LED Astray

Technological innovation has cut many of the trickiest environmental Gordian knots. As readers of SuperFreakonomics know, at the end of the 19th century the American city was tottering on the brink of environmental catastrophe. The cause: byproducts of transportation, specifically the horse-drawn variety (more here). Manure, flies, accidents and disease abounded, until the timely arrival of a technology which at the time was hailed as an environmental savior — the internal combustion automobile.

Of course, those who celebrated did not fully realize that the miracle technology would bring troublesome consequences of its own. Today we face the specter of global warming, to which the automobile is an important contributor.

Just like our ancestors in the fin-de-siecle city, we are using new technology to fight the shortcomings of the old. And unsurprisingly, like our ancestors, we are bound to find that our technological fixes can have unintended, malign consequences of their own.

Consider the seemingly noncontroversial replacement of conventional traffic signals with LED bulbs. What’s not to like?

LEDs last up to ten times longer than the old bulbs, saving on equipment and labor costs. They drastically cut energy consumption. (One Illinois jurisdiction found that the conversion of an intersection reduced electricity bills from over $60 per month to under $10.) Finally, LEDs are brighter than conventional bulbs, increasing visibility. Pretty cool, huh? Exactly – and that’s the problem.

The biggest weakness of LEDs is their biggest strength – they don’t radiate much heat. What on earth could be wrong with that? Depends on which part of the earth you inhabit. In the upper Midwest, LEDs can have deadly consequences.

LEDs’ energy conservation creates a problem in case of a – literal – perfect storm. Low temperatures, wet snow, and driving wind can coat – and obscure – traffic signals. Traditional bulbs throw off heat which melts the snow. LEDs don’t. The result is intersections without visible traffic lights which are hazardous and sometimes deadly, as was recently the case in Chicago.

None of the proposed solutions are foolproof. Heating up the LEDs would help defeat the purpose of saving energy. Lights can be cleaned manually, but with about 2900 traffic signals in Chicago alone, finding, dispatching teams to, and cleaning obscured lights is a fairly big task.

We shouldn’t overstate this problem. It takes just the right conditions to occlude lights, and many times the situation resolves itself. When lights are out, motorists should know to exercise caution and treat them as if they are stop signs.

However, to this point dozens of accidents have been attributed to the LEDs. Is it worth it? Probably, but that is little consolation for those hurt and killed in these crashes.

Perhaps new technological solutions will be found, as has happened to an extent with the automobile. In the mean time, hopefully motorists in colder regions will lengthen their lives by adjusting to this new hazard. Or by moving from Chicago to LA, the way I did.

(They’d have to give up real Chicago dogs and pizza, but this would undoubtedly lengthen their lives as well. Whether a life without these delicacies is worth lengthening is another question.)

Despite the crashes, departments of transportation have no intention of giving up LEDs, any more than their 19th century counterparts would have considered giving up the internal combustion engine. As with the auto, some consequences, even deadly ones, will have to be lived with.

This may seem callous, but transportation often calls for tough decisions that involve sacrificing human life. Surprisingly, these often receive very little public scrutiny compared to more conspicuous life-and-death issues like the death penalty or the wars in Iraq and Afghanistan. More on one such overlooked transportation question coming soon.

(Hat tip: Gary LaBelle)


This is pretty funny since hand held devices used by drivers cause far more deadly accidents than LED lighting...by the way the LED products do not leach toxic mercury into the water supply.

If people paid attention to safe driving habits and observed rules of the road and common courtesy, we may not need all the signal lights and distracting signage people require so they know how to behave behind the wheel.

Rupert Goodwins

Also, hyperbright LEDs have saved many lives in automotive accidents. Because they turn on instantly and don't need the 300mS or so that an incadescent filament needs to warm up, when used as brakelights they give following drivers an edge, especially with the eye-level lights that became common as LEDs spread. It's not many lives saved - one estimate is around fifty a year, worldwide - but the inventor of the hyperbright LED is on record as saying that this is the single most satisfying aspect of his work.

Not sure what he feels about frozen traffic lights.



Just add a wiper, like a windshield wiper.

J Omega T

Seems to me this calls for an application of Einstein's exhortation to the effect that things should be made as simple
as possible, and no simpler.

In places where heat as well as light is desirable, some of the time, simply fashion a hybrid signal, with a thermostatic
control that turns on the [primarily] heat radiating elements
only when needed. For the runway application, why cannot
the lighting units be arrays of both visible and infra-red emitting devices?

And anyway, while I was on a tour of a prominent lighting research center (admittedly this was about 8 years ago), the scientists indicated they were not ready to release their LED traffic signals because they got too hot. Apparently they sovled this problem. But, I did wonder at the time how the LED arrays were going to save electricity, if the efficiency ratio of light to heat (or visible to infra-red) was so low...

David Leppik

Here in Minnesota, we've had LED traffic signals for years. I started noticing them a decade ago. As far as I can tell, they don't seem to get occluded significantly more than traditional lights. Perhaps the folks in Chicago should find out what (if anything) the Minnesotans are doing differently.


LEDs are much brighter than incandescent lights... surely that has prevented accidents where with the old lights, drivers would not know which light is on.

Then there's the task of figuring out how much less air pollution is generated by switching over to LED lights (don't need to burn as much coal, after all.). I suspect you'll come out roughly even in terms of overall effect to human health.


Please. LED traffic lights are not the only ones that have problems in heavy snow storms. I grew up in northern Wisconsin and the Upper Peninsula of Michigan. Heavy snow storms and/or blowing snow often blocked the lenses of the old incandescent traffic lights as well. Everyone knew to be extra careful and alert for traffic under these conditions.

Further, installing heaters or other such devices to clear the snow is unnecessary - assuming that we really need to do anything besides take responsibility for our own safety. A redesign of the lens to deflect the snow rather than allowing it to accumulate is all that is necessary - and then only in areas where the problem may occur.

the Gooch

Professor Victor Petrenko at Dartmouth has come up with electrical de-icing technology that doesn't depend on heating the ice to melt it (changing the state of water is energy intense), but by changing the adhesive properties of ice using low power electric fields. This seems like a good possible application.


the Gooch

"If you can't see the traffic light or it is not functioning, you are supposed to treat the intersection as an all-way stop."

Vincent, in most cases it would only be 1 or 2 directions of the stoplight that would be covered in snow due to the prevailing wind.

So if you can't see the light and are treating it as a 4-way stop, chances are one of the directions perpendicular to you has no problem seeing their light and will be cruising through at full speed.


My dad is color blind and can't see the green LED lights. Thankfully they haven't switched positions on him, so he knows which one is green. Just another thought to add to the discussion.


"Perhaps new technological solutions will be found, as has happened to an extent with the automobile. In the mean time, hopefully motorists in colder regions will lengthen their lives by adjusting to this new hazard. Or by moving from Chicago to LA, the way I did."

Yes, let them eat cake.

Kai Howells

LEDs don't emit heat – this is a myth.
LEDs actually emit quite a fair bit of heat, and it's in proportion to their current draw. They are more efficient, and turn more of the incoming electricity into light, but there's still a fair bit of wasted electricity that's converted to heat.

It's where the heat is emitted that's different to incandescent lights.

If you have any LED light fixtures, look at them closely - there will be a decent sized heatsink on the rear of the unit. This is where they differ from regular lights that pour heat out the front, LEDs (by design) emit their heat from the rear.

It would be a relatively simple engineering exercise to channel this heat back out the front of the units in cold climates.

chaz larson

Heaters? Electric fields? So much complexity and over-engineering. More points of failure.

The lights already have hoods. Put a glass window in the hood, sloping from the base of the light to the tip of the hood. Show and ice can't collect on it. Maybe vent it so the waste heat from the back of the LED array can seep into the enclosed space for some warming.

chaz larson

I should add; or, better yet, rebuild the intersection as a roundabout. No lights to maintain or de-ice. Just yield signs which, fortunately, keep their shape when iced over.


#25: The snow is colder in Minnesota and therefore less sticky.

There are ~40000 traffic fatalities in the US each year, so perhaps 0.01% are due to this problem.

However, since there is a simple and cheap technical solution there is not reason to to implement it:

On the shroud over each light, a number of LED/photodiode pairs should be place. The LED can be blinked once every few seconds. The photodiode will detect significanly more light during the blink if there is snow packed into the shroud. If the controller detects packed snow it turns on a heater to melt it.

Even though this may sound complex, the parts required are cheap, especially compared to the cost of the traffic light as a whole.


LED stop lights can easily be redesigned to shed snow. The problem is using LEDs as replacement bulbs without modifying the lamp.

Overlooked in your equation is the extra pollution created to power conventional incandescent or halogen lights. An estimated 10,000 to 30,000 US citizens die every year because of pollution emitted by coal power plants. But air pollution and the people killed by it is merely an externality according to economists.


The first problems you stated -- manure, disease, global warming -- all have an environmental impact. The problem of crashing due to obscured lights is more of a problem of human fragility, and actually benefits the environment by slowing overpopulation. Just sayin'...


We should bring the lights in out of the cold. I propose installing wireless access points at intersections. The WAPs will broadcast appropriate information that can be displayed on a small panel on the dashboard.

This seems like a lot of work, but it's a great step towards allowing cars to automatically negotiate intersections. It would also require much less energy and maintenance than even LED lights. As an added bonus, we could completely get rid of the huge ugly poles.

The WAPs could also be configured to give much richer information than a standard light: traffic and road conditions, speed limits, or any other info the DoT would want to broadcast.


One problem it appears (I think) people are missing is that they spent all sorts of real money replacing the old lights with LED ones.

While it's easy to imagine a technological fix, it seems quite likely that that will cost extra money to implement (and take a fair amount of time too).

Anyway, "stuff" happens. Even if this change incurs an unexpected extra cost (one that will now have to be figured into the cost savings), in the long run, the LEDs will likely be better (and cheaper).

Of course, the problem that is occurring in some places isn't necessarily occurring everywhere.


" don't understand why an additional heat coil is such a problem in terms of energy."

A heating coil would consume energy somewhere in between the old lights and LED (without the coil.

Presumably, it would be turned-on based on temperature, which means it would be on even when it isn't needed (ie, when there is no precipitation).

Of course, the device would have to be paid-for (on top of the expensive LED lights already in-place) and there would be a cost associated with visiting each LED light and installing this new heating coil.