Click here for DISCLAIMER Document starts on next page TITLE: Technical Guidance Manual for Performing Wasteload Allocations, Book VI: Design Conditions- Chapter 1: Stream Design Flow for Steady-State Modeling EPA DOCUMENT NUMBER: EPA440/4/86-014 DATE: 1986 ABSTRACT As part of ongoing efforts to keep EPA's technical guidance readily accessible to water quality practitioners, selected publications on Water Quality Modeling and TMDL Guidance available at http://www.epa.gov/waterscience/pc/watqual.html have been enhanced for easier access. This manual describes two techniques to estimate the stream design flow at or above which EPA's two-number aquatic life protection criteria must be met. The two-number water quality criteria are: (1) a criteria maximum concentration (CMC), and (2) a continuous criteria concentration (CCC). Generally, EPA's CMC represents an acute pollutant concentration that should be exceeded no more than once in an average of three years for a period of an hour. Similarly, the CCC represents a chronic pollutant concentration that should be exceeded no more than once in an average of three years for a continuous period of four days. Historically, a majority of States in the USA have required that EPA's aquatic life protection criteria must be met at all flows that are equal to or greater than a critical flow condition of 7-day 10-year low flow (7Q10). In a steady-state modeling framework, the critical flow conditions can be used to estimate the maximum amount of any pollutant that can be discharged without violating water quality criteria (WQC). The critical low flow that is used to define the maximum amount of any pollutant that can be discharged without violating ambient water quality standards (WQS) is called the stream design flow. The design flow represents the required level of treatment or the size of the wastewater treatment facility. This manual recommends a biologically-based stream design flow estimating technique that is strictly consistent with the duration and frequency criteria of EPA's two-number WQC. The biologically-based stream design flows 1B3 and 4B3 are synonymous with the 1-hour, 3-year and 4-day, 3-year duration and frequency criteria of CMC and CCC, respectively. The second estimating technique included in this manual is for a hydrologically-based stream design flow represented in the xQy format, such as 7Q10, 1Q10, 30Q10, etc. This hydrologic flow estimation is consistent with the log Pearson Type III frequency curve approach described in USGS Surface Water Branch Technical Memorandum NO. 79.06, "PROGRAMS AND PLANS - Low-Flow Programs", ------- available online at http://water.usgs.gov/admin/memo/SW/sw79.06.html. These two estimating techniques are implemented using EPA's DFLOW computer program, which is available online at http://www.epa.gov/waterscience/dflow/. Version 3.1 of DFLOW includes a graphical user interface, and directly incorporates the USGS implementation of the log Pearson Type III frequency curve approach and EPA's biologically-based stream design flow technique. The guidance manual includes a comparison of stream design flows of 60 randomly selected US streams, calculated using both approaches. KEYWORDS: Wasteload Allocations, Design Conditions, Design Flow, Steady- State, Models, Water Quality Criteria, Acute, Chronic, Biological, Hydrologic ------- TECHNICAL GUIDANCE MANUAL FOR PERFORMING WASTELOAD ALLOCATION Book VI Design Conditions Chapter 1 Stream Design Flow for Steady-State Modeling September 1986 EPA Publication: 440/4-86-014 ------- Transmittal OFFICE OF WATER September 29, 1986 MEMORANDUM Subject: Technical Guidance Manual for Performing Wasteload Allocations Book VI, Design Conditions: Chapter 1 - Stream Design Flow for Steady-State Modeling From: William A. Whittington, Director Office of Water Regulations and Standards (WH-551) To: Addressees Attached, for national use, is the final version of the Technical Guidance Manual for Performing Wasteload Allocations, Book VI, Design Conditions: Chapter 1 - Stream Design Flow for Steady- State Modeling. This manual replaces the interim stream design flow recommendations included in Appendix D of our Technical Support Document for Water Quality-based Toxics Control, September 1985. We are sending extra copies of this manual to the Regional Waste Load Allocation Coordinators for distribution to the States to use in conducting waste load allocations. If you have any questions or desire additional information please contact Tim. S. Stuart, Chief, Monitoring Branch, Monitoring and Data Support Division (WH-553) on (FTS) 382-7074. Attachment Addressees: Regional Water Management Division Directors Regional Environmental Services Division Directors Regional Wasteload Allocation Coordinators ------- ACKNOWLEDGEMENT The contents of this section have been removed to comply with current EPA practice. ------- Table of Contents SECTION 1. INTRODUCTION 1-1 1.1 Purpose 1-1 1. 2 Background 1-1 1. 3 Scope 1-3 SECTION 2. HYDROLOGICALLY-BASED DESIGN FLOW 2-1 2 .1 Introduction 2-1 2 .2 Rationale 2-2 2 . 3 Example Cases 2-2 Table 2-1. Hydrologically-based design flows (ft3/sec) for 60 streams 2-3 SECTION 3. BIOLOGICALLY-BASED DESIGN FLOW 3-1 3 .1 Introduction 3-1 3.1.1 Exceedances and Excursions 3-2 3.1.2 Features of Calculation 3-4 Figure 3-1: Illustration of biologically-based design flow 3-5 3 . 2 Procedure 3-6 3 .3 Rationale 3-7 3 .4 Example Cases 3-8 Table 3-1. Biologically-based design flows (ft3/sec) for 63 rivers 3-9 SECTION 4 . COMPARISON OF THE TWO METHODS 4-1 4.1 Design Flows 4-1 Table 4-1. Comparison of 1Q10 and 7Q10 with 1-day 3-yr and 4-day 3-yr low flows (all flows in ft3/sec) 4-2 4 . 2 Excursions 4-3 4.3 Comparison of the Two Methods 4-4 Table 4-2. Comparison of number of excursions of 1Q10 and 7Q10 with number of excursions of 1-day 3-yr and 4-day 3-yr design flows 4-5 SECTION 5. RECOMMENDATIONS 5-1 SECTION 6 . REFERENCES 6-1 APPENDIX A. CALCULATION OF HYDROLOGICALLY-BASED DESIGN FLOWS ... A-l ------- APPENDIX B. AN EXAMPLE USE OF DFLOW FOR AMMONIA DISCHARGES FROM POTWs B-l APPENDIX C. CALCULATION OF BIOLOGICALLY-BASED DESIGN FLOWS C-l APPENDIX D. DFLOW 2.0 USER'S GUIDE (OMITTED) APPENDIX E. QUESTIONS AND ANSWERS CONCERNING THE BIOLOGICALLY-BASED METHOD E-l ------- GLOSSARY IBS - 1-hour 3-year duration and frequency criteria of Criterion Maximum Concentration. 4B3 - 4-day 3-year duration and frequency criteria of Criterion Continuous Concentration. 7Q10 - EPA's aquatic life protection criteria to meet all flows equal to or greater than critical flow condition of 7-day 10-year low flow. xQy - Hydrologically-based design flows expressed as x-day average low flows whose return period is y years. For example, 1Q4 is the daily low flow that is exceeded once every four years. Other xQy values commonly encountered are 1Q10, 4Q3, 7Q5, and 7Q10. xBy - Biological design flow parameter defined in DFLOW 3. Averaging period - Specified in national water quality criteria are one hour for the CMC and four days for the CCC. The primary use of the averaging periods in criteria is for averaging ambient concentrations of pollutants in receiving waters in order that the averages can be compared to the CMC and CCC to identify "exceedances" i.e., one-hour average concentrations that exceed the CMC and four-day average concentrations that exceed the CCC. Biologically-based stream design flow - Design flow based on the averaging periods and frequencies specified in water quality criteria for individual pollutants and whole effluents. Criteria - Descriptive factors taken into account by EPA in setting standards for various pollutants. These factors are used to determine limits on allowable concentration levels, and to limit the number of violations per year. When issued by EPA, the criteria provide guidance to the states on how to establish their standards.1 Criterion Continuous Concentration (CCC) - The CCC is the 4-day average concentration of a pollutant in ambient water that should not be exceeded more than once every three years on the average. Criterion Maximum Concentration (CMC) - The one-hour average concentration in ambient water should not exceed the CMC more than once every three years on the average. Design flow - A flow rate resulting from an applied waste load allocation process, used to ensure ambient water quality compliance. DFLOW - The DFLOW (Design FLOW) program was originally developed by EPA to support design flow analysis as described in this manual. At the time of original publication, Version 2.0 of DFLOW was available to support this analysis on personal computers. The latest version of DFLOW at the time of enhancement of this manual was Version 3.1, which combines a graphical user interface with the USGS implementation of the Log Pearson Type III calculation. Additional information may be available at http://www.epa.gov/waterscience/dflow/. Duration - The time during which something exists or lasts EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms" http://www.epa.gov/OCEPAterms/cterms.html March 8, 2006. 2 Merriam-Webster Online Dictionary, http://www.m-w.com/dictionary/duration Accessed April 14, 2006. ------- Dynamic models - Preferred for the application of aquatic life criteria in order to make best use of the specified concentrations, durations, and frequencies. Use of aquatic life criteria for developing water quality-based permit limits and for designing waste treatment facilities requires the selection of an appropriate wasteload allocation model. Effluent - Wastewater--treated or untreated--that flows out of a treatment plant, sewer, or industrial outfall. Generally refers to wastes discharged into surface waters.3 Exceedance - Violation of the pollutant levels and average concentration frequencies that are permitted by environmental protection standards Excursion - An unfavorable conditions, e.g. low flow or noncompliant pollutant concentrations Flow Averaging Period - Monitor of water flow for a specific number of days, usually 30, to determine pollutant concentrations. Frequency - The number of repetitions of a particular event in a unit of time. Gaging stations - The locations at which measurements are recorded, generally hydraulic or hydrologic in nature, usually referring to stream flow gages or rain gages. Harmonic Mean - Set of numbers that is the reciprocal of the arithmetic mean of the reciprocals of the numbers. Hydrologic Flow - The characteristic behaviour and the total quantity of water involved in a drainage basin, determined by measuring such quantities as rainfall, surface and subsurface storage and flow, and evapotranspiration.(Source: BJGEO)4 Hydrologically-based design flow - Design flow calculated solely from the hydrologic record. Log Pearson Type III - Statistical analysis, recommended by the US Water Resources Council Bulletin #17B, used to gage natural flood data.4 Model - Either a steady-state or dynamic simulation that uses hydraulic and biological criteria to design regulatory compliant waste loads. Non-exceedance - Relating to steady-state models as flow rates that are less than design flows. Percentile - One of a set of points on a scale arrived at by dividing a group into parts in order of magnitude.5 Pollutant - Generally, any substance introduced into the environment that adversely affects the usefulness of a resource or the health of humans, animals, or ecosystems.6 EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms" http://www.epa.gov/OCEPAterms/. March 8,2006. EPA. "Terminology Reference System: Hydrologic Flow" http://iaspub.epa.gov/trs/trs proc qry.navigate term?p term id=26386&p term cd=TERMDIS May 2, 2006. Texas Department of Transportation. "Hydraulic Design Manual, Section 10. Statistical Analysis of Stream Gauge Data". http://manuals.dot.state.tx.us/dynaweb/colbridg/hyd/@Generic BookTextView/14106;cs=defau lt;ts=default;pt=14873 March 2004. The American Heritage® Dictionary of the English Language,4th Ed. "Percentile." Accessed at www.Dictionary.com July 3, 2006. ------- Receiving waters - Bodies of water into which effluents are discharged. Return Period - Annual x-day average low flow repeated in y-years. Site-specific criteria - Regulatory concentrations, parameters, or frequencies that are exclusive to proximity or spatially exclusive. Steady-state models - Models that assume a constant average flow rate. Toxicity - The degree to which a substance or mixture of substances can harm humans or animals. Acute toxicity involves harmful effects in an organism through a single or short-term exposure. Chronic toxicity is the ability of a substance or mixture of substances to cause harmful effects over an extended period, usually upon repeated or continuous exposure sometimes lasting for the entire life of the exposed organism.6 Two-number water quality criterion - Criterion Continuous and Criterion Maximum Concentrations (CCC, CMC) that are used as the basis for the both hydrologically- and biologically-based design flows Variability - Reference to the consistency of pollutant concentrations in effluent and ambient waters. Wasteload allocation - 1.) The maximum load of pollutants each discharger of waste is allowed to release into a particular waterway. Discharge limits are usually required for each specific water quality criterion being, or expected to be, violated. 2.) The portion of a stream's total assimilative capacity assigned to an individual discharge.7 Water quality criteria (WQC) - Levels of water quality expected to render a body of water suitable for its designated use. Criteria are based on specific levels of pollutants that would make the water harmful if used for drinking, swimming, farming, fish production, or industrial processes. Water quality-based permit - A permit with an effluent limit more stringent than one based on technology performance. Such limits may be necessary to protect the designated use of receiving waters (e.g. recreation, irrigation, industry or water supply). EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms". http://www.epa.gov/OCEPAterms/pterms.html March 8, 2006. EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms" http://www.epa.gov/OCEPAterms/tterms.html March 8, 2006. EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms" http://www.epa.gov/OCEPAterms/wterms.html March 8, 2006. EPA. "Terms of Environment: Glossary, Abbreviations, and Acronyms" http://www.epa.gov/OCEPAterms/wterms.html March 8, 2006. ------- Section 1. Introduction 1.1 Purpose The purpose of this guidance is to describe and compare two methods that can be used to calculate stream design flows for any pollutant or effluent for which a two-number water quality criterion (WOC) for the protection of aquatic life is available. The two methods described are: 1. The hydrologically-based design flow method recommended for interim use in the Technical Support document for Water Quality- based Toxics Control (1); and 2. A biologically-based design flow method that was developed by the Office of Research and Development of the U.S. EPA. 1.2 Background National water quality criteria for aquatic life (2) are derived on the basis of the best available biological, ecological and toxicological information concerning the effects of pollutants on aquatic organisms and their uses (3,4). To account for local conditions, site-specific criteria may he derived whenever adequately justified (4). In addition, criteria may be derived from the results of toxicity tests on whole effluents (1). National, site-specific, and effluent toxicity criteria specify concentrations of pollutants, durations of averaging periods, and frequencies of allowed exceedances. If these criteria are to achieve their intended purpose, decisions concerning not only their derivation, but also their use, must be based on the biological, ecological, and toxicological characteristics of aquatic organisms and ecosystems, and their uses, whenever possible. National, site-specific, and effluent toxicity criteria are expressed as two concentrations, rather than one, so that the ------- criteria can more accurately reflect toxicological and practical realities (1-4) : a. The lower concentration is called the Criterion Continuous Concentration (CCC). The CCC is the 4-day average* concentration of a pollutant in ambient water that should not be exceeded more than once every three years on the average. b. The higher concentration is called the Criterion Maximum Concentration (CMC). The one-hour average concentration in ambient water should not exceed the CMC more than once every three years on the average. Use of aquatic life criteria for developing water quality-based permit limits and for designing waste treatment facilities requires the selection of an appropriate wasteload allocation model. Dynamic models are preferred for the application of aquatic life criteria in order to make best use of the specified concentrations, durations, and frequencies (2). If none of the dynamic models can be used, then an alternative is steady-state modeling. Because steady-state modeling is based on various simplifying assumptions, it is less complex, and may be less realistic, than dynamic modeling. An important step in the application of steady-state modeling to stream is the selection of the design flow. One way of using the CCC and the CMC in steady-state modeling requires calculation of the two design flows (i.e., a CCC design flow and a CMC design flow). Whether the CCC and its design flow or the CMC: and its design flow is more restrictive, and therefore controlling, must be determined individually for each pollutant of Although a 4-day averaging period should be used for the CCC in most situations, an averaging period as long as 30 days may be used in situations involving POTWs designed to remove ammonia when low variability of effluent pollutant concentration and resultant concentrations in receiving waters can be demonstrated. In cases where low variability can be demonstrated, longer averaging periods for the ammonia CCC (e.g., a 30-day averaging period) would be acceptable because the magnitudes and durations of excursions above the CCC would be sufficiently limited (5). ------- concern in each effluent because the CCC and CMC are pollutant- specific, whereas the two design flows are specific to the receiving waters. Wasteload allocation modeling for stream usually uses flow data obtained from the United States Geological Survey gaging stations. If sufficient flow data are not available for a stream of interest, data must be extrapolated from other streams having hydrologic characteristics similar to those of the stream of interest. This guidance is limited to (a) describing two methods that can be used for calculating stream design flows for any pollutant or effluent for which a two-number aquatic life water quality criterion is available, and (b) making recommendations concerning the use of these methods in steady-state modeling. The water quality criterion for dissolved oxygen was revised very recently and the assessment of the appropriate design flow for dissolved oxygen modeling has not yet been completed. Therefore, the state-specified design flows that traditionally have been used for conventional pollutants should not be affected by this guidance. ------- State-specified design flows necessarily preempt any design flow that is recommended in this guidance unless the state chooses to use either of these two methods. The choice of design flows for the protection of human health has been discussed in the Technical Support Document for Water Quality-based Toxics Control (1). Aquatic life criteria of some pollutants are affected by environmental variables such as water temperature, pH, and hardness. In addition to the design flow, such other stream variables as pH and temperature might increase or decrease the allowable in-stream concentrations of some pollutants (e.g., ammonia). The need to consider other variables when determining the design flow for those pollutants should be emphasized. This document will provide guidance for the calculation of design flow; pH, temperature, and hardness will likely be addressed later. ------- SECTION 2. Hydrologically-Based Design Flow 2 .1 Introduction The purpose of this section is to describe the hydrologically- based design flow calculation method and provide some examples of its use. The Technical Support Document for Water Quality-based Toxics Control (1) provides Agency guidance on control of both generic and pollutant-specific toxicity and recommended interim use of the hydrologically-based method. In addition, the Agency also recommended (1, 2) that the frequencies of allowed exceedances and the durations of the averaging periods specified in aquatic life criteria should not be used directly to calculate steady- state, design flows using an extreme value analysis. For example, if a criterion specifies that the four-day average concentration should not exceed a particular value more than once every three years on the average, this should be interpreted as implying that the 4Q3 low flow is appropriate for use as the design flow. Because a procedure had not been developed for calculating design flow based on the durations and frequencies specified in aquatic life criteria, the U.S. EPA recommended interim use of the 1Q5 and 1Q10 low flows as the CMC design flow and the 7Q5 and 7Q10 low flows as the CCC design flow for unstressed and stressed systems, respectively (1). Further consideration of stress placed on aquatic ecosystems resulting from exceedances of water quality criteria indicates that there is little justification for different design flows for unstressed and stressed system. All ecosystems have been changed as a result of man's activities. These changes have resulted in stress being placed on the ecosystem before a pollutant stress. In addition, it is not possible to predict 2-1 ------- the degree of pollutant stress when one considers both the timing and variability of flows, effluent discharges, and ecosystem sensitivity and resilience. 2.2 Rationale The following provides a rationale for the hydrologically-based design flow calculation method: • About half of the states in the nation use 7Q10 as the design low flow. • The log-Pearson Type III flow estimating technique of other extreme value analytical techniques that are used to calculate flow statistics from daily flow data are consistent with past engineering and statistical practice. • Most users are familiar with the log-Pearson Type III flow estimating procedure and the USGS provides technical support for this technique. • Analyses of 60 rivers indicate that, on the average, the biologically-based CMC and CCC design flows are nearly equal to the 1Q10 and the 7Q10 low flows. 2.3 Example Cases In order to illustrate the calculation of hydrologically-based design flows, sixty rivers with flows of various magnitudes and variabilities were chosen from around the country. The 1Q10 and 7Q10 low flow of the sixty rivers are presented in Table 2-1. The list of rivers in this table is arranged in increasing magnitude of the 7Q10 low flows. The estimates of the 1Q10 and 7Q10 low flows were made using the USGS daily flow database and the FLOSTAT program (6) which employs the log-Pearson Type III technique. 2-2 ------- The estimates of 1Q10 and 7Q10 low flows could have been made using EPA-ORD's DFLOW program, which uses a simplified version of the log-Pearson Type III method. The simplified version of the log- Pearson Type III estimating technique for any xQy design flow is presented in Appendix A. Although the Log-Pearson Type III is in general use it should be recognized that there are other distributions that may be more appropriate to use on a case-by-case basis. The hydrologically-based design flow for ammonia is discussed in Appendix B. Analyses of the 1Q10 and 7Q10 low flow in Table 2-1 indicate that the mean of the ratios of 7Q10 to 1Q10 is 1.3. The median of the ratios is 1.1, whereas the range of the ratios is 1.0 to 3.85. Thus, 7Q10 low flows are generally 10 to 30% greater than the corresponding 1Q10 low flows, although in one case the 7Q10 is 3.85 tines greater than the corresponding 1Q10. Table 2-1. Hydrologically-based design flows (ft3/sec) for 60 streams Station ID 01657000 02092500 06026000 12449600 05522000 09490800 14372500 05381000 10291500 05585000 12321500 01111500 River Name Bull Run Trent Birch Cr Beaver Cr Iroquois N Fk White E FK Illinois Black Buckeye LaMoine Boundary Cr Branch State VA NC MT WA IN AZ OR WI CA IL ID RI Period of Record 1951-82 1951-82 1946-77 1960-78 1949-78 1966-78 1942-03 1905-83 1911-78 1921-83 1928-84 1940-82 CV* 4 .48 1.77 1.32 1.77 1.33 1.24 2 .03 2 .51 1.30 1.99 1.65 1.16 Design flow (ft 3/sec) 1Q10 0.3 1.4 1.7 2 .4 3 .4 4 .8 6 .4 5.5 7.1 9.3 11.7 8 .8 7Q10 0.4 1.6 2 .4 3 .2 3 .9 5.3 6 .7 6 .7 7 .7 9.9 13 .1 13 .3 7Q10 1Q10 1.33 1.14 1.41 1.22 1.15 1.10 1.05 1.22 1.08 1.06 1.12 1.51 2-3 ------- Table 2-1 (continued) Station ID River Name Period of QV* State Record Design flow (ft 3/sec) 1Q10 7Q10 7Q10 1Q10 02138500 05053000 02083000 01196500 02133500 06280300 09149500 02296750 07018500 02217530 01481000 09497500 01144000 01600000 09359500 01403060 02413500 01421000 07298500 07013300 01531000 07096000 09070000 01011000 03528000 13023000 02424000 05515500 02490500 01315500 01610000 05386000 02369000 07378500 06465500 02135000 08110200 02076000 03455000 05333500 06287000 03107500 Linville Sheyenne Fishing Cr Quinnipiac Drowning Cr S ho shone Uncompahgre Peace Rig Middle Oconee Brandywine Salt White N Br Potomac Animas Raritan L Tallapoosa E B Delaware Big Sunflower Meramec Chemung Arkansas Eagle Allegash Clinch Greys Cahaba Kankakee Bouge Chitto Hudson Potomac Root Shoal Amite Niobrara Little Pee Dee Brazos Dan French Broad St. Croix Bighorn Beaver NC ND NC CT NC WY CO FL MO GA PA AZ VT MD CO NJ AL NY MS MO NY CO CO ME TN WY AL IN MS NY WV MN FL LA NE SC TX VA TN WI MT PA 1922- 1951- 1927- 1931- 1940- 1957- 1939- 1931- 1922- 1902- 1912- 1925- 1915- 1939- 1946- 1904- 1940- 1915- 1936- 1923- 1915- 1901- 1947- 1932- 1919- 1937- 1902- 1926- 1945- 1908- 1939- 1938- 1939- 1939- 1939- 1942- 1966- 1924- 1901- 1914- 1935- 1957- 84 81 82 84 78 84 80 84 84 84 84 80 84 83 56 83 51 78 80 78 78 81 80 03 78 83 78 78 81 78 83 61 82 83 83 78 70 52 78 81 79 83 1 . 2 . 1. 1. 0. 1 . 0. 1. 2 . 1 . 1. 2 . 1 . 1 . 1. 1. 1 . 1 . 1. 2 . 1 . 1 . 1. 1. 1 . 1 . 2 . 0. 1 . 1 . 1. 1. 0. 1 . 0. 0. 1 . 1 . 0. 0. 0. 1. .74 .10 .48 .02 .80 .54 .86 .54 .16 .37 .17 .05 .43 .42 .56 .64 .31 .41 .42 .41 .91 .12 .36 .39 .55 .16 .07 .48 .89 .10 .48 .65 .95 .98 .59 .94 .48 .25 .93 .61 .82 .10 13 . 15. 17. 17. 38 . 41. 35. 49. 46 . 49. 61. 64 . 75. 54 . 54 . 54 . 72 . 80. 89. 88 . 89. 107 116 124 120 122 151 179 188 207 209 229 280 298 160 306 311 329 473 505 327 571 4 9 0 5 8 8 6 0 4 4 4 6 3 7 8 2 7 8 4 8 7 .9 .9 .5 .7 .9 .9 .0 .6 .7 .6 .7 .1 .1 .9 .7 .6 .6 .6 .9 .1 .3 16 13 19 32 43 46 50 155 55 57 67 68 85 61 62 67 8 . 89 91 92 97 126 131 134 135 144 156 184 191 211 220 245 291 303 322 322 344 387 532 536 557 594 .4 .3 .4 .3 .4 .8 .8 .3 .3 .4 .2 .7 .2 .6 .3 .1 3 .7 .9 .2 .5 .1 .0 .1 .2 .5 .4 .3 .6 .0 .7 .6 .4 .4 .0 .4 .9 .3 .2 .0 .0 .2 1.22 1.15 1.14 1.85 1.12 1.12 1.43 1.13 1.19 1.16 1.09 1.06 1.13 1.13 1.15 1.24 1.21 1 .11 1.03 1.05 1.09 1.17 1.12 1.38 1.05 1.13 1.03 1.33 1.02 1.02 1.05 1.07 1.04 1.02 2 .00 1.09 1 .11 1.18 1.12 1.06 1.70 1.04 2-1 ------- Table 2-1 (continued). Station ID 13341000 07341500 02350500 01536500 01100000 14233430 River Name N P Clearwater Red Flint Susquehanna Merrimack Cowlitz State ID AR GA PA MA WA Period of Record 1927-68 1928-81 1930-58 1901-83 1924-83 1968-78 CV* 1.16 1.41 1.00 1.34 1.01 0.93 Design flow (ft 3/sec) 1Q10 529.2 691.0 207.8 782 .0 270.2 901.5 7Q10 648 .6 769.2 799.8 814 .3 929.3 958 .7 7Q10 1Q10 1.23 1.11 3 .85 1.04 3 .44 1.07 *CV = Coefficient of Variation 2-2 ------- SECTION 3. Biologically-Based Design Flow 3 .1 Introduction The purpose of this section is to describe the biologically- based design flow calculation method and provide some examples of its use. This method was developed by the Office of Research and Development of the U.S. EPA in order to provide a way of directly using EPA's two-number aquatic life water quality criteria (WQC) for individual pollutants and whole effluents to calculate the design flow for performing a wasteload allocation using steady- state modeling. The two-number WQC are in the intensity-duration- frequency format, in that they specify intensity as criteria concentrations, duration as averaging periods, and frequency as average frequency of allowed excursions. Because the flow of, and concentrations of pollutants in, effluents and stream are easily considered in terms of intensity, duration, and frequency, use of this format for expressing WQC allows a direct application to effluents and streams. Because steady-state modeling assumes that the composition and flow of the effluent of concern is constant, the ambient (instream) concentration of a pollutant can be considered to be inversely proportional to stream flow. Thus by applying a specified averaging period and frequency to a record of the historical flow of the stream of concern, the design flow can be calculated as the highest flow that will not cause exceedances to occur more often than allowed by the specified average frequency, based on historical data. The allowed exceedances are intended to be small enough and far enough apart, on the average, that the resulting small stresses on aquatic organisms will not cause unacceptable effects, except in those cases when a drought itself would cause unacceptable effects. The averaging periods specified in national water quality criteria are one hour for the CMC and four days for the CCC. The primary use of the averaging periods in criteria is for averaging ambient concentrations of pollutants in receiving waters in order that the averages can be compared to the CMC and CCC to identify 3-1 ------- "exceedances" i.e., one-hour average concentrations that exceed the CMC and four-day average concentrations that exceed the CCC. However, in steady-state modeling, flow is averaged over a given period to identify "non-exceedances", i.e., average flows that are below a specified flow. 3.1.1 Exceedances and Excursions Use of the term "exceedance" and "non-exceedance" neither of which are in the dictionary, can be a cause of confusion. Water quality criteria are usually expressed as upper limits on concentrations in ambient water and the periods of concern are when the ambient concentration exceeds a criterion concentration, i.e., when there is an exceedance. In steady-state modeling, the averaging is of flows, not concentrations. Because a low flow results in a high pollutant concentration, the period of concern for flow is when the flow is less than the design flow, i.e., when there is non-exceedance of a given flow. A non-exceedance of a design flow corresponds to an exceedance of a criterion. Use of the non-directional term "excursion", which is in the dictionary, avoids this confusion. Use of the term "excursion" also avoids the problem that some water quality criteria, such as those for dissolved oxygen and low pH, must be stated as lower limits, not upper limits. An exceedance of a dissolved oxygen criterion is favorable, not unfavorable. "Excursions", in this guidance manual, will henceforth be used to imply 3-2 ------- an unfavorable condition, e.g., a low flow or a pollutant concentration above an upper limit or below a lower limit. The national water quality criteria specify that, if R is the calculated number of excursions occurring in a period of S years, then S/R should be equal to or greater than 3 years. Most excursions will be small and most aquatic ecosystems will probably recover from the resulting minor stress in less than three years. However, the three years is meant to be longer than the average recovery period so that ecosystems cannot be in a constant state of recovery even if excursions are evenly spaced over time. Although 3 years appears to be appropriate for small excursions that are somewhat isolated, it appears to be excessively long when many excursions occur in a short period of time, such as would be caused by a drought. Droughts are rare events, characterized by long periods of low flow and should not be allowed to unnecessarily lower design flows. Although droughts do severely stress aquatic ecosystems, both directly, because of low flow, and indirectly, because of the resulting high concentrations or pollutants, many ecosystems apparently recover from severe stresses in more than 5, but less than 10 years (1). Because it is not adequately protective to keep ecosystems in a constant state of recovery 15 years seem like an appropriate stress-free period of time, on the average to allow after a severe stress caused by a drought situation. Because three years are allowed for each excursion on the average, counting no more than 5 excursions for any low flow period will 3-3 ------- provide no more than 15 years, on the average, for severe stresses caused by droughts. Thus, for each low flow period, the number of excursions cannot be less than 1.0 or greater than 5.0. The maximum duration of a low-flow period was set at 120 days because it is not too uncommon for excursions to occur within 120 days of each other, whereas it is very rare for excursions to occur during days 121 to 240 after the beginning of a low-flow period. 3.1.2 Features of Calculation Figure 3-1 illustrates the features of the biologically-based design flow calculation method. Intervals a-b and c-d are excursion periods and each day in these intervals is part of an average flow that is below the design flow. The number of excursions in an excursion period is calculated as the number of days in the excursion period divided by the duration (in days) of the averaging period (e.g., 1 day for the CMC and 4 days for CCC). A low-flow period is defined as one or more excursion periods occurring within a 120-day interval. As discussed above, if the calculated number of excursions that occur in a 120-day low-flow period is greater than 5, the number is set at 5 for the purposes of calculating the design flow. Because biologically-based design flows are based on the averaging periods and frequencies specified in water quality criteria for individual pollutants and whole effluents, they can be based on the available biological, ecological, and toxicological information concerning the stresses that aquatic organisms, ecosystems, and their uses can tolerate. The biologically-based calculation method is flexible enough to make full use 3-4 ------- Exclusion Time Figure 3-1: Illustration of biologically-based design flow 3-5 ------- The CMC and CCC design flows are calculated in almost the same manner. The differences result from the fact that the CMC is expressed as a one-hour average, whereas the CCC is expressed as a four-day average. However, the flow records that are available consist of one-day average flows. For streams with naturally occurring low flows, calculation of the CMC design flow from one- day averages, rather than one-hour averages, should be reasonably acceptable because naturally occurring low flows of receiving streams are usually very similar from one hour to the next. In regulated streams, such as those affected by hydroelectric or irrigation projects, hour-to-hour variation of low flows could be significant and in those situations, use of hourly values, when available, is appropriate. Both the pollutant concentrations and the flows of most effluents are expected to change much more from one hour to the next than the naturally occurring flows of streams. 3.3 Rationale The following provides a rationale for the biologically-based design flow calculation method: * It allows the use of the new two-number WQC for aquatic life in the calculation of design flow. If water quality criteria for aquatic life are to achieve their intended purpose, decisions concerning their derivation and use should be based on the biological, ecological, and toxicological characteristics of aquatic organisms and ecosystems and their uses whenever possible. * It takes into account all excursions in the flow record. * It provides the necessary design flow directly without requiring any design flow statistics in the xQy format. * It is flexible enough so that any averaging period and frequency selected for particular pollutants, effluents, or site-specific criteria can be used directly in design flow calculations. 3-7 ------- 3 .4 Example Cases The sixty flow records that were analyzed using the hydrologically-based method (see Table 2-1) were also analyzed using the biologically-based design flow method. The CMC design flow was calculated for a 1-day averaging period and the CCC design flow was calculated using the 4-day averaging period. Both were calculated using a frequency of once every three years on the average. Table 3-1 presents biologically-based design flows for these sixty rivers. In addition to the hydrologically-based design flows, Table B- 1 in Appendix B also includes biologically-based CMC and CCC design flows for 13 streams for 30-day averaging periods and a frequency of once every three years on the average. The purpose of the biologically-based design flows for ammonia (5) in Appendix B is to illustrate how this method might be used for site-specific and pollutant-specific situations where the durations and frequencies in aquatic life criteria might be different from those specified in national two-number aquatic life criteria. Analyses of the 1-day 3-year and the 4-day 3-year low flows in Table 3-1 indicate that the mean ratio of the 4-day 3-year low flows to the corresponding 1-day 3-year low flows is 1.23. The median of the ratios is 1.11, whereas the range of the ratios is 1.0 to 2.81. Thus, 4-day 3-year low flows are generally 11 to 23% greater than the corresponding 1-day 3-year low flows, although in one case, the 4-day 3-year low flow is 2.91 times greater than the corresponding 1-day 3year low flow. ------- Table 3-1. Biologically-based design flows (ft3/sec) for 63 rivers Station ID 01657000 02092500 06026000 12449600 05522000 09490800 14372500 05381000 10291500 05585000 12321500 01111500 02138500 05053000 or 03059000 02083000 01196500 02133500 06280300 09149500 02296750 07018500 02217530 01600000 09359500 01403060 01481000 09497500 01144000 02413500 01421000 07288500 07013300 01531000 07096000 09070000 01011000 03528000 13023000 02424000 River Name Bull Run Trent Birch Cr Beaver Cr Iroquois N Fk White E FK Illinois Black Buckeye LaMoine Boundary Cr Branch Linville Sheyenne Fishing Cr Quinnipiac Drowning Cr S ho shone Uncompahgre Peace Big Middle Oconee N Br Potomac Animas Raritan Brandy wine Salt White L Tallapoosa E B Delaware Big Sunflower Meramec Chemung Arkansas Eagle Allegash Clinch Greys Cahaba State VA NC MT WA IN AZ OR WI CA IL ID RI NC ND NC CT NC WY CO FL MO GA MD CO NJ PA AZ VT AL NY MS MO NY CO CO ME TN WY AL Period of Record 1951-82 1951-82 1946-77 1960-78 1949-78 1966-78 1942-03 1905-83 1911-78 1921-83 1928-84 1940-82 1922-84 1951-81 1927-82 1931-84 1940-78 1957-84 1939-80 1931-84 1922-84 1902-84 1939-83 1946-56 1904-83 1912-84 1925-80 1915-84 1940-51 1915-78 1936-80 1923-78 1915-78 1901-81 1947-80 1932-03 1919-78 1937-83 1902-78 CV* 4 .48 1 .77 1 .32 1.77 1.33 1 .24 2 .03 2 .51 1.30 1 .99 1 .65 1.16 1.74 2 .10 1.48 1 .02 0 .80 1.54 0.86 1 .54 2 .16 1 .37 1.42 1.56 1.64 1 .17 2 .05 1.43 1.33 1.41 1 .42 2 .41 1.91 1.12 1.36 1 .39 1.55 1.16 2 .07 Design flow (itybsec) 1-day 3 -year 0.20 1.40 1. 70 2.80 2.40 4 . 80 5. 80 5.00 7.00 8 . 90 12 .00 10.00 13 .00 15.40 12 .00 14 .90 33 .90 42 .90 29.90 48 .00 45 .00 33 .00 42 .90 60.00 46 .90 55 .80 63 .00 75.90 57.90 82 .00 82 .70 89 .90 85.70 89.90 120.00 134 .00 127.70 124 .80 122 .80 4 -day 3 -year 0.40 1. 60 2 .40 3.40 3.00 5.30 6 . 90 6.10 7.20 9.40 13 . 00 13.20 15.00 17.60 13.50 34 . 00 36 .20 45.80 49.00 55.20 51. 50 45. 70 49.00 61.10 53.60 59.30 59. 50 86.00 70.20 91.40 85.40 92 . 70 92.50 114.00 126.00 138 .40 132.20 135.80 149. 80 ccc CMC 08 2 .00 1 .14 1 .41 1.21 1.25 1 .10 1 .19 1.22 1.03 1 .06 1 .08 1.32 1.15 1.14 1.13 2 .25 1 .07 1.07 1.26 1 .15 1 .14 1 .38 1.17 1.02 1.14 1 .06 1 .10 1.13 1.21 1.11 1 .03 1 .03 1.08 1.27 1.05 1 .03 1.04 1.09 1 .22 ------- Table 3-1 (continued). Station ID 05515500 02490500 01315500 01610000 05386000 02369000 07378500 06465500 02135000 08110200 02076000 03455000 05333500 06287000 03107500 13341000 07341500 02350500 01100000 14233430 River Name Kankakee Bouge Chitto Hudson Potomac Root Shoal Amite Nebraska Little Pee Dee Brazos Dan French Broad St . Croix Bighorn Beaver N P Clearwater Red Flint Merrimack Cowl it z State IN MS NY WV MN FL LA NE SC TX VA TN WI MT PA ID AR GA MA WA Period of Record 1926-78 1945-81 1908-78 1939-83 1938-61 1939-82 1939-83 1939-83 1942-78 1966-70 1924-52 1901-78 1914-81 1935-79 1957-83 1927-68 1928-81 1930-58 1924-83 1968-78 CV* 0.48 1.89 1.10 1.48 1.65 0. 95 1.98 0.59 0. 94 1.48 1.25 0.93 0.61 0.82 1.10 1.16 1.41 1.00 1.01 0.93 Design flow (ft /sec) 1-day 3-year 167.60 187.50 170.00 22.20 239.30 270.50 282.10 199.70 298.70 277.70 321.60 494 .30 477.50 364 .00 539.90 429.60 537.40 262 .50 284.00 934 .70 4-day 3-year 174.20 189.60 191.90 219.60 23937.00 2860.00 295.50 304.30 298.90 305.30 380.40 535.50 508.50 520.20 557.50 613.00 603 .30 731.00 797.30 959.90 ccc CMC08 1.04 1.13 1.13 1.09 1.00 1.06 1.05 1.52 1.00 1.10 1.18 1.08 1.06 1.43 1.07 1.31 1.12 2 .78 2.81 1.03 = coefficient of variation 3-10 ------- For further clarification of the biologically-based method, refer to Appendix E, Questions and Answers. 3-11 ------- of special averaging periods and frequencies that might be selected for specific pollutants (e.g., ammonia) or in site-specific criteria. This method is empirical, not statistical, because it deals with the actual flow record itself, not with a statistical distribution that is intended to describe the flow record. In addition, this method provides an understanding of how many excursions of the CCC or CMC are likely to occur, and during what time of the year, based on actual historical flow data. Thus, it is possible to examine the pattern and magnitudes of what would have been historical excursions. This method makes it clear that criteria concentrations should not be interpreted as values that are never to be exceeded "at any time or place" in the receiving waters. An understanding of what level of protection actually is provided should aid in the use of criteria. 3 .2 Procedure Although the calculation procedure described in Appendix C might look complicated, it merely consists of a sequence of steps that are quite simple. Because flow records usually consist of daily flows for 20 to 80 years, manual calculation of design flow is very time-consuming. The DFLOW computer program (Appendix D (OMITTED) - DFLOW 2.0 has been superseded by newer versions. The current versions of DFLOW and its documentation are available online at http://www.epa.gov/waterscience/dflow.) will calculate biologically-based design flows and display the dates, durations, and magnitudes of the excursions within each low flow period. 3-6 ------- SECTION 4. COMPARISON OF THE TWO METHODS 4 .1 Design Flows Table 4-1 shows the biologically-based 1-day 3-year low flows and the hydrologically-based 1Q10 low flows for the sixty example rivers. The table also presents the difference between 4-day 3-year low flows and the 7Q10 low flows. For 39 of the 60 streams, the 1-day 3-year low flows are less than the 1Q10 low flows. For 18 streams, the 1-day 3-year low flows are greater than the 1Q10 low flows, and for the remaining 3 streams the differences are less than 0.1%. Thus, for the majority of the streams the 1-day 3-year low flow is lower than the 1Q10 low flow. For all sixty streams, the difference between 1-day 3-year low flows and 1Q10 low flows ((1-day 3-year)-(1Q10))/(1-day 3-year) ranges from -50.0% to 20.8%, with the mean and median equal to -4.9% and - 3.1%, respectively. 4-1 ------- Table 4-1. Comparison of 1Q10 and 7Q10 with 1-day 3-yr and 4-day 3- yr low flows (all flows in ft3/sec) River Name State Bull Run VA Trent NC Birch Cr MT Beaver Cr WA Iroquois IN N Fk White AZ E FK Illinois OR Black WI Buckeye CA LaMoine IL Boundary Cr ID Branch RI Linville NC Sheyenne ND Fishing Cr NC Quinnipiac CT Drowning Cr NC Shoshone WY Uncompahgre CO Peace FL Big MO Middle Oconee GA N Br Potomac MD Animas CO Raritan NJ Brandywine PA Salt AZ White VT L Tallapoosa AL E B Delaware NY Big Sunflower MS Meramec MO Chemung NY Arkansas CO Eagle CO Allegash ME Clinch TN Greys WY Cahaba AL Comparison of CMC Design Flows 1Q10 1-day %DIFF* 3-yr 0.3 0.2 -50.0 1.4 1.4 0.0 1.7 1.7 0.0 2.4 2.8 14 .3 3.4 2.4 -41.7 4.8 4.8 0.0 6.4 5.8 -10.3 5.5 5.0 -10.0 7.1 7.0 -1.4 9.3 8.9 -4.5 11.7 12 .0 2.5 8.8 10.0 12 .0 13 .4 13 .0 -3.1 15.9 15.4 -3.2 17.0 12.0 -41.7 17.5 14.9 -17.4 38.8 33.9 -14.4 41.8 42 .9 2.6 35.6 39.9 10.8 49.0 48 .0 -2.1 46 .4 45.0 -3.1 49.4 33 .0 -49.7 54.7 42.9 -27.5 54 .8 60.0 8.7 54.2 46.9 -15.6 61.4 55.8 -10.0 64 .6 63 .0 -2.5 75.3 75.9 0.8 72.7 57.9 -25.6 80.8 82 .0 1.5 89.4 82 .7 -8.1 88 .8 89.9 1.2 89.7 85.7 -4.7 99.9 89.9 -11.1 116.9 120.0 2.6 124.5 134.0 7.1 128.7 127.7 -0.8 122.9 124.8 1.5 151.9 122.0 -23.7 Comparison of CCC Design Flows 7Q10 4 -day 3- %DIFF* yr 0.4 0.4 0.0 1.6 1.6 0.0 2.4 2.4 0.0 3.2 3.4 5.9 3.9 3.0 -30.0 5.3 5.3 0.0 6.7 6.9 2.9 6.7 6.1 -9.8 7.7 7.2 -6.9 9.9 9.4 -5.3 13 .1 13 .0 -0.8 13 .3 13 .2 -0.8 16 .4 15.0 -9.3 18 .3 17.6 -4.0 19.4 13.5 -43.7 32 .3 34 .0 5.0 43.4 36.2 -19.9 46 .8 45.8 -2.2 50.8 49.0 -3.7 55.3 55.2 -0.2 55.3 51.5 -7.4 57.4 45.7 -25.6 61.6 49.0 -25.7 62 .3 61.1 -2.6 67.1 53.6 -25.2 67.2 59.3 -13.3 68 .7 69.5 1.2 85.2 86 .0 0.9 88.3 70.2 -25.8 89.7 91.4 1.9 91.9 85.4 -7.6 92.2 92.7 0.5 97.5 92.5 -5.4 120.1 114.0 -9.3 131.0 126.0 -4.0 134.1 138.4 3.1 135.2 132.2 -2.3 144.5 135.8 -6.4 156.4 149.8 -5.4 * %Difference * %Difference -day 3-year flow) day 3-year flow) (1Q10)),100 / (4-day (7Q10)),100 / ((4-day 3-year flow) 3-year flow) 4-2 ------- Table 4-1. (continued). River Name State Kankakee IN Bouge Chitto MS Hudson NY Potomac WV Root MN Shoal FL Amite LA Niobrara NE Little Pee Dee SC Brazos TX Dan VA French Broad TN St. Croix WI Bighorn MT Beaver PA N P Clearwater ID Red AR Flint GA Merrimack MA Cowlitz WA Comparison of CMC Design Flows 1Q10 1-day 3-yr %DIFF* 179.0 167.6 -6.8 188.6 167.5 -0.6 207.7 170.0 -22.2 209.6 202.2 -3.7 229.7 239.3 4 .0 280.1 270.5 -3.5 298.1 202.1 -5.7 160.9 199.7 19.4 306.7 298.7 -2.7 311.6 277.7 -12.2 329.6 321.6 -2.5 473.6 494.3 4.2 505.9 477.5 -5.9 327.1 364.0 10.1 571.3 539.9 -5.8 529.2 469.6 -12.7 691 537.4 -29.6 207.8 262.5 20.8 270.2 284.0 3.6 901.5 934.7 4.9 Comparison of CCC Design Flows 7Q10 4 -day 3-yr %DIFF* 184.3 174.2 -5.8 191 .6 189 .6 -1.1 211 . 0 191 . 9 -10 . 0 220.7 219.6 -0.5 245.6 239.7 -2.5 291.4 286.0 -1.9 303.4 295.5 -2.7 322.0 304.3 -5.8 322.4 298.9 -7.9 344.9 305.3 -13.0 307.3 380.4 -1.8 532.2 535.5 0.6 536.0 508.5 -5.4 557.0 520.2 -7.1 594.2 557.5 -6.6 648.6 613.0 -5.9 769.2 603.3 -27.5 799.8 731.3 -9.4 929.3 797.3 -16.6 968.7 959.9 -0.9 * %Difference - ((1 * %Difference - (4- -day 3-year flow) day 3-year flow) (1Q10)),100 / (1 (7Q10)),100 / ((4 -day 3-year flow) -day 3-year flow) Similar comparisons can be made between the 4-day 3-year low flows and the 7Q10 low flows based on Table 4-1. For 46 of the 60 streams, the 4-day 3-year low flows are less than the 7Q10 low flows. For nine streams, 4-day 3-year low flows are greater than the 7Q10 low flows, and for the remaining four streams, the differences are less than 0.1%. Thus, the 4-day 3-year low flow is usually lower than the 7Q10 low flow. For all sixty streams, the difference between the 4-day 3-year low flows and 7Q10 low flows ((4-day 3-year) - (7Q10))/(4-day 3-year)) ranges from -44% to 6%, with the mean and median equal to - 7.0% and - 4.4%, respectively. 4.2 Excursions 4-3 ------- Table 4-2 presents the calculated number of excursions that occurred in the 60 streams for the low flows calculated using the hydrologically- and biologically-based methods. The table demonstrates the impact of the choice of one design flow method over the other in terms of number of excursions. For any stream, a higher flow will always result in the same or a greater number of excursions than a lower flow. Occasionally, the difference in the number of excursions of the two design flows is quite dramatic even if the difference between the two design flows is quite small. For example, the 1Q10 and the 1-day 3-year design flow of the Quinnipiac River in Connecticut are 17.5 ft3/sec and 14.9 ft2/sec, respectively, but the corresponding numbers of excursions were 39 and 13. Similar observations could be made for many other streams in Table 4-2. A small difference in design flow may not have a significant impact in wasteload allocations for these streams but may result in a larger number of excursions that desired during the period of flow record. 4.3 Comparison of the Two Methods The comparisons of the design flows show that the magnitudes of the 1-day 3-year and 1Q10 low flows, and the 4-day 3-year and 7Q10 low flows are, on an average basis, similar in magnitude. Although these flows are similar on the average, there may be large differences in the values of these flows for individual streams. More importantly, there can be a significant difference in the number of excursions that result, even if the magnitudes of the flows calculated by the two methods are nearly equal. 4-4 ------- Table 4-2. Comparison of number of excursions of 1Q10 and 7Q10 with number of excursions of 1-day 3-yr and 4-day 3-yr design flows. River Name Bull Run Trent Birch Cr Beaver Cr Iroquois N Fk White E FK Illinois Black Buckeye LaMoine Boundary Cr Branch Linville Sheyenne Fishing Cr Quinnipiac Drowning Cr Shoshone Uncompahgre Peace Big Middle Oconee N Br Potomac Animas Raritan Brandywine Salt White L Tallapoosa E B Delaware Big Sunflower Meramec Chemung State VA NC MT WA IN AZ OR WI CA IL ID RI NC ND NC CT NC WY CO PL MO GA MD CO NJ PA AZ VT AL NY MS MO NY Comparison of CMC Design Flows Comparison of CCC Design Flows 1Q10 % Excur 1-day 3-yr % Excur 7Q10 % Excur 1-day 3-yr % Excur 0.3 19 0.2 10 0.4 8.5 0.4 8.5 1.4 9 1.4 9 1.6 9.3 1.6 9.2 1.7 8 1.7 8 2.4 9.3 2.4 9.2 2.4 1 2.8 6 3.2 4.0 3.4 6.0 3.4 18 2.4 9 3.9 16.8 3.0 9.7 4.8 2 4.8 2 5.3 4.0 5.3 4.0 6.4 13 5.8 12 6.7 11.3 6.9 11.5 5.5 27 5.0 21 6.7 26.0 6.1 24.5 7.1 13 7.0 7 7.7 10.0 7.2 8.5 9.3 33 8.9 20 9.9 24.5 9.4 20.5 11.7 15 12.0 15 13.1 15.8 13.0 15.7 8.8 10 10.0 13 13.3 18.3 13.2 14.0 13.4 21 13.0 15 16.4 25.0 15.0 15.9 11 15.4 6 18.3 14.5 17.6 17.0 17 12.0 15 19.4 29.3 13.5 17.2 17.5 39 14.9 13 32.3 11.3 34.0 13.0 38.8 26 33.9 12 43.4 27.8 36.2 12.7 41.8 3 42.9 6 46.8 9.3 45.8 6.3 35.6 7 39.9 13 50.8 17.5 49.0 49.0 17 48.0 16 55.3 17.3 55.2 46.4 23 45.0 15 55.3 27.8 51.5 49.4 25 33.0 11 57.4 23.3 45.7 14.3 54.7 29 42.9 14 61.6 28.0 49.0 14.8 54.8 0 60.0 2 62.3 6.8 61.1 2.5 54.2 25 46.9 13 67.1 24.3 53.6 13.3 61.4 30 55.8 14 67.2 33.0 59.3 64.6 21 63.0 18 68.7 17.3 6935.0 75.3 20 75.9 20 85.2 20.8 86.0 21.5 72.7 6 57.9 3 88.3 7.0 70.2 3.8 80.8 17 82.0 20 89.7 19.0 91.4 20.5 89.4 31 82.7 8 91.9 30.3 85.4 13.8 88.8 17 89.9 18 92.2 16.5 92.7 17.0 89.7 26 85.7 18 97.5 25.0 92.5 20.5 4-5 ------- Table 4-2. (Continued) River Name Arkansas Eagle Allegash Clinch Greys Cahaba Kankakee Bouge Chitto Hudson Potomac Root Shoal Amite Niobrara Little Pee Dee Brazos Dan French Broad St. Croix Bighorn Beaver N P Clearwater Red Flint Merrimack Cowlitz State CO CO ME TN WY AL IN MS NY WV MN FL LA NE SC TX VA TN WI MT PA ID AR GA MA WA Comparison of CMC Design 1Q10 107.9 116.9 124.5 128.7 122.9 151.9 179.0 188.6 207.7 209.6 229.7 280.1 298.1 160.9 306.7 311.6 329.6 473.6 505.9 327.1 571.3 529.2 691.0 207.8 270.2 901.5 % Excur 23 9 15 23 10 33 34 13 30 19 7 20 19 4 15 11 11 13 34 12 15 20 28 7 13 0 1-day 115 120 134 127 1234 122 167 187 170 202 239 270 282 199 299 277 321 494 477 364 539 469 537 2625 284 934 3-yr 8 0 0 7 .8 8 6 5 0 2 3 5 1 7 7 7 6 3 5 0 9 6 4 .0 0 7 Flows % Excur 26 11 17 17 10 10 14 10 29 14 7 12 14 8 12 4 9 18 2 2 14 4 13 17 9 18 2 Comparison of CCC Design 7Q10 126.1 131.0 134.1 135.2 144.5 156.4 184.3 191.6 211.0 220.7 245.6 291.4 303.4 322.0 322.4 344.9 387.3 532.2 536.0 557.0 594.2 643.6 769.2 799.8 929.3 963.7 % Excur 28.0 17.5 13.0 25.0 18.8 24.8 29.5 19.3 27.8 15.0 10.8 19.3 14.0 11.3 15.0 6.8 10.3 16.0 34.5 16.5 13.3 14.8 28.8 20.3 41.8 4.5 1-day 123 126 138 132 135 149 174 189 191 219 239 286 295 304 298 305 380 535 508 520 557 613 603 731 797 959 3-yr 8 0 4 2 8 8 2 6 9 6 7 0 5 3 9 3 4 5 5 2 5 0 3 0 3 3 Flows % Excur 26.0 11.0 12.0 10.0 16.0 14.0 11.0 24.0 14.0 7.0 17.0 4.0 8.0 19.0 3.0 4-6 ------- The hydrologically-based design flows may actually provide a greater degree of protection of water quality in cases where the value of the design flows are less than that of the corresponding biologically-based design flows. Hydrologically-based design flows have been used successfully in the past in many water quality-based permits. In addition, on an average basis, the values of hydrologically-based design flows are not greatly different from the corresponding values of biologically-based design flows. The biologically-based design flows are not always smaller than the corresponding hydrologically-based design flows for a given stream. Thus, it cannot be stated that choosing one method over the other will always result in the most protective wasteload allocation (and therefore the fewest number of excursions over the period of record). However, the biologically-based method will always provide insurance that the design flow calculated will have resulted in no more than the required number of excursions. Based upon the above, both the hydrologically-based and the biologically-based methods for calculating stream design flows are recommended for use in steady-state modeling. 4-7 ------- SECTION 5. RECOMMENDATIONS 1. If steady-state modeling is used, the hydrologically-based or the biologically-based stream design flow method should be used. If the hydrologically-based method is used, the 1Q10 and 7Q10 low flows should be used as the CMC and CCC design flow, except that the 30Q10 low flow should be used as the CCC design flow for ammonia is situations involving POTWs designed to remove ammonia where limited variability of effluent pollutant concentrations and resulting concentrations the receiving water can be demonstrated. 2. Other technically defensible methods may also be used. 5-1 ------- SECTION 6. REFERENCES 1. U.S. EPA 1985. Technical support document for water-quality based toxics control. Office of Water, Washington D.C. September, 1985. 2. U.S. EPA. Water Quality Criteria. 50 FR 30784 July 29, 1985. 3. Stephan, C.E., D.I. Mount, D.J, Hansen, J.H. Gentile, G.A. Chapman and W.A. Brungs. 1985. Guidelines for deriving numerical national water quality criteria for the protection of aquatic organisms and their uses. PB85-227049. National Technical Information Service, Springfield, VA. 4. U.S. EPA. 1984. Water Quality Standards Handbook. Office of Water regulations and Standards, Washington D.C. 5. U.S. EPA. 1985. Ambient water quality criteria for ammonia - 1984. EPA 440/5-85-001. National Technical Information Service, Springfield, VA. 6. U.S. EPA. 1985. STORET User Handbook, Part FL, Flow Data File. ------- APPENDIX A. Calculation of Hydrologically-Based Design Flows Design flows can be calculated as annual x-day average low flows whose return period is y years, i.e., the xQy low flow. These flows can be estimated from a historical flow record of n years using two different methods. The first is a distribution-free method which makes no assumption about the true probability distribution of annual low flows. The expression for xQy is xQy = (1-e) X(ml) + eX(m2) where X(m) = the m-th lowest annual low flow of record ml = [(n+l)/y] m2 = [(n+l)/y] + 1 [z] = the largest integer less than or equal to z e = This method is only appropriate when the desired return period is less than n/5 years (1) . The second method fits the historical low flow data to a specific probability density function and then computes from this function the flow whose probability of not being exceeded is 1/y. The log Pearson Type III distribution is a convenient function to use because it can accommodate a large variety of distributional shapes and has seen wide-spread use in stream flow frequency analysis. However, there is no physically based rationale for choosing one distribution over another. The xQy low flow based on the log Pearson Type III method is xQy = exp(u + K(g,y) s) where u = mean of the logarithms (base e) of the historical annual low flows, s = standard deviation of the logarithms of the historical low flows, g = skewness coefficient of the logarithms of the historical low flows, K = frequency factor for skewness g and return period y. A-l ------- A sample listing of frequency factors is given in Table A-l. These factors can also be approximated as K = (2/g) [ (1 + (g z)/6 - g2/36)3 - 1] for g < 3 where z is the standard normal variate with cumulative probability 1/y (2). Tables of the normal variates are available in most elementary statistics texts. An appropriate value (3) can be found from z = 4.91 [(1/y)-14 -(1-1/y)-14] To illustrate the use of the two xQy low flow estimation methods, the data in Table A-2 will be analyzed for the 7Q5. The flow values in this table represent the lowest 7-day average flow for each year of record. Also shown are the rankings of these flows from lowest (rank 1) to highest (rank 45). The mean, standard deviation, and skewness coefficient of the logarithms of these annual low flow are shown at the bottom of the table. For the distribution-free approach, the value of (n+l)/y is (45+1)/5 or 9.2. Therefore, the 7Q5 low flow lies between the 9-th and 10-th lowest annual flow. The interpolation factor, e, is 9.2 - 9=0.2 Thus we have 7Q5 = (1. - .20) X(9) + (.20) X(10) = (.80(335) + (.20) (338) = 335.6 cfs A-2 ------- For the log Pearson Type III method, the frequency factor K will be estimated from Table A-l. For skewness of 0.409 and a 5- year return period interpolation results in K = -0.956. The 7Q5 low flow is: 7Q5 = exp(6.01 + (-.856) ( .24) ) = 331.8 cfs For purposes of comparison, K will be estimated using the formulae given above: z = 4.91 [(0.2) -14- (1-0.2)-14] = -0.840 K= (2/.409) [l+( .409) (-.840)/5- (.409)/36)3-l] = -.853 7Q5 = exp(6.01+(-.853) ( .24) ) = 331.8 cfs The difference in the three estimates of the 7Q5 low flow is less than 2 percent. A-3 ------- Table A-l. Frequency Factors (K) for the Log Pearson Type III Distribution Skewness Coefficient 3. 2. 2. 2. 2. 2. 1. 1. 1. 1. 1. 0. 0. 0. 0. 0. -0 -0 -0 -0 -1 -1 -1 -1 -1 -2 -2 -2 -2 -2 -3 0 8 6 4 2 0 8 6 4 2 0 8 6 4 2 0 .2 .4 .6 .8 .0 .2 .4 .6 .8 .0 .2 .4 .6 .8 .0 Return 5 -0 -0 -0 -0 -10 -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 Period, .636 .666 .696 .725 .752 .777 .799 .817 .832 .844 .852 .856 .857 .855 .850 .842 .830 .816 .800 .758 .758 .732 .705 .675 .643 .609 .574 .537 .499 .460 .420 Years 10 -0.660 -0.702 -0.747 -0.795 -0.844 -0.895 -0.945 -0.994 -1.041 -1.086 -1.128 -1.166 -1.200 -1.231 -1.258 -1.282 -1.301 -1.317 -1.328 -1.336 -1.340 -1.340 -1.337 -1.329 -1.318 -1.302 -1.284 -1.262 -1.238 -1.210 -1.180 A-4 ------- Table A-2. Annual 7-Day Low Flows (ft3/sec) for the Amite River Near Denham Springs, LA Year 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 Flow 299 338 355 439 371 410 407 508 450 424 574 489 406 291 352 309 322 278 369 483 523 385 474 Rank 5 10 15 30 20 28 27 38 33 29 41 36 26 4 13 7 8 2 19 35 39 21 34 Year 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 n = 45 u = 6.0 s = 0.23 q = 0.385 Flow 396 275 392 348 385 335 306 280 354 388 357 499 448 650 356 364 648 619 567 445 349 595 Rank 25 1 24 11 22 9 6 3 14 23 17 37 32 45 16 18 44 43 40 31 12 42 A-5 ------- References 1. Linsley, R.K., et al. Hydrology for Engineers, 2nd Edition. McGraw- Hill, New York, NY, 1977. 2. Loucks, D.P., et al., Water Resource Systems Planning and Analysis. Prentice-Hall, Englewood Cliffs, NJ, 1981. 3. Joiner and Rosenblatt, JASA, 66:394, 1971. ------- Appendix B. An Example Use of DFLOW for Ammonia Discharges From POTWs The purpose of this Appendix is to illustrate the use of the DFLOW program to calculate biologically-based design flows for ammonia and compare them with the hydrologically-based design flows of 30Q10 for the 13 streams with the lowest coefficients of variations shown in Table 2-1. B.1 Introduction As stated in the two-number WQC for ammonia (1) , a CCC averaging period of as long as 30 days may be used in situations involving POTWs designed to remove ammonia where low variability of effluent pollutant concentration and resultant concentrations in receiving waters can be demonstrated. In cases where low variability can be demonstrated, longer averaging periods for the ammonia CCC (e.g., a 30-day averaging period) would be acceptable because the magnitudes and durations of excursions above the CCC would be sufficiently limited (1). B.2 Hydrologically-based Design Flow The 30Q10 low flows of the 13 streams with the lowest coefficients of variation (CV) are presented in Table B-l. B-l ------- Table B-l. Design flows and resulting number of excursions using 30-day averaging period (all flows in ft3/sec). River Name Quinnipiac Drowning Cr Uncompahgre Greys Kankakee Hudson Shoal Little Pee Dee St. Croix Niobrara French Broad Bighorn Flint State CT NC CO WY IN NY FL SC WI NE TN MT GA Coeff of Variation 1.02 0.8 0.86 1.16 0.48 1 . 1 0.95 0.94 0.61 0.59 0.93 0.82 1 30Q10 Flow 42.3 54.7 71 160.7 201.8 288 323.5 366.3 571.8 613.2 636.2 913.6 1000 %Excursions 7.8 8.5 6.9 5.7 10 13.4 10.2 7 .4 16.2 6.4 11.9 8.1 6.4 30 -day 3 -year Flow 46.5 65.5 77 .3 166.9 213.6 340.7 339 450 598.6 673.6 715.7 1103 1097 %Excursions 15 15 14 .6 9.9 16.7 24.3 12.1 11.8 21.9 8.1 20.3 14.3 9.6 %Diff* 9 16.5 8.2 3.7 5.5 13.5 4.5 18.6 4.5 9 11.1 17.2 8.8 *%Difference = ((30-day 3-year flow) - (30Q10)) * 100 / (30-day 3-year flow) B.3 Biologically-based Design Flow The 30-day 3-year flows for 13 streams arc presented in Table B-l. To obtain the biologically-based design flow for these streams, an averaging period of 30 days instead of 4 days was entered into the DFLOW program (Appendix D (OMITTED). DFLOW 2.0 has been superseded by newer versions. The current versions of DFLOW and its documentation are available online at http://www.epa.gov./waterscience/dflow). Table B-l also includes the number of excursions that occurred in each of 13 flow records for the hydrologically and biologically-based design flows. B.4 Comparison of Design Flows Table B-l shows that for all 13 streams the 30Q10 low flow is always less than the 30-day 3-year low flow. The difference between the low flows ((30-day 3-year - 30Q10)/30-day 3-year)) 3.7% to 18.6% with the mean equal to 10.2%. Because the 30Q10 low flow is always lower, it results in fewer excursions than the 30-day 3-year low flow. B-2 ------- B.5 Use of Biologically-Based Design Flows for Ammonia Discharges from POTWs As stated earlier, an averaging period of 4 days and a frequency of occurrence of once every three years is used for the CCC. However, for ammonia discharges from POTWs, a longer averaging period may be used in certain cases. According to the national WQC for ammonia, an averaging period as long as 30 days may be used in situations involving POTWs designed to remove ammonia where low variability of effluent concentrations and the resulting concentrations in the receiving waters can be demonstrated. In cases where low variability can be demonstrated, longer averaging periods for the ammonia CCC (e.g., a 30-day averaging period) would be acceptable because the magnitudes and durations of excursions above the CCC would be sufficiently limited. In Section 4.1, the hydrologically-based design flows have been compared with the biologically-based design flows for the 4-day averaging period for all pollutants. Appendix B shows a comparison between the biologically-based 30-day 3-year low flows and the hydrologically-based 30Q10 low flows for 13 streams for ammonia. For these 13 streams, the 30Q10 flow was always less than the 30-day 3-year flow, by an average to 10.2%. Thus, the use of the 30Q10 as the design flow is relatively more protective for these streams. Reference 1. US EPA. 1985d. Ambient water quality criteria for ammonia. 1984. EPA 440/5-85-001. National Technical Information Service, Springfield, VA. B-3 ------- APPENDIX C. Calculation of Biologically-Based Design Flows The biologically-based design flow calculation method is an iterative convergence procedure consisting of five parts. In Part I, Z (the allowed number of excursions) is calculated. In Part II, the set of X-day running averages is calculated from the record of daily flows. Because the ambient (instream) concentration of a pollutant can be considered to be inversely proportional to stream flow, the appropriate "running averages" of stream flow are actually "running harmonic means." (The harmonic mean of a set of numbers is the reciprocal of the arithmetic mean of the reciprocals of the numbers). Thus, "X-day running averages "should be calculated as X/E (1/F), not as (E 5)/X, where F is the flow for an individual day. Throughout this Appendix C, the term "running average" will mean "running harmonic mean." Part III describes the calculation of N (the total number of excursions of a specified flow in the flow record). The calculations described in Part III will be performed for a number of different flows that are specified in Parts IV and V. In Part IV, initial lower and upper limits on the design flow are calculated, the number of excursions at each limit are calculated using Part III, and an initial trial flow is calculated by interpolation between the lower and upper limits. In Part V, successive iterations are performed using the method of false position (1) to calculate the design flow as the highest flow that results in no more than the number of allowed excursions calculated in Part I. Part I. Calculation of allowed number of excursions. 1-1. Calculate Z = D/[(Y) (365.25 days/year)] where D = the number of days in the flow record; Y = the average number of years specified in the frequency and C-l ------- Z = the allowed number of excursions. Part II. Calculation of X-day running averages, i.e., x-day running harmonic means. II-l. Where X = the specified duration (in days) of the average period, calculate the set of X-day running averages for the entice flow record, i.e., calculate an X-day average starting with day 1, day 2, day 3, etc. Each average will have X-l days in common with the next average, and the number of X-day averages calculated from the flow record will be (D+l-X). Part III. Determination of the number of excursions of a specified flow in a set of running averages, i.e., running harmonic means. III-l. Obtain a specified flow of interest from either Part IV or Part V. III-2. In the set of X-day running averages for the entire flow record, record the date for which the first average is below the specified flow and record the number of consecutive days that are part of at least one or more of the X-day averages that are below the specified flow. (Note that whether a day is counted as an excursion day does not depend exclusively on whether the X-day average for that day is below the specified flow of interest. Instead, it depends entirely on whether that day is part of any X-day average that is below the specified flow. Table C-l provides examples of the counting of excursion days.) C-2 ------- Table C-l. Counting excursion days for a specified flow of 100 ft3/sec using 4-day averages Date 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Daily Flow 130 120 110 90 90 100 130 150 70 60 130 90 80 110 100 100 200 500 4 -day avg flow 112.5 102.5 97.5 102.5 112.5 112.5 102.5 102.5 87.5 90.0 102.5 95.0 97.5 127.5 225.0 >100 >100 >100 Is the 4 -day average below 100? No No Yes No No No No No Yes Yea No Yes Yes No No No No No Date Is the date of Number of Number of part of any Date of Number of start excursion excursions 4 -day avg. start of days in of low days in in low that is below excursion excursion flow low flow flow 100? period period period period period No No Yes 3 4 3 12 3 Yes Yes Yes No No Yes 9 8 Yes Yes Yes Yes Yes Yes Yes Yes Yes The daily flows and four-day average flows for days 19 to 200 are all above 100 ft3/sec C-3 ------- Thus the starting date and the duration (in days) of the first excursion period will be recorded. By definition, the minimum duration is X days. III-3. Determine the starting dates of, and number of days in, each succeeding excursion period in the flow record. III-4. Identify all of the excursion periods that begin within 120 days after the beginning of the first excursion period. (Although the first excursion period is often the only one in the 120-day period, two or three sometimes occur within the 120 days. Rarely do any excursion periods occur during days 121 to 240.) All of these excursion periods are considered to be in the first low flow period. Add up the total number of excursion days in the first low period and divide the sum by X to obtain the number of excursions in the first low period. If the number of excursions is calculated to be greater than 5.0, set it equal to 5.0. III-5. Identify the first excursion period that begins after the end of the first low flow period, and start the beginning of the second 120-day low flow period on the first day of this excursion period. Determine the number of excursion days and excursions in the second flow period. III-6. Determine the starting dates of and the number of excursions in each succeeding 120-day low flow period. III-7. Sum the number of excursions in all the low-flow periods to C-4 ------- determine S = the total number of excursions of the specified flow of interest. Part IV. Calculation of initial limits of the design flow and initial trial flow. IV-1. Use L = 0 as the initial lower limit. IV-2. Use U = the xQy low flow as the initial upper limit. IV-3. Use NL = 0 as the number of excursions (see Part III) of the initial lower limit. IV-4. Calculate Nn = the number of excursions (See Part III) of the initial upper limit. IV-5. Calculate T = the initial trial flow as T = L + (Z - NL) (U-L) (ND-NL) Part V. Iterative convergence to the design flow. V-l. Calculate NT = the number of excursions (see Part III) of the trial flow. V-2. If -0.005 <= (NT-Z)/Z) <= +0.005, use T as the design flow and stop. If NT > Z, set U = T and Nn = NT. If NT < Z, set L = T and NL = NT. V-3. If ((U-L)/U)<0.005, use L as the design flow and stop. Otherwise, calculate a new trial flow as T = L + (Z - NL) (U-L) , and repeat steps V-l, V-2, and V-3 as necessary. (Nu-NL) REFERENCE 1. Carnahan, B., H.A. Luther, and J.O. Wilkes. 1969. Applied numerical methods. Wiley, New York. C-5 ------- APPENDIX D. DFLOW 2.0 User's Guide (OMITTED) NOTE: DFLOW 2.0 has been superseded by newer versions. The current versions of DFLOW and its documentation are available online at http://www.epa.gov/waterscience/dflow. D-l ------- APPENDIX E. Questions and Answers Concerning the Bi ologically-Based Method Q #1: New aquatic life protection criteria specify that the acute criteria (CMC) and the chronic criteria (CCC) may be exceeded no more than once every three years on the average by 1-hour and 4-day averages, respectively. They also state that extreme value analyses may not be appropriate- for estimating the ambient exposure condition. What is an extreme value analysis? A. This is a very broad question. There are many types of extreme value analyses. But all extreme value analytical techniques have something in common. Let's consider a time- series of daily flow data in order to explain extreme value techniques. A low-flow water year starts on April 1 of each year and ends on March 30 of the following year. If we perform an extreme value analysis for a 4-day average condition we should estimate 4-day running averages for each water year, then determine which running average is the lowest (extreme) for each water year. Finally, we rank the extreme value of each year for frequency analyses. Q #2: Would you explain how running averages are estimated? A. Starting with April 1, our first running average will be the arithmetic mean of flow data for April 1, 2, 3 and 4: the second running average will be the arithmetic mean of April 2, 3, 4 and 5; and the third running average will be the 3,4, etc. Thus, there will be 362 4-day running averages for each water year of 365 days. Q #3: By extreme value, do you mean lowest running average of the water year? A. In low-flow analyses, the extreme value for a water year is the lowest running average for that year. Q #4: So, do I have 30 extreme values from 30 years' flow record considering one extreme value for each water year? A. Exactly. Q #5: You said about ranking the extreme values. How do you rank them and why do you rank them? A. For low flow analysis, ranking can be done from lowest to highest. For a low-flow analysis of a 30-year flow record, we The biologically-based design flow method has been supported by an overwhelming majority of water quality coordinators at Regional and Headquarter levels. But the method, being totally new, tends to raise a lot of questions which we have heard over time from many reviewers. Some of these questions and related answers are listed here for additional clarification to Appendices C and D of the Guidance. If this paper becomes too long, in a way it defeats its purpose. So we chose questions based on their importance. We encourage our readers to be critical about our answers and raise other questions which they may consider important. This will help us to improve both the method itself and its presentation. In this context, readers may contact Hiranmay Bisuas (FTS-382-7012) or Nelson Thomas (FTS-780-5702). E-l ------- have 30 extreme values. If we rank them from the lowest to the highest value, and no two extreme values are equal, then we have one value for each of 30 ranks, and the return period of the first ranked low is approximately 30 years, and that of the 10th ranked flow is approximately 3 years. Q #6: The frequency analysis using the ranked extreme values seems to be quite straight forward. Why are various kinds of distribution used for frequency analysis? A. If we are concerned with a prediction of low flow for a return period that is equal or less than the flow record, then we will not have to use any distribution at all. The distribution-free, or non-parametric technique is the best for frequency analyses. But, suppose you need 100- 200- or 500- year flood and drought forecasts for the design of a dam (for use power production and irrigation) and we do not have a flow record of such a long period; then, we need to use some form of distribution to extrapolate to 100, 200 or 500 years. There are many well known distributions which can be chosen on a case-by-case basis. Q #7: The new WQC also make some reference to the Log-Pearson Type III distribution as an example of the extreme value analysis. While we are on the subject of distribution, is it the only distribution that is currently in use in the water quality analytical field? A. The United States Geological Survey uses the Log-Pearson Type III distribution in low-flow as well as flood-flow analyses. They made this choice after conducting a study of flood flow analyses using various other techniques. The choice of techniques should be based on the nature of the distribution of extreme values. But, for national consistency of estimates, the USGS chose this technique. Q #8: Extreme value analytical techniques are often used in the hydrologic field, and seem to be quite reasonable. Is there any biological/ ecological reason why extreme value analyses are not appropriate for estimating design flow using the ambient duration and frequency of the new WQC? A. Yes, a direct use of extreme value analyses is not appropriate because biological effects are cumulative. Q #9: Would you elaborate how the cumulative nature of biological effects is related to extreme value analyses? A. In extreme value analytical techniques, only the most extreme drought exposure event is considered, but other, less severe within-year exposure events are totally ignored, although their cumulative efforts could be severe. The severity of those smaller within-year exposure events of extreme drought conditions that are ignored may outrank in severity the extreme exposure events of other less-than-most severe drought conditions. Since the biological effects are cumulative we must find a way to account for all within-year exposures in addition to the most extreme event of each year. Q #10: Your answer is difficult to follow; would you give an example? A. Hydrologists know that we had, in various parts of the US, extreme drought events during the water years 1925-1932, 1955-1956, and during a few years in the late seventies. In other years, drought was not as severe. Suppose that in E-2 ------- water year 1925, there were 4 very low 4-day running averages of which only one was acceptable as the extreme value of that year; the 2nd, 3rd, and the 4th values were ignored. Similarly, one extreme value was estimated for each of the other Water years. But, some of the extreme values of other water years are less severe than 2nd, 3rd or the 4th running averages of the year 1925. Thus, by ignoring these 3 running averages of the water year 1925, the extreme value method has ignored potential severe effects that may result from those exposure events. In addition, the inclusion of other extreme values that are less severe than the 2nd, 3rd, and the 4th running averages of the year 1925, and exclusion of more severe excursion events (2nd, 3rd and 4th excursions of water-year 1925) result in a skewed estimate of low flow. Q #11: The method described to implement the two-number aquatic life criteria is called a biologically-based method. What is biological about it? A. Almost every parameter that is used in this method is derived on the basis of either biological, toxicological or ecological considerations. Whereas the parameters used in the extreme value analysis are unrelated to biological, toxicological or ecological consideration. Q #12: Would you name the things that you think are biological, toxicological, or ecological in nature? - Durations of acceptable exposure conditions: 1 hour for CMC and 4 days for CCC are biologically derived. - 3 years on the average is the allowed ecological recovery period after a single excursion (see Table D-2 of Appendix D of the Technical Support Document for Water Quality-based Toxics Control (TSD)). - 15 years is selected for ecological recovery after a total of 5 or more excursions within a low flow period (see reference Table D-2 in Appendix D of TSD). Q #13: I see neither 15 years nor 5 exposure events in the referenced Table D-2. Could you explain the discrepancy? E-3 ------- A. It is true that neither 15 years nor 5 excursions are found in the reference Table. But what is available is that rivers and streams are fully recovered 5 to 10 years after a severe exposure event. Aquatic biologists consider that repeated within-year exposures can result in catastrophic affects. In their judgment, 10 years exposure interval is inadequate because under that situation the ecology of the receiving system will be under constant stress and recovery. By the same token, a 20-year interval was considered to be unnecessarily stringent for attaining healthy biota. After these considerations and debates among biologists and wasteload allocation coordinators, we decided to use 15 years as an acceptable interval after a severe exposure event consisting of several within-year exposures. Q #14: Have you anything to say about how you decided to allow 5 excursions in an interval of 15 years? A. WDC allow an excursion once every three years on the average. Since the effects of excursions are cumulative, ecological recovery from a severe exposure event requires about 15 years and the recovery period from a single exposure event, according to the national WDC is 3 years. Therefore 15/3 or 5 excursions are accepted as the upper limit of within-year excursion counts. Q #15: Why did you not choose a 12-year interval for 4 within-year exposure events? Or could you not choose an 18-year interval for 6 within-year exposure events (based on the information available in Table D-2 of TSD) ? A. One could make various other choices based on site-specific knowledge but we made our choice for average conditions. Q #16: If 12- or 18- year intervals are chosen for 4 or 6 within- year exposure conditions, would the design flow be different from that of the 15-year interval choice? Do we have any idea about how different the CCC or CMC flow will be for the choices of 12- or 13-year interval? A. No, we did not perform such analyses or comparisons but our guess is that the difference will not be substantial. Q #17: It is understood that, if a 15-year interval is chosen for ecological recovery, then 5 within-year exposures may be allowed because WQC specify 1 exposure on the average of every 3 years. But some extreme drought related low flow periods might include less than 5 within-year exposures, and some more severe low flow periods include more than 5 within- year exposures. If exposure effects are cumulative, why not include all exposures within a year, why limit it to 5? A. The biological method accounts for all within-year excursions when the number of excursions during a low-flow period is 5 or less. So, 5 is the upper limit, and the lower limit is 1. Q #18: What if the within-year excursions for a given flow based on the biological method is naturally greater than 5 during say, a 50- or 100-year drought? In those years, flows may remain low for a long time, such as for 40-50 days, not necessarily E-4 ------- for just 20 days for 5 excursions. After all, we cannot change nature, can we? A. No, we cannot change nature. But we can modify our approach to suit our objective after understanding the consequences of severe events. We made a number of analyses to find out what happens if we account for all, not just 5, excursions that one may expect from those most severe drought years. We found that inclusion of all excursions from those years results in the following: - Design flows of all return periods of say 3, 5, 10, 20, 50 years, etc. are completely dominated by those most severe drought years; and - This leads to extremely stringent design flows. Q #19: There is nothing biological in these analyses. Since the exposure effects are cumulative, should we not count all exposures regardless of how rarely one may expect them, or how stringent the resulting design flow is? A. This is where a little understanding of ecological recovery and familiarity with the North American aquatic life are necessary to make a reasonable choice. The upper bounds of the life cycles and life spans of most North American aquatic species are 2 and 10 years, respectively. An exposure event of 20- or 50-year interval may not be meaningful, particularly when one considers other ways, for example recruitment from the surrounding ecosystem, in which recovery may take place. So, in our judgment, a recovery period of 15 years is adequate for situations where the number of exposures in a low flow period is 5 or more. Q #20: What is described here in the biological method is similar to what is done by hydrologists for partial duration series. They address the problem using traditional statistical approach. Why did you not use a classical statistical method? A. First, the statistical science of partial duration series, particularly in the hydrologic field, is not well developed. Not many people understand it. Although the biological method lacks statistical elegance, it is simple and can be used and understood by field biologists and engineers, alike. We would not be surprised if a statistician comes up with a better statistical answer for the problem that we have in hand. But it would be important for the regions to understand most aspects of the method if we expected them to use it. Q #21: Over the last 20-25 years, the majority of the states in the U.S. used the 7Q10 low flow as the design flow for what we essentially had as a not-to-be exceeded single number WQC value. It seems that it worked fine, although a rationale for such a choice is hard to come by. Why is it so important now to have a rational biolocially-based method to implement the two-number WQC? A. It is important to provide a rational method for three major reasons. First, lack of a biologically-based method in the past led to the adoption of design flows such as 3Q20, 7Q10, 30Q10, 30Q2, and even the annual average flow for identical E-5 ------- water use. A technically defensible method will bring about technical consistency for any desired level of protection. Second, the introduction of the two-number national WQC, whole effluent toxicity, and the guidance on site-specific water quality standards have unalterably changed the environment of toxics control. In these situations, a biologically-based method is necessary that can be applied not only to national two-numbered WQC, but also to other sites-and use-specific durations and frequencies of pollutants and whole effluent toxicities. Third, since WQC and their field use have become complex, it is very important that we develop a simple method that is easily understandable to field biologists and engineers alike. In the past, very few understood the relation between the WQC and the corresponding 7Q10 or other xQy design flow. Q #22: Why is the biologically-based method considered to be more directly based on the water quality criteria than the hydrologically-based method? A. In the biologically-based method, both the averaging period and the frequency (for example, 4 days and 3 years) are taken directly from the criterion, whereas in the hydrologically- based approach, the two number in xQy are not. Most of the other aspects of the biologically-based approach are also based on biological, ecological, and toxicological considerations. One of the major technical differences between the methods is that the 3 years in the biologically- based method is an average frequency, whereas the 10 years in the hydrologically-based approach is a return period. Q #23: Does it make any difference whether biologists, ecologists, and toxicologists understand how design flow is calculated? A. Yes, for three major reasons. First, these are the people who derive the aquatic life criteria. If the criteria are not used in a manner that is consistent with their derivation, the intended level of protection will probably not be achieved. Second, site-specific frequencies and durations will not correctly affect design flow if the duration and frequency are not directly used in the calculation. Third, if they understand what parameters affect design flow, biologists, ecologists, and toxicologists can gather data that might allow them to refine their estimates of such values as one hours, four days, three years, and fifteen years. E-6 ------- Q #24: Let us discuss the simplicity of the biologically-based method. I am not clear how an excursion is counted. Would you explain how you count excursions and estimate design flows? A. This is the key to understanding the biologically-based method. Since the stream flow is inversely proportional to instream concentration, any consecutive 4-day average of low- flow that is lower than the design flow is counted as one excursion of the CCC. The following is the step-by-step explanation of how excursions are counted in estimating x-day y-year design flow: 1. An excursion period is defined as a sequence of consecutive days where each day belongs to an x-day average flow that is below the design flow. For example, if the three running averages of a consecutive 6-day period are less than the 4-day 3-year design flow, then those 6 days belong to an excursion period. 2. The number of excursions in an excursion period is the length of the period divided by the criteria averaging period. For example, if an excursion period is 6 days long, then the number of excursions for the 4-day averaging period for CCC is 6/4 or 1.5. 3. The total number of excursions is limited to 5 within a low flow period. Usually a low flow period lasts 120 days or less. In some rare stream situations, more than one low flow period within a water year is possible. 4. The allowed total number of excursions over the period of record is the number of years of record divided by the frequency of aquatic life criteria (3 years for the CCC of the new national two-number criteria). For example, if we have a 30-year flow record, then total number of excursions that are allowed for x-day 3-year criteria is equal to 30/3 or 10. 5. The 4-day 3-year design flow for the 4-day 3-year CCC based on a 30-year flow record of a given river is equal that flow which results in no more than the allowable number of excursions. For example, the total allowable number of excursions for the given record is 10. The design flow is the highest flow that results in no more than 10 excursions calculated as defined in steps 1 through 4 above. Q #25. Let us take the example printout (from page D-5) for the Amite River as presented below. Will you explain the procedure using this example? A. As shown in the following printout, we have a flow record from 1937 to 1983 which is approximately 42 years. Since we are allowed to have no more than one excursion in every 3 years on the average, we have 42/3 or about 14 excursions. In October 1952, we encountered the first excursion for a continuous period of 6 days. Thus, we calculate 6/4 or 1.5 excursions for that low flow event. The next excursion period occurs, starting from October 10, 1956, for 30 consecutive days. Since the upper limit of excursions in a low flow period (a low flow period is usually 120 days long) is 5, we E-l ------- obtained a total of 5 excursions only, although in reality there were altogether 30/4 or 7.50 excursions in that low flow period. Similarly, we found only 5 excursions for total period of 30 days during the low flow period of 1963. Q #26: It seems like the accuracy of the design flow estimates is totally dependent on the length of the flow record. Do you agree with this observation? A. Absolutely. This is true about any analysis. More relevant data are necessary to provide more accurate information. Q #27: What minimum length of flow record is recommended? A. The longer the flow record, the more reliable the estimated design conditions will be. Figure E-l (OMITTED) shows how the spread in the 90% confidence limits on the extreme value-based design load with 10-year return period decreases with increasing period of record. (This figure was derived on the basis of lognormal statistics, not log Pearson type 3). Results are shown for both low variability (CV=0.2) and high variability (CV=0.8) situations. Based on the behavior of these curves, it appears that 20 to 30 years of record is a reasonable minimum requirement for extreme value analysis at a 10-year return period. The case for the biologically-based excursion criterion is less definitive. However, since it considers all days within the period of record as its sample (not just the worst condition of each year), its sample size is much larger than that of an extreme value analysis. Thus, it may be possible to use periods of record less than 20 years with this criterion and still have a good level of confidence in the results. E-2 ------- Q #28: What would you do for intermittent streams where low flow is zero during low flow periods? Also, how will you use the biologically-based method in situations where flow data are not available? A. These are problems that are generic to all flow estimating techniques. For intermittent streams for which the low flow is zero, the design flows for CMC as well as CCC are equal to zero. In situations where flow data are not available, field hydrologists and engineers sometimes use flow data from hydrologically comparable drainage basins. Q #29: The table given in Question 23 looks simple. How much time does it take to conduct a biologically-based analysis for any stream of interest? A. The analysis is performed in two steps. First, daily flow data are retrieved from the daily flow file in STORET, by submitting a batch job. This will take a few minutes of time at the computer. However, the job run might take anywhere from a few minutes to several hours, depending on how busy the computer system is at the time of submittal. Once the data has been retrieved, the analysis can be performed in five or ten minutes. Q #30: It seems that the foundation of the information about ecological recovery periods for the two-number WQC is all that are listed in Table D-2 of the TSD. But, anybody familiar with these references will tell you that the recovery periods listed in that table are related to recovery from catastrophic exposures caused by spills, not by effluents of malfunctioned advanced treatment facilities. Would you agree that this is not a satisfactory set of information to make such an important decision? E-3 ------- A. This is the best available information that we could use to estimate ecological recovery. Considering the complexities involved in the implementation of the two number WQC, and the site-specific WQC for pollutants and whole effluent toxicity, we could not leave the recovery question open to anyone's interpretation. Considering the potential for misuse of the WQC in their implementation phase, we had to use our best judgment and the best information available, although we recognize that our best judgment would be debatable. Since the information base is not as strong we want to have, in keeping with the Agency policy and legal background, we had to go in the direction of protection in the over-all decision making process. Q #31: What are you doing to improve the information base? A. ORD is planning to undertake a major effort before the next update of the WQC. But, this is an area in which success is dependent more on cooperative efforts in which field biologists, ecologists, toxicologists, engineers and hydrologists share their experience than doing mere literature reviews and/or gathering laboratory-generated information. REFERENCE 1. Stedinger, J.R., 'Confidence Intervals for Design Events', Jour. Hyd. Eng. Div., ASCE, Vol. 109, M. 1, January 1983. E-4 ------- DISCLAIMER We have made efforts to ensure that this electronic document is an accurate reproduction of the original paper document. However, this document does not substitute for EPA regulations; nor is it a regulation itself. Thus, it does not and cannot impose legally binding requirements on EPA, the states, tribes or the regulated community, and may not apply to a particular situation based on the circumstances. If there are any differences between this web document and the statute or regulations related to this document, or the original (paper) document, the statute, regulations, and original document govern. We may change this guidance in the future. Supplemental material such as this disclaimer, a document abstract and glossary entries may have been added to the electronic document. ------- |