section. infiltration rates and areas to which they are applicable. it should be so stated in the conclusions and referred to the Section Head. land use. and any major obstructions to flow such as highway or railroad embankments that are located in the drainage basin. soil and geologic conditions. A recommendation rameters. vegetal cover.REPORTS l Applicable references to formal and informal prompted the field reconnaissance. In the event that members of the reconnaissance team cannot reach mutual agreement on the selection of these parameters. for resolution. These observations should relate to the route traveled. and the relative forest cover used for snowmelt analyses. Denver Office. dimensionless unit hydrograph. Conclusions as to the selection of hydrologic parameters including the unit hydrograph lag time coefficient used in the general lag equation. statement to the effect that the hydrologic pato be appropriate. topographic conditions. and the discussion should be presented in sufficient detail so that it provides all necessary support for the conclusions reached relative to the on-site selection of hydrologic parameters.


American Society of Civil Engineers. [7] Bernard. DC. U. B. A. 1952. NY. Birthplace of Systematic Stream Gaging. “The Role of Infiltration in the Hydrologic Cycle. DC. vol. 108. Bureau of Reclamation. Washington. Water Resources Council.” Washington. 1968. USGS. L. Water and Power Resources Service (Bureau of Reclamation). vol. O. 1976.. Catalog of Dam Disasters. Merrill. E. pp. Bulletin 15. Los Angeles. E. pp. U. H. “Evolution of PMP Cooperative Studies. “Streamflow from Rainfall by the Unit-Graph Method. U. 14. Handbook of Applied Hydrology. McGraw-Hill New York. Ven T. Washington. No. Book Co. [9] Barnes. Corps of Engineers. Dams and Public Safety. [4] Babb.” Transactions of the American Society of Civil Engineers. 516-526. revised June 1977. Improved Procedure jw Determining Drainage Area Lag V&es. et al.S. December 1967.” Engineering News Record. Bulletin 17A. co. Failures. F. pp.. [lo] Tatum. A.. 109. 501-505.S. [8] Stallings. vol. U. [5] Sherman. Bulletin 17. vol. [3] Frazier... Heckler.S. 1933. DC.. Water Resources Council. Denver. ! 231 . Washington.. R.. Department of the Interior. April 7. pp. 446-460. A. Army Engineer District. 112. October 1986. U. [2] Chow.” Journal of Water ResourcesPlanning and Management. July 1962. Water Resources Council. 1944. [ 121 Guidelines for Determining Flood Flow Frequency. [ 111 A Uniform Techniquefor Determining Flood Flow Frequencies. 311-382. S. E. New Mexico. “Embudo. Robert B. Professional Paper 778.. 4. 1964. and W. Unit-Graph Procedures. 1972. K. “The Primary Role of Meteorology in Flood Flow Estimating. and Accidents. [6] Horton.. DC. 1932. Denver.S. CO. DC. 1980. Bureau of Reclamation.” Transactions of the American Geophysical Union.[l] Jansen. Washington.S. [ 131 Guidelines for Determining Flood Flow Frequency.

Tracey.S.. National Weather Service. [2O] “Generalized Estimates United States West of the U. M. et al. 1982... 638 pp.S. Environment Service. and J. 1969. U. “Application of Probable Maximum Precipitation Estimates-United States East of the 105th Meridian. H. Canada. F. Hydraulics.. vol. Open Channel 1959. J..S. World Geneva. 237. Washington. E. Stream-Gaging Procedure-A Manual Describing Methods and Practices of the Geological Survey. R. F. T. vol. Boston. U. vol. 1960. Washington. 1: Montana. of Probable Maximum Precipitation for the 105th Meridian. NOAA Technical Memorandum Depths for the Contiguous NWS Hydro-33. J. 114 pp. 1945. National Oceanic and Atmospheric Administration. for Depth-Area-Duration Organization. National Weather Service. Washington. Washington. [ 151 Corbett. NY. Department of Commerce. 5: Idaho. Jr.. Geological Survey Water Supply Paper 888. Washington. D. Silver Spring. Dutu. [ 171 Barnes. 1973. [16] Chow York. National Weather Service. WMO Analysis of Storm Precipitation. Department of Commerce. H. [ 191 Glossry ofM e t eorology. Department of Commerce.S. TP 129. 7: Nevada. MA. 1959. American Meteorological Society. Ontario. 1967. L. DC. Bulletin Determining Flood Flow Frequency.S. 1943 (reprinted in 1962). [25] Storm Rainfull in Canada. Miller. 2: Wyoming. T. Depth-Area-Duration DC. V. vol. vol. Army Corps of Engineers. 174 pp. [24] Shipe. [23] Storm Rainfall in the United States. 169 pp.. Precipitation Frequency Atlas of the Western United States: vol.[ 141 Guidelinesfor Council. 11: California. and R. National Oceanic and Atmospheric Administration. Geo- [18] Hansen. Silver Spring. Department of Commerce... H. Schreiner. 6: Utah. 66 pp. National Oceanic and Atmospheric Administration. 9: Washington. Frederick. and J. vol. C. U. U. Washington.S. No. vol.. [ 2 1 ] Manual Meteorological Switzerland. Downs- 232 . Riedel. revised September 1981. DC. 38. A. MD. Roughness Chnructeristics logical Survey Water Supply Paper 1849. vol. 52. P..” Technical Paper No.. MD. vol. 4: New Mexico. U. M. Greatest Known Area1 Storm Rainfall United States. DC. DC. 8: Arizona. 3: Colorado. vol. New of Naturul Chunnels. 1976.. Water Resources 17B. [22] Miller. DC.. NOAA Atlas 2. McGraw-Hill Book Co. Weather Bureau.” Hydrometeorological Report No. Atmospheric view. 10: Oregon. 1961.

Hansen. [33] Hansen..” Hydrometeorological Report No. U.S. J. L. DC. Washington. National Weather Service.” Hydrometeorological 233 ..” Hydrometeorological Report No. Washington. and Moisture in the Intense Rainstorms in Eastern Colorado. Operational Hydrology Report No. National Oceanic and Atmospheric Administration. Weather Bureau.. 50. and J. U. U. “Probable Maximum Precipitation Estimates-United States Between the Continental Divide and the 103d Meridian. Schwarz. F. C.S. DC. U. Department of Commerce.” Hydrometeorological Report No. Department of Commerce. Chapman London.S. and Hall LTD. World Meteorological Organization. 232 pp. “Comparison of Generalized Estimates of Probable Maximum Precipitation with Greatest Observed Rainfalls. 1951.. U. 27 pp.S. [35] Schreiner. Environmental Science Services Administration.. [31] Miller. 1970. Department of Commerce. Environmental Science Services Administration. ESSA Technical Memorandum WBTM HYDRO-3. Instability.[26] Schwarz.” NOAA Technical Report NWS 25. [29] Manual for Estimation of Probable Maximum Precipitation. 1986. K. Fenn.S. Department of Commerce. 66 pp. D. WMO No. 21 pp. 1947. 1968. M. Weather Bureau. and L...S. E. 77 pp. The Role of Persistence.. Department of Commerce. DC. L. (In press). National Oceanic and Atmospheric Administration. and D. National Oceanic and Atmospheric Administration. 1967. Geneva. F. Switzerland. Department of Commerce. National Weather Service. DC. Hydrometeorology. [32] Climatic Atlas of the United States.. England.200 and 500 Square Miles.269 pp. 1981.S. DC. T. Schreiner. T.. [30] Wiesner. U. M. [27] Tables $ Precipitable Water. 1980. 80 pp.. E.S. T. “Probable Maximum Precipitation Estimates-United States East of the 105th Meridian. J. National Weather Service. “Meteorology of Important Rainstorms in the Colorado River and Great Basin Drainages. Riedel. MD. DC. Environmental Data Service. 55A. U. Department of Commerce. 23. 332. 167 pp. C. K. Schreiner. Washington. D. U. and F.S. U. Washington. J. 1. C. C. Silver Spring.. Technical Paper No. Jensen. Weather Bureau. Washington. Washington.. [28] “Generalized Estimates of Maximum Possible Precipitation Over the United States East of the 105th Meridian for Areas of 10. 14.. [34] Riedel.

Commerce. 228 pp. in California. Weather Bureau.S. V. Washington.S.. A. 53.. Washington. 1980. 89 pp. U. Environmental Science Services Administration.. “Probable Maximum Precipitation in the Hawaiian Islands. Department of Commerce. and J. U. A. Washington. U. National Weather Service. “Criteria and Limitations for the Transposition of Various Watersheds.S. K. 1963. Department of Commerce. National Atmospheric Administration. 105 pp. Washington. F. Department of Commerce. Department of Commerce. 33.” Report No. MD. U. 161 pp. 69 pp. 36.. [41] Myers. U. MD.” Hydrometeorological Report No.” Hydrometeorological Report No.S. “ 1980 Seasonal Variation of 1 O-Square Mile Probable Maximum Precipitation Estimates. 234 . U. 1951. Paper presented at the Symposium on ConAspects of Storms and Floods in Water Planning.S. “Probable Maximum Precipitation and Rainfall Frequency Data for Alaska. F. DC.Report No.” Technical Memorandum WBTM Hydro-5... 1977. Spring. U. Texas Board. 1966. Washington. [43] Ho. DC. American Society of Civil Engineers. Large Storms Over Water Development sideration of Some October 7-9. 87 pp. National Oceanic and Atmospheric Administration. 43. 47. 20 pp. F. DC. Department National Oceanic and of Commerce. [39] “Interim Report-Probable Maximum Precipitation Hydrometeorological Report No. P.S. [37] Myers. 1966. Texas Section. DC.S.S. Riedel. Weather Bureau. U. Department of Commerce.” Hydrometeorological Report No. U. 1967.S. E. United States East of the 105th Meridian. T. U. Weather Bureau. Department of Commerce. Washington. 1978. 39. 51. [38] Hansen. Silver [36] “Kansas-Missouri Floods of June-July 17. DC. “Probable Maximum Precipitation Estimates. National Oceanic and Atmospheric Administration. MD. U. J.” Technical Department of Paper No. Weather Department of Commerce.S.. Weather Bureau. Colorado River and Great Basin Drainages... U. 29 pp. [40] “Probable Maximum Precipitation Estimates-Northwest States.. 1965.. T. 1952. 1961. Schwarz.S. “Meteorological Estimation of Extreme Precipitation for Spillway Design Floods. 202 pp. Weather Bureau. and J.” Hydrometeorological Report No. M.S. K.. 1963. [42] Miller. Silver Spring.. Weather Service. 98 pp. F. [44] Schwarz. U. 49.” Technical Paper No.. V..S.” Bureau. Environmental Science Services Administration. National Weather Service.. Riedel. Silver Spring. DC.

U. [53] Weaver. U. “Rainfall Frequency Atlas of the United States. DC. p.S..” Hydrometeorological Report No. 82 pp. DC. Auciello.S. DC.S. H. W. Weather Bureau. Department of Commerce. Weather Bureau. 115 pp. J.S. and E.S. 69 pp. Department of Commerce. 1963. and F. Average Precipitation in the United Statesfor the Period 1906 to 1936 Inclusive. “Maximum 24-Hour Precipitation in the United States.” Technical Paper No.. National Weather Service.” NOAA Technical Memorandum NWS HYDRO 35. Department of Commerce. DC. F.” Technical Paper No.S.. 47.S. [54] Horner. 1. 235 . Department of Commerce.” Transactions of the American Geophysics Union. [46] Manual for Depth-Area-Duration Analysis of Storm Precipitation. 115 pp.. U. L.S. 1938.. 101. “Probable Maximum Precipitation and Snowmelt Criteria for Southeast Alaska.S. MD.29 pp. Department of Commerce. [52] Frederick. F... “Synthetic Unit Hydrographs. Technical Paper No. Weather Bureau. 37. “Meteorology of Hydrologically Critical Storms in California.S.” Transactions of the American Society of Civil Engineers. 1931-1960. Myers. U. vol. pp.S. National Oceanic and Atmospheric Administration.” Climatography of the United States No. U. Washington. U. Weather Bureau. A. M. R. Cooperative Studies Technical Paper No. 1964. U.. “Five. DC. Washington. part 1. Weather Bureau.. U. W. F. “Two to Ten-Day Precipitation for Return Periods of 2 to 100 years in the Contiguous United States. National Oceanic and Atmospheric Administration. D. Department of Commerce. Department of Commerce. H. W.. 85. U. L. Weather Bureau. [48] McDonald.to 60minute Precipitation Frequency for Eastern and Central United States. 1961. Washington. Silver Spring. R.S.S. U. Flynt. DC. “Relation Between Rainfall and Runoff from Small Urban Areas. 1944. 140. P. [47] Jennings. 1962. [50] Hershfield. V. 213 pp. 1944. Washington. F. [51] Miller..S.[45] Schwarz. Washington. Department of Commerce. 16. and J.S.. 19. Washington.” Hydrometeorological Report No.. 1983. 1946. K. Miller. F. 447-454. Washington. U. 54. A.. 1936. MD.S. U. U. 40. 1977. [55] Snyder. U. U. vol. F. DC.. U. Department of Commerce. Weather Bureau... National Weather Service. [49] “Decennial Census of United States Climate-Monthly Average for State Climatic Divisions. 284 pp. Silver Spring. 33 pp.

” U.” DC. Iowa State University Regulation of the Tennessee 185. T. Engineer Manual [62] Routing of Floods Through 1408. 1940. 56.000 Square Miles in Area. 1957. London. The Biometric by Cambridge Karl. 3d ed. ward Brops. 35.” 70th Con- 236 .. “Probable Maximum and TVA Precipitation Estimates with Area1 Distribution for Tennessee River Drainages Less Than 3. Bureau from Rain on Snow. Part 1.S. Ebling. University Press..” Hydrometeorological Report No. Army 1936. app. D. NOAA. Flood Flows.S. February TB 5-550-3. 1930. IA. Statistical Press. Washington.. River Channels.[56] “Hydrology. Compaction on Runoff No. NY. “A Graphical Geophysical Union. MD (In press). W. Ed- [59] “Effect of Snow gineering Monograph CO. England. March [63] Wilson. DC. L. A. app. C. MI. [61] “Precipitation Antecedent to the 24-Hour Probable Maximum cipitation for Small Basins in Texas. [60] Zurneorfer. VA. [67] Engineering 5.” Geological Survey Open-File Report 82-426. Method. Transactions of the American [64] Hazen.. T. 20. New York. Publication 101. Army Corps of Engineers.” PreHydro 1110-21960. Surface Runoff Ann Arbor. L. London.S. 1977.S. Memo. Army Corps of Engineers. [69] Puls. and Estuaries. Phenonmena. 1928. Denver. Wiley. Ames. Engineer Canals School. [58] Horton.. and J. Flood-Routing part 3. [66] Elliot. University College. Techniques. G. 1986. Silver Spring. HD Methods in Hydrology. NWS. Dec. Fort Belvoir.. Department of Commerce.. U. “Annual Snowmelt and Rainfall Peak-Flow Data on Selected Foothill Region Streams. part 2. MD. A. [57] “Flood Prediction the Army. et al.. 1941. Whittier Los An- Department of Robert E. San Gabriel River and the Rio Hondo Above Narrows Flood Control Basin. “Flood gress.S. 1966.” Laboratory. NWS Report. U. U. geles District. Washington. B.. “Tables for Statisticians and Biometricians. Silver Spring. 1944. E. G. River. 45 pp. [68] Haan.. 1930.” Enof Reclamation. U. R. J.” NOAA Tech. Printed [65] Pearson. 1st Session. 1982. Jarrett. Procedures as Applied to Tidal Corps of Engineers.

H. 232 Bureau of Reclamation. 202-203 Colorado River. 208. 27 moisture. 232 Base flow. V. T. 16-17. 28 frontal lifting.16 indirect discharge measurements. A. 14. 29 lifting mechanism. 115. 37-38 Water Data Storage and Retrieval System (WATSTORE). 18-19 systematic. 226 Basin factor. 232 Antecedent flood. 30 Atmospheric Environment Service (of Canada). 10 Cumulative distribution function. 11-24 field reconnaissance. 16 transposition. 21. 231. 25. 25. 116-117. 23 Cross drainage. 25 Coast and Cascade ranges dimensionless unit hydrograph. 31. Jr. 236 Concrete Dams Branch. 16 supplemental (bucket survey). 8-10.. M. S.. M. 16-20 form. 29 convergence. 1 I. North Fork. 112-114. Water Resources Council. atmospheric. 29-30 convection. 30 Bernard. 2-3 Atmospheric condensation. 28 Corbett. 177 Columbia River. 232 Auciello. atmospheric. H. 235 Average basin precipitation. 210-211 transposition of storm data. 231 Balanced flood hydrograph. 28 instability. 120-122 Arkansas River. 207. vegetal. 155. 96-97 Coefficient of skew. 219-221 Bubbler gauge. 140. 236 Flood Section. 180. 27-30 water droplet-ice crystal growth. 18-19 Chow. 116-l 18. D. 197-199. 236 Cover. 25. 72 Bergeron-Findeisen theory. 29-30 Confidence limits. 231-232 Climatic trends. 217-219 Convection. 8 Data. 9. 16-20 form. 35. 8. 232 Corps of Engineers (COE). 10. 83.. Bureau of Reclamation. 37-38 transposition of frequency data. 10. B. P.. 18-19 Bulletin 17B. E. 25 Babb. 8. 85. 183-184. 12 Bucket survey. 231 Barnes. 138 Engineering Monograph No. 192-194 Current meter. 28 processes. 20-24. 1 l.. 184. 57 storm survey form. 232.. 205. 22. 228-229 hydrologic. 68. 227 Arbuckle dimensionless unit hydrograph. 1. 15. 138 Condensation. 185-189 meteorologic. supplemental meteorological data.INDEX American Meteorological Society. 2 Dam failure. 29 Convergence. 199-201 generalized. 223. 77. O. atmospheric. 231 Big Thompson River.14 statistical analysis of.. 209-2 10 Barnes. 14-16 recorded streamflow data. 179180 237 . 11. 27 orographic lifting. 165-166 Concrete Dams Branch..

196 generalized extreme value distribution. 150. 225 probabilities (see Frequency) record. 76. 220. 184. 177-181. 195 Wakeby distribution.FLOOD HYDROLOGY MANUAL Density function. rainfall. 191. 197-204 mixed population. 9 frequency (see Frequency) historical floods. 2 11-2 12 routing. 225 regional flood frequency. 10. D. 2-8. frequency analysis. 12 Fixed interval analysis. 154-155 238 . 192-193 Department of the Army. 236 Envelope curves depth-area-duration envelopment. 187 Ebling. 205-208. 187 partial. 810. l-10 hydrograph reconstitutions. 9 Dewpoint-elevation adjustment. 189-197 (see also Frequency) cumulative distribution function. 25 (SPPalso Probable maximum storm) Deterministic flood approach. 196 Diversions. 227 general considerations. 204-205 Extreme value distribution. 228-229 Fisher-Porter punch tape unit. 133-136 modified Puls flood routing. 233 Envelopment depth-area-duration envelopment. L. 228 Evaporation. 109 Exceedance probability. 219-220 normal distribution. D. 20-24. 64 Elliot. 225 historical perspective. 80. J. 196 Gumbel distribution. 140. 196 Failure. 192-194 generalized extreme value distribution. 228-229 report. 195 log Pearson type III distribution. 14-16 Discharge probability (see Frequency) Distribution. 74. 177-181. 191. 115. Bureau of Reclamation. 26. 38-40 recorded data. 7’7. 80. 236 Effective rainfall. 32-36 Dewpoint temperature. 31. 88-89 Rocky Mountains. 223. G. 188. 72-73. 191. 82. 195 presentation. dam. 233 Field reconnaissance.. 98-99 Discharge measurement data recorded. 196 lognormal distribution. frequency analysis. 75. 8 Fenn. 25-61. 189. 83-84 Southwest Desert. 10 Drainage network. 2 Engineering Monograph No. 38-40 recorded data. 9093 Sierra Nevada. 9495 Urban. 191-194 probability density function. 83. 83. J. 21-23 Duration analysis. probability. 183 Excess. 6870. 209-210 deterministic approach. 8’7-102 Arbuckle. 96-97 Great Plains. 236 (see also Corps of Engineers) Depth-Area-Duration (DAD) relationships. 133-175. 86-87. 80-81. 11-14 indirect. frequency analysis. 120-122 Coast and Cascade ranges. flow. 27 maximum. 223.. 36 Dimensionless unit hydrograph. 35. 144-154 modified Wilson flood routing. 203 Extrapolation. 228 Environmental Data Service. 236 Empirical formulas. 79.. 85. 32 defined. 104109 hydrometeorological approach. 196 Gumbel distribution. 38-40 Design storm. 140. 192193 three parameter lognormal distribution. 64-66 Expected probability adjustment. 2 10 Flood balanced hydrograph.

14. 23 Geological Survey (USGS). 2 17-2 19 cumulative distribution function. 171-175 probable maximum flood hydrograph determination. 191. 183-221 analysis by flood type. 203 lognormal distribution. F. 210 flow duration analysis. 205-208. 227 river. 189. 136-140. 2 13-2 19 flood routing. 191. 219-220 normal distribution. Bureau of Reclamation. 183-221 by flood type. 203 extrapolation. 187 partial duration analysis. 28 Frontal storm. 188. 155-171 reservoir. 10. 197 method of moments estimation. 195 outliers. 196 fixed interval analysis. 2. 28 Gage (see Gauge) Gauge bubbler.. 235 Foss Dam flood frequency. 189-197 exceedance probability. 223228 synthetic data. 187 Flynt. 140-144 statistical approach. 191-194 probability density function. 4647 Frequency. 207210 Wakeby distribution. 9-10 statistics (see Frequency) unit hydrograph (see Unit hydrograph) upper limit (see Probable maximum flood) Flood Section. 195 log Pearson type III distribution. 17. 196 historic data adjustment. 179180 Water Supply Papers (WSP). 205. 195 transposition of frequency data. 187 generalized extreme value distribution. 210-212 volume frequency analysis. 187-188 statistical analysis. 2 m-ecinitation stations. 209210 confidence limits. 140. 196-197. 1314. 231-232 Water Data Storage and Retrieval System (WATSTORE). 196 Frontal lifting. 184. 212-213 Gumbel distribution. 202-203 Geologic conditions. 227 successive average lag routing. 16 ‘streamflow. 188 balanced flood hydrograph. 2.. 235 Free atmospheric forcing (FAF). R. 231 Frederick. 12 crest stage. 188-189 distributions. 187 precipitation. 16. 201-202 mixed population. 15 first streim gauge. 188 in study report. 57 Flood study report. 140-171. 204-205 extreme value distribution. 212 three parameter lognormal distribution. 16. 210-211 ungauged area flood frequency. 196-197 partial duration analysis. 223-228 Flow duration analysis. atmospheric. H. 2 1 l-2 12 seasonal analysis. 56 presentation. 11-14 ’ Generalized extreme value distribution. 179 239 .. 183 expected probability adjustment. 192193 regional flood frequency. 118-l 32 Frazier. 232. 192-194 data accuracy. H. A. 196 general trends in frequency distributions. 196 Generalized skew. 228 study report requirements.Muskingum flood routing. 197-204 maximum likelihood estimation. 203 parameter estimation. 21. 14. L. 191. 67-68. 2.

. 114-l 15. 44-46 HMR 52.. 235 Horton. 191.. 151. frequency analysis. 226 Initial loss (see Infiltration) Initial infiltration rate. D. 235 Historical data adjustment. 25-61. 226 Isohyets. 225 Ho. unit hydrograph. 234 Horner. 41-44 storm maximization approach. 203 Historical floods. 44-46. 88-89 Gumbel distribution. 26 envelopment. 8-9. 148 Maximum dewpoint. 22. 8 Lag time. 8 Hazen. 30-58 (see also Hydrometeorological Report Series) definition. 8 Hydrograph. 16. F. atmospheric. 205. 15 Water Supply Paper 1849.. 810. 9. 40-41 Maximum storm. 14-16 statistical analysis of. 8-10. P. II. 236 Heckler. R. W. 59 HMR 53. 67-68. 236 Hansen. 236 Jennings. 109 Interflow. 196 Haan. 226 Manning’s equation. 11-14 indirect discharge measurements. A. 53. 209-210 Hydrograph reconstitution. 111. 148. 49-57. 9. flood. 26. C. 5961 use of. 68. 79. E. National Weather Service. W. 8-9. 109-112. balanced flood. 26. 3140 storm model technique. 36 Maximum flood. 225 Hydrometeorological Branch. 7 Glen Canyon Dam. 46-49. 25 definition.. M.. B. 195 three parameter lognormal distribution. 15.. 60-61 HMR 49.FLOOD HYDROLOGY MANUAL Water Supply Paper 888. 236 Hurricane. 49. 116. G. 197-204 Loss rates. 231 Hershfield. 4143 regionalized studies. 185-l 89 Hydrometeorological analysis. 11. 43-61 Indirect 14-16 discharge measurement data. 67-87 Land use. 53 HMR 43. 25 Jansen. 27 Lognormal distribution. 72-73. 26. 29 Interagency Advisory Committee on Water Data (see Water Resources Council) Interception. 53-57. 55. 23-24 Lifting mechanism. 231 Jarrett. 55-57. 57 HMR 51. 71. 114 hydrographs. 227-228 Maximum likelihood estimation. 13. 60 HMR 50. D. 111 Instability. 22 Gibson Dam. T. 8-9. 225 Hydrometeorological approach to flood estimation. 177 Great Plains dimensionless unit hydrograph.. 235 Johnstown flood. 44-46 HMR 55A. E. 8-9. R. M. Infiltration. 38-40 individual drainage estimates. 159 Manning’s n value.. 231. A. 8 HMR 36.. 197 Maximum precipitation. 60 HMR 37. 9-10. probable (PMP). 195 Log Pearson type III distribution. 25-26. 112-l 14. 116. 191. atmospheric. 1. 189. W. 104-109 Hydrologic data. probable (PMF). 23.. infiltration. 25-26 240 . R. 109112. probable (PMS). 23. 232-234 Hathaway. 184. 31 Hydrometeorological Report (HMR) Series history of.. 11-16 recorded streamflow data.

. K. 27 Precipitation excess. 180. 105 Orographic. 58-59 McDonald. 47 lifting. F. flood hydrograph. 227-228 Probable maximum precipitation (PMP). effective. 116. 810. 140. 4. 3140 storm model technique. 26. 8-9. 223. 24 Nile River. 28 Outliers.. 140. L. W. 18-19 Meter. 234-235 National Water Resources Council (see Water Resources Council) National Weather Service. 59-61 temporal distribution. 16 systematic. 154-155 Moisture. 40-41 Probable maximum storm (PMS). 7 Polygon. 232-236 Hydrometeorological Branch. 140. 50. 5. 236 Puls method. 150.. F. 196197. 16-20 form. 197. 232-235 Min River. 205-208. 31 Technical Paper Series. 2 Precipitable water. first Pharaoh of Egypt. 32-35 defined. 2 Method of moments estimation. 235 Measurement (see Data) Menes. 228-229 report. dam. 192-193 Pseudo-adiabatic chart.. 30-58 spatial distribution. 4143 regionalized studies. 1 Meteorological data. 25. 219-220 analysis by flood type. 144 Rainfall. 50-52 factor. 32 Muskingum flood routing. 63-l 14 Reconnaissance. 2 Probable maximum flood (PMF). 59-61 temporal distribution. 64-66 Rainfall runoff. 11 l-l 12 Mixed population. 196-197. 105 Plum Creek. 144-154 Modified Wilson flood routing. 30-58 (see also Hydrometeorological Report Series) definition. 43 Nearby basins. 144. 191. 26 individual drainage estimates. 201-202 Miller.INDEX (see also Probable maximum precipitation) derivation. 223.. G. 50-52 Normal distribution. 114 hydrographs. 236 Pecos River. V. W. 25-26. Thiessen. J. 105-l 06 Powell.. 114-l 15. 228-229 Reconstitution. 54 Precipitation station. 64-66 Precipitation frequency. 58-59 Probability (see Frequency) Probability density function. 1 Nonorographic. 104-109 Recorded streamflow data. 11-14 envelope curves. 155171 Myers. 188. 16 supplemental (bucket). 228 7-8 201- Parameter estimation. 177-181. 2526 (see also Probable maximum precipitation) derivation. 38-40 estimates. 26 envelopment. 33 Puls. 203 Overtopping. field. 16 Price current meter. A. 7 Planimeter. 202 Partial duration analysis. 30-58 spatial distribution. 16. Price current. 41-44 storm maximization approach. 1 Minimum infiltration rate. 20-24. atmospheric. 27 Moisture maximization. Pearson. 195 Opisometer. 187 241 . 25 definition. 64 Rainfall excess. 56 Precipitation/moisture (P/M) ratio. 9-10. J. 188 Modified Puls flood routing.

140. 232 Transposition of frequency data. 23 Soil Conservation Service (SCS). 227 general considerations. 140-171. 116 Soil groups. 233-235 Seasonal. L. 11-14 Stream routing. 69-71. 40-41 spatial distribution. F. 29 Tracey. 67-68. 227 (SPPalso Routing) Storm frontal. flood. 140-171. 3 l-40 model technique. 154-155 Muskingum flood routing. E. 228 Statistics (see Frequency) Stevens type A-35 recorder. K. 27 Schreiner. 109 Riedel.. 199-201 generalized. 223. 231 Statistical analysis. 80. 227 river. 133-175. 117 Tennessee Valley Authority (TVA). 71-104 Tatum. 10. 2 River routing. 12 Streamflow data recorded. 232 Sierra Nevada dimensionless unit hydrograph. 14-l 6 statistical analysis of. 117 Thiessen polygon. 225 envelope curves. J. 6-7 Supplemental meteorological data (bucket survey). flood. A. 231 Temporal distribution. 80-81. 2-8. 16-20 form. 41-44 Report. F. 226-227 Western Mountain equation. T. 29 transposition. P. 228-229 Reservoir routing. frequency analysis. storm. 16 Synthetic data. 228 Regional flood frequency. 177-181. 105-106 Three parameter lognormal distribution. 156-157. 188 in study report. 16-20 survey form. 227 (see also Routing) Retention.. L.. 76. storm. frequency analysis. 10. F. K. 74. 136-140. 8 isohyetal pattern. 226 Snyder. 212 Synthetic unit hydrograph. R. C. 185-189 Streamflow gauge. 195 Thunderstorm. flood. 140-171. 90-93 Routing. 28 hurricane. 58-59 Tennessee River. 16-20 form. 140-144 Saturation. J. 25 maximization approach. 136-140. 140.. 140. flood study. 7. 18-19 Systematic meteorological data. 232-233 Schwarz. 115116. 210-211 storm. 144-154 modified Wilson flood routing. 87. 94-95 Spatial distribution. 232-234 Rio Grande River. 187-188 S-graph. 82. precipitation. 140144 Sun River. flood. North Fork. 183-221 by flood type. 150. field. 53.. 228-229 Report. storm. 223. 155-171 reservoir. 8-9. 83-84 Skew. 223-228 Reconnaissance report. 162 242 . 11 l-l 12 Southwest Desert dimensionless unit hydrograph. 59-61 survey (bucket). 59-61 Stallings. 11-14 indirect measurement. 37-38 Tuscarawas River. 75. 7. 102-104 Sherman. A. 231 Shipe. 37-38 Study report. 140. 235 Soil conditions. 227 (SPYalso Routing) Rocky Mountains dimensionless unit hydrograph. 63. 133-136 modified Puls flood routing. 9. 227 successive average lag routing. 202-203 Skewness coefficient (see Skew) Snowmelt..FLOOD HYDROLOGY MANUAL Record floods. 80.. F.. 211-212 Regionalized probable maximum precipitation studies. 18-19 Survey (bucket).. 223-228 Successive average lag routing. field reconnaissance. 2. E. 18-19 thunderstorm. 111.

13. 83. 231 Bulletin 17B. 87 Unit hydrograph. 14. 198. 11 l-l 12 area flood frequency. 196 Ware River. 102-104 synthetic. 183-184. 198. 207-210 243 . 75.Ultimate infiltration rate. 71-104 Upper limit flood (see Probable maximum flood) Upper limit rainfall (see Probable maximum precipitation) Urban dimensionless unit hydrograph. 120-l 22 dimensionless. 82. 10. 236 World Meteorological Organization (WMO). 210Ungauged 212 Unit duration. R. 7 Water Resources Council (WRC). 235 Western Mountain snowmelt equation. 96-97 Great Plains. 236 analysis.. 226 Wiesner. 150. 115-116. 207. T. 83. 94-95 Urban. atmospheric. 10. 88-89 Rocky Mountains. 67:87 s-graph. 86-87. 80-81. 83. J. 98-99 Vegetal Volume cover. 197199. 87-102 Coast and Cascade ranges.. 8-9. W. 63104. 41 Weather Bureau. 15 Water Supply Paper 1849. A. 231 Bulletin 17A. 90-93 Sierra Nevada. 68-70. 77. 231 Bulletin 17. 13-14. 80. 86-87. 198. 30 Water Supply Papers (WSP). 23 frequency Wakeby distribution. 69-71. 64. 118. 171. 10. 179 Water Supply Paper 888. 232-235 Weaver. 1 I. 179-180 Water droplet-ice crystal growth. 232 Washita River.. 83-84 Southwest Desert. 79. 231-232 Bulletin 15. 225226 Arbuckle. 80.. 85. 22 Water vapor. 98-99 lag time. 76. 232-233 Zurneorfer. L. 72-73. C. 74. 233 Wilson. 10. 214 Water Data Storage and Retrieval System (WATSTORE). 10. 87. E.

Sign up to vote on this title
UsefulNot useful