Spatio-temporal networks of light pollution

Abstract

Light pollution is an environmental problem produced by an unjustified escalation of upward propagated artificial lighting from inaccurately situated or designed artificial night lighting systems. The radiative transfer of the nocturnal artificial lighting in cities spreads spatio-temporally with a certain randomness, where the modelling through network theory allows to describe invariants of this phenomenon. A critical part of light pollution studies is to quantify how spatial heterogeneity relates to the lighting system. Since satellite images are the visual representation of the information captured by a sensor, they can be used to obtain information data on artificial night radiation. In this sense, the utilization of such images must be justified with the correct use algorithms for data calibration and noise reduction in order to establish the baseline of the data to be used. In this work we analyse satellite images that represent light pollution in a specific time frame. Using these images, we propose to represent the change in light pollution as a network. We calculated some classical network measurements to characterize the networks. In doing so we demonstrate the utility of this representation as an urban construct with respect to radiative transfer of artificial light. The results indicate that largest cliques and transitivity are a good interpretive structural modelling-based approach to determine aspects of light pollution that may be related to urban processes. Therefore, we hope this document will motivate future research that links urban processes with the increase of nocturnal artificial lighting.