Analysis of Web-Based Solar Photovoltaic Mapping Tools

Abstract

As the demand for renewable energy has grown, so too has the need to quantify the potential for these resources. Understanding the potential for a particular energy source can help inform policy decisions, educate consumers, drive technological development, increase manufacturing capacity, and improve marketing methods. In response to the desire to better understand the potential of clean energy technologies, several approaches have been developed to help inform decisions. One technology-specific example is the use of solar photovoltaic (PV) maps. A solar PV mapping tool visually represents a specific site and calculates PV system size and projected electricity production. This paper identifies the commercially available solar mapping tools and provides a thorough summary of the source data type and resolution, the visualization software program being used, user inputs, calculation methodology and algorithms, map outputs, and development costs for each map.

Full text

The United States Government retains, and by accepting the article for publication, the publisher acknowledges that the United States Government retains, a non-
exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for United States Government
purposes. 1
Proceedings of the 3rd International Conference on Energy Sustainability
ES2009
July 19-23, 2009, San Francisco, California, USA
ES2009-90461
ANALYSIS OF WEB-BASED SOLAR PHOTOVOLTAIC MAPPING TOOLS
Jesse Dean
National Renewable Energy Lab
Golden, CO, USA
Alicen Kandt
National Renewable Energy Lab
Golden, CO, USA
Kari Burman
National Renewable Energy Lab
Golden, CO, USA
Lars Lisell
National Renewable Energy Lab
Golden, CO, USA
Christopher Helm
National Renewable Energy Lab
Golden, CO, USA
ABSTRACT
As the demand for renewable energy has grown, so too has
the need to quantify the potential for these resources.
Understanding the potential for a particular energy source can
help inform policy decisions, educate consumers, drive
technological development, increase manufacturing capacity,
and improve marketing methods. In response to the desire to
better understand the potential of clean energy technologies,
several approaches have been developed to help inform
decisions. One technology-specific example is the use of solar
photovoltaic (PV) maps.
A solar PV mapping tool visually represents a specific site
and calculates PV system size and projected electricity
production. This paper identifies the commercially available
solar mapping tools and provides a thorough summary of the
source data type and resolution, the visualization software
program being used, user inputs, calculation methodology and
algorithms, map outputs, and development costs for each map.
NOMENCLATURE
ηa = Efficiency of the PV array
η0 = Measured efficiency at the reference cell temperature
β = Rate of change of efficiency with respect to Tc
Tc = Calculated cell temperature
Tr = Reference cell temperature
Pdc = Direct current power
POA = Plane of Array irradiance, W/m2
ηpr = Efficiency of the power conversion unit
F = Fraction of total rated load
ηp = Actual efficiency of the power conversion unit
ηRL = Efficiency of the power conversion unit at full load
INTRODUCTION
Visual, web-based solar (PV) mapping products are
increasing in prevalence. These tools quantify the potential for
solar PV at a specific location to educate the user about the
benefits of solar PV and its associated costs and savings. In an
effort to inform city officials, as well as the general public, this
paper details the layers of information that are used in solar
mapping applications and outlines the commercially available
solar mapping tools. Finally, the paper summarizes the results
of a comparative analysis between the tools and outlines
potential improvements that could be made to the current solar
maps.
This paper serves as a valuable resource for municipalities
and developers evaluating various software tools to increase the
installed capacity of solar within a given city or region.
Most of these tools are being developed as a part of the
U.S. Department of Energy’s (DOE) Solar America Initiative
(SAI). This initiative aims to make solar electricity from PV
cost competitive with conventional forms of electricity from the
utility grid by 2015 through R&D and market transformation.
Many of the 25 Solar America Cities, part of the SAI, are
pursing solar mapping to educate their populaces.
These maps empower a resident, business owner, or
decision maker to take the first step in analyzing the potential
for solar PV at a particular location.

2
LAYERS OF INFORMATION
Web-based solar PV mapping tools contain three levels of
input data that are used to estimate the performance of a PV
array at a given location. The first level is topographical data
associated with a given location or city. Some of the maps use
three-dimensional digital elevation models (DEMs) to analyze
the impacts of shading obstructions, identify roof tilt, and
estimate the amount of roof area that can be used for a
particular installation. Some of the simplified maps skip this
step and do not take into account local topographical
interactions associated with shading, roof tilt or orientation. The
user is then responsible for defining roof area, tilt, azimuth
angle, and an appropriate derate factor to account for the
impacts of any shading obstructions.
The second layer consists of the meteorological data that
are used to estimate the solar resource at a given site. Some
maps make simplifying assumptions to calculate an annual
solar resource estimate; others use hourly meteorological data
that are derived from ground-based meteorological stations or
satellite-derived meteorological data.
The third layer consists of the financial and incentive data
that is used to calculate the economics associated with a given
installation. Some tools have predefined financial and incentive
data built into the model, some of which cannot be changed.
The financial and incentive data typically consist of:
Electricity rate ($/kilowatt-hour [kWh])
Electricity escalation rate (%/yr)
Installed cost ($)
Federal tax credit ($)
State, local, and utility incentives
The three layers of input data are then processed to provide
an estimate of system size, electricity production, installed cost,
and various levels of financial and environmental data. Some of
the solar maps have additional features that serve as an all-
encompassing source of renewable energy information for
consumers in a given city. They will link consumers to local
installers, provide information about how to capture local
incentives, and provide educational information associated with
the given technology. Some maps are also used to track the
total number of PV installations within a given city, which
helps the city understand how well it is meeting its solar
installation goals.
SOLAR RESOURCE DATA
Similar to localized weather patterns, solar radiation
characteristics vary with geographic location and time. A
significant amount of work has gone into the development of
standardized tools and models that can be used to understand
the spatial and temporal variations in solar radiation. In terms
of collection techniques, solar resource data can be collected
from ground-based meteorological stations or derived from
satellites.
A key requirement of any solar PV mapping tool is its
ability to accurately calculate the spatial and time-dependent
characteristics of the solar resource at a given location. The
National Renewable Energy Laboratory (NREL) and the
National Climatic Data Center were among the first to develop
a set of standard solar resource models through the
development of the National Solar Radiation Data Base
(NSRDB) [1]. The database used meteorological and cloud
cover observations at National Weather Service stations around
the country as inputs into models to simulate the solar resource
at a site. The database, published in the early 1990s, contains
solar resource estimates for 239 stations within the United
States between 1961 and 1990 [1]. Of the 239 stations, 56 are
primary stations and used some ground-based solar
measurements; the remaining 183 stations used only modeled
solar radiation data derived from meteorological data including
cloud cover observations. The datasets contain 8,760 hourly
records selected from the NSRDB to represent a typical single
meteorological year (TMY) at a given location. The NSRDB
provides solar analysts, designers, building architects, and
countless others with all the solar radiation information needed
to analyze the resource available for solar PV systems.
TMY2
TMY datasets are derived from the 1961–1990 NSRDB.
The designation of TMY2 was given to differentiate the dataset
from earlier datasets derived between 1952 and 1975 from the
SOLMET/ERSATZ database [2]. The TMY2 datasets provide
hourly values of solar radiation and meteorological data for a
TMY at a given location. The datasets are intended to be used
in computer simulations of solar energy conversion systems.
The hourly values are intended to be average values and are not
suited for worst-case design condition analysis. The typical
values for a given month at a specific location are taken by
examining all 30 years of weather data in a specific month; the
one judged most typical is selected for use in the TMY dataset.
The other months are selected in a similar fashion. The 12
selected typical months for a given location were chosen based
on the following parameters: global horizontal radiation, direct
normal radiation, dry bulb temperature, dew point temperature,
and wind speed [2].
TMY3
The TMY3 dataset was created based on updated weather
data from the NSRDB between 1991 and 2005. It was created
with recent data from the 239 historic ground-based
meteorological sites used in the TMY2 dataset and a number of
additional sites. The TMY3 dataset currently includes data from
1,454 weather stations [3]. A number of improvements were
made to the TMY3 dataset, including a significant increase in
the number of sites. The solar radiation data in the TMY3
dataset include satellite-modeled data for 1998 to 2005 and
surface-modeled data for earlier years. The satellite-modeled
hourly solar data are also available for any location on a 10-
kilometer (km) grid. These data sets were created by the

3
Atmospheric Sciences Research Center at the State University
of New York – Albany (SUNY) for 1998 to 2005.
Satellite-Derived Weather Data
In many locations throughout the United States, the
absence of ground-based meteorological stations has led to the
development of modeled weather data from geostationary
satellites (GEOS). Currently three GEOS satellites monitor the
Western Hemisphere, one at 75 degrees west longitude to
monitor the East, one at 60 degrees west longitude for South
America support, and a third at 135 degrees west longitude over
the Pacific Ocean. The satellites are positioned at an exact
height above the earth so they orbit around the Earth at the
same speed that the Earth rotates around its axis. This results in
stationary positioning relative to the Earth. The satellites are
used to continuously monitor the atmospheric characteristics,
including cloud coverage, of a location [4]. The satellites have
a resolution approaching 1 km in the visible irradiance range
[4]. Satellite-derived solar resource estimates are the most
accurate form of solar data beyond 25 km from the closest
ground-based station. The ability to accurately characterize
solar microclimates becomes important when analyzing solar
energy systems at locations with no nearby ground-based solar
measurement station. The models that process these data
provide hourly estimates of global horizontal, direct normal,
and diffuse horizontal irradiance levels.
Data Resolution and Accuracy
The satellite-derived weather data discussed previously
have a mean basis error of only 2% to 5%, when compared to
ground-based meteorological stations [6]. This high level of
accuracy validates the sophistication of the algorithms that
calculate the hourly estimates of solar irradiance. Based on
these results, solar measurements from ground-based
meteorological stations would provide the most accurate
representation of solar irradiance, and satellite-derived solar
estimates would provide the most accurate representation of
solar irradiance when the closest ground-based weather station
is more than 25 km from the location being analyzed.
Regardless of the solar radiation data source being used,
TMY2, TMY3, and satellite-derived solar data taken at a 10-km
grid should provide a similar characterization of solar radiation
at a given site.
ELEVATION AND SURFACE MODEL DATA
Most of the web-based solar PV mapping tools discussed
here incorporate topographical elevation data in a city to
analyze the solar potential of building rooftops. In lieu of using
a solar pathfinder to analyze every rooftop within a city, this is
one of the most accurate ways of identifying the rooftop solar
potential. A light detection and ranging technique or stereo pair
imagery is used to create three-dimensional maps of the city.
Light Detection and Ranging
Light detection and ranging (LIDAR) technology uses
laser pulses to measure elevation at a remote site and produces
a three-dimensional elevation image file. The distance to an
object is measured from the time delay between the pulse that is
transmitted and the reflected signal. LIDAR technology is
similar to radar; however, it uses light from laser pulses rather
than radio waves.
Methodology
The data are collected from a LIDAR laser scanner
mounted on the bottom of an aircraft. The scanning system
requires a ground-base location determined from the global
positioning system (GIS) associated with the plane. The plane
generally travels at 60 meters per second and records
measurements at a rate of 2,000 to 5,000 pulses per second [7].
The datasets contain vast amounts of information and may have
as many as 350,000 points per square mile, depending on the
area and density of vegetation. The scan area covers
approximately 300 meters in width from an altitude of around
600 meters.
The time delay of the reflectance data depends on the
distance to the surface and the type of surface that is reflected.
The percentage that is reflected is known as the LIDAR
intensity data. Light can reflect off of metal and nonmetal
objects such as snow or leaves. Thus, the datasets contain
discernible features such as trees, buildings, and power lines.
This technology may be used to scan the elevation for x, y, z
coordinates and distinguish from the intensity of the reflections
whether the object is a building or a tree. The LIDAR scan can
be done any time of the day or night as long as the sky is clear.
LIDAR datasets can be straightforward to interpret, as
from the beginning of the data collection scan the information
is referenced with the GIS system and thus can interface with
other GIS applications. The laser beam will detect the tree and
building canopy and will detect through the foliage and reach
the ground. The scanning system collects the first and last
returns or reflections. The first returns are the reflection off the
highest points; the last returns are the reflection off the lowest
point, which is generally the ground level. Steep terrain and
areas that are often inaccessible are captured by the datasets.
This makes LIDAR well suited for accurate DEMs.
Resolution
The laser scanner uses a very narrow beam that allows very
high-resolution elevation mapping of terrain. The LIDAR uses
short wavelengths in the ultraviolet, visible, or near infrared
areas of the electromagnetic spectrum. Images created from a
reflective scanning technique can generally capture only objects
at the same size or larger than the wavelength used. Because
LIDAR uses wavelengths that range from 10 micrometers to
250 nanometers, the waves reflected can detect extremely small

4
objects. The vertical precision from a LIDAR scan is 15 cm (6
inches) [8].
Stereo Pair Imagery
Stereo pair imagery consists of two photographic images of
a single location taken from two offset vantage points. The
imagery can be taken from satellite-based cameras or from
cameras mounted to the bottom of an aircraft. Sequential
photographs need to be taken along common flight lines that
overlap by at least 60% [9]. The accuracy of the final product
is directly tied to the resolution of the original stereo pair
imagery. Once the imagery is collected, it is radiometrically and
geometrically corrected to create a three-dimensional image of
the city. Exact contour lines of buildings and objects are
acquired from vector geodata and are used to create a three-
dimensional digital elevation map. The imagery can then be
used to extract elevation and contour data needed to analyze
impacts of shading, orientation, and slope. The primary
advantage of stereo pair imagery is that it can capture
geographical characteristics of man-made and naturally
occurring structures.
CALCULATION / ALGORITHMS
PVWatts
PVWatts is an online calculation tool used for estimating
the output of grid connected PV systems. The tool was
developed by NREL and is used to estimate the electricity
produced from a crystalline silicon PV array at any of the 239
locations listed in the TMY2 dataset. PVWatts Version 1
(PVWatts V.1) uses a set of internal calculation algorithms
originally developed by Sandia National Laboratories called
PVFORM. The PVFORM calculation module is built from a
series of individual calculation modules. Each module is
configured according to the following equations [10]:
PV array efficiency:
(1)
Direct current (DC) power model:
(2)
The Perez anisotropic diffuse radiation model is used to
compute the POA irradiance [11].
Alternating current (AC) power conversion model:
Power conversion unit (PCU) efficiency
(3)
(4)
PVWatts then uses a set of predefined inputs to populate the
program with the rest of the data needed to run the calculation
algorithm:
Location (state and city)
Electricity rate ($/kWh)
DC size (kilowatts [kW])
Derate factor
Tilt angle (degrees)
Azimuth angle (degrees)
PVWatts V.1 is one of the most widely used PV system
calculation tools in the United States. PVWatts Version 2,
(PVWatts V.2) uses the same calculation algorithms as
PVWatts V.1 with a few corrections associated with the use of
40-km resolution solar resource data. In My Backyard (IMBY)
uses PVWatts V.2 to calculate the performance of a given PV
array [12].
SOLAR AUTOMATED FEATURE EXTRACTION™
CH2M Hill developed the Solar Automated Feature
Extraction (S.A.F.E.)TM methodology to quantify roof area
exposed to year-round solar radiation for specified locations. To
calculate this area, this technique uses aerial imagery, either
LIDAR or other two-dimensional stereo pair images, to build
three-dimensional models. It uses an integrated time-series
analysis that combines individual snapshots of the shadows cast
from the three-dimensional model at a point in time. These
images are combined into an annual shade-free image used to
compute the rooftop area that does not receive shade
throughout the year. This methodology can account for shading
that is attributable to chimneys, air-conditioning units, or other
structures, as well as the slope and orientation of the roof. The
process does not currently account for shading from trees, but
the inclusion of vegetation in the shade simulations is currently
under development. The output from this analysis is the shade-
free area on a rooftop. This information is presented through a
Web mapping portal that enables users to enter an address to
retrieve the data about shade-free area on their rooftops.
ESRI ArcGIS Solar Analyst Module
The Solar Analysis Tools of ArcGIS, which were
introduced in ESRI’s ArcGIS version 9.2, calculate solar
insolation (W-h/m2) at a location on the Earth’s surface.
Insolation maps are calculated with inputs from DEMs. This
tool uses point-based imagery of local level elevation, slope,
and aspect to determine the amount of energy available.
Optimized algorithms account for variations in surface
orientation and atmospheric weather data.
Total global radiation (Globaltot) is calculated from the sum
of the direct and diffuse radiation of all sectors on the
topographic surface. These are calculated separately for each
location and the total produces an insolation map for the whole
study area. Detailed models and algorithms used to calculate

5
the direct and diffuse solar radiation can be found in the Solar
Analyst design document [13]. The outputs from the Solar
Analysis Tools include a map of direct, diffuse, and global
radiation along with direct radiation duration. The tool also
calculates sky maps and horizontal angles for specific cells over
the entire DEM.
IN MY BACKYARD
Tool Overview
In My Backyard (IMBY) is a Web-based solar simulation
tool, and is meant to introduce homeowners to the possible
benefits of renewable energy. The main purpose of IMBY is to
provide an easy-to-use interface to estimate the hour-by-hour
amount of electricity produced by a PV system over a year.
IMBY provides a map-based interface and allows a user to
specify an address at which to place a PV system. The user
interface for IMBY is shown in Figure 1.
FIGURE 1 – IMBY WEB BASED INTERFACE
The map centers itself on that address and the user may draw a
potential PV system anywhere on the map. An example PV
rendering is shown in Figure 2.
FIGURE 2 – IMBY BUILDING RENDERING WITH PV ARRAY
After the user has drawn a system, several default values
are used to populate information about the PV system’s
configuration. These values are the size, derate, tilt, and
azimuth of the PV system. The size represents the DC rating of
system, the derate is the amount of energy lost in the
conversion from DC to AC, the tilt represents the angle at
which the system is to be tilted (this defaults to the latitude of
the user ’s location to maximize output), and the azimuth is the
primary direction that the system is facing (this is a range of 0
to 360 where both 0 and 360 equal north).
The user then selects the data year. This is the year of
resource data used to drive the simulation of the system’s
output. After the user has reviewed the inputs and made
changes, the simulation may be performed. When the
simulation is a complete, the user sees a summary window that
shows a monthly breakdown of energy generated by the system,
as well as a series of inputs used to calculate the system’s
payback in years. The user can select a second tab that shows
an interactive graph of the system’s hourly energy output.
Finally, the user can select an example load profile that
aims to represent a household’s hourly electricity use. The user
can select one from a pre-generated list of cities or upload a
personal profile that is used to calculate the amount of energy
that the PV system might feed back onto the grid. IMBY uses a
local utility’s residential purchase rate to determine the user’s
monthly electricity costs and shaves the cost based on the
amount of electricity that is fed back onto the grid.
Model Assumptions
IMBY makes no assumptions about local shading or
topography; the map is used only as a guide for placing PV
systems. Systems may be drawn anywhere in the map space,
and are therefore not always realistically placed.
Calculation Algorithm/Methodology
The calculation for the IMBY solar power estimate is
based on a modified version of NREL’s PVWatts calculator.
NREL’s SUNY/Perez solar resource data are used to calculate
the solar resource. The SUNY/Perez data are included in a
satellite-derived hourly dataset that has a spatial resolution of
10 km. The hourly data for the user’s location and year are fed
into PVWatts and used to generate an hourly time series of AC
energy.
This time series represents the estimated output from the
user-defined PV system, and is used to generate several
statistics that are presented to the user. One is a table that shows
month-by-month the sum of AC energy output and the
corresponding dollar value that is based on a local utility
electricity rate. Another generated statistic is the PV system’s
calculated payback. This number represents the number of
years until the system has generated the same amount of
savings as it cost to pay for the PV system. This value takes
into account several values:
The total cost of the PV system, the multiplication of
the system size by the cost per Watt

6
A rebate value taken from the DSIRE database of
renewable energy incentives [14]
Tax credits (state and federal), also taken from DSIRE
[14]
The local utility’s residential electricity rate
User Inputs
The user specifies an address at which to place a PV
system and a drawing tool is used to draw the outline for the
potential PV array on a map. The resulting polygon is used to
pre-populate several needed inputs (all of which may be
adjusted by the user):
DC size (kW)
Derate factor
Tilt angle (degrees)
Azimuth angle (degrees)
Data year
Model Outputs
IMBY outputs the following values:
Initial cost, rebates, and tax credits ($)
Simple PV payback period (years)
Monthly production of electricity and respective dollar
value of electricity produced (kWh/month and
$/month)
If a load profile is chosen and a comparison is done,
IMBY provides a bar graph of the monthly bill
reduction after PV is added.
An example output file from IMBY is shown in Figure 3.
FIGURE 3 – IMBY SOLAR SIMULATION RESULTS
Future IMBY Enhancements
Two primary activities focus on making IMBY a versatile
and robust tool:
Using more realistic building load profiles.
o NREL is developing the capability to
generate several types of building loads for each
Solar America City to allow for a more accurate
estimate of how the PV system might affect the
user’s load profile.
Creating an IMBY version 2.
o This will provide a more user-centric
platform so city planners and developers can
return to IMBY again and again. Each time they
return to IMBY, their previous PV systems will
be available. A user could run many simulations
of the same PV system against many load
profiles and aggregate PV systems to explore
with greater detail the impact of several PV
systems on a particular load profile.
CH2M HILL SOLAR MAP AND SOLAR ESTIMATE
Tool Overview
CH2M Hill has developed two products for estimating PV
potential on roofs in defined geographic areas: the Solar Map
and the Solar Estimate. Both use Google Maps as the
visualization platform, enabling users to view an aerial image
of a location. These tools allow the user to define an address
and output the quantity of PV that could be installed on the
roof. They can also project energy and cost savings.
CH2M Hill is currently developing maps for many entities
and cities, and has completed the development of the San
Francisco Solar Map, which provides mapping analysis of 48
mi2 and cost the city approximately $250,000 [15]. CH2M Hill
is currently developing maps for: the Cities of Berkeley,
Portland, Sacramento, and San Diego as well as Forest City
military communities.
Model Assumptions
Both the Solar Map and the Solar Estimate incorporate the
PV cost assumptions listed in Table 1.
TABLE 1: PV COST ASSUMPTIONS
PV System Size (kW)
Cost ($/Watt)
0–5
10.50
5–10
9.80
10–50
9.25
50–100+
8.50
The San Francisco Solar Map algorithms include an assumption
that 100 to 200 ft2 of roof space is needed per kW. Annual
electricity savings were calculated assuming an electricity tariff
equal to Pacific Gas and Electric’s average total rate of

7
$0.16474/kWh for residential E1 customers as of May 2008.
Carbon savings were calculated based on an assumption that
0.746 lb of carbon dioxide are offset per kWh produced by PV
[16].
Calculation Algorithm/Methodology
The CH2M Hill Solar Map is a Web portal that uses the
S.A.F.E.TM analysis and other calculations to assess the solar
PV potential on rooftops. The S.A.F.E.TM methodology
quantifies the roof area exposed to solar radiation throughout
the year for a specified roof. The data produced by S.A.F.E.TM
are then stored in a database and accessed through a portal that
can include Google Maps. CH2M Hill relies on tools such as
PVWatts or the Clean Power Estimator to compute the size of
the PV system and the amount of electricity that would result
from a PV system installed in these shade-free roof areas.
The Solar Estimate is also a Web portal that bases its
estimates on the area of structures. These data are usually
procured from a city’s or locality’s assessor database. From this
data, the Solar Estimate tool calculates potential available roof
area for solar PV. The Solar Estimate does not take into account
items such as chimneys, air-conditioning units, other structures,
or trees that could shade the roof. It also does not consider the
slope or orientation of the roof. The resulting roof area value is
then used in PVWatts or Clean Power Estimator to determine
the size of the PV system and amount of electricity that could
be produced for the given roof space.
The San Francisco Solar Map employs the S.A.F.E.TM
methodology. The user interface for the San Francisco Solar
Map is shown in Figure 4.
FIGURE 4 – USER INTERFACE FOR THE SAN FRANCISCO
SOLAR MAP
Each building’s estimated roof square footage, as obtained
from the San Francisco Office of the Assessor-Recorder, was
used to estimate available roof area. The S.A.F.E.TM
methodology was then used to calculate the shade-free roof
area for each location. The PV system was sized and the system
electricity production was estimated by applying the value of
peak sun-hours per day. The average peak sun-hours per day
were measured in each neighborhood by the San Francisco
Public Utility Commission’s 11 solar monitoring stations. The
solar insolation by neighborhood in San Francisco by
neighborhood range from 4.1 to 4.6 kWh/m2/day [17].
User Inputs
The user enters an address for examination of PV potential.
Model Outputs
The CH2M Hill Solar Map and Solar Estimator output the
following values:
Roof size (ft2)
Usable roof area (ft2)
Estimated solar PV potential (kW)
Estimated electricity produced (kWh/yr)
Estimated electricity savings ($/yr)
Estimated carbon savings (lb/yr)
An example output file from the San Francisco Solar Map is
shown in Figure 5.
FIGURE 5 – OUTPUT FILE FROM THE SAN FRANCISCO
SOLAR MAP
The San Francisco map also outputs these values:
Currently installed solar PV systems (some or all of
these)
o Building owner type (municipal,
residential, commercial, schools/libraries,
nonprofits, monitoring stations, Environmental
Justice Program)
o Location
o System size (kW)
o System output (kWh/yr)
o Electric savings ($/yr)

8
o Installer
o Picture of system
Case studies of local businesses and homeowners who
have already installed solar PV systems
Information about installing a solar PV system, including
contact information for local solar installers
Future Enhancements
CH2M Hill is currently developing maps for the Cities of
Berkeley, Portland, Sacramento, Pasadena, Anaheim, and San
Diego, as well as Forest City military communities and Los
Angeles County. The Berkeley map will be a Solar Map that
uses the S.A.F.E.TM methodology and will provide analyses of
12 mi2. It cost the city $74,000 and is being developed as part
of the Solar America Cities activities. The Portland map will
use the Solar Estimate methodology. It will cost the City
$25,000 to develop; it is also being developed as a component
of SAI. The Sacramento map will leverage work done through
the Sacramento Municipal Utility District Safe Solar Mapping
and will result in a homeowner self-assessment tool. It will cost
the City $46,000 and is part of SAI. San Diego is having a map
developed that will be a Solar Map and that will use the
S.A.F.E. methodology to analyze 8.4 mi2. It is part of SAI and
is costing the City $65,000.
SOLAR BOSTON MAP
Tool Overview
The City of Boston, in cooperation with the Boston
Redevelopment Authority, has developed the Solar Boston Map
to help track its solar initiative goals and to help residents,
business owners, and decision makers calculate the potential
solar power available at a given location [18]. Boston’s Web
site was built entirely with ESRI ArcGIS software tools. The
user interface for the San Boston Map is shown in Figure 6.
FIGURE 6 – USER INTERFACE FOR THE BOSTON MAP
The Spatial Analyst extension was used to calculate solar
radiation. The tool allows the user to define an address for
consideration and the output includes usable roof area, potential
size PV system (kW), potential annual output, and cost savings
resulting from the PV system.
Model Assumptions
The Solar Boston Map algorithms assume that the roof is
flat. The calculations for potential PV system size assume the
Evergreen Spruce Line solar panel is used, which delivers 11.8
W/ft2 of available roof space. The user selects roof obstructions
and shading with a variable roof percent slider. The maximum
usable area of the roof is 75% and is assumed to be south facing
and free from shading. Annual electricity output is calculated
assuming 1,200 kWh per installed kilowatt. The potential
annual cost savings are determined from the potential annual
output (MWh) and an electricity rate of $0.18/kWh. The Boston
Solar Map also calculates the potential annual avoided
emissions by using the multipliers developed for Massachusetts
by Segue Consulting under subcontract to NREL through the
City of Boston’s Solar City Partnership with DOE [17]. The
multipliers for Massachusetts are: carbon dioxide 1,146 lb,
sulfur dioxide 2.4 lb, and nitrogen oxide 1.1 lb for every
megawatt-hour of solar electricity produced.
Calculation Algorithm/Methodology
The Boston Redevelopment Authority used in-house staff
to develop the Solar Boston Web site. An existing three meter
bare-earth DEM was used as the foundation of the spatial
analysis. A supplementary DEM was created using building
elevation attributes from a building footprint feature class that
were tagged with first return LIDAR values. The resulting
DEM reflects bare earth conditions and building structures. The
algorithm for calculating the solar radiance does not account for
shading from the trees. The actual pitch of the roof is also not
considered and all roofs are assumed to be flat. The resulting
roof area is used by the solar tools to determine the size of the
PV system.
User Inputs
The user may enter an address or select a rooftop to
examine PV potential. The tool also has a drawing feature that
can be used to outline the area of the roof for the PV array.
Model Outputs
The Boston Solar Map outputs the following values:
Chart with Monthly Solar Radiation (kWh/ m2)
Roof size (m2)
Usable roof percent (max 75%) – adjustable slider
Usable roof area (m2)
Estimated solar PV potential (kW)
Incoming solar radiation (kWh/m2)
Estimated electricity produced (kWh/yr)
Estimated electricity savings ($/yr)
Estimated carbon savings (lb/yr)
Currently installed solar PV systems (some or all of

9
these)
o Location
o System size (kW)
o Installer
o Picture of system
Information on installing a solar PV system, including
contact information for local solar installers
An example output file from the Solar Boston Map is shown in
Figure 7.
FIGURE 7 – EXAMPLE OUTPUT FILE FROM THE SOLAR
BOSTON MAP
Future Enhancements
The team that developed the Boston Solar Map hopes to
update its LIDAR scan data to include the first and last returns
with higher resolution. A more detailed DEM would distinguish
trees and other objects that could shade the roof. The software
calculations for solar radiance on the roof could be enhanced to
include pitch and shading.
SOLAR SONOMA COUNTY
Tool Overview
The City of Santa Rosa, in cooperation with Sonoma
County, has developed the Solar Sonoma County Solar Map to
help residents, business owners, and decision makers calculate
the solar potential power available at a given location. The map
was developed by Project DX. The Project DX solar mapping
tool was designed to be used by non-technical commercial and
residential property owners to show the system costs, cost
savings, and payback rates for three solar energy technologies.
The Solar Sonoma County Web site prompts the user for their
address, and the system retrieves information estimates on the
property. This information can be modified as necessary, with
25 information categories ranging from monthly bills to usable
roof area [19].
Once the property information has been verified by the
user there are three types of systems that can be configured: PV,
Solar Hot Water Heating (SHW), or Solar Pool Heating (SPH).
For the PV system, a slider can be adjusted to determine what
percentage of demand the user would like to meet with the solar
system. For the SHW, the slider adjuster determines how many
gallons of water the system holding tank will contain. For the
SPH, there is an input box for the size of the pool (ft2).
Depending on the solar system size input, the various outputs
change accordingly. The user interface for the Solar Sonoma
County Map is shown in Figure 8.
FIGURE 8 – USER INTERFACE FOR THE SOLAR SONOMA
COUNTY MAP
The property energy footprint screen is shown in Figure 9.
FIGURE 9 – PROPERTY ENERGY FOOTPRINT
The output consists of three tabs for each system: the
monthly bill savings, cost savings, and a learn more tab that
provides general information about the system. Once the
analysis is finished, the tool will link the user with a list of local
contractors, lenders, retailers, consultants, maintenance, and
wholesalers that can be contacted to assist with the solar
solutions that the tool outlined.
Model Assumptions
Energy usage at each location:

10
0.5 kW/ft2/month multifamily residence
0.6 kW/ft2/month single family residence
0.7 kW/ft2/month commercial
General Property Information:
Roof area is 80% of building sq. ft.
40% of the total roof area is available for solar
installations
Estimated System Costs:
25-year cost savings calculated with annual utility
inflation rate
Cost estimates could be +/- 20% based on
installation and site characteristics
PV costs for roof installation are $8.5/W
PV costs for ground mount installation are $9.5/W
Sonoma Fuel Costs:
Natural gas costs $1.40/therm
Propane costs $3.25/gallon
Electricity rates equivalent to PG&E E6 Tiered
Rates
Solar Hot Water:
Assumed cost of solar hot water system =
(number of gallons of the system) x ($67.50)
Solar Data:
Yearly averages used for savings analysis
NREL insolation data
Carbon Reduction:
PG&E natural gas carbon dioxide emissions
equivalent to .46 lb/kWh
PG&E CO2 emissions: 0.52lb/kWh electric energy
Solar Pool Energy Usage:
70% of pool area required in solar panel area
Thermal collector area is ~20W/sq ft
Incentives
California Solar Initiative Incentives (www.sgip-
ca.com)
Calculation Algorithm/Methodology
Monthly System Payments:
25-year loan at 6.5% interest for a PV system
15-year loan at 6.5% interest for SHW
15-year loan at6.5% interest for SPH
Cost Savings:
Cost of energy produced by system with utility annual
inflation rate of 4.5%
Benefits Calculations:
Carbon reduction (%)
Carbon reduction (tons/yr)
Grid energy reduction (%)
System efficiency
User Inputs
The only component that the user is required to input is the
address of the residence being considered. All other inputs are
assigned property specific values. These values can be changed
by the user if desired.
Once the default information is accepted or adjusted by the
user, the final input is the desired system size. For PV and
SHW, the slider can be moved to denote the size of the system.
For the SPH, the size of the pool serves as the input.
Model Outputs
The Project DX tool outputs the following values:
Grid energy reduction (with each system)
Carbon reduction (with each system)
Total system cost, state and federal incentives, net
system cost
Monthly savings (on energy bill)
New monthly energy bill
Monthly payment (on system)
Reduction in energy usage
Equivalent number of cars removed from the road
Payback time
Average monthly savings
Cost savings over 25 years (on energy bills)
An example output file from the Solar Sonoma County Map is
shown in Figure 10.
FIGURE 10 – EXAMPLE OUTPUT FILE FROM THE SOLAR
SONOMA COUNTY MAP
Additional Features
Local Contractor Locator
Community Solar Installation Goal Meter
Cost of Financing the System
Cost
The developers of this tool have offered the tool to Sonoma
County for one year, after which they are asking for
$20,000/month paid for by Sonoma County.

11
ADDITIONAL USES OF SOLAR MAPS
NREL analyzed ten Con Edison networks representing the
five boroughs of New York City to determine the maximum
technical PV deployment possible in each network area.
NREL’s IMBY tool was used to estimate the power that could
be produced if all suitable rooftop space in each network area
were covered with PV arrays. The PV generation levels were
then compared to actual hourly load levels in each network. It
was found that in some hours in some networks, under full PV
deployment, PV generation could exceed network load. The
data was further analyzed to determine in which hours PV
generation exceeded network load, and by how much. The
analysis is intended to help New York City and Con Edison
plan for increased deployment of rooftop PV systems, by
providing a better understanding of how full PV deployment
would impact New York City networks.
The City of San Francisco and CH2M Hill are using the San
Francisco Solar Map to analyze 300 apartment buildings in San
Francisco. The map is being used to analyze each individual
rooftop and develop a list of prioritized installations. Once the
installations are prioritized, the city will issue a request for
proposals for the top installations to a set of local solar
installers.
COMPARATIVE ANALYSIS
A comparative analysis was performed to examine the
output results from the different mapping applications. The
Solar Boston Map, the San Francisco Solar Map, and the Solar
Sonoma County map are location-specific, and could therefore
not be compared against each other. Thus, three analyses were
performed, comparing their outputs separately against those of
IMBY and PVWatts. The same PV system size was used for the
three analyses in each comparison.
For the purpose of this paper a preliminary comparison was
done in each city. Further studies are needed to analyze a
variety of system sizes for each mapping tool and statistically
analyze the potential difference between calculated solutions.
The variations in the calculated potential of the PV systems and
the calculated annual electricity produced requires additional
analysis that is outside the scope of this paper.
The San Francisco Solar Map estimated that 319,375 kWh/yr of
electricity would be produced by a 175-kW PV system; this
was the highest output value of the three tools. The lowest
value was 208,059 kWh/yr which was generated by the IMBY
tool. The difference between these highest and lowest output
numbers is 42%. This is not negligible. The discrepancy in
numbers could be attributed to an overestimate in solar resource
or in PV system efficiency by the San Francisco Solar Map, or
to an underestimate by the other tools.
TABLE 2: SAN FRANCISCO TOOL COMPARISON
Sample Address:
211 Main Street
(Commercial)
SF
Solar
Map
IMBY
PVWatts
PV potential (kW)
175
175
175
Elect. Produced
(kWh/yr)
319,375
208,059
219,902
Elect. Cost
Savings ($/yr)
52,614
26,842
27,487
Assumed Elect.
Rate ($/kWh)
0.16474
0.13
0.125
The highest projected electricity output from the Boston Tool
Comparison was 128,647 kWh/yr, which was the output from
the Solar Boston Map. The lowest value was 117,621 kWh/yr,
which was generated by the IMBY tool. The difference between
these numbers is 9%. This is not a large discrepancy.
TABLE 3: BOSTON TOOL COMPARISON
Sample Address:
61 Eutaw Street
(Commercial)
Solar
Boston
Map
IMBY
PVWatts
PV potential (kW)
118
118
118
Elect. Produced
(kWh/yr)
128,647
117,621
121,851
Elect. Cost
Savings ($/yr)
23,156
17,229
14,378
Assumed Elect.
Rate ($/kWh)
0.18
0.15
0.118
The highest projected electricity output from the Project DX
Comparison was 156,302 kWh/yr, which was the output from
the PVWatts tool. The lowest value was 144,818 kWh/yr which
was generated by the IMBY tool. The difference between these
numbers is 8%. This is not a large discrepancy.
TABLE 4: PROJECT DX TOOL COMPARISON
Sample Address:
85 Santa Rosa
Ave.
(Commercial)
Project
DX
Map
IMBY
PVWatts
PV potential (kW)
108.1
108.1
108.1
Elect. Produced
(kWh/yr)
149,121
144,818
156,302
Elect. Cost
Savings ($/yr)
24,456
18,681
19,537
Assumed Elect.
Rate ($/kWh)
0.164
0.13
0.125
All three tool comparisons show a fairly large range in
projected electricity cost savings—a difference of 65% between
the highest and lowest values for the San Francisco tool
comparison, a difference of 47% for the Boston tool
comparison, and a difference of 28% for the Project DX tool

12
comparison. These variations can be attributed to the differing
electricity rates that the tools assume, as well as the varying
estimated amounts of electricity produced.
POTENTIAL AREAS OF IMPROVEMENT
Standardized inputs could be developed for each solar map.
The input data used to categorize usable roof area, PV power
density (W/ft2), installation angle, installed cost, incentives, and
electric rates were significantly different from one map to
another. These discrepancies could be eliminated through the
use of standardized model inputs. The model inputs could
accurately reflect assumptions local installers use when
installing and prioritizing installations.
For each city map that has been developed or is in
development, the city would benefit from clearly defining a set
of metrics of success for the map, depending on the cities
desired outputs. This set of metrics could help define
marketing and outreach activities as well as tracking and
accounting mechanisms that can be used to track the number of
installations that are a direct result of the use of a solar map.
CONCLUSION
Solar mapping applications are increasing in prevalence
and maps are being developed for geographic areas ranging
from cities to the entire United States. Although these tools are
still in their infancy, their potential for informing decisions is
quite large. As an example, in just one month more than 3,700
people have visited the San Francisco Solar Mapping Web site
[20]. However, the number of installed solar PV systems that
have resulted from these maps is currently unknown. In the
future, as cities and private entities make tough decisions about
how to make the largest impact toward renewable energy
technology adoption with minimal funds, they will need to
weigh the costs associated with map development against the
benefits, many of which are currently unknown.
ACKNOWLEDGMENTS
The authors express their appreciation to Donna Heimiller
for providing very constructive insights throughout the process
of developing this report.
REFERENCES
[1] Renne, D., George, R., Wilcox, S., Stoffel, T.,
Myers, D., and Heimiller ,D., 2008, ―Solar
Resource Assessment,‖ NREL/TP-581-42301.
[2] Marion, W., and Urban, K., 1995, ―User’s Manual
for TMY2s,‖ NREL Contract Number: DEAC36-
83CH10093.
[3] Wilcox, S., and Marion W., 2005, ―Users Manual
for TMY3 Data Sets,‖ NREL/TP-581-43156.
[4] Perez ,R., Ineichen, P., Moore ,K., Kmiecik, M.,
Chain, C., George, R., and Vignola F., 2002, ―A
New Operational Model for Satellite-
Derived Irradiances: Description and
Validation,‖ Solar Energy Vol. 73, No.5,
pp. 307–317.
[5] Mehos, M., and Perez, R., 2005, ―Mining for
Solar Resources – U.S. Southwest
Provides Vast Potential.‖
[6] Vignola, F., Harlan, P., and Perez , R., ―Analysis
of Satellite Derived Beam and Global
Solar Radiation Data.‖ US DOE: DE-
FC26-00NT41011.
[7] Spencer B. Gross, Inc. Mapping & Aerial
photography,
http://www.sbgmaps.com/lidar.htm.
[8] The National Oceanic and Atmospheric
Administration (NOAA) Coastal Services
Center, South Carolina's Coast: A
Remote Sensing Perspective,
http://www.csc.noaa.gov/products/sccoasts/html/pdecript.htm.
[9] Hearne, L., and Mathews, D., 2004, ―Improving
the Geospatial Data Extraction and
Analysis Process Using Stereo
Imagery Datasets‖.
[10] Menicucci, D., 1985, ―Photovoltaic Array
Performance Simulation Models,‖ Solar Cells 18,
pp. 383–392.
[11] Perez, R.,―An Anisotropic Model of Diffuse Solar
Radiation with Applications To An Optimization
of Compound Parabolic Concentrators,‖ DOE,
―DEFG0577ET20182, 184.
[12] In My Backyard (IMBY),
http://www.nrel.gov/eis/imby/.
[13] Fu, P., and Rich, P., 2000, ―The Solar Analyst 1.0
User Manual, ―Helios Environmental Modeling
Institute, LLC.
http://www.fs.fed.us/informs/solaranalyst/solar_an
alyst_users_guide.pdf.
[14] Database of State Incentives for Renewable
Energy (DSIRE), http://www.dsireusa.org/.
[15] San Francisco Solar Map, http://sf.solarmap.org/.
[16] International Council for Local Environmental
Initiatives (ICLEI) Clean Air and Climate
Protection Software as applied to the Western
Systems Coordinating Council/CNV sub-region
for 2006.
[17] San Francisco Public Utilities Commission,
http://sfwater.org/custom/solar/solarmap1.cfm
[18] Boston Solar Map,
http://gis.cityofboston.gov/solarboston/.
[19] Solar Sonoma County
http://sonomacountyenergyaction.org.
[20] San Francisco Solar Map website visits during the
time period of January 19, 2009 through February
18, 2009.