Recovering the city street lighting fraction from skyglow measurements in a large-scale municipal dimming experiment

Highlights

• An experiment was conducted involving the dimming of nearly 20,000 municipally owned street lights in a U.S. city of half a million inhabitants over ten nights in the spring of 2019.

• The signal of the dimming tests was successfully observed in measurements of zenith night sky radiance obtained at various radii from the city center.

• Observedzenith radiance changes of $\le -5$% during the tests imply that known street lights account for (26 ± 4)%

• The results can be used to inform future optimization of the city’s street lighting system to further reductions in skyglow while preserving public safety.

Abstract

Anthropogenic skyglow dominates views of the natural night sky in most urban settings, and the associated emission of artificial light at night (ALAN) into the environment of cities involves a number of known and suspected negative externalities. One approach to lowering consumption of ALAN in cities is dimming or extinguishing publicly owned outdoor lighting during overnight hours; however, there are few reports in the literature about the efficacy of these programs. Here we report the results of one of the largest municipal lighting dimming experiments to date, involving  ~ 20,000 roadway luminaires owned and operated by the City of Tucson, Arizona, U.S. We analyzed both single-channel and spatially resolved ground-based measurements of broadband night sky radiance obtained during the tests, determining that the zenith sky brightness during the tests decreased by ($-5.4\pm 0.9$)% near the city center and ($-3.6\pm 0.9$)% at an adjacent suburban location on nights when the output of the street lighting system was dimmed from 90% of its full power draw to 30% after local midnight. Modeling these changes with a radiative transfer code yields results suggesting that street lights account for about (14 ± 1)% of light emissions resulting in skyglow seen over the city. A separate derivation from first principles implies that street lighting contributes only $2-3$% of light seen at the zenith over Tucson. We discuss this inconsistency and suggest routes for future work.

Introduction

Light pollution is a global phenomenon caused by the prolific use of artificial light at night (ALAN) [1], [2]. ALAN offers clear benefits to human society by ensuring safe transit at night, enabling the nighttime economy, and enhancing public perception of outdoor spaces at night through placemaking, [3] but its use entails a number of known and suspected hazards to the natural nocturnal environment, e.g., [4], [5], [6], including potentially significant disruption of ecosystem services [7], [8], [9] and threats to biodiversity [10], [11]. On the other hand, actively preserving natural nighttime darkness not only appears to convey environmental benefits, but also can support sustainable rural economic development through ‘astrotourism.’ [12], [13], [14]

Given the negative environmental influence of ALAN, activists have called for addressing the problem through various lighting engineering and public policy means. Achieving these goals requires identifying best practices for reducing light emissions to levels strictly necessary to ensure public safety. The efficacy of these practices is in part determined by the participation rate of owners of light sources: the more sources whose output is reduced or eliminated presumably results in a greater reduction in the amount of ALAN in the nighttime environment. Public lighting is an attractive target for reduction efforts, especially in urban settings, given that many thousands of luminaires are often under the control of a single administrative entity.

Limited evidence exists to date suggesting that modifications to existing municipal lighting systems, in particular, can yield measurable environmental effects [15], [16]. However, current models are acutely deficient in the sense that the fraction of total light emissions comprised by municipal lighting in cities is not well constrained. While some cities have created inventories of publicly owned lighting, including light emission parameters of luminaires, a complete accounting of privately owned lighting is available only for the smallest municipalities. Currently available remote sensing data do not have sufficient spatial resolution to reliably determine the ownership status of lighting, although one approach to solving that problem is to make an initial guess of the function of outdoor lighting by matching the distribution of remotely sensed night lights to maps of land use patterns [17]. However, even high-resolution aerial imagery of cities is not presently subject to analysis techniques that can reliably discern between lighting types.

In order to better inform models, there is a need to determine the relative contributions of public and private sources to the total light emissions of cities. Reducing the influence of light pollution in cities may be well served by focusing on public lighting; however, if public lighting were a small fraction of total lighting, then the impact of such efforts would be proportionately less. Determining the fraction of total lighting attributable to publicly owned sources in cities is therefore important to guide decisions on how best to direct mitigation efforts.

Previous attempts to determine the relative contributions of public and private lighting in cities have found public lighting fractions ranging from $\sim 10-75$% of total urban light emissions [16], [17], [18], [19], [20]. Several methods have been employed to determine these values, most of which rely on certain model assumptions. Perhaps the most robust approach to distinguish public lighting from private lighting is to change the output of only public light sources in some known way and measure by how much the total city light emission changes. This is relatively easy to arrange for smaller villages (e.g., [21]), but it is difficult for large cities. We arranged a test with the municipal government of Tucson, Arizona, U.S., that involved dimming the municipal street lighting system by certain amounts on a series of test nights. By controlling the output of street lights, we recovered the fraction of the total light emission of Tucson represented by the street lights through a combination of measurements of the brightness of the night sky and modeling skyglow with a radiative transfer code.

This paper is organized as follows. First, in Section 2, we outline the parameters of the street lighting dimming experiment and the intended goals of the project. Next, in Section 3, we describe the measurements made during each of the test nights. Then, in Section 4, we present the results of radiative transfer model runs used to predict skyglow changes over the city consequent to the dimming tests. We analyze the observations in the context of the model results in Section 5, and finally we summarize our work, point out its limitations, and offer suggestions to guide future experiments of this nature in Section 6.