Measuring and Modelling Rural skyglow Photometry of skyglow

Measuring and Modelling Rural skyglow Photometry of skyglow in rural areas: Introduction: The current situation in the UK and the problem of low angle unshielded illumination. 1, Some luminance profiles from rural locations. 2, Modelling luminaires, scatter, and skyglow C J Baddiley British Astronomical Association Campaign for Dark Skies.

NOAA data UK light pollution increase from 1993 to 2000 NOAA CPRE UK changes Map NOAA UK isophotic data for 1993 and 2000 Black 0 -1. 7. . can see MW when clear, Blue 1. 7 -50 MW visible occasionally, light blue 50 -150 can see MW very occasionally, yellow 150 -240. . And above. . never see the Milky Way, Red. . . instrument saturation 24 -255. The increase in the ambient lighting at the second isophot level in just 7 years is alarming. (NOAA/ CPRE)

Atmospheric light pollution Europe now and prediction for 2025 Europe prediction from DMSP data Computer modelled atmospheric skyglow isophotes for Europe, Left 1995. Right prediction for 2025 From a few stars only visible (red) to a few hundred (yellow to green), to near a thousand , some Milky Way visibility, to dark sky (black). Based on DMSP nadir data. (Cinzano P. , Falchi F. , Elvidge C. D. , Baugh K. )

Highways agency Highways Agency. Responsible for major roads and Motor Ways. HA has long since adopted Full Cut-off lighting on all new Motor Ways, Rural roundabout at Beckington, Somerset, on the A 36 between Bath & Warminster. before then re-lit with HCO lights (John Ball) Also found now on new roundabouts in open areas. NATA environmental impact appraisal. New Approach To Appraisal - a high priority for the effect of the environmental impact of lamps when choosing them. Is it working ?

ILE Zones light levels Luminaires and installation information

UK parliament S&T ctte and Gov initiatives 2003 The Science and Technology Committee inquiry into light pollution and astronomy. A lot of CFDS input written and aural evidence. • • Many recommendations, Central recommendation. . That there should be legislation against stray lighting. On 2004 Feb 12 th There was a Parliamentary debate on the report and its findings, at Westminster Hall. • A discussion in the House of Lords was held in mid June. ‘The Government is working on the issue at the speed of light’ (!) • ODPM to produce an annex to PPS 23, obliging councils to control external lighting through planning requirements. Re. PPS 23, Discussions are being held at the Office of the Deputy Minister between stakeholders and Government departments. Inc. ILE, RAS, and Cf. DS. DEFRA intention to make lighting intrusion a statutory nuisance. • •

Measuring rural skyglow Photometry of skyglow : 1, Some luminance profiles from rural locations. C J Baddiley Images by James Weightman, Mike Tabb and Bob Mizon

Summary of finding dispelling myths Sky-glow in rural areas, under atmospheric conditions where stars might be visible. Dispelling assumptions and myths……. · Sky glow is due to light sent directly upwards. No, it is not. · It is also due to reflected light straight up into the sky. No, it is not. · In order to control it we need to cut-out all directly upward light. No, we do not. · When we see skyglow on the horizon from a town, it is from the sky just above the town. No, it is not. · The reflected light is mostly from road surfaces. No, it is not. · It doesn't matter what sort of luminaire is used, scattering in the atmosphere guarantees it ends up as sky glow. Not true. · Switching to white light will reduce sky glow because such luminaries can be run at scotopically matched lower intensity levels. Not necessarily. · Shallow bowl lights luminaires cause less sky glow than fuller cut-off types, because less of them are needed per given road length. No, not convincingly.

Elevated view of LA Los Angeles in 1988 (IDA). Individual light sources dominate this scene, and the illumination of the sky behind the viewer

Elevated view of Belfast form an elevated location (Peter Paice).

Elevated view of Worcester (12 Km away) with its skyglow, the M 5 motorway across the Severn valley looking East, south of Worcester, from the Malvern Hills. (Chris Baddiley) What you see is direct light from many luminaires, right to the far side of Worcester. That is what is illuminating the sky and clouds behind the viewer in the same sight line in rural Herefordshire. All this is wasted light from just above the horizontal.

Direct and reflected rays diagram Air molecules Reflected Aerosols Radiated. Skyglow is caused by the downward scattering of upward light by air molecules and also aerosols, mostly water droplets and dust. The longer the path length through the lowest part of the atmosphere, the more the scattering. Light that goes straight up is mostly reflected, and has shorter paths through the lower scattering layers. The low angle light is mostly directly radiated, and it is that causes most of the sky glow well away from the source.

Visibility limit, star light and skyglow tables A table of stellar magnitude, number of stars per square degree that are visible to those limits and the total number of visible stars sky and total illuminance. Compared with sky glow figures in the lower table. Birmingham sky is about magnitude 17 per square arc second, which is magnitude -1 per square degree and for the whole sky is 57 millilux on the ground. Darkest Holland is magnitude 20 per square arc second and that is 3. 6 millilux. Sky glow is still exceeding the total starlight. In Birmingham it exceeds it by factor of 20.

Visibility limit map for Cotswolds Dark sky areas of the Cotswolds (P. Cinzano - Dipartimento di Astronomia Padova, Italy), Philips – Maps publication for Cf. DS. Commercial publication, 2004 September

Illuminated cloud base Cotswolds 2003/8/20 at location SO 967102 looking south (main glow Swindon 30+ Km SSE); Mars visible left. Sagittarius right. Illumination of clouds from underneath Canon D 60 digital SLR + 15 mm f/2. 8 lens. ISO 800 15 secs. (James Weightman)

Skyglow in Cotswolds, location SO 967102. Skyglow in the Cotswolds Western quadrant, composite transformed to equal area projection, CCD images (James Weightman) Milky Way, Scutum, Sagittarius etc taken last night 2003/8/2 at SO 967102 near Cirencester, Glos; Unguided Olympus C 5050 Z Skyglow in Cotswolds, location SO 967102. Western quadrant, composite transformed to equal area projection, CCD images (James Weightman) Milky Way, Scutum, Sagittarius etc taken last night 2003/8/2 at SO 967102 near Cirencester, Glos; Unguided Olympus Even in the countryside, there is no escape ! Note the gradation from the horizon upwards, due to the concentration of scattering aerosols at lower altitudes. The skyglow from very distant sources is still bight.

Analysis picture Skyglow in the Cotswolds CCD composite of 20 x 15 second exposures. Cotswold Hills, 6 miles south of Cheltenham. It can be seen, perhaps aided by the rising early morning mist, that a glow can be seen all around the horizon, particularly at 12 o' clock. Location SO 967102 (Cheltenham), 7 o'clock (Cirencester). By James Weightman, BAA Made from 12 images covering the whole sky taken with wide angle zoom 7 mm fl lens, setting of Casio QV 3500 digital camera over a period of approx 15 minutes.

Cotswolds Weightman image, photometry Intensity profiles of horizon to zenith lines as shown. The towns are from the top 12 o'clock Cheltenham, 7 o'clock Cirencester.

Cotswolds Weightman image, analysis The apparent elevations of identified stars is compared with true elevations and corrected Sky brightness Mag 18. 5 /sq arcsec at Polaris.

Cotswolds, Mike Tabb image Sky glow in the countryside, from distant towns from the darkest Cotswolds. The lens compresses objects to the horizon (see the tree), so the dark hole overhead appears actually smaller than shown. Location ST 829 743 (Mike Tabb). The towns, are clockwise from the top, Leigh Delamere, Bristol, Bath, Trowbridge, Melksham, Chippenham, Swindon

Cotswolds, photometry Mike Tabb image Intensity profiles of horizon to zenith lines as shown. The towns, are clockwise from the top, Leigh Delamere, Bristol, Bath, Trowbridge, Melksham, Chippenham, Swindon

Visibility limit map for Dorset Skyglow maps of the areas of the Darkest Hampshire and Dorset, The New Forest (National Park) (P. Cinzano - Dipartimento di Astronomia Padova, Italy), Philips – Maps publication for Cf. DS. Commercial publication, 2004 September

Sky glow from Poole, and ferry terminal Bob Mizon image

Poole photometry Intensity profiles of horizon to zenith lines as shown. The curves fit well, despite this being a limited field of view in which has a (Cos angle from the centre )^3 sensitivity law , and the gamma of the

Rural skyglow estimation summary Sky brightness lower limit estimates Cotswolds Weightman data Mag 18. 6 /sq arcsec at 52 deg elevation. Cotswolds Tabb data Mag 18. 2 /sq arcsec at 52 deg elevation. Poole Mizon data Mag 18. 2 /sq arcsec at 30 deg elevation.

Modelling luminaires and skyglow Photometry of skyglow : - 2, Modelling luminaires, scatter, and skyglow C J Baddiley

Luminaire distributions Luminaire polar distributions and ground reflections

Luminaires polar plot gamma for c=90/180 Luminaire light distribution profile for a SOX luminaire (deep orange), shallow bowl (light orange) and a full horizontal cutoff light (pink) of the same total luminosity

Luminaires polar plot gamma and C Skyglow modelling Luminaire light distribution profile for a SOX, (deep orange), shallow bowl (light orange) and a full horizontal cutoff light (pink) of the same total luminosity Note the radiance at horizontal in azimuth of the LPS SOX luminaire shows its selfobscuration. 27% upward light ratio.

Skyglow modelling Direct and surface reflected rays diagram Direct view radiated. Reflected All direct light above horizontal not blocked. Same view angle reflected Reflectivities: at 589 nm wavelength Shallow angle reflected Soil 0. 1 light blocked. Grass 0. 08 At shallow angles there is a reflected component that is enhanced by the increased reflectivity of with incidence angle. Surfaces near grazing incidence are blocked by hedges, buildings terrain etc. There is also the direct upward component. Foliage 0. 06 Asphalt 0. 04 to 0. 08 (dirty) Concrete 0. 25

Surface reflectivity vs angle Reflectivity as a function of angle of incidence to normal. Reflectivites of surfaces increase towards grazing incidence. Smooth surfaces go to unity. The Brewster angle is before this which blocks the horizontally polarised components. Most surfaces, even roughness are very reflective beyond 70 degrees, i. e. below 20 degrees to the horizontal. Low angel light is reflected, smooth ones become a mirror. Grass reflectivity 0. 1 Rises to 0. 2 at 80 degrees, asphalt goes from 0. 04 to 1 at grazing angle, as does water

Surface reflectivity spectrum Reflectivities: at 589 nm wavelength Reflectivities vary with wavelength. Grass is less reflective in yellow. Switching to white light increases its value. The minimum reflectivity is at about 670 nm where light is absorbed by photosynthesis, in common with all vegetation. Reflectivities rise rapidly in the near infrared for thermal rejection. Soil 0. 1 Grass 0. 08 Foliage 0. 06 Asphalt 0. 04 to 0. 08 (dirty) Concrete 0. 25

Luminaires reflection and upper radiance polar plot gamma for c=90/180 Luminaire light distribution profile showing downward and sideways direct component and reflected components profile for a LPS SOX (deep orange), shallow bowl HPS SON (light orange) and a HPS SON full horizontal cutoff light (pink) of the same total luminosity Elevation at azimuth 0/180,

Luminaires reflection and upper radiance polar plot gamma for c Skyglow modelling Luminaire light distribution profile showing downward and sideways direct component and reflected component, lumaires as before. Top line elevation view for orthogonal azimuths, bottom line horizontal azimuthal plot left and same but at 110 nadir, (20 degs above horizontal) right. Elevation at azimuth 0/180, Elevation at azimuth 90/270, Azimuth plot at horizontal Azimuth plot 5 deg above horiz. Azimuth plot 20 deg above horiz.

SOX LPS SOX reflection and upper radiance polar plot gamma for c Skyglow modelling Radiated (orange), reflected (green), total (red) This frame shows the polar plot for the reflected and direct upward radiation in the vertical plane along the road, for the three light Elevation at azimuth 0/180, Elevation at azimuth 90/270, types previously described. The low pressure sodium SOX light source has a dominant direct radiated component which exceeds the reflective component at low angles to the horizontal. Reflected components have been blocked below 15° above horizontal as previously described Azimuth plot 5 deg above horiz. Azimuth plot 20 deg above horiz.

Shallow bowl reflection and upper radiance polar plot gamma and C Skyglow modelling Shallow bowl SON Radiated (orange), reflected (green), total (red) SOX LPS Elevation at azimuth 0/180, Elevation at azimuth 90/270, The SON cases above the horizontal are all reflective. The reflection pattern is nearly a mirror image of the ground luminance distribution, convolved with the scatter distribution on the ground surface. The asymmetry in the across road reflection is due to the rd and verge boundary. Azimuth plot 5 deg above horiz. Azimuth plot 20 deg above horiz.

HCO reflection and upper radiance polar plot gamma and C Skyglow modelling HCO SON Elevation at azimuth 0/180, Elevation at azimuth 90/270, Radiated (orange), reflected (green), total (red) The asymmetry in the across road reflection is due to the rd and verge boundary. Azimuth plot 5 deg above horiz. Azimuth plot 20 deg above horiz.

Integrated luminance tables Luminance integrated over a sphere, and upper hemisphere, direct, reflected and both, without and with obstacle blocking of reflected rays below 15 degree elevation. Note both the direct upward figures are high for the SOX luminaire. The SOX reflection value is most reduced by the 15 degree elevation obstacle blocking due to its wider spread on the ground compared with the others.

Atmospheric scattering

Atmospheric scattering pictures Daylight Rayleigh scattering by air molecules, smaller than the wavelength of light. Equal forward and backward scatter, also sideways, Varies as 1 / wavelength 4 Blue haze Rayleigh scattering by particles smaller then the wavelength of light Mie scattering by aerosols. . water droplets and dust, similar or larger than the wavelength of light. No wavelength dependence and very directional.

Sunset scatterring simulation Sky scatter from Sun (sun omitted) with elevation Rayleigh scatter air molecules Mie scatter from aerosols This is a sequence of a simulated sunset. The Sun has been removed Total Scatter from this picture, just Sun showing the light scatter. Elev. 5 o 0 o -2. 5 o -10 o

Scatter density with altitude The next plot shows the variation of density of air molecules and aerosols, as a function of altitude in the atmosphere. At 10 km altitude the density off the air has reduced to 2/3 of its ground value. The equivalent height of the total atmosphere brought to constant density is only a few Km.

Molecular scatter probability for scatter angle (phase function). Rayleigh scattering. Light is coming in from below and being scattered in the direction of the grid angles. The curve gives the probability of scatter in that direction. Equal probability forwards and backwards, 50% of that sideways. The probability over all angles =1 It varies in intensity as wavelength ^4 (blue biased). It is why the sky is blue by day

Aerosol Mie scatter probability Aerosol scatter probability for scatter angle (phase) functions (Mie scattering) Dependence on particle size (Heye-Greenstein function with added back scatter). The forward scatter is very peaked, increasing with particle size from 1 nm to 10 microns The lower scatter probability profile is the one used. There is practically no sideways scatter and back scatter is tiny. No wavelength dependence, it is why clouds are white

Incremental and cumulative scatter The proportion of light scattered out of the beam is proportional to the length of the increment. (proportional constant k) The input of the next increment is the output of the previous one. This is the starting differential equation that has the solution of an exponential accumulated scatter with distance. What is scattered out of the beam is replaced by what is scattered into the beam. They have to be the same over the path length. So 1 - the negative exponential is the scattering into path.

Path geometry Geometry Path geometry

Path geometry = elevation z Re = Earth radius h h = height of point z = total slant path k = scattering coefficient Km-1 The path distance x is replaced with a function of elevation. The hypotenuse of the triangle shown. Maximum path length are at the lowest elevation. The negative exponential of (k x path length) is then taken, but the path density decreases with height

Curved Earth geometry ’ r’ r h ’ ’ s’ s Re d d’ g’ g Re t Re ’ q

Results Photometry of skyglow Result plots … Incremental vs. distance along a given elevation path Total for vs. elevation at a given distance Total for a given elevation vs. distance

Skyglow at high elevations. All large scatter angles, very few molecules up here Air molecules scattering at these large angles Ground reflection All large scatter angles source angle Aerosols here but don’t scatter at much at these high angles, only direct radiation View elevation Direct radiance (SOX) Skyglow at high elevations. As all scatter is at large angles it is mostly by air molecules, maximally in the blue. Little is from the upper parts of the path, as there the scattering is near orthogonal and the aerosol density is much lower.

Skyglow at low elevations. Air molecules dominate scatter at these high angles, but only at a low level. Peak of scatter is here, aerosols dominating Large scatter angles Aerosols dominating at low angles, forward scattering of direct radiance Ground reflection source angle Direct radiance (SOX) Small scatter angles View elevation Skyglow at low elevations. Much of the path has small scatter angles, so is predominantly aerosol single Mie scattering, for particles >50 microns in diameter. The scatter is strongly in the forward direction. The peak contribution is forwards of the source and is from aerosols. Little is from the upper parts of the path, as there the scattering is near orthogonal and the aerosol density is much lower.

Incremental scatter plot for 45 deg elvn path Over source (15 Km horiz, 17 Km at 30 degree elevation view path) Aerosol scatter peak contribution To viewer The brightness contribution to a 30 deg elevation view path, scatter at each point along the inclined path in the direction of a source (15 Km horiz. , 17 Km inclined path). Reflection for grass normal incidence reflectivity 0. 1. Obstruction of reflected rays up to 15 degrees. Luminaires types SOX LPS (orange) and horizontal cutoff SON (pink). Scatter just from aerosols (dots) and molecules (lower thin curves) and combined (thick lines). Aerosols dominate at low elevations. Molecules take over just above the source.

Scatter into line of sight for all elevations All large scatter angles, scatter from molecules only Ground reflection Scatter peak from aerosols All large scatter angles, scatter from molecules only but at low level Direct radiance (SOX) Forward scatter from aerosols View elevation Skyglow at all elevations, the sum of all the scattering along the path. Scatter at low angles is from aerosols. Scatter at large angles it is mostly by air molecules, maximally in the blue. Little is from the upper parts of the path, as there the scattering is near orthogonal and the aerosol density is much lower.

Total line of sight scatter vs elevation view. Thin lines represent the scatter due to air molecules, while the dots are those from aerosols. The broad lines are the combination. The case of the low pressure sodium and the horizontal cutoff lights are shown, together was a reference curve fit to one of the measured data sets discussed earlier. For the purpose of this calculation sixteen light sources were evenly distributed about the horizon representing the contribution from a number of towns, here The relative brightness from integrating along the full slant path from 16 ground sources equialso to 15 km distance. spaced about the horizon, 15 km away, for elevations from horizontal to the zenith. A crude correlation function was also Luminaires types SOX LPS (orange) and horizontal cutoff SON (pink). (obstruction of reflected rays applied to cover the up to 15 degrees, normal incidence reflectivity 0. 1). Scatter just from aerosols (dots) and molecules effect of multiple (lower thin curves) respectively. Aerosols dominate at low elevations. Red thin curve is a scattering. measurement. Scatter components into line of sight for all elevations

Scatter for alll luminaires into line of sight for all elevations The relative brightness from integrating along the full slant path from 16 ground sources equispaced about the horizon, 15 km away, for elevations from horizontal to the zenith. Luminaires types SOX LPS (deep orange) , shallow bowls SON (light orange), horizontal cutoff SON (pink), horizontal cutoff white (grey) and shallow bowls white (pale blue). The relative brightness from integrating along the full slant path from 16 ground sources equispaced about the horizon, 15 km away, for elevations from horizontal to the zenith. Luminaires types SOX LPS (deep orange) , shallow bowls SON (light orange), horizontal cutoff SON (pink), horizontal cutoff white (grey) and shallow bowls white (pale blue).

Scatter for grass and snow comparison Reflection off grass as before, and reflection off snow, increases the brightness from the HCOs and well directed types, as it is all from ground reflections

Diagram of scatter into line of sight, at 45 degs elevation, as function of distance from the source. Air molecules scatter but there are few of them Ground reflection Scatter from Aerosols dominate 45 deg. View elevation Direct radiance (SOX) Skyglow at 45 deg. elevation, as a function of distance form the source.

Scatter into line of sight, The relative brightness at 45 degs elevation, as function of distance from the source. contribution of scatter at each point along a 45 degree elevation path in the direction of a source as a function of source distance. Reflection for grass. Luminaires types SOX LPS (orange) and horizontal cutoff SON (pink). (obstruction of reflected rays up to 15 degrees). Scatter just from aerosols (dots) and molecules (lower thin curves) respectively. Walker’s law has the brightness fall as 1/distance^2. 5 (a constant slope on this plot), but here the slope increases with distance. The relative brightness contribution of scatter at each point along a 45 degree elevation path in the direction of a source as a function of source distance. Reflection for grass. Luminaires types SOX LPS (orange) and horizontal cutoff SON (pink). (obstruction of reflected rays up to 15 degrees). Scatter just from aerosols (dots) and molecules (lower thin curves) respectively. Walker’s law has the brightness fall as 1/distance^2. 5 (a constant slope on this plot), but here the slope increases with distance.

Table of skyglow cause and effect Table of dominant and secondary scattering for high and low elevations

Simulation and Summary

The next frame follows the Light Pollution from different luminare types simulation Sky scatter from type of luminaires (light source omitted), from horizon upwards. same simulation scheme Rayleigh scatter air molecules Mie scatter from aerosols as was the sunset Total Scatter sequence. but here we are NOT just looking at the effects of the different types of HCO light previously discussed. LPS Again from left or right we have Rayleigh molecular scattering, Mie aerosol scattering, and the Shallow combination. From top to bowls bottom we have luminaire types low pressure sodium, shallow bowl, and HCO full cut off, then HCO HPS white, and shallow bowl white. This is only indicative HCO white Shallow bowls white

Summary • Skyglow in the countryside is dominated by stray light from towns, even 30 Km away. This is still mostly from road lights. Beams from unshielded luminaires just above the horizontal travel though long path lengths and it is that is mostly responsible for rural skyglow in rural areas. • These low angle beams of light that and are scattered by aerosols in the forwards direction within a few degrees. They illuminate the clouds and cause sky glow at great distances, even 30 Km away. • Air molecule scattering only dominates at high elevations • The reduction in skyglow from adopting horizontal cutoff lighting in all areas outside of town centres can be a factor of 10 at 15 Km. • Shallow bowls cause more skyglow then HCOs due to the larger ground area of illuminance at angles closer to horizontal and this is reflected into the sky. They are visually more obtrusive too. • If white light is adopted rather than HPS, then the skyglow increases slightly though Rayleigh scattering by the air at high elevations. This reduces some of the advantage in running them at lower intensity levels for acceptable visual scotopic perception, but only slightly.

Concluding points • Glare and skyglow can be reduced in the countryside by adoption of well controlled shielded lighting. • Big benefits local to the lights from cutting out glare and controlling light spillage. Especially from sports facilities and amenities. • The best way to reduce skyglow is the adoption of HCOs in preference to other types, even if this means more per road length. But all measures are needed, such as dimming, as is now being tried, and switching off when not necessary. • Government funds are available for improved quality lighting. But there is more and more lighting for new housing and amenities. using HCOs alone may not reduce sky glow sufficiently, against this growing amount of lighting in the UK. • There is soon to be a Government requirement on all Councils to have lighting control measures in their planning requirements (ODPM PPS 23). • The sale of domestic high power intensity ‘insecurity’ lights may also be restricted or banned (DEFRA consultation).

Ikeya Zhang Comet Ikeya Zhang in skyglow. 2002. (anon) Comet Ikeya Zhang and tree from a dark site. 2002 (Chris Baddiley).

Hale Bopp Two pictures under dark sky conditions, comet Hale Bopp (50 mm lens), and The Milky Way Cygnus region (35 mm lens) , . . CJB

Milky Way from La Palma Milky Way Cygnus and North American nebula from La Palma (Chris Baddiley).

Information Addresses : The British Astronomical Association Campaign for Dark Skies, Burlington House, Piccadilly, London, W 1 J 0 DU. Tel +44 120 7734 4145 Cf. DS Website mizar-astro. freeserve. co. uk or. The Coordinator, Cf. DS, Bob Mizon, 38 The Vineries, Colehill, Wimborne, Dorset BH 21 2 PX Email: For the CD ROM image library …. of good and bad lighting and its effect on the environment…(roads / security/ amenity/ building floodlighting / Towns and countryside skyglow, views for space, information and resources. …. compiled by Chris Baddiley, Email: Home