Are LEDs good for dark skies?
Unfortunately, no… they could be, but as they are presently being used, they are almost always not. Not even close.
increase from a high-pressure sodium to conversion.|
(For the much more dramatic increase with an to conversion, see here)
But how can that be?
In recent years the press and the U.S. Department of Energy are effusive about the revolution in lighting being brought about by LEDs. In remarks in a press release announcing the 2014 Nobel Prize in physics (awarded to researchers who invented the blue that underlies the ultra-bright white LEDs) Nobelprize.org states:
New light to illuminate the world
This year’s Nobel Laureates are rewarded for having invented a new energy-efficient and environment-friendly light source – the blue light-emitting diode (). In the spirit of Alfred Nobel the Prize rewards an invention of greatest benefit to mankind; using blue LEDs, white light can be created in a new way. With the advent of lamps we now have more long-lasting and more efficient alternatives to older light sources.
And in their announcement of the Prize, the International Dark-Sky Association notes:
If we light properly we can use LEDs to save energy, improve visibility, and lowerlevels.
The DOE, power companies and other governmental agencies have provided substantial financial assistance to communities in replacing current roadway lighting with Los Angeles, which has replaced over 140,000 high-pressure sodium streetlights with . The Los Angeles Bureau of Street Lighting reports that energy use has been reduced over 60%, saving $7.5 million per year; maintenance costs are predicted to fall by $2.5 million per year.. A typical high-profile example is
|Los Angeles, Ventura Blvd|
|Before retrofit||After retrofit|
So – the buzz is that LEDs promise not only huge energy savings but also reduced. How can we claim otherwise?
Looking closely at the facts and numbers, a different story emerges. Regarding energy savings the story is not as simple – and not as dramatic – as described (see below). Regardingthe conclusions are clear, and not positive.
For , the conclusion is unambiguous: research shows that white light – of any kind () – has substantially greater impact on than yellow light (see our page Lamp Spectrum and ). Compared to the currently prevalent high-pressure sodium lights, even the lowest-impact white LEDs (though higher-impact LEDs are far more widely used) will increase visible lumen-for-lumen 2.2x – more than double. Compared to the increase is over 3x, and compared to the most night-friendly light, low-pressure sodium or , the increase is 6x. This is not 6% nor even 60% – it is 6x or 500%! The most optimistic predictions for decreased lighting amounts that might be achieved with LEDs (due to better optical control) cannot compensate for this.
We examine twostreet lighting retrofit projects below. The first – Los Angeles, California – replaced with 4000 K white , and with modest illumination reductions. The LA example is typical of most retrofits.
The second example – Tucson, Arizona – replaced a mixture ofand with 3000 K white . Total lumen amounts and average illumination levels were more aggressively reduced. The Tucson example shows about the best that can be done without using amber LEDs, and as far as we know the magnitude of the Tucson lumen reduction has not been matched in any other community.
In the Los Angeles retrofit, the LEDs used (4000K) cause, lumen-for-lumen, about 2.7x (or 170%) more than the lighting they replaced. Even when total lumen amounts are reduced proportional to the illuminance reduction seen on Ventura Blvd (41%1 – see figures on the images above), the retrofit has still increased from streetlighting by about 60%2. And in a fair equal-illuminance comparison, the system has about 170% greater impact. Again, more than double.
1 The needed lighting level, and no more, should always be used – with any lighting technology. It is hard to understand why LA had Ventura Blvd illuminated to 3.99 fc, as that level is higher than any professional lighting recommendation.
2 (The actual changes will be smaller due to “dilution” by the other lighting in Los Angeles, which was not changed. See the discussion below under the Tucson retrofit.
In the Tucson retrofit, the LEDs used are 3000 K Barentine et al., 2018 – but given its small magnitude it is unsurprising that definitive results were not obtained.), which cause lumen-for-lumen about 2.1x (or 110%) more visual than , but 5.4x (or 440%) more than . But here the total amount of installed “fixture” lumens was reduced from 481 mega lumens to 179 mega lumens3 – a 63% reduction. Taking both the “step backward” effect of the spectrum change and the “step forward” lumen reduction, we estimate the visual arising from street lighting only in Tucson has been decreased by 21%. Due to the dilution effect of the other urban lighting that has not been changed, we estimate that this translates to an actual overall visual sky brightness reduction of 5%. (An attempt to measure this change is reported in
While this is an improvement (though it would be visually imperceptible to most people), by choosing white(even if 3000 K ), Tucson has forgone the possibility of an overall reduction more than 3X as great – 18% – if they had used a good . If they had used good , matching the beneficial spectral characteristics of their former lighting, the reduction would have been 21%. If Tucson was willing to reduce illumination levels to this extent, it is hard to understand why they would sacrifice their previous amber lighting standards and thus leave substantial benefits on the table.
We don’t want to diminish what Tucson has accomplished – in the world of whiteretrofits, they are an unprecedented example. But they have degraded their amber lighting standards and missed an opportunity for a substantial improvement in their night skies. This degradation of the formerly world-leading Tucson amber lighting standards will mean that other lighting users, including the tremendous number of retrofit projects that are ongoing, will now likely follow the lead of the City and go to low- white . These other users, accounting together for a much greater lighting amount than on roadways, will be much less likely to pursue significant lumen reductions. It is hard to see how Tucson skies will not continue to degrade, even without growth in the city and concomitant growth in lighting amounts.
3 The needed lighting level, and no more, should always be used – with any lighting technology.
LEDs and energy savings
Though LEDs offer opportunities for modest energy savings, it appears in many of the most-publicized examples energy savings arise primarily because lighting levels are being reduced. In the example above on LA’s Ventura Blvd, average illuminance was reduced from 2.99 to 1.76 fc – a 41% reduction. That’s a real energy benefit, but most of it comes from reduced lighting levels (available using any light type!), not from increased energy efficiency of LEDs. If LA had reduced light levels 41% while retaininglamps (or with an -like spectrum), they would have reduced energy use but also reduced from the streetlighting by 41%, instead of increasing it 60%.
This is not a characteristic of just the LA retrofit, either. An analysis by Kostic et al. and published 2013 in Lighting Research and Technology concludes:
[T]he average energy savings when usinginstead of high-pressure sodium ( ) luminaires amount to 19–26% for single-sided, staggered and opposite layouts, although they are frequently negligible if mesopic effects are not included. The total costs of the lighting solutions, even including mesopic effects, are 1.36 to 6.44 times higher than those of the comparable lighting solutions. Therefore, LEDs are questionable for street and roadway lighting.
Tucson has demonstrated a more aggressive approach, but still left significant benefits on the table by switching to whiteinstead of an amber , and Tucson had more to lose due to their now abandoned amber lighting standards.
So the conclusions are clear – switching to white LEDs from the current yellow light technologies using sodium lamps saves only a little energy, much less than is being touted, unless lighting is dramatically reduced. Butin the vast majority of retrofits is dramatically increasing.
- First and foremost, use yellow LEDs (not low- white LEDs)! with a spectrum similar to (0.4x the impact of ), or with a spectrum similar to (0.7-1.6x the impact of ) should be used, or at worst filtered warm-white LEDs (1.5x the impact of ). These are not the most efficient LEDs at present, but can be a good tradeoff to balance energy and dark sky impacts.
- Take advantage of dimming technologies to dim or turn off lights when activity or traffic levels justify it.
…And for any light type:
- Only light roadways when a full assessment of costs (both fiscal and environmental) and benefits, including fiscal, environmental, and safety, shows overall benefits.
- Always use fixtures.
- Use the lowest illumination level possible.
Come to for the best in dark skies information and recommendations!