ARIZONA STATE UNIVERSITY
SCHOOL OF SUSTAINABLE ENGINEERING AND THE BUILT ENVIRONMENT

Methodology for Estimating Electricity Generation Vulnerability
to Climate Change
using a Physically-based Modelling System

Matthew Bartos and Mikhail Chester

ASU-SSEBE-CESEM-2014-WPS-002
Working Paper Series
21 July 2014

WORKING PAPER

METHODOLOGY FOR ESTIMATING ELECTRICITY GENERATION
VULNERABILITY TO CLIMATE CHANGE USING A PHYSICALLY-BASED
MODELING SYSTEM
21 JULY 2014

Matthew D. Bartos
Research Engineer
matthew.bartos@asu.edu

Mikhail V. Chester, Ph.D.
Assistant Professor
mchester@asu.edu

Civil, Environmental, and Sustainable Engineering
Arizona State University
www.urbantransitions.org/southwestheat/

A research project “Prioritizing Cooling Infrastructure Investments for Vulnerable
Southwest Populations” (www.urbantransitions.org/southwestheat/) between
Arizona State University and the University of California Los Angeles. Supported
by the National Science Foundation (Award Nos. 1335556 and 1335640).

Table of Contents
1.

Introduction ...................................................................................................................................... 2

2.

Methodology..................................................................................................................................... 3
2.1.

Relating Climate to Power Generation ..................................................................................... 4

2.2.

Estimating Climactic and Hydrological Conditions.................................................................... 8

References .............................................................................................................................................. 13

Page 1 of 15

1. Introduction
In recent years, concerns have grown over the risks posed by climate change on the U.S. electricity
grid.6 The availability of water resources is integral to the production of electric power,16 and droughts
are expected to become more frequent, severe, and longer-lasting over the course of the twenty-first
century.33 The American Southwest, in particular, is expected to experience large deficits in
streamflow.29 Studies on the Colorado River anticipate streamflow declines of 20-45% by 2050.29,30
Other climactic shifts—such as higher water and air temperatures—may also adversely affect power
generation.8,10,12 As extreme weather becomes more common, better methods are needed to assess the
impact of climate change on power generation. This study uses a physically-based modeling system to
assess the vulnerability of power infrastructure in the Southwestern United States at a policy-relevant
scale.
Thermoelectric power—which satisfies a majority of U.S. electricity demand—is vulnerable to
drought. Thermoelectric power represents the backbone of the U.S. power sector, accounting for
roughly 91% of generation. Thermoelectric power also accounts for roughly 39% of all water
withdrawals in the U.S.—roughly equivalent to the amount of water used for agriculture.16 Water use in
power plants is primarily dictated by the needs of the cooling system. During the power generation
process, thermoelectric power plants build up waste heat, which must be discharged in order for the
generation process to continue. Traditionally, water is used for this purpose, because it is safe, plentiful,
and can absorb a large amount of heat. However, when water availability is constrained, power
generation may also be adversely affected. Thermoelectric power plants are particularly susceptible to
changes in streamflow and water temperature. These vulnerabilities are exacerbated by environmental
regulations, which govern both the amount of water withdrawn, and the temperatures of the water
discharged.7 In 2003, extreme drought and heat impaired the generating capacity of more than 30
European nuclear power plants, which were unable to comply with environmental regulations governing
discharge temperatures.32 Similarly, many large base-load thermoelectric facilities in the Southeastern
United States were threatened by a prolonged drought in 2007 and 2008.6 During this period, the
Tennessee Valley Authority (TVA) reduced generation at several facilities, and one major facility was
shut down entirely.34 To meet demand, the TVA was forced to purchase electricity from the grid, causing
electricity prices to rise.6
Although thermoelectric power plants currently produce most of the electric power consumed
in the United States, other sources of power are also vulnerable to changes in climate. Renewables are
largely dependent on natural resources like rain, wind, and sunlight. As the quantity and distribution of
these resources begins to change, renewable generation is also likely to be affected. Hydroelectric dams
represent the largest source of renewable energy currently in use throughout the United States.15 Under
drought conditions, when streamflow attenuates and reservoir levels drop, hydroelectric plants are
unable to operate at normal capacity. In 2001, severe drought in California and the Pacific Northwest
restricted hydroelectric power generation, causing a steep increase in electricity prices.35 Although
blackouts and brownouts were largely avoided, the Northwest Power and Conservation Council
estimated a regional economic impact of roughly $2.5 to $6 billion.36 In addition to hydroelectric power,
Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 2 of 15

it has also been theorized that solar energy resources may also be susceptible to predicted increases in
surface temperature and atmospheric albedo. One study predicts that solar facilities in the
Southwestern U.S. may suffer losses of 2-5%.31
The aim of this study is to estimate the extent to which climate change may impact power
generation in the Southwestern United States. This analysis will focus on the Western Interconnection,
which comprises the states of Washington, Oregon, California, Idaho, Nevada, Utah, Arizona, Colorado,
Wyoming, Montana, South Dakota, New Mexico and Texas. First, climactic and hydrologic parameters
relevant to power generation are identified for five types of generation technologies. A series of
functional relationships are developed such that impacts to power generation can be estimated directly
from changes in certain meteorological and hydrological parameters. Next, climate forcings from the
CMIP3 multi-model ensemble are used as inputs to a physically-based modeling system (consisting of a
hydrological model, an offline routing model, and a one-dimensional stream temperature model). The
modeling system is used to estimate changes in climactic and hydrologic parameters relevant to
electricity generation for various generation technologies. Climactic and hydrologic parameters are then
combined with the functional relationships developed in the first step to estimate impacts to power
generation over the twenty-first century.

2. Methodology
Changes to electric power generating capacity are estimated based on variations in relevant
hydrological and meteorological parameters. These parameters vary depending on the type of power
system under investigation. In this study, six power generation technologies are considered: (1)
hydroelectric power, (2) steam-condensing thermoelectric facilities employing open-loop cooling, (3)
steam condensing thermoelectric facilities employing recirculating cooling, (4) combustion turbine
facilities, (5) solar photovoltaic facilities, and (6) wind turbines. Quantitative relationships between
climactic parameters and power generation are discussed for each technology in Section 1.3.1. Relevant
climactic conditions are determined for the historical period (1949-2010) using daily gridded (1/8degree) observed meteorological forcings.1 Future climactic conditions are then modeled using daily
gridded (1/8-degree) forcings from the CMIP3 multi-model ensemble of downscaled general circulation
model (GCM) outputs.2 A physically-based modeling framework is used to estimate streamflows, water
temperatures and other climactic variables that may constrain power generation (including incoming
shortwave radiation, ambient wet-bulb temperature, and ambient dry-bulb temperature). The modeling
framework consists of (1) the Variable Infiltration Capacity (VIC) macro-scale hydrological model,3 (2) an
offline routing model,4 and (3) a one-dimensional stream temperature model, RBM.5
This section is divided into two parts. The first part of this section (2.1.) identifies how climactic
conditions may affect particular generation technologies, and presents quantitative relationships to help
estimate the effect of climate change on power generation. In the second part of this section (2.2.), the

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 3 of 15

methods for determining future and historical climactic and hydrological conditions are outlined for
each generation technology.
2.1.

Relating Climate to Power Generation

Each power generation technology is vulnerable to different climactic and hydrological
conditions. In this section, adverse climatological conditions are identified for each generation
technology. Quantitative relationships are then developed to relate climactic changes to power
generation impacts. Hydroelectric power is susceptible primarily to decreases in streamflow.6
Thermoelectric power plants are susceptible to a number of climatological and hydrological conditions,
depending on the fuel source and cooling system used (adverse conditions include low streamflows,7
high water temperatures,7,8, 9,10 and high ambient wet/dry bulb temperatures).7, 8, 9, 11 Solar photovoltaic
systems are vulnerable to decreases in shortwave solar radiation and increases in ambient air
temperature.12 Wind turbines are vulnerable to changes in wind speed.13 Climactic effects on particular
technologies are discussed in detail in the following sections.

2.1.1.

Hydroelectric Power

Hydroelectric power harnesses the power of falling water to generate electricity. The generating
capacity of a hydroelectric facility can be described as a function of two hydraulic variables: the flow
rate of the water passing through the turbine and the total hydraulic head acting on the turbine.14
𝑃 = 𝜂𝑡𝑢𝑟𝑏𝑖𝑛𝑒 ∙ 𝜌𝑤 ∙ 𝑄 ∙ 𝑔 ∙ ℎ
Where P is the power output (in Watts), 𝜂𝑡𝑢𝑟𝑏𝑖𝑛𝑒 is the dimensionless turbine efficiency, 𝜌𝑤 is
the density of water (in kilograms per cubic meter), 𝑄 is the flow rate through the turbine (in cubic
meters per second), 𝑔 is the acceleration due to gravity, and ℎ is the net hydraulic head acting on the
turbine (in meters) where the net head is approximately equal to the elevation difference between the
top of the reservoir pool and the elevation of the turbine.14 For this study, it is assumed that the net
head remains relatively constant over the analysis period: in other words, the amount of water stored in
reservoir systems does not accumulate or attenuate significantly over time. This assumption implies that
dam operators will coordinate future releases such that reservoirs operate close to the normal
operating conditions. Keeping the net head constant, the instantaneous power output from a particular
hydroelectric facility can be expressed as the average historical firm power output (𝑃𝑚 ) times the
instantaneous streamflow (𝑄𝑖 ) divided by the average historical streamflow (𝑄𝑚 ):
𝑃 = 𝑃𝑚 ∙

𝑄𝑖
𝑄𝑚

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 4 of 15

2.1.2.

Thermoelectric Power

Thermoelectric power represents the bulk of the power sector, generating 91% of electricity
used in the U.S.15 The majority of thermoelectric power in the U.S. is produced by steam condensing
plants.15 These plants require cooling water in order to reject heat.8,16,17 Access to cooling water can
present an operating constraint for thermoelectric power plants if cooling water is either (1) not
available in sufficient quantity or (2) too hot to meet operational or environmental demands. In the
southern U.S., recent warm and dry weather patterns have caused shortages in cooling water
availability and cooling capacity.18 Cooling water withdrawals and discharge temperatures are often also
constrained by state and federal regulations aimed at preserving riparian ecosystems. Under drought
conditions—when water is scarce and ambient water temperature is high—plants often reduce
generation capacity or shut down operations entirely in order to avoid non-compliance.10
Climatological impacts on thermoelectric power generation depend on the type of cooling
system employed. For power plants using open-loop cooling, usable capacity is constrained by low
streamflow conditions and high water temperatures.7 For power plants using recirculating cooling
systems, ambient wet bulb temperature presents an additional constraint.7,8,9
“Open-loop cooling” (also known as “once-through cooling”) is a relatively simple cooling
scheme in which cold water is removed from the environment (typically a stream), passed through a
heat exchanger (where it absorbs heat from the condenser), then discharged back into the environment
(several degrees warmer than before).9 Because these facilities rely on a constant stream of cool water,
facilities employing open-loop cooling are vulnerable to periods of low streamflow and high inlet water
temperature. To determine the water demand and reduction in capacity for these facilities, the
following approach from van Vliet et al. (2012) is first used to determine the water requirements for
each open-loop facility:7

𝑞=𝑃∙

1 − 𝜂𝑡𝑜𝑡𝑎𝑙
(1 − 𝛼)
∙
𝜂𝑒𝑙𝑒𝑐
𝜌𝑤 ∙ 𝐶𝑝 ∙ max(min((𝑇𝑙𝑚𝑎𝑥 − 𝑇𝑤), ∆𝑇𝑙𝑚𝑎𝑥 ) , 0)

Where 𝑞 is the required water withdrawal of the power plant (cubic meters per second), 𝑃 is the
installed capacity (kW), 𝜂𝑡𝑜𝑡𝑎𝑙 is the net plant efficiency, 𝜂𝑒𝑙𝑒𝑐 is the electrical efficiency, 𝛼 is the share of
waste heat not discharged by cooling water (i.e. flue gas losses), 𝜌𝑤 is the density of liquid water, 𝐶𝑝 is
the heat capacity of water, 𝑇𝑙𝑚𝑎𝑥 is the maximum permissible intake water temperature (℃), 𝑇𝑤 is the
ambient stream temperature (℃), and ∆𝑇𝑙𝑚𝑎𝑥 is the maximum permissible temperature rise of the
water (℃). 𝑇𝑙𝑚𝑎𝑥 and ∆𝑇𝑙𝑚𝑎𝑥 are determined for each thermoelectric power plant using cooling
system data from EIA Form 767.19 Similarly, 𝜂𝑡𝑜𝑡𝑎𝑙 and 𝜂𝑒𝑙𝑒𝑐 are determined using plant and generator
Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 5 of 15

data from EIA Form 767.19 The parameter 𝛼 varies depending on the fuel type, and is roughly 12% for
coal-fired facilities, and 20% for natural gas-fired facilities.9 Knowing the water demand for each facility,
the maximum usable capacity is determined using the following formula:7

𝑃𝑚𝑎𝑥 =

min((𝛾 ∙ 𝑄), 𝑞) ∙ 𝜌𝑤 ∙ 𝐶𝑝 ∙ max(min((𝑇𝑙𝑚𝑎𝑥 − 𝑇𝑤), ∆𝑇𝑙𝑚𝑎𝑥 ) , 0)
1 − 𝜂𝑡𝑜𝑡𝑎𝑙
𝜂𝑒𝑙𝑒𝑐 ∙ 𝜆 ∙ (1 − 𝛼)

Where 𝛾 is the maximum fraction of streamflow available for power generation, 𝑄 is the natural
streamflow, and 𝜆 is a correction factor accounting for changes in efficiencies. The above equation
shows that as 𝑇𝑤 approaches 𝑇𝑙𝑚𝑎𝑥 , once-through cooling systems must withdraw additional water to
maintain the same generating capacity.7 If sufficient additional water is not available, then the usable
capacity of the plant is reduced. Furthermore, if 𝑇𝑤 is greater than or equal to 𝑇𝑙𝑚𝑎𝑥 , plant operation
must be curtailed entirely.7
Recirculating cooling systems reject heat by evaporating water (rather than discharging it
directly into a nearby water body).8,9,11 Water that is not evaporated during the cooling process is reused, meaning that much less water is withdrawn overall. For thermoelectric facilities employing a
recirculating cooling system, the above equations must be modified to account for (1) the effect of
water re-use, and (2) the effects of additional climatological and physical constraints. In recirculating
cooling systems, the ambient wet bulb temperature presents an important design constraint.8,9
Recirculating cooling systems reject heat primarily through latent heat transfer (in other words, by
evaporating water). The ambient wet bulb temperature represents the lowest temperature to which
process water can be cooled using evaporative cooling.11 When the ambient wet bulb temperature is
high, recirculating cooling systems will suffer a performance penalty which can decrease the output of
the plant.8 Accounting for additional constraints of recirculating cooling, the water demand and usable
capacity of recirculating cooling plants can be described using the following equations:7

𝑞=𝑃∙

𝑃𝑚𝑎𝑥 =

(1 − 𝛼) ∙ (1 − 𝛽) ∙ 𝜔 ∙ 𝐸𝑍
1 − 𝜂𝑡𝑜𝑡𝑎𝑙
∙
𝜂𝑒𝑙𝑒𝑐
𝜌𝑤 ∙ 𝐶𝑝 ∙ max(min((𝑇𝑙𝑚𝑎𝑥 − 𝑇𝑤), ∆𝑇𝑙𝑚𝑎𝑥 ) , 0)

min((𝛾 ∙ 𝑄), 𝑞) ∙ 𝜌𝑤 ∙ 𝐶𝑝 ∙ max(min((𝑇𝑙𝑚𝑎𝑥 − 𝑇𝑤), ∆𝑇𝑙𝑚𝑎𝑥 ) , 0)
1 − 𝜂𝑡𝑜𝑡𝑎𝑙
(1 − 𝛼) ∙ (1 − 𝛽) ∙ 𝜔 ∙ 𝐸𝑍
𝜂𝑒𝑙𝑒𝑐 ∙ 𝜆 ∙

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 6 of 15

Where 𝛽 represents the fraction of waste heat released into the air (i.e. sensible heat
transfer), 𝜔 represents a correction factor to adjust for changes in air temperature and humidity, and
𝐸𝑍 is the densification factor that accounts for blowdown due to increases in cooling water salinity.
Thermoelectric facilities employing a combustion turbine generator require little to no water for
cooling. With respect to climactic effects, power generation at combustion turbine facilities is affected
only by the ambient dry bulb temperature of the air.10 In general, as ambient air temperature increases,
the heat rate of the plant increases and the power output decreases.10 However, the relationship
between ambient temperature and plant efficiency depends on individual turbine characteristics. At this
time, it is not possible to determine temperature-efficiency curves for each gas-combustion turbine
serving the WECC region. Thus, combustion turbines are not included in this analysis.
8

2.1.3.

Solar Photovoltaic Power

For solar power facilities, usable generating capacity depends on two climactic variables: net
shortwave radiation and ambient dry-bulb temperature. Net shortwave radiation refers to the amount
of visible light supplied to the solar cell. Solar cell performance also decreases with increasing
temperature “owing to increased internal carrier recombination rates, caused by increased carrier
concentrations”.12 Solar power output is calculated on a daily basis using the following equation: 12

𝑃𝑚𝑝 = 𝑃𝑆𝑇𝐶

𝐽𝑡𝑜𝑡
(1 − 𝛽[𝑇 − 25])
1000

Where 𝑃𝑚𝑝 is the instantaneous power generation, 𝑃𝑆𝑇𝐶 is the power generation under
standard conditions (25 ℃, 1000 W/m2 incident solar radiation), 𝛽 is the power-temperature coefficient
(which varies depending on the technology),12 and 𝑇 is the temperature in Celsius.

2.1.4. Wind Power

Wind power capacity depends primarily on wind velocity. For wind turbines, power generation
can be estimated using the following formula:13
𝑃=

1
𝐶 𝜌𝐴𝑣 3
2 𝑝

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 7 of 15

Where 𝑃 is the generating capacity, 𝐶𝑝 is the maximum power coefficient, 𝜌 is the density of air,
𝐴 is the area swept by the rotor, and 𝑣 is the wind velocity. Assuming that 𝐶𝑝 , 𝜌, and 𝐴 remain constant,
power generation for wind turbines can be expressed using the following equation:
𝑃 = 𝑃𝑚 ∙

𝑣3
3
𝑣𝑚

Where 𝑃𝑚 is the average historical generating capacity, and 𝑣𝑚 is the average historical wind
velocity.

2.2.

Estimating Climactic and Hydrological Conditions

A physically-based modeling framework is used to estimate climactic and hydrological
conditions that may impact power generation. Six climactic and hydrological conditions are needed to
assess impacts on power generation. These climactic variables are listed below, along with the power
generation technologies they affect:
1)
2)
3)
4)
5)
6)

Streamflow (hydropower and thermoelectric power)
Stream temperature (thermoelectric power)
Ambient wet-bulb temperature (thermoelectric power with recirculating cooling)
Net shortwave radiation (solar photovoltaic power)
Ambient dry-bulb temperature (combustion turbine and solar photovoltaic power)
Wind speed (wind turbines)

Each of the six climactic and hydrological conditions listed above are modeled at a daily time
step for both the historical period (1949-2010) and the future period (2010-2100) using 1/8-degree
gridded data. Streamflows are modeled using the Variable Infiltration Capacity (VIC) macro-scale
hydrological model,3 along with an offline routing model.4 To estimate stream temperatures, the
previous two models are joined with a one-dimensional stream-temperature model (RBM).5 Net
shortwave radiation and ambient wet-bulb temperature are obtained by running the VIC model in fullenergy mode. Ambient dry-bulb temperature and wind speed are obtained from input data to the VIC
model. Historical climate inputs to the VIC model are derived from gridded observed meteorological
data.1 Future climate inputs are obtained from the CMIP3 multi-model ensemble, using downscaled
outputs from the ukmo-hadcm3.1 and mpi-echam5.3 GCM models.2 Emissions scenarios include the
SRES a1b, a2 and b1 scenarios.
The VIC model works by dividing the study area into discrete grid cells (typically 1/8-degree) and
performing water and energy balances at each cell. Using the water balance mode of the VIC model,
daily estimates of runoff and baseflow are obtained for each grid cell. Gridded runoff and baseflow are
then fed into a routing model, which produces estimates of streamflow by routing these inputs through
a stream network at the same resolution of the VIC hydrological model. The RBM model estimates

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 8 of 15

stream temperature by solving the one-dimensional heat advection equation using a mixed EulerLagrangian approach.7 Stream temperatures at each stream segment are calculated based on three
characteristics: “upstream water temperature and inflow into the stream segment, the dominant
heat exchange at the air-water surface, and the inflow and temperature of water advected from
tributaries and (anthropogenic) point sources of heat”.7
Two types of input files are required to run the VIC model: (1) meteorological forcing files and
(2) land cover parameter files. Meteorological forcing files contain meteorological characteristics at a
specified time step (typically daily) for each grid cell. VIC requires a minimum of four meteorological
forcing inputs: precipitation, daily maximum temperature, daily minimum temperature, and wind speed.
For the historical period, daily observed values for these four inputs are obtained at a 1/8-degree spatial
resolution.1 For the future period, daily meteorological inputs for the VIC model are obtained using
downscaled General Circulation Model (GCM) outputs from the CMIP3 multi-model ensemble.2 Two
representative GCMs were chosen: ukmo-hadcm3.1 and mpi-echam5.3. These GCMs were chosen
because they have performed favorably compared to many other GCM models in previous
assessments.20 GCM model output was obtained for three SRES scenarios: a1b, a2, and b1. The CMIP3
multi-model ensemble was chosen over the newer CMIP5 ensemble because CMIP5 currently does not
contain all the meteorological forcings needed to run the VIC model. In addition to meteorological
forcings, the VIC model also requires land cover parameter files. Land cover parameter files contain land
cover characteristics needed to perform water and energy balances under the VIC model (such as the
saturated hydraulic conductivity of the soil). These land cover characteristics are divided into three
categories: soil characteristics, vegetation characteristics, and snowband (elevation band)
characteristics. For this study, 1/8-degree gridded land cover files (soil, vegetation, and snowband) are
obtained using NASA’s National Land Data Assimilation Systems (NLDAS) datasets.21
The offline routing model produces estimates of streamflow using gridded runoff and baseflow
outputs from the VIC hydrological model. In addition to the VIC runoff and baseflow outputs, the routing
model requires a flow direction file at the same resolution as the VIC output files. For this purpose, 1/8degree upscaled flow direction networks are used.22
The RBM model combines the routing model with a heat budget model in order to calculate
stream temperatures in the routing network. In addition to gridded runoff and baseflow inputs, the RBM
model requires the following forcings to perform heat budget calculations: air temperature, vapor
pressure, incoming shortwave radiation, incoming longwave radiation, near-surface atmospheric
density, near surface atmospheric pressure, and near-surface wind speed. These forcings are obtained
at a daily timestep by executing the VIC hydrological model in full-energy mode. The RBM model
includes a modified version of the offline routing model. The modified version of the routing model
(rout-DA) includes subroutines that calculate the hydraulic geometry (the width and depth) of the
stream for a given streamflow value, using the methods of Leopold and Maddock.23 Hydraulic geometry
parameters (sometimes referred to as Leopold coefficients) are required as inputs for this part of the
RBM model. These parameters describe how the width and depth of a stream change with respect to
streamflow, and are determined empirically using power regressions on historical streamflow and river
geometry data. For this study, Leopold coefficients are obtained using empirical observations from 674
Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 9 of 15

gauging stations throughout the U.S., as collected by Allen et al.24 The boundary conditions for the RBM
temperature model (headwater temperatures) are determined using a water temperature regression
model used by Mohseni.25 Mohseni’s regression model takes the form of the following logistic function:
𝑇𝑠 = 𝜇 +

𝛼−𝜇
1 + 𝑒 𝛾(𝛽−𝑇)

Where 𝑇𝑠 is the modeled water temperature, 𝜇 is the minimum water temperature, 𝛼 is the
maximum water temperature, 𝛾 is the steepest slope of the function, 𝛽 is the air temperature at the
inflection point of the curve, and 𝑇 is the measured air temperature. Maximum and minimum
temperatures (𝛼 and 𝜇, respectively) are determined for relevant streams, using historical water
temperature data from USGS. A Levenberg-Marquardt curve-fitting algorithm is then used to fit the
additional parameters to the historical data.

The general process for determining relevant climactic characteristics at power generating
stations is outlined below:
a) Power generating stations in the WECC region are identified and mapped by latitude-longitude
coordinate using eGRID coordinate data.26
b) Watersheds containing hydroelectric and thermoelectric powerplants in the WECC region are
delineated using a 15-second resolution digital elevation raster.27
c) The watersheds delineated in step (b) are upscaled to a 1/8 degree spatial resolution by
intersecting the watershed’s contributing area with a 1/8 degree grid.
d) A 1/8-degree flow direction raster is clipped to the delineated basin to produce the 1/8-degree
flow direction file required by the routing model.22
e) To account for the partial contribution of boundary cells, a flow fraction raster is generated. The
flow fraction raster consists of a gridded ascii data file containing the fraction of runoff
contributed by each 1/8-degree grid cell. The flow fraction for each 1/8 degree grid cell is equal
to the internal area of the 15-second resolution contributing area divided by the total area of
the 1/8 degree grid cell.
f) The Variable Infiltration-Capacity (VIC) hydrological model is executed in water-balance mode to
generate runoff and baseflow data for each 1/8-degree grid cell.
g) The offline routing model is used along with the output from the VIC hydrological model to
produce estimates of streamflow at each generating station.
h) For thermoelectric generating stations, the VIC model is executed again in full-energy mode to
produce the following outputs: air temperature, vapor pressure, incoming shortwave radiation,

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 10 of 15

incoming longwave radiation, near-surface atmospheric density, near surface atmospheric
pressure, and near-surface wind speed. These parameters are used as inputs to the RBM model.
i) The RBM model is excecuted using the forcings generated in (f) and (h) to produce estimates of
stream temperature at each stream segment.
j) For solar photovoltaic facilities, the VIC model is executed in full energy mode to obtain
incoming shortwave radiation at a daily time step for each grid cell containing solar photovoltaic
facilities.
k) For wind turbines, near-surface wind speed inputs are extracted for all grid cells containing wind
turbines.
Steps (a) through (k) above are automated using a series of helper scripts. These scripts are
produced by the author and are available online.28 A full workflow schematic—including all inputs,
outputs, executables and helper scripts—is shown in Figure 1.

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 11 of 15

regional
netcdf forcing

Delineate Basin
in ArcGIS

colo_basins.p
unzip_raw

ca_sr_basins.p
nc_to_vic
region_latlon.p

nc_to_vic_ca
basin dbf
copy_forcings

dir ascii
latlon_stns
fix_columns

historical ascii
forcing

frac ascii

vic_dir_converter

vic_frac_converter

param_clip

hydrostn.p

control files

forcing input

param input

dir_frac_clip

Legend

executable

basin frac

rout

vicNl

basin_dir

text data

binary data

station_locs

RBM
rout_post_all

python script

Figure 1. Workflow for pre- and post-processing VIC, rout, and RBM inputs/outputs.

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 12 of 15

References
1. Maurer, E.P., A.W. Wood, J.C. Adam, D.P. Lettenmaier, and B. Nijssen, 2002, A Long-Term
Hydrologically-Based Data Set of Land Surface Fluxes and States for the Conterminous United
States, J. Climate 15, 3237-3251.
2. Reclamation, 2013. 'Downscaled CMIP3 and CMIP5 Climate and Hydrology Projections: Release
of Downscaled CMIP5 Climate Projections, Comparison with preceding Information, and
Summary of User Needs', prepared by the U.S. Department of the Interior, Bureau of
Reclamation, Technical Services Center, Denver, Colorado. 47pp.
3. Liang, X., Lettenmaier, D. P., Wood, E. F. & Burges, S. J. A Simple Hydrologically Based Model of
Land-Surface Water and Energy Fluxes for General-Circulation Models. Journal of
Geophysical Research-Atmospheres 99, 14415-14428 (1994).
4. Lohmann, D., Raschke, E., Nijssen, B. & Lettenmaier, D. P. Regional scale hydrology: I.
Formulation of the VIC-2L model coupled to a routing model. Hydrological Sciences Journal 43,
131-141 (1998).
5. Yearsley, J. R. A semi-Lagrangian water temperature model for advection-dominated river
systems. Water Resources Research 45, W12405 (2009).
6. Harto, C.B. and Yan, Y.E., “Analysis of Drought Impacts on Electricity Production in the Western
and Texas Interconnections of the United States,” Environmental Science Division, Argonne
National Laboratory. December 2011.
7. Van Vliet, M.T.H., Yearsley, J.R., Ludwig, F., Vögele, S., Lettenmaier, D.P., Kabat, P. Vulnerability
of US and European electricity supply to climate change. Nature Climate Change 2,676–681
(2012). doi:10.1038/nclimate1546
8. Comparison of Alternate Cooling Technologies for California Power Plants: Economic,
Environmental, and Other Tradeoffs; EPRI: Palo Alto, CA and California Energy Commission:
Sacramento, CA, 2002.
9. Rutberg, M. J. Modeling Water Use at Thermoelectric Power Plants. MS Thesis. Massachusetts
Institute of Technology, State College, MA. 2003.
10. Colman, J. The Effect of Ambient Air and Water Temperature on Thermoelectric Power
Efficiency. 2013. MS Thesis. Duke University, State College, NC. 2013.
11. Bhatia, A. Cooling Towers. Course No: M07-001. CED Engineering, Continuing Education and
Development.
12. Dubey, S., Sarvaiya J.N., Seshadri, B. Temperature Dependent Photovoltaic (PV) Efficiency and Its
Effect on PV Production in the World—A Review. Energy Procedia, 33 (2013) 311 – 321.
13. Royal Academy of Engineering. Wind Turbine Power Calculations.
https://www.raeng.org.uk/education/diploma/maths/pdf/exemplars_advanced/23_Wind_Turbi
ne.pdf
14. Mays, Larry W. Water Resources Engineering. Hoboken, NJ: John Wiley, 2011. Print.
15. EIA 2012. Annual Energy Review. Released September 2012. Online access:
http://www.eia.gov/totalenergy/data/annual/showtext.cfm?t=ptb0709.

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 13 of 15

16. Energy Demands on Water Resources: Report to Congress on the Interdependency of Energy
and Water; U.S. Department of Energy: Washington, DC, 2006.
17. Averyt, K., Fisher, J.; A. Huber-Lee, Lewis, A.; Macknick, J.;Madden, N.; Rogers, J.; , and S.
Tellinghuisen. Freshwater Use by U.S. Power Plants: Electricity’S Thirst for a Precious Resource,
A report of the Energy and Water in a Warming World Initiative; Union of Concerned Scientists:
Cambridge, MA, November 2011.
18. NETL 2009. Impact of Drought on US Steam Electric Power Plant Cooling Water Intakes and
Related Water Resource Management Issues (National Energy Technology Laboratory, 2009).
19. U.S. Energy Information Administration. EIA Form 767. Detailed Data, Year 2005.
http://www.eia.gov/electricity/data/eia767/ (accessed June 2014).
20. Covey, C., AchutaRao, K.M., Cubasch, U., Jones, P., Lambert, S.J., Mann, M.E., Phillips, T.J.,
Taylor, K.E. 2003. An Overview of Results from the Coupled Model Intercomparison Project
(CMIP). Global and Planetary Change. 37, 103-133.
21. Mitchell, K.E., D. Lohmann, P.R. Houser, E.F. Wood, J.C. Schaake, A. Robock, B.A. Cosgrove, J.
Sheffield, Q. Duan, L. Luo, R.W. Higgins, R.T. Pinker, J.D. Tarpley, D.P. Lettenmaier, C.H. Marshall,
J.K. Entin, M. Pan, W. Shi, V. Koren, J. Meng, B. H. Ramsay, and A.A. Bailey (2004), The multiinstitution North American Land Data Assimilation System (NLDAS): Utilizing multiple GCIP
products and partners in a continental distributed hydrological modeling system, J. Geophys.
Res., 109, D07S90, doi:10.1029/2003JD003823.
22. Wu, H., Kimball J. S., Mantua N., and Stanford J. A. Automated Upscaling of River Networks for
Macroscale Hydrological Modeling. Water Resources Research. 47. 2011.
http://secure.ntsg.umt.edu/publications/2011/WKMS11
23. Leopold, L. B. & Maddock., T. The hydraulic geometry of stream channels and some
physiographic implications. U.S. Geological Survey Professional Paper 252. U.S. Geological
Survey (1953).
24. Allen, P. M., Arnold, J. G. & Bruce, W. B. Downstream channel geometry for use in planning level
models. Water resources bulletin 30, 663-671 (1994).
25. Mohseni, O., Stefan, H. G. & Erickson, T. R. A nonlinear regression model for weekly stream
temperatures. Water Resources Research 34, 2685-2692 (1998).
26. Environmental Protection Agency. Emissions & Generation Resource Integrated Database, Year
2007−2009 Data Files. http://www.epa.gov/cleanenergy/energy-resources/egrid/index.html.
27. Lehner, B., Verdin, K., Jarvis, A. (2008): New global hydrography derived from spaceborne
elevation data. Eos, Transactions, AGU, 89(10): 93-94.
28. Bartos, M. Github Repository: vic_utils. https://github.com/mdbartos/vic_utils
29. Milly PCD, Dunne KA, Vecchia AV (2005) Global pattern of trends in streamflow and water
availability in a changing climate. Nature 438: 347-350
30. Hoerling, M., and J. Eischeid, 2007: Past peak water in the West. Southwest Hydrology, Vol 6,
No. 1, 18-19.
31. Crooks, J.A., Jones, L.A., Forster, P.M., Crook, R. Climate change impacts on future photovoltaic
and concentrated solar power energy output. Energy Environ. Sci., 2011,4, 3101-3109. DOI:
10.1039/C1EE01495A

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 14 of 15

32. IAEA (2004): Operating Experience with Nuclear Power Stations in Member States in 2003,
Vienna.
33. IPCC (2008): Climate Change 2007: Impacts, Adaptation and Vulnerability, Cambridge
University Press, Cambridge.
34. Union of Concerned Scientists, 2007, “Rising Temperatures Undermine Nuclear Power’s
Promise,” Union of Concerned Scientists backgrounder, Washington, DC, August 23. Available at
http://www.nirs.org/climate/background/ucsrisingtemps82307.pdf.
35. Bonneville Power Administration (BPA), 2002, “Guide to Tools and Principles for a Dry Year
Strategy,” available at: http://www.bpa.gov/power/pgp/dryyear/08-2002_Draft_Guide.pdf,
August.
36. Northwest Power and Conservation Council, 2005, “The Fifth Northwest Electric Power and
Conservation Plan,” available at:
http://www.nwcouncil.org/energy/powerplan/5/Volume1_screen.pdf.

Methodology for Estimating Electricity Generation Vulnerability to Climate Change using a Physically-based Modeling System

M Bartos and M Chester | Arizona State University | www.urbantransitions.org/southwestheat/

Page 15 of 15