Ground-based hyperspectral analysis of the urban nightscape
Introduction
Light pollution is an issue of growing interest for urban planners, city managers and environmental protection agencies. In its broadest sense, this term encompasses the undesired side-effects associated with the production and use of artificial light, especially at nighttime (Falchi et al., 2011, Bará, 2013, Bará, 2014). Excessive or misdirected light leads to unnecessary energy waste and increased greenhouse gas emissions (Gallaway et al., 2010, Kyba et al., 2014), and poses, according to recent findings, non-negligible threats to the nocturnal environment (Hölker et al., 2010, Gaston et al., 2013, Gaston et al., 2015) as well as potential health hazards (Cho et al., 2015, Haim and Zubidat, 2015, Stevens et al., 2013).
Making strategic decisions on outdoor lighting policy at local and regional levels requires an accurate knowledge of the actual light emissions, ideally with enough spatial and spectral resolution to allow the identification of individual light sources, their directional emission patterns and their detailed spectral composition (Elvidge et al., 2010).
Urban radiance data with moderate spatial resolution are currently available from several instruments on board of space platforms, such as the historical time series of the Operational Linescan System (OLS) of the U.S. Defense Meteorological Satellite Program (DMSP) (Cinzano and Elvidge, 2004, Elvidge et al., 2013, Yang et al., 2014), the Suomi-NPP VIIRS instrument Day-Night Band (DBN) (Miller et al., 2013), and the Earth images taken with commercial grade cameras from the International Space Station (ISS) (Castiglione et al., 2012). The DMSP and Suomi-NPP satellites, located in near polar orbits at altitudes close to 850 km, provide whole Earth coverage with radiance detection limits of order 5 × 10−10 and 2 × 10−11 W cm−2 sr−1, and ground footprint sizes of 5 km and 742 m, respectively. They give no spectrally resolved information, because the detection is performed in a single 0.5–0.9 μm panchromatic band (Elvidge et al., 2013). NightPod-based ISS nighttime imagery (Castiglione et al., 2012), acquired from an orbit at about 400 km altitude with 51.6° inclination and variable footprint pixel size, is restricted in turn to the conventional Bayer matrix RGB bands. Despite this spectral limitation, satellite data are the primary source of information about urban wasted light at planetary scale (Cinzano et al., 2000, Cinzano et al., 2001, Cinzano and Elvidge, 2004, Sánchez de Miguel et al., 2014), correlate well with on-site flux measurements, and are useful tools for monitoring changes in upward light emissions after remodeling urban lighting systems (Estrada-García et al., 2015). The number of spaceborne instruments monitoring the city lights has been, however, comparatively very small until now (Belward and Skøien, 2015). Given the useful information they may provide and the relevance of the study of the anthropogenic emissions of light it is highly desirable that new satellite programs, ideally with nighttime multispectral capabilities, be planned in the near future.
Considerably better spatial and spectral resolution can be achieved with airborne hyperspectral imaging spectrometers, at the expense of a smaller area coverage. Hyperspectral imaging, that has found many applications in Earth imagery (see, e.g., Aasen et al., 2015, Clark and Kilham, 2016) including urban areas (Demarchi et al., 2014, Kotthaus et al., 2014), has also been successfully applied to the study of nighttime artificial lights (Barducci et al., 2003a, Barducci et al., 2003b, Barducci et al., 2003b, Barducci et al., 2006, Kruse and Elvidge, 2011). Airborne hyperspectral measurements combined with photogrammetric imagery allow to obtain reliable estimates of the upward luminance from urban areas with spatial resolutions well below one meter, as we have shown in previous works (Pipia et al., 2014, Corbera et al., 2015).
The overall radiant flux emitted towards the upper hemisphere by urban areas is a measure of the degree of lighting energy waste and is the leading cause of increased skyglow. This flux, however, is only part of the whole light pollution picture. Several detrimental effects of light pollution are essentially ground-level phenomena. They include, among others, light intrusion, disability and discomfort glare, light clutter, and the diverse photobiological effects of light at night. To evaluate the severity of these effects one needs to know the spectral radiance of the urban nightscape, as seen from the city dwellers standpoint. This radiance distribution allows the assessment of glare and light cluttering. It is also instrumental for computing the spectral irradiance in a plane tangent to the human corneal vertex for arbitrary gaze directions, which is the primary input for evaluating the physiological effects of light at night by means of suitable phototransduction models (CIE, 2015, Rea et al., 2005, Rea et al., 2010, Rea et al., 2012). The vertical irradiance on window panes allows, in turn, to quantify the severity of light trespass effects.
As in other areas of remote sensing (Chen et al., 2016), a comprehensive assessment of the overall levels of light pollution can only be achieved if the airborne and satellite data are complemented with ground based measurements acquired from within the town. In this work we describe the adaptation and operation of a hyperspectral imaging sensor routinely used for airborne surveys to acquire these additional datasets from the city streets. Hyperspectral light pollution measurements require dealing with spatially uneven, high dynamic range radiance distributions, and keeping the noise propagation within admissible levels. In Section 2 we describe the hyperspectral measurement system. The data reduction process implemented to get the desired magnitudes is summarized in Section 3. The results of a proof-of-concept measurement campaign carried out in the city of Barcelona are reported in Section 4. Discussion and conclusions are finally drawn in Sections 5 and 6, respectively.
Section snippets
Site selection
The field campaign was carried out at the city of Barcelona (3 million inh.), capital of Catalonia, in the period around the new Moon in February 19th to 20th, 2015. Its public outdoor lighting is mainly based on high pressure sodium vapor and metal halide lamps, with an increasing presence of phosphor-coated white Light Emitting Diode (LED) streetlights and self-luminous billboards. Several downtown areas were selected for measurements, including Plaça d'Espanya (site 1) and Jardinets de
Methods
In this section it is described on the one hand the theoretical frame of the work, including the products derived from the hyperspectral images, and, on the other hand, some image processing methods to deal with several image artifacts inherent to night images.
Results
The outcomes presented in this section correspond to the Jardinets de Gràcia site, and are based on the data taken in the night of February 19th to 20th, 2015. This date was chosen to minimize the contribution of moonlight to the nightscape radiance. The new Moon occured at 00:49 h of February 19th (local time CET = UTC + 1). Sunset took place at 18:29 of February 19th and the Moon was below the horizon from 19:25 h of this day until 08:20 h of Feb 20th. The measurements were taken in the first part
Discussion
The results presented in this work are affected by several limitations. One of them stems from the fact that the irradiances at vertical planes have been computed from radiance distributions with a vertical swath limited to about 40° (i.e., 20° above and below the horizontal) instead of the nominal 180° angular span extending from zenith to nadir. While in fact these radiance data capture the essential features of the urban nightscape from the viewpoint of pedestrians gazing horizontally, and
Conclusions
We described the use of a hyperspectral camera to characterize the urban emissions of light at night from a ground-based perspective, complementary to airborne or spaceborne remote sensing. Whereas the latter provide valuable data to estimate the amount of energy wasted skywards by public and private outdoor lighting systems, our present approach provides supplementary information from the city dwellers standpoint, allowing to assess other relevant aspects of the light pollution problem, as
Funding
This work has been developed within the framework of the Spanish Network for Light Pollution Studies (AYA2015-71542-REDT).
Conflicts of interests
The Institut Cartogràfic i Geològic de Catalunya (ICGC) is a public body of the Catalonian Government that provides photometric and photogrammetric services to local communities, public administration bodies and private customers, including airborne and ground-based hyperspectral imaging.
Acknowledgments
We acknowledge the useful suggestions and comments made by the anonymous reviewers of this manuscript.
References (53)
- et al.
Generating 3D hyperspectral information with lightweight UAV snapshot cameras for vegetation monitoring: from camera calibration to quality assurance
ISPRS J. Photogramm. Remote Sens.
(2015) - et al.
The spectral amplification effect of clouds to the night sky radiance in Madrid
J. Quant. Spectrosc. Radiat. Transfer
(2016) - et al.
Who launched what, when and why; trends in global land-cover observation capacity from civilian earth observation satellites
ISPRS J. Photogramm. Remote Sens.
(2015) - et al.
Information from imagery: ISPRS scientific vision and research agenda
ISPRS J. Photogramm. Remote Sens.
(2016) - et al.
Mapping of land cover in northern California with simulated hyperspectral satellite imagery
ISPRS J. Photogramm. Rem. Sens.
(2016) - et al.
Assessing the performance of two unsupervised dimensionality reduction techniques on hyperspectral APEX data for high resolution urban land-cover mapping
ISPRS J. Photogramm. Remote Sens.
(2014) - et al.
Limiting the impact of light pollution on human health, environment and stellar visibility
J. Environ. Manage.
(2011) - et al.
The economics of global light pollution
Ecol. Econ.
(2010) - et al.
Light pollution as a biodiversity threat
Trends Ecol. Evol.
(2010) - et al.
Derivation of an urban materials spectral library through emittance and reflectance spectroscopy
ISPRS J. Photogramm. Remote Sens.
(2014)
Measuring and using light in the melanopsin age
Trends Neurosci.
A model of phototransduction by the human circadian system
Brain Res. Rev.
Evolution of the energy consumed by street lighting in Spain estimated with DMSP-OLS data
J. Quant. Spectrosc. Radiat. Transfer
Adverse health effects of nighttime lighting
Am. J. Prev. Med.
Mapping the human footprint from satellite measurements in Japan
ISPRS J. Photogramm. Remote Sens.
Radiometric calibration of the compact airborne spectrographic imager (CASI)
Can. J. Remote Sens.
Anthropogenic disruption of the night sky darkness in urban and rural areas
Roy. Soc. Open Sci.
Effects of light pollution revealed during a nocturnal aerial survey by two hyperspectral imagers
Appl. Opt.
Light pollution observed by MIVIS and VIRS 200 during a nocturnal remote sensing campaign over Italy
In Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II, Manfred Ehlers, Editor, Proceedings of SPIE
Hyperspectral remote sensing for light pollution monitoring
Ann. Geophys.
Protecting the melatonin rhythm through circadian healthy light exposure
Int. J. Mol. Sci.
High dynamic range photogrammetry for synchronous luminance and geometry measurement
Light. Res. Technol.
Effects of artificial light at night on human health: a literature review of observational and experimental studies applied to exposure assessment
Chronobiol. Int.
Cited by (33)
Spectral-level assessment of light pollution from urban façade lighting
2023, Sustainable Cities and SocietyHow to measure light pollution—A systematic review of methods and applications
2023, Sustainable Cities and SocietyQuantification of ecosystem services supply-demand and the impact of demographic change on cultural services in Shenzhen, China
2022, Journal of Environmental ManagementCitation Excerpt :Therefore, for areas that show a surplus to deficit from day to night, the forms of night tourism can be enriched and supporting facilities can be added. Secondly, for areas that show a deficit to surplus from day to night, consider reducing unnecessary illumination to achieve the goal of reducing light pollution and energy consumption (Alamús et al., 2017). Finally, sound tourism infrastructure is essential to attract tourists (Kanwal et al., 2020; Mou et al., 2022), however, some scenic spots in Shenzhen are located within the ecological control line (Fig. S4), which restricts infrastructure construction.
Keeping light pollution at bay: A red-lines, target values, top-down approach
2021, Environmental ChallengesCollaborative learning of lightweight convolutional neural network and deep clustering for hyperspectral image semi-supervised classification with limited training samples
2020, ISPRS Journal of Photogrammetry and Remote SensingRemote sensing of night lights: A review and an outlook for the future
2020, Remote Sensing of EnvironmentCitation Excerpt :The ideal system would be to use a hyperspectral imaging spectrometer at low altitudes. Few studies with limited spatial coverage exist, such as the first ever performed over 1998 in Las Vegas, USA (Elvidge and Green, 2005; Alamús et al., 2017, see section 2.3 for more). Elvidge et al. (2010) showed that it was in principle possible to discriminate light sources with multispectral sensors, using detailed spectral field measurements and a modeling approach for the pre LED technologies.