United States EPA Region 10
Environmental Protection Agency (OEA-095)
910-R-03-003
December 2003
oEPA
Developing a Temperature Total Maximum Daily
Load for the Columbia and Snake Rivers:
Simulation Methods
-------
vvEPA
Developing a Temperature Total Maximum Daily Load
for the Columbia and Snake Rivers:
Simulation Methods
John Yearsley
EPA Region 10
Seattle, Washington
December 2003
-------
Table of Contents
Introduction 1
Model Description 1
Conceptual Approach 2
Model Development 2
Model Domain 3
Data Requirements 3
Parameter Estimation 4
Model Acceptance 4
TMDL Analysis 5
\Appendix_A\Forcing_Functions 5
\Appendix_A\TMDL\Site_Potential 6
\Appendix_A\TMDL\Point_Sources 8
\Appendix_A\TMDL\Dam_lmpacts 8
\Appendix_A\TMDL\Obverse_lmpacts 9
\Appendix_A\TMDL\Work_Space 9
\Appendix_A\TMDL\Hourly_Max 10
References 11
-------
List of Figures
Figure 1. Location map for Columbia TMDL 12
List of Tables
Table 1. Data sources foradvected energy in the Columbia and Snake rivers 13
Table 2. Temperatures monitoring sites in the Columbia River 15
Table 3. Temperatures monitoring sites in the Columbia River 15
Table 4. Meteorological station used to estimate heat budget 16
Table 5. Model performance statistics for RBM10 17
Table 6. Model applications in TMDL 18
Table 7. Point sources of thermal energy in the Columbia River 24
Table 8. Point sources of thermal energy in the Snake River 30
-------
Developing a Temperature Total Maximum Daily Load
for the Columbia and Snakes Rivers:
Simulation Methods
Introduction
The States of Idaho, Oregon and Washington and the U. S. Environmental Protection Agency
(EPA) are working in coordination with the Columbia Basin Tribes to develop Total Maximum
Daily Loads (TMDL) for Temperature and Total Dissolved Gas (TDG) on the Columbia and
Snake Rivers.
A TMDL for a water body is a document that identifies the amount of a pollutant that the water
body can receive and still meet Water Quality Standards (WQS). A TMDL also allocates
responsibility for reductions in the pollutant load that are necessary to achieve WQS. A TMDL is
required by the Clean Water Act for any stream reaches included by States or Tribes on their
lists of impaired waters required under Section 303(d) of the Clean Water Act. Impaired waters
are those that do not attain State or Tribal Water Quality Standards (WQS).
The Snake River from its confluence with the Salmon River at RM 188 to its confluence with the
Columbia River has been included on the 303(d) list of impaired waters for Temperature and
TDG by Idaho, Oregon or Washington as appropriate. Oregon and Washington included all of
the Columbia River on their 303(d) lists for TDG and most of the Columbia River on their lists for
Temperature. The Columbia River also exceeds the WQS of the Colville Confederated Tribes
for Temperature and TDG. The Spokane Tribe of Indians has WQS for the Columbia River that
have been adopted by the Tribe but not yet approved by EPA. These standards are also
exceeded in the Columbia River.
The states of Idaho, Oregon and Washington have assumed responsibility for developing
TMDL's for total dissolved gas for their respective waters in cooperation with the dam operators
within their boundaries. EPA is working with the Colville Tribe and the Spokane Tribes for the
portion of the dissolved gas TMDL within reservation boundaries. Oregon DEQ and
Washington DOE will collaborate on the total dissolved gas TMDL for the interstate portions of
the Columbia River.
The purpose of the Columbia and Snake River main stem temperature TMDL is to understand
the sources of temperature loadings and to allocate those loadings to meet state and tribal
water quality standards. EPA Region 10 is the technical lead for the temperature TMDL. EPA
Region 10 has chosen the mathematical model, RBM10, developed by EPA Region 10
(Yearsley et al, 2001) as the technical basis for developing a TMDL for temperature for the
Snake/Columbia Main stem.
Model Description
RBM10 (Yearsley et al, 2001) is a dynamic, one-dimensional model that simulates water
temperature using the energy budget method. It was originally developed to perform a
-------
temperature assessment of the Snake River from Lewiston, Idaho to its confluence with the
Columbia and of the Columbia River from Grand Coulee Dam to Bonneville Dam. The model
implements a mixed Eulerian-Lagrangian method for solving the dynamic energy budget
equation. The model uses reverse particle tracking to locate the starting point of a water parcel
at each computational time step. The water temperature at the starting point of each time step
for a parcel is determined by polynomial interpolation of simulated temperatures stored on a
fixed grid. The energy budget method (Wunderlich and Gras, 1967) is used to simulate the time
history of temperature as the parcel moves from its starting point at time, t-At, to ending point at
time, t. Kalman filtering is used to account for uncertainty in the water temperature data used
to develop the model.
Conceptual Approach
One-dimensional models have been used to assess water temperature in the Columbia River
system for a number of important environmental analyses. The Federal Water Pollution Control
Administration developed and applied a one-dimensional thermal energy budget model to the
Columbia River as part of the Columbia River Thermal Effects Study (Yearsley, 1969). The
Bonneville Power Administration and others used HEC-5Q, a one-dimensional water quality
model, to provide the temperature assessment for the Columbia River System Operation
Review (BPA, 1994). Normandeau Associates (1999) used a one-dimensional model to assess
temperature conditions in the Lower Snake River for the US Army Corps of Engineers. Perkins
and Richmond (2001) used the one-dimensional temperature model, MASS1, to simulate both
the impounded and unimpounded Snake rivers.
The water quality standards for most of the subject river reaches are written so as to limit the
increase in water temperatures as a result of human activities (Washington WQS) or
anthropogenic activities (Oregon WQS). This requires an estimate of temperature conditions in
the absence of the human activities. The conceptual approach used in the development of the
temperature TMDL is based on the notion that the effect of "human activities" can be estimated
by simulating conditions in the unimpounded river segments with no point sources present.
These results can then be used to determine the impacts of human activities associated with
hydroelectric projects, water withdrawals and point source discharges. An important
assumption in this approach is that impacts of "human activities" on water temperature outside
the geographical limits of this analysis will be addressed by other TMDL's or water quality plans;
that water quality and quantity at the boundaries of this TMDL are the result of existing
upstream activities.
Model Development
Much of the model development was done in the problem assessment phase of the TMDL and
is described in Yearsley et al (2001). Although the basic mathematical structure of the model
was not changed, the model framework was changed in a number of ways to accommodate the
needs of the TMDL.
-------
Model Domain
The Columbia River and the Snake River (Figure 1) are listed by the states of Oregon and
Washington as water-quality limited under Section 303(d) of the Clean Water Act. Listed
segments of these rivers in the model domain for the TMDL include the Columbia River from the
International Boundary (Columbia River Mile 745.0) to the Pacific Ocean near Astoria, Oregon
and the Snake River from its confluence with the Salmon River (Snake River Mile 188.2) to its
confluence with the Columbia River near Pasco, Washington (Columbia River Mile 324.0). In
addition, the Clearwater River from Orofino, Idaho (Clearwater River Mile 44.6) to its confluence
with the Snake River near Lewiston, Idaho (Snake River Mile 139.3) was included in the model
domain. The Clearwater River was included because of the influence releases from Dworshak
Dam on the North Fork of the Clearwater have on water temperatures of the Snake River
downstream from Lewiston. Although the Clearwater is not listed as water-quality limited under
Section 303(d), it may have an important role in any implementation plans developed from the
TMDL.
Major tributaries to the Columbia River and Snake River (Table 1) are included in the model
domain simply as point source inputs. That is the temperatures are not simulated, rather the
advected energy is treated as data input. While some of these tributaries are listed as water-
quality limited for temperature, any improvement in temperature that may result from TMDL's
written for these segments is not considered in this analysis. There are two reasons for this.
The size of these tributaries is such that their impact on the well-mixed temperature of the
Columbia is small. Furthermore, any temperature improvement in the development of TMDL's
on the tributaries will be included in the interpretation of the States' water quality standards as
described below.
Data Requirements
Data requirements for simulating water temperatures with RBM10 include the following
• The speed of the parcel along its characteristic path and the geometric properties of the
river are estimated from functional relationships between flow and geometry. A
gradually varied, steady flow model (USACE-HEC 1995) is used to establish the
functional relationships between flow and geometry. The basic data needed to establish
these relationships are depth as a function of width at various cross-sections. For the
purposes of the TMDL, data of this type were acquired from a number of sources as
described in Yearsley et al (2001).
• The energy budget is developed from meteorological data. The data are wind speed, dry
bulb temperature, relative humidity (or similar measure of water content), cloud cover,
and station pressure as a function of time.
• Advected thermal energy is defined by the stream flow and water temperature of
headwaters, tributaries and points sources.
-------
Parameter Estimation
The basic model framework for the TMDL was developed in the problem assessment and
described in Yearsley et al (2001). In the problem assessment the parameter estimation
process was implemented using a smaller model domain and water temperature data from the
period 1990-1994. For the TMDL, the parameter estimates were updated using the larger
model domain and water temperature data from the period 1995-1999. The water temperature
data are from monitoring sites below the dams and appear to be of higher quality and more
representative of well-mixed river temperatures. Station descriptions for the Columbia and
Snake rivers are given in Tables 2 and 3, respectively. The only parameter estimated was the
coefficient, Ke, in the relationship describing the rate of heat transfer due to evaporation
where,
qevap = the heat flux across the air-water interface due to evaporation,
p = the density of water,
L = the latent heat of vaporization,
ew = the saturated vapor pressure at the water surface temperature,
ea = the vapor pressure of the air above the water.
The energy budget for the model domain of the TMDL analysis is characterized by five different
meteorological provinces as described above (Table 4). The coefficient, Ke, was treated as a
variable for each meteorological province. The parameter estimation process was designed to
select the set of coefficients, Ke , that resulted in the minimum mean squared difference,
between simulated and observed for the monitoring sites shown in Table 5.
Model Acceptance
Statistics used to assess performance of the one-dimensional mathematical model, RBM10, are
similar to those described as appropriate for temperature models (Bartholow, 1989) and
recommended by van der Heijde and Elnawawy (1992) in EPA's guidance for selecting
groundwater models. The performance measures calculated for the TMDL simulations include:
Mean Difference
n
"-"mean
. .
-------
Absolute Mean Difference
N
YI T -T I
£-11 sim ' obs I
Damd = —
N
Root-Mean-Squared Difference
N
where,
N = the number of matched pairs of simulated and observed temperatures,
Tsim = the simulated temperature at the time of the nth observation
Tobs = the observed temperature.
The model performance statistics for the five-year (1995-1999) simulation period are given in
Table 5.
TMDL Analysis
Several types of simulations were used in the development of the temperature TMDL for the
Columbia and Snake rivers. Table 6 gives a summary of the simulation types. Simulation
results are reported at the compliance points as described in the TMDL . The compliance points
are just downstream from hydroelectric projects or, in the case of the unimpounded portion of
the Columbia River below Bonneville Dam, the compliance points are generally downstream
from major discharges. All the data and model source codes for developing the TMDL are on
the data CD (Appendix A).
The following discussion describes the contents of the directories on the data CD. The
computer programs and data files can be used to reproduce all the results used in the Final
Draft Temperature TMDL for the Columbia and Snake rivers. Each of the headings below is the
name of a directory on the data CD, Appendix A. File and directory names are given in
boldface.
\Appendix_A\Forcing_Functions
The files containing thirty-year record (1970 through 1999) of energy inputs to the system are
stored in this directory. Thermal energy inputs to the river system are from advected sources
(main stem boundaries and tributaries) and heat transfer across the air-water interface.
5
-------
Advected thermal energy from tributaries and main stem boundaries are estimated from river
flow and water temperatures. Data for advected thermal energy were obtained from the
sources shown in Tables 2 and 3. Missing water temperature data were filled by linear
interpolation when data gaps were of the order of a week or less. For larger gaps, a lag-one
Markov model was used to fill in missing data.
Heat transfer across the air-water interface is estimated from the meteorological data.
Meteorological data from six weather stations are used to estimate the energy budget for the
TMDL. The weather stations used in the TMDL and the segments of river are defined in
Table 4. Weather data for these stations are in the directory,\Appendix_A\Meteorology\. Only
three of the weather stations, Lewiston, Portland and Yakima, are primary stations, ones where
all the required meteorological variables are measured and reported. The other three weather
stations, Coulee Dam, Wenatchee and Richland, report only air temperature. The remaining
meteorological data for these stations was synthesized from the the primary station as shown in
Table 4. The energy budget files were created in the folder, ..\System_iv\setup, using the
programs, build_heat.exe, and energy.exe. The source code for build_heat.exe, and
energy.exe has hard-wired coding that looks for weather data in specific directories. The code
should be modified to ensure that the pathways specified in the coding are correct for the
particular application The output files with energy budget are named, CityName.budget.avg,
as in, Portland.budget.avg.
The file with thermal energy from advective sources (main stem boundary conditions and
tributaries) is named, No_Ocean.advect. The file with elevation data is named,
No_Ocean.elev. These files were created in the folder, ..\System_iv\setup using the program,
start_iv.exe in conjunction with the control file. no_ocean.control. These advection and
elevation files were used as the forcing functions for all the scenarios simulated for the TMDL.
\Appendix_A\TMDL\Site_Potential
The framework for implementing the State of Washington's water quality standards is
constructed around the concept of "site potential." Site potential, in the case of the temperature
TMDL, is defined as the daily-averaged, cross-sectional average temperature that would result
in the absence of impoundments and discharges of thermal energy from municipal and
industrial point sources as well as from various nonpoint sources. As described above, those
impacts on the thermal energy budget external to the defined boundaries of the temperature
TMDL are considered to be part of site potential. These impacts include those changes in flow
and temperature at the boundaries of the TMDL resulting from human activities. Non-stationary
impacts on climate such as global warming from industrial carbon dioxide production may also
be present in site potential as defined and realized with the inputs described below. Site-
potential is not, therefore, the temperature of the river prior to human development. Rather it is
the temperature that would result in the absence of major human activities in the listed river
segments.
Human activities in the existing river system configuration that have altered the thermal regime
of the Columbia and Snake rivers are:
-------
1. Construction of impoundments for hydroelectric facilities and navigational locks, which
increase the time waters of the Columbia and Snake are exposed to high summer
temperatures, increase the surface area exposed energy transfer across the air-water
interface and change the system's thermal response time.
2. Discharge of thermal energy from industrial and municipal point sources and agricultural
and urban nonpoint sources
3. Hydrologic modifications to the natural river system to generate electricity, provide
irrigation water for farmlands, and facilitate navigation.
4. Modifications of the watershed by urban development and agricultural and silvicultural
practices that reduce riparian vegetation, increase sediment loads, and change stream
or river geometry.
The TMDL focuses on those activities associated with the construction of impoundments,
thermal discharges from point and nonpoint sources and, implicitly, on the effects of hydrologic
modifications. The TMDL's developed for the listed tributaries of the Columbia and Snake
rivers should develop water quality plans that address thermal effects of modifications of the
watershed.
The impacts of impoundments on the thermal regime of the Columbia and Snake rivers are due
to both the change in river geometry and to operation of the hydroelectric facilities. All of the
hydroelectric projects within the model domain, with the exception of Grand Coulee Dam, are
run-of-the river projects. That is, the projects are operated such that approximately all the water
entering the reservoir is passed through the reservoir and released. As a result, the water level
in these reservoirs fluctuates very little. This does not mean the effects of the operation do not
have ecological impacts. It is well known, for example, that daily fluctuations in tailwater
elevations at Priest Rapids affect spawning and rearing habitat of fall Chinook and can cause
stranding of juvenile fish in the Hanford Reach of the Columbia River (Tiffan, 2003). However,
the impact of these operations on the daily-averaged, cross-sectional average temperature is
small. The major impact on the daily-average, cross-section water temperature is due to the
increase in width and depth resulting from the construction and operation of the impoundment.
Flood control is an operational feature of Lake Roosevelt, the reservoir behind Grand Coulee
Dam. As a result, the fluctuations in reservoir elevation are significant. Therefore, simulations
of water temperature for the existing conditions include the effects of storage for this project.
Point source inputs for the TMDL analysis are based on permit numbers provided by the State
of Oregon's Department of Environmental Quality (DEQ) and the State of Washington's
Department of Ecology (DOE). The energy inputs associated with these sources are given in
Table 8. Major discharges are shown individually while smaller discharges are aggregated and
shown as aggregated sources at the end of certain river reaches. In addition, a 20 megawatt
allowance of thermal energy is provided at each compliance point for general permits, general
permit includes impacts from stormwater discharge
The model domain for simulating site potential was created with the hydraulic properties in
7
-------
the file crtes.model.input.no_dams in the directory. A 30-year period of site potential
temperatures were simulated for the model domain using hydrologic data and weather data for
the period 1970 through 1999 and output to the files, Columbia.no_dams.avg and
Snake.no_dams.avg for the Columbia River and the Snake River, respectively.
A 30-year period of daily cross-sectionally averaged temperatures for existing conditions were
simulated for the model domain using hydrologic data and weather data for the period 1970
through 1999 using the executable rbm10_iii.exe. The control file used for simulating existing
conditions is crtes.model.input.dams. The simulation results were output to the files,
Columbia.dams.avg and Snake.dams.avg for the Columbia River and the Snake River,
respectively.
\Appendix_A\TMDI_\Point_Sources
The impact of point sources at compliance points was simulated by comparing the simulated
results from existing conditions, described above, with those same conditions when permitted
thermal sources are removed. Environmental forcing functions and parameters were the same
as those used for simulations of site potential and existing conditions. The basic source code
for RBM10 was modified to accommodate the addition of point sources. The source code is in
..\Model_iii_pnt\rbm10_iii and is named rbm10_pnt.f. The executable is named
rbm10_iii_pnt.exe.
Simulations were performed in two directories, ..\Existing_Sources and ..\Zero_Discharge
using rbm10_iii_pnt.exe (point source version) in conjuncton with the control file,
crtes.model.input.dams. The Fortran source code, rbm10_iii_pnt.f, in the directory,
..\Existing_Sources, differs slightly from that in the directory, ..\Zero_Discharge. The
difference is due to hardwired coding that ignores point sources in the directory,
..\Zero_Discharge. The source code for each version is stored in the appropriate directory.
This version of the control file has also been modified to accommodate the point sources.
Simulated results are output at the compliance points as,
..\Existing_Sources\Columbia_Exist.RM_xxx, ..\Zero_Discharge\Columbia_Zero.RM_xxx,
..\Existing_Sources\Snake_Exist.RM_xxx, ..\Zero_Discharge\Snake_Zero.RM_xxx,
where "xxx" is the river mile of the compliance point. The directory labeled, Existing_Sources,
incorporates the thermal loadings associated with the point sources, while the directory labeled,
Zero_Discharge, simulated the impounded system with no thermal discharges from point
sources.
\Appendix_A\TMDI_\Dam_lmpacts
The effect of adding individual hydroelectric projects to the unimpounded river was simulated by
starting with the river systems in their present configuration of hydroelectric projects.
Simulations of the system were then performed by changing, one hydroelectric project at a time,
the hydraulic coefficients of the portion of the river upstream of the dam from freely-flowing river
type to reservoir type. This assumes that the impounded section of the river associated with a
specific hydroelectric project will not affect the hydraulic characteristics of the unimpounded
8
-------
river both upstream and downstream of the of the project being evaluated. Environmental
forcing functions and parameters were the same as those used for other simulations. Results
are in,..\DamName\, where "DamName" is the name of the hydroelectric project.
Simulations were performed with the same forcing functions used for other scenarios and the
version of the source code used for the characterization of site potential. The version of the
source code is labeled, rbm10_ iii.f, in the directory, ..\Model_lll\rbm_iii\Original_Code. The
executable associated with this source code, rbm10_iii.exe, was used in conjunction with
control file for each each dam and labeled crtes.model.final.nnn, where, nnn, is the symbol for
the specific dam as in the example, crtes.model.final.BON, the file containing simulated effect
of adding Bonneville Dam to the unimpounded river.
\Appendix_A\TMDI_\Obverse_lmpacts
For purposes of the TMDL, the impact of individual dams was simulated by changing, one
project at a time, the hydraulic properties of the reservoir behind the dam to hydraulic properties
representing the freely-flowing river. As in the case above where individual dams were added to
the natural river system, this set of scenarios assumes that the hydraulic properties of the freely-
flowing river will not be affected significantly by hydroelectric projects upstream or downstream
from the one being evaluated. Environmental forcing functions and parameters were the same
as those used for other simulations. Results are in,..\DamName\, where "DamName" is the
name of the hydroelectric project.
Simulations were performed with the same forcing functions used for other scenarios and the
version of the source code used for the characterization of site potential. The version of the
source code is labeled, rbm10_ iii.f, in the directory, ..\Model_lll\rbm_iii\Original_Code. The
executable associated with this source code, rbm10_iii.exe, was used in conjunction with
control file for each each dam and labeled DamName.Obverse, where, DamName, is the
symbol for the specific dam as in the example, Bonneville, for Bonneville Dam.
Output for the simulations in the Columbia and Snake rivers is to files named
RiverName.nnn.Obv. RiverName is either Columbia or Snake and nnn, is the symbol for the
specific dam as in the example, Columbia.BON.Obv, for the file with the simulated effects of
removing Bonneville Dam from the impounded river.
\Appendix_A\TM DL\Work_Space
The software that implements RBM10, the time-dependent, one-dimensional energy budget
model, was modified such that simulated results could be compared to the water quality
standards of Washington and Oregon. The reference data sets used for making comparisons
were the simulations based on site potential (COLUMBIA.NO_DAMS.AVG and
SNAKE.NO_DAMS.AVG). The modified program is named RBM10_TMDL.F and is located in
the directory \Appendix_A\TMDL\Work_Space\RBM_TMDL.
-------
Several TMDL scenarios were evaluated using the RBM10 model framework. 21 of these
scenarios, including the scenario used for the draft final TMDL are in the directory,
\Appendix_A\TMDI_\Work_Space\TMDI__final. All but the scenario used for the draft final
TMDL, Scenario_21a, are archived in the compressed file, Scenario_Archive.zip. A brief
description of the 21 scenarios is in the document,
\Appendix_A\TMDL\Work_Space\Work_Space_log.doc.
The first step in the TMDL was to allocate loads to the permitted discharges. The simulations of
point sources described above provided estimates showing that the far-field temperature effects
of permitted discharges did not exceed the water quality standards of the states of Oregon and
Washington. The point sources were, therefore, allocated thermal loads based on their
National Pollution Discharge Elimination System (NPDES) permits. The allocations for the
dams were determined from the results of the obverse impacts analysis. That is, each dam was
allocated a temperature change based on the daily-averaged difference between simulated
results for the existing system and the results when the particular dam was removed from the
system. The file containing the allocations is named "DELTA.TMDL". The results were
compared with the water quality standards of the states or Oregon and Washington to assure
compliance.
\Appendix_A\TMDL\Hourly_Max
Hourly water temperatures in the Columbia and Snake rivers were simulated using hourly
meteorological data and daily averaged temperature and flow data. Hourly simulations were
performed for calendar years 1992 and 1997 and the results saved in the directory,
\Appendix_A\TMDL\Hourly_Max\Results. The forcing functions for advection and energy
transfer across the air-water interface are the advection file, No_Ocean.advect, and
meteorological data files CityName.199x.hourly, where "CityName" is the name of the
weather station "x" is either "2" or "7", as in the example, Lewiston.1992.hourly.
The source code, rbm10_iii.f, is the same as that used to develop the site potential.
10
-------
References
Bartholow, J. M. 1989. Stream temperature investigations—Field and analytic methods.
Instream Flow and Info. Paper No. 13. U.S. Fish and Wildlife Service.
Bonneville Power Administration et al. 1994. Columbia River system operation review.
Appendix M, Water quality. DOE/EIS-0170. Bonneville Power Administration, U.S. Army
Corps of Engineers, and U.S. Bureau of Reclamation, Portland, Oregon.
Normandeau Associates. 1999. Lower Snake River temperature and biological productivity
modeling. R-16031.007. Preliminary review draft. Prepared for the Department of the Army,
Corps of Engineers, Walla Walla, Washington.
Perkins, W.A. and M.C. Richmond. 2001. Long-term, one-dimensional simulation of Lower
Snake River temperatures for current and unimpounded conditions. Battelle Pacific
Northwest Laboratory, Richland, Washington. 86 pp.
Tiffan, K. 2003. Evaluation of the effect of water management on Fall Chinook spawning and
rearing habitat and on stranding of juvenile Fall Chinook in the Hanford Reach of the
Columbia River. http://wfrc.usgs.gov/research/fish%20populations/STRondorf8.htm viewed on
10/16/2003.
USACE-HEC (U.S. Army Corps of Engineers). 1995. HEC-RAS: River analysis system. U.S.
Army Corps of Engineers, Hydrologic Engineering Center, Davis, California.
Van de Heijde, P.K.M. and O.A. Elnawawy. 1992. Quality assurance and quality control in the
development and application of ground-water models. EPA/600/R-93/011. US
Environmental Protection Agency, Office of Research and Development, Ada, Oklahoma.
68pp.
Wunderlich, W.O., and R. Gras. 1967. Heat and mass transfer between a water surface and
the atmosphere. Tennessee Valley Authority, Division of Water Cont. Planning, Norris,
Tennessee
Yearsley, J.R. 1969. A mathematical model for predicting temperatures in rivers and river-run
reservoirs. Working Paper No. 65, Federal Water Pollution Control Agency, Portland,
Oregon.
Yearsley, J., D. Kama, S. Peene and B. Watson. 2001. Application of a 1-D heat budget model
to the Columbia River system. Final report 901-R-01-001 by the U.S. Environmental
Protection Agency, Region 10, Seattle, Washington
11
-------
CnliHHtxa - ilake Hl^cr IMUt
<(,„,,.,,.. i ,,/iil, TDMI
1 1(7
" 0«n
^ Location? E«£««Jltn9
W«tcrQudily Criteria
1->
V - S
_
't~^""^
Psrlland
Figure 1. Location map for Columbia TMDL
12
-------
Table 1. Data sources for advected energy in the Columbia and Snake rivers
Salmon River at White Bird,
Idaho
Grande Ronde River at Troy,
Oregon
Snake River near Anatone,
Washington
Clearwater River at Orofino,
Idaho
North Fork Clearwater at
Dworshak Dam
Tucannon near Starbuck,
Washington
Palouse River near Hooper,
Washington
Columbia River at the
International Boundary
Kettle River near Laurier,
Washington
Colville River at Kettle Falls,
Washington
Spokane River at Long Lake
Feeder Canal at Grand Coulee,
Washington
Okanogan River at Malott,
Washington
Methow River near Pateros,
Washington
Wenatchee River at Monitor,
Washington
Crab Creek near Moses Lake,
Washington
Yakima River at Kiona,
Washington
Walla Walla River at Touchet,
Washington
Umatilla River near Umatilla,
Oregon
John Day River at McDonald
Ferry, Oregon
Data Sources
Flow
USGS 13317000
USGS 13333000
USGS 13334300
USGS 13340000
DART Site DWR
USGS 13344500
USGS 13351000
USGS 12399500
USGS 12404500
USGS 12409000
USGS 12433000
USGS 12435500
USGS 12447200
USGS 12449950
USGS 12462500
USGS 12467000
USGS 12510500
USGS 14018500
USGS 14033500
USGS 14048000
Temperature
USGS 133 17000
Washington DOE 35C070
USGS 13334300; DART Site
ANQW
USGS 13340000
DART Site DWR
Washington DOE 35B060
Washington DOE 34A070
DART Site CIBW
Washington DOE 59A070
Washington DOE 60A070
USGS 12433000
Washington DOE 49A070
Washington DOE 48A070
Washington DOE 45A070
Washington DOE 41A070
Washington DOE 37A090
USGS 14018500
USGS 14033500
Oregon DEQ 404065
13
-------
Table 1 (continued). Data sources for advected energy on the main stem Columbia and
Snake Rivers
Deschutes River at Moody, near
Biggs, Oregon
Klickitat River
Hood River
Sandy below Bull Run Reservoir,
Oregon
Willamette River at Portland,
Oregon
Lewis River at Ariel, Washington
Cowlitz River at Castle Rock,
Oregon
USGS 14103000
USGS 141 05700
USGS 14120000
USGS 14142500
USGS 14211720
USGS 14220500
USGS 14243000
Oregon DEQ 402081
USGS 141 13000
Oregon DEQ 402081
Oregon DEQ 402349
Constrained to Columbia River
Oregon DEQ 402081
Oregon DEQ402081
14
-------
Table 2. Temperatures monitoring sites in the Columbia River
Station
Bonneville Dam tailwater
The Dalles Dam tailwater
John Day Dam tailwater
McNary Dam tailwater-Washington
Priest Rapids tailwater
Wanapum Dam tailwater
Rock Island Dam tailwater
Rocky Reach Dam tailwater
Wells Dam tailwater
Chief Joseph Dam tailwater
Grand Coulee Dam tailwater
Station
Identifier
BON
TDDO
JHAW
MCPW
PRXW
WANW
RIGW
RRDW
WELW
CHQW
GCGW
Station Description
Columbia RM 146: Right end
of spillway near center of dam
Columbia RM 190: Left bank
one mile d/s of dam
Columbia RM 215: Dam
tailwater Right bank of river
Columbia RM 291: Dam
Tailwater Right bank of river
Columbia RM 396: Tailwater
D/s of dam
Columbia RM 415: Tailwater
D/s of dam
Columbia RM 452: Tailwater
D/s of dam
Columbia RM 472 Tailwater
D/s of dam
Columbia RM 514: Tailwater
D/s of dam
Columbia RM 545: Tailwater
D/s of dam
Columbia RM 590: Six miles
D/s of dam
Table 3. Temperatures monitoring sites in the Columbia River
Station
Ice Harbor Dam tailwater
Lower Monumental Dam tailwater
Little Goose Dam tailwater
Lower Granite Dan tailwater
Station
Identifier
IDSW
LMNW
LGSW
LGNW
Station Description
Snake RM 6.8: Right bank
15, 400 feet d/s of dam
Snake RM 40.8: Left bank
4, 300 feet d/s of dam
Snake RM 69.5:Right bank
3, 900 feet d/s of dam
Snake RM 106.8: Right bank
3, 500 feet d/s of dam
15
-------
Table 4. Meteorological station used to estimate heat budget
Station Name Station #
Station Type
Station Used to
Synthesize
River Segments
Lewiston, Idaho WBAN 24149
WBAN 24229
WBAN 24157
WBAN 24243
NCDC451767
Portland,
Oregon
Spokane,
Washington
Yakima,
Washington
Coulee Dam
Richland
Wenatchee
NCDC457015
NCDC 459074
Surface Airways
Surface Airways
Surface Airways
Surface Airways
Local
Climatological
Data
Local
Climatological
Data
Local
Climatological
Data
Spokane
Yakima
Spokane
Snake RM 0.0-188.0
Clearwater RM 0-42.0
Columbia RM 0.0-1245.5
Columbia RM 292.0-453.4
Columbia RM 738.0 - 596.5
Columbia RM 292.0-145.5
Columbia RM 596.5-453.4
16
-------
Table 5. Model performance statistics for RBM10
Columbia River
# of
Samples
Absolute
Mean
Difference
Average RMS Standard
Difference Difference Deviation
Grand Coulee
Chief Joseph
Wells
Rocky Reach
Rock Island
Wanapum
Priest Rapids
McNary
John Day
The Dalles
Bonneville
1150
678
348
512
534
889
773
1222
666
703
493
0.73
1.05
0.52
0.67
0.64
0.81
0.78
0.56
0.46
0.43
1.07
-0.08
0.81
0.40
0.57
0.53
0.05
-0.09
-0.34
0.02
0.06
-0.59
0.97
1.46
0.70
0.86
0.85
1.25
1.03
0.72
0.59
0.54
1.36
0.94
1.46
0.32
0.42
0.43
1.56
1.06
0.40
0.35
0.29
1.49
Snake River
Lower Granite
Little Goose
Lower
Monumental
Ice Harbor
1144
746
819
1222
0.83
0.77
0.73
0.78
-0.64
-0.29
-0.19
-0.30
1.03
1.13
0.93
0.93
0.65
1.19
0.82
0.78
17
-------
Table 6. Model applications in TMDL
RBM10
Application
Model Setup
Output Files
Findings
1. Site Potential
Temperature
Daily Time Step
Un-impounded River
Existing tributary/boundary inflows
No point sources
Columbia.no_dams.avg
Snake.no_dams.avg
Temperatures exceed numeric
criteria (e.g., 20 deg C in lower
Columbia) in absence of
human activity on mainstems
2. Actual
Temperature
Daily Time Step
Impounded River
Existing tributary/boundary inflows
All point sources
Columbia.dams.avg
Snake.dams.avg
Actual temperatures are higher
than site potential
temperatures in late
summer/fall
(e.g., 3.5 deg C warmer at
John Day dam)
3. Point Source
Cumulative
Impacts
Daily Time Step
Impounded River
Existing tributary/boundary inflows
Point Sources - 2 scenarios
Scenario 1: No point sources
Scenario 2: All point sources
.\Zero_Discharge\
Columbia_Zero.RM_xxx
Snake_Zero.RM_xxx
AExisting_Sources\
Columbia_Exist.RM_xxx
Snake Exist.RM xxx
Maximum, cumulative point
source impact less than 0.14
degC.
4. Individual
Dam Impacts
Daily Time Step
Impounded River
Existing tributary/boundary inflows
All point sources
Dams -16 scenarios -
Scenario 1: all dams included
Scenarios 2-16: one dam removed
and effects evaluated
. AO bve rse_l m pacts
Columbia.nnn.Obv
Snake.nn.Obv
18
Maximum temperature
increases due to dams range
from 0.1 deg C (Rock Island)
to 6.2 deg C (Grand Coulee)
-------
19
-------
Table 6 (continued). Model Applications Used in Development of TMDL
RBM10
Application
Model Setup
Outputs
Findings
5. TMDL Target
Temperatures
Daily Time Step
Un-impounded River
Existing tributary/boundary inflows
All point sources
Dams - 2 seasons:
Aug-Oct
Mean daily effect from 5 dams (Wells,
Rocky Reach, Rock Island, Priest
Rapids, and The Dalles), zero effect
from remaining dams
Nov-Feb
Mean daily effect from 5 dams (Wells,
Rocky Reach, Rock Island, Priest
Rapids, and The Dalles), 0.12 deg C
effect from remaining dams
.\TMDL_Final\Scenario_21 a\
Columbia_TMDL.RM_xxx
Snake TMDL.RM xxx
This model setup represents a
fully allocated temperature
increment, based on
compliance with standards at
RM42
6. Diurnal
Fluctuation
Hourly time step
Impounded and Un-impounded
Existing tributary/boundary inflows
.\Hourly_max\Results
Columbia.no_dams.yyyy.hourly
Columbia.dams.yyyy.hourly
Snake.no_dams.yyyy.hourly
Snake.dams.yyyy.hourly
20
Greater diurnal fluctuations in
un-impounded river than
impounded river
-------
21
-------
22
-------
23
-------
Table 7. Point sources of thermal energy in the Columbia River
Facility
Avista - Kettle Falls
Grand Coulee - Chief
Joseph
Grand Coulee Dam
Grand Coulee
City of Coulee Dam
Chief Joseph-Wells
Chief Joseph Dam
Bridgeport STP
Brewster
Patterns STP
Wells - Rocky Reach
Wells Dam
Wells Hydro Project
Chelan STP
Entiat STP
Rocky Reach - Rock Island
Rocky Reach Dam
Tree Top
Naumes Processing
Columbia Cold Storage
E Wenatchee STP
KB Alloys
Specialty Chemical
Alcoa Wenatchee
Rock Island - Wanapum
Rock Island
Rock Island West
Powerhouse
Vantage STP
Wanapum - Priest Rapids
Priest Rapids - McNary
Columbia Generating Sta
Fluor Daniel Hanford, Inc
Richland STP
Baker Produce
Twin City Foods
Kennewick
Pasco
River
Mile
702.4
596.6
596.6
596.0
545.1
543.7
529.8
524.1
515.8
515.0
503.5
485.0
474.9
470.8
470.5
466.3
465.7
458.5
456.3
455.2
453.4
453.4
420.6
351.8
347.0
337.1
329.2
328.3
328.0
327.6
Thermal
Load
(MW)
1.374
0.906
2.518
1.099
0.030
1.510
1.832
0.414
0.004
0.015
7.399
0.604
0.020
0.331
10.543
5.990
19.126
1.484
15.464
17.847
0.008
0.008
0.438
53.697
27.902
57.378
0.040
0.041
61.405
22.752
Allocation
Bubble
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Flow, Q
(cfs)
0.360
0.360
0.279
1.063
0.464
1.806
0.009
0.464
0.563
0.127
1.164
0.002
0.006
2.274
0.186
2.468
0.006
0.127
2.674
0.928
5.879
0.464
6.189
6.962
23.230
0.002
0.002
0.135
0.139
15.114
8.475
17.638
0.012
0.012
18.876
6.993
Temperature
(deg C)
32.2
32.2
27.5
20.0
20.0
21.2
27.5
22.0
23.0
23.0
22.6
20.0
20.0
25.0
23.0
24.8
27.5
22.0
33.3
23.9
23.5
27.0
21.1
21.6
23.5
27.5
27.5
26.0
26.0
30.0
27.8
23.5
27.5
28.3
23.0
27.5
24
-------
Total
67.121
25
-------
Table 7(continued). Point sources of thermal energy in the Columbia River
Facility River
Mile
Agrium Bowles Road 322.6
Agrium Game Farm Road 321.0
Sanvik Metals 321.0
Boise Cascade Walulla 316.0
McNary to John Day
Umatilla STP 289.0
Goldendale 216.7
John Day- The Dalles
Biggs OR 208.8
Wishram STP 200.8
The Dalles - Bonneville
Dalles/Oregon Cherry OR 189.5
Northwest Aluminum OR 188.9
Cascade Fruit OR 188.3
Lyle 183.2
MosierOR 174.6
SDS Lumber 170.2
BingenSTP 170.2
Hood River OR 168.4
Cascade Locks OR 151.0
Stevenson STP 150.0
Bonneville - Coast
Tanner OR 144.2
North Bonneville STP 144.0
Multnomah Falls OR 134.2
BBA Nonwovens
Washougal 124.0
Exterior Wood, Inc. 123.8
Washougal STP 123.5
CamasSTP 121.2
Georgia Pacific 120.0
Thermal
Load
(MW)
405.821
484.694
0.920
234.905
39.813
0.236
0.488
7.877
8.793
0.875
0.008
0.131
160.323
4.027
0.438
0.381
1.830
1.113
0.508
0.186
0.336
0.295
9.111
24.812
313.206
26
Allocation
Individual
Individual
Individual
Individual
Individual
Individual
Bubble
Bubble
Tot
Bubble
Bubble
Bubble
Bubble
Bubble
Total
Individual
Bubble
Bubble
Bubble
Bubble
Total
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Total
Flow, Q
(cfs)
62.150
62.150
102.410
102.410
0.388
0.388
51.200
51.200
0.780
12.888
12.888
0.084
0.234
2.784
2.228
0.309
0.930
0.046
6.298
16.240
16.240
1.238
0.155
0.135
0.696
Temperature
(deg C)
2.223
0.392
0.193
0.059
0.155
0.077
3.466
9.438
13.780
22.9
Individual 93.230
24.0
22.2
26.7
18.3
32.2
22.2
22.2
22.3
30.6
-------
27
-------
Table 7 (continued). Point sources of thermal energy in the Columbia River
Facility
Toyo Tanso USA OR
Gresham OR
Marine Park
Vancouver Ice & Fuel Oil
Graphic Packaging OR
Northwest Packing Co.
Portland STP OR
Great Western Malting
Vancouver Westside STP
Support Terminal Services
Clark County PUD
Van Alco
Salmon Creek STP
Boise/St Helens OR
Columbia River Carbonates 83.5
Coastal St Helens OR
Clariant Corp
Kalama STP
Noveon Kalama, Inc
Steelscape, Inc.
PGE Trojan OR
Port of Kalama
Riverwood OR
Cowlitz STP
Longview Fiber
Rainier OR
Cytec Industries
Houghton International
Longview STP
Weyerhauser Longview
River
Mile
118.1
117.4
109.5
106.0
105.6
105.2
105.0
105.0
105.0
104.8
103.2
103.0
95.5
85.8
83.5
82.6
76.0
75.0
74.0
73.5
Thermal
Load
(MW)
0.196
106.708
64.431
0.005
31.503
0.348
521.939
36.278
183.024
0.008
5.198
25.321
38.236
219.555
5.898
365.094
5.894
1.627
7.450
1.885
Allocation
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Total
Individual
Individual
Individual
Bubble
Bubble
Bubble
Bubble
Total
Flow, Q
(cfs)
0.071
39.176
24.755
0.002
9.852
0.077
195.881
15.317
71.093
0.002
1.099
7.705
14.544
379.574
52.970
1.547
77.266
1.547
0.619
1.547
0.278
3.992
Temperature
(deg C)
23.4
23.0
22.0
20.0
27.0
30.0
22.5
20.0
21.7
16.0
40.0
27.8
22.2
22.5
35.0
32.2
39.9
32.2
22.2
40.7
57.2
35.7
72.7
511.152
Individual 0.035
22.0
72.2
70.2
68.0
67.4
67.1
67.0
67.0
66.0
64.0
0.081
0.072
109.027
540.993
2.436
3.232
0.008
10.983
398.626
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Total
Individual
0.031
0.025
41.766
116.530
0.979
1.516
0.016
160.864
4.177
73.610
22.2
24.0
22.0
33.0
21.0
22.0
27.0
30.0
22.2
38.9
28
-------
Reynolds 63.0 58.208
Stella STP 56.4 0.014
PGE Beaver OR 53.4 7.026
New Source OR 52.8 24.841
GPWaunaOR 42.3 301.706
Cathlamet STP 32.0 0.549
Astoria OR 11.8 23.383
Ft. Columbia State Park 7.2 0.020
Bell Buoy Crab Co. 6.0 0.329
Warrenton OR 5.0 2.505
llwaco STP 2.0 3.523
Jessies llwaco Fish Co. 2.0 2.748
Coast Guard Sta. 1.0 0.010
Individual
Bubble
Bubble
Bubble
Bubble
Total
Individual
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Bubble
Total
73.610
24.600
0.005
1.695
6.992
33.293
76.27
76.277
0.209
8.227
0.008
0.139
0.881
1.083
1.160
0.003
38.9
20.0
22.2
35.0
30.0
20.0
33.4
22.2
24.0
22.2
20.0
24.0
20.0
16.0
27.5
22.8
29
-------
Table 8. Point sources of thermal energy in the Snake River
Facility
Salmon R - Lower Granite
Asotin STP
Clarkston STP
Potlatch
Lower Granite to Little
Goose
Lower Granite Dam
River
Mile
145.0
138.0
139.3
107.5
Little Goose - Lower Monumental
Little Goose Dam 70.3
Lyon's Ferry 59.1
Lower Monumental - Ice Harbor
Lower Monumental Dam 44.6
Ice Harbor- Columbia R.
Ice Harbor Dam
9.7
Thermal
Load
(MW)
4.016E
6.265E
298.8
0.0194
0.0116
1.381
0.00392
0.00395
Allocation Flow'Q Temperature
Allocation (cfe) ((Jeg Q)
Bubble
Bubble
rota I
1.5626438 21.7
2.0267955 26.1
3.5894393 24.
Individual 75.697228 33.3
Individual 0.0077689 21.1
0.0077689 21.1
Bubble 0.0045907 21.3
Bubble 0.4484799 26.0
Total 0.4530706 26.0
Individual 0.0015472 21.4
Individual 0.0015472 21.5
30
------- |