EPA Fact Sheet: River Basin Model-10

Introduction

The River Basin Model-10 (RBM10) is a one-dimensional mathematical model of the thermal energy
budget of the mainstem Columbia and Snake Rivers. It simulates daily average water temperature under
conditions of gradually varied flow. Similar models of this type have been used since the 1960s to assess
temperature conditions in the Columbia and Snake Rivers.

The technical underpinning of RBM10 has been peer-reviewed, documented, and applied in a number of
settings. Scientists have successfully applied versions of this model framework to rivers in the United
States and abroad, including published studies by researchers at the University of Washington, U.S.
Geological Survey, University of California at Los Angeles, and Wageningen University in the
Netherlands.

Model Scope

The RBM10 model of the Columbia and Snake River mainstems was initially developed and peer-
reviewed by the U.S. EPA in 2001. It has been updated to include more recent data. The geographic area
of the model (see Figure 1) is the Columbia River from the international boundary with Canada (River
Mile 745.0) to the mouth at Astoria, Oregon; the Snake River from Anatone, Washington (Snake River
Mile 168) to its confluence with the Columbia River near Pasco, Washington; and 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). The Clearwater is included in the model domain to account for
substantial cold water releases from Dworshak Dam that strongly influence lower Snake River
temperatures.

Figure 1


-------
The model incorporates the following physical and thermal processes: hydrodynamics within each
model segment (flow, velocity, channel geometry upstream boundary inputs), upstream boundary
inputs (flow, temperature), heat inputs from tributaries, and surface heat exchange within each model
segment.

Building blocks - ID model

Upstream

~ Calibration Data

Figure 2

The tributaries included as flow and heat inputs to the model are shown in red in Figure 3 below. The
model also includes water withdrawal at Grand Coulee dam.



Legend

	I*	' t/'	r	r



	2018 RBM10 Simulated Reaches





Tributaries

Jr ^\ V* I



Withdrawal Tobutafy

\V / fr # ) S\ i idaho I



Canada

f i 1 j i



I United States

_) jivrV Vv _ m j \



Flow Station

) t t \ T W| 3





¦v, i j 1 j J <



USAGE





• US6S

C - ¦ —) \1 ^ >—



• USGS Withdrawal



** WA' HIN ITOW

> V—f

•l' y

< V—1 	3

& v





\ ^X\i ,r' J V \

j %

— T at*** J I

f .Q! / -



* ¦, i —^

t \V. ¦	/

1 f
jp

¦ft/ - r ¦

Figure 3


-------
Using regional weather data, the model simulates the heat exchange processes that occur at the surface
of the rivers and strongly influence water temperature (see Figure 4 below).

solar sc
longwave (direct) (dif

fuse) convection evaporation l

u

l Stream Cross a
Section

_J

Figure 4

The updated RBM10 model simulates temperatures from 1970 through 2016, an unusually long
simulation period for models of this kind. The simplicity of the model, combined with a novel
methodology to solve the mathematical equations, allows for very fast simulation times compared to
most other river temperature models. The RBM10 model can run the fuii forty-seven year simulation of
daily temperatures in less than a minute on a standard laptop.

How Good are the Predictions?

We can check the model simulated temperatures against measured temperatures from the U.S. Army
Corps of Engineers monitoring network (see Figure 5).

Figure 5


-------
Comparing Simulated and Measured River Temperatures

Below are example plots at two locations: Bonneville Dam on the Columbia and Ice Harbor Dam on the
Snake River for the most recent six years (2011-2016).

Bonneville Dam (BON). Columbia River RM 146

2012	2013	2014	2015

—Simulated	• Observed

Figure 6

Ice Harbor Dam (IDSW), Snake River RM 6.8

25

Figure 7

2013	2014	2015

—Simulated	« Obsetved

2016


-------
Long Term Average Conditions

It is difficult to show forty-seven years of results on a single sequential plot like the ones above, but we
can plot the years on top of each other (shown below) and evaluate long term average conditions
(yellow line) and the range of variation (thickness of the band of temperatures). The plot below shows
the thirty-five years from 1970-2004. One can also see specific years where temperatures were
substantially hotter and colder than the norm, such as September 1998 (hot) and August 1985 (cold).

Preliminary RBM10 Results
SimulatedTemperatureat BonnevilleDam (1970-2004)

•	1970 • 1971	• 1972	¦ 1973	• 1974 • 1975 • 1976 • 1977 • 1978 • 1979 • 1980 • 1981
¦ 1982 • 1983 • 1984 1985 • 1986 • 1987 • 1988 • 1989 • 1990 • 1991 ¦ 1992	• 1993

•	1994 • 1995	1996	1997	• 1998	1999 • 2000 • 2001 • 2002 • 2003 • 2004	Average

Figure 8

Impact Assessment

Once we have completed a review of the latest update of the model, we plan to use RBM10 to estimate
changes on river temperatures caused by human activities while accounting for natural variability. For
example, we can change the river channel to represent a free flowing river in the model. We can then
compare "with dams" and "without dams" results to estimate the effect of the dams on river
temperature. We can also use long-term model simulations to estimate the long-term rate of warming
in river temperatures due to climate change. Other impacts include point source inputs and changes to
tributary inflows.

Further Information

The full package of computer programs, supporting data files, and a report on how the model was
developed can be obtained from:

Ben Cope
U.S. EPA Region 10
Modeling Lead
cope.ben@epa.gov
(206) 553-1442


-------