2.

3 Public Transport Study

2.3.1 General

2.3.1.1 The role of a public transport study is to devise a system of models at appropriate level for evaluating
public transport network in a comprehensive manner which would facilitate network development and
rationalisation as well as assessment on financial viability.

2.3.1.2 Public transport study normally covers a region or sub-region. It involves an analytical process which
builds on a number of quantitative techniques.

2.3.1.3 Analytical transport study tries to apply general mathematical relationships or models to simulate the
travel behaviour of individuals whose particular patterns obey no fixed mathematical rule. The
applications of quantitative techniques are often restricted by certain assumptions, availability of
databases and limitation of computer software. They are merely planning tools for planners to test
different sets of assumptions or scenarios with a view to building up an understanding of the key
factors and the effects of their variations on the forecasts of travel demand and pattern. The techniques
can form a basis for assessment of alternative actions and evaluation of transport proposals.

2.3.2 Planning Horizon

2.3.2.1 Public transport study covers short, medium and long term planning. Normally, short term planning
covers planning period of one to two years. For medium term planning, its planning period will cover
up to five years. As for long term planning, the planning period is above five years.

2.3.3 Integrated Planning Process and Planning Hierarchy

2.3.3.1 The position of a public transport study in the integrated planning process is illustrated schematically in
Diagram 2.3.3.1.The planning hierarchy for the territory includes the Hong Kong 2030 Vision and
Development Strategy which reviews and updates strategic land use/transport plans, CTS-3 for
identifying the strategic transport projects and policy requirements and the Railway Development
Study on rail strategy. These are then followed by appropriate regional transport studies.

2.3.3.2 The main transport planning tool within Government is the CTS-3 model developed in 1999. The CTS-
3 model is developed by combining and building upon two existing territory-wide models - the
Enhanced CTS-2 Model produced by the Model Enhancement Study, and the Freight Transport Model
(FTM) produced by the Freight Transport Study (FTS). It is designed to deal with strategic territory-
wide issues, concentrating on large scale corridor and highway movements. It is the basis of most
transportation planning in the territory.

2.3.3.3 CTS-3 model is used to forecast demand by each main type of transport mode and the traffic conditions
in future years under different assumptions of land-use, economic growth, transport network and
transport policy. New transport projects or changes in the transport policy can be analysed by
incorporating them in the CTS-3 model and then comparing the projected traffic flows from the model
runs with and without these projects or policies.

2.3.3.4 The models and procedures needs to be refined, where appropriate, to meet the objectives of individual
study.

2.3.3.5 Apart from territory-wide study level, public transport study will also be carried out at region or district
level, if necessary. The study area of a regional based public transport study normally covers a region,
or one to two new towns located proximately. The review on public transport services between Tai Wai
(Sha Tin) and Ma On Shan in relation to the development of Ma On Shan Rail Link is one of the
examples. The focus of a district based public transport study will be on local area only. The re-
organisation of bus services in Tsing Yi upon opening of Tung Chung Line in 1997 is an example of
 this type of study.

Diagram 2.3.3.1: Integrated Planning Process and Planning Hierarchy

2.3.4 Administrative procedures

2.3.4.1 The administrative procedures for conducting a large-scale transport study are broadly as follows :

(i) Seeking funds for appointment of consultants for conducting the study;

(ii) Preparation of study brief;

(iii) Selection of consultants;

(iv) Setting up of Steering Group;

(v) Setting up of Working Group;

(vi) Preparation of Study Reports and Deliverables; and

(vii) Public consultation.

2.3.5 Study Objectives

2.3.5.1 In conducting a public transport study, its aim is to review the existing public transport services and
assess the future development of public transport services in the Study Area.

2.3.5.2 The major outputs are recommendations on a viable overall transport plan and route development
programmes for various modes taking into account the transport policy, demand, major developments
and infrastructural changes in the Study Area.

2.3.5.3 The typical study objectives are :

(i) identify problems and issues;

(ii) review the provision of public transport services in the Study Area;

(iii) review the patronage of public transport services in the Study Area;

(iv) forecast future patronage demand;

(v) formulate alternative public transport development strategies;

 (vi) establish evaluation framework;

(vii) assess the effect of implementing alternative public transport development strategies;

(viii) recommend a viable development programme;

(ix) reassess transport infrastructure requirements to be compatible with the recommended

network;

(x) examine the need for any legislative change required for the recommended strategy;

(xi) propose the consultation strategy and a implementation time frame; and

(xii) carry out broad assessment of the impact on the total public transport system throughout
the study period.

2.3.6 Public Transport Demand

2.3.6.1 The planning for new rail infrastructure is conducted as part of the territorial transport planning
process. The rail operators are required to forecast their patronage and plan for improvement of their
services.

2.3.6.2 The planning for bus services are carried out by the franchised operators in collaboration with TD.
They are required to prepare an annual “Five-Year Forward Planning Programme” including a route
development programme with a view improving and optimising their services to meet changes in
demand and public expectation.

2.3.6.3 Green minibuses (GMB) are to supplement the mass carriers serving areas where physical accessibility
constraints or demand does not justify the provision of high capacity modes. TD is responsible for
planning of green minibus services, including introduction of new routes and changes to existing
services. It is Government’s policy to limit the total number of minibuses and to encourage the
conversion of red minibuses (RMB) to GMB.

2.3.6.4 Ferry is an essential mode for outlying islands and a supplementary mode in the urban area. Most of
the ferry services in Hong Kong are now operated under a licensing system. The licensed operators are
not required to prepare forward planning programmes. Normally, TD is responsible for planning of
ferry services. The factors which will be taken into account in planning of new routes include:
availability of piers, financial viability of the route, provision of alternative services and any impact on
existing routes.

2.3.6.5 One of the main tasks for a public transport study is to forecast public transport demand. It may be
predicted by the land-use transportation model, time-series trend analysis and the regression analysis.

2.3.6.6 Land -use Transportation Model

Public transport demand may be projected as a by-product of forecasts for all modes of transport
produced within a land-use transportation study. Such forecast usually relates to a specific design date
of about 5 - 10 years which are based on forecasts on population, employment, education, car
ownership level, income level, economic growth, value of time, public transport fares and comfort.
Alternatively, a model specific approach can be adopted. In the long run, it is related to variables such
as population size, income level, economic variable, public transport fares and car ownership level. In
the short run, aspects of service quality such as frequency, speed and fare which can be influenced
markedly by the operator are more emphasised. A common feature of such model is the lack of
importance of variables reflecting characteristics of competing modes. In addition, it is fairly difficult
to include variables such as comfort and convenience.

2.3.6.7 Time-series Trend Analysis

Of specific forecasting methods, extrapolation is the simplest method. However, there is limitation to
 such method as the same trend and relationship will continue to be assumed in those external variables
affecting demand. It is difficult to select the past period from which a trend is to be drawn for
projection. Nevertheless, extrapolation may be useful in the short run, particularly if the projection is
broken down into meaningful categories.

2.3.6.8 Regression Analysis

There are demand forecasting models based on multiple regression in which elasticity of demand of
major quality variables such as frequency and fares are incorporated. With the regression model, the
relationship between the variables can be evaluated. However, the development of a regression model
requires an extensive data collection in order to meet the needs of individual organization which often
involves considerable time and expense. It also requires expertise in the selection of variables to ensure
the relationship that has a logical and meaningful interpretation. For these reasons, regression models
are appropriate for making medium and long-term forecasts.

2.3.7 Peak hour demand

2.3.7.1 Like all major cities, Hong Kong has to face the problem of providing sufficient public transport
capacity to meet peak period demand. The magnitude of problem depends on the level of demand and
the capacity constraints. The peak hour demand of a new housing estate of a developed/developing area
can be forecasted either by projecting the trips generated during peak hour or by applying the peak hour
factor to calculate the peak hour demand of similar estate in the same area. For new area, it may have
to apply peak hour factor of other area with similar demand. A public transport study is normally aimed
at making recommendations for resolving the problems during peak hours.

2.3.8 General approach

The general approach to conduct a public transport study comprises nine stages :

(i) Inception period;

(ii) Data Collection and Analysis;

(iii) Input Assumptions;

(iv) Model Development;

(v) Strategy Development and Testing;

(vi) Evaluation of Strategies;

(vii) Recommendations on Public Transport Strategy;

(viii) Consultation with relevant parties on the recommendations; and

(ix) Implementation Programme.

2.3.8.1 Inception
During this stage, the study objectives, the resources requirement, the study approach and the basic
methodology for the study are set out. Key issues and constraints related to the Study will be reviewed.
In addition, all available relevant data are collected for model calibration, problem identification,
option development and evaluation purposes. In addition, a study programme is drawn up with the
critical paths identified.

2.3.8.2 Data Collection/Analysis

(i) It is necessary to acquire up-to-date information for conducting qualitative and quantitative
analysis of existing services. There is also a need to identify the planned and committed
developments. The information may be readily available or has to be collected through
 surveys.

(ii) The existing available information may include:

(a) Existing public transport services;

(b) Travel pattern and travel time of passengers;

(c) Existing and proposed transport infrastructures;

(d) Land use and development proposals;

(e) Patronage data;

(f) Operation and cost data for alternative modes of transport;

(g) Traffic data;

(h) Route development programmes of public transport operators;

(i) Social and economic data such as population, employment, education and
inflation etc.;

(j) Recommendations of sub-regional traffic studies and traffic impact assessment

of developments;

(k) Existing, planned and committed transport facilities; and

(l) Strategy and recommendation of public transport schemes such as Bus Only
Lane.

(iii) Additional information to be collected through surveys is often required, e.g. :

(a) Origin and destination survey;

(b) Stop/terminus boarding /alighting survey;

(c) Screenline survey;

(d) Passenger interview survey (access/egress mode and reason for mode
selection); and

(e) Stated preference survey (to project passenger behaviour by a survey conducted
on hypothetical situation with a set of alternative scenarios and choice
questions to solicit response for target respondents).

(iv) The information collected or collated are used to develop a database for the study. It is
common that reference can be made to the database of relevant studies conducted in the
past. If the database of a previous study is to be adopted, the information is used to review
the previous model performance and for refining the model formulation. The information is
also used to identify problems, develop feasible options and evaluate options.

2.3.8.3 Input Assumptions

(i) It is necessary to adopt various input assumptions to conduct a study. The assumptions may
be derived from empirical data collected. The accuracy and appropriateness of the
assumptions will affect the quality of the study findings and recommendations directly. To
assess the impact of change in assumptions on the recommendations, sensitivity tests on
different sets of assumption would need to be conducted.
 (ii) Policy input
The study’s policy assumptions normally require the endorsement by the Steering Group.
For public transport studies, the assumptions are normally related to the guidelines on
inter-modal co-ordination, fare policy, role of public light bus and viability of modes etc.
However, there may be other consideration such as the land use and the limitation on
development of a particular area.

(iii) Land-use and Socio-economic Data

The assumptions on land use distribution and development are based on the agreed land-
use and socio-economic data in terms of the land use plan, projected population,
employment and education enrollment figures, age distribution, income level and
population distribution published by the Working Group on Population Distribution,
Census and Statistics Department and Planning Department.

(iv) Fare Index

The fare levels of public transport services play an influential role in forecasting trip
demand, route and mode choice. Fares are vital to derive the revenue estimates for
evaluation of different transport options.

(v) Major Highway Projects/Infrastructure Programme

The major planned and committed highway projects and infrastructure programmes as
given in the Development Programmes of Highways Department and Territory
Development Department, as well as other relevant information are used for input into the
model.

(vi) Public Transport Network

The inputs on public transport network are based on the latest information on TD’s official
schedule of services, the operators’ forward planning programmes and the
recommendations of the recent transport studies. These inputs should be reviewed against
the latest transport policy, land use and socio-economic assumptions.

2.3.8.4 Model Development

(i) This entails an updating/refinement of model for the study plus additional procedures to
evaluate options associated with new development.

(ii) The CTS model is designed to deal with strategic territory-wide issues, concentrating on
the large scale corridor and highway movements. It is the basis of most transportation
planning in the Territory. For public transport studies, a more detailed zoning system which
is different from CTS model may be required.

(iii) Public Transport Model

In general, a public transport model consists of the following :

(a) A refined zoning system, including regional and sub-regional basis;

(b) Trip matrices consisting of zone to zone public transport movements;

(c) A detailed road and public transport network and assignment model to
determine costs and hence paths between zone pairs;

(d) A public transport modal split model to determine the allocation of trips
between modes; and

(e) A public transport assignment model to estimate the demand for individual
route.
 (iv) Network and Assignment Model
A network and assignment model is used to determine costs and hence paths between zone
pairs. Its application is as follows:

(a) The public transport network for CTS is for strategic purpose which needs to be
refined for a specific public transport study. In some circumstances, the
strategic CTS network needs to be extended and refined to cover new zones
like reclamation areas and to enable accurate modelling of the sub-modal split
between various modes.

(b) The base network needs to be converted to suit the format of the computer
software for running the model.

(c) The public transport services in the network are updated to incorporate the
committed network changes for the future years.

(d) Various methods can be used to obtain the modal split. One of the methods to
develop the modal split model is to base on the results of stated preference (SP)
surveys and recalibrated the model on the basis of the revised costs from the
updated networks and the overall observed modal split. The following variables
are commonly included in the model:

 model preference (a constant)

 waiting time
 access/egress time
 time in mode
 fare

(e) For each zone pair, utilities are calculated from network costs for all modes and
the proportion by each mode is calculated.

(f) The inter-zonal times and costs are obtained from model assignment paths
restricted to each mode. The preferred paths may be determined by penalising
trips requiring interchanges.

(g) The resultant cost matrices are used to recalibrate the modal split model.

(v) A more detailed explanation of the public transport modelling process is given in section
2.3.9.

2.3.8.5 Strategy Development and Testing

(i) Prior to establishing strategies for implementation, it is necessary to consider various

options. This stage involves:

(a) Strategy development

(b) Identification of options

(c) Development of modal strategy

(d) Evaluation of strategies

(e) Recommended strategies

(ii) Strategy Development

A number of strategies and options for different modes will be assessed through the
 following process:

(a) identification of options for each mode;

(b) development of strategies from the modal options; and

(c) combination of modal strategies into multi-modal strategies for testing and
evaluation.

(iii) Identification of Modal Options

The reference networks and initial options are tested and the output is analysed so that
problems with the reference cases can be assessed. The ability of the initial options to
overcome these problems and fulfil the requirements of the evaluation criteria are
identified. Options are then developed, refined and evaluated.

(iv) The options are directed towards the optimisation and rationalisation of services in order to
the make best use of the planned infrastructures, to suit the changing land use pattern and
to address particular policy concerns. The options are identified in terms of new routes,
route withdrawal, frequency and fare changes. During the process, discussion will be made
with the public transport operators with reference to their route development programmes.
The evaluation procedures may be designed to evaluate the identified options as well as to
allow additional options to emerge for subsequent evaluation.

(v) Development of Modal Strategies

The modal strategies are built up from the modal options. The basic approach is to develop
three broad strategies for each mode :

(a) a base case;

(b) a low level of service; and

(c) a high level of service.

(vi) The level of service is defined in terms of network coverage, fare level and frequency. By
combining these strategies of each mode into multi-modal strategies, it will generate a
number of scenarios for testing.

(vii) The basic testing framework is modified to reflect the key policy concerns which are
identified during the course of study. An example of the modal strategy definition, in
conceptual terms, is shown below :
Modal Strategy "High" Level of Service
Mode Base (1)
"Low" Level of Service (2) (3)
Keep existing routes in the
Withdrawal of specified
study area, new routes,
Reference routes in the study area, new
frequency adjustments on
network routes, frequency
Ferry selected routes, low fare
suggested by adjustments on selected
assumption and addition of
TD. routes and high fare
land interchange service as
assumption.
found appropriate.
Route
development
programmes
Partial relaxation of policy Complete relaxation of
Bus and other
on competition. policy on competition.
proposals
conforming to
policy.
(viii) Testing and Evaluation of strategies
Based on the modal options and strategies, a program of test can be devised to examine the
 impact of different combinations of modal strategies. The testing can be divided into stages
and an example illustrating the combination of strategies to be tested by stages is shown
below :

Stage Strategy Test Description

The bus/ferry/MTR services included in the tests are in
1 T1 (Base) accordance with the agreed items in the route development
programmes.
2 T2 "High" level of service for bus.
T3 "High" level of service for rail.
T4 "High" level of service for ferry.
3 T5 "High" level of service for bus & rail.
T6 "High" level of service for bus & ferry.
T7 "High" level of service for rail & ferry.
4 T8 "Low" level of service for bus & ferry.
T9 "Low" level of service for bus, rail & ferry.
T10 "Low" level of service for bus & rail.
T11 "Low" level of service for rail & ferry.

High level of service for rail but low level of service for bus
5 T12
and ferry.
High level of service for bus but low level of service for rail
T13
and ferry.
Refined strategy network based on results of T1-T13 tests
6 T14
described above.
(ix) At each stage, the basic elements of each modal strategy are reviewed so that refinements
to the inputs can be made. The tests will demonstrate the implications of all potential
combinations of modal strategies and the impact of different emphasis in provision of
public transport services.

(x) Sensitivity tests are undertaken on the recommended options to reflect the possible changes
in planning assumptions so as to determine the robustness and transport infrastructure
requirements of the route development programmes.

2.3.8.6 Evaluation of Strategies

(i) The evaluation of the multi-modal strategies will be undertaken using a framework
approach.

(ii) The framework involves the evaluation of strategies for three separate groups, i.e., public
transport users, public transport operators and the community. The evaluation for each
group will be based on a number of factors, including the operational, financial, economic,
social effects and public acceptability.

(iii) In certain cases, it will be possible to place a quantifiable value on the evaluation criteria,
for example, average fare and revenue etc. As for other factors, it may need to have
qualitative judgements and some ranking would have to be given to the strategy.

(iv) The main effect of alternative networks on public transport users is the change in levels of
service arising from different options. These include :

(a) average fare;

(b) average travel time(i.e. in-vehicle time);

(c) average waiting time; and

 (d) average number of interchange involved.

(v) The main impact on the operators is the financial viability in respect of their overall
network or on an individual route basis. The assessment is as follows:

(a) changes in revenue;

(b) change in fleet requirements;

(c) changes in operating costs; and

(d) changes in cost recovery ratio (CRR).

(vi) The basic criteria for evaluation of the strategies’ effect on the community are :

(a) economic evaluation;

(b) policy consideration; and

(c) traffic consideration.

(vii) Economic evaluation - The economic analysis of each option will be based on the
following estimates :

(a) number of vehicle kilometres and hours;

(b) passenger hours;

(c) operating costs;

(d) changes in fleet and depot requirements; and

(e) interchange requirements.

(viii) Values of time and the determination of capital and operating costs should be based on the
most recently available information.

(ix) Policy Consideration - An indicator reflecting the degree of conformity to the stated policy
objectives can be applied to individual strategy.

(x) Traffic Consideration - Different strategies will lead to different volumes of public
transport vehicles and hence cause different levels of traffic congestion on road links.
Accordingly, the traffic implications will be assessed on an overall basis.

(xi) The assessment includes a comparison between the financial impact on public transport
operators, in terms of the resources consumed, and the savings in travel times in order to
demonstrate the trade-off.

2.3.8.7 Recommendations
To develop a coherent future strategy, it is necessary to maintain a balance between the costs and
benefits to the public transport users, operators and the community. In addition, flexibility should be
provided to modify the details of the strategy whenever situation warrants.

2.3.8.8 Consultation Strategy and Implementation Programme

(i) To facilitate smooth implementation of the recommendations of the study, a consultation

strategy to seek views of the concerned parties (e.g. District Councils) should be
formulated before finalising the proposals.
 (ii) Having completed the consultation and duly considered the views of relevant parties, the
study will put forward the recommended route development programmes for the public
transport operators with the suggested implementation programme.

2.3.9 Public Transport Modelling Process

2.3.9.1 This section provides more details on the public transport modelling process on a regional scale. The
public transport model projects the volume of patronage as a function of land use. Hence, a change in
either the land use data or the assumed network configuration will have impact on the travel patterns.
As population grows, the number of passenger trips is expected to increase. If the road and highway
network change, the travel pattern will change.

2.3.9.2 A public transport model comprises of three main components:

(i) a traffic zone system and associated demographic data;

(ii) a base network (public transport and road); and

(iii) an algorithm (four-stage transportation modelling procedure).

2.3.9.3 The general form of the public transport model is depicted in Diagram 2.3.9.3.The approach starts by
considering a zoning and network system, and the collection and coding of planning, calibration and
validation data. The model is presented as a sequence of four-stage sub-models, which includes trip
generation, distribution, modal split and trip assignment.

2.3.9.4 It should be noted in the public transport modelling process that the highway modelling should also be
considered simultaneously as the patronage forecasts will be significantly affected by the road traffic
conditions.

2.3.9.5 Patronage forecasts for strategic and local area links within the territory involve two levels of
modelling, i.e. the territory-wide strategic model and local or sub-area model. The hierarchy of public
transport model involving the two levels of modelling is illustrated in Diagram 2.3.9.5.

2.3.9.6 Refinements are necessary in the sub-area modelling process to detail the network and zonal
characteristics. Network expansion and matrix disaggregation are the procedure for the conversion of a
strategic model to a sub-area model. Diagram 2.3.9.6 shows the disaggregation of a matrix in CTS
level to sub-area level.
Input Parameters

Zoning System

2.3.9.7 A zoning system is used to aggregate the individual households and premises into manageable chunks
for modelling purposes. For a particular study, a zoning system is normally developed with the model
of the CTS as the basis and combined with the more detailed zoning system developed in previous
studies. The disaggregation of zones is sometimes necessary due to special condition such as the
importance of the areas around the ferry piers for a cross harbour public transport study.

Land Use and Demography Data

2.3.9.8 Land use planning data indicate where people live, where they work, shop and go to school. For each
zone the following data are specified:

(i) Population

(ii) Households

(iii) Household income

(iv) Resident workers

(v) Resident students

(vi) Age distribution in 10 categories

(vii) Residential parking spaces per household

(viii) School places in 4 categories

(ix) Employment places in 12 categories

(x) Hotel rooms

(xi) Market stalls

(xii) Cinema and theatre seats

(xiii) Hospital beds

(xiv) Container vehicle parking

(xv) Container yards

(xvi) Container depots

(xvii) Container vehicle repair yards

(xviii) Open storage sites

 Networks

2.3.9.9 As mentioned in paragraph 2.3.9.4, public transport and highway models should be considered
simultaneously in the modelling process. Therefore, the model requires the highway network and
public transport to be described. A network is described in terms of nodes and the road links
connecting them. In addition, nodes are required to represent zone centroids, the points where trips start
and finish. They are connected to the network by notional links such as ‘centroid connectors’.

2.3.9.10 The highway network includes all roads in Hong Kong down to the level of local distributor if they
carry through traffic. The following information is required for each road link :
(i) Road type

(ii) Frontage type

(iii) Geographical sector

(iv) Area volume delay function

(v) Link volume delay function

(vi) Distance

(vii) Peak and off-peak speeds

(viii) Number of intermediate traffic signals on link

(ix) Carriageway type and width

(x) Junction type and width

(xi) Junction type and characteristic

(xii) Capacity

2.3.9.11 Public transport services and their accessibility to the public are represented in the public transport
network by route and walk links. All franchised bus, ferry, rail and GMB/RMB services are included in
the route data which consist of the operating frequency and structure of the service (in terms of a
sequence of nodes representing interchange points with other routes or modes). The following
information is required for each public transport service :
(i) Fare

(ii) Vehicle capacity

(iii) Headway

(iv) Operating speed

(v) Dwell time

(vi) Layover time

 Modelling Algorithm

2.3.9.12 The control of the modelling process is represented by a four-stage transport model, which includes trip
generation, trip distribution, modal split and traffic assignment. The first two stages i.e., trip generation
and trip distribution, are aimed at producing person trips generated/attracted in each zone and are
represented in a matrix format. Once the trip matrices have been obtained, they are then splitted into
vehicle and public transport matrices for further process. These matrices are then assigned to public
transport and highway networks simultaneously to produce volumes of patronage and vehicle on public
transport and highway networks respectively.

Trip Generation

2.3.9.13 The aim of the trip generation is to project the number of trips likely to attract to a zone, irrespective of
origin and leaving a zone, irrespective of destination.

2.3.9.14 The standard approach to obtain the required zonal trip totals is to derive relationships which estimate
household trip generation rates from observed data. The number of trip will be dependent on land use
data.

Trip Distribution

2.3.9.15 The process of distributing modelling may be viewed as building up a matrix of person-movements
and, in general terms, the number of trips in the matrix cell is likely to be related to:
(i) Characteristics of the origin/production zone;

(ii) Characteristics of the destination/attraction zone; and

(iii) Characteristics of the separation, i.e., the perceived ‘generalised cost’ of travel between
zones.

Trip Matrices

2.3.9.16 There are various methods for compilation of a new set of trip matrices such as the gravity model, logit
model and disaggregation model. Alternatively, the trip matrices of CTS or previous studies can be
used as the basis and updated to include the new planning data and additional zones. The public
transport trip matrices are divided into modal matrices for different modes i.e. MTR, KCR heavy &
light rail, bus, ferry, public light bus, tram and taxi.

2.3.9.17 To devise the public transport matrices, it may be necessary to pre-determine the combined loading to
the public transport networks. The operators’ return and public transport surveys like origin and
destination surveys and screenline survey undertaken are used to control the overall number and
distribution of public transport trips. Where necessary, adjustment (factoring) to the previous matrices
should be made to reflect the latest information and through factoring process to build the future
matrices. Very often, the trip matrices produced will be crossed checked with the forecast by CTS.

Modal Split

2.3.9.18 The separation of trips by private car and those by public transport. For this process it is customary to
assume that people who do not have a car available will be captive public transport passengers. These
trips would be assigned, without further consideration to the public transport system.

2.3.9.19 People with a car available to them, however, do not necessarily always use it – they have a choice of
travelling either by car or by public transport – a modal choice. This choice is assumed to be based on
their perception of the generalized cost of each of the alternative modes of travel and this is
customarily express in the form:
2.3.9.20 Another methods of modelling the modal split are by the use of “diversion curve”, usually determined
from the survey data, and “nested logit” model. Within a nested structure, the first choice may be
between private car and public transport. For those choosing public transport, there may be a further
choice between bus and rail.

Traffic Assignment

2.3.9.21 An assignment model takes a matrix of trips (which is the output of the trip distribution process) and
assigns it as traffic onto an appropriate network i.e., the public transport and highway network. The
assignment process is as follows:

(i) Choice of route for each zone pair;

(ii) Aggregation of path flows on the links of the chosen paths;

(iii) Introduction of supply-side constraints, as a result of the volume of link flows nearing or
exceeding capacity; and

(iv) Estimation of the resulting generalised cost for each zone pair.

2.3.9.22 According to the concept of optimal strategy, multipath transit assignment technique is best suited for
public transport assignment model and high-frequency urban transit networks.
 Diagram 2.3.9.3: The Transport Model

Diagram 2.3.9.5: Hierarchy of Public Transport Models

Diagram 2.3.9.6: CTS and Subarea Traffic Zone Systems
2.3.10 Model Validation

2.3.10.1 The objective of the model validation process is to ensure that the model accurately replicates the
existing patterns of movements for design years forecasting in respect of the areas and modes which
are of particular interest to the study. The main validation criteria to assess the ability of the model is to
project patronage on a corridor basis and make comparison between the observed and modelled
patronage by service. The public transport elements of the model have to be validated as follows:

(i) Daily public transport boardings

(ii) Daily public transport persons trips across screenlines

(iii) Daily rail link flows

(iv) AM peak rail link flows

(v) PM peak rail link flows

2.3.10.2 The model is validated through the comparison of modelled patronage with the observed flows. If there
are discrepancies between the observed and synthetic flows, the model should be recalibrated or the
output matrices should be adjusted to reflect actual situation.

2.3.10.3 There is no hard and fast rule to validate the degree of accuracy of a model and it mainly depends on
the intention of its application. A model which is developed to compare alternative solutions to a
particular problem should be most adept at reproducing those parts of the study area in the immediate
region of the problem and those which may be significantly affected by any possible solution.

2.3.11 Model Application

2.3.11.1 The final modal split model is incorporated into the computer software for running the program and
testing alternative networks. From the base network, subsequent strategy test networks are created and
revised travel costs are obtained from each transport mode. These are then used to rerun the modal
choice model which produces the revised trip matrices for each mode. These matrices are then
separately reassigned to the network to obtain the link and service loading for each mode.

2.3.11.2 Issues that may be examined include :

(i) Impact of new public transport investments e.g. new rail systems and new stations;

(ii) Potential demand for major new transport initiatives, including park and ride schemes;

(iii) Results of different transport strategies, network structure and service level;

(iv) Effects of overcrowding both on vehicles and inside stations;

(v) Implications of different fare policies and structures;

(vi) Extent of mode shift between private car and public transport in response to road
congestion, traffic restraint or public transport improvements within multi-mode models;
and

(vii) Impact of improvement to a public transport interchange.