Ferry Logo

Ferry Logo

Wednesday, 30 July 2014

Ferry Connections: Part 2 - Interchanges

A ferry operation can be viewed as a system made up of infrastructure elements (fleet and terminals), its customers, staff and a timetable. It is an open system, heavily influenced by the maritime environment, Government policy and regulation, the operation of other modes of transport and many other factors.  

Few aspects of this system are as important as ferry terminals, especially interchanges where passengers transfer from one line to another, or to or from another mode. In a harsh maritime environment, there can be a preoccupation with engineering issues and perhaps less attention on efficient ingress and egress of passengers or how the terminal fits with the overall network plan. 

Eagle Street Ferry Terminal Brisbane
The reliability of the network and the convenience of transfers between boats is largely a function of interchange design and how well it matches the vessel fleet.

Where an interchange is well designed, passengers can disembark and board quickly. If this happens, and the design of the fleet is compatible with the terminal design, there will only be a small variation in boarding time between a light loading and a heavy loading. This in turn helps ensure scheduled connections are met.  

To achieve this happy situation, there are two main requirements of the interchange:
  • capacity to deploy multiple, wide gangways. A single, narrow gangway only allows one passenger at a time to board or disembark. It takes one minute for 50 passengers to board via a single narrow gangway, and longer if there are passengers with strollers or special needs. If 100 are disembarking and 100 are boarding, then the boarding process takes four minutes or more. This is an unacceptably long dwell, not including the time needed to plank/ remove the gangway and tie lines. The solution is to provide for two wide gangways at all high demand wharves, which enable four to six passengers to board simultaneously.
  • keeping disembarking and boarding passengers separate at high demand terminals. If the wharf pontoon is too small, the crowd of passengers waiting to board interferes with the egress of disembarking passengers. Passengers transferring between vessels also need to be accommodated efficiently.
Each interchange must be an appropriate adaptation to the particular location and circumstances of its location. A "one size fits all" approach to designing wharf upgrades is doomed to failure.   

In last week's post, a network plan for Sydney Ferries was outlined with six locations where dual berthing wharf facilities are needed. It is not enough to simply say "this wharf needs to be dual berthing". There must be a visualisation of how these interchanges will work in practice, taking into account the proposed timetable and the implications for vessel and passenger movements. The goal should be to make sure dwell time is minimised and passengers can move safely and quickly between vessels.

At the outset, it should be noted that a classical 'four way' interchange that you might expect in a rail network, where two lines intersect at a four platform station and trains in both directions on both lines arrive at the same time, is not something you would want to replicate in the Sydney Ferry network, outside of Circular Quay. The cost of infrastructure to support this would be significant and not necessary unless transfers in all directions are imperative.

Some less grandiose models are suggested here.

1. Corner Connection

Network requirement:  two lines intersect at a terminal; transfers are required in some, but not all directions. Cremorne Point is an example of a corner connection in the proposed Sydney Ferry network described in the previous post. In this case, transfers from Barangaroo and Rose Bay in the direction of Mosman (and return) need to be possible, but it is not necessary for there to be transfers in the direction of Circular Quay. This means the Mosman to Circular Quay vessel can depart ahead of the other three.
Terminal design solution: required transfers can be supported, without conflicts, with a double length dual berth pontoon. The following diagrams show the order of vessels berthing and departing. Note that the connecting 225 bus arrives first allowing passengers to transfer to Circular Quay, Rose Bay and Barangaroo and departs last with passengers disembarking the ferries. In this case, a single ramp is shown, but separate ramps for boarding and disembarking passengers could be provided if needed.


2. T Connection

Network requirements: A T connection occurs where a wharf is the terminus of one line and an intermediate stop on another. An example is Rose Bay, assuming it is the terminus of a Ring Line connecting the Eastern Suburbs with the North Shore and Barangaroo, as well as an intermediate stop in both directions for the Watsons Bay line. In a situation like this, it is preferable to minimise the dwell for vessels on the Watsons Bay line to avoid inconveniencing passengers who have not completed their journey. There is scope, however, for a layover for the Ring Line as Rose Bay is the terminus. 
Terminal design solution: In this T connection, the inbound and outbound Watsons Bay vessels berth simultaneously at Rose Bay. As the vessel from Barangaroo is also in a layover at the same time, three berths are required. The simplest way to accommodate this is to use a double sided pontoon which is perpendicular to the shoreline and long enough to allow the Barangaroo vessel to berth "in the corner". 

These examples of terminal design show the importance of long term planning in the design of the network and timetable. Unless this is done, the supporting infrastructure may not meet future requirements.The Swiss Federal Railways develops its timetable 20 years in advance. This creates a sufficient lead time for infrastructure improvements to be put in place and minimises reworking and waste.  

Tuesday, 22 July 2014

Ferry Connections: Part 1 - Network Design

Like many public transport systems, the Sydney Ferry network is CBD centric. All lines converge on the city, terminating at either Circular Quay or Darling Harbour. The limitation of CBD centricity was articulated many years ago by the American transport planner Greg Thompson:
"At that time (1969), the prevailing notion that all transit routes must converge on the downtown seemed wrong. What appeared to be a mandatory requirement for a well-used transit was a  network of routes allowing people to travel conveniently to many destinations, of which the downtown was only one".      
The Sydney Ferry network currently has 36 terminals, spread over eight routes in a conventional radial pattern. This means there could be up to 630 unique origin-destination pair connections, leaving aside intermodal journeys.

That's a lot of trips traversing Sydney Harbour where other realistic public transport options are not always available. It could be a trip from Chiswick to Luna Park, or Rose Bay to Taronga Zoo or Balmain to Double Bay. But these are only potential connections, because if the timetable does not provide for timed transfers between lines, or the connection is only available for part of the day, or on week-days but not week-ends, then legibility is lost and it becomes all too hard for the customer. They will find another way to reach their destination, other than by public transport.         

Under the current timetable, just 96 origin-destination pairs can be connected conveniently all day, seven days a week[1]. That’s 16% of the total possible OD pairs.
Increasing the number of convenient OD pair connections is difficult to do in a low frequency network. It has been done best by the Swiss Federal Railways which, since the early 1980's, adopted an integrated regular interval timetable or Taktfahrplan. These principles can also be applied to an urban transit ferry network.

In a network based on Taktfahrplan principles, there are not necessarily more lines, just better connections. If the service interval of the network is 30 minutes:
  • all services are scheduled to arrive at a hub a few minutes before the hour and half hour. This is sometimes referred to as a pulse timetable, allowing passengers to transfer to a different line with only a short wait before their next departure a few minutes after the hour and half hour.
  • between hubs, inbound and outbound ferries cross at half the service interval (15' in this case), which presents the opportunity for a pulse interchange with other modes. If it is a ferry terminal which accommodates dual berthing, then one bus can conveniently connect with the inbound and outbound ferry in both directions.      

There are some rules which need to be followed to achieve consistent connectivity. The cycle time of each return trip, including layovers, must be a whole integer multiple of the service interval; a regular interval must be retained all day (supplementary departures can be scheduled in the peaks if needed); and stopping patterns of inbound and outbound trips must be consistent and symmetrical. 

                            A Stylised Ferry Network with 30' Minute Service Intervals 

Providing these rules are followed, network connections can be dramatically increased, with minimal additional operating costs. 

The good news for Sydney Ferries is that Taktfahrplan rules can be applied with relatively minor changes to the existing network. This is because most lines already have cycle times which are a multiple of the service interval. The remainder can be changed with little difficulty.

So what would a Sydney Ferry network look like if Taktfahrplan principles were applied? There is no one single solution, but it could look something like this:

This new design takes advantage of the new larger ferry hub to be constructed at Barangaroo replacing the existing Darling Harbour King Street terminal. This will become a more attractive destination for residents west of the city with the completion of the Wynyard Walk development. 

With the exception of the silver line between Sydney Olympic Park and Barangaroo (commuter peaks only), all stopping patterns would remain consistent all day:
  • Double Bay is separated from the Watsons Bay line. The Double Bay line has stops all day at Elizabeth Bay (new stop) and Darling Point.
  • Darling Harbour services through line to Taronga Zoo.
  • Manly, Mosman, Neutral Bay, Double Bay, Darling Harbour and Taronga Zoo lines all hub at Circular Quay, with arrivals just before the hour and half hour and departures just after the hour and half hour. This creates convenient transfers between all lines, all day. 
  • The Cockatoo Island line terminates at Barangaroo, with convenient transfers at Balmain East (six minute wait in both directions) for passengers travelling to McMahons Point, Milsons Point or Circular Quay.
  • Parramatta River services through line to Watsons Bay via Circular Quay. All departures stop at McMahons Point and Milsons Point and provide convenient transfers at McMahons Point for passengers from River stops who need to travel to Barangaroo. 
  • As a Stage 2 development, a new ring line is introduced which connects Rose Bay with Taronga Zoo, Cremorne Point, Milsons Point, McMahons Point, Balmain East, Barnagaroo and Jacksons Landing (new stop). This provides convenient connections to the Watsons Bay line at Rose Bay, to the Mosman line at Cremorne Point and Parramatta River line at McMahons Point. 
  • Another element of Stage 2 is to convert the Manly Ferry to a fast ferry service at a 15 minute interval all day.        

Not including Stage 2, these changes add a modest 6% to ferry service hours (up from 80,000 to 84,700), but massively increase the number of convenient OD pair connections from 96 to 332. Including the Manly fast ferry and ring lines, the convenient connections increase to 443, but with an additional 15,250 service hours.

The legibility and connectivity of a Taktfahrplan style timetable would bring enormous benefits for ferry customers. Imagine, for example, a passenger from Balmain Thames Street who may, at different times, need to travel to Barangaroo, Milsons Point, Circular Quay or even Double Bay or Manly. In the timetable proposed here, the passenger can either travel directly to Barangaroo or transfer conveniently at Balmain East to reach Milsons Point and Circular Quay. They could also transfer conveniently at Circular Quay to Double Bay, Manly, Mosman and other locations, or stay on board for travel to Taronga Zoo.

Because it is a regular interval timetable, the connections work for every departure, which significantly adds to the legibility of the network. All that the passenger from Balmain Thames Street needs to remember is that the ferry departs in the direction of Barangaroo at 3 and 33 minutes past the hour. All the connections will be in place whichever ferry is taken.
These changes also have important operational advantages:
  • in a regular interval timetable, the pattern in one hour is repeated over all hours so vessels always cross at the same location. In this proposal, berthing conflicts are eliminated. Departures are separated by at least four minutes at all wharves except those designated for duaI berthing. This provides time for a vessel to berth and depart with a two minute buffer before the next vessel arrives.
  • Safety is improved because vessel movements are more predictable. At Circular Quay, boats will not arrive at the same time as departing vessels, reducing the risk of collisions.
  • Extra departures can be easily added in peak periods, including special event services, because of the modular structure of the timetable.
For this model to work, action must be taken to reduce variation in dwell times and improve the design of high demand wharves, including provision for dual berthing at key interchanges. These two issues are the subject of next week's post.      


[1] A convenient connection is defined as a journey between an OD pair where (a) there is no more than one transfer; (b) the scheduled wait time at a transfer point is between three and ten minutes; (c) the service interval is no more than 60’; and (d) all of the above criteria apply Monday to Saturday between 7 am and 10 pm and between 8 am and 8 pm Sundays. 

Thursday, 17 July 2014

The Secret of Successful Public Transport

Driving a car in Sydney has one big advantage over public transport. We can jump into our car and go wherever we want, at a time that suits us. For most Sydney residents, public transport can’t match this. Services don’t run often enough, or for only part of the day, or our destination lies on a different line which we can’t conveniently transfer to.   

But what if the planning of public transport was truly integrated? Imagine if the near CBD network, where population densities are high, was a high frequency grid, with services running from early in the morning until late at night, seven days a week. And in areas of lower population density, what if there were timed connections at all interchanges, so that even if departures are only twice an hour, transfers to other lines or modes are very convenient. Anywhere to anywhere, at any time, suddenly becomes possible.
This is not a fantasy, because it is exactly what happens in Zurich and many other cities in Europe. Surprisingly, it happens at low cost to the taxpayer. The Zurich public transport system recovers 63% of its operating costs and in Munich there is full cost recovery. According to the NSW Auditor-General, cost recovery here is less than 30%.

The quality of the Zurich public transport network means that more people use it and fewer trips are made by car. The average resident in the Canton of Zurich makes 400 trips by public transport each year. In Sydney, we make 120. Sydneysiders travel 4.9 km per day by public transport and 28.5 km by private motor vehicle. Zurich residents on average travel 11.7 km by public transport and 20.7 km by private motor vehicle. 

Visitors to Zurich or Munich may be surprised at the grim utilitarianism of transport infrastructure in these cities. Interchanges are functional and generally without architectural flourish; there are no smartcards or ticket gate barriers; and a lot of the rolling stock is old. This is because transport planners consider the “customer offer” – a  network design and timetable which allow more than 80% of residents to go from anywhere, to anywhere – to be the most important ingredient of a successful transit system. It is also because the science of timetabling is better understood in Europe and perhaps better in Switzerland than anywhere else.

Tram stop in Zurich: seat, shelter, timetable and ticket machine. What more do you need?

There is one more thing. Around 80% of public transport trips in Zurich are made by passengers using a periodical ticket: a yearly, monthly or weekly ticket which provides unlimited travel on all modes within nominated zones. Pricing does not differentiate between modes or punish transfers.

The periodical tickets are attractive as they provide big discounts compared to single journey tickets. This makes public transport pricing more akin to car ownership. While there are marginal costs in driving a car, we tend to think a short drive is "free". Unlimited travel with a periodical transit ticket creates a similar perception about public transport, in contrast to “pay as you use” systems, like the Opal Card.

So the secret to successful public transport is less about money and technology and more about philosophy and technique. The good news is that these techniques are explained in a wonderfully clear and comprehensive manual on designing good public transport, the “Hi Trans Guide”. We would all benefit if its sage advice was more widely followed.