High resolution map of light pollution over Poland

https://doi.org/10.1016/j.jqsrt.2016.03.014Get rights and content

Highlights

  • A simple model and high-resolution data were used to create a map of light pollution over Poland.

  • Resultant map has 100 m resolution and was calculated in a reasonable time.

  • This method is a new approach to investigate night sky brightness on large areas.

Abstract

In 1976 Berry introduced a simple mathematical equation to calculate artificial night sky brightness at zenith. In the original model cities, considered as points with given population, are only sources of light emission. In contrary to Berry׳s model, we assumed that all terrain surface can be a source of light. Emission of light depends on percent of built up area in a given cell. We based on Berry׳s model. Using field measurements and high-resolution data we obtained the map of night sky brightness over Poland in 100-m resolution. High resolution input data, combined with a very simple model, makes it possible to obtain detailed structures of the night sky brightness without complicating the calculations.

Introduction

The idea behind modeling of light pollution is to calculate brightness of night sky caused by artificial lighting. To build a model of light pollution, it is necessary to do many simplifications and assumptions. Modeling of light propagation through an atmosphere needs information or assumptions about the state of an atmosphere (in particular state of the boundary layer of an atmosphere) and environment (e.g. albedo of surface). Additionally it is necessary to add extra restrictions to made the model calculable within reasonable time.

First models of light pollution were very simple. One of the first models is the model introduced by Treanor in 1973 [11]. He used simplified formula for light propagation through an atmosphere. Light sources are exclusively cities and they are assumed to be point sources. Brightness is proportional to the population of a city. In Treanor׳s model, atmosphere is homogeneous. Light coming from a city obey inverse square law with distance and atmospheric extinction. Light beam reaching zenith above an observer is a sum of two beams: direct beam coming from a city and second beam subjected to scattering limited to a cone of small angle. A portion of the total beam is scattered down to the observer.

In 1976, Berry [3] modified Treanor׳s model and created the map of night sky brightness over Ontario in Canada. This map has resolution 8 km×8 km. He assumed that output light of a city is proportional to square root of its population, instead of direct proportionality assumed by Treanor. The square root proportionality was in a better agreement with Berry׳s observations. Square root proportionality increases importance of satellite cities. Berry also assumed that downward scattering takes place at a given height above an observer.

These two models are based on the assumption that all light comes from urban areas. Input data for calculations are position and population of cities which are considered to be points. This is the main reason of discrepancies between calculated values of brightness and observations. Light emission of cities is not exactly proportional to their population. In these models, light sources other than habitable zones are neglected. In large scale maps of artificial brightness it is a good approximation, but when we go into smaller scales and more detailed maps this description is inadequate.

Current trend in modeling of light pollution is to obtain better description of light propagation in an atmosphere. Also a breakthrough was when satellite data became available. Measurements obtained from ground, which were used in Treanor׳s and Berry׳s model, are available only for limited sites. Also these are spread over a certain time range. An alternative for ground-based measurements are data of an irradiance on top of an atmosphere measured by satellites. Cinzano [4] used data from the US Air Force Defense Meteorological Satellite System (DMSP) Operational Linescan System (OLS) to create a global map of artificial night sky brightness [5]. The method of calculation of light propagation was similar to the method introduced by Garstang [6]. Scattering on molecules and aerosols, atmospheric extinction and Earth curvature were taken into account. State of an atmosphere (e.g. content and distribution of molecules and aerosols) must be assumed for computations of light propagation. Resultant map has 30 arcsec resolution, which is 0.927 km at the equator.

It is necessary to take into account many conditions to obtain realistic description of light pollution at certain area. Such detailed modeling is implemented in ILLUMINA [1]. ILLUMINA simulates the contribution of artificial light source to the total night sky brightness at given wavelengths. The model takes into account both 1st and 2nd order molecular and aerosol scattering and aerosol absorption. It also takes into account shape of a city and distribution of light inside it, which might not be homogeneous. Information about topography of a terrain is used. In this way, the shadowing effect – important factor for complex topography – is included in the model. Necessary input data for the software are light luminosity and angular photo-metric functions at given wavelength, lamp heights, elevation map, ground albedo for the same wavelength, aerosol optical depth, cross-sections for aerosol scattering and absorption [2]. Such detailed analysis needs a lot of computational power. This is the model dedicated for studying light pollution at selected small areas – a case study rather than a large-area survey.

We present a solution, which allows to create a map of night sky brightness for a large area in reasonable computing time. In our solution, we use high-resolution map of human settlements and urbanization (GHSL) as input data. The GHSL is available as a raster map with spatial resolution of 100 m. Each cell in the GHSL raster contains percentage of built-up area coverage per spatial unit. Our method is based on Berry׳s model. The simple form of the model is preserved, but we have introduced different assumptions. Instead of using cities and its populations as input data, we used values of each cell of the GHSL. Therefore resolution increases, but simplicity of calculations remains.

The result of using our model is the map of light pollution over Poland. Calculated values are in a good agreement with observational data collected with Sky Quality Meter. Comparison with satellite data 3 reveals that such solution is a very promising method for large-scale calculations. Additionally, we have created a software module which implements our approach. The module can be used as an extension of GRASS (Geographic Resources Analysis Support System). GRASS is the GIS (Geographical Information System) software which has been used to calculate the map of light pollution.

Section snippets

Global human settlements layer

Instead of using population data, according to the original Berry׳s assumptions, we used the Global Human Settlements Layer – GHSL [10]. The GHSL is a raster map, which contains information about percentage of built-up area per spatial unit. The GHSL was created by the European Commission, Joint Research Centre, Institute for the Protection and Security of the Citizen, Global Security and Crisis Management Unit (http://ghslsys.jrc.ec.europa.eu/).

Panchromatic and multi-spectral satellite images

Results

In Fig. 4 the map of night sky brightness over Poland is presented. Night sky brightness on the map is a sum of natural and artificial brightness. It is color-coded from black, through blue, yellow to white. Resolution of the map is 100 m. It means that the output raster map consists of 73 610 720 cells. Darkest areas have brightness 21.8 magarcsec−2. Darkest areas are situated mainly in the north part of Poland. In the center of Poland areas are not so dark, because of light emission from villages

Conclusions

Modeling light pollution is a complex task. Many factors need to be taken into account. Complicated models describe physical situation in more detail, but these are limited to small areas, because of limited computational power. We proposed a new method to investigate the light pollution on large areas in high resolution. Good quality, high resolution data and simple model are used instead of models with complex physics. With this method we created a map of night sky brightness over Poland in

Acknowledgments

This work was supported by the University of Cincinnati Space Exploration Institute. We also want to thank Dr. Sylwester Kołomański and Dr. Tomasz Mrozek for help with data collection.

References (12)

  • Aubé M, Franchomme-Fossé L, Robert-Staehler P, Houle V. Light pollution modeling and detection in a heterogenous...
  • M. Aubé et al.

    Using two light-pollution models to investigate artificial sky radiances at canary islands observatories

    Mon Not R Astron Soc

    (2012)
  • R.L. Berry

    Light pollution in southern Ontario

    J R Astron Soc Canada

    (1976)
  • P. Cinzano et al.

    The artificial night sky brightness mapped from DMSP satellite operational linescan system measurements

    Mon Not R Astron Soc

    (2000)
  • P. Cinzano et al.

    The first world atlas of the artificial night sky brightness

    Mon Not R Astron Soc

    (2001)
  • R.H. Garstang

    Model of artificial night-sky illumination

    Publ Astron Soc Pac

    (1986)
There are more references available in the full text version of this article.

Cited by (12)

  • Analysis of Light Pollution in Ticino region during the period 2011-2016

    2020, Sustainable Cities and Society
    Citation Excerpt :

    In recent years, it has been used in a large number of studies for several sky brightness evaluations (Hänel et al., 2018). For instance, to measure the different level of light pollution in urban and rural locations around Hong Kong (Haim & Portnov, 2020); to assess the cloud coverage amplification of the sky luminance in Berlin urban area and in the rural areas nearby (Kyba et al., 2011); to validate mathematical models for mapping night sky brightness on large territorial scale over Poland (Netzel & Netzel, 2016); and to analyse the light pollution levels offshore based on the distance from the Mediterranean coastline near Barcelona (Ges et al., 2018). The SQM detector consists of a solid-state light-to-frequency detector (TAOS - TSL237) able to measure the brightness, as a sum of both the sky background brightness and the stars, in an interval between 15 and 23 mag/arcsec2 (SQM brightness values are reported in the unit mag/arcsec2, but can be converted into cd/m2 using Eq. (1)) with a precision of 0.01 mag/arcsec2.

  • A photometric mapping of the night sky brightness of the Maltese islands

    2020, Journal of Environmental Management
    Citation Excerpt :

    Whilst such images help to point out zones where there is upward light emission, they represent a different (and complementary) picture to the ‘view from the ground’ as captured by ground-based measurements of the NSB, which is affected by partial backscattering from air particles, particulate matter (e.g.aerosols), and other environmental factors. Similar NSB-mapping studies have been carried out in other regions, e.g.Perth(Biggs et al., 2012), Madrid(Zamorano et al., 2016), Poland(Netzel and Netzel, 2016), and Eastern Austria(Posch et al., 2018). Such studies are, in turn, well placed to validate modelling approaches to the NSB.

  • High-resolution map of light pollution

    2018, Journal of Quantitative Spectroscopy and Radiative Transfer
    Citation Excerpt :

    Model was used with different input data and recalibrated using genetic algorithms. Interestingly, in the previous analysis [7], the value of V parameter, corresponding to contribution of light pollution from distant cities, was zero. In this analysis, the V parameter is no longer zero, but has value 0.02625 km.

  • A simplified model of all-sky artificial sky glow derived from VIIRS Day/Night band data

    2018, Journal of Quantitative Spectroscopy and Radiative Transfer
    Citation Excerpt :

    This relationship was expanded by Albers and Duriscoe [14] who derived a constant to convert from population to predicted artificial sky brightness for cities. Netzel and Netzel [15] used high resolution daytime satellite imagery classified as "percent of built up area" as a surrogate for escaped artificial light and Berry's empirical model to produce an estimate of sky brightness at zenith over the country of Poland at 100 m resolution with modest computer resources using GIS tools. The model was calibrated with observational sky brightness data from 65 sites.

View all citing articles on Scopus
View full text