Methodology

Production and Investment in Solar and Wind Projects

The U.S. Clean Energy Progress Map includes investments, electricity generation and jobs from solar and wind projects as well as jobs for energy efficiency. Several corroborating and complementing data sources are used to create the maps. Most of the data are from surveys conducted in 2016, though Kevala has been updating the database with additional methods described below.

Solar Projects

Solar projects’ location and size and date of installation are identified from several sources including:

• U.S. Energy Information Administration (EIA) publications
• Solar Energy Industries Association publications
• Building permit records
• NREL’s Open PV database
• Computer vision using satellite imagery
• Project data contributed by developers

Investment in solar projects is estimated by applying year-specific dollar per watt ($/watt) values to project nameplate capacity based on the year projects were built. Annual $/watt values come from NREL’s benchmark reports on solar costs[1]. The most recent year values are used and default values for size by project type are applied if that information isn’t available.

Solar energy production for each project is estimated dynamically for the user-selected search area using NREL’s PVWatts calculator[2]. Inputs used for the PVWatts calculator include actual values for system size, tilt and azimuth. Default values are used if project details are not available.

Wind Projects

Like solar projects, wind projects are identified in reports filed with EIA and the US Geological Survey, and from American Wind Energy Association (AWEA). Project locations are confirmed using satellite imagery.

To estimate total dollars of investment in wind projects, the (1) nameplate capacity (available in US EIA form 860[3]) is multiplied by (2) the average capacity-weighted cost for wind projects ($/kW) for the year the project was put into service. Lawrence Berkeley National Laboratory publishes an annual Wind Technology Report with historical information on capacity-weighted cost ($/kW) for wind projects that provides the source of cost information (see figure 41, 2016 report)[4].

Wind production data by month are published by EIA on a project by project basis for systems equal to or greater than 1MW and are available in EIA form 923[5]. Those data are summed to calculate annual production for each wind project. To account for seasonal variations, the last full calendar year of production data are used. EIA collects data annually for all generators and monthly for a sampling of these generators.

Jobs from Investments in Solar, Wind and Energy Efficiency

Energy Efficiency Jobs

Energy efficiency job data are derived from two data sources and were provided by Environmental Entrepreneurs (E2):

• The Bureau of Labor Statistics (BLS) Quarterly Census of Employment and Wages (QCEW) labor market data for 2016[6].
• Detailed jobs estimated by BW Research, named the Energy Employment Index[7], which builds upon and augments BLS data gathered in 2015. When BW Energy Employment Index publishes the next version based on BLS data for 2016, the map will be refreshed.  The report is anticipated for release in April 2018.

The BW Energy Employment Index relies on the most recently available data from the Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW, Quarter 2), together with a detailed survey of business establishments across the U.S. Taken together, the data provide the most comprehensive calculation of energy-related employment available. The methodology has been used for local, state and federal energy related data collection and analysis for nearly a decade, including The Solar Foundation’s National Solar Jobs Census series (discussed below), clean energy reports for state agencies in the Commonwealth of Massachusetts, State of Vermont, and State of Rhode Island and numerous non-profit agencies across the U.S. The BW Energy Employment Index was also used for U.S. Department of Energy’s U.S. Energy and Employment Report[8].

The Energy Employment Index survey uses a stratified sampling plan that is representative by industry code (NAICS or ANAICS), establishment size and geography. These data are then analyzed and applied to existing public data published by the Bureau of Labor Statistics, effectively constraining the potential universe of energy establishments and employment. BW Research Partnership believes that the methodology used for the Index could be adopted as a supplemental series to the BLS’s QCEW with only minor revision.

The Energy Employment Index survey was administered by telephone via more than 300,000 outbound calls, and by web, with more than 50,000 emails sent to participants throughout the U.S. The phone survey was conducted by I/H/R Research Group and The Castelton Polling Institute. The web instrument was programmed internally and each respondent was required to use a unique ID in order to prevent duplication.

In conducting surveys for the Energy Employment Index, the sample was split into two categories, referred to as the known and unknown universes. The known universe includes establishments that have previously identified as energy-related, either in prior research or some other manner, such as membership in an industry association or participation in government programs.

These known establishments were surveyed census style, and their associated establishment and employment totals were removed from the unknown universe for both sampling and for resulting employment calculations and estimates. The unknown universe includes hundreds of thousands of businesses in potentially energy-related NAICS codes, across agriculture, mining, utilities, construction, manufacturing, wholesale trade, professional services and repair and maintenance. Each of these segments and their total reported establishments (within the Bureau of Labor Statistics QCEW) were analyzed by state to develop representative clusters for sampling. In total, approximately 20,000 business establishments participated in the survey effort, with more than 8,500 providing full responses. These responses were used to develop incidence rates among industries (by state) as well as to apportion employment across various industry categories in ways currently not provided by state and federal labor market information agencies.

Of important note, the Energy Employment Index expressly excludes any employment in retail trade NAICS codes. This excludes gasoline stations, fuel dealers, appliance and hardware stores and other retail establishments. All data in the index rely on the Bureau of Labor Statistics Quarterly Census of Employment and Wages data for the second quarter of 2015. The survey was administered between September 15, 2015 and November 24, 2015 and averaged 14 minutes in length.

Solar Jobs

Solar Job data are provided from the Solar Foundation’s National Solar Jobs Census [9]. The primary data for the Solar Foundation census is the U.S. Department of Energy (OMB Control No. 1910-5179) 2017 U.S. Energy and Employment Report (USEER). U.S. DOE conducted surveys and collected employer-reports similar how the Bureau of Labor Statistics develops the Quarterly Census of Employment and Wages (QCEW) and Current Employment Statistics (CES). The 2016 National Solar Jobs Census uses a statistically significant sample of actual employers that involved more than 500,000 telephone calls and 60,000 emails to get 3,888 fully completed surveys by establishments involved in solar activity in the U.S.

Like the research for energy efficiency jobs, BW Research Partnership administered the surveys based on known and unknown categories of employers:

• The “known universe” of approximately 68,000 energy companies in the Solar Energy Industries Association’s National Solar Database was contacted to obtain 3,595 full or substantially completed surveys.
• The “unknown universe” of approximately 410,000 potentially energy-related industries was sampled using stratified, clustered, random sampling to obtain completed surveys from 293 establishments.

The sampling rigor in the known and unknown universes are reported to provide a margin of error for establishment counts at +/-0.85% and employment at +/-2.30% at a 95% confidence interval.

Furthermore, the National Solar Jobs census includes data from CEB TalentNeuron™, a source of job market data utilizing aggregated job postings from the Internet.  This database provided the basis for the number of new solar-related job postings per quarter and median salaries by aggregating job postings from corporate and online recruiting sites.

Wind Jobs

Job data related to the wind industry are provided via annual report from the American Wind Energy Association (AWEA) as reported in the 9th annual U.S Wind Industry Annual Market Report for Year Ending 2016 on April 11, 2017.  The extensive report includes the follow data:

• Review of regional electricity markets
• Regional age profile of U.S. wind projects
• Rural economic impacts
• Wind jobs by type and by state
• Emission reductions; water consumption avoided; wind energy activity by congressional district.

Summary of Data Sources

Spatial Precision of the Map

The table below shows the level of detail mapped by each technology type: wind, solar or energy efficiency. Solar and wind projects are visualized as chloropleths. Energy efficiency projects are not shown because they are not known. The * for Projects is a reminder that actual projects are not mapped, rather the chloropleths use a color scale to indicate the level of project development activity compared to neighboring areas.

Actual project locations are used to inform the distribution of all projects by census block based on statistical models of solar adoption. Actual project locations are not identified. In addition to analyzing actual projects, the probable location of solar projects, by census geography, is inferred when project information does not include precise location (e.g., street address).

√ = available at cleanenergyprogress.com; X = not available; * = Actual projects are not mapped, rather project counts are inferred and maps show chloropleths of the inferred project count.