Thursday, November 12, 2015

Precinct's Downtown Laneway: How Wide?

This posting is about the laneway proposed in Precinct's plans for redevelopment of the Auckland Downtown block bounded by Lower Queen Street, Quay Street, Lower Albert Street and Custom Street West. The picture below shows a rough plan of this laneway.....


The laneway I want to discuss here, is the one connecting Lower Albert Street through the development to Lower Queen Street. The laneway will be about 100 metres long if the whole of Queen Square (shaded in green) is sold for development, and about 60 metres long if Queen Square stays in public ownership as public space. (The other proposed laneway, running north-south, is more of a dog leg than shown here, as it has to side-step the HSBC building to get to Quay Street.)

In June this year, Precinct obtained resource consent for a development on the site that does not use Queen Square. This consent was non-notified. It includes consent for the 60 metre section of laneway. The question I want to address in this posting is how wide should this laneway be?

At the hearings that have occurred in the past couple of weeks, there have been varying widths discussed by the applicant's experts and by Auckland Council. The Road Stop notice stipulates a 5 metre width east-west laneway. Other witnesses have said 6.5 metres, and that this provides a 5 metre wide central section, allowing shoppers to look at the shops using the other 1.5 metres of width. Various submitters have raised questions about the safety of this - because the laneway will be used by passengers to walk between the bus terminal in Albert Street and the Britomart rail terminal or other buses that terminate there.

At the Plan Change 79 hearing, Blair Johnston of Warren and Mahoney Architects, (who are doing master plan and architecture work for Precinct) stated that he had worked with Auckland Transport, and that modelling showed that a 5 metre wide corridor was safe and appropriate, even up to conservative estimates of 16,000 people going through per hour.

The world authority on these matters is John J Fruin. You can see a summary of his work for New York Transit here. This is a bit dense, so I've unpacked relevant sections....
PEDESTRIAN CHARACTERISTICS
The primary characteristics needed to evaluate pedestrian facilities are walking speeds, walking distances, demand patterns, and traffic-flow relationships. The ability of pedestrians to select their own individual walking pace and speed is a qualitative measure of convenience. Walking distances define the effective service area of transit stations, with shorter distances improving passenger perceptions of service and convenience. The patterns of passenger demand affect the methods used to analyze pedestrian facilities and the applications of service standards.
WALKING SPEEDS
Pedestrian speeds, in addition to being directly related to traffic density, have been found to vary for a wide range of conditions, including individual age, sex, personal disabilities, environmental factors, and trip purpose. Normal walking speeds unimpeded by pedestrian crowding have been found to vary between 150 and 350 ft/min (0.76 and 1.76 m/s), with the average at about 270 ft/min. As a point of comparison, running the 4-min mile is equivalent to a speed of 1320 ft/min or almost 5 times normal walking speed. Walking speeds decline with age, particularly after age 65, but healthy older adults are capable of increasing their walking speed by 40% for short distances. Dense pedestrian traffic has the effect of reducing walking speed for all persons. The smaller personal space limits pacing distances and the ability to pass slower moving pedestrians or to cross the traffic stream.
Photographic studies of pedestrian traffic flow on walkways have shown that individual area occupancies of at least 35 ft2/pr (3 m2/pr) are required for pedestrians to attain normal walking speeds and to avoid conflicts with others. Interestingly, the maximum pedestrian traffic-flow volume is not obtained when people can walk the fastest, but when average area occupancies are at about 5 ft2/pr, and pedestrians are limited to an uncomfortable shuffling gait less than half normal walking speed. At individual space occupancies below 2 ft2/pr, approaching the plan area of the human body, virtually all movement is stopped. When there is a large crowd in a confined space, this density can result in shock waves and potentially fatal crowd pressures.
Fruin has developed a set of useful metrics and measures that allow walking speeds and infrastructure capacity to be calculated. These are depicted in the following two diagrams and tables:


What this tabulates is the different levels of service (LOS) experienced by pedestrians, as the space between pedestrians gets more cramped, in order to get more people through the same corridor, per hour. LOS A is the most comfortable, and LOS E and LOS F are where things get tricky. In between is a LOS that gets the most people through, and retains an optimal level of comfort - but you can see that cross flows become difficult (eg people using the north-south laneway) etc.
This graph shows this relationship quantitatively. At the top are the different LOS, with "A" to the right. Which is described on the graph as "commuter bi-directional" - people easily going in both directions. Then you have LOS "B" described as "shoppers multi-directional", and then LOS "C" described as "commuter uni-directional" - people going in similar directions, and taking up between 15 and 25 square feet/person. And so on.

Crucially, Fruin, in his piece poses this question:
Determine the recommended width of a 200-ft (61-m)-long corridor connection to a transit station with a forecasted two-way, peak-hour pedestrian traffic of 10,000 pr/h under the following conditions:
Alternative 1: commuters only, no doors entering the corridor, no columns, no other services.
Alternative 2: same volume, but with retail development along the corridor edges in a shopping mall configuration.
And provides the answers (which you can look at in the link) as follows:

Solution for Alternative 1: This requires the selection of a design LOS and appropriate design peak. Unless there are significant restraints, the approximate mid-range of LOS C, or 12.5 pr/ft min, is an appropriate standard. The 15-min peak, typically about 40% of peak-hour volume, is an appropriate design period. (Ans=6.8 metres) 
Solution for Alternative 2: The shopping mall alternative can be analyzed as (a) a corridor with additions at the edges to allow window-shopping and door-opening "lanes" or (b) on a time-space basis, assuming some percentage of the commuters will spend additional time in the corridor. For illustrative purposes, it will be assumed that 100% of the commuters walk through the corridor, but that 30% of this total stop to window-shop for an additional 1 min each. It is desired to provide a density of 20 ft2/p average within the corridor during the 15-min peak. (Ans for 'a'=8.3 metres, or 'b'=8.9 metres).
The laneway proposed by Precinct is the "shopping mall" alternative with shops lining both sides of the corridor/laneway. Thus, according to Fruin's calculations, that laneway should be somewhere between 8.3 and 8.9 metres wide, and that's for 10,000 passenger movements/hour - rather less than the "conservative" number of 16,000 mentioned by Blair Johnston for Precinct.

Until we are very clear about this, extreme caution should be exercised in the planning of this crucial link in Auckland's downtown passenger transport station infrastructure.


2 comments:

Joel Cayford said...

Thinking about the 16,000 "conservative" figure in Precinct's pedestrian estimates, this figure is what you get when you apply the Fruin formula assuming that, within an hour, 40% of people pass through in 15 minutes. A peak flow point within the hour. This peak flow = a flow rate of 16,000/hour - the figure quoted to the hearing. This suggests AT did use the Fruin formulae - but have NOT taken into account the fact pedestrians are in a retail mall environment. Perhaps they assume shops must close, and all window displays are switched off to prevent any window-shopping. Seems unlikely. Not what Precinct retail tenants would value....

Joel Cayford said...

Thus 16,000/hour is the peak flow rate, based on 10,000 passing through in a full hour. Meant to state that in my other comment...

Thursday, November 12, 2015

Precinct's Downtown Laneway: How Wide?

This posting is about the laneway proposed in Precinct's plans for redevelopment of the Auckland Downtown block bounded by Lower Queen Street, Quay Street, Lower Albert Street and Custom Street West. The picture below shows a rough plan of this laneway.....


The laneway I want to discuss here, is the one connecting Lower Albert Street through the development to Lower Queen Street. The laneway will be about 100 metres long if the whole of Queen Square (shaded in green) is sold for development, and about 60 metres long if Queen Square stays in public ownership as public space. (The other proposed laneway, running north-south, is more of a dog leg than shown here, as it has to side-step the HSBC building to get to Quay Street.)

In June this year, Precinct obtained resource consent for a development on the site that does not use Queen Square. This consent was non-notified. It includes consent for the 60 metre section of laneway. The question I want to address in this posting is how wide should this laneway be?

At the hearings that have occurred in the past couple of weeks, there have been varying widths discussed by the applicant's experts and by Auckland Council. The Road Stop notice stipulates a 5 metre width east-west laneway. Other witnesses have said 6.5 metres, and that this provides a 5 metre wide central section, allowing shoppers to look at the shops using the other 1.5 metres of width. Various submitters have raised questions about the safety of this - because the laneway will be used by passengers to walk between the bus terminal in Albert Street and the Britomart rail terminal or other buses that terminate there.

At the Plan Change 79 hearing, Blair Johnston of Warren and Mahoney Architects, (who are doing master plan and architecture work for Precinct) stated that he had worked with Auckland Transport, and that modelling showed that a 5 metre wide corridor was safe and appropriate, even up to conservative estimates of 16,000 people going through per hour.

The world authority on these matters is John J Fruin. You can see a summary of his work for New York Transit here. This is a bit dense, so I've unpacked relevant sections....
PEDESTRIAN CHARACTERISTICS
The primary characteristics needed to evaluate pedestrian facilities are walking speeds, walking distances, demand patterns, and traffic-flow relationships. The ability of pedestrians to select their own individual walking pace and speed is a qualitative measure of convenience. Walking distances define the effective service area of transit stations, with shorter distances improving passenger perceptions of service and convenience. The patterns of passenger demand affect the methods used to analyze pedestrian facilities and the applications of service standards.
WALKING SPEEDS
Pedestrian speeds, in addition to being directly related to traffic density, have been found to vary for a wide range of conditions, including individual age, sex, personal disabilities, environmental factors, and trip purpose. Normal walking speeds unimpeded by pedestrian crowding have been found to vary between 150 and 350 ft/min (0.76 and 1.76 m/s), with the average at about 270 ft/min. As a point of comparison, running the 4-min mile is equivalent to a speed of 1320 ft/min or almost 5 times normal walking speed. Walking speeds decline with age, particularly after age 65, but healthy older adults are capable of increasing their walking speed by 40% for short distances. Dense pedestrian traffic has the effect of reducing walking speed for all persons. The smaller personal space limits pacing distances and the ability to pass slower moving pedestrians or to cross the traffic stream.
Photographic studies of pedestrian traffic flow on walkways have shown that individual area occupancies of at least 35 ft2/pr (3 m2/pr) are required for pedestrians to attain normal walking speeds and to avoid conflicts with others. Interestingly, the maximum pedestrian traffic-flow volume is not obtained when people can walk the fastest, but when average area occupancies are at about 5 ft2/pr, and pedestrians are limited to an uncomfortable shuffling gait less than half normal walking speed. At individual space occupancies below 2 ft2/pr, approaching the plan area of the human body, virtually all movement is stopped. When there is a large crowd in a confined space, this density can result in shock waves and potentially fatal crowd pressures.
Fruin has developed a set of useful metrics and measures that allow walking speeds and infrastructure capacity to be calculated. These are depicted in the following two diagrams and tables:


What this tabulates is the different levels of service (LOS) experienced by pedestrians, as the space between pedestrians gets more cramped, in order to get more people through the same corridor, per hour. LOS A is the most comfortable, and LOS E and LOS F are where things get tricky. In between is a LOS that gets the most people through, and retains an optimal level of comfort - but you can see that cross flows become difficult (eg people using the north-south laneway) etc.
This graph shows this relationship quantitatively. At the top are the different LOS, with "A" to the right. Which is described on the graph as "commuter bi-directional" - people easily going in both directions. Then you have LOS "B" described as "shoppers multi-directional", and then LOS "C" described as "commuter uni-directional" - people going in similar directions, and taking up between 15 and 25 square feet/person. And so on.

Crucially, Fruin, in his piece poses this question:
Determine the recommended width of a 200-ft (61-m)-long corridor connection to a transit station with a forecasted two-way, peak-hour pedestrian traffic of 10,000 pr/h under the following conditions:
Alternative 1: commuters only, no doors entering the corridor, no columns, no other services.
Alternative 2: same volume, but with retail development along the corridor edges in a shopping mall configuration.
And provides the answers (which you can look at in the link) as follows:

Solution for Alternative 1: This requires the selection of a design LOS and appropriate design peak. Unless there are significant restraints, the approximate mid-range of LOS C, or 12.5 pr/ft min, is an appropriate standard. The 15-min peak, typically about 40% of peak-hour volume, is an appropriate design period. (Ans=6.8 metres) 
Solution for Alternative 2: The shopping mall alternative can be analyzed as (a) a corridor with additions at the edges to allow window-shopping and door-opening "lanes" or (b) on a time-space basis, assuming some percentage of the commuters will spend additional time in the corridor. For illustrative purposes, it will be assumed that 100% of the commuters walk through the corridor, but that 30% of this total stop to window-shop for an additional 1 min each. It is desired to provide a density of 20 ft2/p average within the corridor during the 15-min peak. (Ans for 'a'=8.3 metres, or 'b'=8.9 metres).
The laneway proposed by Precinct is the "shopping mall" alternative with shops lining both sides of the corridor/laneway. Thus, according to Fruin's calculations, that laneway should be somewhere between 8.3 and 8.9 metres wide, and that's for 10,000 passenger movements/hour - rather less than the "conservative" number of 16,000 mentioned by Blair Johnston for Precinct.

Until we are very clear about this, extreme caution should be exercised in the planning of this crucial link in Auckland's downtown passenger transport station infrastructure.


2 comments:

Joel Cayford said...

Thinking about the 16,000 "conservative" figure in Precinct's pedestrian estimates, this figure is what you get when you apply the Fruin formula assuming that, within an hour, 40% of people pass through in 15 minutes. A peak flow point within the hour. This peak flow = a flow rate of 16,000/hour - the figure quoted to the hearing. This suggests AT did use the Fruin formulae - but have NOT taken into account the fact pedestrians are in a retail mall environment. Perhaps they assume shops must close, and all window displays are switched off to prevent any window-shopping. Seems unlikely. Not what Precinct retail tenants would value....

Joel Cayford said...

Thus 16,000/hour is the peak flow rate, based on 10,000 passing through in a full hour. Meant to state that in my other comment...