Artificial lighting at earth contributes via tropospheric scattering to the night sky brightness over a large area around the source of light. Both a continuous component as well as distinct emission lines are present in the light pollution spectrum. A recent review of sky pollution is given in McNally (1994).
'').
With this distance one can enter Fig. 32 (click here) and
obtain a scaling for the (arbitrary) intensity axis of this figure.
Thus the artificially
caused sky brightness at 45 deg altitude at 
distance from
the city can be estimated from this figure.
Figure 31: Variation with city population of the distance at which the
lights of a city produce an artificial increase of the night sky brightness
at 45 deg altitude toward the city by 0.20 mag. This increase refers to an
assumed natural sky brightness of V = 21.9 mag/''.
Observations by Walker (1977) are indicated
by dots. Two models by Garstang (1986) are shown as
solid lines. K is a measure for the relative importance of aerosols for
scattering light. The uppermost dot refers to Los Angeles County, the cross below
it to Los Angeles City. From Garstang (1986)
Figure 32: Variation with distance from the city of the sky brightness
at 45 deg altitude in the direction of the city.
The dots indicate observations in V band by Walker
(1977) near the city of Salinas. The solid curves
are according to models by Garstang (1986).
The brightness ratio is defined as
, where b
= sky brightness. Zenith distance +45
is towards and
away from the city. The solid curves are according to models
by Garstang (1986). Curve 1: L0 = 986 lumens per
head, K =0.43, 
. Curve 2: L0 = 1000 lumens per head, K
=0.5, 
. L0 is the artificial lighting in lumens produced per
head of the population. K is a measure for the relative importance of
aerosols for scattering light. F: a fraction F of the light produced
by the city is radiated directly into the sky at angles above the horizontal
plane, and the remainder (1-F) is radiated toward the ground.
The dashed line is the relation 
. From Garstang
(1986)
Figure 33: Zenith distance dependence of sky pollution light according
to the model calculations of Garstang (1986). Results
are for sky pollution due to Denver as
seen from a distance of 40 km in the vertical plane containing the observer
and the center of Denver. Curve 1: sky background; Curve 2: Denver only;
Curve 3: Denver and sky background. Negative zenith distances are away
from Denver. From Garstang (1986)
Treanor (1973) and Bertiau et al. (1973) have used an empirical formula, based on a simplified model of the tropospheric scattering, to fit the sky pollution observations near cities. Garstang (1986, 1989a,b, 1991) has used radiative transfer models including 1st and 2nd order Rayleigh and aerosol scattering, effects of ground albedo and curvature of the earth's surface, and the areal distribution of the light source to calculate the sky pollution light intensity. He has compared and scaled his model results against the above mentioned observational results. Garstang's fitted models are shown in Figs. 31 (click here) and 32 (click here). superimposed on the observational points of Walker (1977). Garstang (1986, 1989a,b) gives also the calculated zenith distance dependence of the sky pollution light intensity both towards and away from the source of light. These results are reproduced in Fig. 33 (click here).
The emission line spectra of the different types of street lamps are
visible in the night sky light even at good observatory sites, such as
Kitt Peak in Arizona. While the most commonly used street lamps until
the 1970's were filled with Hg there has been since then a general
change over to sodium lamps, both of the high pressure (HPS) and low
pressure sodium (LPS) types. The most important sky pollution lines
are given in Table 14 (click here) according to Osterbrock et al. (1976),
Osterbrock & Martel (1992) and Massey et al.
(1990). At good sites (e.g. Kitt Peak), the strongest pollution
lines are about a factor of two weaker than the strongest airglow lines.
The opposite is true for strongly contaminated sites (e.g. Mt Hamilton).
Whereas the pollution lines are normally restricted to a relatively narrow
wavelenth range the Na D line wings produced by the HPS lamps are extremely
broad, extending over 
. Thus the LPS lamps are highly
preferable over the HPS ones from the astronomer's point of view.
Other studies of the night sky spectrum, including the artificial pollution lines, have been presented by Broadfoot & Kendall (1968) for Kitt Peak, Turnrose (1974) for Mt. Palomar and Mt. Wilson, and Louistisserand et al. (1987) for Pic du Midi.
| Line | Sources |
| Hg I 3650 | Hg lighting |
| Hg I 3663 | Hg lighting |
| Hg I 4047 | Hg lighting |
| Hg I 4078 | Hg lighting |
| Hg I 4358 | Hg lighting |
| Na I 4665, 4669 | HPS |
| Na I 4748, 4752 | HPS |
| Na I 4978, 4983 | HPS |
| Na I 5149, 5153 | HPS |
| Hg I 5461 | Hg lighting |
| Na I 5683, 5688 | HPS, LPS |
| Hg I 5770 | Hg lighting |
| Hg I 5791 | Hg lighting |
| Na I 5890, 5896 | HPS, LPS, airglow |
| Na I 5700 - 6100 | HPS |
| (wings) | |
| Na I 6154, 6161 | HPS, LPS |
| K I 7665 | HPS, LPS |
| K I 7699 | HPS, LPS |
| Na I 8183 | HPS, LPS |
| Na I 8195 | HPS, LPS |
. The most intense features are shown in
boldface