CRWR Online Report 06-11
Assessment of Hydrologic Alteration Software
by
Eric S. Hersh, B. S.
Graduate Research Assistant
and
David R. Maidment, PhD
Principal Investigator
October 2006
CENTER FOR RESEARCH IN WATER RESOURCES
Bureau of Engineering Research • The University of Texas at Austin
J.J. Pickle Research Campus • Austin, TX 78712-4497
This document is available online via World Wide Web at
http://www.ce.utexas.edu/centers/crwr/reports/online.html
2
Executive Summary
The purpose of this project was to assess the applicability of hydrologic analysis tools such as
The Nature Conservancy’s Indicators of Hydrologic Alteration (IHA) and US Geological
Survey’s (USGS’s) Hydrologic Assessment Tool (HAT) for use in the Texas Instream Flow
Program (TIFP) and to evaluate the flow regime in six priority subbasins: Lower Sabine, Middle
Trinity, Middle and Lower Brazos, Lower Guadalupe, and Lower San Antonio.
As general methods of streamflow hydrograph characterization, both IHA and HAT offer many
useful functions that illuminate the nature of streamflow patterns through time at a stream gaging
site. In their current versions, however, neither IHA nor HAT is directly suitable for use in the
TIFP; both would require modifications to successfully be implemented to define flow
component statistics and wet, dry, and normal years. IHA, as its name implies, is best suited to
assess hydrologic alteration and to quantify the effects of dam construction and other such water
management development projects on the flow regime via two-period analyses and the Range of
Variability Approach. HAT, as its name implies, is focused on characterizing streamflow,
particularly in the context of a regional analysis of factors that influence streamflow properties.
If the task of choosing between the two programs is framed as which tool better characterizes
streamflow hydrographs in general, then there is little difference between these two software
packages. But if the task of choosing between them is framed more narrowly as which program
will best support the four-level flow characterization (subsistence flow, base flow, high flow
pulses, and overbank flows) as for the Texas Instream Flow Program, then the HAT program is a
better choice than IHA as it allows for more flexibility in the determination of flow component
thresholds and has a greater capacity for regionalization, as described below. Note that the
indices calculated in IHA and HAT can all be calculated using independent software such as
Microsoft Excel, Matlab, or SAS, oftentimes allowing for greater flexibility.
Using daily flow information available from USGS gaging stations, IHA, HAT, and other tools
were used to evaluate the flow regime at 24 selected gages within the 6 TIFP priority subbasins
for a period-of-record averaging 68 years. The study has demonstrated that there are pronounced
3
regional patterns in the seasonality of the streamflow hydrograph, with little seasonal variation in
central Texas grading to a strongly seasonal variation in East Texas. The degree of spatial
variability in the seasonal pattern of streamflow in East Texas is significantly greater than spatial
variation in precipitation alone can explain.
When considered over a long time scale, the seasonal variation in the flow regime of various
percentile levels of flow follows a consistent pattern, such that if the median flow goes down
then both high and low flow percentiles reduce proportionately to some extent, and vice versa
when the median flow increases. Although this pattern holds true across a sample of priority
gages studied, the high and low portions of the flow regime also each exhibit their own
secondary patterns of seasonality. For the hydrologic conditions present in Texas, the median
flow is the most robust streamflow characteristic that is available (i.e., it is most consistently
estimated with a given number of data values), and it can reasonably be estimated both from
daily and from monthly streamflow data.
Components of the TIFP flow component model may be better delineated using alternatives to
IHA or HAT, such as by using the Standard Institute of Hydrology Method for base flow
separation using the United States Bureau of Reclamation’s BFI computer program. The
overbank flow component has a specific, physically-based flow threshold for any given stream
cross-section based on the stream slope, roughness, and channel capacity. Looking at additional
published gage data (in addition to discharge) provides additional insight into the delineation of
this flow component that discharge data alone can not provide, and can be used to estimate the
incipient point of overbank flooding.
From this study we conclude that a Texas-customized version of HAT (as part of the
Hydroecological Integrity Assessment Process, or HIP) is suitable and preferable to IHA for
application in the Texas Instream Flow Program. Further work is necessary as part of the TIFP
guidance or within the subbasin studies to define the specific role of hydrologic assessment tools,
particularly with respect to flow component delineation and the definition of wet, dry, and
normal years.
4
Table of Contents
Executive Summary........................................................................................................................ 2
Table of Content s............................................................................................................................ 4
List of Figures ................................................................................................................................. 6
List of Tables .................................................................................................................................. 8
List of Acronyms ............................................................................................................................ 9
Acknowledgements....................................................................................................................... 10
I. Texas Instream Flow Program.................................................................................................. 11
Background ................................................................................................................................11
History and Goals.................................................................................................................. 11
Disciplines............................................................................................................................. 12
Hydrology ............................................................................................................................. 13
Instream Flow Regime...............................................................................................................14
The Natural Flow Regime..................................................................................................... 14
Flow Components ................................................................................................................. 15
Hydrologic Characteristics.................................................................................................... 19
II. Instream Flow Models............................................................................................................. 20
Categories...................................................................................................................................20
Model Development...................................................................................................................21
IHA........................................................................................................................................ 21
HAT ...................................................................................................................................... 22
Input Data Requirements and Sources.......................................................................................24
Note on Model Documentation............................................................................................. 24
IHA........................................................................................................................................ 24
HAT ...................................................................................................................................... 25
Model Use and Application.......................................................................................................26
Temporal Scale ..................................................................................................................... 26
External Verification............................................................................................................. 26
Statistical Definitions ............................................................................................................ 26
Environmental Flow Component Definitions ....................................................................... 27
Model Output .............................................................................................................................34
IHA........................................................................................................................................ 34
HAT ...................................................................................................................................... 35
III. Model Application in the Texas Instream Flow Program...................................................... 39
Hydrologic Study Needs ............................................................................................................39
Temporal Resolution..................................................................................................................39
Additional Potential Enhancements...........................................................................................40
Calculation of Medians ......................................................................................................... 40
Environmental Flow Component Thresholds ....................................................................... 42
HAT Display Indices ............................................................................................................ 43
IV. Flow Regime Evaluation ....................................................................................................... 44
Data Sources and Metadata........................................................................................................44
Methodology ..............................................................................................................................48
Problem Definition.....................................................................................................................49
5
Impact and Extent of Flow Regulation......................................................................................50
Measures of Central Tendency ..................................................................................................54
Precipitation and Streamflow.....................................................................................................58
Seven-Day Two-Year Low Flow (7Q2) ....................................................................................62
Range of Flow Regime ..............................................................................................................64
Base flow....................................................................................................................................69
Overbank Flow...........................................................................................................................69
V. Conclusions and Recommendations ....................................................................................... 73
Appendix A – Scope of Work....................................................................................................... 78
Appendix B – IHA Parameters ..................................................................................................... 80
Appendix C – HAT Parameters .................................................................................................... 85
Appendix D – Electronic Files...................................................................................................... 95
IHA and HAT Output for 24 Priority Gages .............................................................................95
Study Data..................................................................................................................................95
Study Maps, Models and Programs ...........................................................................................95
References..................................................................................................................................96
6
List of Figures
Figure 1. TIFP priority studies (TIFP 2002). ............................................................................... 12
Figure 2. TIFP sub-basin study flowchart (TIFP 2006)............................................................... 14
Figure 3. Example daily streamflow hydrograph depicting flow components (from NRC 2005).
............................................................................................................................................... 16
Figure 5. Location and stream classification of the 420 gages of Olden and Poff (2003)........... 23
Figure 6. IHA EFC definitions screen, displaying default thresholds. ........................................ 28
Figure 7. IHA Environmental Flow Component algorithm flow chart. ...................................... 29
Figure 8. Daily Environmental Flow Components for USGS Gage No. 08171000, Blanco Rv at
Wimberley, TX, with a close-up view of a case where low flows (dark green) are of greater
magnitude than high flow pulse events (blue). ..................................................................... 30
Figure 9. Quantitative comparison of IHA EFCs for USGS Gage No. 08171000, Blanco Rv at
Wimberley, TX for the period of record............................................................................... 31
Figure 10. (a) Example instream flow prescription to be developed under SB2 depicting
integration of the flow components, and (b) tabular representation of (a) (TIFP 2006). ..... 33
Figure 11. (a) Example of graphics capture bug present in version 3.0 of NATHAT; (b) same
plot as viewed on-screen in the NATHAT software and captured for comparison here using
screen-capture tools............................................................................................................... 36
Figure 12. Example NATHAT default hydrologic index selection menu depicting selection
options for one of the six stream types. ................................................................................ 38
Figure 13. Various procedures for the calculation of median streamflows: (a) method employed
in IHA, (b) alternative method for daily medians, and (c) alternative method for monthly
medians. ................................................................................................................................ 41
Figure 14. Locus map of 24 study gages in the State of Texas. .................................................. 46
Figure 15. Map of study gages in the Lower Guadalupe and Lower San Antonio River
subbasins. .............................................................................................................................. 46
Figure 16. Map of study gages in the Middle and Lower Brazos River subbasins. .................... 47
Figure 17. Map of study gages in the Middle Trinity River subbasin. ........................................ 47
Figure 18. Map of study gages in the Lower Sabine River subbasin. ......................................... 48
Figure 19. Flow regulation at study gages, where regulated gages are defined as those where
runoff from greater than ten percent of their contributory drainage area is affected by
regulation. ............................................................................................................................. 51
Figure 20. Time trend of maximum 1-day flow in Lower Guadalupe River downstream of
Canyon Dam, which began impounding water in 1964........................................................ 52
Figure 21. Time trend of minimum 1-day flow in the Middle Trinity River downstream of
Dallas-Fort Worth. ................................................................................................................ 52
Figure 22. Median daily streamflow in the Lower Sabine River subbasin.................................. 54
Figure 23. Median daily streamflow in the Middle Trinity River subbasin. ............................... 55
Figure 24. Median daily streamflow in the Middle Brazos River subbasin. ............................... 55
Figure 25. Median daily streamflow in the Lower Brazos River subbasin. Note: Due to the
logarithmic scale, flows shown along the abscissa are 1 cfs or less, and not necessarily 0 cfs.
............................................................................................................................................... 56
Figure 26. Median daily streamflow in the Lower Guadalupe River subbasin. .......................... 56
Figure 27. Median daily streamflow in the Lower San Antonio River subbasin. ....................... 57
7
Figure 28. Streamflow and precipitation patterns for the Lower Sabine River basin.................. 60
Figure 29. Streamflow and precipitation patterns for the Middle Trinity River basin. ............... 60
Figure 30. Streamflow and precipitation patterns for the Middle Brazos River basin. ............... 61
Figure 31. Streamflow and precipitation patterns for the Lower Guadalupe River basin. ........... 61
Figure 32. Modified box and whisker plot for USGS Gage No. 0802000, Sabine Rv nr
Gladewater, TX. .................................................................................................................... 64
Figure 33. Modified box and whisker plot for USGS Gage No. 08065000, Trinity Rv nr
Oakwood, TX........................................................................................................................ 65
Figure 34. Modified box and whisker plot for USGS Gage No. 08168500, Guadalupe Rv abv
Comal Rv at New Braunfels, TX. ......................................................................................... 65
Figure 35. Modified box and whisker plot for USGS Gage No. 08188500, San Antonio Rv at
Goliad, TX. ........................................................................................................................... 66
Figure 36. Modified box and whisker plot for USGS Gage No. 08168500, Guadalupe Rv abv
Comal Rv at New Braunfels, TX, normalized by dividing by the median flow. .................. 67
Figure 37. Modified box and whisker plot for USGS Gage No. 0802000, Sabine Rv nr
Gladewater, TX, minus the monthly median and normalized by the interquartile range. .... 68
Figure 38. Modified box and whisker plot for USGS Gage No. 08188500, San Antonio Rv at
Goliad, TX, minus the monthly median and normalized by the interquartile range. ........... 68
Figure 39. Base flow separation using the BFI program for USGS Gage No. 08171000, Blanco
Rv nr Wimberley, TX. .......................................................................................................... 70
Figure 40. Discharge versus wetted channel width for USGS Gage No. 0802000, Sabine Rv nr
Gladewater, TX, plotted by decade....................................................................................... 71
Figure 41. Discharge versus wetted channel width for USGS Gage No. 08110500, Navasota Rv
nr Easterly, TX. ..................................................................................................................... 72
8
List of Tables
Table 1. Comparison of conceptual models for flow components. ............................................. 17
Table 2. Common characteristics for each flow component for each discipline (TIFP 2006). ... 18
Table 3. Environmental flow model types................................................................................... 20
Table 4. HAT indices requiring the input of peak flow data. ...................................................... 25
Table 5. HAT indices useful for the delineation of the TIFP flow components (refer to Appendix
C for explanation of indices)................................................................................................. 32
Table 6. Ten default hydrologic indices for each stream type in NATHAT. .............................. 37
Table 7. Selected USGS gages for flow regime evaluation. ........................................................ 45
Table 8. Major dams and reservoirs regulating flow at the study gages...................................... 53
Table 9. Comparison of measures of central tendency by proportionally-weighted months. ..... 59
Table 10. Seven-day average, two-year recurrence interval low flow discharge (7Q2).............. 63
9
List of Acronyms
7Q2 Seven-day average, two-year recurrence interval discharge
CCEFN Consensus Criteria for Environmental Flow Needs
CFS Cubic feet per second
CRWR Center for Research in Water Resources
EFC Environmental Flow Components
ESRI Environmental Systems Research Institute
HAT Hydrologic Assessment Tool
HIP Hydroecological Integrity Assessment Process
HIT Hydrologic Index Tool
IHA Indicators of Hydrologic Alteration
NATHAT National Hydrologic Assessment Tool
NJHAT New Jersey Hydrologic Assessment Tool
NJSCT New Jersey Stream Classification Tool
NWIS National Water Information System
PCA Principle Components Analysis
PRISM Parameter-elevation Regressions on Independent Slopes Model
RVA Range of Variability Approach
SB2 Senate Bill 2
SCT Stream Classification Tool
TCEQ Texas Commission on Environmental Quality
TIFP Texas Instream Flow Program
TPWD Texas Parks and Wildlife Department
TWDB Texas Water Development Board
TXHAT Texas Hydrologic Assessment Tool
USGS United States Geological Survey
VBA Visual Basic for Applications
WAM Water Availability Model
WRAP Water Rights Analysis Package
10
Acknowledgements
The authors would like to thank the Texas Water Development Board (TWDB) for
funding and supporting this research under TWDB Project #2005-483-029, and the Texas
Commission on Environmental Quality (TCEQ) and Texas Parks and Wildlife Department
(TPWD) for their valuable input. In particular, the assistance of Jordan Furnans and Mark
Wentzel of the TWDB, Wendy Gordon of the TCEQ, and Kevin Mayes of the TPWD is
appreciated.
11
I. Texas Instream Flow Program
Background
History and Goals
The 77th session of the Texas Legislature passed Senate Bill 2 (SB2) in 2001, directing the
Texas Commission on Environmental Quality, Texas Parks and Wildlife Department, and Texas
Water Development Board (hereinafter referred to as “the agencies”) to “…jointly establish and
continuously maintain an instream flow data collection and evaluation program…” and to
“…conduct studies and analyses to determine appropriate methodologies for determining flow
conditions in the state rivers and streams necessary to support a sound ecological environment."
The Texas Instream Flow Program (TIFP) was developed by the agencies in response (Senate
Bill 2, TIFP 2006).
Six subbasins were identified by the agencies for TIFP priority study based on potential water
development projects, water rights permitting issues, and other factors: Lower Sabine River,
Middle Trinity River, Middle and Lower Brazos River, Lower Guadalupe River, and Lower San
Antonio River (Figure 1). Basin-specific instream flow studies are scheduled to be completed
for each priority basin by December 31, 2010 (TIFP 2002).
12
Figure 1. TIFP priority studies (TIFP 2002).
Disciplines
Instream flow science is both multidisciplinary, relying on multiple different scientific fields, and
interdisciplinary, relying on the interconnectedness of these various fields. The TIFP will
include analyses of hydrology and hydraulics, geomorphology and physical processes, water
quality, and biology, and the connectivity between and among the four primary disciplines. The
integration of sometimes disparate findings from these disciplines is one of the most challenging
and one of the most important steps of developing instream flow recommendations.
Environmental flows encompass instream flows, and the terms are often used interchangeably.
Despite the terminology, “instream flows” is typically used to describe the entire spectrum of the
flow regime including overbank flows, which are technically not “instream.” The term
“instream” flow requirement is used to distinguish these flow requirements from legal
prescriptions on water withdrawals – those taking the water out of the stream.
13
Hydrology
Recent work has come to recognize streamflow as the “master variable” controlling riverine
physical, biological, and chemical processes (Poff et al. 1997, Annear et al. 2004). In other
words, flow is often the driving force in the complex web of fluvial processes. This is because
the quantity and timing of flow are critical to the function, health, and ecological integrity of
riverine systems as they affect nearly every other process that occurs within the system.
Streambank erosion, habitat availability, and dissolved oxygen concentrations are examples of
physical, biological, and chemical components that all are highly impacted by the amount of
flow in a river as flow affects the stream power, water depth, habitat connectivity, water
temperature, reaeration, and numerous other factors that regulate in this example. As such, an
understanding of the linkages between the flow regime and the natural processes that occur in the
river is essential to developing environmental flow recommendations to preserve those natural
processes to the maximum extent practicable. Within TIFP, evaluations of the hydrology and
hydraulics are to be conducted along with evaluations of the other three disciplines immediately
following the Study Design element for the sub-basin studies (Figure 2).
14
Figure 2. TIFP sub-basin study flowchart (TIFP 2006).
Instream Flow Regime
The Natural Flow Regime
In the 1950s, instream flow methods were developed that identified a minimum flow value
practitioners believed to be protective of aquatic resources. However, scientific thinking on the
topic of instream flows has progressed over the last half-century to the point of recognizing that
the temporal variability of flow across hours, days, months, seasons, years, and even decades is
critical in sustaining biodiversity and ecosystem integrity (Poff et al. 1997, Richter et al. 1997).
This variability is represented by consideration of the natural flow regime, which is the full range
of magnitude, timing, and variability of streamflows that have occurred historically or that could
15
reasonably be expected to occur naturally within a given reach. The natural flow regime is
characterized by five components: magnitude, frequency, duration, timing, and rate of change of
hydrologic conditions, and a large body of research demonstrates that protecting or recreating the
natural range of flow variation in a river system serves to protect or recreate a healthy river
(Arthington et al. 1991, Sparks 1995, Stanford et al. 1996. Richter et al. 1997).
Flow Components
The National Research Council of the National Academy of Sciences performed a scientific peer
review of the TIFP in which, among other recommendations, they put forth a conceptual model
to characterize the natural flow regime into four flow components: subsistence flow, base flow,
high flow pulses, and overbank flows (NRC 2005). According to the NRC (2005) document The
Science of Instream Flows: A Review of the Texas Instream Flow Program,
Subsistence flow is the minimum streamflow needed during critical drought
periods to maintain tolerable water quality conditions and to provide minimal
aquatic habitat space for the survival of aquatic organisms. Base flow is the
‘normal’ flow conditions found in a river in between storms, and base flows
provide adequate habitat for the support of diverse, native aquatic communities
and maintain ground water levels to support riparian vegetation. High flow pulses
are short-duration, high flows within the stream channel that occur during or
immediately following a storm event; they flush fine sediment deposits and waste
products, restore normal water quality following prolonged low flows, and
provide longitudinal connectivity for species movement along the river. Lastly,
overbank flow is an infrequent, high flow event that breaches riverbanks.
Overbank flows can drastically restructure the channel and floodplain, recharge
groundwater tables, deliver nutrients to riparian vegetation, and connect the
channel with floodplain habitats that provide additional food for aquatic
organisms.
16
An example daily streamflow hydrograph from the NRC report for the Guadalupe River at
Victoria, Texas (USGS Gage No. 08176500) for water year 2000 is presented in Figure 3 with
flow components identified.
Figure 3. Example daily streamflow hydrograph depi cting flow components (from NRC 2005).
Although the TIFP has adopted the conceptual model and terminology for flow components from
the NRC report, previous iterations of this model had been presented (Richter et al. 1998),
including one within the State of Texas (Mosier and Ray 1992) (Table 1).
17
Table 1. Comparison of conceptual models for flow components.
As discussed above, each of the four components of the natural flow regime controls and affects
different purposes, functions, and processes of the riverine ecosystem, evident in each of the four
TIFP disciplines (Table 2). Many of the flow pattern influences are specific to certain species,
communities, regions, and rivers, and sound instream flow policy lies within the successful
integration of these spatial, temporal, and interdisciplinary factors. Within the TIFP, the specific
characteristics and their relative ecological significance will be identified as part of each sub-
basin study, and specific flows may be recommended that provide specific ecological benefits
(TIFP 2006).
It is important to note that hydrology occurs in four dimensions of space and time. Flow must be
considered longitudinally (down a river’s length), laterally (between river channel and
floodplain), vertically (surface water-groundwater interactions), and through the course of time.
18
Table 2. Common characteristics for each flow component for each discipline (TIFP 2006).
Component Hydrology Geomorphology Biology Water Quality
Subsistence
Flows
Infrequent, low
flows
Increased deposition
of fine & organic
particles
Restricted aquatic
habitat
Limited
connectivity
Elevated
temperature
Reduced levels
of dissolved
oxygen
Base
Flows
Normal flow
conditions with
variability
Maintain soil
moisture &
groundwater table
Maintain diversity of
habitats
Suitable aquatic
habitat
Connectivity along
channel corridor
Suitable in-
channel water
quality
High Flow
Pulses
In-channel, short
duration, high
flows
Maintain channel &
substrate
characteristics
Prevent
encroachment of
riparian vegetation
Recruitment events
for organisms
Connectivity to
near-channel water
bodies
Restore in-
channel
water quality
after prolonged
low-flow
Overbank
Flows
Infrequent, high
flows that exceed
normal channel
Floodplain
maintenance
Lateral channel
movement
New habitat
construction
Flush organic
material into channel
Deposit nutrients in
floodplain
Life phase cues for
Organisms
Riparian
recruitment &
maintena nce
Connectivity with
floodplain
Restore water
quality in
floodplain water
bodies
19
Hydrologic Characteristics
The range and variability of a flow regime is commonly characterized by the magnitude,
frequency, duration, timing, and rate of change of hydrologic events (Karr 1991, Richter et al.
1996, Poff et al. 1997). Magnitude is a measure of the flow volume associated with a particular
hydrologic event; frequency describes how often events occur within a specific time period;
duration is how long an event occurred (above a certain flow rate threshold); timing is when the
events occur within a specific time period; and rate of change is how rapidly the hydrograph rises
and falls (Figure 4).
Magnitude
Timing
Duration
Rate of
Change
Figure 4. Hydrograph depicting hydrologic event characteristics (not shown: frequency).
20
II. Instream Flow Models
Categories
The science of quantifying environmental flow needs has progressed remarkably in the last half
century, and hundreds of methods and models ha ve emerged that seek to answer the question of
how much water a river needs (Richter et al. 1997, Tharme 2003). Based on available data and
resources and desired goals and confidence level, scientists have developed and applied
methodologies from four broad categories (Table 3).
Table 3. Environmental flow model types.
n Hydrologic (Desktop) Models
n Simple, cheap, inexpensive
n Use flow as an indicator for ecological and biological functions
n Examples:
n Indicators of Hydrologic Alteration (IHA), The Nature Conservancy, 1997
n Hydrologic Assessment Tool (HAT), USGS, 2006
n Tennant Method (a.k.a. Montana Method), U.S. Fish and Wildlife Service,
1976
n Lyons’ Method, Texas Parks and Wildlife Dept., 1979
n Hydraulic Models
n Correlate flow with available habitat area based on river channel geometry
n Physical proxy for in-stream ecology and biology, i.e., does not account for
overbank processes
n Examples:
n Wetted Perimeter Method, Montana Dept. of Fish, Wildlife, and Parks, 1970s
n R2-Cross Method, Colorado Div. of Wildlife, 1980s
21
n Habitat Models
n Complex, data intensive
n Use target species population data with hydraulic data to determine optimal
habitat
n Mainly used for economically valuable or endangered species
n Has proven legal credibility in the United States
n Examples:
n Instream Flow Incremental Methodology (IFIM), U.S. FWS, 1970s. Includes
Physical Habitat Simulation Model (PHABSIM)
n Holistic Models
n Very complex, resource and data intensive
n Comprehensive ecosystem assessment
n Based on multidisciplinary scientific consensus
n Examples:
n Building Block Methodology (BBM), South Africa Dept. of Water Affairs and
Forestry and Univ. of Cape Town, 1990s. Top-down approach.
n Downstream Response to Imposed Flow Transformation (DRIFT), above plus
Southern Waters Ecological Research and Consulting, 1990s
Model Development
IHA
According to The Nature Conservancy’s Sustainable Waters Program website (TNC 2006):
22
The Indicators of Hydrologic Alteration (IHA) is a software program that
provides useful information for those trying to understand the hydrologic impacts
of human activities or trying to develop environmental flow recommendations for
water managers. More than 1,000 water resource managers, hydrologists,
ecologists, researchers and policy makers from around the world have used this
program to assess how rivers, lakes, and groundwater basins have been affected
by human activities over time, or to evaluate future water management scenarios.
IHA was originally developed as a result of work done by Brian Richter and others from 1996
through 1998, and is intended to model “what the fish feels” (Richter et al. 1996, Richter et al.
1997, Poff et al. 1997, Richter et al. 1998). In this study, IHA version 7.0.0 Beta 4.10 was used.
HAT
The Hydrologic Assessment Tool (HAT) is a primary component of the USGS Hydroecological
Integrity Assessment Process (HIP) software. HIP was developed in 2004 to 2006 by a team led
by James Henriksen of the USGS Fort Collins [Colorado] Science Center along with the New
Jersey Water Science Center and the New Jersey Department of Environmental Protection. The
current HIP software suite is composed of four components:
1. Hydrologic Index Tool (HIT), version 1.0. HIT is a generic tool (i.e., not developed for
any particular geographic region) to calculate the 171 statistical hydrologic indices
presented in Olden and Poff (2003) based solely on the input of USGS streamflow data
for any gaged site.
2. National Hydrologic Assessment Tool (NATHAT), version 3.0. Commonly referred to
in this document as “HAT,” NATHAT is a nationwide customization of HIT based on the
six stream classifications of Olden and Poff (2003). The six stream classes were pared
down from ten original classes developed by Poff (1996) based on a study of 420 gages
across the contiguous United States (Figure 5). Although NATHAT performs the full
complement of 171 statistical routines, the default graphical presentation of results is
limited to the ten non-redundant, critical indices identified by Olden and Poff (2003) for
each of their six national stream classes.
23
Figure 5. Location and stream classification of the 420 gages of Olden and Poff (2003).
3. New Jersey Stream Classification Tool (NJSCT), version 1.0. NJSCT is a New Jersey-
specific tool to partition that state’s gaged streams into four stream classes, termed A, B,
C, and D, by their relative degree of skewness of daily flows (high versus low) and by the
relative frequency of low flow events (high versus low). The USGS currently has no
plans to develop a national stream classification tool on this basis.
4. New Jersey Hydrologic Assessment Tool (NJHAT), version 3.0. NJHAT is a New
Jersey-specific regionalization of NATHAT incorporating the results of the NJSCT and
the identification of ten primary flow indices for each of the State’s four stream classes.
The analyses presented in this report were predominantly performed using a 2005 beta-released
version of NATHAT, version 2.15. HIP was officially released to the public in June 2006.
Currently, New Jersey is the only state to have completed the Hydroecological Integrity
Assessment Process. The Missouri Department of Conservation is in the midst of the process
24
with USGS as is the Commonwealth of Massachusetts, and the TCEQ is set to embark on the
process as well in 2006-2007 (Henriksen personal communication 2006, Henriksen et al 2006).
Input Data Requirements and Sources
Note on Model Documentation
This document is not intended to be a user’s manual for the IHA and HAT programs; refer to
The Nature Conservancy (2006a) and Henriksen et al (2006) for that purpose. Use of the tools
will be described only where an important distinction exists that affects the functionality and/or
suitability of the program for use in the TIFP.
IHA
IHA requires an input file of daily streamflow data; in the United States, these data are typically
obtained from the USGS National Water Information System (NWIS). To adequately capture
annual and interannual variations in the flow regime, 20 or more years of continuous daily flow
data are recommended (Richter et al. 1997). Currently, there are NWIS records for 957 stream
gaging sites in the State of Texas, of which 406 have 20 or more years of daily flow data.
If data are missing from the input files, IHA performs a linear interpolation across the missing
data gap. If a particular year in the input flow record has no data values, the entire year will be
excluded from analysis. The IHA program will issue a warning message if there is a consecutive
block of missing data greater than a user-defined length, defaulted as 10 days; a warning
message is also issued to identify water years with more than 30 missing daily values.
25
HAT
HAT requires the same input file of daily streamflow as IHA, and the above-discussed issue of
record length remains applicable. If data are missing from the input files, HAT will not
interpolate data gaps and will not use missing flow value days in the statistical calculations;
however, all years within the user-specified period of record will be used in the analyses. In both
IHA and HAT, data gaps of significant length will cause the programs to return peculiar results.
The number of missing flow values that might invalidate or otherwise cast excessive doubt on
the flow statistics returned by the program results is debatable, but the user must be aware of the
effect of missing data when performing analyses. HAT also requires a file of annual
instantaneous peak flow data that is also commonly available from NWIS. If peak data are not
available, the analysis can still be performed to generate all but eight indices (Table 4).
Table 4. HAT indices requiring the input of peak flow data.
Index Definition
FH11 Flood frequency
DH22 Flood interval
DH23 Flood duration
DH24 Flood-free days
TA3 Seasonal predictability of flooding
TH3 Seasonal predictability of non-flooding
TL3 Seasonal predictability of low flow
TL4 Seasonal predictability of non-low flow
Detailed definitions and calculation instructions for these indices along with an explanation of
the index naming convention can be found in Appendix C – HAT Parameters. Of these eight
indices, the first six are based on the definition of a flood as a flow event exceeding a 1.67-year
recurrence interval, calculated by assuming the peak flow values are a sample from a lognormal
distribution of flows; the last two are based on the same principle, but use a 5-year recurrence
interval for low flow calculations (Poff 1996). The 1.67-year flood is used as an assumption of
bankfull flow (Dunne and Leopold 1978).
26
Model Use and Application
Temporal Scale
Both IHA and HAT include indices calculated across multiple time scales ranging from daily to
annual and intra-annual statistics. This range of time scales is important to adequately capture
the timing, duration, frequency, and rate of change aspects of a flow regime. Both IHA and
HAT include tools to define analyses based on either one or two specific time periods that are
not necessarily the period of record. This is a powerful feature of the tools in that they can be
used to isolate and characterize the flow regime for particular time periods of interest, including:
before and after water development project implementation, before and after major water
withdrawals or transfers, and specific droughts or wet periods, among others. IHA allows for
batch processing of multiple data sets at once, a feature that HAT does not currently support.
External Verification
For both programs, flow statistics are calculated such that independent external verification is
possible. Henriksen (2006) includes a discussion of HAT verification results performed by
USGS and Colorado State University using commercially-available software including:
MATLAB, SAS, and Microsoft Excel. Calculations of the 33 IHA indices (not including the 34
Environmental Flow Components (EFCs)) were replicated using Microsoft Excel in unpublished
work by Joe Trungale (2006) of Trungale Engineering and Science for a sample Texas
streamflow gage.
Statistical Definitions
Both IHA and HAT allow the user to choose between parametric (characterized by a normal
distribution around the mean with a standard deviation) and non-parametric statistics (no a priori
frequency distribution, characterized by the median and percentiles). “For most [hydrologic]
situations non-parametric statistics are a better choice, because of the skewed (non-normal)
nature of many hydrologic datasets…but for certain situations, such as flood frequency or
27
average monthly flow volumes, parametric statistics may be preferable.” (The Nature
Conservancy 2006a).
Environmental Flow Component Definitions
The four flow component conceptual model recommended in the NRC Report and subsequently
adopted into the TIFP Technical Overview bears strong resemblance to the IHA Environmental
Flow Component (EFC) model as the lead IHA developer, Brian Richter of The Nature
Conservancy, served on the NRC committee. The stated goal of delineating the hydrograph in
this manner is linking specific flow events with their ecological purposes so that the benefits of
each flow component are preserved or recreated (The Nature Conservancy 2006a). One
difficulty of this flow component model is that the specific linkages, particularly biological, must
be understood in a particular sub-basin so that the appropriate flow components may be
incorporated into the environmental flow prescription. In the absence of detailed understanding
and exhaustive data on various riverine systems, a first-pass approach may be used in
conjunction with a program of adaptive management, where indicators of system health are
monitored for a period of time following instream flow implementation and the flow criteria
reevaluated as needed. Nonetheless, it behooves the tri-agencies and stakeholders to make the
‘first-pass’ at flow component delineation as ecologically-significant as possible by optimizing
the flow thresholds based on realistic, physically-based principles.
The IHA EFC parameters (Appendix B) are broken down into five categories: extreme low flow,
low flow, high flow pulses, small floods, and large floods. The software incorporates default
parameters for the delineation of the EFCs as well as an interface for some user modification of
the defaults (Figure 6).
28
Figure 6. IHA EFC definitions screen, displaying default thresholds.
Under the IHA EFC model, all daily flows fall within one of the five categories and a complex
algorithm parses the hydrograph accordingly based on the delineation thresholds being employed
(Figure 7). The program logic is to separate flow into base flows and flow pulse periods (i.e.
partition in time) using a base flow separation method, then take the pulses and classify them by
the rate of change of flow (from percent difference from previous day) and take the base flows
and classify them by the magnitude (from recurrence intervals). The thresholds include: flow
magnitude (e.g.: 10th percentile and median), recurrence intervals (e.g.: 2-year event), and rate of
change (e.g.: 25 percent flow increase from previous day). The default threshold values were
recently reformulated based on the scientific judgment of the software developers from their
extensive collective experience to be “ecologically-responsive” to model “what a fish feels,” but
not all the consequences of this reformulation on flow characterization have been worked out as
of yet (Richter personal communication 2006).
29
Daily
Flow
Data
> median?
high flow
> 75th
percentile?
YES
NO
low flow
25% greater
than previous
day?
high flow pulse
next day
decreases by
>10%?
high flow pulse
>median?
high flow pulse
low flow
low flow
YES
YES
YES
YES
NO
NO
NO
low flow
high flow pulse
NO
<10th
percentile?
extreme low
flow
low flow
recurrence
interval (from daily
high flow data) > 2
years?
small or large
flood
R.I. > 10
years?
large flood
small flood
high flow pulse
YES
YES
NO
NO
Note: all thresholds are default and adjustable in magnitude but not format (i.e.
large floods can be defined at > 5 years R.I., but not as 7 times the median
flow).
NO
YES
next day
Figure 7. IHA Environmental Flow Component algorithm flow chart.
30
The software allows for the user to change the magnitude of each threshold but not the type of
threshold employed. For example, small floods can be adjusted from the 2-year recurrence
interval to a 1.5-year or 3-year event but not pegged to the 80th percentile of all daily flows or
three times the median flow. In addition, the algorithm methodology with respect to the rate of
flow onset or recession often defines daily streamflows with smaller absolute magnitude than
other recent days being considered as high flow pulses, whereas the ‘higher flows’ are
considered low flows (Figure 8). The default rate of change parameters can be tweaked to
minimize this occurrence, but this anomaly likely results in a flow characterization too complex
for effective regulation and the communication of said regulation. A quantitative comparison of
cases where the IHA EFCs overlap is presented in Figure 9. Likewise, this increased level of
complexity of flow component delineation does not mesh well with the goal for flow
prescriptions presented in the TIFP Technical Overview (2006).
Figure 8. Daily Environmental Flow Components for USGS Gage No. 08171000, Blanco Rv at Wimberley,
TX, with a close-up view of a case where low flows (dark green) are of greater magnitude than high flow
pulse events (blue).
31
Figure 9. Quantitative comparison of IHA EFCs for USGS Gage No. 08171000, Blanco Rv at Wimberley, TX
for the period of record.
The hydrologic stream classification of NATHAT includes six stream types that arose from the
hydrogeographic classification scheme of Poff (1996): (1) harsh intermittent, (2) intermittent
flashy or runoff, (3) snowmelt, (4) snow and rain, (5) superstable or stable groundwater, and (6)
perennial flashy or runoff. The stream types are based on Pearson correlation coefficients for 13
variables on 420 streams nationwide. HAT includes flow indices based on: percentiles,
multipliers of the median flow, and recurrence intervals, among others (Appendix C). None of
the statistics are internally adjustable with respect to flow component thresholds, but HAT also
does not define Environmental Flow Components (i.e., “what the fish feels”) as IHA does, so no
specific indices or thresholds are tied to any particular flow component. In other words, it is up
to the user to explore the results produced by HAT (often using statistical tools) to select
appropriate indices for flow component delineation or otherwise for hydrograph characterization.
One example of this selection process is the multivariate statistics Principal Component Analysis
32
(PCA) that was employed by USGS and NJDEP to determine a subset of non-redundant indices
for the NJSCT and NJHAT. A number of the 171 HAT indices describe streamflow in a manner
that closely parallels the TIFP flow component model (Table 5).
Table 5. HAT indices useful for the delineation of the TIFP flow components (refer to Appendix C for
explanation of indices).
The intended outcome for SB2 is an instream flow prescription that integrates the results of the
multi-disciplinary studies, broken down by month, by flow component, and by hydrologic
conditions (i.e. wet, dry, normal years) (Figure 10). Each ‘building block’ depicted in Figure 10a
is intended to provide the flow conditions necessary to provide the associated ecosystem function
(e.g. channel maintenance, seed dispersal, fish spawning, etc). Although a number of hydrologic
33
Figure 10. (a) Example instream flow prescription to be developed under SB2 depicting integration of the
flow components, and (b) tabular representation of (a) (TIFP 2006).
34
indices in NATHAT can be used to delineation the TIFP flow components (Table 5), HAT is not
explicitly designed in accordance with the flow component model of IHA and subsequently
incorporated into TIFP based on the NRC recommendations (NRC 2005). As such,
modifications would need to be made to HAT to increase its suitability for use in the TIFP that
would focus on assessing the appropriate hydrologic indices to delineation each of the four flow
components. The indices presented in Table 5 could be evaluated for application for Texas
rivers, modified or amended as needed, and the NATHAT software could be revised to present
analyses and results grouped by flow component.
In IHA, all of the flow magnitude indices are absolute and expressed as the volumetric flow rate
in units of length cubed per time, often cubic feet per second (cfs) or cubic meters per second
(cms). IHA allows for the normalization of some flow statistics by the contributory drainage
area. In HAT, some of the indices are expressed in similar fashion, whereas others are
normalized by another flow, often the median daily streamflow, resulting in a dimensionless
ratio. These ratios give an indication of the range of the flow regime and the occurrence of
various flows within that regime, yet are decoupled from the watershed size and characteristics.
This allows for an isolated characterization of the hydrograph shape and pattern and also for an
equivalent comparison of these dimensionless ratios across disparate drainage basins or regions
of the State, a feature highly valuable for TIFP. Although IHA allows for normalization by
drainage area, varying impacts of factors such as land use, land cover, soils, surficial geology,
and local climatology cause varying hydrologic responses that are not necessarily explained by
the sheer area of flow contribution. By normalizing by another flow at the same gage, the
indices in HAT account for (and thus remove) many of the impacts of the physical and
climatological factors associated with flow generation.
Model Output
IHA
IHA tabular output is displayed on-screen in Microsoft Spreadsheet format and can easily be
exported to spreadsheet software such as Microsoft Excel. The output includes tabbed pages for:
35
(1) annual summary statistics, (2) (non-) parametric IHA summary scorecard, (3) linear
regression, for identifying trends in the data, (4) IHA percentile data, (5) EFC daily flow
characterization, and (6) messages and warnings regarding the results generated. The output is
also available as text (.txt) files.
IHA graphical output includes numerous plotting options for various indices with various
presentation styles. The plots are displayed individually on-screen and can be exported or
captured to image software (such as Microsoft Paint) or saved in various graphics file types (e.g.
.jpg). The current version of IHA appears to have a software bug for certain operating systems
that affects the quality of exported images (Figure 11a). Screen-capture tools can be used as a
workaround to this problem, but the exporting bug should be addressed in future versions of IHA
(Figure 11b).
HAT
HAT tabular output is saved in comma-separated value (.csv) files that can also be read into
standard spreadsheet software. The single output table consists of the median value of each of
the 171 statistical indices across all years of the period analyzed along with the lower and upper
limits of that particular value, where applicable. The defaults for the limits are the 25th and 75th
percentiles of the results, but the percentiles can be user-defined. Some of the indices do not
have limits associated with them. For example, TL1, the median Julian day of the annual
minimum flow across all years, is a single, finite numeric value, has no distribution of results,
and thus has no related percentiles.
Additional tables and graphs of the ten default ‘non-redundant’ indices for the selected stream
type are available within the HAT program; the ten default indices for the six stream types are
based on the hydrogeographic classification scheme of Poff (1996). In lieu of the ten default
indices, the user has two alternatives. First, the ten default indices can be user-selected from
menus of 18 to 30 indices (number based on stream type), with one index representing each of
the following categories: (1) magnitude of flow events under average flow conditions, (2)
magnitude of flow events under low flow conditions, (3) magnitude of flow events under high
36
Figure 11. (a) Example of graphics capture bug present in version 3.0 of NATHAT; (b) same plot as viewed
on-screen in the NATHAT software and captured for comparison here using screen-capture tools.
flow conditions, (4) frequency of flow events under low flow conditions, (5) frequency of flow
events under high flow conditions, (6) duration of flow events under low flow conditions, (7)
duration of flow events under high flow conditions, (8) timing of flow events under low flow
conditions, (9) timing of flow events under high flow conditions, and (10) rate of change in flow
(Table 6, Figure 12).
37
Table 6. Ten default hydrologic indices for each stream type in NATHAT.
Stream Type Magnitude Frequency Duration Timing Rate of Change
Intermittent (harsh) MA34, ML13, MH23 FL2, FH3 DL13, DH10 TL2, TH1 RA4
Intermittent (flashy or runoff) MA37, ML16, MH23 FL2, FH3 DL18, DH13 TL1, TH3 RA9
Perennial (snowmelt) MA29, ML13, MH1 FL3, FH8 DL5, DH19 TL1, TH1 RA1
Perennial (snow and rain) MA3, ML13, MH17 FL3, FH3 DL6, DH12 TL1, TH1 RA9
Perennial (superstable/
stable groundwater)
MA3, ML18,
MH17 FL3, FH3
DL9,
DH11 TL2, TH1 RA9
Perennial (flashy or runoff) MA26, ML17, MH23 FL3, FH4 DL10, DH13 TL1, TH3 RA9
38
Figure 12. Example NATHAT default hydrologic index selection menu depicting selection options for one of
the six stream types.
The selection of ten replacement indices may be saved in the NATHAT software as the program
default. Second, up to twelve indices of the entire list of 171 can be selected for graphical
presentation, but this selection can not be saved within the software. Output presented within the
program can be exported similarly to IHA.
39
III. Model Application in the Texas Instream Flow Program
Hydrologic Study Needs
Hydrologic and hydraulic analyses are an integral part of all of the steps of the TIFP sub-basin
studies, and a robust hydrologic statistics tool is essential for many of the study steps. For the
purposes of the flow regime, there is a need to calculate subsistence flow, base flow, high flow
pulse, and overbank flow (flood) frequency statistics and to define wet, dry, and normal years.
During the Reconnaissance and Information Evaluation step recommended in TIFP (2006), there
is a need to calculate historic and current flow statistics and to identify existing features (e.g.,
tributaries) and existing and proposed alterations (diversions, impoundments, land uses, etc)
affecting hydrologic character. During the Data Integration to Generate Flow Recommendations
step, there is a need to: (1) calculate the occurrence of various flow rates during historical and
current conditions, (2) determine annual variability of hydrologic characteristics, including
description of wet, normal and dry years, (3) develop hydrologic time series to evaluate habitat
suitability of a proposed flow regime, (4) calculate variability a of proposed flow regime and
compare with historic/current conditions, and (5) evaluate how proposed flow regimes would
impact current operating conditions (TIFP 2006).
Temporal Resolution
A focus of current environmental flow science is on the importance of flow events that operate
on a relatively small (i.e., daily) time step. This represents a paradigm shift from historical
methods applied in Texas such as the Lyons Method (Bounds and Lyons 1979) that uses daily
gaged flow data as input but the resulting flow prescriptions are presented on a monthly basis.
Likewise, the Consensus Criteria for Environmental Flow Needs (CCEFN) (TCEQ 2004) are
developed using monthly flow data from the Water Rights Analysis Package (WRAP) Water
Availability Models (WAM) (Wurbs 2003, TCEQ 2006).
Both IHA and HAT require the input of daily flow data, and both models generate statistical
indices that represent a broad spectrum of time scales: daily, weekly, monthly, seasonal, annual,
40
and inter-annual. This point is critical in that neither model can ingest the monthly naturalized
flows derived from the Water Availability Models, nor can they ingest any other flow data that is
measured or modeled on a monthly time step. If the Texas Instream Flow Program is to use
monthly flow data, neither program is applicable. Thus, daily flow data must be available at the
site of interest or must be reconstructed from monthly flow data. It is possible to synthesize a
daily flow record via apportionment of modeled monthly flows, but this transformation
inherently introduces uncertainty into the input data.
If the IHA and HAT models were modified to require the input of monthly flow data instead of
daily, some of the statistical indices of the two tools would be impossible to calculate and others
would have less relevance due to the decreased temporal resolution. In general, indices of flow
magnitude would be largely preserved but indices of flow frequency, duration, timing, and rate
of change would be largely destroyed. This is because hydrologic events that are driven by
runoff occur on a time scale much smaller than one month, so any signal generated by individual
flow events within a month would be muted when viewed over the entire month.
Additional Potential Enhancements
Calculation of Medians
Measures of central tendency are critical to hydrologic analysis and various calculatio n methods
were explored in detail in this study; a discussion of these methods can be found in the following
chapter. From this work, however, arose a concern over the methodology IHA and HAT use to
calculate the monthly median flow.
As depicted in Figure 13a, the two software tools first calculate the median for all daily flows
within one month of one year (here, January 1940), repeat this process for every January, then
calculate the median of all the January values to get the long-term median flow for the month of
January across the entire time period of analysis. Given the typical large positive skew in the
distribution of streamflow data, particularly for streams in Texas, calculating the median of a
median likely results in a downward bias. More accurate methods for calculating the long-term
41
daily median and the long-term monthly median are shown in Figure 13b and Figure 13c; these
methods are considered ‘more accurate’ because they eliminate the likely downward bias that
results from the multi-step calculation process of Figure 13a. The concern discussed above is a
moot point if parametric statistics are being calculated, as the mean of a mean preserves the true
measure of central tendency. The method shown in Figure 13b was employed in this study to
calculate median daily flows, as discussed in the Measures of Central Tendency section.
Figure 13. Various procedures for the calculation of median streamflows: (a) method employed in IHA, (b)
alternative method for daily medians, and (c) alternative method for monthly medians.
42
Environmental Flow Component Thresholds
As previously discussed, the default IHA EFC flow thresholds that define the five flow
components (extreme flow, low flow, high flow pulses, small floods, and large floods) arose
from the judgment of the ecologists and hydrologists that is based on and field-truthed over time
at multiple rivers and regions across the United States and the world. Oftentimes, the EFCs and
environmental flow prescriptions are tailored for specific applications such as the protection or
restoration of a particular species, habitat, or life cycle cue. A large body of research (see
discussion in Poff et al (1997)) has established the ecological significance of certain flow
components for certain species for certain purposes in certain regions of the world. However,
TIFP faces a different and broader challenge. Rather than a specific, narrow goal, the language
of SB2 calls for the “support of a sound ecological environment,” a much more holistic
viewpoint that has been interpreted to account for all endemic species over all of their lifecycles
for all habitat types across all regions of the State.
Given the diversity of species, habitats, and other conditions within the large State of Texas, the
difficulty of establishing suitable flow component thresholds becomes obvious. Thus, the default
EFC delineations in IHA may not be representative nor suitable for widespread application in
Texas, and additional field studies and sampling would be required to truly ascertain the
ecological significance of various components of the flow regime.
As such, there is a need to establish the ecological significance of the EFC thresholds for Texas,
or redefine the thresholds accordingly. It would be ideal to customize the thresholds to suitable
Texas flow component delineations, such as recurrence intervals, percentiles, or fixed flow rates
at specific points, like the streamflow required to reach the incipient point of overbank flooding
at a particular cross section of interest. The development of such basin- or reach-specific
thresholds is likely to be highly resource intensive, and this work is likely to be developed on a
study-specific case as deemed necessary and beneficial.
43
HAT Display Indices
The choice of non-redundant indices in NATHAT is based on the six stream classifications and
associated analyses from Poff (1996). Regardless of selected stream classification, the model
calculates all 171 indices and returns the results in spreadsheet format. A user-defined selection
of up to twelve alternative indices is possible for graphical display, but there is no current
capability in the program to save the suite of manually-selected indices. The addition of this
functionality would likely be simple to accomplish and would streamline the creation of
graphical output.
44
IV. Flow Regime Evaluation
Data Sources and Metadata
One goal of this study was to evaluate hydrologic flow regimes using hydrologic assessment
tools such as IHA and HAT. A cohort of 24 USGS streamflow gages were selected for study by
the TWDB in conjunction with the Center for Research in Water Resources (CRWR) with four
gages located within each of the six TIFP priority study subbasins: Lower Sabine, Middle
Trinity, Middle and Lower Brazos, Lower Guadalupe, and Lower San Antonio. A list of the
study gages is presented in Table 7. The flow data obtained from the USGS NWIS and analyzed
in this study is presented in electronic format in Appendix D. A general locus map of the study
gages may be found in Figure 14, and specific locus maps of the study gages by subbasin may be
found in Figure 15 through Figure 18. These 24 gages were selected because they: (1) are
believed to be representative of hydrologic conditions in the main stems and tributaries of the
priority subbasins; (2) each have 20 or more years of continuous daily mean streamflow data and
instantaneous peak streamflow data; and (3) encompass contributory drainage areas and time
periods where streamflow is both minimally and highly altered by water development projects.
The 24 gages have contributory drainage areas ranging from 42.8 to 35,773 square miles with a
mean of 8,618 square miles and continuous periods of record ranging from 20 to 87 years with a
mean of 64 years, typically through Water Year 2004.
45
Table 7. Selected USGS gages for flow regime evaluation.
Site
Number Site Name
Period of
Record
(Years)
Contributory
Drainage
Area (mi2)
08020000 Sabine Rv nr Gladewater, TX 72 2,791
08028500 Sabine Rv nr Bon Wier, TX 81 8,229
08029500 Big Cow Ck nr Newton, TX 52 128
08030500 Sabine Rv nr Ruliff, TX 80 9,329
08062500 Trinity Rv nr Rosser, TX 65 8,147
08062700 Trinity Rv at Trinidad, TX 40 8,538
08065000 Trinity Rv nr Oakwood, TX 81 12,833
08066500 Trinity Rv at Romayor, TX 80 17,186
08093100 Brazos Rv nr Aquilla, TX 66 17,678
08098290 Brazos Rv nr Highbank, TX 39 20,870
08100500 Leon Rv at Gatesville, TX 54 2,342
08106500 Little Rv nr Cameron, TX 87 7,065
08110500 Navasota Rv nr Easterly, TX 80 968
08111500 Brazos Rv nr Hempstead, TX 66 34,314
08115000 Big Ck nr Needville, TX 52 42.8
08116650 Brazos Rv nr Rosharon, TX 20 35,773
08168500 Guadalupe Rv abv Comal Rv at New Braunfels, TX 76 1,518
08171000 Blanco Rv at Wimberley, TX 76 355
08175800 Guadalupe Rv at Cuero, TX 39 4,934
08176500 Guadalupe Rv at Victoria, TX 69 5,198
08181800 San Antonio Rv nr Elmendorf, TX 42 1,743
08183500 San Antonio Rv nr Falls City, TX 79 2,113
08186000 Cibolo Ck nr Falls City, TX 74 827
08188500 San Antonio Rv at Goliad, TX 65 3,921
46
Figure 14. Locus map of 24 study gages in the State of Texas.
!(
!( !(
!(
!(
!(
!(
!(
08188500
08186000
08183500
08181800
08175800
08176500
08171000
08168500
0 5025
Miles
±
Figure 15. Map of study gages in the Lower Guadalupe and Lower San Antonio River subbasins.
Sabine
Trinity Brazos
San Antonio
Guadalupe
47
!(
!(
!(
!(!(
!(
!(
!(08111500
08116650
08115000
08098290 08110500
08106500
08100500
08093100
0 10050
Miles
±
Figure 16. Map of study gages in the Middle and Lower Brazos River subbasins.
!(
!(
!(
!(08066500
08065000
08062700
08062500
0 10050
Miles
±
Figure 17. Map of study gages in the Middle Trinity River subbasin.
48
!(
!(!(
!(08030500
08029500
08028500
08020000
0 10050
Miles
±
Figure 18. Map of study gages in the Lower Sabine River subbasin.
In addition to streamflow, the flow regime assessment performed included an analysis of
precipitation. Data for this analysis were obtained from digital climate maps created by using the
Parameter-elevation Regressions on Independent Slopes Model (PRISM) by Oregon State
University’s PRISM Group for the period 1961 through 1990 (PRISM 2006). Electronic copies
of the data and maps of the data in raster format are included in Appendix D.
In an analysis of systematic methodologies to delineate flow components this study also made
use of published USGS stream gage measurement data (Appendix D). In particular, a time series
of the field-measured (as opposed to gaged) streamflow and the measured active width of flow at
the time of streamflow measurement was used to identify the river channel and overbank
morphology.
Methodology
Both IHA and NATHAT were used to analyze the flow regime at each of the 24 gages; the
results of these analyses in included in Appendix D. The initial period-of-record results were
49
examined for spatial and temporal trends and patterns and a number of intriguing results were
revisited using IHA, HAT, an external tool, or more commonly, multiple tools for comparison
purposes. External analyses were performed using Microsoft Excel spreadsheets, customized
Microsoft Visual Basic for Applications programs, Environmental Systems Research Institute
(ESRI) ArcGIS ArcMap version 9.1, and Minitab statistical software package version 14.1.
These tools were employed in instances where they provided greater flexibility for analyses or
where their specific capabilities were more valuable in addressing specific issues than IHA, HAT
or both. Individual case methodologies are discussed in the sections below.
Problem Definition
IHA includes 67 statistical routines and HAT includes 171. Both tools include capabilities to
view the hydrograph for each stream gage analyzed, a prudent starting point for any flow regime
evaluation. However, in the case of an engineer or scientist tasked with evaluating the flow
regime of a river and armed with one or both of these tools and little other information and data,
the user can quickly become overwhelmed by the sheer number and diverse meaning and
purpose of the various hydrologic indices. Furthermore, the user’s ability to identify spatial and
temporal trends and patterns is only as good as their understanding of the particular indices of
interest or the particular goal of their assessment.
For example, which flow characteristics are of primary import: magnitude, duration, frequency,
timing, and/or rate of change? Is the user concerned with seasonal, annual, or inter-annual
variation? Is the purpose of the analysis to assess the impacts of change or define acceptable
limits of change? Are generalized long-term patterns or the evaluation of individual flood events
more important? As the problem at hand becomes better defined, the capabilities of IHA and
HAT best suited to solving of that problem become better defined as well. In the absence of a
narrowly-defined problem for this study, external tools were sometimes relied upon to identify
macro-scale trends and patterns across space and time.
50
Impact and Extent of Flow Regulation
A major capability of the tools, particularly IHA, is the ability to assess hydrologic alteration,
and this is typically how IHA has been employed to-date (Nature Conservancy 2006); HAT is
too new to have an understanding of its typical usage Thus, an analysis of hydrologic alteration
became a logical starting point in this study and served both as a means of evaluating and
comparing the two tools and as a means of understanding how water development projects such
as dams, major diversions, and return flows have impacted flow regimes at the study gages.
Of the 24 gages, 19 are now considered by the USGS to be highly impacted due to upstream
flow regulation, meaning that runoff from greater than ten percent of their contributory drainage
area is affected by regulation (Figure 19) (USGS 2004). Eight major dams impound reservoirs
that impact a number of the rivers studied (Table 8).
Besides all providing water supply capabilities, the reservoirs variously support hydropower,
irrigation, recreation, and flood control. As such, individual operating rules of each dam serve as
the primary measure of when, how, and to what extent the downstream flow regime is altered.
For example, the construction of a water supply or flood control dam is commonly associated
with a decreased magnitude of peak flows. Figure 20 illustrates this case for the Lower
Guadalupe River above the Comal River at New Braunfels, Texas following the development of
Canyon Lake. Also, irrigation reservoirs or wastewater treatment plant return flows from a
major city may cause artificially elevated summertime low flows, as is evident in the Middle
Trinity River downstream of the Dallas-Fort Worth Metroplex (Figure 21). The temporal
analysis capabilities of both IHA and HAT are useful for illustrating the extent and quantifying
the magnitude of such flow regulation. In addition, both programs can analyze historic gaged
streamflow data to assess actual historic change, and both can accept modeled streamflow data
(in a compatible file format) to assess proposed future change.
51
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
TOLEDO BEND RESERVOIR
LAKE LIVINGSTON
LAKE WHITNEYRICHLAND-CHAMBERS RESERVOIR
CEDAR CREEK RESERVOIR (HENDERSON)
LAKE LIMESTONE
CANYON LAKE
VICTOR BRAUNIG LAKE
LAKE GLADEWATER
Big Ck nr Needville, TX
Big Cow Ck nr Newton, TX
Blanco Rv at Wimberley, TX
Cibolo Ck nr Falls City, TX
San Antonio Rv at Goliad, TX
Sabine Rv nr Ruliff, TX
Trinity Rv nr Rosser, TX
Little Rv at Cameron, TX
Brazos Rv nr Aquilla, TX
Guadalupe Rv at Cuero, TX
Trinity Rv at Romayor, TX
Trinity Rv nr Oakwood, TX
Sabine Rv nr Bon Wier, TX
Brazos Rv nr Rosharon, TX
Leon Rv at Gatesville, TX
Brazos Rv nr Highbank, TX
Trinity Rv at Trinidad, TX
Brazos Rv nr Hempstead, TX
Sabine Rv nr Gladewater, TX
Navasota Rv nr Easterly, TX
Guadalupe Rv at Victoria, TX
Guadalupe Rv abv Comal Rv at
San Antonio Rv nr Falls City
0 50 10025
Miles
±
Legend
!( Regulated Gages
!( Unregulated Gages
Figure 19. Flow regulation at study gages, where regulated gages are defined as those where runoff from
greater than ten percent of their contributory drainage area is affected by regulation.
52
Figure 20. Time trend of maximum 1 -day flow in Lower Guadalupe River downstream of Canyon Dam,
which began impounding water in 1964.
Figure 21. Time trend of minimum 1-day flow in the Middle Trinity River downstream of Dallas-Fort
Worth.
53
Table 8. Major dams and reservoirs regulating flow at the study gages.
Reservoir
Name Dam Name Owner Purpose River D/S Gage
Date
Impound
Normal
Height
(ft)
Normal
Storage
(ac-ft)
Cons
Height
(ft)
Cons
Storage
(ac-ft)
Flood
Height
(ft)
Flood
Storage
(ac-ft)
Max
Height
(ft)
Length
(ft)
DA
(mi2)
Lake
Gladewater
Gladewater
Dam
City of
Glade-
water
WS Glade Creek
08020000,
08028500,
08030500
8/1952 36 6,950 36 6,950 N/A N/A 48 1,203 35
Cedar
Creek
Reservoir
(Trinity)
Joe B.
Hogsett
Dam
Tarrant
Regional
Water
District
WS, R Trinity River 08065000 7/1965 53 N/A 73 679,200 N/A N/A 91 17,539 1,007
Lake
Whitney
Whitney
Dam
Army
Corps of
Eng.
Fort
Worth
District
FC,
WS, HP
Brazos
River
08093100,
08092890? 12/1951 49 4,270 108 679,200 146 1,999,500 159 17,695 17,656
Toledo
Bend
Reservoir
Toledo
Bend Dam
Sabine
River
Auth. of
TX & LA
WS,
HP, R
Sabine
River
08028500,
08030500 10/1966 72 1,161,800 99 4,477,000 102 5,102,000 112 11,200 7,178
Lake
Livingston
Livingston
Dam
Trinity
River
Authority
WS Trinity River 08066500 10/1968 54 N/A 86 1,750,000 89 N/A 100 14,400 16,616
Canyon
Lake
Canyon
Dam
Army
Corps of
Eng.
Fort
Worth
District
FC, WS Guad. River 08168500 6/1964 125 N/A 159 386,200 193 737,444 224 6,830 1,432
Richland-
Chambers
Reservoir
Richland-
Chambers
Dam
Tarrant
Regional
Water
District
WS, R Richland Creek 08065000 11/1987 N/A N/A 83 1,135,000 86 N/A 96 31,000 1,957
Lake
Limestone
Sterling C.
Roberston
Dam
Brazos
River
Authority
WS, I, R Navosta River 08110500 10/1978 54.6 325,670 48 225,400 61 458,603 65 9,100 675
Purpose: Water Supply, Recreation, Flood Control, Hydropower, Irrigation
54
Measures of Central Tendency
For each of the 24 study gages, the long-term daily median flow for each calendar day was
calculated across the period of record (average = 68 years) of continuous daily flow data. The
calculation was performed using a Visual Basic for Applications (VBA) macro in Microsoft
Excel (Appendix D). External to IHA and HAT, this analysis was performed as part of the flow
regime assessment as a means to evaluate the long-term seasonal patterns of streamflow within
each subbasin by filtering out the effects of individual flow events and thus the variability and
flashiness associated with the daily time series hydrograph. As such, the following figures are
useful for visualizing the underlying patterns of daily streamflow but are not suitable for the
development of environmental flow prescriptions, as they do not include any of the shorter
duration variability that has been recognized to be so critical to ecosystem preservation. Figure
22 through Figure 27 are plotted by subbasin with streamflow depicted on equivalent logarithmic
scales on the ordinate and calendar day depicted on the abscissa.
Sabine Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
Sabine Rv nr Gladewater, TX
Sabine Rv nr Bon Wier, TX
Big Cow Ck nr Newton, TX
Sabine Rv nr Ruliff, TX
Figure 22. Median daily streamflow in the Lower Sabine River subbasin.
55
Trinity Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
Trinity Rv nr Rosser, TX
Trinity Rv at Trinidad, TX
Trinity Rv nr Oakwood, TX
Trinity Rv at Romayor, TX
Figure 23. Median daily streamflow in the Middle Trinity River subbasin.
Middle Brazos Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
Brazos Rv nr Aquilla, TX
Brazos Rv nr Highbank, TX
Leon Rv at Gatesville, TX
Little Rv nr Cameron, TX
Figure 24. Median daily streamflow in the Middle Brazos River subbasin.
56
Lower Brazos Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
Navasota Rv nr Easterly, TX
Brazos Rv nr Hempstead, TX
Big Ck nr Needville, TX
Brazos Rv nr Rosharon, TX
Figure 25. Median daily streamflow in the Lower Brazos River subbasin. Note: Due to the logarithmic scale,
flows shown along the abscissa are 1 cfs or less, and not necessarily 0 cfs.
Guadalupe Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
Guadalupe Rv abv Comal Rv at New Braunfels, TX
Blanco Rv at Wimberley, TX
Guadalupe Rv at Cuero, TX
Guadalupe Rv at Victoria, TX
Figure 26. Median daily streamflow in the Lower Guadalupe River subbasin.
57
San Antonio Basin
1
10
100
1,000
10,000
100,000
1-Jan 31-Jan 2-Mar 1-Apr 2-May 1-Jun 2-Jul 1-Aug 1-Sep 1-Oct 1-Nov 1-Dec
Median Daily Flow (cfs)
San Antonio Rv nr Elmendorf, TX
San Antonio Rv nr Falls City, TX
Cibolo Ck nr Falls City, TX
San Antonio Rv at Goliad, TX
Figure 27. Median daily streamflow in the Lower San Antonio River subbasin.
As can be seen from the graphs, many of the hydrographs within each subbasin follow a similar
pattern to other streams within that subbasin, irrespective of the contributory drainage area to
each gage. The streamflows generally follow a sinusoidal pattern, with a spring peak (May-
June) and a fall trough (August-September) when viewed on a semi-log plot, with the signal
being stronger to the east (e.g. Lower Sabine and Middle Trinity) and muted to the west (e.g.
Lower Guadalupe and Lower San Antonio) along the Te xas Gulf Coast, thus demonstrating that
there are pronounced regional patterns in the seasonality of the streamflow hydrograph within
the State of Texas. Because of the semi-log plotting style, the seasonal pattern is not a true
sinusoid in time and streamflow.
The hydrographs depicted above were developed from flow data that typically encompasses time
periods of both relatively unaltered and highly altered hydrologic conditions as the periods of
record for many of the study gages bracket the construction of numerous major dams, the rapid
growth of multiple urban areas, and the creation of many other communities. Consideration of
the long-term seasonal patterns would prove beneficial to the TIFP in the development of
58
instream flow requirements as the patterns reflect the underlying seasonality of flow in many of
the rivers of interest. In addition, the signals expressed here also provide clues regarding the
likely streamflow pattern based on a stream’s geographic location within the State.
Given the importance of measures of central tendency to an evaluation of flow regimes, this
project incorporated a study comparing means and medians and also comparing monthly median
flows to daily median flows (Table 9). Based on an analysis of the period of record data for the
24 study gages, the long-term daily median streamflow can be reasonably approximated by the
long-term monthly median streamflow, with an average error of 1.3 percent and error values
ranging from 0 to 7 percent. There is no difference in the calculation of the long-term mean
daily flow from either daily or monthly mean flow data. The mean daily flow for each of the 24
gages was higher than the median daily flow, an indication of the large positive skew of daily
streamflow data, with a typical difference of 56 percent and difference ranging from 25 to 95
percent. This study shows that there is minimal loss in accuracy when measures of central
tendency are calculated based on a monthly time step as opposed to a daily time step.
Precipitation and Streamflow
In conjunction with the analysis of long-term seasonal streamflow patterns discussed above, a
parallel analysis of long-term precipitation patterns was conducted. This analysis was performed
using Spatial Analyst tools within ESRI ArcGIS version 9.1 and PRISM data for the period 1961
through 1990; note that the time period of this analysis does not match exactly with the periods
of record for each of the 24 study gages, but is assumed to be a reasonable approximation of
long-term conditions based on the considerable length of the time period being analyzed. In this
case, the long-term streamflow was aggregated into monthly medians instead of daily to more
closely match the monthly precipitation estimates, and the streamflow was also normalized by
the contributory drainage area to arrive at a unit monthly median flow represented in inches of
flow depth across the drainage area, thus directly comparable to inches of precipitation depth.
The results of this analysis are presented in Figure 28 through Figure 31 for four of the priority
subbasins.
59
Table 9. Comparison of measures of central tendency by proportionally-weighted months.
gage
number
period of
record
(yr)
mean daily
median
(cfs)
mean
weighted
monthly
median
(cfs)
percent
difference
abs value
percent
difference
mean daily
mean
mean
weighted
monthly
mean
percent
difference
percent
difference
(mean daily
median to
mean daily
mean)
8020000 72 848 836 1% 1% 1918 1918 0.0% 56%
8028500 81 4919 4798 2% 2% 6996 6996 0.0% 30%
8029500 52 71 70 1% 1% 132 132 0.0% 46%
8030500 80 6276 6255 0% 0% 8398 8398 0.0% 25%
8062500 65 1336 1313 2% 2% 3206 3206 0.0% 58%
8062700 40 2067 2065 0% 0% 4452 4452 0.0% 54%
8065000 81 2258 2236 1% 1% 5259 5259 0.0% 57%
8066500 80 3984 3959 1% 1% 7905 7905 0.0% 50%
8093100 66 593 597 -1% 1% 1541 1541 0.0% 62%
8098290 39 1147 1127 2% 2% 2775 2775 0.0% 59%
8100500 54 53 52 2% 2% 304 304 0.0% 83%
8106500 87 619 612 1% 1% 1751 1751 0.0% 65%
8110500 80 37 34 7% 7% 425 425 0.0% 91%
8111500 66 3191 3212 -1% 1% 6914 6914 0.0% 54%
8115000 52 2.0 2 3% 3% 37 37 0.0% 95%
8116650 20 4828 4653 4% 4% 8813 8813 0.0% 45%
8168500 76 242 240 1% 1% 488 488 0.0% 50%
8171000 76 62 62 1% 1% 144 144 0.0% 57%
8175800 39 1137 1127 1% 1% 2148 2148 0.0% 47%
8176500 69 1053 1050 0% 0% 1963 1963 0.0% 46%
8181800 42 321 321 0% 0% 592 592 0.0% 46%
8183500 79 259 257 1% 1% 488 488 0.0% 47%
8186000 74 28 28 0% 0% 139 139 0.0% 80%
8188500 65 366 364 0% 0% 808 808 0.0% 55%
average 64 1.3% 0.0% 56%
60
Figure 28. Streamflow and precipitation patterns for the Lower Sabine River basin.
Figure 29. Streamflow and precipitation patterns for the Middle Trinity River basin.
61
Figure 30. Streamflow and precipitation patterns for the Middle Brazos River basin.
Figure 31. Streamflow and precipitation patterns for the Lowe r Guadalupe River basin.
62
Based on the above graphs for a 30-year period of record, the precipitation signal is bimodal,
irrespective of geography, with spring (i.e. May to June) and fall (i.e. September to October)
peaks. The precipitation gradient from East to Central Texas is expressed in the magnitude of
precipitation (more to the east) but doesn’t change the timing of spring and fall peaks. Also,
there appears to be a correlation between the spring peaks in streamflow and precipitation
whereby the increased precipitation accounts for increased runoff in the waterways. However,
there does not appear to be a similar correlation for the fall peak in precipitation; multiple factors
might explain this discrepancy, but for the purposes of this study is it important to note that the
spatial variability of the seasonality in long-term streamflow patterns cannot be explained by
spatial variability long-term patterns in precipitation alone.
Seven-Day Two-Year Low Flow (7Q2)
The seven-day average, two-year recurrence interval low flow discharge (7Q2) is an important
flow statistic in the State of Texas due to its statutory designation in the Texas Surface Water
Quality Standards (TCEQ 2000) as the defining low-flow condition variable to determine the
‘critical low flow,’ the flow below which some water quality standards no longer apply and the
flow at which the impacts of permitted discharges are analyzed. Here, the 7Q2 values were
calculated for each of the 24 study gages based on the entire period of record using IHA and
verified with Microsoft Excel (Table 10). Neither IHA nor HAT explicitly has the capability to
calculate the 7Q2. IHA calculates the seven-day minimum flows for each year (7Q1), and the
user can then externally calculate the median of these values to obtain the 7Q2. ML17, the ‘base
flow’ index in HAT, is the average of the seven-day minimum flow (7Q1) divided by the median
(or mean) annual flow for that year; the user can externally modify ML17 to obtain the 7Q2.
Note that the official TCEQ Surface Water Quality Standards values for the 7Q2 are calculated
from a sliding, limited time period that does not necessarily include the entire period of record of
gaged flow data.
63
Table 10. Seven-day average, two-year recurrence interval low flow discharge (7Q2).
Site
Number Site Name
Contributory
Drainage Area
(mi2)
7Q2
(cfs)
8020000 Sabine Rv nr Gladewater, TX 2,791 38.4
8028500 Sabine Rv nr Bon Wier, TX 8,229 457.9
8029500 Big Cow Ck nr Newton, TX 128 26.8
8030500 Sabine Rv nr Ruliff, TX 9,329 788.6
8062500 Trinity Rv nr Rosser, TX 8,147 370.1
8062700 Trinity Rv at Trinidad, TX 8,538 615.6
8065000 Trinity Rv nr Oakwood, TX 12,833 384.7
8066500 Trinity Rv at Romayor, TX 17,186 552.4
8093100 Brazos Rv nr Aquilla, TX 17,678 42.8
8098290 Brazos Rv nr Highbank, TX 20,870 179.0
8100500 Leon Rv at Gatesville, TX 2,342 2.0
8106500 Little Rv nr Cameron, TX 7,065 54.7
8110500 Navasota Rv nr Easterly, TX 968 1.2
8111500 Brazos Rv nr Hempstead, TX 34,314 620.6
8115000 Big Ck nr Needville, TX 42.8 0.4
8116650 Brazos Rv nr Rosharon, TX 35,773 573.1
8168500
Guadalupe Rv abv Comal Rv at
New Braunfels, TX 1,518 76.9
8171000 Blanco Rv at Wimberley, TX 355 18.9
8175800 Guadalupe Rv at Cuero, TX 4,934 489.4
8176500 Guadalupe Rv at Victoria, TX 5,198 530.6
8181800 San Antonio Rv nr Elmendorf, TX 1,743 141.9
8183500 San Antonio Rv nr Falls City, TX 2,113 110.0
8186000 Cibolo Ck nr Falls City, TX 827 9.9
8188500 San Antonio Rv at Goliad, TX 3,921 164.6
64
Range of Flow Regime
Similar to the analysis of long-term seasonal patterns of the daily median flow, additional
analyses were performed for various other magnitudes of flow, including the 5th, 25th, 75th, and
95th percentile flows across the period of record. Additional VBA macros were developed for
each of these analyses (Appendix D), and the output was then aggregated into modified box and
whisker plots whereby the four flow magnitudes discussed above were plotted alongside the
median flow on a monthly basis; results for selected study gages are presented in Figure 32 to
Figure 35. When considered over a long time scale, the seasonal variation in the flow regime of
various percentile levels of flow follows a consistent pattern, such that if the median flow goes
down then both high and low flow percentiles reduce proportionately to some extent, and vice
versa when the median flow increases. This means that the high and low flows within a
particular flow regime are pegged to the median flow to some degree.
Figure 32. Modified box and whisker plot for USGS Gage No. 0802000, Sabine Rv nr Gladewater, TX.
65
Figure 33. Modified box and whisker plot for USGS Gage No. 08065000, Trinity Rv nr Oakwood, TX.
Figure 34. Modified box and whisker plot for USGS Gage No. 08168500, Guadalupe Rv abv Comal Rv at
New Braunfels, TX.
66
Figure 35. Modified box and whisker plot for USGS Gage No. 08188500, San Antonio Rv at Goliad, TX.
If the pattern of flow regime variation as a function of the median flow was able to completely
explain the seasonal variation of the entire flow spectrum, then the filtering of the seasonal
pattern of the median flow signal from the entire flow regime would result in a series of parallel
lines representing the various high and low flows flanking the median. As can be seen in Figure
36, this explanation is partly true in this test case, particularly for the 25th and 75th percentile
flows, but secondary factors are also contributing to the statistical variation of the atypical flows.
67
Monthly Streamflow Percentiles Normalized by Monthly Median
08168500 - Guadalupe Rv abv Comal Rv at New Braunfels, TX
0.01
0.10
1.00
10.00
100.00
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Flow (cfs/cfs)
95th
75th
median
25th
5th
Figure 36. Modified box and whisker plot for USGS Gage No. 08168500, Guadalupe Rv abv Comal Rv at
New Braunfels, TX, normalized by dividing by the median flow.
To explain some of the secondary variations still exhibited following the filtering of flow values
by the median, an alternative step was performed that entailed subtracting the median flow and
then dividing by the interquartile range, which is the 75th percentile flow minus the 25th
percentile flow. The result of this test for two gages is presented in Figure 37 and Figure 38.
68
Monthly Streamflow Percentiles
Minus Monthly Median and Normalized by Interquartile Range
08020000 Sabine Rv nr Gladewater, TX
-2
0
2
4
6
8
10
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Flow (cfs/cfs)
95th
75th
50th
25th
5th
Figure 37. Modified box and whisker plot for USGS Gage No. 0802000, Sabine Rv nr Gladewater, TX,
minus the monthly median and normalized by the interquartile range.
Monthly Streamflow Percentiles
Minus Monthly Median and Normalized by Interquartile Range
08188500 San Antonio at Goliad, TX
-1
1
3
5
7
9
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Flow (cfs/cfs)
95th
75th
median
25th
5th
Figure 38. Modified box and whisker plot for USGS Gage No. 08188500, San Antonio Rv at Goliad, TX,
minus the monthly median and normalized by the interquartile range.
69
From a subset of the study gages, it is evident that the last normalization step was able to account
for a large majority of the seasonal variation for the 5th, 25th, and 75th percentile flows, as each of
these flows is depicted in the above plots as a nearly straight line parallel to the median.
However, the occurrence of 95th percentile streamflows is not explained at all by this statistical
technique and is thus likely a factor of the skewness of higher order statistical moments that
affect the distribution of extreme flood events.
Base flow
Base flow is the part of the stream discharge that is not attributable to direct runoff from
precipitation or snowmelt and is usually sustained by throughflow and groundwater flow; base
flow can be thought of as the typical flow condition of a river in the absence of a rain event. The
concept of base flow is not new to the environmental flow component model, nor is the task of
calculating base flow statistics a new one to the field on hydrology. As such, there are multiple
existing hydrograph-separation techniques that could be suitable for the determination of the
base flow component of the TIFP flow regime model.
One of these techniques, the Standard Institute of Hydrology Method, was applied to a subset of
the study gages using the United States Bureau of Reclamation’s BFI computer program
(Institute of Hydrology 1980, USBR 2004). An example of the results of this analysis is
presented in Figure 39. An adoption or incorporation of the principles of base flow separation or
the comparison of results from this technique of hydrograph analysis would be beneficial to the
TIFP as a means to better define and evaluate the base flow component of the four flow
component model. Furthermore, the BFI program has seen widespread use and is becoming the
standard base flow separation tool in the United States.
Overbank Flow
At the opposite end of the flow spectrum lies overbank flows, those streamflows where the water
level surpasses the height of the channel banks and results in flow over the floodplain. Unlike
70
the other flow components, overbank flows have a finite, physically-defined lower bound. Given
the channel slope, roughness, and cross-sectional area, it is easy to determine the flow volume
required to overtop the banks and cause flooding, and all flows up to this threshold will be
entirely contained with the stream channel. Also, bankfull flow is a common metric within the
field of geomorphology (Leopold et al 1964).
Blanco Rv nr Wimberley, TX - #08171000
WY1929 to 2004
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
Mar-86 May-86 Jul-86 Sep-86 Nov-86 Jan-87 Mar-87 May-87 Jul-87
Flow (cfs)
Total Flow
Baseflow
Figure 39. Base flow separation using the BFI program for USGS Gage No. 08171000, Blanco Rv nr
Wimberley, TX.
Methods to ascertain the incipient point of overbank flooding were evaluated as part of this
study. It was found that looking at additional published gage measurement data (in addition to
discharge) from the USGS provides additional insight into the delineation of the overbank flow
component that discharge data alone can not provide.
The USGS maintains an extensive nationwide network of real-time streamflow gaging stations
across the United States; these stations measure real-time stage and that information is combined
71
with an established rating curve to estimate the volumetric streamflow. The rating curves are
continually calibrated and verified by USGS employees who visit the gage sites to measure the
stream velocity and the flow cross-sectional area. Records of these site calibration visits are
available on the USGS website and the pertinent data for the 24 study gages is included in
Appendix D.
By plotting the discharge versus the measured active flow width and identifying the discharge
range where the flow width increases significantly, it is possible to estimate the streamflow
corresponding to the incipient point of overbank flooding. Examples of this technique are shown
in Figure 40 and Figure 41.
Sabine Rv nr Gladewater - 08020000
1
10
100
1,000
10,000
100,000
0 200 400 600 800 1000 1200 1400
channel width (ft)
streamflow (cfs)
2000s
1990s
1980s
1970s
1960s
1950s
1940s
1930s
Figure 40. Discharge versus wetted channel width for USGS Gage No. 0802000, Sabine Rv nr Gladewater,
TX, plotted by decade.
72
Navasota Rv nr Easterly - 08110500
0.1
1.0
10.0
100.0
1,000.0
10,000.0
100,000.0
0 200 400 600 800 1000 1200 1400
channel width (ft)
streamflow (cfs)
Figure 41. Discharge versus wetted channel width for USGS Gage No. 08110500, Navasota Rv nr Easterly,
TX.
This technique is currently under development and subsequent refinements may increase its
effectiveness and value as an analytical methodology for the estimation of overbank flows,
particularly the development of a methodology to quantitatively determine a break in slope. A
limitation of this method is that the gaging stations are often sited at altered river cross sections,
such as at bridges or culverts, or the gage itself causes alteration of the river, such as via the
installation of a weir. Thus, the streamflow required to cause overbank flooding at the gage site
may not be representative of that required at the broader reach or river. Nonetheless, techniques
for the evaluation of overbank flows based on geomorphic conditions may prove valuable in the
understanding of the overbank flow component and for the development of statistical thresholds
for defining overbank flows.
73
V. Conclusions and Recommendations
As general methods of streamflow hydrograph characterization, both IHA and HAT offer many
useful functions that illuminate the nature of streamflow patterns through time at a stream gaging
site. Both are simple to learn and to use, require the same readily-available input data, and are
based on a consistent theory of hydrograph characterization that incorporates the magnitude,
duration, timing, frequency, and rate of change of various hydrologic events. As such, the use of
either analysis package to support the range of natural flow variability in making instream flow
determinations represents a vast improvement over the old paradigm of a flat-line, minimum-
flow approach to environmental flow prescriptions. Statistics generated by each program have
many similarities, chiefly among them the inclusion of the 33 indicators of hydrologic alteration
in the 171 HAT indices. In their current versions, however, neither IHA nor HAT is directly
suitable for use in the TIFP; both would require modifications to be able to define flow
component statistics for wet, dry, and normal years.
IHA, as its name implies, is best suited to assess hydrologic alteration and to quantify the effects
of dam construction and other such water management development projects on the flow regime
via two-period analyses and the Range of Variability Approach. HAT, as its name implies, is
focused on characterizing streamflow, particularly in the context of a regional analysis of factors
that influence streamflow properties. The USGS Hydroecological Integrity Assessment Protocol
provides for the customization of NATHAT to specific regions for optimal results via the
development of region-specific, hydrologically-defined stream classifications. The stream
classifications and corresponding indices derived from a national dataset within NATHAT are
likely neither ecologically significant nor relevant for application in the State of Texas. Both
tools currently include temporal comparison tools (i.e., both tools can be used to compare results
for the same stream gage over different periods of time), though neither tool currently includes
spatial comparison tools (e.g., upstream versus downstream analysis of a specific reach or water
withdrawal point).
If the task of choosing between the two programs is framed as the question of which tool better
characterizes streamflow hydrographs in general, then there is little difference between these two
74
packages. But if the task of choosing between them is framed more narrowly as the question of
which program will best support the four-level flow characterization (subsistence flow, base
flow, high flow pulses, and overbank flows) of the Texas Instream Flow Program, then the HAT
program is a better choice than IHA as it allows for more flexibility in the determination of flow
component thresholds and greater capacity for regionalization. Note that the indices calculated
in IHA and HAT can all be calculated using independent software such as Microsoft Excel,
Matlab, or SAS, oftentimes allowing for greater flexibility.
It would be of great benefit to stakeholders and those performing the subbasin studies if TIFP
were to provide guidance and suggest methodologies to determine thresholds with ecological
and/or biological significance of the four flow components. This guidance will help to define
how a hydrologic assessment tool will be used to calculate statistics for the entire range of the
flow regime. Guidance on the definition of wet, dry, and normal years would be similarly
valuable.
The IHA Environmental Flow Component algorithm subdivides the daily streamflow hydrograph
into a set of discrete flow regimes based on the magnitude and rate of change of the discharge.
At any time, the flow must be in one of the five specified flow regimes (extreme low flow, low
flow, high flow pulses, small floods, and large floods, which are analogous to the four TIFP flow
components when small and large floods are considered in aggregate). The decision tree that
makes the distinction between flow components is complex and at times the criteria that
characterize the magnitude of flow and those that characterize the rate of change of flow get
combined in ways that are hard to justify or understand. For example, it is possible for high
flow pulses to have a lower daily discharge than surrounding periods of low flow due to the rate
of hydrograph recession. As such, it will likely be hard to apply and explain this algorithm to a
general audience and ever more difficult to codify into water allocation permits. We have
concluded that the procedure within IHA for identifying high flow pulses on the basis of percent
change in flow leads to illogical results where high flow pulses can have lower peak discharges
than surrounding periods of low flow. We have discussed these issues with Brian Richter, the
developer of IHA, and have found that he concurs and has indicated that he will address this
issue in a future version of IHA, along with changing the name of the IHA “low flow”
75
component to “base flow” to more accurately describe this component of the hydrograph and
also to ensure semantic consistency with the NRC recommendations and thus the TIFP flow
component model (Richter personal communication 2006).
HAT determines a set of streamflow indicators that characterize the flow hydrograph without
requiring it to be partitioned into a discrete number of flow regimes. This is a more flexible
approach that does not require the complex flow process used by IHA, but also does not
internally incorporate the flow component model of IHA and TIFP. As discussed above, HIP
provides for regionalization via the application of multivariate statistical analyses to a large
population of HAT results to identify a subset of no n-redundant indices that best characterize the
flow regime. Based on preliminary exploratory analyses at a limited number of streamflow
gages in the six priority subbasins, HAT can generate streamflow characteristics that will meet
the TIFP criteria and that correlate well with similarly derived characteristics by other hydrologic
methods such as for the determination of base flow. Unlike IHA, HAT makes use of non-
dimensional indices, many of which are normalized by the median daily streamflow. These
indices have great potential to be applied regionally as they are not direct measures of magnitude
and thus are somewhat decoupled from contributory drainage area. The seven-day, two-year low
flow (7Q2) is an important regulatory threshold in the Texas Surface Water Quality Standards;
Neither IHA nor HAT currently has the ability to explicitly calculate the 7Q2, but it can be
derived easily from indices within each of the programs.
The study of streamflow time series at the 24 selected gages within the 6 TIFP priority subbasins
for a period-of-record averaging 68 years has demonstrated that there are pronounced regional
patterns in the seasonality of the streamflow hydrograph, with little seasonal variation in central
Texas (e.g., Lower San Antonio and Lower Guadalupe subbasins) grading to a strongly seasonal
variation in East Texas (e.g., Lower Sabine and Middle Trinity subbasins). The degree of spatial
variability in the seasonal pattern of streamflow in East Texas is significantly greater than spatial
variation in precipitation alone can explain. In general, the long-term average daily hydrographs
in East Texas exhibited a sinusoidal pattern with a spring peak (i.e., May to June) and a fall
trough (i.e., August to September) when viewed on a semi-log plot. For a 30-year period of
record, precipitation is bimodal, irrespective of geography, with spring (i.e., May to June) and
76
fall (i.e., September to October) peaks; the precipitation gradient from East to Central Texas is
expressed in the magnitude of precipitation (more to the east) but not in the timing.
When considered over a long time scale, the seasonal variation in the flow regime of various
percentile levels of flow follows a consistent pattern, such that if the median flow goes down
then both high and low flow percentiles reduce proportionately to some extent, and vice versa
when the median flow increases. Although this pattern holds true across a sample of priority
gages studied, the high and low portions of the flow regime also each exhibit their own
secondary patterns of seasonality. This suggests that if the seasonal variation of the flow regime
is normalized by dividing by the median flow, the n some of the statistical variation of percentiles
of low and high flow ratios throughout the year may be explained. Also, it will likely be easier
to characterize the seasonal variation of within-bank high flow pulses and flood flows as ratios to
the seasonal variation of median flow rather than via rate of change of flow (i.e., hydrograph
onset and recession) arguments. As previously discussed, HAT makes use of such ratios to the
median flow as indices of regional streamflow characteristics but IHA does not.
For the hydrologic conditions present in Texas, the median flow is the most robust streamflow
characteristic that is available (i.e., it is most consistently estimated with a given number of data
values), and it can reasonably be estimated both from daily and from monthly streamflow data.
This means that either gaged flows or WAM-derived naturalized flows can be used to
recommend median flows. This is not so for definitions of instream flow pulses using rate of
change arguments where daily data (perhaps even hourly data on small watersheds) are required.
The time scales selected for evaluation and implementation in the TIFP are critical to the success
of environmental flow prescriptions as various riverine physical, chemical, and biological
processes operate on highly variable time scales.
The overbank flow component has a specific, physically-based flow threshold for any given
stream cross-section based on the stream slope, roughness, and channel capacity. Looking at
additional published gage data (in addition to discharge) provides additional insight into the
delineation of this flow component that discharge data alone can not provide. For example,
flood flows out of the stream banks can be characterized by plotting the discharge versus the
77
measured active flow width in the records available at all USGS streamflow gage sites and
identifying the discharge range where the flow width increases significantly, thus likely
indicating the incipient point of overbank flooding. The social, political, and economic issues of
prescribing flood events for ecological purposes are real and complex; work done as part of this
study has focused simply on promising analytical methodologies to identify the portion of the
historic flow regime that is overbank.
From this study we conclude that a Texas-customized version of HAT (as part of the HIP) is
suitable and preferable to IHA for application in the Texas Instream Flow Program. Further
work is necessary as part of the TIFP guidance or within the subbasin studies to define the
specific role of hydrologic assessment tools, particularly with respect to flow component
delineation and the definition of wet, dry, and normal years. Regardless of the tool selected, the
consideration of the variability of the natural flow regime and the development of instream flow
prescriptions accordingly represents a leap forward in the science and policy of instream flow
and in the ability to protect and/or restore a “sound ecological environment.”
78
Appendix A – Scope of Work
The purpose of this project is to assess the applicability of hydrologic analysis tools such as The
Nature Conservancy’s Indicators of Hydrologic Alteration (IHA) and US Geological Survey’s
Hydrologic Assessment Tool (HAT) for use in the Texas Instream Flow Program. Additionally,
the Principal Investigator may recommend improvements or changes to these software tools that
would make them more useful to the Texas Instream Flow Program.
Both IHA and HAT software are designed to help users analyze flow regimes and easily identify
the hydrologic impacts of human activities in a time series of flows. They are intended to provide
users with tools to characterize and compare hydrologic regimes in ecologically-meaningful
terms. Parameters are based on five fundamental characteristics of hydrologic flow regimes:
magnitude, timing, frequency, duration, and rate of change. Different time periods in the flow
record can be separated for analyzing the impact of such activities as reservoir construction and
operation, major water rights, or long-term changes in climate. The software may also prove
helpful for identifying subsistence and base flows, high flow pulses, and overbanking flows.
While IHA includes 67 statistical routines, HAT incorporates the 171 hydrologic indices of
Olden and Poff (2003). HAT offers the added utility of guiding users to tailor the selection of
hydrologic indices based on the type of hydrologic regime (i.e., stable, groundwater-fed stream
versus intermittent, flashy stream). HAT has been customized and adopted by the state of New
Jersey for use in its instream flow program. Similar customization is underway for use by other
states and Canada (Henriksen 2006).
Using daily flow information available from USGS gaging stations, Dr. Maidment and his
research staff will use IHA and HAT to evaluate the flow regime of at least four reaches in each
of six sub-basins currently under investigation by the state agencies: Guadalupe, Lower Sabine,
middle and lower Brazos, Trinity, and San Antonio rivers. Reaches will be selected in
coordination with state agencies and should include mainstem and tributary segments, as well as
regulated and non-impacted sites.
79
At the culmination of this exercise, Center for Research in Water Resources (CRWR) staff will
have assessed the appropriateness of these software packages for use in the Texas Instream Flow
Program, determined whether HAT or IHA is more suitable for Texas’ purposes, and
recommended any changes or enhancements that might make the software more effective for
Texas studies. Subject to time and resource availability and the ability to obtain program source
code, the contractor may modify IHA and/or HAT software as recommended and provide
training for state agency staff.
80
Appendix B – IHA Parameters
(from Nature Conservancy 2005)
81
82
83
84
85
Appendix C – HAT Parameters
(from Henriksen 2006)
Exp lanation – The following information is for the 171 hydrologic indices is from Olden and Poff (2003). The
USGS revised a limited number of the formula and (or) definitions when deemed appropriate. A USGS Scientific
Investigations Report in preparation will document these changes. The Olden and Poff (2003) article contains 12
additional references from which the indices were derived. Two of these articles are referenced here because they
provide examples and additional explanation for complex indices.
The alphanumeric code preceding each definition refers to the category of the flow regime (magnitude, frequency,
duration, timing, and rate of change) and type of flow event (A, average, L, low, and H, high) the hydrologic index
was developed to describe. Indices are numbered successively within each category. For example, MA1 is the first
index describing magnitude of the average flow condition.
MA# - Magnitude, average flow event
ML# -Magnitude, low flow event
MH# - Magnitude, high flow event
FL# - Frequency, low flow event
FH# - Frequency, high flow event
DL# - Duration, low flow event
DH# - Duration, high flow event
TA# - Timing, average flow event
TL# -Timing, low flow event
TH# -Timing, high flow event
RA# - Rate of change, average event
Following each definition, in parentheses, are (1) the units of the index, and (2) the type of data, temporal or spatial
data, from which the upper and lower percentiles limits (for example, 75/25) are derived. Temporal data are from a
multiyear daily flow record from a single stream gage. For example, index MA1- mean for the entire flow record -
uses 365 mean daily flow values for each year in the flow record to calculate the mean for the entire flow record.
Consequently, there are 365 values for each year to calculate upper and lower percentile limits. However, formulas
for 60 of the indices do not produce a range of values from which percentile limits can be calculated. MA5
(skewness), for example, the mean for the entire flow record divided by the median for the entire record results in a
single value, and thus, upper and lower percentile limits cannot be calculated.
Exceedance and percentile are used in the calculation for a number of indices. Note the difference - a 90 percent
exceedance means that 90 percent of the values are equal to or greater than the 90 percent exceedance value, while a
90th percentile means that 10 percent of the values are equal to or greater than the 90th percentile value.
MA1 Mean of the daily mean flow values for the entire flow record (cubic feet per second – temporal).
MA2 Median of the daily mean flow values for the entire flow record (cubic feet per second –
temporal).
MA3 Mean (or median - Use Preference option) of the coefficients of variation (standard
deviation/mean) for each year. Compute the coefficient of variation for each year of daily flows.
Compute the mean of the annual coefficients of variation (percent - temporal).
MA4 Standard deviation of the percentiles of the logs of the entire flow record divided by the mean of
percentiles of the logs. Compute the log10 of the daily flows for the entire record. Compute the
5th, 10th, 15th, 20th, 25th, 30th, 35th, 40th, 45th, 50th, 55th, 60th, 65th, 70th, 75th, 80th, 85th,
90th, and 95th percentiles for the logs of the entire flow record. Percentiles are computed by
interpolating between the ordered (ascending) logs of the flow values. Compute the standard
86
deviation and mean for the percentile values. Divide the standard deviation by the mean (percent -
spatial).
MA5 The skewness of the entire flow record is computed as the mean for the entire flow record (MA1)
divided by the median (MA2) for the entire flow record (dimensionless - spatial).
MA6 Range in daily flows is the ratio of the 10 percent to 90 percent exc eedance values for the entire
flow record. Compute the 5 percent to 95 percent exceedance values for the entire flow record.
Exceedance is computed by interpolating between the ordered (descending) flow values. Divide
the 10 percent exceedance value by the 90 percent value (dimensionless – spatial).
MA7 Range in daily flows is computed like MA6 except using the 20 percent and 80 percent
exceedance values. Divide the 20 percent exceedance value by the 80 percent value
(dimensionless - spatial).
MA8 Range in daily flows is computed like MA6 except using the 25 percent and 75 percent
exceedance values. Divide the 25 percent exceedance value by the 75 percent value
(dimensionless – spatial).
MA9 Spread in daily flows is the ratio of the difference between the 90th and 10th percentile of the logs
of the flow data to the log of the median of the entire flow record. Compute the log10 of the daily
flows for the entire record. Compute the 5th, 10th, 15th, 20th, 25th, 30th, 35th, 40th, 45th, 50th,
55th, 60th, 65th, 70th, 75th, 80th, 85th, 90th, and 95th percentiles for the logs of the entire flow
record. Percentiles are computed by interpolating between the ordered (ascending) logs of the flow
values. Compute MA9 as (90th – 10th) /log10(MA2) (dimensionless – spatial).
MA10 Spread in daily flows is computed like MA9 except using the 20th and 80th percentiles
(dimensionless – spatial).
MA11 Spread in daily flows is computed like MA9 except using the 25th and 75th percentiles
(dimensionless – spatial).
MA12-23 Means (or medians - Use Preference option) of monthly flow values.
Compute the means for each month over the entire flow record. For example, MA12 is the mean of
all January flow values over the entire record (cubic feet per second – temporal).
MA24-35 Variability (coefficient of variation) of monthly flow values. Compute the
standard deviation for each month in each year over the entire flow record. Divide the standard
deviation by the mean for each month. Average (or median - Use Preference option) thes e values
for each month across all years (percent – temporal).
MA36 Variability across monthly flows. Compute the minimum, maximum, and mean flows for each
month in the entire flow record. MA36 is the maximum monthly flow minus the minimum
monthly flow divided by the median monthly flow (dimensionless – spatial).
MA37 Variability across monthly flows. Compute the first (25th percentile) and the third (75th
percentile) quartiles (every month in the flow record). MA37 is the third quartile minus the first
quartile divided by the median of the monthly means (dimensionless – spatial).
MA38 Variability across monthly flows. Compute the 10th and 90th percentiles for the monthly means
(every month in the flow record). MA38 is the 90th percentile minus the 10th percentile divided
by the median of the monthly means (dimensionless – spatial).
MA39 Variability across monthly flows. Compute the standard deviation for the monthly means. MA39
is the standard deviation times 100 divided by the mean of the monthly means (percent – spatial).
MA40 Skewness in the monthly flows. MA40 is the mean of the monthly flow means minus the median
of the monthly means divided by the median of the monthly means (dimensionless – spatial).
87
MA41 Annual runoff. Compute the annual mean daily flows. MA41 is the mean of the annual means
divided by the drainage area (cubic feet per second/square mile – temporal).
MA42 Variability across annual flows. MA42 is the maximum annual flow minus the minimum annual
flow divided by the median annual flow (dimensionless – spatial).
MA43 Variability across annual flows. Compute the first (25th percentile) and third (75th percentile)
quartiles and the 10th and 90th percentiles for the annual means (every year in the flow record).
MA43 is the third quartile minus the first quartile divided by the median of the annual means
(dimensionless –spatial).
MA44 Variability across annual flows. Compute the first (25th percentile) and third (75th percentile)
quartiles and the 10th and 90th percentiles for the annual means (every year in the flow record).
MA44 is the 90th percentile minus the 10th percentile divided by the median of the annual means
(dimensionless – spatial).
MA45 Skewness in the annual flows. MA45 is the mean of the annual flow means minus the median of
the annual means divided by the median of the annual means (dimensionless – spatial).
ML1-12 Mean (or median - Use Preference option) minimum flows for each month across all years.
Compute the minimums for each month over the entire flow record. For example, ML1 is the
mean of the minimums of all January flow values over the entire record (cubic feet per second –
temporal).
ML13 Variability (coefficient of variation) across minimum monthly flow values. Compute the mean and
standard deviation for the minimum monthly flows over the entire flow record. ML13 is the
standard deviation times 100 divided by the mean minimum monthly flow for all years (percent –
spatial).
ML14 Compute the minimum annual flow for each year. ML14 is the mean of the ratios of minimum
annual flows to the median flow for each year (dimensionless – temporal).
ML15 Low flow index. ML15 is the mean of the ratios of minimum annual flows to the mean flow for
each year (dimensionless – temporal).
ML16 Median of annual minimum flows. ML16 is the median of the ratios of minimum annual flows to
the median flow for each year (dimensionless – temporal).
ML17 Base flow. Compute the mean annual flows. Compute the minimum of a 7-day moving average
flow for each year and divide them by the mean annual flow for that year. ML17 is the mean (or
median - Use Preference option) of those ratios (dimensionless – temporal).
ML18 Variability in base flow. Compute the standard deviation for the ratios of 7-day moving average
flows to mean annual flows for each year. ML18 is the standard deviation times 100 divided by
the mean of the ratios (percent – spatial).
ML19 Base flow. Compute the ratios of the minimum annual flow to mean annual flow for each year.
ML19 is the mean (or median - Use Preference option) of these ratios times 100 (dimensionless –
temporal).
ML20 Base flow. Divide the daily flow record into 5-day blocks. Find the minimum flow for each block.
Assign the minimum flow as a base flow for that block if 90 percent of that minimum flow is less
than the minimum flows for the blocks on either side. Otherwise, set it to zero. Fill in the zero
values using linear interpolation. Compute the total flow for the entire record and the total base
flow for the entire record. ML20 is the ratio of total flow to total base flow (dimensionless –
spatial).
ML21 Variability across annual minimum flows. Compute the mean and standard deviation for the
annual minimum flows. ML21 is the standard deviation times 100 divided by the mean (percent –
spatial).
88
ML22 Specific mean annual minimum flow. ML22 is the mean (or median - Use Preference option) of
the annual minimum flows divided by the drainage area (cubic feet per second/square mile –
temporal).
MH1-12 Mean (or median - Use Preference option) maximum flows for each month across all years.
Compute the maximums for each month over the entire flow record. For example, MH1 is the
mean of the maximums of all January flow values over the entire record (cubic feet per second –
temporal).
MH13 Variability (coefficient of variation) across maximum monthly flow values. Compute the mean
and standard deviation for the maximum monthly flows over the entire flow record. MH13 is the
standard deviation times 100 divided by the mean maximum monthly flow for all years (percent –
spatial).
MH14 Median of annual maximum flows. Compute the annual maximum flows from monthly maximum
flows. Compute the ratio of annual maximum flow to median annual flow for each year. MH14 is
the median of these ratios (dimensionless – temporal).
MH15 High flow discharge index. Compute the 1 percent exceedance value for the entire data record.
MH15 is the 1 percent exceedance value divided by the median flow for the entire record
(dimensionless – spatial).
MH16 High flow dis charge index. Compute the 10 percent exceedance value for the entire data record.
MH16 is the 10 percent exceedance value divided by the median flow for the entire record
(dimensionless – spatial).
MH17 High flow discharge index. Compute the 25 percent exceedance value for the entire data record.
MH17 is the 25 percent exceedance value divided by the median flow for the entire record
(dimensionless – spatial).
MH18 Variability across annual maximum flows. Compute the logs (log10) of the maximum annual
flows. Find the standard deviation and mean for these values. MH18 is the standard deviation
times 100 divided by the mean (percent – spatial).
MH19 Skewness in annual maximum flows. Use the equation:
MH19 = N2 x sum(qm3)-3N x sum(qm) x sum(qm2) + 2 x (sum(qm))3
N x (N-1) x (N-2) x stddev3
where: N = Number of years qm = Log10(annual maximum flows) stddev = Standard deviation of
the annual maximum flows. (dimensionless – spatial).
MH20 Specific mean annual maximum flow. MH20 is the mean (or median - Use Preference option) of
the annual maximum flows divided by the drainage area (cubic feet per second/square mile –
temporal).
MH21 High flow volume index. Compute the average volume for flow events above a threshold equal to
the median flow for the entire record. MH21 is the average volume divided by the median flow for
the entire record (days – temporal).
MH22 High flow volume. Compute the average volume for flow events above a threshold equal to three
times the median flow for the entire record. MH22 is the average volume divided by the median
flow for the entire record (days - temporal).
MH23 High flow volume. Compute the average volume for flow events above a threshold equal to seven
times the median flow for the entire record. MH23 is the average volume divided by the median
flow for the entire record (days - temporal).
MH24 High peak flow. Compute the average peak flow value for flow events above a threshold equal to
the median flow for the entire record. MH24 is the average peak flow divided by the median flow
for the entire record (dimensionless – temporal).
89
MH25 High peak flow. Compute the average peak-flow value for flow events above a threshold equal to
three times the median flow for the entire record. MH25 is the average peak flow divided by the
median flow for the entire record (dimensionless – temporal).
MH26 High peak flow. Compute the average peak flow value for flow events above a threshold equal to
seven times the median flow for the entire record. MH26 is the average peak flo w divided by the
median flow for the entire record (dimensionless – temporal).
MH27 High peak flow. Compute the average peak flow value for flow events above a threshold equal to
75th percentile value for the entire flow record. MH27 is the average peak flow divided by the
median flow for the entire record (dimensionless – temporal).
FL1 Low flood pulse count. Compute the average number of flow events with flows below a threshold
equal to the 25th percentile value for the entire flow record. FL1 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
FL2 Variability in low pulse count. Compute the standard deviation in the annual pulse counts for FL1.
FL2 is 100 times the standard deviation divided by the mean pulse count (percent – spatial).
FL3 Frequency of low pulse spells. Compute the average number of flow events with flows below a
threshold equal to 5 percent of the mean flow value for the entire flow record. FL3 is the average
(or median - Use Pre ference option) number of events (number of events/year – temporal).
FH1 High flood pulse count. Compute the average number of flow events with flows above a threshold
equal to the 75th percentile value for the entire flow record. FH1 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
FH2 Variability in high pulse count. Compute the standard deviation in the annual pulse counts for
FH1. FH2 is 100 times the standard deviation divided by the mean pulse count (number of
events/year – spatial).
FH3 High flood pulse count. Compute the average number of days per year that the flow is above a
threshold equal to three times the median flow for the entire record. FH3 is the mean (or median –
Use Preference option) of the annual number of days for all years (number of days/year –
temporal).
FH4 High flood pulse count. Compute the average number of days per year that the flow is above a
threshold equal to seven times the median flow for the entire record. FH4 is the mean (or median -
Use Preference option) of the annual number of days for all years (number of days/year –
temporal).
FH5 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to the median flow value for the entire flow record. FH5 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
FH6 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to three times the median flow value for the entire flow record. FH6 is the average (or median -
Use Preference option) number of events (number of events/year – temporal).
FH7 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to seven times the median flow value for the entire flow record. FH6 is the average (or median -
Use Preference option) number of events (number of events/year – temporal).
FH8 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to 25 percent exceedance value for the entire flow record. FH8 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
FH9 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to 75 percent exceedance value for the entire flow record. FH9 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
90
FH10 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to median of the annual minima for the entire flow record. FH10 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
Note - 1.67-year flood threshold (Poff, 1996) - For indices FH11, DH22, DH23, DH24, TA3, and TH3
compute the log10 of the peak annual flows. Compute the log10 of the daily flows for the peak
annual flow days. Calculate the coefficients for a linear regression equation for logs of peak
annual flow versus logs of average daily flow for peak days. Using the log peak flow for the 1.67
year recurrence interval (60th percentile) as input to the regression equation, predict the log10 of
the average daily flow. The threshold is 10 to the log10 (average daily flow) power (cubic
feet/second).
FH11 Flood frequency. Compute the average number of flow events with flows above a threshold equal
to flow corresponding to a 1.67-year recurrence interval. FH11 is the average (or median - Use
Preference option) number of events (number of events/year – temporal).
DL1 Annual minimum daily flow. Compute the minimum 1-day average flow for each year. DL1 is the
mean (or median - Use Preference option) of these values (cubic feet per second – temporal).
DL2 Annual minimum of 3-day moving average flow. Compute the minimum of a 3-day moving
average flow for each year. DL2 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DL3 Annual minimum of 7-day moving average flow. Compute the minimum of a 7-day moving
average flow for each year. DL3 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DL4 Annual minimum of 30-day moving average flow. Compute the minimum of a 30day moving
average flow for each year. DL4 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DL5 Annual minimum of 90-day moving average flow. Compute the minimum of a 90day moving
average flow for each year. DL5 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DL6 Variability of annual minimum daily average flow. Compute the standard deviation for the
minimum daily average flow. DL6 is 100 times the standard deviation divided by the mean
(percent – spatial).
DL7 Variability of annual minimum of 3-day moving average flow. Compute the standard deviation for
the minimum 3-day moving averages. DL7 is 100 times the standard deviation divided by the
mean (percent - spatial).
DL8 Variability of annual minimum of 7-day moving average flow. Compute the standard deviation for
the minimum 7-day moving averages. DL8 is 100 times the standard deviation divided by the
mean (percent - spatial).
DL9 Variability of annual minimum of 30-day moving average flow. Compute the standard deviation
for the minimum 30-day moving averages. DL9 is 100 times the standard deviation divided by the
mean (percent - spatial).
DL10 Variability of annual minimum of 90-day moving average flow. Compute the standard deviation
for the minimum 90-day moving averages. DL10 is 100 times the standard deviation divided by
the mean (percent - spatial).
DL11 Annual minimum daily flow divided by the median for the entire record. Compute the minimum
daily flow for each year. DL11 is the mean of these values divided by the median for the entire
record (dimensionless – temporal).
DL12 Annual minimum of 7-day moving average flow divided by the median for the entire record.
Compute the minimum of a 7-day moving average flow for each year. DL12 is the mean of these
values divided by the median for the entire record (dimensionless – temporal).
91
DL13 Annual minimum of 30-day moving average flow divided by the median for the entire record.
Compute the minimum of a 30-day moving average flow for each year. DL13 is the mean of these
values divided by the median for the entire record (dimensionless – temporal).
DL14 Low exceedance flows. Compute the 75 percent exceedance value for the entire flow record.
DL14 is the exceedance value divided by the median for the entire record (dimensionless –
spatial).
DL15 Low exceedance flows. Compute the 90 percent exceedance value for the entire flow record.
DL14 is the exceedance value divided by the median for the entire record (dimensionless –
spatial).
DL16 Low flow pulse duration. Compute the average pulse duration for each year for flow events below
a threshold equal to the 25th percentile value for the entire flow record. DL16 is the median of the
yearly average durations (numb er of days – temporal).
DL17 Variability in low pulse duration. Compute the standard deviation for the yearly average low pulse
durations. DL17 is 100 times the standard deviation divided by the mean of the yearly average low
pulse durations (percent – spatial).
DL18 Number of zero-flow days. Count the number of zero-flow days for the entire flow record. DL18
is the mean (or median - Use Preference option) annual number of zero flow days (number of
days/year – temporal).
DL19 Variability in the number of zero-flow days. Compute the standard deviation for the annual
number of zero -flow days. DL19 is 100 times the standard deviation divided by the mean annual
number of zero-flow days (percent – spatial).
DL20 Number of zero-flow months. While computing the mean monthly flow values, count the number
of months in which there was no flow over the entire flow record (percent – spatial).
DH1 Annual maximum daily flow. Compute the maximum of a 1-day moving average flow for each
year. DH1 is the mean (or median - Use Preference option) of these values (cubic feet per second
– temporal).
DH2 Annual maximum of 3-day moving average flows. Compute the maximum of a 3day moving
average flow for each year. DH2 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DH3 Annual maximum of 7-day moving average flows. Compute the maximum of a 7day moving
average flow for each year. DH3 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DH4 Annual maximum of 30-day moving average flows. Compute the maximum of 30day moving
average flows. Compute the maximum of a 30-day moving average flow for each year. DH4 is the
mean (or median - Use Preference option) of these values (cubic feet per second – temporal).
DH5 Annual maximum of 90-day moving average flows. Compute the maximum of a 90day moving
average flow for each year. DH5 is the mean (or median - Use Preference option) of these values
(cubic feet per second – temporal).
DH6 Variability of annual maximum daily flows. Compute the standard deviation for the maximum 1-
day moving averages. DH6 is 100 times the standard deviation divided by the mean (percent –
spatial).
DH7 Variability of annual maximum of 3-day moving average flows. Compute the standard deviation
for the maximum 3-day moving averages. DH7 is 100 times the standard deviation divided by the
mean (percent – spatial).
DH8 Variability of annual maximum of 7-day moving average flows. Compute the standard deviation
for the maximum 7-day moving averages. DH8 is 100 times the standard deviation divided by the
mean (percent – spatial).
92
DH9 Variability of annual maximum of 30-day moving average flows. Compute the standard deviation
for the maximum 30-day moving averages. DH9 is 100 times the standard deviation divided by
the mean (percent – spatial).
DH10 Variability of annual maximum of 90-day moving average flows. Compute the standard deviation
for the maximum 90-day moving averages. DH10 is 100 times the standard deviation divided by
the mean (percent – spatial).
DH11 Annual maximum of 1-day moving average flows divided by the median for the entire record.
Compute the maximum of a 1-day moving average flow for each year. DL11 is the mean of these
values divided by the median for the entire record (dimensionless – temporal).
DH12 Annual maximum of 7-day moving average flows divided by the median for the entire record.
Compute the maximum daily average flow for each year. DL12 is the mean of these values
divided by the median for the entire record (dimensionless – temporal).
DH13 Annual maximum of 30-day moving average flows divided by the median for the entire record.
Compute the maximum of a 30-day moving average flow for each year. DL13 is the mean of these
values divided by the median for the entire record (dimensionless – temporal).
DH14 Flood duration. Compute the mean of the mean monthly flow values. Find the 95th percentile for
the mean monthly flows. DH14 is the 95th percentile value divided by the mean of the monthly
means (dimensionless – spatial).
DH15 High flow pulse duration. Compute the average duration for flow events with flows above a
threshold equal to the 75th percentile value for each year in the flow record. DH15 is the median
of the yearly average durations (days/year – temporal).
DH16 Variability in high flow pulse duration. Compute the standard deviation for the yearly average
high pulse durations. DH16 is 100 times the standard deviation divided by the mean of the yearly
average high pulse durations (percent – spatial).
DH17 High flow duration. Compute the average duration of flow events with flows above a threshold
equal to the median flow value for the entire flow record. DH17 is the average (or median - Use
Preference option) duration of the events (days – temporal).
DH18 High flow duration. Compute the average duration of flow events with flows above a threshold
equal to three times the median flow value for the entire flow record. DH18 is the average (or
median - Use Preference option) duration of the events (days – temporal).
DH19 High flow duration. Compute the average duration of flow events with flows above a threshold
equal to seven times the median flow value for the entire flow record. DH19 is the average (or
median - Use Preference option) duration of the events (days – temporal).
DH20 High flow duration. Compute the 75th percentile value for the entire flow record. Compute the
average duration of flow events with flows above a threshold equal to the 75th percentile value for
the median annual flows. DH20 is the average (or median - Use Preference option) duration of the
events (days – temporal).
DH21 High flow duration. Compute the 25th percentile value for the entire flow record. Compute the
average duration of flow events with flows above a threshold equal to the 25th percentile value for
the entire set of flows. DH21 is the average (or median - Use Preference option) duration of the
events (days – temporal).
DH22 Flood interval. Compute the flood threshold as the flow equivalent for a flood recurrence of 1.67
years. Determine the median number of days between flood events for each year. DH22 is the
mean (or median - Use Preference option) of the yearly median number of days between flood
events (days – temporal).
DH23 Flood duration. Compute the flood threshold as the flow equivalent for a flood recurrence of 1.67
years. Determine the number of days each year that the flow remains above the flood threshold.
93
DH23 is the mean (or median - Use Preference option) of the number of flood days for years in
which floods occur (days – temporal).
DH24 Flood-free days. Compute the flood threshold as the flow equivalent for a flood recurrence of 1.67
years. Compute the maximum number of days that the flow is below the threshold for each year.
DH24 is the mean (or median - Use Preference option) of the maximum yearly no-flood days
(days – temporal).
TA1 Constancy. Constancy is computed via the formulation of Colwell (see example in Colwell, 1974).
A matrix of values is compiled where the rows are 11 flow categories and the columns are 365 (no
February 29th) days of the year. The cell values are the number of times that a flow falls into a
category on each day. The categories are:
log(flow) < .1 x log(mean flow), .1 x log(mean flow) <= log(flow) < .25 x log(mean flow) .25 x
log(mean flow) <= log(flow) < .5 x log(mean flow) .5 x log(mean flow) <= log(flow) < .75 x
log(mean flow) .75 x log(mean flow) <= log(flow) < 1.0 x log(mean flow)
1.0 x log(mean flow) <= log(flow) < 1.25 x log(mean flow)
1.25 x log(mean flow) <= log(flow) < 1.5 x log(mean flow)
1.5 x log(mean flow) <= log(flow) < 1.75 x log(mean flow)
1.75 x log(mean flow) <= log(flow) < 2.0 x log(mean flow)
2.0 x log(mean flow) <= log(flow) < 2.25 x log(mean flow) log(flow) >= 2.25 x log(mean flow)
The row totals, column totals, and grand total are computed. Using the equations for Shannon
information theory parameters, constancy is computed as:
1 - (uncertainty with respect to state)
log (number of state) (dimensionless – spatial).
TA2 Predictability. Predictability is computed from the same matrix as constancy (see example in
Colwell, 1974). It is computed as: 1-(uncertainty with respect to interaction of time and state -
uncertainty with respect to time
1 - (uncertainty with respect to interaction of time and state - uncertainty with respect to time)
log (number of state)
(dimensionless – spatial).
TA3 Seasonal predictability of flooding. Divide years up into 2-month periods (that is, Oct-Nov, Dec-
Jan, and so forth). Count the number of flood days (flow events with flows > 1.67-year flood) in
each period over the entire flow record. TA3 is the maximum number of flood days in any one
period divided by the total number of flood days (dimensionless – temporal).
TL1 Julian date of annual minimum. Determine the Julian date that the minimum flow occurs for each
water year. Transform the dates to relative values on a circular scale (radians or degrees). Compute
the x and y components for each year and average them across all years. Compute the mean angle
as the arc tangent of y-mean divided by x-mean. Transform the resultant angle back to Julian date
(Julian day – spatial).
TL2 Variability in Julian date of annual minima. Compute the coefficient of variation for the mean x
and y components and convert to a date (Julian day – spatial).
Note - 5-year flood threshold (Poff, 1996) – For TL3 and TH3, compute the log10 of the peak annual
flows. Compute the log10 of the daily flows for the peak annual flow days. Calculate the
coefficients for a linear regression equation for logs of peak annual flow versus logs of average
daily flow for peak days. Using the log peak flow for the 5-year recurrence interval (20th
percentile) as input to the regression equation, predict the log10 of the average daily flow. The
threshold is 10 to the log10 (average daily flow) power (cubic feet per second).
94
TL3 Seasonal predictability of low flow. Divide years up into 2-month periods (that is, Oct-Nov, Dec-
Jan, and so forth). Count the number of low flow events (flow events with flows <= 5 year flood
threshold) in each period over the entire flow record. TL3 is the maximum number of low flow
events in any one period divided by the total number of low flow events (dimensionless – spatial).
TL4 Seasonal predictability of non-low flow. Compute the number of days that flow is above the 5-
year flood threshold as the ratio of number of days to 365 or 366 (leap year) for each year. TL4 is
the maximum of the yearly ratios (dimensionless – spatial).
TH1 Julian date of annual maximum. Determine the Julian date that the maximum flow occurs for each
year. Transform the dates to relative values on a circular scale (radians or degrees). Compute the x
and y components for each year and average them across all years. Compute the mean angle as the
arc tangent of y-mean divided by x-mean. Transform the resultant angle back to Julian date (Julian
day – spatial).
TH2 Variability in Julian date of annual maxima. Compute the coefficient of variation for the mean x
and y components and convert to a date (Julian days - spatial).
TH3 Seasonal predictability of nonflooding. Computed as the maximum proportion of a 365-day year
that the flow is less than the 1.67-year flood threshold and also occurs in all years. Accumulate
nonflood days that span all years. TH3 is maximum length of those flood-free periods divided by
365 (dimensionless – spatial).
RA1 Rise rate. Compute the change in flow for days in which the change is positive for the entire flow
record. RA1 is the mean (or median - Use Preference option) of these values (cubic feet per
second/day – temporal).
RA2 Variability in rise rate. Compute the standard deviation for the positive flow changes. RA2 is 100
times the standard deviation divided by the mean (percent – spatial).
RA3 Fall rate. Compute the change in flow for days in which the change is negative for the entire flow
record. RA3 is the mean (or median – Use Preference option) of these values (cubic feet per
second/day – temporal).
RA4 Variability in fall rate. Compute the standard deviation for the negative flow changes. RA4 is 100
times the standard deviation divided by the mean (percent – spatial).
RA5 Number of day rises. Compute the number of days in which the flow is greater than the previous
day. RA5 is the number of positive gain days divided by the total number of days in the flow
record (dimensionless – spatial).
RA6 Change of flow. Compute the log10 of the flows for the entire flow record. Compute the change in
log of flow for days in which the change is positive for the entire flow record. RA6 is the median
of these values (cubic feet per second – temporal).
RA7 Change of flow. Compute the log10 of the flows for the entire flow record. Compute the change in
log of flow for days in which the change is negative for the entire flow record. RA7 is the median
of these log values (cubic feet per second/day – temporal).
RA8 Number of reversals. Compute the number of days in each year when the change in flow from one
day to the next changes direction. RA8 is the average (or median - Use Preference option) of the
yearly values (days - temporal).
RA9 Variability in reversals. Compute the standard deviation for the yearly reversal values. RA9 is 100
times the standard deviation divided by the mean (percent – spatial).
95
Appendix D – Electronic Files
IHA and HAT Output for 24 Priority Gages
Study Data
Study Maps, Models and Programs
96
References
77th Texas Legislature. (2001). “Senate Bill 2.”
Acreman, M., and Dunbar, M. J. (2004). "Defining environmental river flow requirements - a
review." Hydrology and Earth System Sciences, 8(5), 861-876.
Anderson, A. A., Hubbs, C., Winemiller, K. O., and Edwards, R. J. (1995). "Texas Fresh-Water
Fish Assemblages Following 3 Decades of Environmental-Change." Southwestern Naturalist,
40(3), 314-321.
Annear, T., Chisholm, I., Beecher, H., Locke, A., et al. (2004). Instream Flows for Riverine
Resource Stewardship: Revised edition. Instream Flow Council: Cheyenne, Wyoming.
Arthington A.H., King J.M., O’Keefe J.H., Bunn S.E., Day J.A., Pusey B.J., Bluhdorn D.R., and
Thame R. (1991). “Development of an holistic approach for assessing environmental flow
requirements of riverine ecosystems.” In Pigram J.J., Hooper B.A., eds. Water allocation for the
environment: proceedings of an international seminar and workshop. University of New
England Armidale (Australia): The Centre for Water Policy Research.
Austin, B. (2005). The Texas Instream Flow Program. Presentation at the Environmental Flows
Conference, Texas State University, San Marcos, October 31.
Bounds R. and Lyons B. (1979). Existing Reservoir and Stream Management: Statewide
Minimum Streamflow Recommendations. Austin, TX: Texas Parks and Wildlife Department.
Bovee K., Lamb B., Bartholow J., Stalnaker C., Taylor J., and Henriksen J. (1998). Stream
habitat analysis using the Instream Flow Incremental Methodology. U.S. Geological Survey
Information and Technical Report USGS/BRD-1998-0004. Fort Collins, CO: USGS Biological
Resources Division.
Bunn, S. E., and Arthington, A. H. (2002). "Basic principles and ecological consequences of
altered flow regimes for aquatic biodiversity." Environmental Management, 30(4), 492-507.
Clausen B, Iversen H.L., and Ovesen N.B. (2000). “Ecological flow indices for Danish streams.”
In Nordic Hydrological Conference, Nilsson T (ed.). Uppsala, Sweden; 3-10.
Clausen, B., and Biggs, B. J. F. (1997). "Relationships between benthic biota and hydrological
indices in New Zealand streams." Freshwater Biology, 38(2), 327-342.
Clausen, B., and Biggs, B. J. F. (2000). "Flow variables for ecological studies in temperate
streams: groupings based on covariance." Journal of Hydrology, 237(3-4), 184-197.
Colwell R.K. (1974). “Predictability, constancy, and contingency of periodic phenomena.”
Ecology, 55, 1148-1153.
97
Dunne T., and Leopold L.B. (1978). Water in Environmental Planning. W.H. Freeman and
Company, New York.
Dyson M., Bergkamp G. and Scanlon J., Ed. (2003). Flow: The Essentials of Environmental
Flow. International Union for Conservation of Nature and Natural Resources (IUCN): Gland,
Switzerland.
Espegren G.D. (1998). Evaluation of the Standards and Methods Used for Quantifying Instream
Flows in Colorado. Final Report. ColoradoWater Conservation Board: Denver, CO.
Henriksen J. (2005). Personal communication.
Henriksen J. A., Heasley J., Kennen J.G., and Niewsand S. (2006). Users’ manual for the
hydroecological integrity assessment process software (including the New Jersey Assessment
Tools). U.S. Geological Survey, Biological Resources Discipline, Open File Report 2006-1093.
http://www.nature.org/initiatives/freshwater/conservationtools/art17004.html
Hughes J.M.R., and James B. (1989). “A hydrological regionalization of streams in Victoria,
Australia, with implication for stream ecology. ” Australian Journal of Marine and Freshwater
Research, 40, 303-326.
Hughes D. A., and Hannart P. (2003). "A desktop model used to provide an initial estimate of the
ecological instream flow requirements of rivers in South Africa." Journal of Hydrology, 270(3-
4), 167-181.
Institute of Hydrology. (1980). Low flow studies: Wallingford, Oxon, United Kingdom. Report
Nos. 1 and 3.
Johnson I. W., and Law F. M. (1995). "Computer-Models For Quantifying The Hydro-Ecology
Of British Rivers." Journal of The Chartered Institution Of Water and Environmental
Management, 9(3), 290-297.
Jowett I. G. (1997). "Instream flow methods: A comparison of approaches." Regulated Rivers-
Research & Management, 13(2), 115-127.
Kaiser R. A., and Binion S. (1998). "Untying the gordian knot: Negotiated strategies for
protecting instream flows in Texas." Natural Resources Journal, 38(1), 157-196.
Karim K.G., Maureen, E., and Goulter, I.C. (1995). Review of determination of instream flow
requirements with special application to Australia. Water Resources Bulletin. Volume 31, No. 6,
pp. 1063-1077.
Karr J.R. (1991). “Biological integrity: a long neglected aspect of water resource management.”
Ecological Applications, 1, 66-84.
98
Kiesling R. (2003). “Applying Indicators of Hydrologic Alteration to Texas Streams- Overview
of Methods with Examples from the Trinity River Basin.” U.S. Geologic Survey, Fact Sheet
128-03.
Lamb B. L. (1995). "Criteria For Evaluating State Instream-Flow Programs - Deciding What
Works." Journal of Water Resources Planning and Management, 121(3), 270-274.
LBJ School of Public Affairs. (1978). Preserving Texas’ Natural Heritage. Research Project
Report 31.
Leopold L.B., Wolman M. G., and Miller J. P. (1964). Fluvial Processes in Geomorphology. W.
H. Freeman and Company, San Francisco.
Maidment D., Montagna P., Sansom A., Ward G., Winemiller K. (2005). Scientific Principles for
Definition of Environmental Flows. Statement for Environmental Flows Conference, Texas State
University, San Marcos, October 31.
Merwade V.M. (2004). “Geospatial Description of River Channels in Three Dimensions.” Dissertation,
The University of Texas at Austin.
Metsi Consultants. (2000). Consulting Services for the Establishment and Monitoring of the
Instream Flow Requirements for River Courses Downstream of LHWP (Lesotho Highlands
Water Project) dams. Final Report: summary of main findings. Report No. 648-F-02. (Authors:
King J, Sabet H, Brown C, Hirst S.) Lesotho Highlands Development Authority Contract 648:
Maseru, Lesotho.
Mosier D. and Ray R.. 1992. Instream Flows for the Lower Colorado River: Reconciling
Traditional Beneficial Uses with the Ecological Requirements of the Native Aquatic Community.
Austin, TX: Lower Colorado River Authority.
National Research Council Committee (2005). The Science of Instream Flows: A Review of the
Texas Instream Flow Program. Committee on Review of Methods for Establishing Instream
Flows for Texas Rivers, National Research Council. The National Academies Press,
Washington, D.C.
Nilsson, C., and Svedmark, M. (2002). "Basic principles and ecological consequences of
changing water regimes: Riparian plant communities." Environmental Management, 30(4), 468-
480.
Olden, J.D., and Poff, N.L. (2003). Redundancy and the choice of hydrologic indices for
characterizing streamflow regimes. River Research and Applications, 19, 101-121.
Pittman, E.G.R. (March 16, 2004). Letter to: The Honorable Kenneth Armbrister, Texas Senate,
and The Honorable Robert R. Puente, Texas House of Representatives, RE: Environmental
Flows.
99
Poff, N. L., Allan, J. D., Bain, M. B., Karr, J. R., Prestegaard, K.L., Richter, B. D., Sparks, R. E.,
and Stromberg, J. C. (1997). The natural flow regime. Bioscience, 47(11), 769-784.
Poff, N. L., and Ward, J. V. (1989). "Implications of Streamflow Variability And Predictability
For Lotic Community Structure - A Regional-Analysis Of Streamflow Patterns." Canadian
Journal Of Fisheries And Aquatic Sciences, 46(10), 1805-1818.
Poff, N.L. (1996). A hydrogeography of unregulated streams in the United States and an
examination of scale-dependence in some hydrological descriptors. Freshwater Biology, 36, 71-
91.
Postel, S. and Richter, B. (2003). Rivers for Life: Managing Water for People and Nature.
Island Press: Washington, D.C.
PRISM Group, Oregon State University, http://www.prismclimate.org, created January 2006.
Puckridge, J. T., Sheldon, F., Walker, K. F., and Boulton, A. J. (1998). "Flow variability and the
ecology of large rivers." Marine and Freshwater Research, 49(1), 55-72.
Resh V.H., Brown A.V., Covich A.P., Gurtz M.E., Li H.W., Minshall G.W., Reice S.R., Sheldon
A.L., Wallace J.B., and Wissmar R. (1988). “The role of disturbance in stream ecology. ” Journal
of the North American Benthological Society, (7), 433-455.
Richards R.P. (1989). “Measures of flow variability for Great Lakes tributaries.” Environmental
Monitoring and Assessment, 12, 361-377.
Richards R.P. (1990). “Measures of flow variability and a new flow-based classification of Great
Lakes tributaries.” Journal of Great Lakes Research, 16, 53-70.
Richter B. (2006). Personal communication.
Richter, B. D., Baumgartner, J. V., Powell, J., and Braun, D. P. (1996). A method for assessing
hydrologic alteration within ecosystems. Conservation Biology, 10(4), 1163-1174.
Richter, B. D., Baumgartner, J. V., Wigington, R., and Braun, D. P. (1997). How much water
does a river need? Freshwater Biology, 37(1), 231-249.
Richter, B. D., Mathews, R., and Wigington, R. (2003). Ecologically sustainable water
management: Managing river flows for ecological integrity. Ecological Applications, 13(1), 206-
224.
Richter, B.D., J.V. Baumgartner, D.P. Braun, and J. Powell. (1998). A Spatial Assessment of
Hydrologic Alteration Within a River Network. Regulated Rivers 14:329-340.
Shiau, J. T., and Wu, F. C. (2004). Feasible diversion and instream flow release using range of
100
variability approach. Journal Of Water Resources Planning And Management-ASCE, 130(5),
395-404.
Sparks R.E. (1995). “Need for ecosystem management of large rivers and their floodplains.”
Bioscience, 45, 168-182.
Stalnaker C.B., Lamb B.L., Henriksen J., Bovee K.D., and Bartholow J. (1994). The Instream
Flow Incremental Methodology: A Primer for IFIM. National Ecology Research Center, Internal
Publication. National Biological Survey: Fort Collins, CO.
Stalnaker C.B., Lamb B.L., Henriksen J., Bovee K.D., and Bartholow, J. (1995). The Instream
Flow Incremental Methodology-A primer for IFIM. National Biological Service, U. S.
Department of the Interior. Biological Report 29.
Stanford J.A., Ward J.V., Liss W.J., Frissell C.A., Williams R.N., Lichatowich J.A., and Coutant
C.C. (1996). “A general protocol for restoration of regulated rivers.” Regulated Rivers: Research
& Management ,12, 391-414.
Teasley, R.L. (2005). “Modeling the Forgotten River Segment of the Rio Grande/Bravo Basin.”
Thesis, The University of Texas at Austin.
Tennant, D.L. (1976), Instream Flow Regiments for Fish, Wildlife, Recreation, and Related
Environmental Resources” in Orsborn, J.F. and C.H. Allman (Editors). Proceedings of the
Symposium and Specialty Conference on Instream Flow Needs II. American Fisheries Society,
Bethesda, MD, pages 359-373.
Texas Commission on Environmental Quality. (2000). “Surface Water Quality Standards.”
Chapter 307 of the Texas Administrative Code.
Texas Commission on Environmental Quality. (2004). “Consensus Criteria for Environmental
Flow Needs.”
Texas Commission on Environmental Quality. (2006). “Water availability models ”.
http://www.tceq.state.tx.us/permitting/water_supply/water_rights/wam.html
Texas Instream Flow Program. (2002). Texas Instream Flow Studies: Programmatic Work Plan.
Texas Instream Flow Program. (2006). Draft Texas Instream Flow Studies: Technical Overview.
Texas Instream Flow Program. (June 15, 2006). Stakeholder Technical Workshop. University
of Texas at San Antonio.
Tharme R.E., and King J.M. (1998). “Development of the Building Block Methodology for
Instream Flow Assessments, and Supporting Research on the Effects of Different Magnitude
Flows on Riverine Ecosystems.” Water Research Commission Report No. 576/1/98.
101
Tharme, R. E. (2003). A global perspective on environmental flow assessment: Emerging trends
in the development and application of environmental flow methodologies for rivers. River
Research and Applications, 19(5-6), 397-441.
The Nature Conservancy. (2006). “IHA Applications Database.”
The Nature Conservancy. (2006a). “Indicators of Hydrologic Alteration – Version 7 User’s
Manual.”
The Nature Conservancy. (2006b). “Sustainable Waters Program website.”
http://www.nature.org/initiatives/freshwater/
Trungale J.F. (2006). “IHA Test.” Unpublished data.
Trungale J.F., Mayes K.B., Moss R.E., and Kleinsasser L.J. (2003). Draft Report: Using Water
Availability Models to Assess Alterations in Instream Flows. Resource Protection Division,
Texas Parks and Wildlife Department, Austin, Texas.
United States Bureau of Reclamation. (2004). “BFI: A Computer Program for Determining an
Index to Base Flow.” Department of the Interior.
United States Geological Survey (2002). Water Resources Data – Texas, Water Year 2001:
Vols. 1, 2, 3, and 5. USGS: Austin.
Ward G.H. (2000). “Texas Water at the Century’s turn — perspectives, reflections and a
comfort bag.” In Water for Texas: 2000 and Beyond.
Whiting, P.J. (2002). “Streamflow necessary for environmental maintenance.” Annual Review
Earth. Planet. Sci., 30,181-206.
Wood P.J., Agnew M.D., and Petts G.E. (2000). “Flow variations and macroinvertebrate
community responses in a small groundwater-dominated stream in south-east England.”
Hydrological Processes, 14, 3133-3147.
Wurbs R.A. (2003). “Water Rights Analysis Package (WRAP) Modeling System Users Manual”
Texas Water Resources Institute, TR-256.