View
231
Download
0
Category
Preview:
Citation preview
8/2/2019 TRB Paper - Final
1/13
Billings, Garrick and Lownes 1
INTRODUCTION1The late 1950s and 1960s were the peak period of freeway construction in North America. At2the time, these freeways were seen not only as a means of increasing mobility between cities but3
also as a way of relieving congestion in cities. This meant that many miles of freeways were4
constructed directly through dense urban centers. In order to construct these freeways, thousands5
of citizens had to be relocated and large swaths of urban land were cleared of buildings. For6 these reasons, citizens protested the further expansion of freeways in many cities including San7
Francisco, Boston, New York and Washington, DC. These protest movements stopped some8expansions of urban freeways, but not before numerous freeways were constructed in hundreds9
of North American cities.10
Now that many of these freeways are reaching the end of their lifespan, governments are11
faced with a critical decision. Repairing or rebuilding urban freeways can be very expensive and12it is becoming increasingly difficult for governments to find the funds for such a major13
undertaking. An emerging idea that is taking hold in scores of cities across the country is that of14
eliminating sections of freeways from within the urban fabric of cities as a means of removing a15 barrier and freeing up land for redevelopment. Often the plan involves replacing the freeway16
with a lower speed boulevard, which can enhance access in the local area. The cost of removing17 the freeway is typically much lower than the cost of rebuilding it, and therefore removal is an18attractive option to governments from an economic standpoint.19
In many cases, however, the idea of freeway removal meets resistance because of20
concerns that reduction in road capacity will have significant negative effects on traffic locally21and throughout the city. Limited empirical evidence exists to explain how removing a link or22
significantly reducing link capacity affects the performance of a street network. However, a23
phenomenon known as Braess Paradox purports to explain how in certain situations adding a24
link to a network can actually decrease the overall efficiency of a network by increasing the25travel cost for the users and the system as a whole. If adding a link can theoretically decrease26
efficiency, it is possible that inversely, removing a link can increase efficiency.27In this study we examine empirical evidence of Braess Paradox in three different case28
studies of urban freeway removal: two located within the city of San Francisco and one located29
within the city of Milwaukee. The street networks surrounding these freeways were analyzed30over time to determine the change in volume-capacity ratios. These ratios were used to31
qualitatively test the existence of Braess Paradox in these cases, serving as indicators of severe32
traffic congestion (and thereby reduced system performance) after freeways were removed.3334
BACKGROUND3536
Braess Paradox37The line of thinking in conventional traffic engineering is that when congestion becomes an issue38
build more roads to relieve it. This thought process was challenged, however, in the 1960s with39
the realization that a Nash Equilibrium (or in terms of traffic networks, a user equilibrium40assignment) is not necessarily optimal from a system performance perspective (1). This concept41
states that adding additional links to a street network can actually result in decreased network42
efficiency thereby increasing the total travel cost for the system. However, this concept did not43immediately influence planning and design decisions as the freeway boomed continued through44
the 1960s and into the 1970s. But now many cities are beginning to realize that adding lanes to45
their existing freeways and even constructing new freeways has not resulted in a decrease in46
8/2/2019 TRB Paper - Final
2/13
Billings, Garrick and Lownes 2
traffic congestion. The realization that new lanes and roads do not necessarily decrease traffic1
congestion brings Braess Paradox back to the forefront of the conversation. Sheffi (2)2demonstrates Braess Paradox with a single-origin, single-destination network with four links3
and two paths. This network is then modified by adding a fifth link and the travel costs are then4
compared. In this example, the travel time for the system and the individual travel time5
increased 10.8% when the fifth link was added. It is important to note that this phenomenon6 does not always occur. Sometimes adding a road can improve the performance of a network, but7
it does point out that the operation of a network is dependent upon the characteristics of the8network in question. This concept aims to get engineers and planners to examine the9
characteristics of a network more carefully before simply adding a link to an already congested10
network as the intended results may not be achieved (2).11
Removing links from a network comes into the picture when looking at Braess Paradox12from a different angle. If adding a link increases travel cost, what happens when a high capacity13
link is removed from the network? Braess Paradox indicates that the network may operate more14
efficiently when the link is removed, but the common fear is that traffic pandemonium would15ensue as there would not be enough capacity to accommodate the redistribution of traffic from16
the removed link. Empirical data is required to determine if severe traffic congestion is the more17 likely outcome of urban freeway removal.1819
Induced Demand20The explanation of Braess Paradox brings up a related question: why does traffic congestion not21decrease when more roads and lanes are built? As stated earlier, conventional traffic design22
philosophy is based on the fact that building more lanes and roads will help alleviate traffic23
congestion. However, many cities that have followed this philosophy are finding that their new,24
wider roads are just as congested as they were prior to being built. Taylor (3) explains one25reason for this by defining the concept of induced demand. When a new lane or road opens,26
traffic congestion initially decreases which results in an increase in travel speed and ultimately a27decrease in travel time for the users. This decrease in travel cost, however, encourages other28
users who previously avoided this particular road because of the congestion to now use it, thus29
resulting in an increase in congestion. Therefore, after a new lane or road is opened, congestion30will often rebound to similar levels that were seen prior to the capacity expansion.31
From an urban freeway removal standpoint, the question is how does traffic in the area32
surrounding the freeway change when capacity is reduced? Based on the inverse of the induced33demand concept, the assumption would be that traffic in the area would decrease as drivers34
encounter the reduced capacity. Drivers would choose alternative routes to avoid the congestion35
that is feared from the reduced capacity. This paper will also assess the applicability of this36assumption.37
38
Case Studies39To assess the applicability of Braess Paradox, three urban freeways were selected as case40studies: two in San Francisco and one in Milwaukee. These freeways were selected because41
they had sufficient traffic data available for analysis for time periods before, during and after the42
capacity removal. The availability of traffic data is the cornerstone of this analysis and was43therefore the driving force behind which case studies were ultimately selected for analysis.44
45
46
8/2/2019 TRB Paper - Final
3/13
Billings, Garrick and Lownes 3
Embarcadero Freeway San Francisco, CA1
Like most other US cities, San Francisco was not immune to the freeway building boom of the21950s and 1960s. The original freeway master plan for San Francisco was crafted in 1948 by the3
California Department of Highways and included miles of freeways that would surround and4
cross the city. However, after residents saw the destructive results that the first sections of5
freeways brought to their city, protests ensued and many freeway projects were either only6 partially completed or cancelled altogether.7
The Embarcadero Freeway was planned to be a double deck freeway running along the8northern waterfront to connect the Golden Gate Bridge with the Oakland Bay Bridge.9
Ultimately, a 1.2 mile section of this freeway built in 1958 was the only part ever completed.10
This section of freeway was built above a surface street which also utilized the corridor. After it11
was built many citizens fought to have the structure removed as they considered it a barrier12dividing the city from the waterfront.13
For years this debate ensued with significant support on both sides of the argument, but14
the Loma Prieta earthquake of 1989 brought the issue to a critical decision point. Due to the15severe structural damage suffered during the earthquake, the Embarcadero was immediately16
closed to traffic. The debate over the fate of the Embarcadero after the earthquake was relatively17 straight forward as the costs of retrofitting the existing structure were financially crippling (4). A18 plan was made to replace the freeway with an at-grade boulevard, which would allow the19
reconnection of the waterfront to the city.20
The damaged freeway was torn down in 1991 and construction of the boulevard was21completed in 2000. It is important to note that this new boulevard functions as a standard part of22
the network. Turns are allowed to and from the boulevard at numerous locations, which allows it23
to blend in with the existing street network and the waterfront. In addition to reconnecting the24
city to the waterfront, this project allowed for the addition of a trolley line connecting the Castro25district to downtown San Francisco and Fishermans Wharf.26
Redevelopment that followed in the area included the remodeling of the historic Ferry27building (vacant for years prior to demolition of the freeway), construction of a multi-block retail28
and office center, development of the Rincon Hill and South Beach residential neighborhoods,29
and development of new recreational parks (5). At its peak, the Embarcadero freeway in30combination with the companion surface street carried approximately 100,000 vehicles per day.31
The replacement boulevard carries roughly 50,000 vehicles per day, 50% of the volume32
previously carried by the freeway (6,7).3334
Central Freeway San Francisco, CA35As was the case for the Embarcadero Freeway, only a portion of the Central Freeway was ever36constructed. Originally, the Central Freeway was designed to provide access to the Civic Center37
from US-101 and to continue north to ultimately connect to the Golden Gate Bridge. A 1.8 mile38
double deck spur west of I-80 was the only portion of this freeway ever constructed. Of this39
segment, 0.8 miles was constructed north of Market Street through the Hayes Valley residential40neighborhood.41
This freeway was also damaged by the Loma Prieta earthquake, but the process of42
removing this freeway was much more complicated than for the Embarcadero Freeway. The43northern most section of the freeway providing connections to Franklin and Gough Streets was44
demolished shortly after the earthquake. In 1996, the rest of the upper deck of the freeway north45
of Market Street was demolished due to structural issues which brought the issue of what to do46
8/2/2019 TRB Paper - Final
4/13
Billings, Garrick and Lownes 4
with the rest of the freeway to the forefront. Figure 1 shows the stages in which the freeway was1
demolished. Traffic was allowed to operate on all sections of the freeway shown in the various2stages of the figure.3
4
56
FIGURE 1 Central Freeway Phased Demolition78
Whereas proponents of removing the Embarcadero included the powerful businesses and9
individuals involved in the downtown waterfront, proponents of removing the Central Freeway10
mainly included residents of largely poor minority neighborhoods which surrounded the freeway.11
Thus, the process of garnering enough support to approve the removal of the Central Freeway12was extremely difficult (8). After much debate, a plan was finally approved in 1999 to replace13
the freeway with a multi-way boulevard named Octavia Boulevard. Multi-way boulevards differ14
from conventional boulevards by having the higher speed through traffic lanes separated by a15median from one-way slower speed local access lanes. Figure 2 illustrates this point by showing16
a typical intersection design of a multi-way boulevard.17
Before
Earthquake
Before
1996Before
2003
8/2/2019 TRB Paper - Final
5/13
Billings, Garrick and Lownes 5
1
23
FIGURE 2 Multi-way Boulevard Intersection4
(Source: https://reader009.{domain}/reader009/html5/0505/5aec973831233/5aec973b8cbfd.56
Based on the design, the integration of this boulevard with the surrounding street networks is7
much different than that of the Embarcadero. Once on Octavia Boulevard, left turns are8prohibited until the end of the boulevard at Fell Street. The reason given for this prohibition is9
related to the number of conflict points caused if left turns are allowed (9). Multi-way10
boulevards are essentially six way intersections as the local access roads are separated by11
medians from the main through lanes. Prohibiting left turns from the through lanes decreased the12number of conflict points thus, in theory, making the intersections safer (9). However, drivers13
are also forced to take less direct routes when accessing areas to their left, which makes the road14
a limited access thoroughfare more similar in operation (for access) to a freeway than to a typical15surface street. The overall design of this boulevard was also done as a compromise between16
commuters who utilized the highway and residents who live in the area.17
Replacing the freeway with a boulevard provided redevelopment opportunities for the18neighborhoods and a more pleasing environment in which to live. At the same time, this specific19
8/2/2019 TRB Paper - Final
6/13
Billings, Garrick and Lownes 6
design allows commuters a quicker trip through the area by limiting turns to and from the1
boulevard. Prior to the earthquake, the Central Freeway carried approximately 90,000 vehicles2per day. Today, Octavia Boulevard carries roughly 52,000 vehicles per day which is 57.8% of3
the volume previously carried by the freeway (6,7).4
5
Park East Freeway Milwaukee, WI6 Similar to San Francisco, Milwaukee also had a plan for surrounding and crisscrossing their7
downtown with freeways that was approved by voters in 1948. Though not part of the original8plan, the Park East Freeway was added to the plan ten years later. The freeway was intended to9
run from I-43 east to Lake Michigan where it would connect with the Lake Freeway and10
ultimately I-794 (10). Construction on the freeway began in the late 1960s and the first section11
of the freeway was opened to traffic in 1971.12Protests were mounted shortly thereafter by local communities and environmental13
activists with the primary concern being that the freeway would cut off Juneau Park from the14
waterfront. These protests were strong enough to cancel the rest of the project in 1972. Only 0.815miles of this elevated freeway were ever constructed. Before the project was canceled, the land16
necessary to complete the rest of the project had already been acquired. This land remained17 unutilized for 20 years until the 1990s when the state finally removed the lands designation as a18transportation corridor. This paved the way for the land to be redeveloped into a mixed-use19
community.20
The success of this redevelopment convinced Mayor John Norquist to start pushing to21remove urban freeways as a way to revitalize areas within downtown Milwaukee. Two freeways22
were originally considered for removal: I-794 and the Park East Freeway. I-794 carried roughly23
twice the traffic that the Park East Freeway carried, so the removal of the Park East Freeway24
came to be seen as a means of achieving urban renewal while minimizing the impact to the25existing transportation system. Ultimately, the freeway was demolished in 2002 and was26
replaced with McKinley Avenue by 2003. This boulevard functions similar to the Embarcadero27in the sense that it is well connected to the surrounding street network (10). Since the freeway28
was never completed to the waterfront, it was significantly underutilized and only saw ADT of29
54,000 vehicles per day at its peak. The replacement boulevard now carries 18,600 vehicles per30day, 34.4% of the volume previously carried by the freeway (11).31
32
RESEARCH APPROACH3334
Time Period and Traffic Data Selection35A key component of the analysis is understanding how traffic distribution changed before, during36and after the removal of the freeways. Therefore, time periods for the before, intermediate37
and after situations had to be selected for each case study. These time periods were selected38
based on the timeline for removal and the availability of average daily traffic (ADT) data.39
For the Embarcadero, limited data was available for the before situation as the Loma40Prieta earthquake occurred over 20 years ago. Only one source of ADT data was found for the41
before situation which was a report created by the San Francisco Department of Parking and42
Traffic in 1992 titled South of Market Cordon Count Pre-Earthquake and Current(5). Based43on this report, the before time period for the Embarcadero assessment was specified as 1988.44
This report was also the driving force behind street selection for the Embarcadero as the same45
streets needed to be used throughout all the time periods to ensure an accurate comparison.46
8/2/2019 TRB Paper - Final
7/13
Billings, Garrick and Lownes 7
Ultimately, there were 13 streets or freeway sections selected for the Embarcadero analysis. The1
intermediate period for the Embarcadero was selected as 1993 1999. During this period, the2freeway was completely removed, but the replacement boulevard was under construction. The3
construction of the boulevard was completed in phases, so portions of it were open to traffic at4
different times throughout this period. The after period was selected as 2000 2008 since the5
boulevard was fully completed in 2000. Data from the San Francisco Municipal Transportation6 Agency (SFMTA) was utilized for the intermediate and after periods (7).7
Time periods for the Central Freeway were more complicated as the process of removing8the freeway occurred in multiple stages over a range of years. More interest in the changes in the9
area around the Embarcadero was evident as no reports or data were available for the Central10
until the mid 1990s. Therefore, data was not obtained for a time period prior to the earthquake.11
It is important to note that Octavia Boulevard did not replace the entire length of the Central12Freeway. The length of freeway demolished shortly after the earthquake was not replaced by a13
road. Only the sections of freeway demolished after 1996 were replaced by the boulevard.14
Therefore, the before period was identified as 1995 1996 which corresponds to the period15right before the first part of the freeway that would ultimately be replaced by Octavia Boulevard16
was removed. The intermediate period was identified as 2005 as this was the time when the17 freeway was completely removed but the boulevard was not yet open to traffic. The after18period was identified as 2006 2008 after Octavia Boulevard was completed. ADT data for all19
three time periods was available for 9 streets/freeways from the SFMTA (7).20
The before period for the Park East Freeway was selected to be 1999 2001 when the21freeway was still in full operation. 2003 was selected as the intermediate period as the22
freeway was completely removed but the boulevard was not fully operational, though some23
sections were open to traffic. 2008 was selected as the after period as the boulevard was24
completely finished by this point in time. ADT data for all three time periods was available from25the Wisconsin Department of Transportation (11).26
27
ADT to Peak Hour Flow Conversion28ADT data cannot be compared directly to peak hour capacity data. The ADT data needed to be29
converted to estimated peak hour flow data in order to proceed with the comparison. Peak hour30flow data was unavailable for the streets under investigation, so a design hour factor had to be31
utilized. Design hour factors identify the portion of ADT which is included in the peak hour of32
traffic flow. AASHTO recommends k-factors in the range of 0.08 0.12 for urban areas and the33Highway Capacity Manual (HCM) recommends a default value of 0.09 be used. The upper end34
values of this range are more appropriate for cities that see travel mainly during peak times (i.e.35
commuter cities). Because San Francisco consists of a combination of commuters from the36suburbs and those living and working within the city limits, the default value of 0.09 was chosen.37
However, a value of 0.12 was selected for Milwaukee because the city has more of a suburban38
demographic than San Francisco. Multiplying this value by the ADT data resulted in peak hour39
traffic flow data.4041
Capacity Data42For San Francisco, the SFMTA utilizes a Synchro traffic flow model, which can be used to43estimate the capacity of any given section in the road network. Capacities for all streets utilized44
in this analysis were obtained from this model. Capacities for the Central and Embarcadero45
Freeways, however, had to be estimated separately as these roads no longer exist and therefore46
8/2/2019 TRB Paper - Final
8/13
Billings, Garrick and Lownes 8
there is no capacity data available. The HCM identifies a detailed process for estimating1
multilane highway capacity which involves utilizing factors to adjust for elements such as lane2width, access points and percentage of heavy vehicles. Data of this detail was unavailable for3
these freeways, so the equation to calculate base capacity, without adjustment factors, fromHCM4
exhibit 21-3 was used.5
6 Base Capacity = 1000 + 20*FFS, FFS
8/2/2019 TRB Paper - Final
9/13
Billings, Garrick and Lownes 9
12
FIGURE 3 Embarcadero Corridor V/C Analysis34
Prior to the removal of the freeway, it can be seen that the distribution of traffic was very5 concentrated on the freeway and very low on the neighboring street network. This indicates6significant traffic congestion on the freeway and minimal congestion on the surrounding urban7
street network. However, after the freeway was removed, the distribution of traffic throughout8
the area actually became more balanced. The concentration of the traffic on the replacement9
boulevard decreased significantly while the concentration on many of the other streets increased10slightly. The removal of the freeway eliminated the high V/C ratio peaks thereby supporting11
Braess Paradox as removing the link did not cause severe traffic congestion, and may have in12fact improved performance from a system perspective. In other words, San Francisco was able13
to utilize existing excess capacity to replace some of the lost freeway capacity. This figure also14
shows that the intermediate and after traffic distributions are essentially the same which suggests15
that the removal of the freeway was the primary driving force behind the redistribution.16 The results of the V/C analysis for the Central Freeway are displayed in Figure 4.17
18
Freeway/Boulevard
8/2/2019 TRB Paper - Final
10/13
Billings, Garrick and Lownes 10
12
FIGURE 4 Central Corridor V/C Analysis34
Similar to the Embarcadero, the area around the Central Freeway experienced very high traffic5 concentrations at the freeway and lower concentrations on the local streets. Oak and Fell Street6
did see fairly high concentrations as well, but these streets led directly to or from Central7
Freeway ramps so this is understandable. When the freeway was removed and replaced with a8
boulevard, the traffic distribution did not change like the situation with the Embarcadero. In fact,9the distribution remained essentially the same with the boulevard having a very high traffic10
concentration and the local streets having significant excess capacity. The distribution did11
change during the intermediate period as neither the freeway nor the boulevard were available12for use during this time. Since the V/C ratio distribution remained the same, it can be inferred13
that traffic congestion in the area did not change significantly. Even though removing the14
boulevard did not improve the situation by lowering the V/C ratios, it did not make the situation15worse. Therefore, these results also offer support for Braess Paradox. Another interesting point16
is that the intermediate distribution appears more balanced as there are no significantly high17
peaks like in the before and after distributions. This suggests that not just freeways, but roads18that operate similar to freeways can make a network operate in an unbalanced fashion. When no19
roads similar in operation to freeways were present in the network (intermediate period) the20
distribution was much more balanced with no excessively high or low V/C ratios.21
The results of the V/C analysis for the Park East Freeway are shown in Figure 5.2223
Freeway/Boulevard
8/2/2019 TRB Paper - Final
11/13
Billings, Garrick and Lownes 11
12
FIGURE 5 Park East Corridor V/C Analysis34
The Park East Freeway is different from the two San Francisco case studies in the fact that the5 freeway was not experiencing high V/C ratios. Even though this freeway was underutilized, its6
removal still caused the distribution of traffic in the area to change. The intermediate data shows7
that significant increases were experienced on the eastern side of the freeway corridor (streets go8
from western-most to eastern-most when reading from left to right) and significant decreases9were experienced on the freeway/boulevard. Once the boulevard was fully opened to traffic,10
traffic rebounded on the western side of the corridor and dipped back down to levels similar to11
the before period in the eastern section of the corridor. Overall, the corridor did not see any12excessively high V/C ratios after the freeway was removed, and the distribution was fairly13
similar to the pre-removal situation. Therefore, the results of this case study lends some support14
to Braess Paradox, in that by this measure there was no significant decline in system15performance due to the link removal.16
Removing the urban freeway segments produced different results in each of the three17
case studies. The Embarcadero Freeway corridor started off with very high V/C ratios on the18freeway and low V/C ratios everywhere else. The removal of the freeway made the V/C ratios19
more balanced thereby making the capacity utilization of the network more balanced. The20
Central Freeway corridor started off similar to the Embarcadero but remain unchanged when the21
freeway was replaced with the boulevard. One possible explanation for this is the difference in22road design from the other two case studies. The replacement boulevard operates similar to a23
Freeway/Boulevard
8/2/2019 TRB Paper - Final
12/13
Billings, Garrick and Lownes 12
freeway which means that the users ability to access the network remains limited as it was when1
the freeway was in operation. The Park East Freeway corridor started out with relatively low2V/C ratios throughout and remained in this state when the freeway was replaced by a boulevard.3
Since this corridor already experienced low V/C ratios, it would be difficult to significantly4
change the balance of capacity utilization in the network. It is important to note, however, that5
none of the case studies experienced increases in V/C ratios after the removal that would be6 indicative of more traffic congestion than was experienced in the area when the freeway was7
present. The V/C ratio distribution either remained the same or became more balanced when the8high capacity link was removed.9
It is also important to note that traffic in the area surrounding the Embarcadero Freeway10
increased by approximately 24,000 vehicles per day (9.0%) from the before to the after time11
period. This goes against the induced demand based assumption mentioned earlier in this paper12that when capacity in an area is reduced, traffic in the area would also decrease. However, this13
increase may be able to be explained when looking at the results of the Central Freeway analysis.14
Traffic in this area decreased by approximately 60,000 vehicles per day (22.5%) from the15before to the after time period. Since the Embarcadero was removed and replaced with a16
boulevard first, it is quite possible that some traffic shifted from the Central Freeway corridor to17 the Embarcadero Freeway corridor and never returned. For the Park East Freeway, traffic18decreased by approximately 37,000 vehicles per day (18.3%) from the before to the after19
time period. Therefore, the results show that two of the case studies support the induced20
demand-based assumption that traffic decreases when capacity decreases. The Embarcadero21may have supported the assumption as the increase in traffic may have been caused by drivers of22
the Central corridor switching over to the Embarcadero corridor; however data was not available23
to support this hypothesis.24
25
CONCLUSIONS26As our road infrastructure ages, more and more cities will be faced with the decision of what to27do with urban freeways and other high capacity roads within their urban core. With local28
budgets tightening, it is likely that more cities will begin considering the removal of these roads29
as a way to save on costly maintenance and rebuilding costs. From a community standpoint,30removing these roads is seen as a way of freeing up land for redevelopment to help revitalize the31
neighborhoods. Without an adequate understanding of how removing a high capacity link32
affects the capacity utilization of a network, it might be difficult for municipalities to garner33enough support to proceed with these projects. This project has shown how the V/C ratio34
distribution within three case study areas changed when segments of urban freeways were35
removed.36Initially, both areas in San Francisco had high traffic concentrations on the freeways and37
significant unused capacity on the surrounding street network. The Milwaukee case study had38
significant unused capacity in both the freeway and urban streets as the freeway was39
underutilized. The V/C ratio distribution for the Embarcadero network became more balanced40and the distributions for the Park East and Central Freeways remained relatively unchanged. It is41
important to note that boulevard design may play a part in the exact nature of the change in V/C42
ratio distribution. The Embarcadero was replaced by a traditional boulevard that was well43connected to the surrounding network and actually saw a more balanced V/C ratio distribution.44
The Central Freeway, however, was replaced by a multi-way boulevard with limited network45
connectivity and saw the V/C ratio distribution remain unchanged. With only a limited number46
8/2/2019 TRB Paper - Final
13/13
Billings, Garrick and Lownes 13
of case studies it is difficult to form a solid conclusion on this, but these results do imply that1
boulevard design is significant.2Even with this implication, the main question still remains: how does the traffic3
distribution in a network change when a high capacity link is removed? Braess Paradox4
suggests that a network can actually operate more efficiently when a link is removed, but the5
common fear is that removing a link will cause gridlock. The results from this analysis suggest6 that the common fear of gridlock may be unwarranted, since the V/C ratio distribution did not7
get worse in any of the case studies.89
ACKNOWLEDGEMENTS10This project was sponsored by the New England University Transportation Center, the11
University of Connecticut Center for Transportation and Livable Systems, and the Dwight David12Eisenhower Transportation Fellowship Program.13
14
REFERENCES151. Braess, D., A. Nagurney, and T. Wakolbinger. On a Paradox of Traffic Planning.16
Transportation Science, Vol. 39, No. 4, 2005, pp. 446 450.17 2. Sheffi, Y. Urban Transportation Networks. Prentice Hall, Inc., New Jersey, 1985.183. Taylor, B. Rethinking Traffic Congestion. Access, No. 21, 2002, pp. 8 16.19
4. Hastrup, S. Battle for a Neighborhood. Places, Vol. 18, No. 2, 2006, pp. 66 71.20
5. San Francisco, CA: Embarcadero Freeway. The Preservation Institute.21http://www.preservenet.com/freeways/FreewaysEmbarcadero.html. Accessed22
October 1, 2009.23
6. South of Market Cordon Count Pre-Earthquake and Current. San Francisco24
Department of Parking and Traffic, 1992.257. Traffic Volume Information. San Francisco Municipal Transportation Agency.26
http://www.sfmta.com/cms/vhome/hometraffic.htm. Accessed June 5, 2010.278. Cervero, R., J. Kang, and K. Shively. From Elevated Freeways to Surface Boulevards:28
Neighborhood, Traffic, and Housing Price Impacts in San Francisco. University of29
California Transportation Center, December 2007.309. MacDonald, E. Building a Boulevard. Access, No. 28, 2006, pp. 2 9.31
10. Milwaukee, WI: Park East Freeway. The Preservation Institute.32
http://www.preservenet.com/freeways/FreewaysParkEast.html. Accessed33October 1, 2009.34
11. Milwaukee County Traffic Counts. Wisconsin Department of Transportation.35
http://www.dot.wisconsin.gov/travel/counts/milwaukee.htm. Accessed April 10, 2010.3637
LIST OF FIGURES38FIGURE 1 Central Freeway Phased Demolition39
FIGURE 2 Multi-way Boulevard Intersection40FIGURE 3 Embarcadero Corridor V/C Analysis41
FIGURE 4 Central Corridor V/C Analysis42
FIGURE 5 Park East Corridor V/C Analysis4344
Recommended