GPO's CGP - Record View
Page 1 of2
1 '"''->P / BASSC / ADVANCED
NFW TrTL.rs
CATALOG OF
U.S. GOVERNMENT PUBLICATIONS
Catalogs to Search:
Congressional Serial Set GPO Access Publications Periodicals
£QDgressignai_Pubiications lntern_eLPiJb.!!fia.tiSO.s Seriate
My Options:
Bookshelf Results list
Preference! EreyiousjSearche^
Catalog of U.S. Government Publications Home Page > National Bibliography of U.S. Government Publications
Add to Bookshelf | E-mail
Choose Record View: Standard | Short | MARC
Record 1 out of 1
Title Technical guidance manual for performing waste load allocations book VI design
conditions-chapter I stream design flow for steady-state modeling.
Publisher Info. [Washington, D.C.]: United States Environmental Protection Agency, Office of
Water. [1986]
Internet Access http://pyri.access.gpQ.gov/GPO/LPS67684
SuDoc Number
Item Number
Description
General Note
Bibliography
System Details
Subject - LC
Added Entry
1200
EP 2.8:W 27/3 «,,„_ &'3/.'f>:V
0607-C (online)
65 p.: digital, PDF file.
Title from title screen (viewed on Mar. 14, 2006). v-"-^c
"September1986."
Includes bibliographical references.
Mode of access: Internet from the EPA web site. Address as of 3/14/2006:
http://www.epa.gov/waterscience/iibrary/modeling/wlabook6chapter1.pdf; current
access available via PURL.
Effluent quality -- Government policy -- United States --Handbooks, manuals, etc.
LJLQJdronmejrjt^
% MVV
. J
Holdings Alljtems
Locate in a Library (online) http://catalog-web2.flpo.gov/LocateLibraries/locate.jsp?ltemNumber=0607-
CaSYS=.OOQ5863.34
OCLC Number (OCoLC)64668564
System Number 000586334
.Session - Preferences - Feedback - Hejfi - Browse - Search - ResuJtsJLjs.t - ElSviQ.ys..Sear,ch6s - Catalogs to.Search - Bookshelf
http://catalog.gpo.gOv/F/CS5GGXD7NILUJA718SIBFG2B31B2HCH7KJ5211KKXCQV... 7/13/2006
-------
GPO's CGP - Record View
Page 2 of2
A service of the U.S. Government Printing Office. 200S-2006.
http://catalog.gpo.gOv/F/CS5GGXD7NILUJA718SIBFG2B3IB2HCH7KJ52HKKXCQV... 7/13/2006
-------
PB92-231778
TECHNICAL GUIDANCE MANUAL
FOR
PERFORMING WASTELOAD ALLOCATION
BOOK VI
DESIGN CONDITIONS
CHAPTER
STREAM DESIGN FLOW FOR STEADY-STATE MODELING
SEPTEMBER 1986
US «f«flTUEHT Of COMMERCE
•UIOUl TECHMC
-------
Click here for
DISCLAIMER
Document starts on next page
L
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON. D.C. 7.9400
SEP 291986
Of
WAT III
MEMORANDUM
Subject: Technical Guidance Manual for Performing Waste load
Allocations Book VI, Design Conditions: Chapter 1 - Stream
Design Flow for Stead-State Modeling
From:
To:
»TTliara A. Wluttington, Director
Office of Water Regulations and Standards (WH-55U
Addressees
Attach^, for national use, is the final torsion of the Technical
Guidance Manual for Performing Wasteload Allocations, Book VI, Design
Conditions: Chapter i - Stream Design Flow for Steady-State Modeling.
This manual replaces the interim stream design flow reconrnendation*
included in Appendix D of our Technical Support Document for Hater
Quality-based Toxics Control, September, 1985. He 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 Tin S. Stuart, Chief, Monitoring Branch, Monitoring and Data
Support Division (WH-553) on (FTS) 382-7074
Attachment
Address«
Regional Water Management Division Directors
Regional Environmental Service* Division Din
Regional Wasteload Allocation Coordinators
-------
ACKNOWLEDGMENT
The preparation of this docunent was a collaborative effort of the Office
of Water and the Office of Research and Development. Tin S. Stuart and
Mark Morris of the Office of Water Regulations and Standards, and Nelson
Thomas of Environmental Research Laboratory-Duluth provided overall guidance
and direction in preparation of this manual. Program managers of Bag ions
III, IV, V, VI and VII, and the Headquarters program managers within the
Office of Municipal Pollution Control and the Office of Water Regulations
and Standards actively participated in developing this guidance.
Hiranmay Biswas of the Monitoring and Data Support Division, Lewis A.
Rossman of water Engineering Research Laboratory-Cincinnati and Charles E.
Stephen of Environmental Research Laboratory-Duluth are the principal
authors of this manual. Sections 1, 2 and 4, and Appendices B and E were
prepared by Hiranmay Biswas. Sections 3 and S of the manual were prepared
by Charles E. Stephen with statistical support from Russell E. Erickson of
Environmental Research Laboratory-Duluth. Appendices A. C and 0 of the
manual were jointly prepared by Lewis A. Rossman and Charles E. Stephan.
Individuals listed below contributed to the preparation of this manual
and their efforts ar« greatly appreciated.
Garret Bondy, U.S. EPA Waste Load Allocation (WLA) Section Region VI
Rick Brandes, U.S. EPA Permit Division
Miriam Goldberg, U.S. EPA Analysis and Evaluation Division
Joseph Gornley, U.S. EPA Municipal Facilities Division
Robert C. Horn, U.S. EPA Criteria and Standard Division
Nortwrt Huang, U.S. EPA Municipal Facilities Division
Sally Marquis, U.S. EPA WLA Coordinator teg ion X
Robert F. McGhee, U.S. WLA Coordinator Region IV
John Maxted, U.S. EPA Criteria and Stardard Division
Rosella O'Conner, U.S. EPA WLA Coordinator Region II
Thomas W. Purcell, U.S. EPA Criteria and Standard Division
Robert J. Steiert, U.S. EPA Coordinator Region VII
Randall E. William, PXI, RTF, NC
Dale Wismer, U.S. EPA WLA Coordinator Region III
Edward H. W», U.S. EPA WLA Coordinator Region I
Phil Woods, U.S. EPA WLA Coordinator Region IX
Bruce Zander, U.S. EPA WLA Coordinator Region VIII
ii
-------
TABUE OF CONTEXTS
SECTION 1. I^m*)OUC^CN
l.l Putpos*
1.2 Background
1.3 Scop*
SECTION 2. HYDROLOGICALLY-BASED DESIGN FLOW METHOD
2.1 Introduction
2.2 Rational*
2.3 Exarple Cases
SECTION 3. BIOLOGICALLY-BASED DESIGN FLCM METHOD
3.1 Introduction
3 2 Procedure
3.3 Rational*
3.4 Exanele Cas*s
SECTION 4. COMPARISON OF RESXTS OF THE TWO METHODS
4.1 Dislgn Flow*
4.1.1 (fee of Biologically-Based Design Flows for Annonia
Discharges fcon
4.2 Excursions
4.3 Gonparison of th* Two Methods
SECTION 5. RECOMMENDATIONS
SECTION 6. REFERENCES
Page
1-1
1-1
l-l
1-3
2-1
2-1
2-2
2-3
3-1
3-1
3-6
3-1
3-7
4-1
4-1
4-4
4-5
4-8
5-1
6-1
iii
-------
TABLE OF CONTENTS
APPENDIX A - Calculation of Hydro logically-based Design "lews
APPENDIX B - Design Flows for Ammonia
APPENDIX C - Calculation of Biologically-based Design Flows
APPENDIX 0 - Description of the Program DFtflW
Pace
A-1
8-1
C-l
D-l
APPENDIX I - Questions and Answers Concerning the Bioloqlcally-based Method E-l
lv
-------
SECTION 1. INTRODUCTION
1.1 Purpose
The purpose of. this guidance i.i to describe and compare two methods
that on be used to calculate stream design flows for any pollutant or
effluent for which a two-nunber water quality criterion (WQC) for the
protection of aquatic life Is available, the two methods described are:
1. Tie hyrirolagically-baseri design flow method teconnenderi for
interim use in the Technical Support Document for Water Quality
based Toxics Control (1); and
2. A biolojically-based design flow method that was developed by
th« Office of Research and Development of the U.S. EP*.
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 be 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 exceedences. 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.
1 - 1
-------
National, site-specific, and effluent toxlclty criteria are
as o concentrations, rather than one* so that the criteria can more
accurately riflect toxlcologlcai and practical realities (1 - 4):
a. The lo-er concentration Is called the Criterion Continuous
Concentration (CCO. The CCC is the 4-day average concentration
of a pollutant in ancient water that should not be exceeded
more tlun once every three years on the average.
b. the higher concentration is called the Criterion Maximum Concen-
tration (C1C). Die one-hour average concentration in *«bient
water should not exceed the WC more than once every three years
on the average.
Use of aquatic life criteria for develoolng water quality-based
Demit limits and for designing waste treatment facilities requires the
selection of an appropriate wasteload allocation model. Cynamic models
are preferred for the application of aquatic life criteria in order to
ma'
-------
Che way of using the CCC and the CMC in steady-state modeling requires
calculation of the two design flows (i.e., a CCC design Clow and a CMC
design flew). Whether the CCC and its design flow or the CMC and its
oeslgn flow Is more restrictive* and therefore controlling, oust be
determined individually for each pollutant of concern in each effluent
because the CCC and CMC are pollutant-specific, whereas the two design
flcvs are specific to the receiving waters.
Uasteload allocation noriellng for streams 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
mst be extrapolated from other streams having hydrologic characteristics
similar to those of the strewn of interest.
1.3
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 recoimendations concerning the use of these methods in stea^/
state modeling.
The water qua'.ity criterion for dissolved oxygen was revised very
recently and the assessment of the appropriate design flow for dissolved
oxyqen modeling has not yet been completed. Therefor*, the state-specified
design flows that traditionally have been used for conventional pollutants
should not be affected ty this guidance.
1 - 3
-------
State-specIfled design Clows necessarily preenpt any design flow
that U recoRimmded in this guidance unless the state chooses to use either
of th-HO two methoit*. Vv» choice of d*il'jn flows for tlw i>eot*:ti
-------
SECTION 2. HYDreLOGICMiTMiASED OESXCH PLOW MCTWOD
2.1 Introduction
Th<» puqme of tills section la to describe the hydrologlcally-based
design Clow calculation method and provide same examples of its use. The
Technical Support Document Cor Water Quality-based Toxics Control (1)
provides Agency guidance on control of both generic and pollutant^pecif ic
toxicIty and rwcowwftJeJ interim use oE the hvdrologlcally-based method.
In addition, the Agency also reconroended (1,2) that the frequencies of
allowed exceedence* and the durations of the averaging periods speclfitd in
aquatic UEe criteria should not be used directly to calculate steady-
stat« >!esign flows using'an extn»*e value analysis. Foe example, if a
criterion specifies that the fou»-day average concentration should not
exceed a particular value more than once every three years on the average,
this should not be interpreted as implying that the 403 low Clow LI
appropriate for use as the design flow.
Because a procedure had not been developed for calculating design
flew based on the durations and frequencies specifier! in aquatic life
criteria, the U.S. EPA recciiiianded interim use of the 105 and 1Q10 low
flows as the CMC design flow and the 7Q5 and 7Q10 low flows as the CCC
design flow for unstressed «nd stressed systems, respectively (I).
Further consideration of stress placed on aquatic ecosystems resulting
from exceedences of water quality criteria indicates that there is little
justification for different design flows for uratr»ws*1 and stressed
systems. 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
-------
th» degree r»f pollutant *trc*s when one considers both the timing and
variability of flows, effluent discharges, and eewyston sensitivity and
reslli-ince.
2.2 Rationale
The following provide* a rationale for the hydrologlcally-based
design flow calculation method:
• *bout half of the states in the nation use 7Q10 as the design low
Clow.
• The log-Pearson type III flow estiniting technliju* or other extresne
value analytical techniques that are used to calculate flow
statistics from daily flow data are consistent with past engine«}rt'vj
and statistical practice*
• Most users are familiar with the log-Peanon Type HZ flow estimating
procedure and the USGS provides technical* support for this technique.
• Analyse* of 63 rivers Indicate that, on the average, the biolcmcally-
Uased CMC and CCC design flows are nearly equal to th- 1QIO and the
7010 low flaw*.
2.3 Sxaffple Cases
In order to illustrate eh* calculation of hydrologically-tjased
design flows, sixty rivers with flows of various magnitudes and variabilities
were chosen fron around the country. The 1010 and 7Q10 low flows of the
sixty rivers are presented in Ibbl* 2-1. the list of riven in this table
is arranged in increasing magnitude of the 7010 low flows. The
estimates of the 1010 and 7Q10 lew flows were made using the USGS
daily Clow database and the FLOSTAT program (6) which enploys the
log-parson type III technique.
2-2
-------
The estimate* of 1QIO and 7QLO low flows could have been made using
EP^-ORD's DFLCW progran, which usos a simplified version of the log-parson
Type III Method. The simplified version of the log-Pearson Type III
estl-natlng technique for any xQy design flow la presented In Appendix A.
Although the Log-Pearson Type til 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 hydrologlctlly-oased design flow for anroonia
is discussed In Appendix B.
Analyses of the 1010 and 7010 lev flows In Table 2-1 indicate that
the man of the ratios of 7Q10 to 1010 Is 1.3. The median of the ratios
Is 1.1, vhereas the range of the ratios is 1.0 to 3.85. Thus, 7010 lev
flews are generally 10 to 30% greater than the corresponding 1010 low
flows, although In one case the 7Q10 Is 3.85 tiMts greater than the
corresponding 1010.
Table 2-1. Hydrologically-based design flows (ftVsec) *or 60 streams
Station 10
01657000
02092500
06026000
12449600
05522000
09490800
14372500
05381000
10291500
OSS8SOOO
12321500
01111500
River name
Bull Rm
Trent
Birch Or
Beaver Cr
Iroquols
N IX tfcite
C Fk Illinois
Black
Buckeye
U Molne
Boundary Cr
Branch
State
VA
NC
MT
W*
IN
AZ
OR
WI
CA
tt
ID
RI
Period of
focon)
1951-82
1951-82
1946-77
1960-78
1949-78
1966-78
1942-83
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 (ft3/«ec)
1010
0.3
1.4
1.7
2.4
3.4
4.8
6.4
5.5
7.1
9.3
11.7
8.8
7QIO
0.4
1.6
2.4
3.2
3.9
5.3
6.7
6.7
7.7
9.9
i3.1
13.3
1
7010
15TZ
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 10
02138500
05059000
02083000
01196500
02133500
06280300
09149500
02296750
07018500
02217300
01481000
09497500
01144000
01600000
09359500
01403060
02413500
01421000
07288500
07013000
01531000
07096000
09070000
01011000
03528000
13023000
02424000
05515500
02490500
01315500
01610000
05386000
02369000
07378500
06465500
02135000
08110200
02076000
03455000
05333500
06287000
03107500
River name
Linville
Sheyenne
Fishing Cr
Ouimiplac
Drcwnlng Cr
Shoahone
UhconiMhgre
Peace
Rig
Middle Cconee
Vandywine
Salt
Miite
H 8r Potomac
AnifliM
Raritan
L lallapcosa
B B Delaware
Rig Sunflower
Meramec
Chenung
Arkansas
Eagle
Alle^ash
Q inch
Greys
Cahabe
Kankakee
Botjge Chitto
Hudson
Potomac
Moot
Shoal
Amite
Mobrara
Little Pee Dee
Brasos
Dan
FteitcJ) Broad
St. Croix
Bighorn
Beaver
State
NC
NO
NC
CT
NC
W
CO
FL
MO
GA
PA
A2
VT
MD
CO
NJ
AL
NY
MS
MO
NY
CO
CO
ME
TO
W
AL
IN
MS
NY
WV
MN
FL
(A
NC
sc
TX
v&
TO
WI
MT
PA
DA ^4 /^4 f\ f
renou oc
Record
1922-84
1951-81
1927-82
1931-84
1940-78
1957-84
1939-80
1931-84
1922-84
1902-84
1912-84
1925-80
1915-84
1939-83
1946-56
1904-83
1940-51
1915-78
1936-80
1923-78
1915-78
1901-81
1947-80
1932-83
1919-78
1937-83
1902-78
1926-78
1945-81
1908-78
1939-83
1938-61
1939-82
1939-83
1939-83
1942-78
1966-78
1924-52
1901-78
1914-81
1935-79
1957-83
cv
1.74
2.10
1.48
1.02
0.80
1.54
0.86
1.54
2.16
1.37
1.17
2.05
1.43
1.42
1.56
1.64
1.32
1.41
1.42
2.41
1.91
1.12
1.36
1.39
1.55
1.16
2.07
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
Design flc*
1010
13.4
15.9
17.0
17.5
38.8
41.8
35.6
49.0
46.4
49.4
61.4
64.6
75.3
54.7
54.8
54.2
72.7
80.8
89.4
88.8
89.7
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
» (ftVsec)
7010
16.4
13.3
19.4
32.3
43.4
46.8
50.8
55.3
55.3
57.4
67.2
68.7
85.2
61.6
62.3
67.1
88.3
89.7
91.9
92.2
97.5
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
T3IU
1.22
1.15
1.14
1.85
1.12
1.12
1.43
1.13
1.19
1.16
i.09
1.06
1.13
1.13
1.15
1.24
1.21
1.11
1.03
1.05
1.09
1. IV
1.12
1.D8
l.OS
1.19
1.03
1.03
1.02
1.02
1.05
1.07
1.04
1.02
2. 00
l.Oi
1.11
l.U
1.12
l.Ofc
1.70
1.04
2-4
-------
Table 2*1 (continued).
Station ID River name
Period of
State Record Or
Design Clow (ftVsec)
1010 j 7Q10 | 17
13341000
07341500
02350500
01S36500
01100000
14233430
N F deacwatec
tod
Hint
Susquehanna
MerrL-nadc
Cowlitz
ID
AR
GA
PA
MA
UA
1927-68
1928-81
1930-58
1901-83
1924-83
1968-78
1.16
1.41
1.00
1.34
1.01
0.93
529.2
691.0
207.8
782.0
270.2
901. 5
648.6
769.2
799.8
814.3
929.3
968.7
1.23
1.11
3.85
1.04
3.44
LOT
*CV • Coefficient of Variation
2-5
-------
SECTION 1. BtOLOGICALLlf-HASSO OF.SIGN FL3W MCTHOD
3.1 rntrcHuction
The purpose of this section is to describ* th* tn
•1*si-jn flow calculation method and provide some examples of its use.
Thl* method was developed by the Office of Research and Development of
the U.S. £?* in order to provide a way of directly using EPA'a two-number
aouatlc life water quality criteria (WQC) for individual pollutant* and
whole effluents to calculate the design flow for performing a wasteload
allocation using steady-state modeling. The two-number MX are in the
intenslty-duration-ereouency fornat, in that they specify Intensity as
criteria concentrat Ions, duration a* averaging periods, and frequency as
average frequency of allowed excursions. Because the flow of, and
concentration* of pollutants in, effluents and streams are easily considered
in terns of intensity, duration, and frequency, use of this format for
expressing XX allows a direct application to effluents and itre*ns.
Recausa stead/estate modeling assmes that the co-position and flew
of the effluent of concern is constant, the ambient (instream) concentcation
of a pollutant can be considered to be inversely proportional, to stream Jlcv.
Thus by applying a specifier] averaging period and freouency to a record
of the historical flow of the stream of concern, the design flow can be
calculated a* the highest flow that will not cause exceedences to occur
more often than allowed by the specified average frequency, based on
historical data. The allowed exceedences are intended to be small enough
and far enough apart, on the average, that the resulting small stresses
on aouatic organisms will not cause unacceptable effects, except in
those cases when a drought Itself would cause unacceptable effects.
3 - 1
-------
~mo averaging iwrioi* specified in national water tf that exceed th* CCC. However, in st.jaJy-*tate nolelini;,
flow is averaged over a given period to Identify "non-nxeeedences",
i.e., *v*r*r* fli>n fiat *e* below a specified flow.
UM of the terns "exceedence" and "non-«xceedence". neither of
which are in the dictionary, can be a cause of confusion. Mater quality
criteria are usually expressed as upper llnlts on concentrations in
*7V3iene watar aivj the periods of concern are when the anbient concentration
exceeds a criurion concentration, I.e., when there is an excevdenc*.
In «te*V~stJta noteUn<;, the averaging is of flows, not concentrations.
ffecause a low flow results in a high pollutant concentration, the period
o! concern for flow is when the flow is lev* than the design flow, i.«.,
when there is a non-oxceedence of a given flow. A non-«xceedence of a
design flew corresponds to an exceedence of a criterion. Use of the
nonfunctional tern "excursion", which is in the dictionary, avoids
this confusion. Oie of the term "excursion* also avoids the problem
that none water quality criteria, such as those for dissolved oxygen
and low pH, »ust be stated as lower limits, not upper limits. An
exceedence of a dissolved oxygen criterion is favorable, not unfavorable.
"Excursion", in this guidance manual, will henceforth be used to i/
3-2
-------
an unfavorable condition, <*.*»vsr, *•„**? three
yean is meant to he longer than the average recovery period so that
ecosystem cannot he In a constant statu of re-wecy even if excursions
are evenly spaced over time.
M though 3 yean 4p-;van to be appropriate Cor s.-aall excursions
that are somewhat Isolated, it appean 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 lev
flow and should not be allowed to unnecessarily l'>«*er design flows.
Although droughts do severely stress aquatic ecosystems, both directly,
because of low flow, and Indirectly, because of t'w wsaltioj high
concentrations of pollutants, many ecosystem apparently recover from
severe stresses in more than S, but less than 10 yean (1). Because it
Is not adequately protective to keep ecosystem In a constant state of
recovery, IS yean seems like an appropriate stress-free period of
tlae, on th» average, to allcv after a severe stress caused by a drought
situation. Because three yean are allowed for each excursion on the
average, counting no more than S excursions for any low flow period will
3 - 3
-------
provide ao nor« thaii 15 yiars, on the averse, for sever? stcj**** csus*.l
by droMnfets* Thus, for each low flow period, the nimber of excursion*
cannot Ivt less than 1.0 or greater t'uin 3.0. TI-J -nixL-un duration of a
IOM-'.AC*' period wan set *t 120 days because It Is not too uncormon for
excursion* to occur within 120 days of each other, wh»r**s It U wry
rare for excursions to occur during days 121 to 240 after the beginning
of a low-flow parlor)*
Figure 3-1 illustrate* the features of the biologically-based design
flow calculation method. Intarvals a-b and c-<5 are excursion periods ancj
each -Jay In these intervals is part of *n average flow that is below the
design flow. The nunber of excursions in an excursion period is calculated
as the numh*r of days In ths excursion period divided by the duration (in
days) of the averaging period (e.o., 1 day for the CMC and 4 days for the
Cr"). A low-flew period ij <1«fi-vjl -is -v>> or ior« excursion periods
occurring within a 120xiiy interval. As discussed above, if the calculated
i-xiber o? excursions that occur in a 120-day low-flow period is greater
than 5, the number is set at 5 far the purposes of calculating the desi^i
flow.
Bectuse 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 toxicologlcal information concerning the stresses tha*
aquatic org«ol»ia, ecosystems, and their uses can toUrate. the
biologically-based calculation method is flexible enough to make full use
3 - 4
-------
r
L
0
w
DESIGN
EXCURSION
PERIOD
EXCURSION
PERIOD
* D C
TIME, (DAYS)
figure 3-1: Illustration of blologicalvly-^>as€d design flow
3-5
-------
of special «*v.«*.il'xi ixicloli 41? Fc\»7.wmcles thJit -nl^Jit ?M «*lo«:f.»>1 f>r
SiVK:lClc *01lMt»nt» f*.o. » amnonta) or Ln site-*p-*ci(ic criteria. thii
:MChoJ I* «rplrioil, not statistical, because it duals wit") the actual
flow record Itself, not with a statistical distribution that is intended
to 'Jtaorlb* th» flow record.
In *JJltlon, thU method provides an understanding of how «\any
of the OCC er OC are likely to occur, and during what t±n» of the yesr,
based on actual historical (low data. Thus, It i* possible to 4Jirvj of yhat level of protection
actually is provided should aid in the use of criteria.
3.2 Procedure
Although the calculation procedure described in *p?«n
-------
The OtC and CCC design flows avo calculated In almost the atom manner.
The differences result fixm the fact that the CMC Is expressed as a one-
hour average, whereas the CCC is expressed as a four-day average. However,
the flo- records that art available consist of on .-day average* Clows.
For stream with naturally occurring low flows, calculation of the CC design
flow frcn one-day averages, rather than one-hour averages, should be
reasonably acceptable because naturally occurring low flows of receiving
streaw are usually very similar from one hour to the next. In regulated
streanv, 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 nuch more from one hour to the next than the naturally occurring
flows of streams.
3.3 Rationale
The following provides a rationale for 'J*.e biologically-based
design flow calculation method:
• Xt allows the us* 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 toxicologlcal characteristics of aquatic organisms and ecosystem
and their uses whenever possible.
• Zt takes into account all excursions in the flow record.
* tt provides the necessary design flow directly without requiring any
design flow statistics in the xQy format.
• Zt 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 Ex angle _C*s*>s
The sixty fl»* records that w*>ro .vwly<*1 usiry t'v» hydrolojicaily
based ntthod (sea Table 2-1} were also analyzed using the L lo logically -
based das ion flow method. The CMC design flow was calculated for A
I -day averaging period and the CCC design flow was calculated using the
4-rtay averaging period. toth w»re calculated using a fre
-------
Tabl* 3*1. Blolojlcally-basad design Clows (ftVsec) for fiO rivers
Station ID
01657000
020*2500
06026000
124496i)0
05522000
09490800
14372500
05381030
10291500
OSMSOOn
12321500
01111500
02138500
05059000
02083000
01196500
02133500
06280300
09149500
02296750
07018500
02217500
01600000
09359500
01403060
01481000
09497500
01144300
02413503
01421000
07288500
07013000
01531000
07096000
09070000
01011000
03528000
1302300)
02424000
Riv*r name
Bull fen
Trtnt
Birch Or
6*av«r Cr
Iroquols
M Fk Whit*
t Fk IlUnots
•Hack
Buckcy*
U loin*
Boundary Cr
Branch
llnvill*
Shcyenrw
Pishing Cr
OuinnlpUe
Drowning Cr
9ioshon*
QnconpaHgn
fct 1C*
Big
N 1r rbtOJMC
Ani-nas
Mr lean
8r and/wine
Salt
Whit*
L Tallapoosa
E 8 Qtiowar*
Dig Sunflow*r
Hiranwc
Ch*nurg
Arkansas
Cagl*
Allvgash
Q inch
Seat*
1A
HC
KT
fA
IN
AZ
OR
WI
CA
IL
ID
RX
NC
MD
NC
cr
NC
WY
CO
FL
HO
GA
no
CO
SJ
PA
AZ
vr
AL
NY
MS
MO
NY
GO
CO
ME
TO
Poriol of
o*c"jr\i C
-------
Ttblt 3-1 (Continued)
Station ID
05515500
02490500
01315500
01610000
0538(000
02369000
07378500
06465500
02135000
03110200
02076000
034S50HO
05333500
06297000
03107500
13341000
07341500
02350500
01100000
14233400
River name
Kankake*
Bouge Chit to
Hudson
Potomac
foot
Shoal
Mite
Niobrara
Little Pee D»e
Brazos
Dan
French Broad
St. Croix
•Hgtvarn
Reaver
S F Clearwater
Md
Flint
Merri-nack
Gowlitz
State
IN
MS
MY
W
Mil
FL
LA
MS
sc
TX
VA
TO
WI
.'TT
PA
10
AR
GA
MA
W*
B._ *» 1 MuJ
Pec ion
MCJCl
1926-78
1945-81
1908-78
1D39-83
1938-61
1939-32
1939-83
1939-83
1942-78
195,1-78
1924-52
1901-78
1914-n
1935-79
1957-83
1927-68
1928-81
1930-58
1924-83
1968-78
-t
oc
c/
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
Cesign flow* (*tVv»c)
1-rJay 3-/«ar
167.6
187.5
170.0
202.2
239.3
270.5
282.1
199.7
298.7
277.7
321.6
494.3
477.5
364.0
539.9
469.6
537.4
262.5
284.0
934.7
J4-\3«y 3-ycar
J
174.2
189.6
191.9
219.6
239.7
29-5.0
295.5
304.3
298.9
305.3
380.4
535.5
508.5
520.2
557.5
613.0
603.3
731.0
797.3
959.9
I
i
,
fff
?r;»
V* ".
1.04
1,01
1.13
1.19
1.00
1.16
l.OS
1.52
1.00
i.10
1.14
1.09
1.06
L.43
LOT
1.31
1.12
2.79
2.81
!.-53
*CV • Coefficient of Variation
3-10
-------
For £urt?\*r cUr ideation of th* Mologioally-*m«J
.x E,
, r.j*«c to
3 - 11
-------
s~crroi 4. COMPARISON or THE rwo METHODS
4.1 fesfr|n Flows
Table 4-1 shows how th«j blolo«jLcally-b«ed 1-day 3-y*ar low JVv.*
and the hydroloaiCJlly-'iased 1010 low flows for the slxt£ exanple rivers.
t!i« Mbl« «l*o presents th* difference betwjen 4-day 3-year low flows and
th» TQIO low flows.
For 39 of the 60 stream, the 1-rlay 3-year low flows are
less than th« 1010 low flows. For 18 streams* the 1-day 3-y«ar low
flows are gn»Jti»c titan Uw lOin low flows, and for the remaining
3 stream the differences are less thai 0.1%. Thus, for the majority of
th* stream the 1-day 3-year low flow is lower than the 1010 low flow.
For all Sixty streams, the difference between 1-day 3-year low flows
arri 1010 low flows (U-iJsy 3-year) -( 1010 »/(l-1*y 3-year) ranges from
-50.0* to 20.8*, with the mean and median equal to -4.91 and -3.11,
respectively.
4 - 1
-------
Tabla 4-1. Grparison of 101'J *nd 7Q10 wit.n 1-day 3-yr anJ 4-day 3-yr
(all flews In ftVsac.)
.Uv«r Nam Staeo
Bull Run
Trent
Birch Cr
Baavar Cr
I rogue is
N Ik Wilta
E Ik Illinois
Black
Buekay*
U Hoina
Boundary Cr
Branch
Llnvlllla
SHayanna
Fishing Cr
Quirmlpiac
Drowning Cr
Shoahona
Unccnpang ra
Paaea
Big
Middca Oconaa
N Br Potomac
Animas
Raritan
Brandyvina
Salt
White
L lallapooaa
E B Dalawara
Big Sunflowar
Mvramac
Qtamjng
Arkansas
Eagla
Allagaah
Clinch
Grays
CaJuba
VJV
NC
Iff
WA
IN
AZ
OR
wi
CA
IL
10
U
NC
MO
NC
CT
NC
w*
CO
FL
HO
GA
NO
CO
NJ
FA
A2
VT
AL
NY
MS
*O
NY
CO
CO
HE
W
W
AL
CoRparison
at CMC Cfesign
IQ10 l-day 3-yr
0.3
1.4
1.7
2.4
3.4
4.B
6.4
S.S
7.1
9.3
11.7
8.8
13.4
15.9
17.0
17.5
38.8
41.8
35.6
49.0
46.4
49.4
54.7
54.8
54.2
61.4
64.6
7S.3
72.7
80.8
89.4
88.8
89.7
99.9
116.9
124.5
128.7
122.9
151.9
0.2
1.4
1.7
2.8
2.4
4.8
5.8
5.0
7.0
8.9
12.0
10.0
13.0
15.4
12.0
14.9
33.9
42.9
39.9
48.0
45.0
33.0
42.9
60.0
46.9
S5.8
63.0
75.9
57.9
82.0
82.7
89.9
85.7
89.9
120.0
134.0
127.7
124.8
122.8
Flows
tOIFF*
-50.0
0.0
0.0
14.3
-41.7
0.0
-10.3
-10.0
-1.4
-4.5
2.5
12.0
-3.1
-3.2
-41.7
-17.4
-14.4
2.6
10.8
-2.1
-3.1
-49.7
-27.5
8.7
-15.6
-10.0
-2.5
0.8
-25.6
1.5
-9.1
1.2
-4.7
-11.1
2.6
7.1
-0.8
1.5
-23.7
Comparison
7QIO
0.4
1.6
2.4
.2
.9
.3
.7
.7
7.7
9.9
13.1
13.3
16.4
18.3
19.4
32.3
43.4
46.8
50.8
55.3
55.3
57.4
61.6
62.3
67.1
67.2
68.7
85.2
88.3
89.7
91.9
92.2
97.5
120.1
131.0
134.1
135.2
144.5
156.4
of CCC Osign
4-day 3-yr
0.4
1.6
2.4
3.4
3.0
5.3
6.9
6.1
7.2
9.4
13.0
13.2
15.0
17.6
13.5
34.0
36.2
45.8
49.0
55.2
51.5
45.7
49.0
61.1
53.6
59.3
69.5
96.0
70.2
91.4
85. 4
92.7
92.5
114.0
126.0
138.4
132.2
135.8
149.8
Flos
ID IFF*
0.0
0.0
0.0
5.9
-30.0
0.0
2.9
-9.3
-6.9
-5.3
-0.8
-0.8
-9.3
-4.0
-43.7
5.0
-19.9
-2.2
-3.7
-0.2
-•J.4
-25.6
-25.7
-2.6
-25.2
-IX 1
1.2
0.9
-25.9
1.9
-7.6
0.5
-5.4
-9.3
-4.0
3.1
-2.3
-S.4
-4.4
* %Diffaranc» • ((1-day 3-yaar flow) - (1010)) • 100 / (1-day 3-ya«r flow)
"tOiffaranoa * ((4-day 3-yaar flow) - (7Q10)) * 100 / (4-day 3-yaar flow)
4-2
-------
Table 4-1. (Co-stinjed)
River N*n» State
Xankakee
Bouge OUtto
Hudson
Potorae
Root
Shoal
taite
Niobrara
Little Pee Dee
Brazos
ten
French HroaJ
St. Croix
Bighorn
Beaver
N F Clears t*r
il*4
flint
Mtrriinac*
Cowlitx
IN
MS
trt
w
f«f
FL
LA
ME
SC
TX
v\
TTJ
WI
HT
?*
10
AR
G*
HA
WV
Comparison
of CiC Qisign t*ows
1010 1-day 3-yr
179.0
188.6
207.7
209.6
229.7
280.1
298.1
160.9
306.7
311.6
329.4
473.6
SOS. 9
327. J
571.3
529. 2
691.0
207.8
270.2
901.5
167.6
187,5
170.0
202.2
239,3
270.5
282.1
199.7
298.7
277.7
321.6
494.3
477.5
364.0
539.9
469.6
537.4
262.5
284.0
934.7
1DIFF*
-6.e
-0.6
-22.2
-3.7
4.0
-3.5
-5.7
19.4
-2.7
-12.2
-2.5
4.2
-5.9
10.1
-5.8
-12.7
-28.6
20.8
3.6
4.9
Coipar i*on of CCC Des
7010
184.
191.
211.
220.
245.
291.
303.
322.
322.
344.
387.
532.
536.
557.
594.
648.
769.
799.
929.
968.7
4-day 3-yr
174.2
189.6
191.9
219.6
239.7
286.0
295.5
304.3
298.9
305.3
380.4
535.5
508.5
520.2
557.5
613.0
603.3
731.0
797.3
959.9
ign Flows
10 IFF"
-5.8
-1.1
-10.0
-0.5
-2.5
-1.9
-2.7
-5.8
-7.9
-13.0
-1.8
0.6
-5.4
-7.1
-6.6
-5.S
-27.5
-9.4
-16.6
-0.9
* I0ife«r*nc« • ((1-day 3-y«ar flow) - (1Q10)) * 100 / (1-day 3-y«*r flow)
**1Dieftr*nc* - {(4-day 3-year flow) - (7Q10)) * 100 / (4-day 3-year flow)
Similar comparisons can be made between the 4«day 3-year low flows
and the 7010 low flows based on table 4-1. Fbr 46 of the 60 strea>
-------
4.2 Recursions
Table 4-2 prosant* t3w calculated nvmber at excursion £>
In the W atro*™ Cot the lew flows calculated usiixj the
enJ biologically-based mutUods. The table demonstrate* the t^act of
the choice of one design flow method over the other In teem of nunber
of excursion!*, foe any stream, a higher flow will always result in the
•awe or a greater nunfotr of excursions than a lower flow. Occasionally,
th* difference in the nunter of «*curslons of the two «l*«lGn flows
1» quite dramatic even if the difference between the two design flcvs is
quit* *tv»ll. For eKanple, the 1Q10 and the i-day 3-year design flow of
the Olinniplac River in Connecticut are 17,5 ftVsec and 14.9 ftVsec,
respectively* hut the corresponding nunber of excursions were 39 and 13.
Similar observations could be made for
-------
Tadl* 4*2. Cvparison of «unl>»r of excursions of 101'' «»>1 7Q10
of «xcjrsions of 1-day 3-yr and 4nJay 3-yr
Rivar Naa»
Bull ftm
rr*it
%irsh Cr
*tav*r Cr
Iroqtjoia
M fk Whit*
•: fk Illinois
Black
luckay*
La Hoina
•Va-wJary Cr
Branch
Linvilltt
Sh«y*nna
Pishing Cr
Quinniplac
Drowning Cr
Shoshom
Uncorpahgr*
Paaca
%i?
Middle Ocona*
ft ir Potomac
Ani-nas
Rtritan
Brandywin*
Sslt
white
L Tallapooaa
C 8 Dalawar*
Big Sunflower
Meramc
Chanting
Stab*
VA
NC
"ft
IA
IS
AZ
OR
m
CA
a
to
RZ
NC
ro
NC
CT
NC
w
CO
n.
HO
G\
MO
CO
HJ
PA
AZ
VT
AL
MTT
MS
HO
MY
Conparison of
1010
0.3
1.4
1.7
2.4
3.4
4.9
6.4
5.S
7.1
9.3
11.7
8.8
13.4
15.9
17.0
17.5
38.8
41.8
35.6
49.0
46.4
49.4
54,7
54.8
54.2
61.4
64.6
75.3
72.7
80.8
89.4
88.8
89.7
tCxcur
19
9
8
1
18
2
13
27
13
33
15
10
21
11
17
39
26
3
7
17
23
25
29
0
25
30
21
20
6
17
31
17
26
CMC Design
l-«Jay 3-yc
0.2
1.4
1.7
2.8
2.4
4.8
5.8
5.0
7.0
8.9
12.0
10.0
13.0
15.4
12.0
14.9
33.9
42.9
39.9
48.0
45.0
33.0
42.9
60.0
46.9
55.8
63.0
75.9
57.9
82.0
82.7
89.9
85.7
Flows
I'xcuc
10
9
a
6
9
2
12
21
7
20
IS
13
15
6
15
13
12
6
13
16
15
11
14
2
13
14
18
20
3
20
8
18
18
Co:Aris>>n of C"C >sigr
7310
0.4
1.6
2.4
3.2
3.9
5.3
6.7
6.7
7.7
9.9
13.1
13.3
16.4
18.3
19.4
32.3
43.4
46.8
50.8
55.3
55.3
57.4
61.6
62.3
67.1
67.2
68.7
85.2
qa.3
89.7
91.9
92.2
97.5
»Excuc 4
8.50
9.25
9.25
4.00
16.75
4.00
11.25
26.00
10.00
24.50
15.75
18.25
25.00
14.50
29.25
11.25
27.75
9.25
17.50
17.25
27.75
23.25
23.00
6,75
24.25
33.00
17.25
20.75
7.00
19.00
30.25
16.50
25.00
-lay 3-/r
0.4
1.6
2.4
3.4
3.0
5.3
6.9
6.1
7.2
9.4
13.0
13.2
15.0
17.6
13.5
34.0
36.2
45.9
49.0
55.2
51.5
45.7
49.0
61.1
53.6
59.3
69.5
66.0
70.2
91.4
85.4
92.7
92.5
i Fl:
.*u-
*.i
3. 1
}. 1
6. 1
9."*
4.0
11. 5
24.5
fl.5
20.5
15.*
14.0
. S. "*:
'.5.'
?.2:
;3.r
: 2. ?!
<;. J-
• • >
.i.l'
3. 2:
•4.1!
14.7;
2.5'
13.2:
19. r
• * •*•
11.5!
3.7;
20. 5C
13.?:
n.o;
JO. 5"
4-5
-------
TJ&!« 4-2. (Continued)
Rlvtr Nans
Arkansas
EM!C
Mlagash
Cl inert
Grvys
Cahiba
K^nkako*
Boo^t Chit to
Hudson
Potomac
Root
SHoal
Anit*
Hiodrara
LittU PC* DM
Brazos
Din
Fr*nch Sroad
St. Crolx
iljHorn
9*JV*r
*i * CltirwAtar
Rtd
Flint
M*rrrnack
Cbwlitz
5tat«
CO
CO
••'£
TM
V*
AL
TN
HS
W
W
**N
FL
LA
NE
SC
IX
VA
TN
wt
MT
PA
ID
W
C\
MA
?A
Conparison of
1010
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
473.8
505.9
327.1
571.3
529.2
(591.1
207.*
270.2
901.5
CtC assign
lExcur l-day 3-yr
23
9
15
23
10
33
34
13
30
19
7
20
19
4
15
11
11
13
34
12
IS
20
29
7
13
0
115.8
120.0
134.0
127.7
124.8
122.8
167.6
187.5
170.0
202.2
239.3
270.5
282.1
199.7
299.7
277.7
321. («
494.3
477.5
364.0
539.9
469. <
537.4
262.5
284.0
934.7
Flows
lExcur
26
11
17
17
10
10
14
10
29
14
7
L2
14
8
12
4
9
18
22
14
4
13
17
9
18
2
Cwrrarison of CCC >si;n .-'.>
7Q10
126.1
131.0
134.1
135.2
144.5
156.4
I'M. 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
tCxcur
28.00
17.50
13.00
25.00
19.75
24.75
29.50
19.25
27.75
15.00
10.75
19.25
14.00
11.25
15.00
6.75
10.25
16.00
34.50
16.50
13.25
14.75
28.75
20.25
41.75
4.50
123.
126.
138.
132.
135.
149.
174.
189.
191.
219.
239.
286.
295.
304.
298.
305.
380.
535.
50fl.
520.
55?.
513.
$33.
"31.
79^.
959.
8 ?«.:
0 !!.'
4 "V
2 :?.:
8 •-••), r
8 '.*.
2 ii.
6 U."
9 ?*.:
6 .4.'
7 T.:
o i-.:
5 4.
3 •<.-
9 .1.:
3 :.
4 >.•
5
5
2
5
^ i
** .• • .
3
0 -I.
3 {?.
9 2.
4-6
-------
The hydrologically-based design flows may actually provide a. greater
of protection oC water quality In cases where the value of the
design flows are less than that ol 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 hydrologieslly-based design flows are not greatly
different fron the corresponding values of biologically-based design flows.
The biologically-based design flu** are not always smaller than the
corresponding hydrologically-based design flows for a given stream. Thus,
it cannot be stated that chousing 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* th*
biologically-based method will always provide insurance that th* design
flow calculated will have resulted in no more than the required nurtoer of
excursions.
Based upon the above, both the hydrologically-based and the bio-
logically-toase<1 methods for calculating stream design Clows are recommended
for use in steady-state modeling.
4-7
-------
SECTION 5. R£aOMM£NaffI>6
1. If steady-state modeling is used, the hydro log Leal ly-bised or the
biologically-based stream design flow method should be used. If the
hydrolcgically-based method is used, the 1Q10 and 7Q10 low flows should
be used as the CMC and CCC design flows, except that the 30QIO low flow
should be used as the CCC design flow for amnonia in situations involving
POrWs designed to remove amonia where limited variability of effluent
pollutant concentrations and resulting concentrations in the receiving
water can be demonstrated.
2. Other technically defensible methods may also be used.
5 - 1
-------
SECTION 6. REFERENCES
1. U.S. EPA. 1995. Technical support coc'jrient tor water-wality ixised
toxics control. Offlo* of '-later, vtashington, OC. Saptr-dwr/ 1935.
2. U.S. *PA. w»ter Quality Criteria. SO FR 307«4 July 29, 1985
1. Stoptan, C.G., D.X. Mount, D.J, H4i«en, J.H. Centll*. C.A. Chapuin and
M.A. Btungs. 1985. Oiictelinea for ctorivlng nuntrical national water
oualtty crit*rial Cor th« protection oE aquatic i>r>jani«
-------
A. .Calculation ot HygLroloqically.-oasedDesign Flows
flo
ml
ml
til
This method is only appropriate when the desired return period is less
than n/S years (1).
the second method fits the historical low flow data to a specific
probability density function and then conputes from this function the
flew whose probability of not being exceeded 1$ 1/y- The log Btarson
Type III distribution is a convenient function ta use because it can
accommodate a large variety of distributional shapes and has seen wide-
spread use In streamflow frequency analysis. However, there is no physically
based rational* for choosing one distribution over another.
TJw xOy lew flew based on the log Pearson Type lit method is
• exp( u + R{g,y> s)
th* a-th lowest annual lew flow of record
Un+D/Vl
l(n*i)/yl + I
th* largest Integer less than or equal to z
(n*l)/y -
where u • mean of the logarithms (base e) of the historical annual
lew flews,
s • standard deviation of the logarithms of the historical low flows,
g • skewnes* coefficient of the logarithms of the historical low
flews,
X » frequency factor for skewness g and return period y.
A - I
-------
A sample listing of frequency (actors is given in Table A-l. These factors can
also be approximated as
K - (2/g)t (1 + {g z>/6 - g2/36)3 - 1]
Cor 111 1 * where z la the standard normal variate with emulative probability
1/y (2). Tables of the normal varlates are available In most elementary
statistics texts. Vi approximate value (3) can be found from
* - 4.91 I U/y).M-(l-l/y)'14]
Tb Illustrate the use of the two xQy low flow estl-nation methods, the data
in Table A-2 will be analyzed for the 7Q5. The flow values In this table
represent the lowest "7-
-------
For the It*] Pearson Type III method, the fr«?uency factor K will be estimated
f- -^ TaMe A-l. For skevness of 0.409 and a 5-year return .rwriol interpolation
results In K • -0.956. The 7QS low flow Is
70S • *xp(6.01 + (-.856)(.24))
1.8 cfs
For purposes of conparison, K will tie estimated using the fonculae given
z - 4.91 [ (0.2) -1<-(1-0.2>-14J
• -0.840
K • (2/.409)[(l * t.409){-.840)/5 - (.409)/36)3 - 1]
- -.8.53
• e»p(6.01 * (-.853H.24))
- 331.8 cfs
"Die difference in the three estimates of the 7QS Low flow is less than 2 per=e~.:
A - 3
-------
Tabl« A-l. Frequency Factors (K) for che log Pearson Type III
Sk«wnesa
Coefficient
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
-1.4
-1.6
-1.8
-2.0
-2.2
-2.4
-2.6
-2.8
-3.0
Return Period,
5
-0.636
-0.666
-0.696
-0.725
-0.752
-0.777
-0.799
-0.817
-0.832
-0.844
-0.852
-0.856
-0.857
-0.855
-0.850
-0.842
-0.830
-0.816
-0.800
-0.780
-0.758
-0.732
-0.705
-0.675
-0.643
-0.609
-0.574
-0.537
-0.499
-0.460
-0.420
Years
10
-0.660
-0.702
-0.747
-0.795
-0.844
-0.895
-0.945
'-0. 5*94
-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-Oay Lev HCVS (ftVsec) for the Anute River Near
Henham Springs, LA
Year
1939
1940
1941
1942
1943
1944
1945
194?
1947
1948
1949
19SO
1951
1952
1953
1954
1955
1956
195?
1958
1959
1960
1961
Flew
299
338
355
439
371
410
407
508
450
424
574
489
406
291
352
309
322
278
369
483
523
385
47«
Rank
5
10
IS
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
Flo*
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
n » 45
u • 6.0
a - 0.23
q - 0.385
A - 5
-------
REFERENCES
1. Linslcy, R.K., et al., Hydrology Cor Erg^.^ecra, 2ne" Edition, «cCrsw-
Hill, New Vork, NY, 1977.
2. Loucks* O.P. , *c al., Water _R«sourc« Sj-Jtans Planning and Analysis,
Prtntict-«all, Cr^lewood Cliffs, NJ, 1981.
3. Joiner arri Rosenblatt, JA5A, 66:394, 1971.
A - 6
-------
APPENDIX 3. Vi Exanole Use Of DFLCW For Affnonia Discharges From
TS\* purpose of this Appendix i* to illustrate the use of the OFLCW
program to calculate biologically-based design flews for ammonia and
compare then with the hydrologically-based des.cn flows of 30010 for
the 13 stream with the lowest coefficients of v«,riatio- shewn
In Table 2-1.
». I Introduction
A* 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 lew 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 amnonia CCC (e.g., a 30-day averaging period)
would be acceptable because the magnitudes and durati^s of excursions
above the CCC would be sufficiently 1 united (1).
3.2 Hydroloqieally-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 - 1
-------
TaDle B-l. Hesiyn flow* and resulting nunber of «l« >3, page D-6). Taole &-1 also includes the njnDer o^ excursions
occurred in eao;i of 13 flow r*coMs for tie hydro logically an-i ii o'..o;i^^'.
hased design flows.
9.4 Comarison off Design Flows
Tabl* &-1 shows that for all 13 str-»-vns the 30Q10 low flow is always
l^ss t-^an the 30-day 3-year low flow. The difference between tie low f
(O0-d*y 3-yoar - 30010)/ 30-l*Y 3-year)) 3.7% to lB.f>\ with the -nean
equal to 10.21. tocause the 30010 lew flow is always lower, it results
in fewer excursions th
-------
8.5 Use of B icloc ica I '. /-^as1 -if 4 ja/s anJ A fre7-e~.cy of
occurrence of ones every three years is used for the CCC. However, foe
arrmonia Ji«c'Mi--j»f< "coi 'OTuJs, a longer ^vnr^jioj i«riol "w/ >* as*1 in
certain cases. According to the national UQC for armonia, an averaging
perio>1 as long as 30 HA/S -nay bo use<1 In situatlois Involving ?OT.4s
lO4igiia.1 to rcnove ammonia where lo^ variability of effluent concentrati
and th* r*«ult iivg concentration* in the receiving waters can '« -'•*''*>'.scr
In cases ^hcro !•» variability can be demonstrated, longer averaging
perio-ls for the acrxxua CCC (e.g., a 30-day averaging periol) ^rjcli 5e
acceptable because the ^agnitules and durations of excursions above the
CCC *>ild be sufficientfy iimite-1.
Section 4.1, "Jr? hydro log ic*Uy-*w«1 -lesi-jn flows have b»
with the biologically-based design flc*»s for the 4-^ay av^ra^-.i;
Cor all ;>^1 lul mts . Appendix R shows a ccrrarison bet jeer. -,-.«
y-based 30-day 3-/ear ICM flcvs and the hydrologic.il ly- Dase^
30-710 Low flows for 13 »treac« for A-wonia. For th'isa 13 itr-4-.s, tw.«
30Q10 flow was always less than the 30-
-------
APPENDIX C.
jljUripJLoXj* J*^
Th* biil<>;icail;/-*>awJ design flow calcjlAtion -wthod is *n iterative
px^lare expiating of five pacts, tn Part I, 1 n th* recor-1 of dally flows. Because the a.i"5i»nt
(instrearO concentration of a pollutant can be considered to be inversely
prrortiixul to ^tr-jn flow, t'w -vjpojortate "runnin.; a^«r*j«^" o" •tir**^
f\'» ^r«» *ct'j-illy "rjnnino harnonic means." (The harmonic rnean of a set
of nj-iner* is t;v» reciprocal of t:w -»rit'\n«tic itean of the reciprocals
of the nunbers.) Thus, "X^lay running averages" should be calculated
*s X/S (l.T] , not as 4?)A, where F is the flo« for an inJi^idual day.
T^roughnut this 'ippendix C, th« term "running Average" <*ill iwan
P^r*. Ill ,1jscri*>»s t'ie calcjlsiion OF N (tit tot.-il ijnber if. exc^rs; i-s
of a specified fl^* in the flo« record). The calculations d*scribe-d ii
Part III will "vt -v-rforwi for a nunaer of '1iff«sr-*nc 'low* tiat *r«
specified in Parts IV and V. In Part IV, initial loner and upper lumts
on the design flew are calculate.'., the njnher of excursions J»t «*ac:ii
limit are calculated using P»rt lit, and an initial trial flow is calculated
by interpolation S»t«*een the low«r and j?per Units. In Part V, success w*
iterations are perfom«»] nunher of excursions.
I- 1. Calculate 2 » O/ I (Y) (365.25 days/year) J
C - I
-------
where D • the nonber of days in the flow record;
Y - the average number of years specified in
the fre
-------
I
!
I
s
I
U
3
S
1
2
~ 2 .2
o S u
Ha
s*£
!:ss
**J1
o "•
!r»
._! C4 A A
8
I
O r>
er i
1 u
«
$
1
-------
Thus the starting date ami the duration (in days) of c.ie
first excursion period will b« recorded. 8y definition, t.">e
mint.iMn duration Is X days.
IIt-3. Determine UK starting datea of, and nunber of days in, each
excursion period in the flow record.
III-4. Identify all of the excursion periods that begin within 120 days
after the beginning of the fint excursion period. {Although
the first excursion period is often the only one in the 120-
day period, t«o 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 lot flow period. Add up the total nuntoer of excursion
days In the first low flow period and divide the sum by X to
obtain the number of excursions in the first, low flow period.
If the nunber of excursions Is calculated to be greater than
5.0, set it equal to 5.0
I I 1-5, identify the first excursion period chat begins after the end
of the first low flow period, and start the beginning of the
second 120-<«ay low flow period on the first day of this
excursion period. Determine the nunber of excursion days and
excursions in the second low flow period.
I I 1-6. Determine the starting dates of, and the nunber of excursions in,
each succeeding 120-rlay low flow period.
III-?. Sen the nunber of excursions in all the low-flow periods to
C - 4
-------
detemin* S • the total number of excursions of Lie
flow \-f ir.t-jrest.
Part IV, Calculation of initial Ibuts of the design flow and initial
trial flow.
IV-l. Us* L • 0 as the initial lower limi".
IV-2. (toe U » the XQY lev (lev as the Initial upper limit.
IV-3. Use *t.O as th* nunber of excursions (see Part III) of the
Initial lower limit.
IV-4. Calculate I^j » the number of excursions (See Part til) of t,-.e
initial upper limit.
IV-5. Calculate T - the initial trial flow as T • L *
Part V. Iterative convergence to the design flow.
v-l. Calculate !*j- • the mincer of excursions (see Part III) of t.ie
trial flow.
V-2. If -0.005 £ «>*i-Z)/Z) £ *0.005, us* T as the design flcv *nZ, set 'J - T and NU • Nj.
If ^
-------
APPENDIX D. Description of the DFL£» Ccnout-er Program
OFUX Is A computer program that can perSom a variety of calculations
related to design flew for any stream Cor which daily flew data are in
STORET. The program Is installed on the U.S. CPA's NCC-IBH computer
and is run under the TSO operating environment. DFLOw consists of two
procedures* the first retrieves the daily flew record for the U.S.
Geological Survey (USCS) gaging station of interest from the U.S. EPA'a
STORET system, whereas the second allows selection of one or more calculations,
After logging on to TSO, the uner invokes the program by entering
the command: exec 'mrfunr.dflcw.clist1.
The following menu will appear:
ESTER THE NUMBER OF THE PROCEDURE YOU WISH TO EXECUTE:
1 RETRIEVE FLOW DATA FBCM STORET
2 PERFORM CALCULATIONS USING RETRIEVED FLOW DATA
3 EXIT THE PROGRAM
If procedure 1 is selected, the user will be asked for the 8-diqit USCS
station nvnber for the Clow gage of interest and a 2-digit state code
(see Table D-l). Gating station nunbers can be obtained from local 'JSGS
offices or through a separate retrieval from the STORET system. After
this information is entered, a batch job is automatically submitted to
the IBM system to carry out the STORET retrieval. The user may log off
the system at this point because the retrieval might take several hours.
An exanple flew retrieval session is shown in Table D-2.
After a period of time, the user can invoke the DFLOW program again
and select procedure 2. If the Clow data have not been successfully
retrieved, the message 'FILE NOT AVAILABLE* will appear. If the retrieval
D - 1
-------
is not successful within about six hours, a new retrieval can be attempted.
After a successful retrieval, procedure 2 will allow one or more of the
following to be calculatedi
1. A biologically-teased OC design flew using a 1-day averaging period
and a frequency of allowed excursion* of once every three years on
the average. After the OC design flow ha* been calculated and the
excursion table printed for that flew, any flew* can be entered
in order to obtain OC excursion table* for those flows.
2. A biologically-based OCC design flew using a 4-day averaging period
and a frequency of allowed excursion* of once every three years on
the average. After the OCC design flew has been calculated and the
excursion table printed for that flew, any flows can be entered m
order to obtain OCC excursion table* for those flews.
3. One or more user-defined design flow*. -If a biologically-based
design flew is selected, ehe user will be asked to input six variable*
so that the desired design flew and excursion table can be printed.
If a hydrologically-toased design flew is selected, the user wi;i be
asked to input four variables so that the desired
-------
A copy of th* FORTRAW aourcv oode for DFtfltf can b* obtained froa
Lewis A. Reassert, WTOL, U.S. EPA. 26 West St. Clair Street, CincLnnati,
OH 45268 (T*l«phcn« $13-684-7603 or FTS - 6S4-76O3).
Tab la 0-1. STOUT Seat* Coda a
01
02
0*
0)
06
01
Of
10
11
12
13
1)
It
17
II
19
20
21
22
23
24
25
2*
27
2B
29
Alabau
Alaaka
Xrk«a*««
California
C«t0ra4e
Coonccticut
D«Uv«r«
Dx«tr cc «( C«lua»i*
Florida
C«ersi«
Hawaii
Idaho
Illiaoia
Iowa
laatuck?
Lowiiiaaa
Main*
Karylaad
Naaaach«aatt»
30
31
32
33
34
33
36
37
31
3t
M)
41
42
44
43
44
47
4«
49
50
51
33
Kooeaaa
fevatfa
(Lavpabira
Jaraajr
Raw Naii ee
Haw Tork
Hortli Carolina
North Dakota
Ohio
OhlahoM
Orago*
Pa»aaf Ivaaia
Di«4a talao^
South Caroli«a
South Dakota
Taaaaaaaa
Taxaa
Utah
Varvomt
Virgiaia
Waibiagto*
NioMioea
Hiaai«»i»
Nitaouri
54 Vaat Virginia
55 Viacoaait
5* Wyo.
0-3
-------
T»bl« 0-2. E»««pl« Flow D«t« Retrieval Using OFLOW (User input it underlined)
ENTER THE NUMBER or THE PROCEDURE YOU WISH TO E*- I:
1 RETRIEVE FLOW DATA FROM STORET
2 PCR/ORrt CALCULATIONS USING RETRIEVED FLOW DATA
3 EXIT THE PROGRAM
Jl,
EKTER 8-01C1T USCS STATION NUMBE1 .... 07378500
EMTEE I-DICXT STORET STATE CODE 13
SAVED
JOi A»C(JOB12345) SUBMITTED
AFTER JOB IS COKPLETED. FLOW DATA WILL RESIDE IN FILE DFLOW.OATA
-------
Table 0-3. Use'of OFLOW for the Amit« Riv«r.
irii M *f-.t m «• xmi « iiw
i N'IIM MI H>< nu vml
i tun wtu'iin 41* «nir~»»
Mil* «i <«MM
I
i - (it. t •
"11 I1«
«• wt" w»
i • «i WIM, • • tut
Mi* tnll *jt KM*
PU>:: * ««•
IKM1B
It I
lt»t tat IIT^CM Pf> Ml -ttnt ntfum •
«t •. I'm I.* «1 », >«• i Ul :
fe » • : «l II. Hk H 4,1
•n 11. i«i » ti
• II. 1*1 U U
> I'. :•* 1.1
«T% l«.« '
i r • ji -ji-it:» ani-u- > i| nuix: M.B •
i • «. i • ent. l
«
i
. • • uit
IKK •« W«l
I • HftMICfel IMI. | • HIM •»«
•MI • IHI ii• Mi •«•<• ^n» M '• at. ••mat
0-5
-------
Tshla 0-3. (Continued)
wi r i«i •'«» inntai <•
» • MM »11> nm 4>ia tim
niH • <•• mam
•in irai «• ••» •>!•, t
111** MB)
UMI cat f ttmtm
i nmt i«
i it. a
m via <«<
i> '• •*•> v •« •!:• >i» ••
• St. ) •«. ) • •» »i«. i • DII
.Ju' "Ml. I • !$!«• «« IV
urn Mmm «>«. M »i
•in «•• «•
••<• • OMB
!•*•>• l««J»
• •on
Ml ••!• ™* "J
D-fo
-------
QUESTIONS AND ANSWERS CONCESNtNG THE BIOLXICUL*-9ASEO ".ET5CD*
0. » Is New aouatic life protection criteria specify that the acute criteria
(dC) and the chronic criteria (CCC) may be exceeded no nore than
one* every three years on the average by 1-^our and 4-day averages,
respectively. They also state that extreme value analyses rnay not t»
appropriate for estimating the ambient exposure condition. Miat is
an extre-Tie 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. Ut's consider a timt-terles 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.
0. I 2s itould you explain how running averages are estimated?
A. Starting with April 1, our first running averago will be the arithmetic
mtsn 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^«ay
running averages for each water year of 365 days.
Q. I 3: By extreme value, do you mean lowest rurning average of the -ater year?
A. In low-flow analyses, the extreme value for a water year is the lowest
r-jnning average for that year.
Q. I 4: So, do I have 30 extreme values from 30 years' flow record consider;:*;
one extreme value for each water year?
A.
Exactly.
• the biologically-based design flow method has been supported by an overwhelming
majority of utter 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 reviewer*. Some of these questions and
related an«wer» are listed here for additional clarification to Appendices
C and 0 of the Guidance. If this paper becomes too long, in a way It defeats
its purpose. So we chose questions baaed on their Ijieortance. We encouraoe
our reader* to be critical about our answers and raise other questions which
they mey consider important. This will help us to improve both the method
itself and its presentation. In this context, readers may contact Hlrantnay
Biswas (FTS-382-7012} or Nelson Thaws (FTS-780-5702)
E - I
-------
I 5:
A.
A.
•7. I 7:
A.
Q, I 3:
A.
You -«*i'l -o •> •-.'•• v»j 4'»Mt r^kirv} th* •sxtrrn v.U .>}••• Hi-* •?•> /TJ rt~
th*i and why do you rank then?
For low flow analys.'-*, CJi'-.iij jvi 'v; '.)•>» fco"» low^r. tj 'u ;'".e-5>- .
For a lov-flow analysis of a 30-year Il-ov record, we have 30 o*tr?->
values. If w* rank then from fclu* lowoit to t'i* 'litest v.il.j*, i.ii
two .^treue valuo* are equal, then wo have one valu<* Cor each of 30
ranks, and the retJru .wrl>J o' t^w fir-it rank*J Clow is ;>rix;ii:
10 />l
that Is equal or less than tn* flow record, then we will not have ta
use any discretion at all. The dIstrihution-Cow, or non-para,-w^r.c
techniijo, is th« best for frequency analyses. 3ut, suppose you n?>i
100-, 200- or 500-year floal and drought for*:.v*ta foe t'w tesign of
a dsm {for use power production and irri7atlon) and we do not have a
flow record of such a loraj period; then, we need to usa scne forn of
distribution to extrapolate to 100, 200 or 500 years. Tv$re are T\ar.y
wtll known Hshri'>itions which can *>e cik>wn on a C43e-ty-cas« sasis.
The new iiCC alsio sake some reference to the Cog-ftsarson typ* III
•Jistrihution -n ^n exanple of th* extre.-ne value analysis. '*nl*
art on the subject of distribution, is it the only distribution
is currently in u^e In the water quality 4iilytic.il field?
the Urtited States Geological Survey us** the Loo-ftsarson ty-p* II!
distribution in low-flov as well *i fl3o1-flow analyse.-!. Ti*/ nade
choice after conducting a study of flood flow analyses using vari?js
other te-r'inique-*. "Pi's choice o' t*:'ini.7u«s5 should be 'Msed 01 t'*.*
nature of the distribution of extreme values. Out. Cor national
consistency of e^ti'Mt**, the USGS chos* this tec'iTi-?;*.
extreme value analytical techniques are often used in the
€ieH, and s*«m to ?» ouit* rdJisonahl*. Is t*itre any biological '
ecological reason *tv/ extreme value analyses are not appropriate
for estimating .i*-*ljo n.>* .^ing the a.^)ienf. -!urati;ical effects
A. In extreme v«lue analytical techniques, only the nost *qx«ure ev«nt is considered, but other, less severe wi thin-year
e.xpusuo* e>Af'it.4 ar>» totally ignored, although their c>r»jl4tiv«« eff-frt
could be sovere. The severity of those snaller wi thin-year exposure
E - 2
-------
:;.»,
>C .so-.rmB drought o>viltl«>v*
severity th* ftxtrrt* axyos'jre events of othar less -than— use
brought coovlini-yu. Sine* th* biological eff-KM if! <: ivil v
nj«t fin-! a w*y to account for all within-yd-ir ensures in
tD the ix>9C extr>tn9 exposure avo-ii: >' o.nj'i year.
Q. « 10: V'jur answer is difficult to follow; would you give an exampl
4. Hydrolajiaw know t!v*t «9 ha1, in various parts of the USA.
dro-ijht «v^nts during tho wat«r years 1925-1932, 1955-1956, ana
• f«w years In the lata seventies. In other ye^n, drought -as rx?t
a* severe. Suppose that in water year 1925, there <*»re 4 v«cy low
running av*r*g«s of which only one was accopt«*l ** tho .» of t'v» /.j*r 1925. Thus
ignoring these 3 running averages of the water year 1925, th* ext
v*lu* vjl:|i« t'wt are lw* *svero tSan th>* 2nd, 3rd ar»J che 4th
running averages of the year 1925, and exclusion of nore severe
excursion *v-»fit* (2nd, 3rd and 4th *xcursioiw of water-ye%r 1925}
result in a skewed estunate of low
4-iay
arly,
li'e enter: -<
what if biological afiout it?
Q. I 11: The iwt'nod vte*crihe-1 to inplement th«
ts call*l a biologically-based mthod.
*. AL-rost *v-»r/ ;>araTiet*r that is u*ed in thi* "wtHal is derivel on t'ie
of either biologicJil, toxicological or ecological consi.-!erjti3".s,
i'u jUr-iMt-irs used in the extronr* vilae •in^ly*-*-? ara jnrela: •••
to biological, toxicological or ecological consideration*.
0. * 12: Nbuld you n*%» t'w £?Ur»7«i that you think are biological, t •xinl.agic*!
or ecolgoical in nature?
of acceptable exposure crvlU. i-xn: 1 hour fnr CX 4.-sd 4
days for CCC are biologically derived.
3 years on tlw average is the allow*] ecological «sc>/ery
after a single excursion (see Table D - 2 of Appendix D of th«
Technical Support Document for Water Quality-toased Toxics Control
jncy?
E - 3
-------
It is true that neither 15 yw arc four*! ir\ r:\g
reference TV>le. But vh.it I* available Is th-it rivers a?*1 scrii-n
are f-ally r->; jv-v-' '>H:*M-I 5 to 10 yean *ftvr .1 -i-iv^r... '<;>wi(ri
went. Aquatic biologists consider that repeated within-yeae
exposures* i;iii c<**ult in catastrophic afftsi.-i-.H. In t'mr ju--J;e"vnt ,
10 years' exposure interval is inadequate because under that sitjatian
the eooloqy of th* recalling iy*te* will he under c-jont-vii; *.;,<•.!*<
attd recovery. 9y tho same token, a 20-year interval was considers!
to 'oe unneownrlly strHjsMt Cor ««ttv»»lthy liota. Vter
the«e consi4>tration!« and Abates anong biologists and wasteloa>1
allocation coorllrvitOM, w* decided to us« 15 yoan as art jjc-
interval after a severe exposure event consisting ot several wi thin-yea:
exposures.
0. I 14: Have you anything to *ay about how you decided to Allow S excursions
in an interval ot 15 years?
A.
w^C allrv an excursion o-i.» e^ry three years on the
Since the effects of excursions are emulative, ecological recover/
fran A «rf*MO? exoosure event r->j-Hr>j.-< .i:>xit IS y*ir* and the
recovery period fron a single exposure event, according to the
national MQC, is 3 years. ?ieref.->re, 15/3 or S excursions are
accepted as the -upper limit of wlthln-year excursion counts.
0- t 15: Why di«1 you not chco#* a 12-/«»jr interval for 4 within-yaar exposure
events? Or could you not choose an 18^/ear interval Car 6 wi^iin
exposure events (biWJ .-xt the Info available in Tahle O-2 of TSOl?
A. One could make various other choices based on slt»-*pecif ic
but w>» ^is-3e our choice for average conditions.
0. * 16: If 12- or 18-year intervals are chosen for 4 or 6 wi thin-year
con*Utii"n, would the Jesiign flow be different from that of tv 15-.'ej
interval choice? Co we have any idea about how different the CCC
or CMC flow will be for the choices of 12- or 13-year iiter/al?
A. So, we >iid not perform such analyses or comparisons hut our guess is
that the difference will not i>*
Q. I 17: It is understood that, if a 15-year interval is chosen for ecological
recovery, then 5 wi thin-year exposures may be allowed because '»QC
specify 1 exposure on th* average of every 3 years. But some extrerx?
drought related low flow periods might include 1«49 than 5 wi thin-year
exposures, and seme more severe lew flow periods include more
than S within-year exposures. Zf exposure effect.** <»rw cuiulative,
why not include all exposures within a year; why liwit 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, S is the upper limit* an4 the lower limit is 1.
E - 4
-------
0. • 19: »'*.at if civ* «*U!un-y«iar excursion* for a
ft.w 'v<».*l
or
r.r
!>iol>>jical :>wthod is naturally greater than 5 during *ay, a 50-
-'-Mr drought? In t!>a« years, 'l-w i»y rsi-iia low ?or a
tin*, such as for 40-50 'lays, not necessarily for just 20
days Cor 5 excursions. After all, we c*«\not change nature, can
A. No, w» cannot change nature. But we can modify our approach in s
our objective aftor umlecstaoUmj the cons*iu«*nc'.-* of *>var» -iv-ii
t* made a number of analyses to find out what happens if we account. '.-><•
•11, not Just 5, exojrtioa* that one may expect frot t-vj** ixjst vts-fci
drought years. v*» found that inclusion of all excursion* from
years renulta in the followingi
- Design flow* of ail return perlaH of *v, 3, 5, 10, 20, 50 years,
etc. are conplately dominated by those most severe Jrouj'it y.^rs; a-:.;
• this lead* to axtwwly *tringent design flow*.
There is nothing biological in these analyses. Since the exposjr-s
effect* •*« ou«»jl4tiv*, *hould we not oxmt ^U «<^>s.j'-»> -'.^-il.jss
of how rarely. one may expect them, or ho-f strinjent the resultir-.j
design flo» is?
to
4-vi
>-?i
far
0. I 19i
A. this i* -her* * littVj jntemandinc.of ecol«>jic^l '^•:v-*r/ an
familiarity with the torth .^mtrlcan aouatic life are necessar/
"lake a rea«»xvi Ww iurroundlrxj eco4y*te^, in -i\i~i r»
-«ay take nlace. So, in our judge-wit, a recovery period of 15
is ade^u^te for situations ^here the nanlxr of exposures in a \
flow period is S or more.
0. • 20: t*iat is d*»crihe-1 here in th* biological method is similar 1> *
is done by hydrologists for partial duration series. They adir
the pcjitlem as ing traditional staci*tic«t approach. Why did you
not use a classical statistical method?
Pint, ihe *t.-ttlstic«l science of partial •1or^l:io^ *»ri?*, >irticjlir'. /
in the hydrologic field, is not well developed. Sot many people
understand It. Although the biological metliod lac'<* .*t^ti*tij«l
elegance, it is simple and can be used and understood by field
biologists art] ongineers, alike. W% -*x»ld not be siir;tri*ed if a
statistician cones up with a better statistical answer for the
problem that we have in hamj. But it *>jlJ be tapxtant for civ*
regions to uMerstand most aspects of the method if we
them to us* it.
r. -
-------
A.
wjority ••>' fi-J *•: •<«-*» -i -.'.* 'J.S. ;«•••>?
the TQio low fix *•» the design Clow for wvit -o «***ritially had as
a not-to-be *xc.»;-V.'J -uvjl.J .vai.>.»r '.I3C value, I* ***"* t:\-st :t oroi
fine, although a rational* for such a choice is hard to cot* ty. '.'.-•
is it -vj Lvjrtnt •«* to 'i.ivc * ratio-nl ')i >l >ji-:»! l/-5i?^i
method to I'iplervnt the t-jo-fluni»r io/l-1e * rational "«tho1 for f.Mrw -vijor
First, lack of a MolojlcJily-fmed .-netxxl in the rxast
led to i;'vj il^tlon of d«nl«jn flows such Vs 3Q20, 70l'\ 30Q10,
33Q2, and «van th« annual av«rag« tlo* for Identical ««ter use. A
technically .1«»F^MHli;l.t .»»t:i.>l *tll briig *bout tec'xuc^l con^i^u-K-/
for any rtwured levol of protection. Second, th« Introduction
of •:'.>* tv*>-nu.iiL>«r •w'-.to-ial "«1X, whoU efflux-it t-xici^/, *«v1 the
guijanc-j on alt»-<4>*cific water quality sta;»33r1s ha-.-* unalLeraoly
charxjwJ tha .*nvir*»'«nt of Lot
only to .n*r.i>Ml twy^-j^bero-i '.JQC, but also to »t:wr j 7Q10 or oi'vjr '
Q. » 22; Why is th« binl»jically->vr* Direct'./
tis-*-l •>» t'w -<»t-»f •piU'-./ criteria th^n tw ii/.1rul'>j'c-«U/-:u«ei
method?
A.
(*or
In the MioV^jiiTaUy-bJ?*! :1, both th«
l«. 4 .lays and 3 years) *« ta'*en directly
in XQY are not. ^*wt of the other aspects of the
approach are also Nit.*1, on biJl.>;i>M'i, •»:>'.
and to*icological considerations. One of the major technical
differences !>jcv«9i the iwthols is that the 3 years in t'\»» .n?
ba*sd .Tethort is an average frequency, whereas the 10 years in ;
hydrologically-bssed approach is a return period.
."^
•rtx-
-.e
Q. I 23: Does it make any difference whether biologists, ecoiogists, and
toxicologlsts understand !KV design flow is calculat«1?
A. Yes, for three major reasons. First, these are the people who
derive the aquatic life criteria. It the criteria are not used io
a manner that is consistent with their derivation, the intended
levol of po)t-j:;tlon will probably not be ac'iie**!. -fecund, site-
specific frequencies and durations will not correctly affect desi;:
flow if the duration -»n>J fw.|j.*.v_y are not directly u*»-1 in t'w
calculation, third, if they understand what parameters affect
data that might allow them to refine their estirutes of
as one hour, four -Uy*, three yean, an! fifteen years.
E - 6
-------
0. t 24:
A.
**', js U*.;ts-< »:'v< •il-pllci'iy of th* biolo^ically-rus * t -nt'o*
I -it nor. clear ho* »n excursion U co-nte-l. r>;»r*. i vwtl :;-i
concentration, any consecutive 4-lay averagu of low-Clow that is
lower than t'w design flow U conit*S as one «j1ay period -*rj la-« thin the 4-day 3-ycar
Clow, then those 6 days belong to an excursion period.
0. I 25:
2. The nunber of wxeursi.vi* In +n 4v*, .nor-f than o»w low Clow period
within a waiter year is possible.
4. The allowed total nuVser of excursions over tha perioS of record
is the nmber of years of record d'ivided by the frecuenc/ of
aouatlc life criteria (3 years for the CCC of the new national
two<«vjnber criteria). For exanpie, if we hava a 30-year fl>*
recar<1, then total nijniwr of excjrslons that are allowed f-jr
x-iay 3-year criteria is equal to 30/3 or 10.
5. Th* 4-Jay 3 -year design flow Cor t'w 4-May 3-year CCC ^aserl or
a 30-year flow record of a given river is erual that fl»» ^.:rw
results in no tore th^n t'ie allow^~?le nunber of excursions.
ror example, the total allowable number of excursions for t^.«
given record is 10. the design flow Is the highest f'.cv fiv,
results in no more than 10 excursions calculated as define? it
steps 1 through 4 above.
f-et us t«
-------
tora
i mi i
i
imam
• *
. l
** f m* «• mm i
«• 4
I • Ull
obtain*! * total of 5 atcursion* ->ily, •lU'TSush In reality
there -*»re altogether 30/4 or 7.50 excursions In that low fltv
period. Similarly, •*» fou-*l only 5 excursion* foe total ^wfiit
30 days during the lev flow period of 1963. In 1969, we had 2.S
excursion* !or .1 V>* flo* period that lwt»1 for 10
26: tt s*errw li'<« the accuracy of the design flow astLitates is total I/
i.vt tw \\ of ti\* flow record. Cb you agree ' '
obsarvatian?
.Absolutely.
are
t*ii< ii tru* shout -*ny *'v»ly»ii. Hsre relevant
to provU* »nore accurate infor»ti>t>n.
of
The longer the Clow e»«orJ, the "we reliable the estimated
cof>iiti->n rfill h>». Figure E-l shows how the a;ic*.-U in fw
liiits on the extreme value-based design load with 10-year return
period .1ecre***» wit'i tncwasi'Vj period of record. (Thi* !i-jjr-
•terived on the basis nf lognornal statistics, not log Pearson type 3).
l-»sjlts ar»» .*'v>^i for both low variability (7/-0.2) and high variibi!::.-
(CV-0.8) situations. Based on the behavior of these curves, it appear »
that 20 n-j 1-1 y-»an of record H a r«»^!V>M;)W -nln «••»!« r-*7utre"Wit for
extr«TM value analysis at a 10-year return period.
The case for th* biologically-basal excursion criterion is less
definitive. However, sine* it considers all days within the period
of recorl as its *npl* (not just the worst condition of *<*ch year),
Its sa.-nple 4lze is nuch larrjer than that of an extreme value analysis.
Thus, It may be possible to uw period* of record Lesa than 20 years
with this criterion arvJ still have a good level of confidence in
the result*.
-------
JC }Q
c«»» cr ICC:M
A.
Figure E-l. Spread in 90% Confidence LUits on EstUMting a Quant Ley
wii'i « l^-^<«r Setum Pariod as a Functi.jn -if r.;u*
Record Length (3ariv*d (ran taoles in Stedinger U9831 i
• 23: "'»!*': *>jll ?>i 1o for i'i*:-%r-nittent str-»-»<« w'v»c^ I *» *V>^ U wro
•luring low fix periods? Mso, now will you use the
based -•«t'x>l in sit'Mtirxvi wlvtc* fl<» data ant i>ot
1ft*so ar* problena that arc generic to all (lev estimating techniques.
For int*r->itt*nt stM*« for which th* lo» Clew is zero, th* lesion
flrvs for dC AS well as CCC an equal to zero. In situations
whers flow iat* ar« not .wxiVibl*, field hydrolooiatA ^o>«. .»•>;•. i«».»r3
*5w»ti."«s use flow data from hylrol'kjLcally corparable drainage
29; The t*hl« •giv'-n in Oueftri^n 23 loj<< -li-ple. How i»uch tirw 1>?s
tat* to conduct a biologically-based analysis for «ny strei-t
0. I Ms
o*
is perfoc^*»d in t«o st«^i.t. Ties':, 1aily fls« data are
retrieved from U» daily flow cile in STORET, by su&nitting a Satr?- :;;.
This -ill taf tb«w *t tw J>«\vj|>-*»<'' "*>'-*«'<5^, Vi-s : X
run -»igh^. take anywhere from a few minutes to several hours,
on how busy the computer 4yqt>M L-* it t'i>t tLne of suhmittal.
the data has been retrieved, the analysis can be performed in five
or ten minutes.
tt S««M that th* foundation of the information «ix»ut «»
recovery period* Cor the tuo-nunber WQC Is all that are Uste.1
in Tabl>* ^2 of the TRO. But, mnyboly fMiiliar with the^e cef*rences
will Mil you JJiat the recovery periods listed In that table are
relst»1 to recovery fron cat*-ito>ii«: -txposucis* cause-\ by spvlln,
not by effluents of malfunction*) advanced treatment facilities.
ttould you Age** tutt this is not a satisfactory set of iifornation
to (« <«uch an important decision?
E - 9
-------
A. this is the best available information that we could use to estimate
ecologies' ri-a.-ory. Considering the caiplexlties involved in the
irplener r.?tL-\ of the two ninber WQC, and the sit.j-specif ic woe for
nolluti -VC* and w; -sle effluent toxlcity, we could not leave the
recove./ a «JOL: :/i open to anyone's interpretation. Considering
the potenti ' Jor misuse of the WCC in their implenentation phase,
we had to use our best Judgement and the best information availaole,
although we recognize that our best judgement would be Abatable.
Since the intonation baso 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*
0. I 31: What are you doing to Improve the information base?
A. ORO 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,
toxicnlegists, engineers and hydologists share their experience
than doing mere literature review* and/or gathering laboratory-
generated information.
REFERENCE
1. Stedinger, J.R., "Confidence Intervals for Design Events',
Jour, Hvd. Eng. Div., ASCE, Vol. 109, No. 1, January 1983.
E - 10
-------
------- |