Detroit Streetlight Effort Dramatically Reduces Ped Deaths

Photo: Public Lighting Authority
Photo: Public Lighting Authority

Detroit is seeing the light.

A new report on streetlights — that most crucial, yet oft-overlooked infrastructure element — shows that pedestrians deaths plummeted in the Motor City from an average of 24 per year to just one, thanks to a two-year, $185-million effort that replaced 65,000 lights [PDF] in a city where 40 percent were not working in 2014.

The results have been more than just illuminating. In 2017, there was just one pedestrian death at night in an unlighted area. And overall, they’re down 40 percent since their height in 2015.

detroit graph


Detroit, as the Detroit Free Press recently reported, has the highest per-capita pedestrian fatality rate of any city in the country. But since 2009, pedestrian deaths have dropped 16 percent. Meanwhile, throughout the rest of Michigan, they increased 47 percent.

This amazing success story can literally be viewed from space, as Politico reported. The use of LED lights also reduced the city’s energy bill by 55-60 percent, according to the city.

“Certainly no pedestrian fatalities or serious injury crashes are acceptable, but city’s safety data is improving,” Todd Scott, director of the advocacy group Detroit Greenways, wrote in a blog post. “We expect this trend to continue as additional Complete Streets treatments, such as road diets, bike lanes and other pedestrian infrastructure improvements reduce motor vehicle speeding and pedestrian exposure leading to a safer, healthier, and more walkable city.”

Mayor Mike Duggan has proposed an ambitious plan to help orient the city’s streets toward biking, walking and transit, which should improve Detroit’s numbers further.

One question for advocates is how applicable is this situation to other cities? How big of a safety issue is lack of or broken streetlights? We saw in the dash cam footage from the AV death of Elaine Herzberg in Tempe, for example, that low light was likely factor.

We asked Robert Schneider, a professor at the University of Wisconsin-Milwaukee, who specializes in pedestrian safety. He says there is a need for more research on this topic but that “Lighting is very important since approximately three-quarters of pedestrians in the U.S. are killed at night,” citing recent findings on the problem from the National Transportation Safety Board [PDF].

Both NTSB and the Insurance Institute for Highway Safety have called for better, brighter headlights to help reverse skyrocketing pedestrian deaths. But there may be a need to pay more attention to street lighting as well, especially in poorer cities or suburban areas.

35 thoughts on Detroit Streetlight Effort Dramatically Reduces Ped Deaths

  1. It’s not just the amount of light, but also the type of light, which is important. The high-pressure sodium lights traditionally used for streetlighting are awful. The spectrum kills both peripheral vision and depth perception. Potholes look a lot shallower than they actually are. In general the lit scene looks “flat”. The yellowish light lulls you to sleep, which is the last thing you want happening to someone driving a multi-ton machine at high speeds. HPS is also awful from an aesthetic perspective. Streets lit by them are drained of color. Everything is cast in a depressing yellow haze.

    The switch to LED streetlights by most major cities not only saves energy, but improves the outdoor environment. The scene now looks crisp, with good contrast and good color rendition. Peripheral vision is enhanced. Drivers can’t avoid hitting what they can’t see. LED streetlights ensure that they see well.

    The only fly in the ointment is the idiotic anti-blue light movement seeking to get municipalities to use either yellower LEDs, or even worse amber LEDs, for streetlighting. Blue light in the spectrum is necessary for seeing well. The negative health effects of blue light are unproven at best, junk science at worst. Even if the science was valid, the screens people view once they get home are a far bigger source of blue light than streetlights. Moreover, blue light keeps people awake, which is exactly what you want when they’re driving on public roads. From an aesthetic perspective, things look best when lit with a 4500K to 5500K light source as this is the same range as natural sunlight. People can use whatever type of light they want in their homes, but public spaces should be lit with the safest type of light possible. That means fairly high intensities in the 4500K to 5500K range, along with good color rendering (LEDs can be mediocre to great in that area, depending upon the type).

  2. Low light was not a factor in the AV death of Elaine Herzberg. The cause was distracted driving. Speaking of which, when will charges be filed?

  3. If better light would have meant that even a distracted driver would see her, it’s legitimately a factor.

  4. Actually, distracted driving was irrelevant here. What was really at fault is the flawed concept of having the driver take over if the autonomous system makes a mistake. In a nutshell, the driver will have to realize the system made an error, decide on an alternate course of action, and then implement that course of action in a time frame of tenths, perhaps hundreds of a second. That’s beyond human capability. If nothing else, the Herzberg incident is a vivid illustration of why having a human driver as a backup is a horrible idea.

    If we’re going to make AVs, then there should be no option for a human to override them, short of maybe a big red stop button which brings the vehicle to a halt as soon as it’s safely possible. The fact automakers are seriously considering letting a human driver function in a backup capacity is frightening. This is exactly the type of task humans are horrible at, namely monitoring something which is mostly uneventful on theory they can correct errors when they happen. Humans always become distracted and/or fall asleep doing such tasks. Arguably, having a human driver backing up an AV is worse than just having the human do all the driving. Chances are nearly 100% that sooner or later the human will in fact override the system, but make a worse decision than either the AV or a full-time driver will make, with potentially disastrous consequences.

    If there are charges to be filed, it should be against Uber itself for basically beta testing an AV which was not ready for prime time on a public road. There was nothing humanly possible the driver could have done given the inevitable complacency which occurs when you’re just monitoring a task instead of participating in it:

  5. Visibility (or conspicuity as the technical term) is a critical element in pedestrian safety after dark. NHTSA research shows that a high percentage of pedestrian fatalities occur at night. Many pedestrians could reduce their risks by wearing light colored and/or reflective clothing as many cyclists do today. This is obviously more critical in areas that do not have good street lighting.
    James C. Walker, National Motorists Association

  6. Many drivers who hit and kill pedestrians could reduce their risks by SLOWING THE F DOWN ON A DARK STREET

  7. Very true, but I am only interested in realities – not wishful thinking behavior that does not and will not happen voluntarily. The fatality rate is now 1.16 fatalities per 100 million vehicle miles traveled.

    That means if you are in a car for about 15,000 miles a year, you will be in a crash with a fatality of a pedestrian, cyclist or vehicle occupant about once every 5,700 years. The risks on any one trip are so miniscule that to expect voluntary compliance of always driving a lot slower at night just in case a pedestrian in dark clothing will cross improperly is not realistic.

    James C. Walker, National Motorists Association

  8. Every person who gets out of a car becomes a pedestrian in that moment. Every place currently selling winter coats is selling racks and racks of black, charcoal gray, navy blue, and other dark colors.

    It is not realistic to tell people to buy special clothing, and that advice assumes the ability to own multiple coats. Many people walking/bicycling do so out of necessity. A fix that relies on them to have money ignores the responsibility of the managers of the public realm to make it work for all.

    Systematic safety approaches work far better than admonishing individuals to change their coats. Better lighting on the street provides a fix for everyone regardless of mode. Better headlight designs on cars would also help drivers see farther down the road. Are you advocating for those through the NMA?

  9. No, victim blaming is not right, but you can also be both “right” and “dead”. Taking simple, easy, precautions is just being smart. I get that dressing all in black is currently “stylish”, for some people (“goths”?), but adding just one item that increases contrast is neither expensive nor onerous. Don’t “win the battle and lose the war”!

  10. Better street lighting is VERY helpful, but difficult to achieve in some cities if the councils don’t allocate budget for it.
    The USA has been behind European rules for vehicle lighting for 50+ years and a LOT of people have begged to adopt ECE rules. But NHTSA is stubborn and refuses to adopt the better rules. The same is true for many of the vehicle rules where the ECE has been far ahead for decades.
    In my area, cyclists over about the last 10 years have dramatically improved their visibility after dark. It would not be difficult for pedestrians to do the same — OR to take greater care in crossing by making the assumption they are not visible enough.
    James C. Walker, National Motorists Association

  11. The NTSB report was quite useless (as usual). Here is what the police investigation concluded:

    A search warrant obtained by Tempe police show her personal phone was streaming an episode of The Voice on Hulu until one minute before 10 p.m., the approximate moment of the crash, the new report shows.

    Tempe investigators conducted lighting and road tests to determine whether an average motorist should have seen Elaine Herzberg walking across. Because she was so distracted, Vasquez missed the chance to see Herzberg and take evasive action, investigators concluded.

  12. You continue to ignore Barb’s comment. When you go out to dinner or other destinations after dark (after driving there of course), do you and those you are with don your reflective vests so you can walk to your final destination or cross the street to get to and from your parked vehicle? Yeah, I didn’t think so.

  13. Wrong. Even light colored clothing provides very little difference in conspicuity of a pedestrian at night. That is a fact borne out in tests.

    Are you saying if I have dark pants I need to select a light colored jacket, or vice versa? Am I supposed to coordinate my wardrobe simply so I can cross the street? Seriously? I’m sure you join James Walker in donning your reflective vests when walking about at night. Hah, I just realized the irony in his name.

  14. Oh, so the reality (not wishful thinking) is that pedestrians should select and wear a specific wardrobe, but motorists can’t be expected to actually obey speed limits and not overdrive their headlights? F*cking seriously?

    Car crashes are the leading cause of death for children, teens, and young adults. It’s one of the riskiest things we do. And that doesn’t account for the many more crashes that result in injuries, not fatalities, including the many thousands of serious injuries every year. And no, it isn’t just “pedestrians in dark clothing” that are hit.

    You’re worse than a troll; you’re daft.

  15. Placement of the lighting is also even more critical. Downcast lighting behind the pedestrian creates a black silhouette that blends into the dark surroundings and shadows, rendering the person virtually invisible. Lighting needs to be between the driver and the pedestrian, illuminating the side of the pedestrian visible to the approaching motorist.

  16. No, but I am exceptionally careful and patient when crossing streets after dark because I know I am less visible.
    James C. Walker, National Motorists Association

  17. Basing wishes for safety improvements on behaviors that do not happen and will not happen is both useless and foolish. I am interested only in realistic solutions that can work.
    James C. Walker, National Motorists Association

  18. Basing wishes for safety improvements on behaviors that do not happen and will not happen is useless and foolish. I am interested in solutions that can work.

    James C. Walker, National Motorists Association

  19. So, the only “correct” answer would be streetlights every 25′-30″, to minimize the chances for shadows and back-lighting?

  20. “If pedestrians don’t want to get killed, they should stop wearing clothes that aren’t fluorescent and reflective”

    It’s literally the exact same

  21. Well designed pedestrian precincts work very well.
    James C. Walker, National Motorists Association

  22. Usually old people dress in darker clothing and cross streets more slowly (thus spending more time in the path of oncoming vehicles) than young people. Older pedestrians often have poorer hearing and eyesight as well as slower reaction times than young people.

    As a result drivers need to take special care around older pedestrians. Source: Any German driving school textbook.

    A lot of the nonsense you hear on American web sites on the subject of traffic comes from the fact that Americans get their licenses without ever having gone through a proper driving training course.

  23. Unfortunately the streets in Detroit are dangerous by design — to wide, too straight and with few clear boundary markers.

  24. I’d argue that people of every age dress in darker clothing. All black seems to be pretty stylish, these days.

  25. The concept of separate arms on streetlight poles to light sidewalks is a good idea. Right now it seems like any light the sidewalks get is incidental. By lighting sidewalks separately, you can have the streetlights focused only on the street (eliminating light trespass issues), and the lower sidewalk lights can be designed to only focus on sidewalks. Their lower height will also minimize light trespass issues.

    As for complaints about the Kelvin number, those will mostly pass as people get used to the new streetlights. It was a huge jump from ~2000K to 4000K+. People unfortunately get used to even poor lighting like the old sodium lights, and any change will elicit complaints. I remember when we went from mercury vapor to sodium in the 1970s lots of people complained. Eventually, the complaints ceased, even though the sodium lights are worse in every aspect than the mercury lights they replaced (other than the absence of mercury). Even their supposed energy efficiency was on paper only. They made seeing worse. When you increased light levels enough to match the seeing ability under the old mercury vapor lamps, energy consumption was the same or greater.

    This is also why the quest for lower Kelvin numbers for LED streetlights is misguided. Blue light helps us to see well. If you decrease the Kelvin number then you need to increase the lumens in order to not degrade the ability to see. Besides increasing energy consumption, you’ll end up with the same amount of blue light as you would using a higher Kelvin number but fewer lumens. There’s really no way to get around this. If you keep the lumens the same but drop the Kelvin number, you’re making things more dangerous for everyone. From an aesthetic standpoint you’re also degrading the nighttime atmosphere. Nights are naturally illuminated with moonlight (~4000K) and starlight (>6000K). Our eyes actually see better under this type of light in low-lighting conditions.

    Besides that, the AMA recommendations are ridiculous because only a small fraction of your night time blue light exposure is from streetlights. Most is from the screens people look at. And the supposed negative effects of blue light exposure mostly fall into the category of junk science. The primary source of blue light exposure is the sun. Perhaps the AMA should also propose that we make the sun yellower.

  26. Yes, by all means please reprint it there. The part about the decreased visual acuity of lower Kelvin streetlights is particularly relevant. If NYC ever decides to start using 2700K or 3000K streetlights, I will mount a lawsuit to insist they increase lighting levels appropriately over and above the current 4100K models so as not to decrease the ability of drivers to see at night. Unfortunately, the AMA is either ignorant of this issue, or conveniently decided to ignore it. When all is said and done, if the ability to see is equal under different Kelvin streetlights, the blue light exposure will be about the same because the lower Kelvin ones will require higher intensity. If you don’t increase the intensity, you’re actually making the street more dangerous.

  27. Two articles you might want to link to:

    Also worth a mention is the fact we used mercury vapor streetlights for decades with none of the issues reported by the AMA. MV lights have more blue content in their spectra than LED. If there’s any problem with LED streetlights, it’s light trespass and possibly intensity (i.e. they may be brighter than is necessary for good visual acuity), not color temperature. Light trespass and intensity can both issues can be addressed by luminaire design. Nobody should have glaring streetlights shining into their home, regardless of color temperature.

Leave a Reply

Your email address will not be published. Required fields are marked *


New Report Maps the Gap Between Pedestrian Risks and Federal Safety Aid

The top 10 most dangerous cities for pedestrians. (Chart: Dangerous by Design report) If the equivalent of one jumbo jet full of Americans died every month, the resulting public outcry would be deafening. Or would it? Anne Canby, the former Delaware transportation secretary who heads the Surface Transportation Policy Partnership (STPP), raised that question today […]

America’s Progress on Street Safety Is Pathetic

A new report from the International Transport Forum shows America is only falling farther behind all of its peer nations on street safety [PDF]. The traffic fatality rate in the United States (10.7 per 100,000 people) is nearly four times higher than in the United Kingdom (2.8 per 100,000) and close to double that of […]