Syncing Traffic Lights No Sure-Fire Way to Reduce Emissions

From a motorist’s perspective, few things are more frustrating than sitting at a red light when the lights ahead are all green. That would help explain the popularity of traffic signal synchronization, neatly timing lights so that someone traveling the speed limit can expect to wait only every several cycles.

Many communities have seized on synchronization as a relatively inexpensive way to reduce congestion and emissions. In fact, it’s one of the approved uses for Congestion Mitgation and Air Quality money from the federal government.

Many communities have seized on traffic signal coordination as a way to reduce vehicle emissions, but they might just be encouraging people to drive more. Photo: ##

But it’s not quite as simple as it seems, says Network blog Transport Nexus. Some research has shown traffic signal synchronization can actually increase emissions, by encouraging more driving:

Traffic engineering has long been full of promises on solving road congestion, with many solutions worse than the problem. When roads have been widened or new roads built, induced demand has been the result. If we make all of our signals “smart” and easy for people to drive in the city, what happens then? We’ll see a mode shift away from alternative transportation options as people discover that synchronized traffic signals make driving easier, thus increasing demand for driving.

Making driving easier brings all the negative externalities to the fore: increased auto emissions, gas consumption and pressure on land uses to accommodate cars. Thus, while traffic signal synchronization seems like a good idea, the unintended consequences are likely to make the existing traffic congestion status quo even worse.

Where I live in Cleveland, traffic signals have never been coordinated (or thinned out as the population has shrunk), and the headache of driving in Cleveland is one of the major reasons I started biking. Too bad our metropolitan planning organization still spends the bulk of its CMAQ money on traffic signal synchronization and not on bike lanes.

Elsewhere on the Network today: Reinventing Parking explains why American parking policies are like vanilla ice cream, European parking policies are like dark chocolate, and Japanese regs are like a sushi bar. Cap’n Transit says that evicting everyday activity from the street — like child’s play — has contributed to chaos on the sidewalks. And Cycling Solution says that bus lanes are now open to cyclists in Budapest.

0 thoughts on Syncing Traffic Lights No Sure-Fire Way to Reduce Emissions

  1. If “cyclist must obey the rules of the road” is such a goal for some, then synchronizing traffic lights specifically for cycling speeds would encourage cyclists to stop for that one traffic light per mile. 

  2. This article ignores one very important fact-namely unsynchronized traffic signals are even more annoying and punitive for cyclists than they are for drivers, assuming of course that cyclists even bother to obey them. Don’t even get me started on how detrimental traffic signals (either synchronized or not) are to pedestrians, assuming again that pedestrians even bother to follow them. In fact, I could make a good case that one reason the majority of cyclists/pedestrians ignore red lights is because of both lack of synchronization, and an excessive number of signalized intersections (both major problems in NYC). While syncing lights for cycling speeds sounds good in theory, cycling speeds vary too widely to make this idea practical. Moreover, you need to take terrain into account. For my riding style I might want lights synced for 18 to 20 mph on level roads, but maybe only 12 mph on 4% upgrades, and perhaps as high as 45 mph on very long, steep downgrades. The numbers other cyclists would prefer would undoubtedly differ wildly from mine. Major headwinds or tailwinds will of course lay the best laid synchronization plans to waste. Cycling works best when traffic lights and stop signs largely don’t exist at all.

    It’s important to remember here that traffic signals aren’t a traffic calming device or a “discouraging driving” device, even though community boards love to have them installed for those purposes. Their primary purpose is to control intersections with limited or no lines of sight where those lines of sight cannot be easily corrected (i.e. due to obstructions like bridge abutments). Their secondary use might be in places with heavy pedestrian traffic where getting across the street is difficult due to heavy motor traffic (however a better solution here is to just reduce the volume of motor traffic rather than install signals). In all other cases we can use yields, stop signs, or roundabouts. It’s the very proliferation of traffic signals due to their misuse which has resulted in the desire to synchronize signals to make driving easier. Had traffic signals only been used as intended, they would be so sparse that synchronization would be irrelevant.

    There are far better ways to discourage driving without impacting other users such as cyclists than unsynchronized signals. Just eliminating or charging heavily for parking works wonders. So do congestion charges. As much as I feel motor vehicles are grossly overused in the US, I also realize that some are needed for society to function. As such, I’d prefer that the necessary motor vehicles stop as infrequently as possible since that makes less pollution, and also allows cheaper delivery of goods. You can largely eliminate the unnecessary motor vehicles (i.e. private passenger cars) by just not having places to park them. With fewer motor vehicles, traffic lights largely aren’t needed.

  3. Synchronize the lights at bicycle speed.  It is called a “green wave,” and they have done it in Copenhagen and Amsterdam. 

    Wikipedia says: “In the UK, it was revealed that the Department for Transport had previously discouraged green waves as they reduced fuel usage, and thus less revenue was raised from fuel taxes.”

  4. If it is true that traffic signals have “never been coordinated” in Northeast Ohio, despite the MPO “spending the bulk of its CMAQ on traffic signal synchronization,” then something is indeed terribly wrong in Northern Ohio. 

  5. As much as I usually seem to agree with Joe R., here I have to mention that one cyclist’s particular “riding style” really has to take a back seat (!) in urban settings to a reasonable, safe-yet-practical speed for cyclists as a class.

    Whatever an ideal timing ends up being, shouldn’t we be encouraging urban cyclists to *always* ride at much less than 45 mph?If I’m any example, the main reason cyclists bristle at traffic lights is *not* the relatively small amount of delay caused by an individual light.  It’s the having to laboriously overcome the inertia of stopping & restarting every few blocks (unlike car drivers, who only have to wiggle their toes, and who therefore mostly seem to grudgingly accept the presence of traffic lights, even when they recklessly try to “beat” them).If traffic lights have to exist (for now), then they should be timed to enable Charles Siegel’s “green wave” so as to encourage the slower-but-steadier ~20mph pace that liveable streets advocates are trying to require of motor vehicles.  This would discourage bike racing style from the streets, which will go a long way toward reassuring the LOLs (Little Old Ladies) that cyclists are not dangerous.We should all be riding schoolmarm style when in town.

  6. @45589687e8df260df565d048dab64df2:disqus In my opinion if we are to going to attempt to sync traffic lights to cycling speeds, then we should be using the cruising speeds (not the top speeds) of the fastest cyclists. This will tend to be in the ~20 mph area you mention. The reason for choosing the fastest cyclists, as opposed to average cyclists, is twofold. One, the fastest cyclists won’t hit a lot of reds as they would under slower light timings. Generally, when cyclists of any speed hit a lot of lights, they’ll tend to start ignoring them. Two, faster light timing won’t be all that detrimental to most slower cyclists. Maybe a 12 mph cyclist will get caught at a light every ten or twenty blocks under 20 mph light timing but that might only impact their travel speeds by 10% (while 12 mph light timing will slow a 20mph cyclist by 40%).

    While I agree to some extent about discouraging bike racing let’s remember that 20 mph is hardly “racing”. If we liveable streets advocates feel 20 mph is a reasonable pace for cars, then it’s an even more reasonable pace for bicycles with their far lower mass. There’s no good reason for faster cyclists to be forced to ride at 10 or 12 mph.  As for the 45 mph I mentioned, that’s more an outlier but there is a good reason to sync lights for higher speeds on long downhills. If a descending cyclist tries to maintain a slower speed than their freewheeling speed by riding the brakes then they risk a blowout as the rim overheats. Note however that most places with long hills where this would be an issue aren’t in urban settings, and 45 mph wouldn’t be an issue in these settings. While we’re on this subject of high speeds, my highest speed ever on a bicycle was 65 mph, not 45 mph. My highest speed in NYC was 61 mph descending the Queensboro Bridge in the mid 1980s (I had a good tailwind to assist me). This wasn’t dangerous in any way to anyone except possibly myself as there were no pedestrians on the bridge. I was actually just keeping up with motor traffic. Lots of fun to be sure but I wouldn’t do those kinds of speeds in pedestrian-rich areas. Still, on some hills in NYC, especially in the outer boroughs, it wouldn’t be entirely unreasonable syncing lights for 30 mph (or just getting rid of lights on downhills altogether on cycling routes). I find under mixed use conditions, the 25 to 35 mph which I typically achieve on downhills on NYC streets is pretty safe and reasonable. It’s certainly no faster than the pace of motor traffic to which pedestrians are already accustomed. I still have plenty of time to react and slow down if anything gets in front of me.

    Minor point but in some cases the delay caused by traffic lights is nearly as much an issue for cyclists as overcoming the inertia of starting/stopping every few blocks. I’ve found on some roads obeying lights (as opposed to treating reds as yields) will slow my average speeds from 16 or 17 mph down to 5 or 6 mph. This is a significant delay any way you look at it.

  7. @2555783a6f62598b6aadd2d882a4830f:disqus You seem out of touch with how the average urban cyclist rides. Come ride down Valencia St in SF where the traffic lights are timed to 13 mph (what has been determined to be the average urban cyclist speed) and watch how well it works in *practice* (not just theory).

    Valencia even has a slight incline so that many cyclists going uphill end up hitting a red eventually and most almost barely have to pedal going downhill to avoid catching up to the next red. So some bicyclists have to go a little slower than they might want and some a little faster in order to ride the green wave, but overall, it works really well, regardless of your theories. And it also discourages cars from using the road (especially since most motorists are completely unable to figure out the wave and race ahead to the next light only to sit there and watch all the cyclists catch up to them … you would think they would figure it out, but apparently driving a car tends to make you think less).

  8. Joe R is obviously a speed demon and hopefully does not represent the majority of us cyclists. It’s not the cycle’s mass that’s important, it’s the tires and the brakes (and the rider’sskill) . Most cycles are terribly unsafe at speeds way less than Joe proposes (try emergency braking, or going over an unexpected bump, or a slick patch); 15-20mph should be the top design speed. 

    Afa synching is concerned, there’s arguments against it. Peds, for example (and I don’t understand the comment above … you can’t synch for walking) may feel more, not less comfortable if lights are randomized. There’s a positive psychological feeling  that vehicular traffic is *not* priortized as when lights are synched. Iae, lights cannot be synched on 2-way streets. Maybe that’s why European cities, with way fewer 1-ways and un-synched lights, seem friendlier to peds.

  9. I like all the talk about syncing certain lights to cycling speeds, but I disagree with the implication that traffic light syncing is bad just because it increases driving. I don’t even own a car, and am pretty viscerally pro-transit, walking, and biking, but arbitrarily making driving worse without any trade-off is the wrong way to do it. If you take a lane away for a 2-way cycle track or a dedicated bus lane, that’s wonderful, but just making driving worse and leaving it at that isn’t the solution.

    That said, you also make the valid point that much more CMAQ money is going into this rather than non-car modes, which is a valid critique.

  10. @jd_x:disqus Most urban cyclists get faster as they become stronger. It makes no sense designing with slow, novice riders in mind when even they will eventually become much faster and more skilled. The example you gave works well because it’s only 9 blocks long. I would imagine there would be issues if you used 13 mph timing over several miles. My larger point though is that slower design speeds affect faster cyclists more than faster design speeds affect slower cyclists. I’ve already done the math. Or put another way, you can better accommodate everyone by syncing lights for 18 to 20 mph.

    @d17552797c96e647a430187ee253d613:disqus I’m hardly a “fast” cyclist. My cruising speeds are about what you might call “medium touring pace”. “Fast” is Bradley Wiggin’s 34 mph time trial pace. I never see those speeds except very occasionally on downgrades. I haven’t fallen in over 16 years riding at the speeds I do despite hitting my share of bad pavement and unexpected obstacles. The vast majority of riders who have spent more than a few years in the saddle seem to ride in the speed range I do.  We’ve had a flood of new riders in the last few years in many cities which tends to drag down the average speeds a casual observer might see. Most will eventually get faster. We really can’t directly translate the speeds of cyclists in European cities to American cities with wider streets and typically longer cycling trips. An average cycling trip in Amsterdam is maybe 2 kilometers? That will barely get you to the next neighborhood in NYC. And many American cyclists typically have road bikes or hybrids which are inherently faster than the upright single or three speeds often seen in Europe. Bottom line-20 mph seems a perfectly reasonable design speed from where I stand. There may even come a time when much faster velomobiles become common and we’ll have to think of how to accommodate them into the mix. From a pedestrian standpoint I’d much rather have to deal with 20 or 25 mph cyclists, or even 35 mph velomobiles, than 40 mph SUVs. I’m less likely to be hit by a bike which easily has room to maneuver around me. I’m much more likely to survive if I am hit. With a 40 mph SUV, all bets are off.

  11. I hear where you’re coming from, but I think you extrapolating from your own preferences rather than thinking about what would be best overall. Average trip distances in the US and Europe actually aren’t all that different despite US cities being “bigger”. And assuming that only novice riders ride under 20mph seems odd considering that very few people in high cycle usage cities (like in Europe) ride at that pace despite riding heavily their entire lives. 10-12 mph is the something of a standard for average biking speed and usually works well to accommodate riders of all levels and ages.

  12. @MaxUtil:disqus Look at the POS bikes they usually use in European cities. Put me on one of those bikes and I’ll have my work cut out maintaining 15 mph. American cyclists tend to have better equipment. I’ll be your average European cyclist on a decent road bike will be riding closer to 20 mph than to 10 mph. And if you haven’t been following trends you’re seeing more and more velomobiles on the cycling “superhighways” being built out in the Netherlands. For example, this guy normally AVERAGES 22 mph on his commute of 30 km:

    These same superhighways also safely accommodate school children. We really need to stop thinking in terms of the speeds traditionally associated with older design upright bicycles here if human-powered transport is to really take off. There’s no safety or other rational reason to limit ALL cyclists to 10 or 12 mph. If someone feels comfortable at that pace, then fine. I no more feel that a slow cyclist should have to ride at my pace than I feel I should have to ride at the pace of a slower cyclist. We can and should accommodate everyone. Remember that the speed differences between a really slow 9 mph cyclist and a really fast 25 mph one are only 16 mph. We can and already do have motor vehicles traveling at greater speed differentials on public roads with few issues. It should be even easier to do this with bikes as they need less room to pass.

    I usually agree with people on this site, but when I start hearing stuff like bikes should only go 9 or 12 mph on the same streets we allow cars to go 30+ mph it just gets under my skin. The only reason given is often just to appease the crowd who feels cyclists are dangerous. I can do that just by riding as I do and continuing to avoid crashes as I have for the last 34 years.

  13.  Joe R – I start my 50 mile commute to work, which averages 17+ MPH on mixed terrain, with the 12 blocks on Valencia and it’s one of the most pleasant parts of my commute. If I were able to ride all of the flat portions of SF at 13 MPH in order to guarantee no red light, I’d be super stoked.

    Valencia operates as a parade of cyclists all rolling at pretty much the same speed, we all know the lights. Then you get a n00b who runs a red, then races ahead and we all catch him at the next light. They learn pretty fast.

    It’s super cool and really works great.

  14.  Also Joe – in order for the green wave to work in both directions, there are only certain speeds that will work, you can’t just pick an arbitrary speed…

  15. @2555783a6f62598b6aadd2d882a4830f:disqus As I said, you do not seem to understand how the average urban cyclist rides. You clearly ride *way* faster than the majority of cyclists, so just realize that you are at the edge of the bell curve and not a target demographic for any urban planner. It would make no sense to time lights based on a statistic that is way off the mean.

    I’m in good shape and have been riding just about everyday for 5 years, and my speed hasn’t changed at all after the first few months. I don’t ride like it’s a workout (well, not unless I’m going out for a recreational ride, which I don’t do in the city where there would be traffic signals). In other words, the way you ride compared to the way I and most urban cyclists ride is the difference between walking down the street to, say, a restaurant versus running there. What you do is the equivalent to running, where every time you get on a bike it’s a workout and challenge to go fast. That’s fine, but designing bike infrastructure (which is utilitarian) around that would be like designing auto roads around Nascar drivers.

    Also, I’ve seen no evidence that as more people cycle, speeds get faster. Can you point me to some study that shows this? In fact, as more people cycle, the opposite happens: speeds will slow down as the bikeways become more congested.

  16. Armchair traffic engineering.  In most urban areas, I guarantee that most synchronization programs do not induce nearly enough trips to offset the needless starting and stopping – that result in needless delays, wasted fuel, and emmissions – of vehicular travel on unsynchronized systems. Under that logic we should make driving SOO painful that a one mile drive takes an hour.  Soon enough no one will drive…..and no one will live in that City……Transportation systems require balance, people.

  17. @twitter-14678929:disqus Yes, I’m aware that only certain speeds work on a bidirectional route.
    On the rest, so you do 12 blocks of your 50 mile (is that each way, or round trip?) commute at 13 mph? Not a huge time penalty here but what would happen if you had to do your entire commute at only 13 mph? You would add literally hours each week to your travel time. That’s my larger point. Long distance bike commuters like yourself are a minority, and always will be, but I also feel those traveling significant distances by bike (say 10 miles or more) will be the largest growth segment in the future. Longer commutes by definition aren’t very compatible with very slow speeds. While a slow green wave over small portions of a commute might work just fine, we still need to let cyclists travel at their own pace over the rest.

    And I’d be super stoked if I could ever ride more than two miles in NYC without having to stop or slow for a red light, pothole, stop sign, or any other of a number of other nonsensical things we have to deal with here. That’s how bad things are here. A few weeks ago when coming back from a ride out of the city at 2 AM I was in bliss for a while. 6.3 miles on NY25 without hitting a single red light. The lights all had traffic detectors and nobody was there, so they all stayed green. Once I hit city limits, it was back to the same old red lights every few blocks nonsense even though there was no traffic. Why can’t NYC use detectors? Lights which only go red when there is cross traffic are even better than synchronized lights. Travel when traffic is light, you’ll probably never see a red light if you stick to arterials.

  18. Synchronizing lights does not work for increasing travel speed, when the effects of road congestion are taken into account. If fact, they substantially reduce the actual speed.

    Suppose during the red cycle, there are a significant number of cars turning onto the roadway. They will stop at the next corner because the light is red. Suppose blocks are fairly short, like NYC, then a significant percentage of the next block will already be occupied by cars when the traffic light turns green. This will result in the max speed the car can go before the next light turns green will not be the cycle rate of the lights. Rather, it will be the cycle rate of the lights reduced by occupancy percentage of cars waiting for the next light to change. Thus, if the signals are timed at 30 mph and the 25% of the block ahead is occupied with cars waiting for the next light to change, then the car’s maximum speed will be only 75% of 30 mph or 22.5 mph.

    The driver will hit red lights going at 22.5 mph, when the lights are timed for 30 mph. The numbers are easier to solve, when converted to ft/sec. I’ll use a nominal 1.5 conversion factor.

    A driver proceeds along an avenue at 33.75 ft/sec. The light cycle proceeds up that same avenue at 45 ft/sec but gives the driver a 50 second head start. When will the light cycle overtake the driver?

    45(t-50) = 33.75t or t = 200 seconds.

    The driver will have travelled 200 x 33.75 ft or 6750 ft. The driver must then wait 40 seconds for the light to change to green before proceeding. Therefore, the driver’s overall speed is:

    6750 ft / 240 sec or 28.125 ft/sec or 18.75 mph.

    Life gets a lot worse for the driver, when the occupancy on the blocks ahead increases. The driver’s overall speed is reduced to 10.7 mph, when the occupancy is 50%, and 7.0 mph when the occupancy is 75%. The last two figures are closer to Manhattan’s rush hour experience than 30 mph. Observers have noted that the average speed of surface transportation in Manhattan has not risen with the introduction of the automobile.

    There is an alternative strategy to increase travel speed. Forget about synchronized lights and have all the lights on the street turn green or red at the same time. All cars will start at the same time so there will be no cars ahead waiting for the next light to change to green. Travelling at the same 30 mph, all cars will move for 50 seconds and wait for red lights for 40 seconds. This works out to an average speed of 16.5 mph.

  19. @jd_x:disqus Just to be clear, I’m referring here to all cyclists, recreational riders included, not solely to urban utility cyclists who might be riding slowly on purpose just so they don’t arrive at work all sweaty (and I might do the same in their shoes). The upper edge of the bell curve isn’t someone like myself who maybe will average 19 or 20 mph on a good day if they never have to slow down, but more typically averages 16 or 17 mph overall. The upper edge are racers who’ll go 25 to 30 mph in packs, often for many miles. That’s the demographic we don’t need to design for because it accounts for probably 0.001% of cyclists. Those similar to myself admittedly may only account for 10% or 15% of the total number of cyclists, but we probably account for 50% of the bike-miles traveled given that those like me might ride 20 or 25 miles at a time. That’s certainly a demographic which should be taken into consideration. For a good analogy look at the Interstate highways. The vast majority of miles driven there are by a relatively small proportion of the drivers. As a result, we design these highways more to cater to the faster, long distance drivers, not a city dweller who might use the highways for a few exits once in a while. The latter might actually be more numerous in terms of total number of users, but not in terms of miles driven. That’s really what I’m getting at here. I’m the equivalent of the faster, longer distance driver here but I’m certainly not a NASCAR driver, either. I expect that my needs will be taken into account, especially on any bike route going any kind of distance. Also, from a practical standpoint speed is second only to safety when judging the utility of most forms of transportation. I want human-powered transport to move beyond the niche of short inner city trips. It does fine here, but it can do so much more. In turn, this means more cycling infrastructure catering to people like myself.
    I should have said that as more people cycle, eventually average speeds will rise as more of the population becomes fitter. In Europe where most everyone on a bike has been cycling their entire lives you just don’t have a large segment of novices in relatively poor shape as in the US. Cycling speeds in Europe will only get faster due to better infrastructure and/or better bicycles, not from the average cyclist getting fitter. In the US we still have a way to go before most of the population plateaus in terms of fitness. This is the basis for my statement.

    By the way, it takes probably ten years of regular riding to start getting near your limits but you’ll never get there unless you push yourself occasionally. Maybe that’s why you have seen your speeds increase much? I noticed by my late 20s that I could ride fast much longer than I could in my early 20s, even though I had already been riding for over 5 years at that point. My top sprint speeds didn’t increase much, if at all, but my average speeds continued to rise. Even now at age 49, my average speeds might still be creeping up ever so slightly even though my sprint speeds are definitely way less than what I could do 25 years ago. Just for reference I’m 50 pounds overweight (5’9″ and 195 pounds).  I’ve never been any great example of physical prowess. If I can ride like I do, so can 90% of the population if they try.

  20. I think having all the lights turn simultaneously is done on Park Avenue if I’m not mistaken.

    Great analysis, by the way, and I agree 100% with your conclusions. Synchronized lights generally work best in places where traffic is mostly free-flowing to start with. In NYC in my opinion the best thing would be 20 mph speed limits, get rid of most traffic lights, and put roundabouts or mini-roundabouts at intersections. Traffic would probably end up averaging about 15 mph without ever exceeding 20 mph. This is way better than it does most of the time with synchronized traffic lights and 30+ mph speeds.

  21. @2555783a6f62598b6aadd2d882a4830f:disqus So where I disagree with you is the part where you think recreational riders who travel at your speed should be considered in the design of traffic synchronization on dense, urban streets with lots of other cyclists, pedestrians, and cars. Of all the MAML’s (middle-aged men in lycra) that I see going for recreational rides, they aren’t using Valencia St, for example (and they are definitely less than 10% on this street), for their ride, but to get to their ride. It makes no sense to design urban streets for that demographic just like it makes no sense to design urban automobile streets for those going for joyrides.

    Personally, there are many times when I would like to go faster than 13 mph, but I don’t mind going slower for the sake of accommodating others, and recreational riders are just going to have to understand that. In the end, a fast rider can ride at 13mph as well as at, say 20mph, but a slow rider *cannot* ride at 20 mph. So the slower speed limit will always be more inclusive, and when it comes to urban cycling as a form of transit, being more inclusive is one of the primary goals.

    Your analogy to highways is used incorrectly. For cars, you have freeways for connecting cities/regions and then you have urban streets for moving within a given city/region. And as you pointed out, the design for each is very different (although, unfortunately, we went through a good couple decades in the 50’s through 80’s where we though freeway design was appropriate for cities, but that’s another story). But the equivalent of “freeways” for bicycles would be rural or suburban roads where bicycles are traveling large distances at high speeds (the kind of speeds you are talking about), not urban streets. And on those rural or suburban roads where the recreational riders are, there are few traffic signals anyway so the issue of synchronization is moot.

  22. @jd_x:disqus Remember that the demographic I speak of isn’t solely limited to MAML going for recreational rides. In fact, that’s not even myself (I ride in regular street clothes). What about commuter cyclists going long distances? Now if the 13 mph portion of their trip is short it doesn’t matter much, but what if suddenly a place like NYC decided it’s a good idea to synchronize all roads for 10 mph travel while simultaneously clamping down on enforcement so that passing red lights would virtually guarantee a fine? The end result would be hell for anyone going from the outer boroughs to Manhattan. I feel the fastest growing demographic in the future will be longer distance travel.If the primary goal is inclusiveness as you say, then you need to set a compromise synchronization speed which doesn’t unduly penalize anyone. I’ve done some spreadsheets where I vary synchronization speed, total cycle time, and cruise speed. Assuming 20 lights per mile and a 60 second cycle, with 20 mph light timing a 13 mph cyclist only has their average speed reduced to 10.2 mph and needs to stop every 7 blocks. This represents a 27% increase in travel time-high but borderline tolerable. On the flip side, a 20 mph cyclist with 13 mph light timing with have an increase in travel time of 54%. This is unacceptable for any other mode, and it should be for bicycles as well. You could instead choose a compromise 17 mph timing. Here 13 mph cyclists only suffer a 19% delay and 20 mph cyclists only suffer an 18% delay. Better yet, it’s not *that* annoying for a 20 mph cyclist to slow to 17 mph compared to slowing to 13 mph, a speed which I personally find puts me to sleep about as quickly as watching paint dry. Which brings me to another point. When we had a national 55 mph speed limit back in the 1970s, falling asleep while driving suddenly became more common. Granted, it’s a lot harder to fall asleep on a bike since you’re actively involved, but forcing a slower speed can easily result in your attention wandering elsewhere, with predictable results. I’ve seen people texting on bikes at 10 mph. They wouldn’t even dare try that at 20 mph or they would end up eating a curb in no time.Really, the best solution where bicycles are involved isn’t to sweat over what speeds to synchronize lights at, but to just remove them altogether as I originally said, even if it requires grade separation in parts. Everyone should be able to ride at their own pace. That’s what the bike superhighways in the Netherlands allow. I’ve actually worked on a few ideas where protected bike lanes go below grade only at intersections to avoid the lights. This gives 100% of the benefits of full grade separation at a fraction of the cost. It would work in places where you have signalized intersections every few blocks or less frequently. This probably covers most urban arterials except those in places like Manhattan. For the unique situation in Manhattan, I would just use elevated bikeways. You may disagree with me here but I feel the equivalent of freeways for bicycles are more sorely needed in urban areas than in suburban areas because urban areas have more things worth riding to but at the same time also a lot more obstacles to biking at street level.

  23. @2555783a6f62598b6aadd2d882a4830f:disqus My experience with roundabouts (both mini and maxi), mostly in the UK, where they luuuuuuuuv them, is that they’re great (fun!) for bicyclists—but make life absolutely miserable for pedestrians.

    I don’t think that’s a fair tradeoff.  Really street planning in urban areas should start with what’s best for pedestrians, and go from there.

  24. @google-9ed3368a6439fa92efd353af4436290d:disqus It all depends on the percentage of pedestrians versus cyclists/autos. In NYC in the outer boroughs there aren’t large numbers of pedestrians anyway in most places, so roundabouts would work fine. In Manhattan that might be a different story. A solution to the dilemma you mentioned is to just get as many pedestrians as possible on bikes. I suspect this is why there doesn’t seem to be much consideration for pedestrians in some European cities. They’re all riding bikes instead of walking!

  25. Anti-driving bias is no substitute for logic or reason in transportation planning.  This article is a disgrace.

  26. If 25% of the road is occupied, I don’t think that translates directly into a 25% reduction in speed.  Can you explain that relationship?

    That 25% that is occupied will make a rolling start and be cleared before the traffic from the next light has arrived.  It seems like it should only become an issue when you start approaching 100% occupancy and vehicles are unable to proceed through an intersection without missing the light because the storage in the receiving lanes is already filled.

    Even if those assumptions were correct, it would only apply to the peak congestion periods of the day.  There would still be benefits for travel speed (and air quality through reduced idling) for the entire remainder of the day.

  27. As a pedestrian, I am a big fan of one-way streets with synchronized signal timing.  The traffic is organized into distinct waves, which often clear the intersection well before the end of the signal phase.  Once they’re gone, I (and everybody else in Manhattan) can feel free to safely cross (against the light).  This is particularly nice, since there are no conflicting turning movements during the end of these phases.

    My sense is that the everything-turns-green-at-once model encourages drivers to hit the gas and mark off as many signal lights as they can.  Synchronization tends to hold them down to the speed limit, but the countdown clock model can do exactly the opposite.  And these speed demons trying to get in one more light show up in a random arrival pattern.  When that’s going on with a two-way street, you’re taking your life in your hands crossing against the light (and with the light too, since there are conflicting turning movements…)

  28. This article inanely distorts a research paper to an absurd level.


    Todd Litman’s hypothesis that traffic signal synchronization leads to
    greater emissions is based on research done by a team at North Carolina State
    University  back in 2001 (the report can
    be found at: 


    Here is an actual quote from the NCSU report, “The results of this
    study also suggest caution regarding the conventional wisdom about some traffic
    control measures, such as those aimed at “traffic calming”.  Measures such as frequent stop signs or speed
    bumps have the potential to increase real-world emissions because each
    occurrence of a stop sign or speed bump is also a likely location of


    So, in other words, not synchronizing signals can cause vehicles to
    excessively stop-and-go on an urban street. 
    Vehicles that spend a considerable amount of time accelerating emit more
    pollutants into the ambient atmosphere.  The main thrust of the NCSU report is that time
    in acceleration has more impact on emissions than time spent idling.

  29. That’s where we often go wrong: thinking that making driving worse doesn’t have inherent value. In fact, anything that makes driving harder relative to other modes will have the effect of making those other modes more competitive. In places like San Francisco or other dense urban areas with good alternatives, this is especially the case.

  30. Guest wrote:

    If 25% of the road is occupied, I don’t think that translates directly
    into a 25% reduction in speed.  Can you explain that relationship?

    You saw the implication when you extrapolated to 100% occupancy.

    Let's start at 1st Street and work uptown. Car A is at the intersection and waiting for the light to change at 1st Street. Car B1 is at the end of the line between 1st and 2nd and Car B2 is at the beginning line waiting for the light to change at 2nd St, Cars C1 and C2 are the same for 3rd St, etc.

    The light turns green at 1st St at time T=0. The light at 2nd St will change at T=6; the light at 3rd St will change at T=12, etc. There are 20 blocks per mile in NYC. Therefore a car travelling with the change to green will travel 20 blocks in 6*20=120 seconds. 30 mph is 1 mile in 2 minutes.

    However, car A will not be able to get to the traffic light at 2nd St until car B1 clears the intersection. How long will it take C1 to clear 2nd St after that light changes? He's 25% of the block length from the light. Assuming cars are travelling at the 30 mph speed limit, that comes to 25% of 6 sec or 1.5 sec. That's 1.5 sec after the light at 2nd St changes to green. So, car A will travel the block in 7.5 (6.0 + 1.5) seconds. Car C2 is in the same position at 3rd St as car A was at 2nd St. Cars B2 and B1 will take 7.5 seconds to travel the block between 2nd and 3rd. Car A, which is blocked by car B1 will also take 7.5 seconds to go between 2nd and 3rd. Clearly, with all the streets ahead being similarly occupied, car A will take 7.5 seconds to traverse each block.

    Car A will take 7.5 x 20 or 150 seconds to travel 1 mile. That comes to 24 mph. However, cars in a queue do not all simultaneously start when the light turns green. They wait until there is space for them to proceed (even cab drivers). Let's assume they wait for a car length before hitting the gas. This means that car B1 will not start until car B2 is twice the occupancy length in front of car B1. Thus, car B1 will not hit the gas until 3 seconds after the light at 2nd St turns green. Car B1 will then take an additional 1.5 seconds to reach the light at 2nd St. This increases the time for car A to 10.5 seconds per block, instead of the 6 second 30 mph traffic light cycle.

    So, car A will take 10.5 x 20 or 210 seconds to travel a single mile for an average rate of 17.1 mph. Of course, he will hit a red light before reaching the mile post to further reduce his average speed.

  31.  Joe R – my commute is 50 miles one way, then I take the train back because we Bay Area folks not only have green waves for cyclists, but we also have racks for 80 bikes on our trains.

    Once I get out of the city I am on roads which don’t have a lot of traffic lights or stop signs. A mixture of Freeway frontage roads, bike paths, or semi-rural roads. Places to step on the gas.

    I don’t think in the city on crowded streets is the place to step on the gas.

    If Valencia was timed for 17 MPH which I could comfortably hold, I would still be riding in a bike lane wide enough for me to avoid doors but not wide enough for me to pass someone riding at 13 MPH without going into the “primary travel lane”. And with the faster light timing more cars would use valencia and travel faster, making such passes more dodgy than they are now (occasionally I will have to pass someone going 8 MPH).

    The slower speeds don’t make for super fast travel for those who can ride faster but make it very civil for everyone. And having seen more than enough crashes, I think in dense areas that civil is more important than fast. If you have a 20 mile commute on dense city streets like Valencia, you might just be doing it wrong, whether in a car or on a bike. Dense areas should be made slow, and because they are so dense, travel distances will typically be shorter so that slowdown isn’t a big deal. My ride to the train station is 15 minutes. It’s not worth a 10% increase in crashes to cut my train commute to 13 minutes.

  32. @twitter-14678929:disqus Thanks for the clarification on your commute. 50 miles a day is impressive enough. If you rode both ways I would think you were super human!

    I agree in principal with this: “My ride to the train station is 15 minutes. It’s not worth a 10% increase in crashes to cut my train commute to 13 minutes.”

    I even said as much earlier. If these slow portions only comprise a small portion of a commute, it’s not really a big deal. That said, when you need to slow all cyclists down to the speed of the slowest ones for safety, then to me this says the infrastructure is inadequate. My attitude is either build top notch infrastructure for bikes or build nothing. Any decent cycling route will take into account the fact that cyclists ride at different speeds. Even most sidewalks are wide enough for faster pedestrians to pass slower ones. Maybe on Valencia it would have been politically difficult to take more space from cars to allow passing. I don’t know anything about your local politics. In any case, since the 13 mph zone is so short, I wouldn’t consider it a major drawback if I had to ride that route. If it went on for miles instead, then there would be a pressing need to do something different. As I said earlier, I really a big fan of grade-separated bicycle infrastructure given the reality that it’s politically difficult to take road space from cars in most places, but at the same time riding at street level is full of obstacles in your average urban environment.

    I have to disagree with you that dense areas should be made slow. Maybe that works OK when the really dense “downtown” only goes on for 12 blocks. Look instead at NYC where basically you can go 20 miles and never get out of a high density urban environment. If you condemn cyclists to going slow in dense areas here, then longer distance commutes, such as from the outer boroughs, just really will never be practical by bike. This is why in places where the dense areas just go on, you need the equivalent of freeways for bicycles to get them out of that dense environment. Unlike freeways for cars, freeways for bicycles won’t divide neighborhoods in two, won’t be eyesores, and would probably be largely welcomed.  They also would probably cost a few percent per mile of what car freeways do while carrying just as many people. In effect, they could serve the same mass transit function as subways in areas where we never bothered to build a subway. Maybe over on the West Coast none of this is much of an issue. Maybe in most cases you can bike though dense areas in 5 minutes and then reach the open road. I can tell you that we sorely needed a better way to travel by bike here on the East Coast than street level bike lanes.

  33. perhaps. but synchornized lights also help bicyclists (you typically get pretty far even if the timing is faster than you can go) pedestrians (synchronized traffic at a reasonable speed is way better than traffic going 50 to catch the light) and maybe transit (if it doesn’t stop every single block. or maybe even if it does, by clearing other traffic out of the way).

  34. The blog that was referenced, Transportation Nexus, pulled a quote from an author that completely falsified the results from a empirical research study.

    The quote from the blog: “…traffic signal synchronization on congested roads provides little measurable benefit, and can increase emissions in some situations (Frey and Rouphail 2001).”

    An actual conclusion from the Frey and Rouphail report: “NCDOT [North Carolina DOT] should be encouraged to carry out projects that reduce congestion on signalized arterials and/or improve coordination and signal timing as those strategies have been shown in this work to have a demonstrable effect on emission rate reduction.”  Report link:

    Those two statements are not the same thing.

    If Streetsblog wants to be regarded as a source for responsible journalism, they would retract that article and state the correct results of the cited research.  Some of your sponsors, who are more objectionable, may question why they are financially supporting an organization that falsifies the research of others.

  35. Yes, making
    driving easier brings all the adverse externalities to the fore, improved auto pollutants,
    gas intake and pressure on land uses to accommodate cars. Thus, while traffic
    signal synchronization seems like a wise decision, the random repercussions are
    likely to make the current traffic congestion status quo even intense.

  36. When a left or right turn is indicated, the electricity flows through a device called the thermal flasher to the front and rear turn indicator bulbs, and to the dashboard turn signal indicator lights.

  37. Traffic lights in the same direction should be synchronized and would alleviate a lot of the congestion that is caused by having one green and the next light red. It doesnt make any sense not having them synchronized and the notion that this is going to increase and encourage driving is ludicrous. People are going to drive regardless. This is something that needs to be changed. Traffic control could easily implement something like this in no time at all. I can’t understand why they haven’t yet

Leave a Reply

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


Syncing Traffic Lights No Sure-Fire Way to Reduce Emissions

From a motorist’s perspective, few things are more frustrating than sitting at a red light when the lights ahead are all green. That would help explain the popularity of traffic signal synchronization, neatly timing lights so that someone traveling the speed limit can expect to wait only every several cycles. Many communities have seized on […]

Sadik-Khan: Many Initiatives Are Under Way…

Today’s Crain’s Insider, available to subscribers only, reports on the next steps for the Bloomberg Administration’s broader Long-Term Sustainability Plan now that congestion pricing has cleared its first hurdle in Albany: 7.30.07 Crain’s NY Business The Insider by Erik Engquist and Anne Michaud GREEN PLAN Closer to home NOW THAT the Legislature has passed a […]

Would Motorists Pay 15 Cents a Mile for No Traffic?

Congestion pricing may be dead in New York, but the discussion about its merits continues elsewhere. Today, David Alpert at Greater Greater Washington looks at how road pricing could dramatically change the traffic situation in the DC area: Photo by derang0. If the Washington region charged around 15 cents per mile for use of the […]