Understanding the Austin Transportation Department’s
West Riverside Drive Traffic Analysis
by Larry Akers
Friends of the Parks of Austin Stakeholder Representative
Town Lake Park Community Project
Accept the findings. Reject the conclusion.
In their West Riverside Drive Traffic Analysis, which considers the impact of closing a portion of Riverside Drive in Town Lake Park between Butler Park and Auditorium Shores, the City of Austin Traffic Department (TD) would have you believe their analysis of the impact of closing a short section of Riverside Drive in the heart of Town Lake Park would have a seriously negative effect on traffic flow. Yet underneath the obfuscation in the presentation of their report, their analysis shows the closure would induce VIRTUALLY NO NOTICEABLE DELAY in the commuting system. This is made clear simply by dividing the ominous “vehicle hours of delay” by the number of vehicles involved. This data is presented below, and it correlates excellently with analysis in previous traffic studies.
Moreover, in its corridor discussion, the report misrepresents east-west delays by not attributing increased net delay to the increased net traffic through those corridors. The findings show be no increased per-car delay.
Furthermore, the data in the report suggests the possibility of scenarios that may be even more favorable to closure than the ones analyzed.
The report has a serious case of confirmation bias. Even before TD did any analysis they took the position that closing Riverside would be adverse to transportation needs in the area. This pre-conception is embedded throughout their report, where its phrasing is tilted toward supporting the department’s predetermined conclusion that Riverside should not be closed to commuters. This conclusion was expressed prior to conducting any analysis in a July, 2011 memo (see appendix) stating that the roadway should be retained for commuters. However, a detailed examination shows that this conclusion is not supported by the data and findings of the study. Moreover, the report presents numerous outright falsehoods and dubious claims to frame the discussion. Some of these are discussed below.
The TD report also misstates Council direction and fails to perform the demanded analysis of well-known measures (see appendix) proposed to mitigate the absorption of Riverside traffic into the rest of the grid. New capacity can be readily developed, contrary to the report’s assertion.
Finally, the report pursues a specific value-based judgment that should not be the role of an engineering study, and in so doing it completely ignores the community values that weigh in favor of closing this small section of roadway in the middle of our new central park.
Interpreting the Scenarios
One troubling aspect of the study is the assumption that the traffic displaced by Riverside closure would be evenly redistributed between Chavez and Barton Springs. Yet Scenario 2 (60% Chavez/40% Barton Springs) yielded very significantly less delay than the 50/50 split. Statistically, drivers adjust rationally over time toward the optimal, so of the three scenarios, the 60/40 is most inevitable. However, that being the case, it should have been chosen as the central redistribution assumption and given the same kind of sensitivity variance as the 50/50 split. What would the study show, then, for a 70/30 distribution? That would be a much more reasonable and valuable assessment than the highly unlikely 40/60 scenario, which the presentation uses to skew the overall picture.
Considering the inevitability of drivers adjusting to the more optimal traffic pattern, there is no point in evaluating any scenario other than Scenario 2, 60% Chavez/40% Barton Springs. The others are simply unrealistic. That is where we shall proceed from here.
Viewing the Results in Understandable Terms
The presentation of delay factors in terms of total vehicle hours is a way of disguising the impact on drivers. The important measure is delay per vehicle trip. Traffic volumes in some cases are given in the report, while others have been presented to project stakeholders in a separate document. Applying this traffic data, the conversions of total intersection delay to delay per vehicle are simple and informative. Without having precise data on all the intersections in question, aggregating the data is one way to expose the real impact to drivers.
According to the study, traffic volumes across the bridges during peak hours are as follows:
Cross-River Traffic Volume
- Lamar — AM-2931 PM-3825
- S. First — AM-2644 PM-3618
- Total — AM-5575 PM-7443
We can believe each number in these counts represents a unique vehicle, as double crossings are unlikely and in any case would constitute separate trips. These numbers do not include traffic entering the grid and traversing east-west without crossing the river, so the actual aggregate flow through the grid is significantly higher than the cross-river volumes.
Nevertheless if we apply these numbers to the delay hours to calculate average delay per vehicle trip, the impact of closure is much clearer:
AM (min. 5575 vehicles)
- Scenario 1 — 4.1 vehicle hrs
- Scenario 2 (likely) — 1.5 vehicle hrs
- Scenario 3 (unlikely) — 7.2 vehicle hrs
PM (min. 7443 vehicles)
- Scenario 1 14.0 vehicle hrs
- Scenario 2 (likely) -0.3 vehicle hrs
- Scenario 3 (unlikely) 33.9 vehicle hrs
Dividing these aggregate delays by the number of vehicles gives:
Delay per Vehicle(seconds)
AM (min. 5575 vehicles)
- Scenario 1 — 2.6 seconds
- Scenario 2 (likely) — .96 second
- Scenario 3 (unlikely) — 4.6 seconds
PM (min. 7443 vehicles)
- Scenario 1 — 6.8 seconds
- Scenario 2 (likely) — -.15 second
- Scenario 3 (unlikely) — 16.4 seconds
Under Scenario 2, the most likely, additional morning delay induced by Riverside closure is less than one second per trip, and afternoon performance actually improves slightly.
Thus there is NO significant induced delay, even averaging over only a portion of the vehicle trips.
Additional traffic count data, taken from a somewhat more comprehensive summary presented to project stakeholders (Appendix 2), gives the following numbers for east-west peak traffic volumes.
East-West Corridor Traffic Volumes
- Barton Springs AM-1650 PM-1902
- Cesar Chavez AM 2226 PM-2357
Adding these east-west corridor volumes into the delay computation reduces the computed average delay per vehicle.
Delay per Vehicle (seconds)
AM (max. 10023 vehicles)
- Scenario 1 1.5 seconds
- Scenario 2 (likely) .5 second
- Scenario 3 (unlikely) 2.6 seconds
PM (max. 12030 vehicles)
- Scenario 1 4.2 seconds
- Scenario 2 (likely) -.1 seconds
- Scenario 3 (unlikely) 10.1 seconds
Given that the number of river crossings establishes a lower bound for the number of vehicle trips, and adding in the east-west corridor numbers nears an upper bound (though the upper bound would be reached only by adding cars traversing Riverside). The average delay per vehicle will lie in the range between the two tables.
The Traffic Department’s presentation of vehicle hours hides the fact that the model shows virtually no perceivable aggregate intersection delay induced by the closure of the road at these intersections.
What, then, of the corridor delays presented in the study? The north-south corridors (South First Street and South Lamar), and their result is as might be intuited. Under Scenario 2, the S. First Street corridor AM peak delay, measured only against vehicles crossing the river, is 4.8 seconds, and Lamar improves by 2.5 seconds, essentially offsetting the induced delay. At PM peak, the South First delay is 12.9 seconds per vehicle, while Lamar improves by 12.3 seconds. Again these numbers essentially offset each other.
But what about those scary east-west corridor numbers? The problem is they are completely misleading. Yes, the cumulative delay times through the intersections in those corridors increase a bit. But that is because those corridors have absorbed all the traffic formerly travelling through the Riverside corridor, including all the delay time experienced by those vehicles. There is no evidence in the report that the per-car delay through the east-west corridors has increased. The selective highlighting of the increased net delay, while not stating that this delay is attributable to the additional vehicles, creates a bogus impression that the level of service has deteriorated. This is plainly dishonest. Indeed, the total east-west delays displayed in the base modeling data from Tables 2 and 3 indicate no net increase in east-west delay within the system.
The Traffic Department wished only to justify their predetermined conclusion that closure of Riverside would induce unacceptable delay for commuters. But since two of the most troublesome intersections are made more efficient, travel time is maintained at virtually current levels, according to this modeling. Having this analytical finding, the authors of the report attempted to disguise it by presenting the numbers in a format whose abstraction (total vehicle hours of delay) hid the reality, and then went further, hiding a portion of the corridor travel times from the presentation of east-west corridor analysis.
The reality of their finding is VIRTUALLY NO INDUCED DELAY, only a slight shifting of delay between intersections and corridors.
Comparison with Other Studies
Except for TD’s conclusions and manner of presentation, the technical results of this study appear to be analytically satisfactory. They correlate very closely with previous, broader studies of the same area. The Traffic Impact Analysis performed by WHM Associates for the 1999 Master Plan indicated closing Riverside Drive would induce only a six second delay to a morning commute and a seven second delay in the evening. That modeling was based on westbound traffic volume of 1200 vehicles per hour on Riverside at PM peak hour. The current volume is 526 vehicles in the evening peak. One would expect those minor induced delays to have fallen, since the actual volume of Riverside traffic needing to be absorbed has fallen by over half relative to the study. The WHM report also forecast slightly improved service in the Lamar corridor and slightly worse service along South First, which correlates with the TD report’s finding.
Moreover, the DAMP study from 2002, which used even higher traffic projections, found that eliminating all eastbound flow on Riverside would streamline the commuting grid, a finding that is reinforced in the Transportation Department study.
False and Dubious Claims in the Presentation
In an attempt to slant the presentation toward its predetermined conclusion, the report makes a number of misleading statements and includes some outright falsehoods.
It begins in the cover letter: “The Austin City Council directed staff to maintain West Riverside Drive until alternative capacity could be identified in either the Barton Springs Road or Cesar Chavez corridors.” Not really. The direction in Council Resolution 020215-66C was “Continue to develop alternatives that would make viable the removal of Riverside Drive as an at-grade roadway through Town Lake Park. (Riverside Drive shall not be permanently closed through the Park until such time as alternatives are implemented to mitigate Riverside Drive’s lost traffic capacity.)” Alternatives were not explored in TD’s study, even though they were known to the department. Mitigation, which could be partial, was the goal of the Council resolution, not necessarily complete replacement of capacity.
The study states, “no other capacity options have been identified over the years.” Completely false. Options suggested and well known to both City Council and the Transportation Department include creating a reversible lane on Drake Bridge, a double left-turn option from Drake Bridge onto westbound Cesar Chavez, and a lengthy list (see Appendix 3) of other options drawn up by the same Transportation Department staff for study as mitigation measures.
Other TD Study statements are as follows:
“Delays generated during peak hours .. are likely replicated throughout the day.” That is hogwash. When streets are not clogged with peak hour traffic, delay is induced only by traffic signals, which are governed by the engineers. As there is virtually no induced delay anyway, the argument becomes moot.
“W. Riverside Drive serves as a primary access route to the entertainment assets such as the PEC, Long and Auditorium Shores”. This is true only for westbound entrance from S. First, which would not be affected by closure. Eastbound approach via Riverside from Lamar is impossible for southbound traffic coming off the Lamar Bridge. Approaches from the south and west are shorter and easier via Barton Springs. There is no reason to use Riverside for access to events.
A “closed Riverside Drive would not be preferable from an emergency response basis”. False. The assumption behind this statement is lack of a paved surface, but retaining a paved surface has been a part of all informed discussions on the Riverside closure for some time. Retaining a drivable hardscape surface between the east and west terminii of Riverside would serve two purposes: 1) emergency vehicle access, and 2) provision for a festival plaza that could host special events now conducted on (and shutting down) downtown streets. Providing this venue would vastly improve access to downtown during these events. Regarding emergency access in particular, it might be improved, since there would be no traffic impediment to the entry and passage of emergency vehicles. There would be no need for the “circuitous routing for emergency responders” cited in the TD report.
Guests departing Long Center events cause “traffic impacts to Barton Hills”. False. At that hour, there is no noticeable impact on access to Barton Hills or, for that matter, the Zilker neighborhood.
“The ceremonial entrance to the PEC is underutilized because of the current design of the circular drive.” False. The current entrance is underutilized not because of the design of Riverside and the driveway, but because the PEC usually locks to north doors to control access. There is nothing wrong with the driveway access.
An alternative design of a retained roadway would allow “improved access to critical parking infrastructure on .. the west end of the park”. False. What critical parking infrastructure? There are only two small surface lots. These are adequately served via a Lamar entrance. Closing the road would actually allow for more parking by allowing adaptive re-use of its right-of-way. Otherwise more parking could be provided only by consuming very precious parkland.
Retaining the road would allow “Improved loading and unloading of events at area venues.” A dubious claim. Long Center and PEC loading is from their joint service yard, a “Cadillac facility” (according to Convention Center Department Director Mark Tester) that is easily reachable from Barton Springs Road. Improved access to the Auditorium Shores performance area as envisioned by the trail head and Auditorium Shores improvements by TBG will not involve the closure section, but will be improved rather by renovations on the site that are now in progress.
Retaining the road would allow “a pedestrian zone that is consistent with the park concept and activities currently and planned at the park”. False. The park was shrunk by the placement of the PEC, in what was to be green space, to accommodate the PEC/Long Center service yard. Using the Riverside ROW for park features to compensate for this loss was understood by the hundreds of Master Plan participants as critical to the success of the park. A core original design guideline was to eliminate vehicular impact on the heart of the park. A commuter road through the heart of the park is the very antithesis of that goal. Commuters passing through a heavy pedestrian zone at the daily period of peak park use is directly contradictory to the park concept and is a danger to all park users, but particularly children.
Interpreting the Evaluation — Positive Directions
What kind of credibility can the conclusion of a report have if its presentation is so biased, obfuscated, and false?
But putting it conclusion aside, the report points the way for solutions that would take the closure of the street from painless to a boon to commuters and possibly help resolve venue access and parking problems.
The report suggests that optimal post-closure performance would occur if a majority of Riverside traffic redirected itself toward Cesar Chavez. It would follow that making that route even more attractive would improve conditions further. A reversible lane on Drake Bridge, an idea that has been endorsed by RECA and the Downtown Austin Alliance, among others, would create extra capacity along the South First/Chavez route, particularly if it included a double left turn onto westbound Chavez.
To the extent that traffic delays may occur at the intersection of Barton Springs Road and South Lamar, there is one particularly easy way to provide new capacity. That intersection is designed for high capacity and functions well. But a bottleneck exists at the railroad and Bouldin Creek bridges on Barton Springs Road that prevents westbound traffic from readily reaching the intersection and blocks vehicles desiring to turn left onto Lamar from reaching the double turning bays. A two-way left turn lane currently exists on Barton Springs from just east of South Lamar to Dawson. The only portion of this lane that is used is a short section providing left turns into the park and Daugherty Arts Center parking lots. The remainder of this lane could be converted to a westbound traffic lane, relieving the bottleneck and signficantly increasing capacity through the area. Despite a statement to the contrary in the TD report, the Traffic Department is aware of this mitigation step.
The report does mention reconfiguration of the Barton Springs/ Riverside/South First triangle, a complex notion developed by Transportation Department engineers, and one which was on the list of measures to be modeled, but which has not been analytically modeled to date. The triangle treatment could potentially reduce travel times and alleviate congestion in the South First and Barton Springs Road corridors.
The physical issues of access to the Long Center, PEC, and Auditorium Shores events were well studied and addressed in the venue master plan, and the operational problems those venues have encountered result almost exclusively from the failure to coordinate schedules among the venues.
The Long Center gained a great deal in the agreement it and its constituent groups negotiated in the Town Lake Park Venue Project Master Plan — a superb service yard, resulting the consumption of much of the heart of the Butler Park site by the PEC, in exchange for the closure of a short section of Riverside Drive to recover and maintain the safety, integrity, and critical mass of the park. All the Long Center lost in the agreement was westbound egress from one end of its parking garage, the end that was designed to serve South First access. That is a loss they should accept in exchange for value requested and received. Their operational concerns can and should be addressed, but not at the cost of abandoning an existentially critical aspect of Town Lake Park.
Despite the hidden data, questionable assumptions, false statements, and erroneous conclusions TD report places on its findings, the study does illustrate once again that Riverside Drive can be closed without significant impact on the commuting grid.
The study fails to address Council direction to model the extent to which some modest mitigation measures can increase the capacity of the area grid. The effectiveness of these measures is yet to be modeled, and should be, as they may more than compensate for any assumed delay induced by the proposed Riverside closure.
As for the induced delays, even those projected under the more unrealistic scenarios are well within the range of tolerance that was established with the adoption of some of the Great Streets projects downtown. The mentality exists that for significant enhancements in the quality, livability, and safety of the central city environment, some tradeoff against commuting efficiency is acceptable. Why is that consideration not applied in this case?
For our premier urban park, the tradeoffs are so small, and the benefit so great (park design, amenity, safety, and quality, remediation of downtown congestion by providing an alternative venue for street festivals, and public faith in the execution of negotiated agreements and master plans), that any recommendation other than closure is completely inconsistent with traffic decisions made north of the river and are out of character with the city we all envision.
After ten years of refusal by the Traffic Department to provide the analysis directed by Council resolution, we now have another report that shows the commuting delays induced by the closure of Riveside Drive would be neglible. We still have no study of the well-known suggestions for improving traffic flow in the area of the park. The Transportation Department has shown itself unwilling to conduct this study and unwilling to present findings in an unbiased manner and without drawing predetermined, unsupported conclusions.
The TD report provides renewed confidence that Riverside closure will not noticably delay commuters, but completing the refinement of the park’s Master Plan with confidence still requires an impartial second look at mitigation measures, parking, and venue access issues. For professional, objective, and impartial analysis, we must turn to an independent and unbiased professional traffic and parking engineering consultant. Park planners and City Council must understand all these related issues to support prudent decisions on Riverside and the future of our possibly great urban park.
Appendix 1: Rob Spillar Memo, July 6, 2011
Excerpt: “My perspective is that the capacity provided by Riverside is vitally important to the overall network both north and south of the river.” It continues in the mode of considering closing the road only for special events, but retaining it for commuting.
Appendix 2: 2011 Traffic Count Summary
Appendix 3: Suggested Mitigation Measures
Reversible lane proposed in the “Near Term CBD Transportation Improvement Projects” circa 2002 and supported by RECA, among others. Summarized in:
linked also from
Text version here:
Development of Mitigation Measures for Riverside Closure
This list of projects anticipates the use of a less sophisticated traffic simulation to allow a number of runs to be made quickly, with a goal of selected a “best” solution, that would then be modeling using CORSIM. The background will be the traffic for 2005 as per the DAMP model process. The mitigation measures will be in addition to those contained in the TPSD Staff recommendation contained in the Near Term Transportation Project list.
After review of the previous modeling work, the three critical areas for the evening peak hour are Barton Springs westbound at Lamar Boulevard, Cesar Chavez westbound from San Antonio to west of B. R. Reynolds and crossing South First Street.
Barton Springs westbound at Lamar Boulevard 1) Use the continuous left turn lane as a trap lane feeding the left turn bays at Lamar. This would begin just past Dawson. 2) Assume there is a third westbound lane on the north side of the street, which will be a trap lane feeding the right onto Lamar or allow traffic to use Lee Barton down to Riverside.
Crossing South First Street 1) At the intersection of Barton Springs and Riverside, assume that the left turns from westbound Riverside are prohibited. At South First, assume Riverside westbound has dual left turn lanes. 2) In the blocks between South First and where Riverside and Barton Springs cross, assume that Riverside is one way westbound and Barton Springs is one-way eastbound.
Cesar Chavez westbound from San Antonio to west of B. R. Reynolds 1) Assume that all of the signalized intersections use called pedestrian phases, therefore the regular phase for the side street would not need to be long enough for a pedestrian to cross the street. 2) Try restricting left turns from Cesar Chavez to B. R. Reynolds in the evening peak hour. Assume that traffic uses 5th Street to get to Lamar. 3) Try contra-flow lane on Cesar Chavez westbound from Lavaca to San Antonio. Cesar Chavez would then have 3 EB and 1 WB lanes in that section. <>
Reallocation of traffic using Riverside now: Although we can start with an assumption that half the traffic will use Barton Springs and half will go north of the river to Cesar Chavez. The final allocation should use the travel time from Riverside and Congress to Cesar Chavez west of B. R. Reynolds to rationalize the split. Is that possible? If not, how would you suggest we make the split?