Remember our description of Montgomery County’s screwed-up system for measuring traffic congestion? Remember our proposal for accurately measuring congestion through massive usage of GPS devices? Well, it turns out that the country’s largest traffic measurement company agrees with us.
INRIX is a traffic measurement firm based outside Seattle that was founded by two former Microsoft executives in 2004. The company just released a report ranking traffic congestion all over the United States. (The Washington area ranks fourth-worst after Los Angeles, New York and Chicago.) How does the company gather its data? The company states:The raw data comes from the historical traffic data warehouse of the INRIX Smart Dust Network. Since 2006, INRIX has acquired billions of discrete “GPS-enabled probe vehicle” reports from commercial fleet vehicles – including taxis, airport shuttles, service delivery vans, long haul trucks – and cellular probe data. Each data report from these GPS-equipped vehicles includes at minimum the speed, location and heading of a particular vehicle at a reported date and time.
How do they measure congestion? The company calculates a “reference speed” based on how fast GPS-equipped vehicles travel on a road in the middle of the night. Presumably, that reflects driving time in non-congested conditions. Then the company draws on more GPS data to calculate average speeds in each hour of the day for every day of the week. Then the company divides the reference speed (representing free flow) by the average peak-hour driving speed to calculate a Travel Time Index. The higher the index, the greater the congestion. For example, an index value of 1.3 indicates that a peak-hour trip will take 30% longer than a free-flow trip because of congestion.
INRIX has developed efficient methods for interpreting probe vehicle reports that are provided in real-time to establish a current estimate of travel patterns in all major cities in the United States. These same methods can aggregate data over periods of time (annually in this report) to provide reliable information on speeds and congestion levels for segments of roads. With the nation’s largest probe vehicle network, INRIX has the ability to generate the most comprehensive congestion analysis to date, covering the nation’s largest 100 metropolitan areas.
Is INRIX’s congestion formula the right one? Maybe yes, maybe no. But more importantly, their calculations are based on billions of actual trips recorded by GPS devices in commercial vehicles all over the U.S. Unlike Montgomery County, INRIX does not base its statistics on fluky critical lane volume measurements that are taken once every four years or so and, according to Park and Planning’s own research, do not actually measure congestion.
Folks, we have to be able to measure traffic congestion accurately in order to plan successful mitigations, including road improvements and transit. Here’s a private sector company that is getting a ton of real-world data and giving it their best shot. So if INRIX is doing it, why can’t we?
Friday, June 20, 2008
Trying to Get it Right on Traffic
Thursday, June 05, 2008
Planning Department Knew About Traffic Measurement Flaws, Part Two
Ten years ago, the Planning Board considered whether to raise the county’s Critical Lane Volume (CLV) standards by up to 100 points. This would have allowed developers to escape traffic mitigation requirements near some intersections exceeding the CLV standards in their policy areas. But citizen activists questioned whether allowing CLV standards to rise would result in more congestion since higher CLVs are thought to indicate more congestion. The surprising answer from the Planning Department’s staff was no. Why? The staff found that CLVs were unrelated to real-world traffic delays.
In a 1998 study entitled, “Measuring Congestion and Delay: The Critical Lane Volume Method,” Planning Department staffers Richard C. Hawthorne and Ronald C. Welke looked at how CLVs compared to other measures of congestion. The authors stated:In researching how to measure congestion, the study group selected actual delay as the best measure as perceived by the roadway user. The average stopped delay per vehicle in the peak hour is the measure used in the operational analysis of intersections in the 1994 HCM [Highway Capacity Manual], and therefore provided a quantifiable standard.
The authors then gathered data on CLVs and actual delays from 27 observations taken at 15 highly-congested intersections to see if the two measures correlated. They plotted each of these observations against each other on the chart below:
If the two measures were directly related, the observations would cluster tightly around a line rising from zero on both axes. Instead, the observations form an amorphous blob. The authors found that the coefficient of determination between delay and CLV, also known as R-squared, was only 14%. That means that only 14% of the variation in one measure is explained by changes in the other. The authors concluded, “There is little relationship between delay and CLV.”
So if CLV is such a poor predictor of “the best [congestion] measure as perceived by the roadway user,” why not stop using it? The authors said, “The problem is that delay data is difficult and expensive to gather and thus not readily available.” So because CLVs were shown to be unrelated to delay, the Planning Board raised the allowable CLV standards – a move that was opposed two years later by then-Council Member Ike Leggett. And rather than search for a better data source that truly measured actual traffic congestion, the Planning Department has continued to rely on CLVs. The situation is compounded by the fact that Richard C. Hawthorne, one of the study’s authors, was then and still is now the department’s Chief of Transportation Planning.
This has potentially severe consequences for traffic management in Montgomery County. CLVs are used by Planning staff to form recommendations on traffic mitigation for new developments. Under Local Area Transportation Review (LATR), if an intersection near a new development exceeds the CLV standard for its policy area, a developer is required to pay for traffic mitigation measures. But what if, as the above study holds, CLV is not a reliable predictor of congestion? That means there is a possibility that mitigation measures have been installed at intersections that do not need them, and have not been installed at intersections that desperately require them. And this has been going on for at least ten years even though the Planning Department KNEW that CLVs by themselves were a flawed measure of actual congestion.
Planning’s knowing reliance on a defective congestion measure is difficult to understand and impossible to excuse. But it can be fixed. Perhaps delay was expensive to collect ten years ago, but that was prior to GPS units being available for rent at $5 per day.
Our Planning Department was once the best in the country. It is important to every one of us that it produce the highest-quality information on traffic and development that is humanly possible. We do not deserve the cheapest traffic measurement system, or the quickest and dirtiest, or the one we have been using for a long time merely because the bureaucracy wants to avoid change. We deserve the best. We put our alternative on the table and now it’s the Planning Department’s turn.
The long-forgotten CLV study is not available online, but we reproduce it in its entirety below.
Wednesday, June 04, 2008
Planning Department Knew About Traffic Measurement Flaws, Part One
After last week’s marathon five-part series on traffic measurement, I was prepared to move on to other things. But an anonymous friend of the blog sent me proof that the issues I have been discussing have been known to the Planning Department for at least a decade. An internal study prepared by Planning staff back in 1998 found that Critical Lane Volume (CLV), the measure they use to estimate traffic congestion, has “little relationship” to actual delays. This revelation throws into question the entire system of traffic mitigation used by Montgomery County.
To understand the magnitude of this statistical debacle, we must first review the role played by CLVs in the county’s planning procedures. As I detailed last week, CLVs are hourly sums of conflicting movements (both through and turns) of cars through an intersection. They reflect volume: the more cars pass through an intersection, the more a CLV count will increase. But some of the county’s most congested road corridors have low CLVs because they are too gridlocked for large numbers of cars to get through. Nevertheless, the Planning Department uses CLVs to construct their regular lists of the county’s most-congested intersections and make recommendations for capital improvements.
CLVs play an even more important role when the Planning staff assesses traffic impacts of new developments. One of the review procedures that a new development must pass is Local Area Transportation Review (LATR), an analysis that explores the development’s traffic impact on a handful of nearby intersections. Under LATR, the developer is required to submit a traffic study to the Planning staff estimating the number of new trips that will be created at peak travel hours. The staff then obtains the CLV estimate for the affected intersection(s) and compares it to the standard set for the development’s policy area. If the CLV exceeds the relevant policy area standard, the developer will be required to pay for traffic mitigation measures to offset that impact. These measures might include purchasing Ride-On buses, building bus shelters, installing turn lanes, widening an intersection or other remedies.
The county’s policy area standards provide for acceptable CLVs to be relatively low in rural areas and higher in dense areas. The rationale is that dense areas are likely to have more transit options, especially when they are near Metro stations. As established in last year’s growth policy, the CLV standards by policy area are:
1350: Rural East, Rural West
1400: Damascus
1425: Clarksburg, Gaithersburg, Germantown East, Germantown West, Montgomery Village/Airpark
1450: Cloverly, North Potomac, Olney, Potomac, R&D Village
1475: Aspen Hill, Derwood, Fairland/White Oak
1500: Rockville City
1550: North Bethesda
1600: Bethesda/Chevy Chase, Kensington/Wheaton, Germantown Town Center, Silver Spring/Takoma Park
1800: Bethesda CBD, Friendship Heights CBD, Glenmont, Grosvenor, Rockville Town Center, Shady Grove, Silver Spring CBD, Twinbrook, Wheaton CBD, White Flint
Here’s an example of how this review procedure might work. Suppose a developer wanted to build a commercial project near the corner of Connecticut Avenue and Jones Bridge Road in North Chevy Chase. The Planning Department estimates CLVs for three nearby intersections: Jones Bridge at Manor Road (679 AM, 676 PM), Jones Bridge at Platt Ridge Drive (773 AM, 963 PM) and Jones Bridge at Connecticut (1731 AM, 2017 PM). The policy area standard for Bethesda/Chevy Chase is 1600. While two intersections fall below that standard, Jones Bridge at Connecticut exceeds it in both the morning and evening. This developer would have to agree to a package of mitigation measures to offset the traffic impact on Jones Bridge at Connecticut.
Now suppose the developer wanted to build the same project near the corner of Connecticut Avenue and Knowles Avenue in Kensington. The Planning Department estimates CLVs for three nearby intersections: Plyers Mill Road at Metropolitan Avenue (687 AM, 866 PM), Connecticut at University Boulevard (1335 AM, 974 PM) and Connecticut at Knowles (1433 AM, 1274 PM). The policy area standard for Kensington/Wheaton is 1600. Since none of the intersections exceed the standard, this developer would not be required to pay for mitigation (at least not under LATR).
The county’s traffic mitigation system under LATR is thus based on the premise that CLVs are reliable measures of congestion. When intersections “fail” – that is, they exceed their policy area standard CLV – traffic mitigation is required for new developments. When intersections “pass” – in other words, fall below their standard – traffic mitigation is deemed unnecessary.
But we called into question whether CLVs truly measure congestion all last week. We even revealed a list of four heavily-congested corridors with low CLVs using the Planning Department's own data.
And now we learn that the Planning Department’s own staff confirmed our suspicions ten years ago. More on their long-forgotten findings tomorrow.
Friday, May 30, 2008
A Better Way to Measure Traffic, Part Three
In Part Two, we listed four corridors with high volumes and low speeds. Today, we list four corridors with modest volumes and low speeds. While these corridors perform differently than the ones we listed yesterday, they may be even more clogged. Several of these roadways contain chains of gridlocked intersections that collectively cut down on auto movement.
The Worst Modest Volume, Low Speed Corridors
These corridors have very low speeds but do not have critical lane volumes that exceed policy area standards (at least not in the evening). The relatively modest volumes combined with the crawling speeds suggest that cars have an extremely difficult time squeezing through these gridlocked areas.
Colesville Road, 16th Street to Spring Street
Test Run #1
5/23/07 at 5:06 PM, Eastbound
0.77 Miles in 7.87 Minutes, 5.9 MPH
Test Run #2
5/23/07 at 7:34 PM, Eastbound
0.66 Miles in 5.57 Minutes, 7.1 MPH
PM Critical Lane Volumes
Colesville at East-West Highway: 1061 (Recorded on 6/2/04)
Colesville at Georgia: 1049 (Recorded on 9/26/06)
Colesville at Spring: 1248 (Recorded on 9/20/06)
Notes: This corridor is located in Downtown Silver Spring and runs past the Metro Station.
Frederick Road, Chestnut Street to Montgomery Village Avenue
Test Run #1
5/15/07 at 5:47 PM, Northbound
0.54 Miles in 4.97 Minutes, 6.5 MPH
Test Run #2
5/15/07 at 6:15 PM, Northbound
0.67 Miles in 5.85 Minutes, 7.1 MPH
PM Critical Lane Volumes
Frederick at Chestnut: 1204 (Recorded on 9/30/04)
Frederick at Odenhal: 1372 (Recorded on 11/10/04)
Frederick at Perry: 974 (Recorded on 3/10/04)
Frederick at Montgomery Village: 1427 (Recorded on 5/2/05)
Policy Area CLV Standard: 1450
Notes: This corridor runs near Lake Forest Mall and the I-270/Quince Orchard Road interchange.
Georgia Avenue, Prince Phillip Drive to Olney-Sandy Spring Road
Test Run
5/30/07 at 5:32 PM, Northbound
0.97 Miles in 5.17 Minutes, 11.3 MPH
PM Critical Lane Volumes
Georgia at Prince Phillip: 1145 (Recorded on 3/6/07)
Georgia at King William: 1095 (Recorded on 12/9/03)
Georgia at Olney-Sandy Spring: 1251 (Recorded on 3/15/07)
Policy Area CLV Standard: 1450
Notes: This corridor is just south of the Olney commercial district.
Georgia Avenue, Spring Street to Dennis Avenue
Test Run #1
6/4/07 at 5:15 PM, Northbound
0.46 Miles in 8.32 Minutes, 3.3 MPH
Test Run #2
6/4/07 at 5:15 PM, Northbound
1.46 Miles in 6.67 Minutes, 13.1 MPH
PM Critical Lane Volumes
Georgia at Spring: 1080 (Recorded on 11/17/05)
Georgia at Seminary: 1374 (Recorded on 4/7/05)
Georgia at Beltway: 1206 (Recorded on 11/20/03)
Georgia at Forest Glen: 1377 (Recorded on 6/6/07)
Georgia at Dennis: 1437 (Recorded on 6/7/07)
Policy Area CLV Standard: 1600
Notes: This corridor includes Montgomery Hills, the Beltway/Georgia Avenue interchange and the Intersection of Death. The Planning Department’s staff witnesses this congestion every day because their headquarters is at the southern tip of the corridor. A redesign of Georgia Avenue between 16th Street and Forest Glen Road is the top county priority for state projects requiring planning. The state announced it would undertake planning for this corridor back in January.
These four corridors, as well as the four listed yesterday, present the most urgent and intractable traffic problems in Montgomery County. The Planning Department should be able to dig into these corridors in great detail if they adopt the measurement system that I recommended in Part One. Some of the questions that need to be answered are:
1. Is some level of congestion actually desirable? Heavily populated down-county areas like Bethesda and Silver Spring have large numbers of pedestrians. Policies that speed cars may cause increases in pedestrian accidents.
2. Are there any transit improvements that could ease the congestion? Five of the eight corridors (Wisconsin, Rockville Pike, Colesville, and two on Georgia) are already near Metro stations. Could some of the traffic be headed to areas beyond Metro’s reach? If it is, Metro extensions (including bus rapid transit) may be warranted. US-29 would be a particularly good candidate.
3. What does this information mean for the Purple Line? Four of the eight corridors (Wisconsin, Connecticut, Colesville and Georgia) would be along or close to the route for the Purple Line. But the congestion measurements we have now apply to north-south traffic during evening rush. Would the Purple Line really be effective in alleviating north-south traffic? Perhaps a GPS test drive should be held on East-West Highway, which is the major auto connection between Bethesda and Silver Spring. If its congestion level matches that found in these eight bad corridors, that would strengthen the case for a direct transit connection.
Dear readers, never in the history of Maryland blogdom (or as far as I know, the mainstream media) has anyone suggested such a radical redesign of our traffic measurement system as I have. The Planning Department’s over-reliance on critical lane volumes without context has produced an incoherent, unreliable measurement system that is plagued with bad data. If we cannot correctly diagnose our problems, we will never arrive at a solution. Our current economic downturn will not last forever. When it ends, development will resume and traffic will get even worse. High gas prices will increase our cost of sitting in congestion. The time to re-evaluate, improve and plan for the inevitable is now.
Thursday, May 29, 2008
A Better Way to Measure Traffic, Part Two
In Part One, we discussed how GPS-based traffic measurement could work on a county-wide basis. The Planning Department has already done test drives on several corridors, plotting out distances, times and speeds in evening rush hour. When combined with critical lane volume (CLV) measurements, we can get a sense of which corridors are really the worst rush hour drives in Montgomery County.
In general, there are two types of problematic corridors. First, there are stretches of major roads characterized by high volumes and low speeds. Second, there are stretches characterized by modest volumes and low speeds. The modest-volume corridors are no less troubled than the high-volume corridors. In fact, some corridors may record lower volumes because chains of clogged intersections prevent large numbers of cars from getting through.
Using Planning’s preliminary data, we have identified four high-volume and four modest-volume corridors that all recorded low travel speeds. These are the true problem areas in Montgomery County, not the “most-congested intersection” list that is based on flawed, one-day CLV estimates.
Today, we report the worst high-volume corridors and tomorrow we report the worse modest-volume corridors. Unlike the Planning staff, we do not rank them; I am sure Wisconsin Avenue drivers are just as miserable as Georgia Avenue drivers like myself. For each corridor, we report the date, time and results for the test drives as well as any CLV measurements taken in its component intersections. We also report the policy area standard CLV. Any intersections with CLVs measured above their policy area standard are rated as “failing” by the Planning Department.
The Worst High Volume, Low Speed Corridors
The following four corridors are characterized by high auto volume (often exceeding allowable policy area standards) and low speeds. Lots of cars can and do get through these areas, but the experience is agonizing.
Connecticut Avenue, Bradley Lane to the Beltway
Test Run
6/4/07 at 6:20 PM, Northbound
2.01 Miles in 19.6 Minutes, 6.1 MPH
PM Critical Lane Volumes
Connecticut at Bradley: 1577 (Recorded on 3/17/04)
Connecticut at East-West Highway: 1829 (Recorded on 3/29/06), Exceeds Standard
Connecticut at Jones Bridge: 2017 (Recorded on 6/6/07), Exceeds Standard
Connecticut at Beltway: 1245 North, 1100 South (Recorded on 3/9-3/10/04)
Policy Area CLV Standard: 1600
Notes: This is the major commuter route between Chevy Chase, Kensington and Downtown D.C.
Georgia Avenue, Arcola Avenue to Randolph Road
Test Run
6/4/07 at 5:15 PM, Northbound
0.69 Miles in 8.03 Minutes, 5.2 MPH
PM Critical Lane Volumes
Georgia at Arcola: 1471 (Recorded on 2/23/06)
Georgia at Randolph: 1910 (Recorded on 2/23/06), Exceeds Standard
Policy Area CLV Standard: 1600
Notes: Georgia at Randolph had the second-highest AM CLV in the county. A grade-separated interchange at this intersection is the county’s top priority among state construction projects. The state announced it would fund the project in January 2008. The Glenmont Metro Station is just to the north.
Rockville Pike, Congressional Lane to Edmonston Drive
Test Run #1
6/7/07 at 5:42 PM, Northbound
0.69 Miles in 4.62 Minutes, 9.0 MPH
Test Run #2
6/7/07 at 6:02 PM, Northbound
0.46 Miles in 2.93 Minutes, 9.4 MPH
PM Critical Lane Volumes
Rockville Pike at Congressional: 1538 (Recorded on 6/3/04), Exceeds Standard
Rockville Pike at Edmonston: 1590 (Recorded on 10/13/04), Exceeds Standard
Policy Area CLV Standard: 1500
Notes: This corridor is just north of the Twinbrook Metro Station and is adjacent to the Woodmont Country Club.
Wisconsin Avenue, Battery Lane to Cedar Lane
Test Run
5/12/05 at 4:52 PM, Northbound
1.24 Miles in 11.0 Minutes, 6.8 MPH
PM Critical Lane Volumes
Wisconsin at Battery: 1745 (Recorded on 2/8/07)
Wisconsin at Jones Bridge: 1536 (Recorded on 12/22/05)
Wisconsin at Cedar: 1996 (Recorded on 9/7/06), Exceeds Standard
Policy Area CLV Standard: 1600 (1800 at Battery)
Notes: This corridor extends from the northern edge of Downtown Bethesda through the National Naval Medical Center and is just south of the Beltway interchange. The extremely-congested, high-volume traffic provides a strong justification for BRAC-related transportation work.
Tomorrow, we will look at corridors with modest Critical Lane Volumes and slow speeds.
Wednesday, May 28, 2008
A Better Way to Measure Traffic, Part One
In our previous two posts, we outlined the county’s reliance on Critical Lane Volume (CLV) to measure traffic and identify problem intersections. We identified two problems with this system. First, the county’s measurement of CLV at each intersection only once every four years or so guaranteed the proliferation of fluky measurements. Second, CLV has different meanings under different circumstances. A low CLV could mean few cars or it could mean that few cars are able to move through an intersection due to excessive congestion.
There is a better way to measure traffic than this.
Any longitudinal data system must accomplish three things. First, the users must have reality-based data that actually measure what the users intend to measure. Second, there must be large numbers of observations to build an adequate sample size. Third, identical procedures must be applied over and over again in different conditions to isolate out their effects. The Planning Department’s current traffic measurement system has none of these characteristics.
But there is hope for something better that is buried deep within the planners’ own documents. In the appendix to the 2008 Highway Mobility Report, the staff has added something new: actual test runs by cars equipped with GPS devices on major corridors in rush hour. On pages 69-82, the staff shows the results of rush hour test drives on Wisconsin Avenue, Rockville Pike, Frederick Road, Georgia Avenue, Norbeck Road, Colesville Road, Columbia Pike, Connecticut Avenue, Clopper Road and Great Seneca Highway. Yes, these drives were only taken on one day each in May 2005, May 2007 or June 2007. But the results are very informative – you can actually see which intersections cause low speeds, and which stretches of these corridors have free flow.
Here is the kind of simulation that reflects reality. People do not drive through problem intersections once every four years, as the CLVs measure. They drive through long, clogged corridors every day, often during rush hour. The challenge is how to collect more of this data so it is not subject to outlier results determined by bad weather, accidents or other unusual conditions.
The solution is to draw on the hundreds of thousands of real live drivers who navigate this county every day. The Planning Department should offer county residents temporary GPS units for their cars in return for a payment of $100 per month. (That is the equivalent of one to two weeks of free gas!) These units would record all driving information and store them in internal memory for download upon return to the staff. Planning could rotate the units among different residents, perhaps 50-100 different people every month. After one year, Planning would have an unrivaled database of tens of thousands of actual drives under every condition imaginable. Planning would be able to judge the performance of any major roadway in the county under any weather condition, on any day, at any time of the day or night. Additionally, matching this traffic data with the police department’s online traffic accident database would enable planners to see the impact of auto collisions on roadway performance for every major route in the county.
How much would all of this cost? GPS units rent for as little as $5 per day and Planning could negotiate a better deal for lots of them. If Planning allocated them to 50 different residents each month and paid them $100 per month, the total rental and payment cost would be $151,250 per year. If Planning allocated them to 100 residents, the cost would be $302,500. Add on the cost of one full-time employee to monitor the program and the cost would be roughly $300,000 to $450,000 per year. The Planning Department’s total budget is approximately $19 million and it is currently suffering from cuts.
But Planning does not necessarily have to request lots of extra money. Why not pay for this, at least in part, by cutting back on CLV measurement? As we have seen, critical lane volume estimates taken in isolation do not by themselves provide reliable measurements of traffic. Do we really need 422 fluky, seldom-updated CLVs, many of which are taken at tertiary intersections with low volume? Why not reduce the number of intersections measured by CLVs and use the money for reality-based GPS measures instead?
Until the Planning Department moves to this sort of system, its existing test drive data provides a tantalizing peek at how a real traffic measurement system could work. In Part Two, we will begin looking at some of the new data Planning has already gathered.
Tuesday, May 27, 2008
Defining Deviancy Down at the Intersection of Death, Part Two
In Part One, we described how Montgomery County’s Planning Department relies on Critical Lane Volume (CLV) to estimate congestion at intersections across the county. Any statistical system that relies on just one measure, taken very infrequently, with little collaborating information is prone to fluky data. And that has happened at the Georgia Avenue-Forest Glen Road intersection, lovingly referred to by its neighbors as the Intersection of Death.
According to the Planning Department, this intersection went from being the most congested in the county to falling below the county’s allowable congestion standard for its policy area. Why? Because its morning CLV declined by 26% between surveys taken on 8/28/03 and 6/6/07. They would like us to believe that one-quarter of our traffic congestion has magically disappeared even though there have been no major engineering changes at the intersection. Similarly, only four of the ten most-congested intersections reported in 2006 have returned to the current 2008 list. Have the six intersections that fell off the list been “fixed,” surpassed by others that have become worse or simply fallen victim to bad CLV measurements?
While the Gazette covered our objections to traffic measurement at Georgia and Forest Glen, they spent a bit more time discussing the intersection’s infamous nickname than exploring the underlying data issues. I presented the following case study of the intersection to the Planning Board two weeks ago. Just looking at this one intersection calls into question the statistical validity of how traffic is measured in this county.
*****
Testimony of Adam Pagnucco
Montgomery County Planning Board, 5/15/08
Item 8: Highway Mobility Report 2008
Good morning. My name is Adam Pagnucco. I am Chairman of the Forest Estates Community Association's Crossing Georgia Committee as well as my civic association's incoming county government liaison. Today I offer my observations on the planning staff's new Critical Lane Volume (CLV) estimates for the Georgia Avenue-Forest Glen Road intersection.
In the 2006 Highway Mobility Report, the Georgia-Forest Glen intersection was ranked as the most congested in the county with an AM CLV of 2106 and a PM CLV of 1643. Those counts were taken on 8/28/03. No one in the surrounding area was surprised. What was surprising were the new CLV counts taken on 6/6/07: an AM CLV of 1553 and a PM CLV of 1377. If those counts are to be believed, then congestion has decreased at this intersection by 26% in the morning and 16% in the evening in just four years.
Now let's remember the location of Georgia at Forest Glen. It is adjacent to the Forest Glen Metro station, one block away from the Beltway interchange and three blocks away from Holy Cross Hospital, the second-biggest hospital in the state. Any car coming from the north to the Beltway must pass through it. Most vehicles heading to Holy Cross, and all of them coming from the Beltway, must pass through it. And hundreds of pedestrians cross the street every day to use Metro. No other intersection in the county, and possibly the state, has this combination of characteristics.
The principal change in recent traffic conditions around the intersection was the 2005 expansion of Holy Cross Hospital. That expansion, the biggest in the hospital's history, added 210,000 square feet of new space to the facility. The hospital reports that it had 128,591 visitors in 2003 and 157,573 visitors in 2007, a 23% increase after the expansion was finished. But now the hospital claims that it is bursting at the seams and it intends to expand again. Its plan calls for a new parking garage that would hold 500-700 more cars, which the hospital says it needs because cars are stacking up in its existing garage. Given the phenomenal increases in patient visits and the hospital's need for another expansion so soon after its last one, how can anyone believe that traffic congestion at Georgia and Forest Glen has really declined by double digits in just four years?
But there is more. In 1995, the planning staff estimated CLV at Georgia and Forest Glen at 1511 in the morning and 1530 in the afternoon. If the 2007 CLV is to be believed, then current traffic congestion is only 3% worse in the morning and is actually 10% less in the evening than it was in 1995. That's right, we are told that evening rush has actually improved by 10% over the last 12 years. Considering the dramatic redevelopment of Downtown Silver Spring and the residential construction upcounty in that period of time, that is extremely difficult to believe.
Last December, I showed you our video of conditions at the intersection. Remember the sight of cars stacked up towards Wheaton as far as the eye can see? Remember the constant illegal left turns? Remember how the pedestrians had to maneuver past cars stopped in the crosswalk? Of course you remember and there is no need to show you that chaos again. As I recall, that video provoked quite a reaction in this room. Many things were said at the time, but I do not remember anyone saying, "Hmmm… now that intersection has great traffic flow!"
Perhaps the issue here is how CLVs are used for purposes of analysis. Critical Lane Volume is after all a volume measure. As the number of cars proceeding through an intersection goes up, the CLV goes up. Now imagine a perfectly gridlocked intersection. No cars can move. What would its CLV be? Exactly zero. After all, no cars would be able to get through. Is it possible that the Georgia-Forest Glen intersection's declining CLV indicates more congestion and not less?
Members of the Planning Board, it is highly unlikely that you will find anyone in my neighborhood who believes that the biggest expansion in the history of the state's second-biggest hospital has led to less traffic congestion over the last four years. It is equally unlikely that you will find anyone who believes that afternoon traffic flow on Georgia Avenue has actually improved since the mid-1990's. I hope that you will ask your staff to explain how their data contradicts the facts I have cited today along with plain common sense. I for one would like to hear their answer.
*****
Tomorrow, we will show you a better way to measure traffic.
Monday, May 26, 2008
Defining Deviancy Down at the Intersection of Death, Part One
Imagine if we measured the American economy by estimating Gross Domestic Product once every four years. Forget about measuring other things like employment, unemployment and inflation. And forget about taking monthly or quarterly measurements.
“That’s ridiculous!” you would yell. “How would we understand the different trends in the components of the economy? How would we track the ups and downs? How would we get a complete picture of what’s going on?”
And you’d be right. But guess what: that is how we measure traffic in Montgomery County.
Montgomery County’s Planning Department measures traffic by calculating Critical Lane Volume (CLV) at each of 422 intersections in the county. CLV is the maximum hourly sum of conflicting auto movements, both through traffic and turns, proceeding through an intersection. (You can see how the measurement is constructed on page 6-26 of the 2006 Edition of the Mass Highway Manual.) CLV is dependent on volume. An intersection without cars would have a CLV of zero. As traffic picks up, CLV rises. But in a perfectly gridlocked intersection, no cars would be able to move and CLV would go back down to zero.
The Planning Department measures CLVs at each intersection by sending out traffic surveyors to count cars in both the morning and evening rush. But since Planning has limited resources, the surveyors can only appear at each of the intersections every four years or so. What if the weather is bad? What if there’s an accident nearby? Too bad, the survey results are in. Whatever happened on that one day is assumed to be the case on every day for at least the next four years.
Some of these measurement days are a bit unusual. Of the 422 reported CLVs in the 2008 Highway Mobility Report, 22 were taken in December, 23 were taken in January, 30 were taken in February and 3 were taken in August. Are these representative months of the year for driving conditions? Also, 29 of the CLVs were taken in 2003, 14 were taken in 2002 and 6 were taken in 2001. Can these measurements really be compared to estimates made in 2006 and 2007?
Moreover, the Planning Department assumes that a high CLV means high congestion. In fact, a high CLV means an intersection is carrying a lot of traffic. Remember – it rises with volume. An extremely congested intersection is one where it takes a long time to get through. In that case, the CLV may actually be quite low. How about measuring average delay times or average speeds? The Planning Department reports average speeds in travel runs taken on MD-355, Georgia Avenue, US-29, Norbeck Road, Connecticut Avenue, Clopper Road and Great Seneca Highway in 2005 and 2007, but those runs were one-day spot-checks covering only a fraction of the county’s intersections. Nevertheless, these travel runs provided valuable data and we will cover them in more detail later this week.
So why should you care whether Planning measures traffic badly? For one thing, an intersection’s CLV plays a role in determining whether developers building nearby will have to pay for traffic mitigation measures. If your CLV was measured on a day in which traffic was abnormally low, it might fall below the allowable traffic standard in that policy area. That would make it easier for a developer to escape responsibility for mitigating the effects of any new traffic generated by additional construction.
In Part Two, we’ll look at how the county’s faulty traffic measures have played out at the intersection of Georgia Avenue and Forest Glen Road, also known as the Intersection of Death.
Traffic Week on Maryland Politics Watch
Dear readers, I am going to tell you something you already know: traffic congestion has reached catastrophic levels in Montgomery County. We measure it badly. We do not plan around it very well. But we can do better.
In a special five-part series, MPW offers an unprecedented look at our county's traffic measurement system. Today and tomorrow, we look at how we currently measure traffic and why we come up short. On Wednesday, Thursday and Friday we propose an alternative measurement system and preview its results. Our recommendations are radical. They are sure to be resisted by the county's Planning Department. But radical problems demand radical solutions. Our plight is truly dire. Let the timid flee, let the entrenched bureaucracy fossilize and let the naysayers be bound and gagged! None of them will be spared from Traffic Week on Maryland Politics Watch.
