Greening Local Energy: Explaining the Geographic Distribution of Household Solar Energy Use in the United States

Abstract

Problem: Solar energy has potential to solve many types of planning problems. Knowing where existing household solar energy users are located and what factors explain this distribution can help craft appropriate local policies.

Purpose: This study analyzes the spatial distribution of households who heat their homes with solar energy across the contiguous United States.

Methods: We use geographic information systems (GIS) and zero-inflated negative binomial statistical techniques to test three sets of geographic predictors at the scale of the county: environmental, economic, and sociopolitical.

Results and conclusions: Descriptive, GIS, and regression results indicate that the expected number of households using solar energy to heat their homes increases significantly with the amount of solar radiation received, but that other environmental, socioeconomic, and political factors are also significant predictors. The number of solar households in a county is a positive function of wealth (operationalized as median home value), urbanization, and the percentage of residents in the peak period of the lifecycle-consumption curve. Having a solar energy technology provider in the county did not appear to be significant. Finally, we confirmed that households heating with solar energy increase with the percentage of persons who vote for the Democratic Party in presidential elections and with local government involvement in the International Council for Local Environmental Initiatives.

Takeaway for practice: Our model can be used to enhance adoption of solar technologies by showing which localities possess most of the attributes that predict solar use, but adoption lags expectations. This can help design efficient and effective local plans and incentives calibrated to local environmental, economic, and sociopolitical conditions.

Research support: Portions of the data collection for this research were supported under Award No. NA03OAR4310164 by the National Oceanic and Atmospheric Administration, U.S. Department of Commerce. The statements, findings, conclusions, and recommendations are those of the authors and do not necessarily reflect the views of the National Oceanic and Atmospheric Administration or the Department of Commerce.

Keywords:

• household solar energy use
• solar heating
• GIS
• green communities

Notes

ψp < .10

*p< .05

**p< .01

Note: a. From the zero-inflated negative binomial regression models shown in Table 2.

1. There are few scientific studies of households' experiences with active solar energy systems, but Sawyer and Wirtshafter (1985) discovered that 73% of solar users reported serious technical problems with system performance. They found that average annual repair expenditures topped 4.5% of initial purchase price, far exceeding what product payback models had assumed. Yet, despite such high malfunction rates, 40 of 60 homeowners using solar energy systems in Florida reported high satisfaction with the products.

2. In addition to the lapse of solar tax credits, the low price of conventional fuels from 1990 to 2000 may account for the observed decline in household solar energy use. Measured in constant 2006 U.S. dollars, the average price of a barrel of crude oil remained below the 50-year (1947–2006) world average of $25.56 and the U. S. average of $23.67 throughout the 1990s; even dipping below the median U. S. and world price of $18.43 in both 1994 and 1998. Researchers note a positive relationship between the price of conventional fuels and household solar energy use (Dukert, 1985; Durham, Colby, & Longstreth, 1988). The historically low price of crude oil during the 1990s stretched the payback period in areas of low solar irradiance beyond the average 15-year useful life of solar heating products, making it uneconomic to adopt such products (Pimentel et al., 1994). Many who adopted solar heating systems in the 1980s likely did not update them in the mid-1990s because conventional fuel costs were so low. We expect household solar use will rise again with the price of crude oil at a historic high.

3. Concern for climate change partially accounts for the renewed interest in solar energy (Gadsden, Rylatt, & Lomas, 2003). The thermometric instrumental record indicates that global average surface temperature is increasing, up by about 0.6 degrees Celsius in the last 100 years. The scientific community is increasingly certain that greenhouse gas emissions are primarily responsible for observed variation in temperature change (Oreskes, 2004). The expected risks of temperature change are many, including coastal flooding and beach erosion, extreme weather events, loss of habitat and species, fluctuations in crop yields, and increased spread of vector-borne diseases like malaria and encephalitis (Hurd, Callaway, Smith, & Kirshen, 2004; Parry et al., 2001; Scheraga & Grambsch, 1998; Smith, Lazo, & Hurd, 2003). To mitigate the expected risks of climate change, planners and policymakers advocate a switch to energy sources that limit carbon dioxide emissions and concentrations in the atmosphere. For example, the state of California recently passed the Million Solar Roofs Plan (SB1), calling for 1 million solar roofs across the state by 2018. The plan explicitly addresses climate change: with 1 million rooftops of solar power, officials expect to reduce the greenhouse gas emissions by 3 million metric tons, the equivalent of removing 1 million cars from the road.

4. The choices for answering the question are: “Gas: from underground pipes serving the neighborhood”; “Gas: bottled, tank, or LP”; “Electricity; Fuel oil, kerosene, etc.”; “Coal or coke”; “Wood”; “Solar energy”; “Other fuel”; “No fuel used.”

5. For Census 2000, this includes all territory, population, and housing units in urbanized areas and urban clusters. The urban classification may cut across other geographic entities; for example, there is generally both urban and rural territory within both metropolitan and nonmetropolitan areas.

6. The NCSS core file merges descriptive information from three cumulative files compiled by the IRS: the business master file, the return transaction file, and the statistics of income file. The NCCS conducts standardized checks on all information, making its core file “the most complete and highest quality data source ever available on nonprofit organizations” (Lampkin & Boris, 2002, p. 1683). However, because it counts only organizations with $25,000 dollars or more in gross receipts, it probably undercounts environmental groups for our purposes, potentially missing highly motivated, but smaller grassroots environmental organizations.

7. To provide a sense of the metric, 1 kWh runs a typical space heater.

8. The conditional variance is computed as: Var(y_i I x_i, z_i) = μ_i (1 – ψ_i) [1 + μ_i(ψ_i + α]