GUIDELINES FOR NATURAL

TERRAIN HAZARD STUDIES

GEO REPORT No. 138

K.C. Ng, S. Parry, J.P. King, C.A.M. Franks & R. Shaw

GEOTECHNICAL ENGINEERING OFFICE

CIVIL ENGINEERING DEPARTMENT

THE GOVERNMENT OF THE HONG KONG

SPECIAL ADMINISTRATIVE REGION

GUIDELINES FOR NATURAL

TERRAIN HAZARD STUDIES

GEO REPORT No. 138

K.C. Ng, S. Parry, J.P. King, C.A.M. Franks & R. Shaw

This report was originally produced in April 2002

as GEO Special Project Report No. SPR 1/2002
 - 2 -

The Government of the Hong Kong Special Administrative Region

First published, December 2003

Prepared by:

Geotechnical Engineering Office,

Civil Engineering Department,

Civil Engineering Building,

101 Princess Margaret Road,

Homantin, Kowloon,

Hong Kong.

 - 3 -

PREFACE

In keeping with our policy of releasing information

which may be of general interest to the geotechnical profession
and the public, we make available selected internal reports in a
series of publications termed the GEO Report series. The GEO
Reports can be downloaded from the website of the Civil
Engineering Department (http://www.ced.gov.hk) on the
Internet. Printed copies are also available for some GEO
Reports. For printed copies, a charge is made to cover the cost
of printing.

The Geotechnical Engineering Office also publishes

guidance documents as GEO Publications. These publications
and the printed GEO Reports may be obtained from the
Governments Information Services Department. Information
on how to purchase these documents is given on the last page of
this report.

R.K.S. Chan

Head, Geotechnical Engineering Office

December 2003

 - 4 -

FOREWORD

Demand for new land for development has not eased in recent years and
land nearer to steep natural slopes has to be considered for building and
infrastructure projects. There is a need to contain the increase in overall natural
terrain landslide risk to the community and to ensure that new developments are
not subject to undue risk.

The potential hazards posed by natural terrain to developments close to

hillsides are studied so that appropriate mitigation measures are incorporated
where necessary into development planning. Over the years, studies of natural
terrain and other relevant research and development work have led to better
understanding of hillside mass wasting processes in Hong Kong. Progressively,
a set of procedures has been developed for evaluating natural terrain hazards. In
setting up these procedures, overseas practice has been studied and there has been
close interaction with the local geotechnical community.

This report describes the current requirements for a natural terrain hazard
study, and outlines the approaches to studying natural terrain hazards and the
design requirements for appropriate mitigation measures. Technical guidelines
on the work involved in natural terrain hazard study and mitigation strategy were
prepared by consolidating current practices.

This document is based on an earlier Geotechnical Engineering Office

(GEO) report No. SPR 5/2000 by K.C. Ng, J.P. King, C.A.M. Franks and R.
Shaw, issued in August 2000. The revision, based on the experience gained and
feedback received over the last year, was carried out by K.C. Ng and S. Parry
under the direction of H.N. Wong. In particular, the following areas have been
updated or added: revision of the criteria for screening of sites subject to natural
terrain hazards, explanation of the differences between Design Event and
Quantitative Risk Assessment approaches, the requirement of a detailed aerial
photograph interpretation at an early stage of the study, review of notable natural
terrain landslide studies, relevance of geomorphology to natural terrain hazard
studies, and environmental consideration and maintenance requirements of
mitigation measures.

Practitioners are encouraged to comment at any time to the GEO on the

contents of this document so that improvements can be made to future editions.
When needed, the GEO will issue Technical Guidance Notes for promulgation of
further amendments to the report.

( R.K.S. Chan )

Head, Geotechnical Engineering Office

April 2002

 - 5 -

ABSTRACT

Natural terrain covers about 60% of the land area of the Hong Kong Special
Administrative Region. Today, most developed areas in Hong Kong extend from the limited
areas of level ground onto sloping ground. The need for natural terrain hazard studies is
increasing as the pressure on land results in building and infrastructure projects being located
nearer to natural hillsides.

This report documents current recommended practice and procedures for studying
natural terrain hazards. A Natural Terrain Hazard Study (NTHS) normally includes desk
study, detailed aerial photograph interpretation, and field mapping. The study should
provide sufficient information to determine the hazard models, their respective design events
and run-out distances. If the risks posed by an identified hazards needs to be lowered,
formulation of a mitigation strategy is required as part of the study.

For practical purposes, it may be expedient to proceed with an NTHS in stages,

relating to the progression of a development project, typically from land use planning,
through engineering feasibility study to the design stage. If it is established at any stage of
the study that the natural terrain hazards are inconsequential, there is no need to proceed with
further studies.

It is much more practical and economical to study and deal with the natural terrain
hazards at the time of development, than to address the problem later. If the potential
hazards are considered at an early stage of the development project, the hazards can be
controlled in a cost-effective manner by such methods as adjusting the facility layout,
providing buffer zones, delineating no-build zones, implementing suitable drainage provisions,
or constructing check dams and boulder fences. When planning the mitigation strategy, due
consideration should also be given to environmental issues and maintenance requirements.

The preferred technical approach for dealing with natural terrain hazards is to mitigate
the risk through adjustments to the layout of new developments, instead of carrying out
stabilization works to large areas of natural terrain, which would be both impractical and
environmentally damaging.

While studies are still on-going to further our understanding about the complex
interplay between the various attributes contributing to natural terrain instability, experienced
judgement by competent professionals remains a key aspect in assessing hazards and in
formulating appropriate mitigation strategies.
 - 6 -

CONTENTS

Page
No.

Title Page 1

PREFACE 3

FOREWORD 4

ABSTRACT 5

CONTENTS 6

1. INTRODUCTION 10

1.1 Purpose 10

1.2 Natural and Quasi-natural Terrain 10

1.3 Natural Terrain Landslides 11

1.4 Natural Terrain Landslide Risk Management 12

1.5 Requirements for the Study of Natural Terrain Hazards 13

1.5.1 Screening of Sites Subject to Natural Terrain Hazards 13

1.5.2 In-principle Objection Criteria 14

1.5.3 Alert Criteria 15

1.6 Objectives of Natural Terrain Hazard Study 16

1.7 Personnel 16

2. APPROACHES TO A NATURAL TERRAIN HAZARD STUDY 17

2.1 Scope of a Natural Terrain Hazard Study 17

2.2 Site and Study Area 17

2.3 Approaches to the Study 18

2.3.1 Introduction 18

2.3.2 Factor of Safety Approach 18

2.3.3 Quantitative Risk Assessment Approach 18

2.3.4 Design Event Approach 19

2.4 Design Requirements for Natural Terrain Hazards 19

2.5 Hazard Models 20

2.5.1 Introduction 20

2.5.2 Open Hillslope Landslide 21

 - 7 -

Page
No.

2.5.3 Channelized Debris Flow 21

2.5.4 Deep-seated Slide 21

2.5.5 Rock Fall 22

2.5.6 Boulder Fall 22

2.6 Approaches to Study Hazard Models 22

2.6.1 Introduction 22

2.6.2 Open Hillslope Landslide and Channelized Debris Flow 22

2.6.3 Deep-seated Slide 22

2.6.4 Rock Fall 23

2.6.5 Boulder Fall 23

3. NATURAL TERRAIN HAZARD STUDY 23

3.1 Introduction 23

3.2 Desk Study 24

3.2.1 General 24

3.2.2 Types of Information 24

3.2.3 Topographical Maps 24

3.2.4 Aerial Photographs 25

3.2.5 Geological Maps 25

3.2.6 Terrain Classification Maps 26

3.2.7 Natural Terrain Landslide Inventory 27

3.2.8 Large Landslide Dataset 27

3.2.9 Boulder Field Inventory 27

3.2.10 Rainfall Data 28

3.3 Aerial Photograph Interpretation 28

3.3.1 Purpose 28

3.3.2 Review of Existing Data 28

3.3.3 Review of the Site Condition 29

3.3.4 Interpretation 29

3.4 Collection of Field Data 32

3.4.1 Introduction 32

3.4.2 Site Reconnaissance 32

 - 8 -

Page
No.

3.4.3 Field Mapping and Site Measurement 33

3.5 Geological and Geomorphological Model 34

3.6 Design Event 35

3.6.1 Overview 35

3.6.2 Determining the Design Event 35

3.6.3 Catchment-specific Landslide Data 37

3.6.4 Regional Landslide Data 37

3.6.5 International Data 38

3.6.6 Special Consideration for Channelized Debris Flows 38

3.7 Run-out Assessment 39

3.8 Presentation of Results 39

3.9 Stages of Natural Terrain Hazard Study 41

3.9.1 Introduction 41

3.9.2 Review 41

3.9.3 Assessment 42

3.9.4 Mitigation Strategy 43

4. NATURAL TERRAIN HAZARD MITIGATION STRATEGY 43

4.1 Introduction 43

4.2 Typical Mitigation Strategies 43

4.2.1 Review 43

4.2.2 Passive Mitigation Strategies 44

4.2.3 Active Mitigation Strategies 44

4.3 Hong Kong Practice 45

4.4 Environmental Considerations 46

4.4.1 Introduction 46

4.4.2 Bio-engineering 46

4.4.3 Visual Impact of Mitigation Measures 47

5 MAINTENANCE OF MITIGATION FACILITIES 47

5.1 Introduction 47

5.2 Types of Natural Terrain Hazard Mitigation Measures 48

 - 9 -

Page
No.

5.3 Maintenance Requirements for Mitigation Measures 48

5.4 Other Measures 49

6. REFERENCES 50

LIST OF TABLES 57

LIST OF FIGURES 67

APPENDIX A: GLOSSARY OF TERMS AND GENERAL 72

CLASSIFICATION OF HILLSIDE FAILURES

APPENDIX B: NOTABLE NATURAL TERRAIN LANDSLIDES 94

IN RECENT YEARS

APPENDIX C: GEOMORPHOLOGY AND NATURAL TERRAIN 109

HAZARDS

APPENDIX D: BIBLIOGRAPHY 130

 - 10 -

1. INTRODUCTION
1.1 Purpose

This document outlines current recommended practice and procedures for studying
natural terrain hazards that may affect land development in the Hong Kong Special
Administrative Region (hereafter referred to as Hong Kong). It is aimed at practitioners who
have some familiarity with the subject but who may not be familiar with local conditions,
practice or procedures.

This report is a guidance document. As such, the methodologies contained within are
not mandatory. It is likely that situations will arise for which the report provides inadequate
or inappropriate guidance, and the designers must use alternative methods of approach.
There will also be improvements in professional practice that will supersede specific
recommendations of this report during its lifetime. For proper recognition to be made of
advances in our knowledge, it is therefore essential that practitioners continue to provide the
Geotechnical Engineering Office (GEO) with suggestions for improvement to the report.

1.2 Natural and Quasi-natural Terrain

Natural terrain (see Appendix A) in this document is defined as terrain that has not
been modified substantially by human activity but includes areas where grazing, hill fires and
deforestation may have occurred. Common anthropogenic features occurring on natural
terrain in Hong Kong include the following:

(a) Hillsides that have previously been used for agriculture.

Overgrown, abandoned agricultural terraces (Figures 1a
& 1b) with small masonry walls at the front are found in
many areas.

(b) Walking tracks, commonly along ridgelines, have in some

cases developed into deep erosion gullies. Four-wheel
drive vehicle tracks with minimal ground cutting have been
used by the military to access remote areas in the past and
remnants of other military artefacts such as dugouts,
foxholes and fortifications (Figures 1c & 1d) also exist.

(c) Hillside streams used for water supply and where small
dams, pipelines and catchwaters may be present.

(d) Traditional village graves of various styles, some with a

very similar arcuate shape to a landslide, occurring in
groups, or as isolated features, on natural terrain.

(e) Previously extensive small scale prospects and mines for

various minerals. Overgrown pits, trenches, adits and spoil
tips are the present day evidence of these and may be found
on natural terrain in many parts of the territory (Figures 1e
& 1f). Some indications of the areas where mining
 - 11 -

activities were more common are provided in Fyfe et al.

(2000) and Sewell et al. (2000).

In Hong Kong, the term quasi-natural terrain is used to describe natural terrain
which largely retains its original profile and regolith cover, but is partly modified by human
activities (see Appendix A). In addition to the above anthropogenic influences, modification
of the slopes above and below by cutting, filling and construction may have affected the
surrounding ground and surface water conditions. The guidelines given in this document are
generally also applicable to studying landslide hazards from quasi-natural terrain. However,
in studying quasi-natural terrain, relevant man-made effects need to be taken into account.

1.3 Natural Terrain Landslides

Natural terrain covers over 60% of the total land area (about 1,080 km2) of Hong Kong.
Nearly 50% of the natural terrain is sloping at 30 or more. As its present day morphology
reflects on-going geological and geomorphological processes, much of the natural terrain is
only marginally stable over large areas. From a review of high altitude aerial photographs
taken between 1945 and 1997 alone, about 27 000 past landslides on natural terrain were
identified, of which about 16 000 occurred within the last 50 years. On average, about one
natural terrain landslide occurs each year for every 2 km2 of the natural hillsides in Hong
Kong.

Several landslide-specific natural terrain studies have been carried out in recent years.
These well-documented incidents provide relevant information on the hazard model, probable
cause and controlling factors of natural and quasi-natural terrain failures in Hong Kong.
Consequently, they provide a useful reference for assessing natural terrain hazards. Some
notable studied natural terrain landslides are summarised in Appendix B. A broad review of
geomorphology with respect to natural terrain landslides is given in Appendix C and further
information on the geomorphology of Hong Kong is given in Fyfe et al. (2000).

Whilst most natural terrain landslides occur in relatively remote areas, some of them
do affect existing developments. Between 1982 and 1999, about 800 natural terrain
landslides affecting developed areas were reported to the GEO. Most of these landslides
were small-scale failures affecting open spaces, minor roads and footpaths, and other less
important facilities. In the past 20 years, failures of undisturbed natural slopes have not
resulted in any fatalities in planned developments, i.e. excluding squatter areas. However, a
boulder fall from natural terrain has resulted in one fatality and a failure of predominantly
natural hillside resulted in two fatalities in planned development. Over the same period, 18
fatalities were caused by failures of man-made slopes affecting planned developments. A
significant number of fatalities are also known to have occurred in squatter areas as a result of
landslides, and current Government policy is to deal with such unplanned developments by
clearance of squatters.

Although the historical natural terrain landslide data give an indication of the scale of
the problem in Hong Kong, they do not fully reflect the inherent landslide risk to the
community. Some landslides were near-miss incidents that could well have resulted in
more serious consequences. This situation will be aggravated as more new developments
take place on or close to steep natural hillsides.
 - 12 -

As is the case with other natural disasters, less frequent, larger-scale natural terrain
landslides, if they were to take place close to developed areas, could result in severe
consequences. These extreme events are under-represented by the limited historical
records available in Hong Kong. For example, the 20 000 m3 debris flow in 1990 on the
eastern hillslope of Tsing Shan, above Tuen Mun, reached the Area 19 platform below and
could have had more serious consequences if this area had been developed. This incident
highlights the potential risk of natural terrain landslides in Hong Kong.

1.4 Natural Terrain Landslide Risk Management

The current practice for natural terrain landslide risk management is aimed at keeping
the natural terrain landslide risk to an as low as practically achievable level. It entails the
following general principles:

(a) For new developments, contain the increase in overall risk

through studying the natural terrain hazards and undertaking
any necessary mitigation actions where such hazards occur.
Avoid new developments on sites that are subject to severe
natural terrain hazards.

(b) For existing developments, undertake mitigation actions

urgently where there exists an immediate and obvious
danger. Study natural terrain hazards where there is reason
to believe that a dangerous situation could develop, such as
persistent landslides affecting an existing development, and
undertake mitigation actions when considered necessary.

The Geotechnical Manual for Slopes (GCO, 1984) recognizes that natural slopes are
frequently close to limiting equilibrium over very large areas, that preventive works can be
expensive and difficult, and that it is not advisable to undertake extensive trimming-back of
natural slopes in order to achieve what may only be marginal improvement in stability.
Therefore, it is considered that disturbance of natural slopes and vegetation and the need for
costly preventive or protective measures should be avoided by siting structures away from
areas that could be affected by landslide debris.

The conventional approach to stabilization works on man-made slopes is generally not

suitable for natural terrain. Given the possible large areas of potentially unstable natural
terrain involved, such works are likely to be exceedingly expensive, and may not be justified.
Also, widespread stabilization works on natural terrain could cause considerable
environmental impact, and are difficult to carry out safely.

For new developments, the most effective means of risk mitigation is to avoid areas
that are exposed to undue natural terrain hazards. Whilst this is the solution commonly
adopted in many countries, it is sometimes not practical in Hong Kong where flat land is
limited. However, if the potential natural terrain hazards are considered at an early stage of
the development project, the hazards can be controlled in a cost-effective manner by such
methods as siting important facilities away from the most vulnerable areas, providing buffer
zones between the hillsides and the developments, delineating no-build zones for the most
 - 13 -

vulnerable parts of the sites, implementing drainage provisions to divert debris-laden

stormwaters from important facilities, specifying minimum clearance of occupied floors from
ground level, integrating debris resisting structures as part of the building structures,
constructing check dams and boulder fences, etc. It is much more practical, economical and
safe to study and deal with the hazards at the time of development, than to attempt to address
the problem later. This approach has been successfully adopted in a number of recent
developments.

Given the current state of knowledge and available technology it is not possible to
predict with confidence where and when natural terrain landslides will occur. For existing
developments, given the vast number of sites involved, it is neither desirable nor realistic to
deal with them all at the same time. However, for existing developments known to be
subject to significant hazards, it is prudent and practical to mitigate the hazards. This
react-to-known-hazard strategy has been adopted in dealing with boulder fall hazards in
Hong Kong, and has recently proved practical in applying to mitigate natural slope failure
hazards.

1.5 Requirements for the Study of Natural Terrain Hazards

1.5.1 Screening of Sites Subject to Natural Terrain Hazards

Not all development sites are affected by natural terrain hazards. Whether a site will
be affected depends on how close the site is to the hillside and how susceptible the hillside is
to landsliding. To facilitate project planning and early identification of potential constraints,
it is good practice to carry out, where necessary, an initial screening as early as possible to
examine whether the site may be subject to natural terrain hazards. The screening is not
demanding either in terms of resources or time.

Sites that are located beyond the influence zone where landslide debris may reach
would not be subject to the natural terrain hazards even if landslides occur on the hillside.
Such sites may be excluded from further screening and study of natural terrain hazards. For
this reason, the GEO has established a set of simple guidelines to assist planners, land
administrators, project managers, etc. in identifying whether a site may require screening in
respect of natural terrain hazards. This set of Inclusion guidelines comprise the following
two conditions:

(a) the proposed development involves Group 1, 2 or 3 facilities

(see Table 1), and

(b) there is a hillside sloping at more than 15 within 100 m

horizontally upslope of the site. For the purpose of
applying the Inclusion guidelines, the term hillside
refers to sloping terrain that is undeveloped or largely
undeveloped. It excludes built-up areas, such as terraces
supported by retaining walls, and a series of paved
platforms separated by man-made slopes. Where there are
man-made slope features adjoining a sloping terrain that is
undeveloped or largely undeveloped, the hillside includes
both the man-made slope features and the sloping terrain.
 - 14 -

Where a development site satisfies the Inclusion guidelines, arrangements should be

made for a screening in respect of natural terrain hazards. The GEO has been adopting two
sets of technical criteria for screening of sites in respect of natural terrain hazards. These
two sets of criteria, viz. the In-principle Objection Criteria and the Alert Criteria
(Figure 2), are described in Sections 1.5.2 and 1.5.3 below.

The GEO applies the technical criteria in screening development proposals, to

determine whether an in-principle objection should be tendered or whether a natural terrain
hazard study (NTHS) is required. The screening process is generally conducted at the land
use planning stage when the case is referred to the GEO from the Planning Department, e.g.
during Section 16 Application under the Town Planning Ordinance (Cap. 131), or at the land
disposal or lease modification stage when Lands Department seeks GEOs input. The
screening should also be carried out as part of the technical feasibility of a proposed public
works project (i.e. Preliminary Project Feasibility Study report or Technical Feasibility
Statement).

The screening will provide information on whether any parts of the site may be
affected by natural terrain hazards, which will give an indication of likely scale of the
problem, facilitate preliminary design of the site layout, and help to determine where further
input on study of natural terrain hazards is required. It should normally come to a
conclusion on the feasibility of the proposed development, any need to avoid areas exposed to
severe hazards, and any provision to be made in the development for natural terrain hazard
study and mitigation.

1.5.2 In-principle Objection Criteria

In line with the current practice, the GEO would object in-principle to proposals for
the zoning and disposal of a site that fall under either (a) or (b) of the following conditions,
denoted as In-principle Objection Criteria:

(a) The site is faced with severe natural terrain hazards. A site
falls within this category if any proposed Facility Groups
1(a), 1(b) and 2(a) (Table 1) is either:

(i) located within an angle of reach of 35 from any natural

terrain at an elevation of 50 m or more above the
proposed site formation level (Figure 2a), or

(ii) located on, or immediately below, terrain that is known

to be affected by active, large-scale, deep-seated
movement (e.g. Area 19, Tuen Mun).

(b) The s ite is proposed for small-scale New Territories

Exempted House (NTEH) and is subject to potential natural
terrain hazards. However, the required NTHS and
mitigation works are disproportionate to the scale of the
proposed development and could thereby render the
proposed development not economically viable. A site
 - 15 -

falls within this category if it meets all of the following

conditions:

(i) the site meets GEOs requirement for study of natural

terrain hazards (see Section 1.5.3 below),

(ii) the site is not located in a cluster of existing

developments of a comparable nature and affected by
similar hazards, and

(iii) it is unlikely that the requirement for study and

mitigation of natural terrain hazards can be lifted by
alternative means, such as adjustment of the layout of
the proposed development, delineation of no-build zone,
simple drainage provisions, confirmation of no need for
NTHS and mitigation by a brief review of the natural
terrain hazards, etc.

Although the criteria are essentially applied to habitable facilities, Group 1(b)
facilities are included because they are comparable in terms of consequence if affected by
natural terrain landslides. Items (b)(i) to (b)(iii) above are to be assessed based on
professional judgement on a case-by-case basis, with account taken of the nature of the
proposed development, its proximity to the hillside, size and geometry of the site, and
available information on the conditions and history of the hillside. The above criteria are not
applicable to sites after disposal.

1.5.3 Alert Criteria

For a site that may be affected by natural terrain hazards, but does not satisfy the
In-principle Objection Criteria, an NTHS is required to study the hazards and identify any
mitigation measures required. The following criteria, denoted as Alert Criteria (Figure 2b),
are used by the GEO to help decide whether a site falls under this category:

(a) It is a new development site involving provision of Group 1

to 3 facilities (Table 1), or it is a redevelopment that
requires modification of the lease conditions and involves
either a significant population at risk or a significant
increase in population at risk.

(b) Where there is natural terrain outside the site, but within the
same catchment, that is at an angular elevation of 20 or
more from the site and where there is ground sloping at
more than 15 within 50 m horizontally upslope of the site,
provided that there is a credible debris flowpath to the site.

An NTHS may be required for sites that lie beyond the area delineated by the above
criteria, as for example for sites where there are historical landslides with long debris run-out
extending beyond these limits, and for sites that are either intersected by, or adjacent to, a
 - 16 -

natural drainage course.

A credible debris flowpath is generally a downhill path followed by surface water.

However, flowpaths that debris could follow, but are deemed unlikely to do so would not be
regarded as credible. For example, a debris flow path down a ridge line, rather than
descending into the catchment on either side of the ridge line, would not be credible.
Another example would be a site that is shielded from debris by a substantial structure such as
a large building.

In applying the Alert Criteria to help decide whether an NTHS is required for a site,
professional judgement is exercised with account taken of the nature of the proposed
development, its proximity to the hillside, size and geometry of the site, and available
information on the conditions and history of the hillside. In addition to recommending that
an NTHS be carried out, consideration is also given to alternative means for dealing with
natural terrain hazards. These include adjustment of the layout of the proposed development,
delineation of no-build zones, simple drainage provisions, etc (see Sections 1.4 & 4.3).

1.6 Objectives of a Natural Terrain Hazard Study

The potential risk from landslides on natural hillsides, unless mitigated, will inevitably
increase as the population of Hong Kong grows and as building and infrastructure
developments continue to spread into areas adjacent to natural hillsides. The relative
significance of the natural terrain hazards to any site will vary from non-existent in the middle
of a flat plain to very high at a site below steeply sloping ground with a history of landslides.

The objectives of an NTHS are to identify any natural terrain hazards that could affect
a proposed development and to enable a decision to be reached as to whether the potential risk
to the site warrants mitigation. In practice, for sites that are found by screening (Section 1.5)
as requiring an NTHS, there is the potential to curtail the NTHS during the study if there is
sufficient information to conclude that no significant hazards exist. Alternatively, as more
data are obtained about the site, it may become possible to establish that whilst hazards may
exist, mitigation measures are unnecessary.

1.7 Personnel

Given the nature of the work involved, a multi-skilled team is normally required to
undertake an NTHS. The make-up of the team will depend on the complexity and level of
the study. In general, the team should include professionals with suitable engineering
geological and geotechnical engineering experience in NTHS. Engineering geological
expertise is needed for aerial photograph interpretation, engineering geological mapping,
interpretation of geological data, and development of geological and geomorphological
models. Geotechnical engineering expertise is required for analysis of engineering data,
assessment of slope stability and design of mitigation measures. Some tasks, such as
determination of design events and development of mitigation strategy would be best carried
out jointly by engineering geologists and geotechnical engineers. In some cases, input from
risk analysts may also be required in detailed quantitative risk assessment.
 - 17 -

As an NTHS forms part of a geotechnical submission, it should be reviewed by a

professionally qualified geotechnical engineer who has suitable local experience in natural
terrain hazard studies.

2. APPROACHES TO A NATURAL TERRAIN HAZARD STUDY

2.1 Scope of a Natural Terrain Hazard Study

The scope of an NTHS comprises:

(a) the determination of the location, type and magnitude of any

natural terrain hazards that may affect the development of a
site, and

(b) where an identified hazard is deemed to affect a site, to

determine the design event for such a hazard and outline a
mitigation strategy to reduce the risk it poses to the
development.

The level of study required to determine if any hazards exist and to select suitable
mitigation measures if required will vary according to the setting of a site. The resources
required to study a site will also vary according to the size of the site and its proposed use.
Siting critical facilities away from vulnerable areas, where possible, will often result in a
substantial saving with respect to the input required to address the natural terrain hazards. It
will also minimize exposure to risk associated with uncertainties or unforeseen conditions.

2.2 Site and Study Area

The study of any natural terrain close to a proposed development should be related to a
well-defined site boundary, preferably recorded on a topographical map of scale 1:5 000 or
larger. In the early stages of a project, there may be a number of possible uses and layouts
for a site, which may be subject to changes. At this stage, the study may have to consider
the entire area inside the site boundary as potentially habitable. Hence, any hazards would
be evaluated in relation to the site boundary. The study should determine if any identified
hazards could result in debris being deposited within the site boundary and if so, to establish
the debris travel path, magnitude and nature of such materials.

The study area for any site is normally the entire upslope catchment of natural terrain
above the site boundary. This may include natural terrain separated from the site by existing
developments such as roads or buildings, unless these developments are judged to provide a
barrier or buffer against debris reaching the site from the natural terrain. It may be possible
during the early stage of a study to exclude parts of the catchment on the basis that simple
topographical relationships show that the terrain could not affect any part of the site.
However, in order to assess the potential natural terrain hazards in its regional context,
catchments in the relevant vicinity also need to be considered.
 - 18 -

2.3 Approaches to the Study

2.3.1 Introduction

Three different approaches, namely, Factor of Safety, Quantitative Risk Assessment

and Design Event, may be used, either individually or in combination, to evaluate natural
terrain hazards.

2.3.2 Factor of Safety Approach

The Factor of Safety approach is set out in the Geotechnical Manual for Slopes (GCO,
1984). This approach requires the study of the stability of the natural terrain and design of
any necessary slope stabilization measures to meet the requirements given in the Manual.

As stipulated in Section 5.2.3 of the Manual, natural terrain does not need to meet the
factors of safety given in Table 5.1 of the Manual if it is undisturbed and a careful
examination is made to determine that there is no evidence of instability or severe surface
erosion. The term evidence of instability used in the Manual refers to evidence relevant
to future instability.

By implication, if one chooses to adopt the Factor of Safety approach and if the natural
terrain does not satisfy the two criteria given in Section 5.2.3 of the Manual, the natural
terrain has to meet the factors of safety stipulated in Table 5.1 of the Manual. In practice,
few natural hillsides satisfy both criteria.

The Factor of Safety approach can be used by designers who opt for prevention of
natural terrain failures. In the past, the Factor of Safety approach was adopted in the study
of natural terrain below development areas to check that the sites would not be adversely
affected by failure of the natural hillside. It has also been used in assessing the stability of
hillsides above development sites against large failures or deep-seated landslides.

2.3.3 Quantitative Risk Assessment Approach

The Quantitative Risk Assessment (QRA) approach entails detailed assessment of the
probability and consequence of natural terrain hazards to a development site and determines
any necessary mitigation measures. A comprehensive review of the QRA concepts as
applied to slope engineering is given in Ho et al. (2000). Lo (2001) gave an interim review
of pilot application of QRA to landslide problems in Hong Kong.

This approach provides a formalized methodology to gauge the relative importance of

different parameters and the uncertainties associated with them, and so to determine the
acceptability of a calculated risk level by comparing it to established guidelines.
Furthermore, it provides a common, definitive basis for applying the As Low as Reasonably
Practicable (ALARP) principle, and for evaluating the cost-benefit of alternative risk
mitigation strategies to optimise design.

The risk guidelines currently adopted by the GEO are given in ERM (1998), and are
for interim use as a basis for the evaluation of QRA results. These risk guidelines may be
 - 19 -

further refined as necessary.

QRA can be applied by designers who opt for mitigation of landslide risk instead of
prevention of natural terrain failures. The study calls for expert input, and the assessment
may be fairly involved and costly. Therefore, it would not be appropriate to apply QRA to
routine problems that can be adequately handled by conventional techniques. Relevant QRA
studies carried out in recent years include Atkins Haswell (1995), Atkins (1998, 1999a),
Halcrow (2000) and Fugro Maunsell Scott Wilson (2001).

2.3.4 Design Event Approach

The Design Event approach is based on a qualitative risk framework, with account
taken of both the susceptibility (or likelihood of failure) of the natural terrain and failure
consequence in a semi-quantitative manner. Under this approach, the relevant design failure
events are assessed for the different natural terrain hazards and any necessary mitigation
measures are evaluated. The design requirements for the Design Event approach are given
in Section 2.4 below.

Like QRA, the Design Event approach can be used by designers who opt for
mitigation of landslide risk instead of prevention of natural terrain failures. The Design
Event is relatively easy to apply as it does not demand formal and rigorous quantification of
risk. Many practitioners in Hong Kong favour the use of the Design Event approach where
the required mitigation measures are not disproportionate to the scale of the development.
However, as this approach does not explicitly consider the risk tolerability, it cannot evaluate
mitigation actions based on the ALARP principle, or justify not taking actions on the grounds
of a broadly acceptable risk level.

2.4 Design Requirements for Natural Terrain Hazards

For the Design Event approach, the need for any further study and the design
requirements for mitigation measures are evaluated, normally in an early stage of the study,
based on Table 2. This involves consideration of the consequence of failure and
susceptibility of the hillside to failure.

(a) Consequence of failure is categorized into five classes based

on the types of facilities (Table 1) affected and their
proximity to the hillside (Table 3).

(b) Susceptibility of the hillside to failure is categorized into

four classes based on historical landslide data and other
information (Table 4).

According to Table 2, further study is not required if the susceptibility of the hillside to
failure and the consequence of failure are insignificant. This is consistent with the
provisions in the Geotechnical Manual for Slopes (GCO, 1984), where there is no evidence of
instability and negligible consequence slopes. Otherwise, further study should be carried out
to establish the need for any mitigation measures to deal with the relevant design failure
 - 20 -

events. Depending on the susceptibility of the hillside to failure and the consequence at the
site, the required design event may be a conservative event or a worst credible event.
The return periods given below are notional, and are intended to reflect the design principles.
They should not be taken as accurate values derived from detailed statistical analysis.

(a) A Conservative event is a reasonably safe but not overly

cautious estimate of the hazard that may affect the site, with
a notional return period in the order of 100 years, and is
generally based on the largest historical landslide over the
past 50 to 100 years (e.g. identified from aerial photographs)
in the catchment and its vicinity as appropriate.

(b) A Worst credible event is a very conservative estimate

such that the occurrence of a more severe event is
sufficiently unlikely. Its notional return period is in the
order of 1 000 years, which indicates an order of magnitude
difference compared with that of a conservative event.
This is intended to exclude historical events that occurred in
the geological past. It generally corresponds to the largest
credible landslide based on interpretation of historical
landslide data, geomorphological evidence in the catchment
and its vicinity as appropriate, and any other relevant
evidence from similar terrain in Hong Kong.

The design requirements for the Design Event approach involve consideration of the
susceptibility of the hillside to landsliding. Hence, in applying the Design Event approach,
conservative assumptions may have to be made in the early stage of an NTHS on the
susceptibility of the hillside, to ensure that the design requirements would not be
underestimated due to lack of information on the susceptibility of the hillside. As further
work is carried out in the NTHS, the assumptions made on the susceptibility of the hillside
should be reviewed in order to confirm the design requirements.

For calibration purposes, the technical framework proposed for determining the design
requirements for the Design Event approach have been applied to some previous cases of
NTHS and to some selected sites of proposed development (Table 5). The framework gives
reasonable results.

2.5 Hazard Models

2.5.1 Introduction

Natural terrain hazards can be grouped into hazard models on the basis of the transport
mechanism, the nature of the displaced material and the topographical location. Five hazard
models that are generally applicable to most sites in Hong Kong are presented below. These
may not cover all possible situations and it may be useful to define sub-models or other
models as appropriate to the prevailing conditions. Some notable natural terrain landslides
occurring in recent years are summarized in Appendix B. Typical landslide types that have
been identified in Hong Kong are discussed in Section C.6 of Appendix C.
 - 21 -

The five hazard models described in the following sections should be considered in the
study of each site. For certain sites it may be possible to discount some or all of them at an
early stage. As a study progresses, the hazard models may need to be refined in the light of
additional information.

2.5.2 Open Hillslope Landslide

This type of hazard results from a landslide that remains wholly on the open hillside
and is not channelized along a stream course. It may include landslides classified as debris
slides, debris avalanche or debris flow (Figure C2). An open hillslope landslide commonly
involves a rapid to extremely rapid, shallow detachment of partially or fully saturated debris
on steep slopes. The initial displaced mass disintegrates and loses its original structure
during displacement but is not channelized in a stream bed. There is minimal additional
water admixed into the displaced material during transport. Debris transport can include the
mechanisms of flow, sliding, rolling and bouncing. These landslides may occur on any part
of a hillside but as the landslide debris moves downhill it will converge into hollows
(especially linear depressions) in its path. Thus, the hazard may be much greater where the
site boundary intersects such hollows. This model may be sub-divided on the basis of
particular initiation points such as coastal slopes, river slopes or previous landslide scars.

2.5.3 Channelized Debris Flow

This type of hazard results from saturated landslide debris being transported by slurry
flow or hyper-concentrated stream flow and can only occur along a stream (Figure C3).
Channelized debris flows generally have much greater mobility than open hillslope landslides
and normally develop when debris from one or more open hillslope landslides enters a stream
channel, and becomes mixed with stream water. The relative volume of debris and the
cross-sectional area of the channel (the channelization ratio) control the degree to which the
debris is confined by the channel. Collapse of the channel sides and erosion and entrainment
of material from the stream bed and banks may add additional debris. Channelized debris
flows may occur repeatedly in the same channel and such locations can sometimes be
recognized by the presence of debris fans from previous events.

2.5.4 Deep-seated Slide

This type of hazard involves displacement of an intact mass by sliding along a basal
rupture surface. Landslides in this model are generally slower moving and with a deeper
surface of rupture than open hillslope debris slides. The model does not include soil creep,
which is not generally considered to be a hazard; however, evidence of soil creep may provide
an indication of potential hazards. Deep-seated, intact sliding is relatively rare in Hong
Kong and usually occurs only as landslides with very short travel distances, where the
displaced material has not moved from the source area (slump) or where tension cracks and
bulging are the sole surface expression of movement, e.g. Siu Sai Wan (Ho & Evans, 1993)
and Island School (Irfan, 1989). In Hong Kong, slow, small sliding slope movements are
particularly associated with sites where natural terrain has been disturbed by excavation, in
some cases into degraded ancient landslide scars.
 - 22 -

2.5.5 Rock Fall

This type of hazard results from one, or several, typically angular rock fragments being
transported initially by free fall but may include sliding, bouncing and rolling. Rock
fragments are generally joint-bounded and displaced from a rock face by sliding, wedge
failure or toppling. Initiation may often be related to transient high water pressure along
joints and where smaller key blocks have been displaced. Locations where this hazard is
present are relatively easy to identify as it is limited to rock faces.

2.5.6 Boulder Fall

This type of hazard results from one, or several, typically rounded rock fragments
being transported by rolling, bouncing and sliding. Boulders originate either as corestones
in a weathered profile or are derived from colluvium. They become exhumed during
hillslope erosion, although in some instances they may have originated higher up the slope
from rock outcrops and been displaced to their present position as a rock fall. They are
usually dislodged from the natural hillside as a result of soil erosion or undercutting due to
landsliding. Where vegetation growth is minimal, locations of boulders on steep slopes can
be easily identified from interpretation of aerial photographs.

2.6 Approaches to Study Hazard Models

2.6.1 Introduction

The three approaches outlined in Section 2.3 above should not be considered as
independent of one another. A site may be exposed to several types of natural terrain hazard
and of varying scales. Some approaches may be more applicable to certain hazards than
others. Therefore, the designer should choose the most suitable approach, or combination of
approaches, for a site.

2.6.2 Open Hillslope Landslide and Channelized Debris Flow

As large areas of natural terrain may be susceptible to open hillslope landslides which
may develop into channelized debris flows, stabilization works on natural terrain to prevent
these hazards are usually impractical. Therefore, the QRA and Design Event approaches are
usually more appropriate for addressing these two types of hazard.

2.6.3 Deep-seated Slide

If the extent of the hillside that is susceptible to this type of hazard can be defined, the
Factor of Safety approach may be applied to determine the need for any slope stabilization
works. However, stabilization works on the hillside may be costly, particularly if a large
volume of potentially unstable ground is involved.
 - 23 -

2.6.4 Rock Fall

The Factor of Safety approach is generally applicable to dealing with rock fall hazard
because potentially unstable blocks can commonly be identified from field mapping.
Alternatively, in cases where the extent of the unstable blocks is difficult to ascertain for
in-situ stabilization works, or if it is more cost-effective to provide mitigation measures, the
Design Event approach may be more suitable. However, implementation of some mitigation
measures, such as rock ditches and slope drapery, is of prescriptive or empirical in nature.

2.6.5 Boulder Fall

There is an established methodology for studying boulder fall hazard and engineering
solutions have been available and adopted for many years (e.g. Au & Chan, 1991).
Therefore, this document does not specifically address boulder fall hazard. Depending on
the hillside conditions, all three approaches may be used to tackle this type of hazard.

3. NATURAL TERRAIN HAZARD STUDY

3.1 Introduction

The purpose of an NTHS is to identify natural terrain hazards that may affect a
particular site and to propose a mitigation strategy if required. The study of natural terrain
hazards involves consideration of the site in the context of its regional geological and
geomorphological settings, any man-made influences that may have modified this setting and
the history of landsliding in the area. This information together with an engineering
geological investigation allows a geological model of the catchment to be developed.

An NTHS generally comprises a desk study of available information about the study
area, an examination of all available sets of aerial photographs, collection of field data,
assessment of the natural terrain hazards, and determination of the hazard events to be
mitigated against. Ground investigation may also be required. Account may have to be
taken of catchments in the relevant vicinity or other similar locations to identify all possible
hazards.

The results of the study should allow a decision to be reached as to whether the
potential risk can be considered negligible, where the risk can be contained by mitigation
works, or whether the risk is so severe that it would render the proposed development not
feasible. Where applicable, the study should include determination of the location, potential
magnitude (i.e. the design events) and run-out characteristics of specific hazard models that
are likely to affect the site being studied. The economic viability of the proposed
development may form part of the study.

The following sub-sections of Section 3 outline the key components and procedures of
good practice recommended for an NTHS. The aspects on mitigation strategy of an NTHS
are given in Section 4.
 - 24 -

3.2 Desk Study

3.2.1 General

The desk study should comprise a thorough review of geological and geotechnical
information pertinent to the study area so as to identify those features that may be hazardous
to the site.

3.2.2 Types of Information

Desk study information relevant to site investigation in Hong Kong is described in

Geoguide 2 Guide to Site Investigation (GEO, 1987). Much of the information described
in Sections 4, 5 and 6 and Appendices A, B and C in the Geoguide is relevant to the desk
study. Appendix B of the Geoguide has been superseded by GEO Technical Guidance Note
(TGN) No. 5. Information concerning the history of developed areas is only likely to be
relevant to the site itself. A useful source of information covering natural catchments in
Hong Kong is the Stormwater Drainage Manual (Drainage Services Department, 1994).

Published maps that provide information on natural terrain for the whole of Hong
Kong include topographical maps, geological maps, and the Geotechnical Area Study
Programme (GASP) series maps based on terrain classification. Another map series based
on high altitude aerial photograph interpretation is the GEO Natural Terrain Landslide
Inventory. An inventory of boulder fields mapped from the 1963 low altitude aerial
photographs is also available. Most of these sources of data are available for consultation by
the public in the Geotechnical Information Unit (GIU) of the Civil Engineering Library (CEL)
maintained by the Civil Engineering Department (CED).

A considerable amount of slope safety-related information is contained in a

computerized Slope Information System (SIS). The SIS displays the locations of registered
man-made slopes, past landslide incidents and ground investigation records along with base
maps and lot boundaries. These are linked to slope inspection records, maintenance
information and photographs. Public access to the SIS is available through the GEO Hong
Kong Slope Safety Website, http://hkss.ced.gov.hk. By early 2002, the website will also
maintain an up-to-date list of geotechnical publications and technical papers on geology and
geotechnical engineering in Hong Kong.

If the site or other parts of the study area have previously been investigated, or
developed, then useful data may be available. This information may include records of past
landslide incidents, ground investigations, building records and details of man-made slopes.
Besides the building records which are held at the Buildings Department, a great deal of this
information is held in the GIU of the CEL and can be searched for by location using the
Geotechnical Information Library System (GILS) database. A bibliography of geological
and geotechnical publications in Hong Kong is given in Brand (1996).

3.2.3 Topographical Maps

Topographical maps of Hong Kong are available from the Map Publications Centre of
the Lands Department Survey and Mapping Office at the North Point Government Offices.
 - 25 -

Full coverage of Hong Kong is available at 1:20 000, 1:5 000 and 1:1 000 scales and each is
appropriate for different aspects of natural terrain hazard studies.

The 1:20 000-scale maps have been used as base plans for various thematic maps that
cover the whole of Hong Kong, including geology and GASP terrain classification. In
general, this scale is too small to show details of a study area but can be used to place the
study area in a regional context. The 1:5 000-scale maps are useful for delineating the study
area, and for large catchments they can be useful base maps to present data and information.
The 1:1 000-scale maps are a suitable base for aerial photograph interpretation and field
mapping. Slope angles derived at this scale are similar to those derived from the
1:5 000-scale maps, but the greater detail provides better resolution of smaller features.

3.2.4 Aerial Photographs

The earliest available aerial photographs of parts of Hong Kong were taken in 1924
with the next set of photographs taken in 1945. Other photographs of variable quality and
coverage are available up to the first complete survey of Hong Kong undertaken in 1963.
This is a set of high quality low altitude aerial photographs taken at a time when there was
generally low vegetation cover in Hong Kong, which allows clear observations of subtle
ground features and consequently provide a useful baseline for comparison with subsequent
observations. Since 1973, reasonable quality photographs covering most of Hong Kong
have been taken annually, although the low altitude coverage of many of the more remote
natural terrain areas of Hong Kong has been sporadic.

In addition to vertical aerial photographs, a collection of oblique aerial photographs is

held in Planning Division of GEO and should also be consulted. Some areas in Hong Kong
have been photographed using infra-red aerial photography, which may in some cases be
useful for identifying areas of seepage. Potential applications of remote sensing techniques
for identifying areas of seepage in Hong Kong and for geotechnical applications are discussed
in Scott Wilson (1999a) and Mason (2001).

Aerial photographs are available for sale from the Lands Departments Map Sales Unit
at the North Point Government Offices. For public works projects, much of the Government
collection can be borrowed from the GEO Aerial Photograph Library, housed in the Planning
Division of the GEO. The lending policy of the library and general background information
on aerial photography in Hong Kong are contained in GEO Circular No. 12.

3.2.5 Geological Maps

Geological maps cover the whole of Hong Kong at 1:20 000 scale and are available
from Government Publications Centres. These maps are regional surveys, and should not be
used to determine site-specific lithology. However, they are useful in forming an initial
geological model as the basis for detailed site mapping and ground investigation. These
1:20 000-scale maps are supported by a series of six memoirs describing the geology of Hong
Kong. The 1:10 000-scale field sheets that record the original field mapping observations
can be viewed at the Hong Kong Geological Survey, which is part of the Planning Division of
the GEO. The geology of Hong Kong is comprehensively described in two memoirs, one of
 - 26 -

which focuses on the Quaternary geology, including aspects of weathering and

geomorphology (Fyfe et al., 2000), and the other provides details on the solid geology (Sewell
et al., 2000).

A limited number of 1:5 000-scale maps, supported by sheet reports, have also been
published. They are mainly located in the development areas of North Lantau, Ma On
Shan and Yuen Long (Appendix D). These sheet reports can be obtained from the Planning
Division of the GEO. Some site-specific geological data are available in the Hong Kong
Geological Survey Archive of Maps and Reports and can be viewed in the Planning Division
of the GEO. Some of the geologically orientated data held by Planning Division of the GEO
have been digitized and are stored in a geographical information system. These include the
1:20 000-scale geological maps, some drillhole locations, and the locations of geophysical
surveys.

3.2.6 Terrain Classification Maps

A series of eleven 1:20 000-scale terrain classification maps (TCM) covering the
whole of Hong Kong were produced by the GEO in the 1980s. Each map covers an irregular
area based on catchment boundaries. Derivative maps and accompanying reports have been
published for each area as the Geotechnical Area Studies Programme (GASP) series. The
series contains 12 volumes (Appendix D), of which Volume XII is a summary report for the
whole of Hong Kong. They are available for consultation in the GIU.

The terrain classification system recorded three attributes, slope angle, landform, and
instability and erosion. This information was derived from interpretation of pre-1980 aerial
photographs. The slope angle information was derived from topographical maps and visual
estimates and should now be replaced with slope angles derived from more updated
large-scale maps. The landform data are still generally relevant although the erosion and
instability characteristics may have changed due to erosion control measures, particularly
revegetation. The terrain classification maps were used to create thematic maps that include
a Physical Constraints Map and the Geotechnical Land Use Map (GLUM). Appendix A to
each GASP report explains the system of terrain evaluation used.

It should be noted that these maps were prepared for general planning and resource
evaluation purposes. Whilst the information they contain provides a good basis for
engineering feasibility studies they should not be used to interpret parcels of land smaller than
3 ha in size.

Nine catchments with extensive colluvial deposits were also identified for GAS
mapping at 1:2 500 scale. These maps contain details of old landslide scars. In 1990, a
1:5 000-scale terrain classification mapping exercise was carried out for eleven 1:5 000-scale
map sheets in North Lantau (Appendix D). These are available for consultation in the CEL
and have been incorporated into three engineering geology reports on the North Lantau Area
(Franks, 1991, 1992; Woods, 1992).
 - 27 -

3.2.7 Natural Terrain Landslide Inventory

The Natural Terrain Landslide Inventory (NTLI) was compiled from the interpretation
of high altitude aerial photographs taken in various years between 1945 and 1994. Emery
(1996) and King (1999) document the methodology and procedures of the study. The NTLI
1:5 000-scale maps are available on the SIS (http://hkss.ced.gov.hk). Factors such as
coverage, scale and resolution of the high altitude aerial photographs impose certain
limitations on the data set. Consequently, some landslides may not have been identified and
some landslides shown may be other features mis-identified as landslides. Furthermore,
identification of landslides on the aerial photographs may have been hindered by poor image
resolution, cloud cover, ground shadows or vegetation cover. These maps are useful for
reviewing the landslide history of an area of natural terrain on a regional basis, but should not
be used as the only basis for individual site assessments. The NTLI has been updated for
1995 to 1997 as reported in Scott Wilson (1999b). An update of the NTLI for 1998 to 2000
is currently being underway and will be available in late 2002.

3.2.8 Large Landslide Dataset

The Large Landslide Dataset, covering the whole of Hong Kong, was prepared under
the Large Landslide Study. In this study, a geomorphological interpretation was conducted
to identify features thought to be landslides with source areas greater than 20 m wide. The
main source of the large landslide data was from the interpretation of the 1963/64 low altitude
aerial photographs (Scott Wilson, 1999b, 1999c). The study also examined the large
landslides recorded in the NTLI, GASP maps and geological maps. The 1:5 000-scale maps
containing the features are available for viewing in the Planning Division of the GEO.

The features that were identified range from fully vegetated, degraded features of
valley-side dimensions to single recent landslide scars. It should be noted that some features
might be the result of multiple smaller landslides. Furthermore, old landslides might have
occurred under environmental conditions different from the present. The classifications of
sliding, flowing or toppling/falling need to be confirmed by field observations.

3.2.9 Boulder Field Inventory

A series of fifteen 1:20 000-scale boulder field inventory maps covering the whole of
Hong Kong was completed in 1998. These maps were derived mainly from interpretation of
the low altitude aerial photographs taken in 1963/64. A multiple attribute mapping
technique was adopted to present the results - an area of land considered to have a relatively
uniform pattern of boulder deposits was delineated and enclosed within a polygon and drawn
onto the 1:20 000-sacle topographical maps (Emery, 1998). The polygons were defined to
ensure that the units were essentially homogeneous for the range of attributes described.
Four attributes were mapped - percentage area covered by boulder, boulder type, boulder size
and boulder shape. These maps, available for reference in CEL, provide information for
identification of boulder fall hazard and on the number, size, shape and type of boulders
present. For site specific assessment, these maps should be supplemented by detailed
interpretation of low altitude aerial photographs and by field mapping.
 - 28 -

3.2.10 Rainfall Data

The GEO operates an extensive network of automatic rain gauges providing real-time
rainfall data to support the Landslip Warning System. This network, which dates back to
1984, has recently been upgraded to improve coverage in the new development areas. The
network now comprises 86 rain gauges located throughout Hong Kong. The data capture,
control and processing system, which has also recently been upgraded, receives data from the
GEO raingauges and from an additional 24 automatic rain gauges operated by the Hong Kong
Observatory and 5 gauges operated by the Drainage Services Department. Since 1984, the
GEO has compiled annual reports to document information on rainfall and landslides
occurring in Hong Kong. Detailed reports on significant landslides, prepared by the GEO,
also analyse the relevant rainfall data.

3.3 Aerial Photograph Interpretation

3.3.1 Purpose

Aerial Photograph Interpretation (API) should be used as the first step in the
development of geomorphological and geological model as well as to examine the history of
development and past instability in the study area. Being one of the most important parts of
an NTHS, interpretation of aerial photographs should be undertaken by an experienced
professional.

The objective of the API is to determine, in as much detail as possible, the terrain
attributes of the study area and its surroundings that may influence hazards, as well as details
of past events. These include the characteristics, magnitude and observed consequences of
past natural terrain hazards, the extent of any boulder fields, drainage characteristics, slope
gradient, slope aspect and morphology, characteristics of vegetation cover, location and types
of rock outcrops and superficial deposits, etc. Table 6 gives further guidance on terrain
attributes that can be mapped from aerial photographs. The API results help identify areas
for field data collection and provides the basis for determining landslide susceptibility.

3.3.2 Review of Existing Data

Many of the datasets available for an NTHS were derived from the interpretation of
aerial photographs (e.g. Terrain Classification Maps, Natural Terrain Landslide Inventory,
Large Landslide Dataset and Boulder Field Inventory). As described in their respective
sections (Section 3.2), these data were prepared under various constraints; some due to data
quality whereas others due to method of mapping. Consequently, it is recommended that
these datasets should be compared with the results of the API in order to evaluate the
relevance of the desk study information for the site of interest. For example, the NTLI
(Section 3.2.7) suffers from certain limitations inherited in the coverage and quality of the
high altitude aerial photographs. Therefore, the NTLI features should be confirmed and
supplemented by detailed interpretation of low altitude aerial photographs as part of the
NTHS. Account should also be taken of the coverage and quality of the datasets,
corresponding rainfall history at the site and any environmental changes (e.g. hill fires,
man-made effects) that have occurred in the area.
 - 29 -

Using the NTLI data (Section 3.2.7), Evans and King (1998) integrated geology and
slope angle to derive the regional susceptibility of debris avalanches in Hong Kong.
However, in applying this regional susceptibility, due consideration should be given to the
methodology and assumptions adopted in deriving the correlation. In general, this regional
correlation should not be regarded as sufficiently reliable for use in site specific assessment.

3.3.3 Review of the Site Condition

The most recent aerial photographs provide information on the current condition of the
study area. This is a useful check on the current status of the base map to be used for the
API and may provide data to update it. Observations from these photographs can also be
useful for planning the site reconnaissance by identifying suitable locations for an overview
of the site and possible access points. The best access is normally along footpaths and
streams and the current vegetation cover in these locations is often the most important factor
that affects access. Vegetation cover can change rapidly in Hong Kong due to hill fires or
clearance by man and regrowth can leave a cleared area difficult to access.

3.3.4 Interpretation

An important part of the API is mapping topographical, geomorphological and

geological features to assist in the development of a geological/geomorphological model.
The 1:1 000-scale topographical maps generally represent sufficient details of the topography
to be used as the base maps for recording information. The observations need to be
transferred onto a suitable base map, preferably at a scale similar to that of the photographs.
The use of ortho-photographs is extremely beneficial in that it allows the accurate location of
features and enables scaled measurements to be made. A number of studies have used
ortho-rectified photographs for measurement of ground features and as base plans for field
mapping.

The mapping should be carried out using a well-defined legend that includes all the
aspects of engineering geology and geomorphology covered by the mapping. A generally
accepted system is described in Anon. (1972). Important aspects to consider include the
scale, flying height, year of photography, time of year and quality of the photographs. Both
high and low altitude photographs should be included in any interpretation. Interpretation of
good quality oblique aerial photographs and site photographs can also provide useful
information.

The interpretation should make use of as complete a history as possible, preferably

using both dry season and wet season coverage as this may allow more precise age-bracketing
of specific landslides, and/or determination of the effect of changed conditions such as slope
excavations or hill fires. It is important to review aerial photographs from as many
consecutive wet seasons for a catchment as possible to build up a history on the likelihood of
landslide initiation.

A glossary of commonly used terms for landslide description and related subjects is
given at Appendix A. Relevant features that can be recorded from API are discussed below.
 - 30 -

(a) Catchment Characteristics - Catchment boundaries can be

delineated on the basis of the 1:1 000 scale maps but details
should be confirmed by API. Hollows, channels and
streams should also be marked.

(b) Superficial Geology - The ground surface should, as far as

possible, be divided into areas of rock outcrop and regolith.
It may be possible to interpret structure or lithology from
the rock outcrop and open joints and loose blocks may be
observed. Regolith may be divided into saprolite and
transported soils. Colluvium can often be recognized by
distinct morphological features such as fans, lobes and
valley infill. The colluvium may be further sub-divided
depending on topographical position, morphology and
material properties (see Section C.4 of Appendix C).
Areas of boulders should be mapped and if possible
interpreted as saprolite with corestones or as colluvial
boulders. Colluvial boulders may form boulder trails or
fans on the surface of distinct colluvial landforms.

(c) Bedrock Geology - With the assistance of geological maps,

it may be possible to interpret some details of the bedrock
geology. Dykes, faults and lithological contacts can often
be seen as linear features in the topography.
Photo-lineaments should also be recorded.

(d) Landslide Scars - In Hong Kong, landslides occur on natural

terrain at an average frequency of about one in every 2 km2
per year (Evans & King, 1998). It is likely that landslide
scars are present in most study areas. As the past
instability of terrain can provide information for the
assessment of future events, it is critical to record the
location and dimensions of all landslide scars and to
estimate their date of occurrence. The terminology used to
describe landslide scars and the parameters recorded should
be clear and unambiguous. A useful reference for
landslide description is given by Cruden and Varnes (1996).
A set of terms that are adopted for the description of
landslides in Hong Kong is given in Appendix A.

A landslide scar is defined as the land surface affected by a

landslide and includes the source, the displaced material and
any trail. Landslide scars can be divided into recent, which
can be observed as a light toned area bare of vegetation on
aerial photographs, and relict, which are covered with
vegetation and their exact date of occurrence cannot be
established as they occurred prior to the earliest available
aerial photographs. Care should be taken in the
identification of recent landslide scars, particularly on high
 - 31 -

altitude aerial photographs, because some features, such as

gully erosion and grave sites, can easily be confused with
landslides. Parameters such as source dimensions and
run-out extent can be recorded from API.

The identification of relict landslide scars is much more

interpretative. Although it may be possible to estimate
source dimensions, it is seldom possible to record run-out
information with confidence. Vegetated concave features
with colluvial deposits downslope are common in the Hong
Kong landscape and may be interpreted as relict landslide
sources. However, such landscape features could result
from a large single past event, a number of smaller landslide
events, protracted surface erosion, or any combinations of
these processes. As such landscape features are often the
largest that may be interpreted as past landslides, this
interpretation can be crucial to an understanding of the
largest events that may have affected the site.
Consideration should also be given to the likely age of these
landslides and their relevance to the current climatic and site
conditions. Great care should therefore be taken with their
interpretation, and it should always be backed up by
thorough field observations wherever possible.

(e) Landslide Deposits - Locations of landslide deposits such as

colluvial lobes and fans, boulder fields or trails, and scree
slopes should be recorded.

(f) Slope Features - Breaks in slope, localized slope inversions,

ridges, knolls, hollows and other small slope features that
may not be reflected on the contour map should also be
mapped. In particular, attention should be paid to locally
steep slopes that could be the source of landslides, and for
hollows and ridges that could channelize flowing landslide
debris or surface runoff.

Areas of hummocky ground, reverse scarps, tension cracks,

and impeded drainage should also be recorded. These may
offer clues to areas where sliding movements have occurred.

(g) Channel/Stream Features - Where possible, particular

attention should be paid to stream beds that could act as
channels for channelized debris flows. The roughness of
the stream bed, as represented by the presence or absence of
vegetation, bedrock or boulders, should be recorded. Local
steep reaches and waterfalls should be noted. The
steepness and character of the channel banks and their
potential to confine and channelize debris should be
established if possible. A note should be made of any
material in the stream bed and on the banks that could be
 - 32 -

eroded and entrained in a debris flow.

(h) Anthropogenic F eatures - A number of anthropogenic

features may occur on natural terrain in Hong Kong and
they are easy to miss from interpretation of aerial
photographs. They may include walking tracks,
abandoned agricultural terraces, traditional village graves,
military fortifications and overgrown trenches and adits left
from previous small scale mining (Figure 1).

(i) Erosion - Unvegetated areas where sheet or gully/rill

erosion occurs are present in some parts of Hong Kong,
mainly on granite bedrock. Eroded areas, bare of
vegetation, were more common in the past and because
much revegetation has occurred in recent years, the
relevance of historical information should not be discounted.
Where erosion has resulted in steep scarps or undercutting
of boulders, these may form landslide sources and should be
mapped as such, even if the surface has become revegetated.

(j) Vegetation - Vegetation patterns should be mapped and may

provide information on ground or surface water flow, and
past erosion and instability. High lush vegetation cover
may indicate areas of shallow groundwater table or
groundwater discharge; linear strips of low vegetation
oriented down slope may indicate overgrown landslide trails.
As hill fires are common in Hong Kong, care should always
be taken to relate the current vegetation pattern to past
conditions.

3.4 Collection of Field Data

3.4.1 Introduction

Field data collection is an integral part of any NTHS and can be broadly differentiated
into site reconnaissance and detailed field mapping. The field work, anticipated to be
predominantly mapping and site measurements, should be of sufficient detail to validate the
findings of the API and to enable a well reasoned assessment of any natural terrain hazards in
the catchment. Ground investigation, including vegetation clearance for access and
excavation of trial pits, may be required to supplement the field mapping.

3.4.2 Site Reconnaissance

For site reconnaissance, it is essential to find a location where an overview of the site,
and preferably the whole study area, can be gained. This is generally from a hill or tall
building close by. Photographs, in some cases oblique stereo pairs, can be taken to illustrate
site conditions for field mapping in the next stage and for further analysis in the office.
 - 33 -

Following the overview, the whole upslope site boundary of the site should be
traversed, as this is the location where any landslide debris would impinge upon it. Any
materials along the boundary, such as colluvium or boulder deposits, that could be debris
from past landslides should be noted. For known recent landslides, particularly those that
reached the site boundary, an initial inspection to determine the slope conditions, slope
processes and material involved should also be carried out at this stage. Careful observation
of the local topography along the site boundary and consideration of any proposed site
formation levels should also be made. This can be one of the most important aspects in
identifying any possible hazards to a site. If the development is located higher than the
adjacent ground or potential debris transport paths such as streams and channels, then the
hazard from any postulated event is likely to be relatively low. The final proposed site
formation level needs to be considered when making these judgements.

3.4.3 Field Mapping and Site Measurement

Although the API will have determined the range of features associated with any
natural terrain hazards in the study area, these features (as indicated in Table 6) should be
confirmed and quantified further during the field mapping. Additional features as shown in
Table 7 may be recorded. The mapping may be required to extend outside the catchment of
concern in order that the geological and geomorphological setting is fully understood.

The contour interval of 1:1 000-scale topographical maps is 2 m and the topography
shown usually bears a closer resemblance to the ground conditions than that on smaller-scale
maps; therefore, this scale is the most useful for field work where reliable positioning is
essential. However, locating specific features on a broad, vegetated catchment can be
difficult and it may be necessary to place surveyed markers across the study area or to use a
Global Positioning System.

Where access is restricted by dense vegetation, emphasis should be placed on

examining features relevant to the potential magnitude of natural terrain hazards. Specific
features to target for detailed field examination will come from the detailed API of the study
area and its environs. The location of all traverses undertaken during the field mapping
should be marked on the map.

The field work should enable relevant information (Table 7) to be collected on all
accessible landslide scars, drainage lines and hollows, regolith (e.g. rock outcrops, boulders
and boulder fields, soil exposures; see Section C.4 of Appendix C), seepages and stream
flows, open slope and drainage line morphologies, and vegetation types and densities.
Traverses should be made both across the open hillside and along important drainage lines, to
examine the nature and extent of superficial materials, and to determine areas that may be
susceptible to failure during heavy rainfall. Traverses along drainage lines should examine
their characteristics, such as channel width and height, slope gradient, and channel bed
materials. The level of detail should be sufficient to assess the likely travel paths, velocities,
depths and transport distances of any failure debris that enters the drainage line.

Information about past landslides should be collected, including measurements of the

recent and relict source areas, such as their depth and length, and width of the main scarps for
volume estimates. Where possible for recent landslides, materials displaced from the source
areas and their degree of disaggregation should be described. In addition, characteristics of
 - 34 -

debris trails should be determined, including their length, width and processes of transport
and deposition, and related erosion.

Other significant observations that should be made, if possible, to determine the cause
and mechanisms of recent landslides and their implications for future landslide hazards.
These may include:

(a) super-elevation of debris trails, based on vegetation

flattening, erosion scars, marks on trees, and/or
characteristics of materials deposited in levees (including
thickness and volume of materials) for estimating velocity
of debris flows (Johnson & Rodine, 1984),

(b) Evidence of scour within stream beds or scars on channel

sides to determine whether entrainment of additional debris
from within the debris trail has occurred, and

(c) the location and character of tension cracks above the main
scarp to establish the possibility of future landslide hazards.

An examination of rock mass characteristics should be made at all accessible bedrock

outcrops. Where pertinent, rock type, material weathering grade, measurements of joint
attitude and spacing, geological features that could affect stability (e.g. bedding, eutaxitic
fabric, clay layers, metamorphic texture), and seepage locations and flow rates should be
recorded. The character of the weathering profile in stream sections, landslide scarps or rock
outcrops should also be recorded, as should descriptions of colluvial materials if present.

Where boulder falls are considered to be a hazard to the development, mapping of

boulder fields should be at sufficient detail to allow an assessment of the locations and sizes
of all boulders that are considered to pose a significant risk to the site under consideration.
Potentially unstable boulders should be identified on site, particularly if in-situ stabilization is
to be adopted for hazard mitigation. The likely travel path of such boulders in the event of
failure should, where possible, also be assessed in the field. The field work should also
attempt to distinguish between rock outcrops and individual corestones/boulders. In order to
estimate the likelihood of boulder fall initiation and subsequent movement, information on
boulder type, boulder shape, boulder volume, boulder location (travel distance to development
site/boundary), local slope angle, embedding material, slope angle, and vegetation density
(Maunsell & ERM, 1999) are useful.

3.5 Geological and Geomorphological Model

Based on the information collected and interpreted, a geological and geomorphological

model of current and past processes in the study area should be developed. This will form a
basis for assessing the likely sources, nature and volume of any instability that could affect
the site. It is beneficial to look beyond the catchment boundary so as to observe processes
and events in adjacent or similar catchments that may be relevant for determining possible
hazards to the site.
 - 35 -

The geological and geomorphological model should be specifically designed for each
assessment and will combine the API with the engineering geological mapping. Historical
landslide data often provide useful information for development and calibration of the model.
Fookes (1997) notes that The strength of the geological model is in providing an
understanding of the geological processes which made the site. This enables predictions to
be made or situations anticipated for which explorations need to be sought in the geological
materials, geological structure and the ancient and active geological processes in the area.

Mass movements have displaced much of the weathered mantle on hillslopes, forming
extensive colluvial deposits on the shallower slopes, and large debris sheets and lobes on the
steeper slopes and in the steep tributary valleys. Recent scars can be identified either in the
field or on aerial photographs, a record that covers most of the territory from 1945 to the
present. On the other end of the spectrum, large relict scars, which have a degraded
morphology, probably extend the historical record to thousands of years.

In the context of an NTHS, the geological/geomorphological model is useful for

assessing the types of landslide hazard that may occur, the likelihood of their occurrence, and
the possible size of failures.

3.6 Design Event

3.6.1 Overview

A design event refers to the magnitude of the hazard model that is selected to be
mitigated against as part of a development. As such, it requires information about the
proposed layout of the site. If the Design Event approach is adopted, the relevant design
requirements (i.e. worst credible event or conservative event) set out in Section 2.4 should be
followed. If the QRA approach is adopted, the design event is normally determined by
reference to risk guidelines (ERM, 1998). In both cases, the determination of the design
event requires professional judgement.

The geological/geomorphological model and design requirements (i.e. conservative or

worst credible event) for individual identified hazards obtained from the technical framework
(Section 2.4), and subsequently refined with field information from the NTHS (Section 3.4),
should be used to determine the design event for each hazard model in each sub-catchment.
The design event needs to be defined in terms of the volume, velocity (or impact load) and
nature of the landslide debris (or boulder) at the site boundary, or any other upslope location
where construction of mitigation measures might be proposed.

3.6.2 Determining the Design Event

The design event can be derived in several ways, largely dependent on the nature of
the natural terrain hazards. For the more frequent, shallow failures it may suffice to
determine the design event based on historical data whereas for the infrequent, large events it
requires the development of an appropriate geological and geomorphological model as well as
the exercise of professional judgement. Some typical means of assessing the design events
are given as follows:
 - 36 -

(a) Open hillslope landslide - Most natural terrain landslides in

Hong Kong are relatively shallow open hillslope failures.
They are well represented by historical data, such as from
landslide records, aerial photograph interpretation and field
mapping. As a result, historical data usually provide a
good basis for assessing the frequency, mechanism and
scale of such landslides.

Given the relatively large amount of data available,

development of site-specific landslide susceptibility models
may be practical in some cases (e.g. Parry, 2001).
Furthermore, the historical landslide data may also be used
to construct the catchment-specific frequency-magnitude
relationships in selecting the source volume for a design
event. Section 3.6.3 summarizes this approach that has
been adopted for a number of cases in Hong Kong.

(b) Deep-seated s lide or large scale (area extent) shallow

failure - Such landslides are less frequent, and may not be
well represented by historical landslides at the site. Hence,
correlation by reference to historical landslide data could be
of limited use. Consideration of the geological and
geomorphological model, as well as any geomorphological
evidence of relic events, is essential to assessing the
likelihood of occurrence of such failures. Historical
landslides in the vicinity of the site, and in other places with
a similar geological/geomorphological setting, could also
give insight to evaluating whether factors contributing to
development of such failures are present at the site. For
example, well documented large failures include the Tsing
Shan debris flow involving colluvium infilled gullies on
steep terrain (King, 1996a) and the quasi-natural terrain
Shum Wan Road landslide in thick weathering profile
related to fault and weak clay seams (GEO, 1996). The
designer should be aware of the geological environments in
which these events have occurred, and make a comparison
with that of the study area.

Given the relatively sparse data for this type of infrequent,

large magnitude events, a catchment-specific cumulative
frequency-landslide volume relationship is likely to be
poorly defined. An attempt to circumvent this limitation
using landslide data from similar geological settings to
extrapolate a catchment-specific frequency-volume
relationship (Section 3.6.4; see also Atkins, 1998; Lo, 2000).
was described in Tse et al. (1999). However, it is essential
that the extrapolation is compatible with the geological and
geomorphological setting of the catchment concerned.
 - 37 -

(c) Channelized debris flow - This type of failure has several

distinctive characteristics that deserve special consideration
(Section 3.6.6). Assessment of the magnitude of the
design event involves consideration of the possibility of
multiple landslide sources, entrainment along travel path,
and development of dam break scenario.

Like deep-seated slides, determination of the design event

for a channelized debris flow requires the establishment of
an appropriate geological and geomorphological model, and
the exercise of professional judgement. For example, API
and field mapping can identify superficial deposits along
drainage lines and hence provide useful indications of
probable entrainment material and its likely volume.

3.6.3 Catchment-specific Landslide Data

This approach makes the assumption that past events in the study area and its relevant
vicinity give an indication of the potential for future events. The approach has been used in
the assessment for the Foothills Bypass, Tuen Mun by Scott Wilson Kirkpatrick (1996), Tsing
Yi North Coastal Road by Lam (1999), Sham Tseng San Tsuen by Maunsell (2000) and Shek
Pik School and Hostel by Parry (2001). The first case was note-worthy for the presence in
the study area of the largest recent debris flow recorded in Hong Kong (King, 1996a). It was
taken to be a reasonable representation of the more extreme events that could occur in the
study area.

Records of the date and source dimensions of past landslides in the study area can be
used to construct local frequency-magnitude relationships that are useful in selecting an
appropriate source volume for a design event. Such catchment-specific data should also be
considered in relation to the rainfall history of the study area. A basic problem of using
historical landslide records to help assess future hazards is that catchment-specific data may
not be entirely representative of all events that could occur at that location. The smaller the
catchment area and the more limited the available geological information is, the greater the
uncertainty becomes. Catchments with a past landslide history are likely to reflect better the
future occurrence of landsliding although even these may not include evidence of more
extreme events. Therefore, it may be useful to consider data from wider regional studies.
However, care needs to be taken to ensure that only data with a similar geological/
geomorphological setting to that being assessed is used.

3.6.4 Regional Landslide Data

Regional landslide data can be used to check that a design event estimated from a
catchment-specific geological model is reasonable in the context of a wider population of past
events from similar events and similar environments. Regional data may also be the only
logical basis for determining a design event where catchment-specific landslide records and
data are insufficient or not available. It may be possible to extend relationships established
from catchment-specific data by incorporating regional data. Where the two sets of data
 - 38 -

give comparable results, this would lend confidence to the determined design event.

Sources of regional data on landslides in Hong Kong include the NTLI, which is a
database of some 27 000 landslides identified on high altitude aerial photographs (Emery,
1996; King, 1999) and the Large Landslide Dataset which contains nearly 2 000 features
identified on low altitude aerial photographs (Scott Wilson, 1999b, 1999c). Compilations of
field data from landslide area studies are reported in Franks (1998) and Wong et al. (1998).
An overview of some landslides in Hong Kong and elsewhere, and their use in selecting a
design event, is given in Atkins (1999b). A number of individual landslide case studies have
also been carried out for natural terrain landslides in Hong Kong (see Appendix B). Again,
care should be taken to ensure that the datasets used are from similar geological/
geomorphological settings.

3.6.5 International Data

Review of data and methods used in other countries can also be a useful aid when
determining a design event. For example, methods of quantitative analysis of debris flows
used in Canada are reported in Hungr et al. (1984) and VanDine (1985). These are further
developed in work carried out at the Cheekye Fan in British Columbia, described in Hungr
and Rawlings (1995) and Sobkowicz et al. (1995). The relationship between angle of reach
and landslide volume is presented for worldwide data by Corominas (1996).

Some methods for predicting design event parameters in Hong Kong and elsewhere are
given in the Proceedings of the Seminar on Planning, Design and Implementation of Debris
Flow and Rockfall Hazards Mitigation Measures (AGS & HKIE, 1998).

3.6.6 Special Consideration for Channelized Debris Flows

If the study has determined that the catchment contains drainage lines that have the
characteristics to channelize debris flows, then the possibility of multiple landslide sources,
the potential entrainment of eroded debris along the travel path, and the probable development
of a dam-break scenario need to be considered. These phenomena can result in a significant
increase in the total volume of the landslide debris and/or a significant increase in the debris
mobility.

For drainage lines that are considered susceptible to channelize debris flows, the debris
yield rate (volume of eroded material per linear metre) of each drainage line should be
determined. This will typically be carried out by reference to the thickness, characteristics
and distribution of the materials present within the sections of the drainage line susceptible to
erosion and entrainment of debris. Debris yield rate is affected by the characteristics of the
stream course. Even for the same stream course, debris yield rate can be affected by the
volume and mechanism of the source failure.

In Hong Kong, debris yield rates for North Lantau were estimated to be 2.3 m3/m to
3.7 m /m, for landslide source volumes of about 450 m3 to 1 500 m3 and where the average
3

stream bed gradients were in the range of 20 to 30 (Franks, 1998). Wong et al. (1998)
estimated that the total debris volume for some landslides at South Lantau could be three
 - 39 -

times the source volume. King (1996a) reported a debris yield rate of up to 24 m3/m for the
Tsing Shan debris flow in 1990, with a source volume of about 3 000 m3 and mean stream bed
gradients of 22. For those sections of a drainage line that have exposed bedrock in the
stream bed, debris yield rates are likely to be much lower. For example, the debris yield rate
of the Sham Tseng San Tsuen debris flow was estimated to be negligible (Fugro Maunsell
Scott Wilson, 2000a).

Halcrow (2001) documented the formation of a debris dam along a drainage line, some
30 m below the source area of a landslide above Leung King Estate, Tuen Mun. The dam,
represents the initial debris lobe resulting from the failure, had a travel angle of about 27.
Subsequently breaching of the dam, presumably due to damming of water behind the debris lobe,
resulted in the formation of a wet channelized debris flow with a travel angle of about 17. This
case demonstrates that debris mobility can be significantly increased in a dam break scenario.

3.7 Run-out Assessment

Where the probable sources for future landsliding have been identified, the potential
run-out of landslide debris from them should be estimated in respect of the required design
event, i.e. a worst credible event or a conservative event. Determination of the debris travel
distance is an important component in assessing if the hazards may affect the site.

A reliable run-out assessment requires information on the likely failure volume,

characteristics of the travel path and the momentum or energy of the moving debris which is
related to the velocity and density of the debris. It should be assessed either with respect to
the site boundary, or other suitable reference point critical to the consideration or mitigation
of the hazards (e.g. check dam or boulder fence).

Catchment-specific data on the extent of debris trails of past landslides can be used if
available. However, it is often difficult to map the extent of past landslide debris from aerial
photographs and such debris may be difficult to recognize on the ground after several years.
Colluvial deposits provide some guidance as to the possible reach of past landslide debris but
the possibility that they were deposited under a different climatic and geomorphological
regime than today must be considered.

Mobility of landslide debris, in terms of the distal point of debris deposition, was
discussed by Lau and Woods (1997). Lo (2000) presented a comprehensive review of
various approaches for debris run-out assessment and suggestions for assessing debris
mobility and debris impact loads in the design of landslide debris-resisting barriers.

3.8 Presentation of Results

The NTHS should establish the following:

(a) Geological/geomorphological model for the area with

particular reference to existing and potential landslide
locations.
 - 40 -

(b) Hazard models and their r espective design requirements

based on the detailed study.

(c) The location, magnitude, and run-out distance of each

specific hazard model that may affect the site or other
relevant facilities (e.g. check dam).

(d) Whether measures are required to mitigate against the

specific hazards that may affect the site.

In the report, the study area should be described in the context of its regional setting
with specific regard to the current and/or proposed development of the site and its topography,
geomorphology, geology, drainage and rainfall characteristics. The limitations of the study
(e.g. access problems, limited aerial photograph coverage) should be stated. The report
should include the scope and methodology adopted and assumptions used to establish items (a)
to (d) above. For cases where mitigation measures are required, sufficient detail for
formulation of mitigation strategy (Section 4) should be given.

Adequate photographs, tables, figures and drawings, including cross-sections, should

be used. The following illustrations are suggested as part of the report:

(a) Cross-sections along each potential debris travel path from

identified natural terrain hazard source locations to the site
boundary.

(b) An Engineering Geology Map of the study area at an

appropriate scale (e.g. 1:1 000) showing:

(i) the solid geology and regolith,

(ii) geomorphological setting,

(iii) the locations of exposures of soil and rock, annotated

where necessary with key information about the nature
of the materials (e.g. stereonet analyses),

(iv) sites of past instability (e.g. landslide scars, boulder fall

locations), annotated with key information about the
nature of the instability,

(v) catchment b oundary and drainage courses with

comments regarding any areas of concentrated surface
water flow or seepage, and

(vi) the location of field traverses.

(c) A Natural Terrain Hazard Map of the study area at an

appropriate scale (e.g. 1:1 000) showing:
 - 41 -

(i) potential landslide sources, for open hillslope landslide,

deep-seated slide or channelized debris flow, annotated
for easy reference in the report,

(ii) potential boulder fall and/or rock fall locations,

(iii) open hillslope landslide paths from identified potential

landslide sources with the likely distance that debris will
travel,

(iv) channelized debris flow paths with their likely

maximum extent, and

(v) boulder/rock fall trajectories with the estimated distance

that boulders and rock blocks will travel.

Where mitigation measures are required, the report should clearly present the rationale
for the mitigation strategy formulated to reduce the risk to the development. The general
types and layout of the measures should be shown on a Mitigation Measures Strategy Map at
an appropriate scale (e.g. 1:1 000 or 1:5 000). If in-situ stabilization measures at the source
area are recommended, their locations should be marked on the map. This is particularly
applicable to boulder and rock falls hazards where the exact location of the hazards should be
indicated on a map of 1:1 000 scale or better. The report should recommend a programme,
including the scope and frequency for maintenance inspections, and should list the parties
responsible for carrying out the inspections and maintenance works.

3.9 Stages of the Natural Terrain Hazard Study

3.9.1 Introduction

For large scale projects, such as site formation and highway projects (GEO, 2000a), it
may be beneficial to divide the study into sequential stages (review, assessment and
mitigation strategy), in increasing details, that meet the requirements for different phases of
the projects from land use planning through engineering feasibility study to design stage.
However, for projects that are relatively straightforward, a staged approach is unlikely to be
necessary. Should project proponents opt to undertake an NTHS in stages, consideration
may be given to splitting the study into stages (Figure 3) as described below.

3.9.2 Review

This first level of study is a desk and reconnaissance level study to identify in broad
terms, any natural terrain hazards that might affect the proposed development. The review is
to decide whether further assessment will be necessary, and also to identify whether the
hazards are so significant that the site is deemed uneconomic to develop. If the review
identifies significant hazards, the study should progress to the next stage.

This process is generally carried out at the earliest possible stage of a project, such as
Building Plan stage for private developments, or as part of the feasibility study or preliminary
 - 42 -

design for public works projects. Consideration should also be given at this stage of the
broad costs that may be involved for future financial estimates.

A natural terrain hazard review should include the following aspects:

(a) desk study of available information (Section 3.2),

(b) detailed interpretation of all available aerial photographs

(Section 3.3), and

(c) site reconnaissance (Section 3.4.2).

At the end of the review stage, the study should conclude the following:

(a) Whether further study is required or whether the proposed

development is considered not feasible.

(b) If further study is required, the recommended design

approach and the design requirements for each type of
hazard models for each sub-catchment above the proposed
site should be provided.

If the Design Event approach is to be followed, the design requirements (i.e.

conservative or worst credible event) should be assessed in accordance with the procedures
given in Section 2.4. The indication of the design requirements provides information about
the probable scale of the natural terrain hazards posed to the proposed development and
relative to the site boundary (Section 2.2). This would give the designer the opportunity to
consider reducing the hazards through adjusting the internal layout of the critical facilities at
an early stage of the development.

The results of the review should be presented in a report that summarises all the
geological and geotechnical information about the study area and, in particular, the results of
the aerial photograph interpretation. Sufficient text should be given to explain the reasoning
and the methodology adopted for the review and to justify the conclusions made. The report
should be supported by relevant photographs, an aerial photograph interpretation map using
1:1 000 or 1:5 000-scale topographical base maps, and representative cross-sections.

3.9.3 Assessment

The assessment requires a detailed field study of the natural terrain, sufficient to
determine the location, type and magnitude of any hazards. Quantifying the design event
from the catchment above the site can be carried out when the detailed design and layout of
the development have been decided. The assessment will establish whether the risk
necessitates mitigation measures.

This process is generally carried out at the design stage for both private and public
works projects.
 - 43 -

A natural terrain hazard assessment should include the following aspects:

(a) all items covered in a natural terrain hazard review

(Section 3.9.2), and

(b) detailed field mapping (Section 3.4.3).

At the end of the assessment stage, the study report should cover all the items
(excepting the mitigation strategy) given in Section 3.8 above.

3.9.4 Mitigation Strategy

The final stage, which is to formulate a suitable mitigation strategy (Section 4), may be
needed once the assessment stage is complete. The NTHS may not proceed through to this
stage because as more data are obtained about the site, it may become possible to establish
that mitigation measures are unnecessary.

This process is generally carried out at the design stage for both private and public
works projects.

At the end of the mitigation strategy stage, the study report should provide the
mitigation strategy formulated for reducing the risk to the development, with all the
supporting rationales.

4. NATURAL TERRAIN HAZARD MITIGATION STRATEGY

4.1 Introduction

Upon determination of the design event, the risk to the development can be assessed
and an appropriate mitigation strategy formulated if the risk has to be reduced. Mitigation
measures can either be passive, for example by re-locating a development site beyond the
area considered at risk and so avoiding the hazard, or active, involving the construction of
measures to reduce the effect of the anticipated hazard and therefore the risk to the
development. Active mitigation measures can be further subdivided into preventive and
protective. The former may include in-situ stabilization (at source areas) and the latter
commonly involves defensive mitigation works in the travel path or within the site to retard,
store or deflect failure debris.

4.2 Typical Mitigation Strategies

4.2.1 Review

Franks and Woods (1997), Greg Wong & Associates (1999) and Lo (2000) have
reviewed typical mitigation measures and their applicability to Hong Kong conditions.
 - 44 -

4.2.2 Passive Mitigation Strategies

Passive mitigation strategies involve precluding development in areas that may be

affected by severe natural terrain hazards. This may be accomplished by land use planning,
statutory regulations and development moratoria. The Mid-Levels construction moratorium
was introduced in 1973 to safeguard the stability of slopes in the district. In some cases,
mitigation can be achieved by closure or relocation of facilities at risk, provision of suitable
buffer areas on the upslope side of the development site or adjacent to drainage lines, or
erection of warning signs. The option of adopting passive mitigation measures should be
considered at an early stage of the project so as to ensure that any necessary arrangements can
be accommodated.

4.2.3 Active Mitigation Strategies

Active mitigation strategies commonly include stabilization in the source area, and
energy dissipation, deflection and containment measures along the travel path, or in the
deposition area. The following active mitigation strategies that may be applicable for the
main hazard models in Hong Kong are extracted primarily from Franks and Woods (1997):

(a) Open Hillslope Landslide

(i) Source - afforestation, surface drainage, surface

protection, shallow subsurface drainage, slope regrading
and soil nailing.

(ii) Deposition - energy dissipators, deflectors and

containment structures to induce deposition of debris
away from the elements at risk.

(b) Deep-seated Slide

(i) Source - afforestation, surface drainage, surface

protection, subsurface drainage, buttresses, slope
regrading and soil nailing.

(ii) Deposition - not relevant for this hazard model.

(c) Channelized Debris Flows

(i) Source - as for open hillslope landslides, and also

including channel works, such as revetment works and
spur dykes, to prevent production of debris due to
erosion of the stream bed and banks, and slope
stabilization measures.

(ii) Deposition - dykes, check dams, deflection walls,

sacrificial baffles and containment basins built along the
path and in the depositional area to dissipate energy and
 - 45 -

allow containment of debris.

(d) Boulder and Rock Falls

(i) Source - excavations, shotcrete, rock bolts, dowels,

anchors, buttresses and wire catch nets to stabilize rock
fragments.

(ii) Deposition - catch fences, deflection walls and catch

ditches along the path and in the deposition area to
mitigate the effects of boulder and rock falls.

4.3 Hong Kong Practice

Lo (2000) provides a review of the types of debris- and boulder-resisting barriers

deployed in Hong Kong. These include rock/boulder fences, gabions, reinforced concrete
retaining walls, earthfill berms and check dams. Boulder fences, gabions and concrete walls
have been used to arrest loose boulders; reinforced concrete retaining walls have been
employed as debris-resisting barriers. Some of these examples are shown in Figure 4.

Besides the typical debris-resisting structures, some alternative measures that have
been successfully implemented are given below for reference:

(a) Adjusting facility layout - This can be an effective action for

a large site. It is generally possible to avoid or reduce the
exposure to natural terrain hazards by siting the critical
facilities (e.g. residential blocks) away from the hillside.

(b) Making use of natural barrier - During site formation

planning stage, the possibility of making use of natural
features (e.g. existing ridge) as a barrier for landslide debris
should be considered.

(c) Delineating non-building area - In some cases, it may be

possible to delineate the most hazardous part of the site as
no-build zone.

(d) Drainage p rovision - For sites intersected by natural

drainage courses that may act as a flow path for channelized
debris flow, it may be possible to deal with the hazard by
having adequate drainage provisions such as debris basins
and boulder straining structures to intercept, retain and/or
convey the debris away form the site.

(e) Elevated platform - For a narrow site, there may be little

room to accommodate debris barriers. A possible option is
to have the superstructure built on an elevated platform so
as to allow debris to pass underneath the building.
 - 46 -

(f) Integrated structure - In some cases, it may be possible to

incorporate debris-resisting wall into the design of the
permanent structure. This is particularly appropriate for
development sited very close to hillside.

(g) Potentially Hazardous Installation (PHI) - For erection of

PHIs (e.g. petrol filling station), a QRA is normally required.
Natural terrain landslide hazard risk, being one of the risks
that the facility would be exposed to, should be taken into
account as part of the QRA. This eliminates duplicating
effort in undertaking a separate natural terrain hazard study.

4.4 Environmental Considerations

4.4.1 Introduction

The majority of natural terrain landslides occur in areas where humankind has had
little or no influence. As a result both mitigation and remedial works as applied to natural
terrain should consider landscape aesthetics in that the measures should blend with their
surroundings as far as practicable. Although both Works Bureau Technical Circular (WBTC)
No. 25/93 and 17/2000 specifically apply to the design and maintenance of man-made slopes,
their principle of minimizing visual impact is also relevant to natural terrain. Emergency
landslide repairs are however exempt from both circulars.

For any mitigation works carried out within the Country Parks or in environmentally
sensitive areas, or for any works classified as designated projects under the Environmental
Impact Assessment (EIA) Ordinance, the potential environmental impacts of these works
should be assessed. If found necessary, appropriate monitoring and/or measures should be
implemented to ensure that the impacts to the environment are kept to an acceptable level.

4.4.2 Bio-engineering

Bio-engineering has been defined in Hong Kong (GEO, 2000b) as the use of living
vegetation, either alone or in conjunction with non-living plant material and civil engineering
structures, to stabilize slopes and/or reduce erosion; and the use of any form of vegetation,
whether a single plant or collection of plants, as an engineering material (i.e. one that has
quantifiable characteristics and behaviour). In Hong Kong, only simple bio-engineering
techniques in the form of vegetation cover have been used. However, more sophisticated
techniques have been used in other countries and information on these techniques is contained
in GEO (2000b).

Natural terrain landslides occur as part of the on-gong, natural processes of erosion and
landform evolution. Therefore, aesthetic reasons alone may not be sufficient to qualify a
natural terrain landslide scar for revegetation. However, where landslide reactivation is
likely, or where the landslide debris may be remobilized, or deterioration of the conditions of
the hillsides needs to be minimized, revegetation may form an integral, cost-effective part of
any mitigation proposals. Bio-engineering may be adopted in reparation of hillsides affected
by landslides or hill fires so as to control surface erosion and soil loss, and to prevent further
 - 47 -

deterioration of the hillsides. There is also scope for application of bio-engineering as

mitigation measures to reduce the consequence of natural terrain landslides by partial
retention of coarse debris or to reduce entrainment by stabilizing loose deposits

Most natural terrain landslides in Hong Kong are relatively shallow and commonly
occur along geological/hydrogeological boundaries such as the colluvium/saprolite interface.
The topsoil and other suitable growth media are commonly removed and natural revegetation
can be slow. Furthermore, natural terrain landslides can take many forms and hence they
result in a wide range of soil damage. Therefore, each location needs to be carefully
analysed to determine the objective of the bio-engineering application and the most
appropriate techniques for that site.

Experience gained from application of vegetation cover to cut slopes in Hong Kong
may be relevant to natural terrain. Characteristics of various plant species commonly used
in Hong Kong as surface protection are summarized in the Geotechnical Manual for Slopes
(GCO, 1984). The General Specification for Civil Engineering Works (Hong Kong
Government, 1992) provides specifications and guidance on landscaping works and
establishment. Wong et al. (1999) noted that it is essential to provide both erosion
protection mats and a biodegradable protection fabric to reduce erosion and protect the seeds
while the vegetation establishes. Requirements for plant growth, typical plant mixes and
prescriptive and potential landscape treatments for soil cut slopes are given in GEO (2000b).
The recent use of landscaping and bio-engineering of cut slopes is summarized in Martin
(2001).

As natural hillside is generally remote, consideration should be given to future access

requirements for maintenance of the mitigation works. If planting of vegetation is to be used,
special consideration should be given to use of watering and maintenance free plant species.

While there is little experience in Hong Kong on revegetating landslide scars on

natural terrain, there is considerable experience in other places (e.g. Ministry of Forests,
1997). It may be possible to adapt some of their practice with suitable modification to cater
for the Hong Kong climate and the natural terrain conditions.

4.4.3 Visual Impact of Mitigation Measures

In addition to bio-engineering approach, the visual impacts of any hard mitigation

measures, such as deflection barriers, should be considered. Where possible visual intrusion
should be reduced, for example, through plantation of vegetation screens to hide the facilities
or blending of the measures with the natural environment.

5. MAINTENANCE OF MITIGATION FACILITIES

5.1 Introduction

Maintenance requirements for the mitigation measures should be proposed by the

designers. Besides the maintenance schedule required for the works, preliminary estimates
of the related capital and recurrent costs should be prepared where necessary. The proposed
maintenance requirements follow the principle that the maintenance agents work is confined
 - 48 -

to maintaining the physical integrity and continued functionality of the Measures. The
maintenance agent is neither required to maintain the natural hillside, nor to review the
adequacy of the Measures provided. However, special cases that warrant more stringent
maintenance requirements would be treated separately such as by imposing Green-hatched-
black geotechnical clause.

5.2 Types of Natural Terrain Hazard Mitigation Measures

Natural Terrain Hazard Mitigation Measures are classified into two categories for
registration in the Slope Catalogue. They are:

(a) Stabilization Measures constructed on the natural hillside to

prevent failure, e.g. boulder buttresses, soil nails, raking
drains and retaining walls.

(b) Defense Measures to contain landslide debris or boulder fall

from the hillside above, e.g. check-dams, earth bunds and
boulder fences.

5.3 Maintenance Requirements for Mitigation Measures

Natural hillsides do not require maintenance. Natural Terrain Hazard Mitigation

Measures do not normally result in substantial modification to the geometry and conditions of
the natural hillside. Where this is the case, the purpose of maintenance is confined to
ensuring the physical integrity and continued satisfactory function of the Measures. Where
stabilization measures have been adopted and the hillside has not been substantially modified
by the Measures, the hillside concerned does not require maintenance. Likewise, defense
Measures are designed to contain landslide debris or boulder falls, and the hillside where
landslides or boulder falls may occur does not require maintenance.

The general principles given in Geoguide 5 (GEO, 1998), such as the recommended
good practice in maintenance management and personnel requirements, and attention to
access and safety precautions, are applicable to maintenance of Natural Terrain Hazard
Mitigation Measures. However, since the Measures are of different nature from man-made
slope features, their technical requirements for maintenance differ and these are set out below.

Routine Maintenance Inspections on stabilization measures and defence measures

should be carried out annually by maintenance staff at the rank of Assistant Clerk of Works,
Technical Officer or Work Supervisor, to identify any maintenance works required. The
inspection should cover the Measures, and the area containing the Measures and the adjoining
ground. The inspection should be completed well before the onset of the wet season. This
would allow sufficient time to carry out the necessary routine maintenance works, which may
typically include:

(a) clearance of debris from drainage channels, catch trenches

and pits, containment basins and straining structures,
 - 49 -

(b) repair or replacement of damaged sections,

(c) unblocking of weepholes and outlet drain pipes,

(d) removal of any vegetation causing severe cracking,

(e) repair or reinstate the ground adjoining the Measures if

affected by severe erosion, and

(f) other routine maintenance works to upkeep the integrity and

continued function of the Measures.

In cases where unusual conditions or problems are observed, e.g. check dam filled with
a large amount of landslide debris or significant movement observed at boulders supported by
buttresses, advice should be sought from a professional geotechnical engineer.

In line with the recommendations of GEO Report No. 56 (Wong et al., 1999) raking
drains used as prescriptive measures should not be considered as Special Measures as
defined by Geoguide 5. Regular inspections and routine maintenance of the raking drains
should however be carried out.

Unless specified otherwise by the designer or by special provisions (e.g. delineation of

Green-hatched-black area), Engineer Inspections for Maintenance are not required for
stabilization measures and defence measures in view of the following:

(a) Unlike in the case of maintenance of man-made slope

features, the natural hillside itself does not require
maintenance. Hence, the guidelines set out in Geoguide 5
on Engineer Inspections for man-made slope features are
not applicable to natural hillside, and are outside the scope
of work required of the maintenance agent.

(b) The maintenance works required for stabilization measures

and defence measures are relatively simple, and do not
normally require input from a professional geotechnical
engineer.

If stabilization measures were adopted and the hillside was substantially modified by
the Measures (e.g. major regrading), then the hillside has been turned into a man-made slope
feature. In such case, it should be registered as a man-made slope feature and maintained in
accordance with the guidelines given in Geoguide 5.

5.4 Other Measures

In some circumstances, dealing with natural terrain landslide hazards may involve the
use of other mitigation measures that do not require registration in the Slope Catalogue.
Examples include:
 - 50 -

(a) provision of a buffer zone (e.g. an open space) between the

hillside and developed facilities, and

(b) incorporation of debris basins, sand traps, etc. as part of the

drainage facilities.

Unless specified otherwise by the designer, there are no maintenance requirements for
such measures from the geotechnical point of view, apart from regular clearance of debris.

Re-vegetation may be carried out in a prescriptive manner to repair hillsides that are
affected by landslides, hill fires, etc. The vegetation species to be adopted in such
circumstances should be maintenance free. Where special bio-engineering measures are
adopted for mitigation of natural terrain hazards, the maintenance requirements should be
specified by the designer.

6. REFERENCES

AGS & HKIE (1998). Proceedings of the One Day Seminar on Planning, Design and
Implementation of Debris Flow and Rockfall Hazards Mitigation Measures. The
Association of Geotechnical Specialists (Hong Kong) Limited and The Hong Kong
Institution of Engineers (Geotechnical Division), 27 October 1998, Hong Kong.

Anon. (1972). The Preparation of Maps and Plans in terms of Engineering Geology: Report
by the Geological Society Engineering Group Working Party. Quarterly Journal of
Engineering Geology, Vol. 5, pp 295-382.

Atkins China Ltd. (1998). Geotechnical Assessment Fanling Area 49A Phases 1 & 3 -
Natural Slope, Vol. 1. Report to Hong Kong Housing Authority, Hong Kong, 88 p.

Atkins China Ltd. (1999a). Geotechnical Assessment for Natural Slope Adjoining Tsing
Yan Temporary Housing Area, Tsing Yi. Report to Housing Department, Hong
Kong.

Atkins China Ltd. (1999b). Scoping Study for a Global Quantitative Risk Assessment of
Natural Terrain Landslides in Hong Kong. Report to Geotechnical Engineering
Office, Hong Kong, 51 p.

Atkins Haswell (1995). Quantitative Landslide Risk Assessment for the Squatter Villages in
Lei Yue Mun. Report to Geotechnical Engineering Office, Hong Kong, 37 p. plus
4 drgs.

Au, S.W.C. & Chan, C.F. (1991). Boulder treatment in Hong Kong. Selected Topics in
Geotechnical Engineering - Lumb Volume, (Edited: Li, K.S.), pp 39-171.

Brand, E.W. (1996). Bibliography on the Geology and Geotechnical Engineering of Hong
Kong to March 1996. GEO Report No. 50, Geotechnical Engineering Office, Hong
Kong, 111 p.
 - 51 -

Corominas, J. (1996). The angle of reach as a mobility index for small and large landslides.
Canadian Geotechnical Journal, Vol. 33, pp 260-271.

Cruden, D.M. & Varnes, D.J. (1996). Landslide Types and Processes. In: A.K. Turner &
R.L. Schuster, Landslides: Investigation and Mitigation. Transportation Research
Board Special Report No. 247, National Research Council, National Academy Press,
Washington, D.C., pp 36-75.

Drainage Services Department (1994). Stormwater Drainage Manual: Planning, Design and
Management. Drainage Services Department, Hong Kong, 145 p.

Emery, K.A. (1996). The Natural Terrain Landslide Inventory. Department of Land and
Water Conservation of New South Wales. Report to Geotechnical Engineering
Office, Hong Kong, 20 p. plus 2 Appendices.

Emery, K.A. (1998). Boulder Study of Hong Kong - Final Report. Maunsell Geotechnical
Services Ltd. Report to the Geotechnical Engineering Office, Hong Kong, 12 p.

Evans, N.C., Huang, S.W. & King, J.P. (1999). Natural Terrain Landslide Study Phases I
& II. GEO Report No. 73, Geotechnical Engineering Office, Hong Kong, 128 p. plus
2 drgs.

Evans, N.C. & King, J.P. (1998). The Natural Terrain Landslide Study - Debris Avalanche
Susceptibility. Technical Note TN 1/98, Geotechnical Engineering Office, Hong
Kong, 96 p.

ERM-Hong Kong Ltd. (1998). Landslides and Boulder Falls from Natural Terrain: Interim
Risk Guidelines. GEO Report No. 75, Geotechnical Engineering Office, Hong Kong,
183 p.

Fookes, P. G. (1997). Geology for Engineers: the Geological Model, Prediction and
Performance. Quarterly Journal of Engineering Geology, Vol. 30, pp 293-431.

Franks, C.A.M. (1991). Engineering Geology of North Lantau: Tsing Chau Tsai Peninsula
to Ta Pang Po. Special Project Report SPR 4/91, Geotechnical Engineering Office,
Hong Kong, 148 p. plus 9 drgs.

Franks, C.A.M. (1992). Engineering Geology of North Lantau: Ta Pang Po to Tai Ho Wan.
Special Project Report SPR 2/92, Geotechnical Engineering Office, Hong Kong, 109 p.
plus 5 drgs.

Franks, C.A.M. (1998). Study of Rainfall Induced Landslides on Natural Slopes in the
Vicinity of Tung Chung New Town, Lantau Island. GEO Report No. 57,
Geotechnical Engineering Office, Hong Kong, 102 p. plus 3 drgs.

Franks, C.A.M. & Woods, N.W. (1997). The Natural Terrain Landslide Study - Preliminary
Review of Natural Terrain Landslide Hazard Mitigation Measures. Technical Note
TN 8/97, Geotechnical Engineering Office, Hong Kong, 52 p.
 - 52 -

Fugro Maunsell Scott Wilson Joint Venture (2000a). Report on the Debris Flow at Sham
Tseng San Tsuen of 23 August 1999 - Findings of the Investigation. Geotechnical
Engineering Office, Hong Kong, 92 p.

Fugro Maunsell Scott Wilson Joint Venture (2000b). Detailed Study of the Landslide above
Kap Lung, Shek Kong, during the Severe Rainstorms between 22 and 24 August 1999.
Landslide Study Report LSR 3/2000, Geotechnical Engineering Office, Hong Kong,
52 p.

Fugro Maunsell Scott Wilson Joint Venture (2000c). Initial Review of the Landslide above
No. 37 Mount Davis Road on 24 August 1999. Geotechnical Engineering Office,
Hong Kong, 52 p.

Fugro Maunsell Scott Wilson Joint Venture (2001). Detailed Study of the Hillside below
Sha Tin Heights Road. Landslide Study Report LSR 4/2001, Geotechnical
Engineering Office, Hong Kong, 204 p.

Fugro Scott Wilson Joint Venture (1999a). Detailed Study of the Natural Terrain Landslide
at No. 30 Pak Sha Wan on 9 June 1998. Landslide Study Report LSR 1/99,
Geotechnical Engineering Office, Hong Kong, 47 p.

Fugro Scott Wilson Joint Venture (1999b). Detailed Study of the Landslide at The Outward
Bound School on 9 June 1998. Landslide Study Report LSR 7/99, Geotechnical
Engineering Office, Hong Kong, 88 p.

Fugro Scott Wilson Joint Venture (1999c). Detailed Study of Slope Distress at Queens Hill,
Burma Lines Camp, Fanling. Landslide Study Report LSR 10/99, Geotechnical
Engineering Office, Hong Kong, 102 p.

Fugro Scott Wilson Joint Venture (1999d). Preliminary Study of the Landslide at Luk Keng
Wong Uk on 24 May 1998. Landslide Study Report LSR 14/99, Geotechnical
Engineering Office, Hong Kong, 33 p.

Fugro Scott Wilson Joint Venture (1999e). Detailed Study of the Landslide behind 7C
Bowen Road on 11 June 1998. Landslide Study Report LSR 19/99, Geotechnical
Engineering Office, Hong Kong, 61 p.

Fyfe, J.A., Shaw, R., Campbell, S.D.G., Lai, K.W. and Kirk, P.A. (2000). The Quaternary
Geology of Hong Kong. Geotechnical Engineering Office, Hong Kong, 208 p. plus
6 maps (In press).

GCO (1984). Geotechnical Manual for Slopes (Second edition). Geotechnical Control
Office, Hong Kong, 295 p.

GEO (1987). Guide to Site Investigation (Geoguide 2). Geotechnical Engineering Office,
Hong Kong, 353 p.

GEO (1996). Report on the Shum Wan Road Landslide of 13 August 1995 - Findings of the
Landslide Investigation (Vol. 2). Geotechnical Engineering Office, Hong Kong, 51 p.
 - 53 -

GEO (1998). Guide to Slope Maintenance (Second Edition). Geotechnical Control Office,
Hong Kong, 91 p.

GEO (2000a). Highway Slope Manual. Geotechnical Engineering Office, Hong Kong,
114 p.

GEO (2000b). Technical Guidelines on Landscape Treatment and Bio-engineering for Man-
made Slopes and Retaining Walls. GEO Publication No. 1/2000. Geotechnical
Engineering Office, Hong Kong, 146 p.

Greg Wong & Associates (1999). Evaluation of Applicability of Active Landslide

Mitigation Measures to Hong Kongs Natural Terrain. Geotechnical Engineering
Office, Hong Kong, 96 p.

Halcrow Asia Partnership (1998a). Detailed Study of the Landslide Near Lido Beach, Castle
Peak Road on 2 July 1997. Landslide Study Report LSR 8/98, Geotechnical
Engineering Office, Hong Kong, 38 p.

Halcrow Asia Partnership (1998b). Detailed Study of the Landslide Behind Block 3 and 4
Wonderland Villas, Kwai Chung on 8 May 1997. Landslide Study Report LSR 11/98,
Geotechnical Engineering Office, Hong Kong, 30 p.

Halcrow Asia Partnership (1998c). Detailed Study of the Landslide Opposite Shing On
Temporary Housing Area, Ma On Shan Road on 2/3 July 1997. Landslide Study
Report LSR 19/98, Geotechnical Engineering Office, Hong Kong, 35 p.

Halcrow Asia Partnership (2000). Natural Terrain Hazard and Risk Area Studies - Mount
Johnston North and Tung Chung East - Final Report, Geotechnical Engineering Office,
Hong Kong, 97 p.

Halcrow China Ltd (2001). Detailed Study of Selected Landslides above Leung King Estate
of 14 April 2000. Landslide Study Report LSR 9/2001, Geotechnical Engineering
Office, Hong Kong, 140 p.

Ho, K., Leroi, E. & Roberds, B. (2000). Quantitative risk assessment - application, myths
and future direction. (Invited Paper). Proceedings of the International Conference on
Geotechnical and Geological Engineering GeoEng2000, Melbourne, Vol. 1,
pp 269-312.

Ho, K.K.S. & Evans, N.C. (1993). Interim Report on the Failure of Slopes 11SE-D/C182 &
11SE-D/C183, Siu Sai Wan, Hong Kong. Special Project Report SPR 3/93.
Geotechnical Engineering Office, Hong Kong.

Hungr, O., Morgan, G.C. & Kellerhals, R. (1984). Quantitative analysis of debris torrent
hazards for design of remedial measures. Canadian Geotechnical Journal, Vol. 21,
pp 663-677.
 - 54 -

Hungr, O. & Rawlings, G. (1995). Assessment of terrain hazards for planning purposes:
Cheekye Fan, British Columbia. In: Canadian Geotechnical Conference, September
25-27, 1995, Vancouver, B.C., pp 509-518.

Irfan, T.Y. (1989). Landslide Studies: Island Road Government School Landslip, Aberdeen.
Special Project Report SPR 4/89. Geotechnical Engineering Office, Hong Kong.

Johnson, A.M. & Rodine, J.R. (1984). Debris flow. Slope Stability, (Edited: Brunsden, D.
& Prior, D.B.). John Wiley and Sons Ltd, NY, pp 257-361.

King, J.P. (1996a). The Tsing Shan Debris Flow. Special Project Report SPR 6/96,
Geotechnical Engineering Office, Hong Kong, 3 volumes, 427 p, 129 p & 166 p.

King, J.P. (1996b). The Tsing Shan Debris Flood of June 1992. Special Project Report
SPR 7/96, Geotechnical Engineering Office, Hong Kong, 83 p. plus 2 drgs.

King, J.P. (1997). Damage to Liu Pok School by a Natural Terrain Landslide. Discussion
Note DN 2/97, Geotechnical Engineering Office, Hong Kong, 66 p.

King, J.P. (1999). Natural Terrain Landslide Study - The Natural Terrain Landslide
Inventory. GEO Report No. 74, Geotechnical Engineering Office, Hong Kong, 127 p.

Lam, A.Y.T. (1999). Tsing Yi North Coastal Road - Natural Terrain Hazards and Mitigation
Measures. Advisory Report ADR 1/99, Geotechnical Engineering Office, Hong
Kong, 6 p. plus 2 Appendices.

Lau, K.C. and Woods, N.W. (1997). Review of Methods for Predicting the Travel Distance
of Debris from Landslides on Natural Terrain. Technical Note TN 7/97,
Geotechnical Engineering Office, Hong Kong, 48 p.

Lo, D.O.K. (2000). Review of Natural Terrain Landslide Debris-resisting Barrier Design.
GEO Report No. 104, Geotechnical Engineering Office, Hong Kong, 91 p.

Lo, D.O.K. (2001). Interim Review of Pilot Application of Quantitative Risk Assessment to
Landslide Problems in Hong Kong. Technical Note TN 4/2001, Geotechnical
Engineering Office, Hong Kong, 69 p.

Martin, R.P. (2001). Panelist Report - Landscaping and bio-engineering of slopes in Hong
Kong. Geotechnical Engineering Meeting Societys Needs, Proceedings of the
Fourteenth Southeast Asian Geotechnical Conference, (Edited: Ho, K.K.S. & Li, K.S.),
10-14 December 2001, Hong Kong, pp 661-670.

Maunsell Geotechnical Services Ltd. (2000). Detailed Design of Check Dam at Sham Tseng
San Tsuen, Debris Flow Barrier (Check Dam) Design, Vol. I & II, Report to
Geotechnical Engineering Office, Hong Kong.

Maunsell Geotechnical Services Ltd. & ERM-Hong Kong Ltd. (1999). Territory Wide
Quantitative Risk Assessment of Boulder Fall Hazards, Stage 1. Report to
Geotechnical Engineering Office, Hong Kong, 116 p. plus Appendices A-H.
 - 55 -

Ng, K.C., King, J.P., Franks, C.A.M. and Shaw, R. (2000). Natural Terrain Hazard Study:
Interim Guidelines. Special Project Report SPR 5/2000, Geotechnical Engineering
Office, Hong Kong, 132 p.

Parry, S. (2001). Natural Terrain Hazard Study Shek Pik. Advisory Report ADR 2/2001,
Geotechnical Engineering Office, Hong Kong, 60 p. plus 3 drgs.

Resources Inventory Committee (1996). Terrain Stability Mapping in British Columbia - A

review and suggested methods for landslide hazard and risk mapping (Final Draft).
Government of British Columbia, British Columbia, Canada, 79 p.

Scott Wilson (Hong Kong) Ltd. (1999a). Specialist API Services for the Natural Terrain
Landslide Study - Potential Application of Remote Sensing Techniques for Identifying
Areas of Seepage in Hong Kong. Report to Geotechnical Engineering Office, Hong
Kong, 10 p.

Scott Wilson (Hong Kong) Ltd. (1999b). Specialist API Services for the Natural Terrain
Landslide Study - Task B Factual Report. Report to Geotechnical Engineering Office,
Hong Kong, 9 p. plus 4 Appendices.

Scott Wilson (Hong Kong) Ltd. (1999c). Specialist API Services for the Natural Terrain
Landslide Study - Interpretive Report. Report to Geotechnical Engineering Office,
Hong Kong, 32 p. plus 6 Appendices.

Scott Wilson Kirkpatrick (1996). Foothills Bypass, Tuen Mun Road/Wong Chu Road
Interchange and Other Road Improvement Road, Geotechnical Appraisal Report.
Report to Territory Development Department, 23 p. plus figures.

Sewell, R.J., Campbell, S.D.G., Fletcher, C.J.N., Lai, K.W. & Kirk, P.A. (2000). The
Pre-Quaternary Geology of Hong Kong. Geotechnical Engineering Office, Hong
Kong, 181 p. plus 4 maps.

Sobkowicz, J., Hungr, O. & Morgan, G. (1995). Probabilistic mapping of a debris flow
hazard area: Cheekye Fan, British Columbia. In: Canadian Geotechnical Conference,
September 25-27, 1995, Vancouver, B.C., pp 519-530.

Tse, C.M., Chu, T., Lee, K., Wu, R., Hungr, O. & Li, F.H. (1999). A risk-based approach to
landslide hazard mitigation measure design. In: Proceedings of the Hong Kong
Institution of Engineers Geotechnical Division Annual Seminar on Geotechnical Risk
Management. Hong Kong Institution of Engineers and Association of Geotechnical
Specialists, Hong Kong, pp 32-42.

VanDine, D.F. (1985). Debris flows and debris torrents in the southern Canadian Cordillera.
Canadian Geotechnical Journal, Vol. 22, pp 44-68.

Wong, C.K.L. (1998). The New Priority Classification Systems for Slopes and Retaining
Walls. GEO Report No. 68, Geotechnical Engineering Office, Hong Kong, 117 p.
 - 56 -

Wong, H.N., Lam, K.C. & Ho K.K.S. (1998). Diagnostic Report on the November 1993
Natural Terrain Landslides on Lantau Island. GEO Report No. 69, Geotechnical
Engineering Office, Hong Kong, 98 p. plus 1 drg.

Wong, H.N., Pang, L.S., Wong, A.C.W., Pun, W.K. & Yu, Y.F. (1999). Application of
Prescriptive Measures to Slopes and Retaining Walls. GEO Report No. 56,
Geotechnical Engineering Office, Hong Kong, 73 p.

Woods, N.W. (1992). Engineering Geology of North Lantau: Tung Chung, Vol I & II.
Special Project Report SPR 1/93, Geotechnical Engineering Office, Hong Kong, 94 p.
plus 8 drgs.
 - 57 -

LIST OF TABLES

Table Page

No. No.

1 Grouping of Facilities 58

2 Design Requirements for the Design Event Approach 59

3 Consequence Classes 60

4 Susceptibility Classes 61

5 Application of the Proposed Framework to Previous 62

NTHS Cases and Proposed Development Sites

6 Natural Terrain Features that Should be Determined as 65

Part of the API, Supported by 1:1 000 Scale

Topographical Maps

7 Natural Terrain Features that Should be Recorded during 66

Field Mapping

 - 58 -

Table 1 - Grouping of Facilities (adapted from Wong, 1998)

Group No. Facilities

(a) Buildings
- any residential building, commercial office, store and shop, hotel, factory,
school, power station, ambulance depot, market, hospital/polyclinic/ clinic,
welfare centre

1 (b) Others
- bus shelter, railway platform and other sheltered public waiting area
- cottage, licensed and squatter area
- dangerous goods storage site (e.g. petrol station)
- road with very heavy vehicular or pedestrian traffic density

(a) Buildings
- built-up area (e.g. indoor car park, building within barracks, abattoir,
incinerator, indoor games sport hall, sewage treatment plant, refuse
transfer station, church, temple, monastery, civic centre, manned
substation)
2
(b) Others
- road with heavy vehicular or pedestrian traffic density
- major infrastructure facility (e.g. railway, tramway, flyover, subway,
tunnel portal, service reservoir)

- densely-used open space and public waiting area (e.g. densely-used

playground, open car park, densely-used sitting out area, horticultural
garden)
3
- quarry
- road with moderate vehicular or pedestrian traffic density

- lightly-used open-air recreation area (e.g. district open space, lightly-used

playground, cemetery, columbarium)
4 - non-dangerous goods storage site
- road with low vehicular or pedestrian traffic density

- remote area (e.g. country park, undeveloped green belt, abandoned quarry)
5 - road with very low vehicular or pedestrian traffic density

Note: For roads, the Facility Group should be based on Figure A13 (Wong, 1998)
taking into account the actual Annual Average Daily Traffic and the number of
road lanes. For footpaths alongside roads, it may be assumed that footpaths
are within the same group as the adjoining roads, except for Expressways (EX),
Urban Trunk Roads (UT) and Rural Trunk Roads (RT). Footpaths alongside
EX, UT and RT roads may be taken, by default, as a Group 5 facility, unless
dictated otherwise by site-specific conditions.
 - 59 -

Table 2 - Design Requirements for the Design Event Approach

Consequence Class (see Table 3)

Susceptibility
Class (see Table 4)
I II III IV

A WCE WCE WCE CE

B WCE WCE CE CE

C WCE CE CE N

D N N N N
Notes: (1) The recommended minimum design requirements are given in this Table.
The designer may adopt a more conservative design or provision of other
precautionary/warning measures if considered necessary.
(2) This Table will be applied to each sub-catchment and normally the
predominant type of hazard will control the design requirements.
(3) WCE = Adopt a worst credible event as the design event.
CE = Adopt a conservative event as the design event
N = Further study is not required
 - 60 -

Table 3 - Consequence Classes

Facility Group
Proximity
1&2 3
Very Close
I II
(e.g. if angular elevation from the site is 30)
Moderately Close
II III
(e.g. if angular elevation from the site is 25)
Far
III IV
(e.g. if angular elevation from the site is < 25)
Notes: (1) Facility groups are given in Table 1. Facilities of Groups 4 and 5
normally do not require natural terrain hazard study and hence they are not
included in Table 3.
(2) For channelized debris flow, if the worst credible event affecting the site is
judged to have a volume exceeding 2000 m3, the angular elevation given in
the above examples should be reduced by 5.
(3) The examples given above are for general guidance only. Other factors,
such as credible debris path, topographical conditions and site-specific
historical data, should also be taken into account in assessing the
proximity of the natural terrain to the site.
(4) This Table is applicable only to sites that require natural terrain hazard
study.
 - 61 -

Table 4 - Susceptibility Classes

Susceptibility
Description
Class
The natural terrain is extremely susceptible to the type of failure under
consideration, with a notional annual probability of occurrence in the
order of 1/10 or higher. For example: there are signs of instability,
A continued movement, or records of repeated recent failures (say over
the past 50 to 100 years as observed from aerial photographs) in the
catchment and its relevant vicinity.

The natural terrain is highly susceptible to the type of failure under

consideration, with a notional annual probability of occurrence within
B the order of 1/10 to 1/100. For example: there are records of occasional
recent failures in the catchment and its relevant vicinity.

The natural terrain is moderately susceptible to the type of failure under

consideration, with a notional annual probability of occurrence within
the order of 1/100 to 1/1000. For example: there are few records of
recent failures, but there are indications of relict failures, or
C
geomorphological evidence of potential problems in the catchment and
its relevant vicinity, or any other evidence from similar terrain in Hong
Kong.

The natural terrain is of low susceptibility to the type of failure under

consideration, with a notional annual probability of occurrence less than
1/1000. For example: there are no records of recent and relict failures,
D little geomorphological and other evidence of potential problems in the
catchment and its relevant vicinity, and little other evidence from similar
terrain in Hong Kong.

Note: In assessing the susceptibility of the hillside to failure, consideration should be

taken of the potential effects of changes in environmental factors, e.g. any
changes to the overall setting of the terrain such as hill fires and construction
upslope, and the relevance of the available historical landslide records. Most
of the information on environmental changes can be obtained from a review of
past aerial photographs.
 Table 5 - Application of the Proposed Framework to Previous NTHS Cases and Proposed Development Sites (sheet 1 of 3)

Design Requirement
Consequence Class (see Table 3)
Predominant (see Table 2)
Actual design
Site Catchment No. Type of Required Remarks
Angular Susceptibility Event
Hazard Facility Group Proximity Class Design
Elevation Class
Event

Actual design event based on extrapolation

using regional data to a 1-in-300 years
1 failure at source together with mobilisation
Area 49A, Channelized VC (WCE
1 (residential 26 I B WCE WCE of material along trail and with no deposition
Fanling debris flow >2000 m3)
building) before reaching the debris barrier. Factor
of Safety Approach adopted in checking
against deep-seated failures.
1 A (given the 1990
Channelized 2 VC (WCE >
(catchment of 29 I large scale debris WCE WCE
debris flow (road) 2000 m3)
1990 failure) flow)

- 62 -
2
Tsing Shan (catchment of Channelized 2 VC (WCE > Actual design event based on the 1990 Tsing
30 I A/B WCE WCE
Foothill 2000 failure - debris flow (road) 2000 m3) Shan debris flow, which was assessed as an
Bypass North) event in the order of 1 in 1000 years.
3
(catchment of Channelized 2 VC (WCE >
30 I A/B WCE WCE
2000 failure - debris flow (road) 2000 m3)
South)
4
1 A (given the 1990
Channelized (lightly- used VC (WCE > Actual design event based on the 1990 Tsing
(catchment of 28 III large scale debris WCE WCE
debris flow open recreation 2000 m3) Shan debris flow, which was assessed as an
Area 19, 1990 failure) flow)
area) event in the order of 1-in-1000 years. Due
Tuen Mun
2 4 to the presence of a buffer zone between the
Golf Centre
(catchment of Channelized (lightly- used VC (WCE > WCE/ site and the hillside, the mitigation measures
28 III A/B WCE provided are not very substantial.
2000 failure - debris flow open recreation 2000 m3) CE
North) area)
 Table 5 - Application of the Proposed Framework to Previous NTHS Cases and Proposed Development Sites (sheet 2 of 3)

Design Requirement
Consequence Class (see Table 3)
Predominant (see Table 2)
Actual design
Site Catchment No. Type of Required Remarks
Angular Susceptibility Event
Hazard Facility Group Proximity Class Design
Elevation Class
Event

Actual design event based on x10 the largest

Channelized 2 F (WCE > historical failure. Since the facility is quite
1 18 III C CE WCE
debris flow (road) 2000 m3) far away from the hillside, the mitigation
measures provided are not very substantial.
N. Tsing Yi Although the debris barriers are designed for
Coastal Road a CE, some stabilisation measures are
Open slope 2 provided to the hillside to reduce the chance
2 28 MC II B WCE CE
failure (road) of open slope failure. In this respect, the
actual design event may also be taken as
equivalent to a WCE.

- 63 -
By provision of buffer zone, angular
Open slope 1 elevation from edge of site set back to 23,
27 MC II C CE CE
failure (factory) which was considered adequate given the
To Kau Wan, design event is not large in volume.
1
N. Lantau By provision of buffer zone, angular
Open slope 1 elevation from edge of site set back to 8.
30 VC I C WCE WCE
failure (factory) This is conservative since the hillside in
question is small in size.
The debris barrier is designed for a CE
VC (if WCE which corresponds to more than twice the
I B WCE
Sham Tseng > 2000 m3) volume of the 1999 failure. Additional
San Tsuen Channelized 1 buffer zone is provided downstream of the
1 28 CE
squatter debris flow (squatters) barrier to cater for possible over-spill from a
village MC (if WCE larger event. The mitigation works were
II B WCE urgent protective measures for an existing
2000 m3)
squatter village in response to a landslide.
Proposed C Proposed
1
development Channelized VC (WCE > (based on development
1 (residential 26 I WCE The design requirements are reasonable.
site, Blacks debris flow 2000 m3) preliminary (NTHS yet to
building)
Link assessment) be carried out)
 Table 5 - Application of the Proposed Framework to Previous NTHS Cases and Proposed Development Sites (sheet 3 of 3)

Design Requirement
Consequence Class (see Table 3)
Predominant (see Table 2)
Actual design
Site Catchment No. Type of Required Remarks
Angular Susceptibility Event
Hazard Facility Group Proximity Class Design
Elevation Class
Event

The less stringent design requirements are

Proposed bus C/D Proposed reasonable. For this site, it may be possible
shelter, Lion Open slope 1 (based on development to show that the landslide debris will not
1 29 MC II CE/N
Rock Tunnel failure (bus shelter) preliminary (NTHS yet to reach the site and hence no mitigation
Road assessment) be carried out) measures are required, if the CE is not of
large volume.
The less stringent design requirements for
Proposed
4 C/D this case which is of less serious
Short Term
Channelized (non-dangerous (based on consequence are reasonable. The critical
Tenancy site, 1 32 VC III CE/N N
debris flow goods storage preliminary facilities will be located farthest away from
Mt. Parker
site) assessment) the hillside, and subsequently the NTHS

- 64 -
Road
requirement has been waived.
Mitigation measures will not be required if the
hillside is assessed as of low susceptibility.
Proposed C/D Proposed Alternatively, if the proposed buildings are
1
development Open slope (based on development placed on the part of the site that is further
1 (residential 29 MC II CE/N
site, South failure preliminary (NTHS yet to away from the hillside and if the CE is not of
building)
Bay Road assessment) be carried out) large volume, it may be demonstrated that the
debris will not reach the buildings and no
mitigation measures are required.
1
Channelized QRA
1 (residential 24 MC II B WCE
debris flow approach
Shatin building) An existing development affected by
Heights 1 several hillside failures in 1997.
Open slope QRA
2 (residential 33 VC I A/B WCE
failure approach
building)

Notes: (1) VC = Very Close; MC = Moderately Close; F = Far. See Table 3 for details.
(2) WCE = Worst credible event; CE = Conservative event; N = Further study is not required. See Table 2 for details.
(3) See Table 3 for details of Consequence Classes, Table 1 for facility types, and Table 4 for Susceptibility Classes.
 - 65 -

Table 6 - Natural Terrain Features that Should be Determined as Part of the API,
Supported by 1:1 000 Scale Topographical Maps (modified from
Resources Inventory Committee, 1996)

Terrain Feature Examples (not exhaustive)

Slope Morphology
Gradient Typical, average or range over a specified slope length
Uniformity of slope Straight, stepped, benched, concave, convex, etc.
Lateral curvature Broad, narrow, re-entrant, ridge
Position Near crest, mid-slope, near toe, etc.
Elevation Typical, range
Length Slope length with similar features
Aspect Quadrant with respect to north
Identified Processes
Rock fall, open hillslope landslide, deep-seated slide,
Landslides
channelized debris flow, etc.
Erosion Sheet, rill, drainage line, etc.
Other processes Flooding
Age Recent, relict, or date if known
Areal extent Width and length
Depth Typical
Incipient Processes
Instability signs Tension cracks, disrupted road, tilted trees, etc.
Abundance Number of indications per area or length
Superficial Material
Origin, genesis Fluvial, colluvial, alluvial, boulder field, etc.
Geomorphological features Debris fan, apron, cone, landslide scar, rock exposure, etc.
Geomorphological processes Gullying, erosion, failing, etc.
Solid Geology
Outcrops Location and extent
Structural features Bedding, faults, folds, dykes, other discontinuities
Drainage Lines
Order and status First, second, third, etc; permanent vs ephemeral or dry
Channel bed material Inorganic (boulders, bare bed rock etc) vs vegetation debris
Channel processes Flood, debris flood, debris avalanche, etc.
Sideslope/Headslope material As for superficial material above
Sideslope/Headslope processes As for identified and incipient processes above
Groundwater
Soil drainage Rapidly, well, moderate, etc.
Seepage regime Recharge area, discharge area, undrained area, etc.
Vegetation
Type Trees, shrub, grass, bare
Age <5 years, 5 - 10 years, 10-20 years, 20 -30 years, >30 years, etc.
Deforestation/fire history Deforestation plus years since, hill fire plus years since
Human Activity
Fill at top of slope, cut, agricultural terraces, graves, footpaths,
Type
road across slope, etc.
Quantity Road length/unit area, etc.

Table 7 - Natural Terrain Features that Should be Recorded during Field Mapping
(modified from Resources Inventory Committee, 1996)
Terrain Feature
Identified Processes
Landslides, Erosion, other
Processes
Areal extent
Depth
Superficial Material
Type
Boulder Shape
Properties of soil
Thickness
Genesis of Contrasting Layers
Solid Geology
Rock Type
Decomposition Grade
Block size and shape
Structural attitudes
Properties of rock mass
Description of Contrasting
Layers
Layer description
Thickness
Substrate description
Nature of contact
Drainage Lines
Channel gradient
Uniformity of gradient
Channel width
Sideslope/Headslope height
Sideslope/Headslope gradient
- 66 -
Examples (not exhaustive)
Width & length (of source area and debris trail, debris run out,
deposition, erosion)
Typical (of source, erosion, etc.)
Saprolite, Colluvium, alluvium, scree
Rounded, sub-rounded, angular, tabular, etc.
Strength, consistency, density, etc.
Typical, average or range
Saprolite overlain by colluvium, etc.
Granite, feldsparphyric rhyolite, eutaxitic tuff, etc.
Fresh, slightly decomposed, moderately decomposed
Spacing of discontinuities, cubes, etc.
Strike, dip, dip direction
Strength, fracture roughness, RQD, PW 90/100, etc.
Loose, dense etc, together with material type
Typical, average or range
Residual soil, saprolite, bedrock, etc.
Weak soil horizon, distinct, etc.
Typical, average or range (in degrees, over typical length)
Uniform, stepped, etc.
Typical, average or range
Typical, average or range
Typical, average or range
 - 67 -

LIST OF FIGURES

Figure Page
No. No.

1 Examples of Anthropogenic Features on Natural Terrain 68

2 Application of In-principle Objection Criteria and Alert 69

Criteria

3 Overall Process of Natural Terrain Hazard Study 70

4 Examples of Debris- and Boulder-resisting Barriers Used 71

in Hong Kong
 - 68 -

(a) Abandoned Agricultural Terraces, (b) Abandoned Agricultural Terraces,

Route Twisk (1963) Route Twisk (1999)

(c) Military Redoubt, Ying Pun, (d) Foxholes from a Military Exercise,
New Territories (1963) Route Twisk (1963)

(e) Mineral Exploration Trenches, (f) Mining Induced Failures,

South Lantau (1963) Ma On Shan (1977)

Figure 1 - Examples of Anthropogenic Features on Natural Terrain

 - 69 -
Figure 2 - Application of In-principle Objection Criteria and Alert Criteria
 - 70 -

Figure 3 - Overall Process of Natural Terrain Hazard Study

(a) Check Dam (Sham Tseng San Tsuen) (b) Landslide Debris/Boulder Barrier (Fanling)

- 71 -
(c) Boulder Fence (Mid-Levels) (d) Gabion Barrier (Shau Kei Wan)

Figure 4 - Examples of Debris- and Boulder-resisting Barriers Used in Hong Kong

 - 72 -

APPENDIX A

GLOSSARY OF TERMS AND

GENERAL CLASSIFICATION OF HILLSIDE FAILURES

 - 73 -

CONTENTS

Page

No.

Title Page 72

CONTENTS 73

A.1 GLOSSARY OF TERMS FOR NATURAL TERRAIN LANDSLIDES 74

A.1.1 Introduction 74

A.1.2 Glossary of Terms 74

A.2 GENERAL CLASSIFICATION OF HILLSIDE FAILURES 86

A.2.1 Introduction 86

A.2.2 Classification Framework 86

A.3 REFERENCES 87

LIST OF FIGURES 91

 - 74 -

A.1 GLOSSARY OF TERMS FOR NATURAL TERRAIN LANDSIDES

A.1.1 Introduction

This glossary of terms has been compiled to help promote a consistent use of terms in
the study of natural terrain hazards in Hong Kong. Much of it is extracted from Cruden and
Varnes (1996) which is a widely accepted reference on landslide types and processes. A
recent review on the landslide terminology is given by Hungr et al. (2001).

For many Hong Kong natural terrain landslides, it is difficult to establish the
mechanism of initial displacement and transport and use of the wider definitions proposed
below may be more practical. The terminology is primarily prepared for study of natural
terrain landslides in Hong Kong. As such, it is biased towards terms used for describing
elongate, relatively shallow natural terrain landslides which are dominated by flow and
alluvial processes. The glossary includes some additional and more detailed terms than
those of Cruden and Varnes (1996), and some new definitions. Additional definitions of
landslide types are based on Hutchinson (1988) and for the description of sediment-water
flows, on Pierson and Costa (1987). Users of the terminology may define and use their own
terms for any specific project although consistency is recommended wherever possible.
Figure A1 illustrates the main features of a typical natural terrain landslide scar.

A.1.2 Glossary of Terms

Angle of Reach

The arc-tangent of the height of a landslide (H) divided by its length (L) where; H = elevation
difference between the crown and the toe and L = the horizontal distance from the crown to
the toe. Note that this is similar to the apparent angle of friction (Sassa, 1987), equivalent
coefficient of friction (Hsu, 1975) and average coefficient of friction (Scheidegger, 1973).
However, these terms have in some cases been used when H and L are defined on the basis of
the centre of gravity of the displaced material. The term reach angle is proposed in
conformity with its use by Corominas (1996) as this reference may be used to help estimate
run out distance. The same angle is termed Travel Angle by Cruden and Varnes (1996).

Blocky Landslide Debris

Landslide debris comprising loose blocks of intact displaced material, commonly rock and/or
soil.

Block

A discrete boulder-sized fragment of rock significantly detached from its outcrop source and
characterized by an angular shape of tabular or blocky form. It may lie on or adjacent to the
source or have travelled downslope.

Boulder

A rock fragment greater than 200 mm diameter that is not part of a rock mass.
 - 75 -

Boulder Fall

The displacement of a boulder where the movement may include free fall through the air,
sliding, leaping and rolling. Note that in Hong Kong boulder falls commonly result from
undercutting by erosion, or landsliding of exhumed corestones or colluvial boulders on steep
natural hillsides.

Boulder Field

An area covered by a veneer of boulders.

Catchment

An area of a hillside with a relatively uniform geomorphology from which one or several
well-defined hazard models could affect the site.

Channel

A linear depression of the ground surface in which water may flow.

Channelize

To convey in a channel. A landslide is considered to be channelized when the moving

landslide debris converges into a channel that confines its lateral extent. Note that this
commonly results in a longer travel distance than for an open slope landslide of the same
volume.

Channelized Debris Flow

A debris flow that has become channelized in a drainage line with significant amounts surface
water mixed into the debris. Note that while debris flows may occasionally develop on
hillslopes and become confined in hollows, these are difficult to distinguish from debris
avalanches on the basis of mobility or by API, or even by field observations soon after the
failure. Therefore, for the purposes of hazard recognition and data collection, the term is

best used for debris flows that are coincident with a streambed.

Channelization Ratio

The width to depth ratio of the cross section area in a channel occupied by a pulse of landslide

debris. Channelization may be defined in relation to this ratio. Hungr et al. (1984)

suggested 5; the Tsing Shan Debris Flow started to deposit at about 7 (King, 1996).

Cliff

Any steep to sub-vertical or over-hanging face of rock.

Colluvium

A general term applied to generally structureless and commonly heterogeneous mass of soil

 - 76 -

and/or rock material and sometimes organic matter, deposited on, and at the base of, natural
slopes predominantly by mass-wasting processes (after Bates & Jackson, 1987). It is usually
landslide debris.

Colluvial Fan

A colluvial deposit shaped like a fan or cone, usually with the apex at a point where a
drainage line ends. Such a feature is usually the result of deposition from a number of debris
flows over a period of time.

Colluvial Lobe

A colluvial deposit with convex sides that forms a positive landscape feature. Often located
along the bottom of a valley. Such a feature is usually interpreted as a landslide debris
deposit from a single landslide.

Confined Scar

A landslide scar in which the surface of rupture comprises a distinct main scarp and floor and
may include a rising downslope part. The source material cannot be displaced until it is
transformed into a kinematically feasible geometry by internal displacements or shears
(Hutchinson, 1988). A confined scar has no trail.

Crown

The practically undisplaced material still in place and adjacent to the highest parts of the main
scarp (IAEG, 1990).

Debris

Predominantly coarse soil. In Hong Kong such material is usually the weathered profile
and/or colluvium and may be referred to as regolith. Landslide displaced material that has
disintegrated is commonly called debris. To retain conformity with international practice
the term landslide debris is used for this in the terminology (see Landslide Debris).

Debris Avalanche

A landslide comprising a rapid to extremely rapid, shallow movement of partially or fully

saturated mainly granular debris on steep slopes with only minimal water mixed into the
displaced material after failure. Movement is by sliding and/or liquefaction and/or rolling
and/or bouncing and/or saturated or inertial granular flow (Pierson and Costa, 1987; Varnes,
1978).

Debris Flood

The transport of significant load of sediment by the dominant mechanisms of stream flow or
hyperconcentrated stream flow. The load can range in grain size from clay to boulders. A
debris flood may originate from outwash of landslide debris or gully erosion.
 - 77 -

Debris Flow

A landslide in which the landslide debris moves by the dominant mechanism of slurry flow.
Note that the presence of remoulded landslide debris is a good indication that slurry flow has
occurred. In Hong Kong, debris flows usually develop from debris slides or debris
avalanches. If the moisture content of the displaced material is high enough then slurry flow
may develop during movement without additional water. More commonly a debris flow
develops when landslide debris mixes with stream water. Debris flows may form open
hillslope failures on relatively planar slopes or in hollows but commonly become confined in
drainage lines forming channelized debris flows.

Debris Slide

A landslide in which an intact displaced mass moves by the dominant mechanism of sliding.
Note that when these occur on steep hillsides they typically form an elongate scar.

Debris Torrent

A term used in Canada by D.F. VanDine (1985) to mean the same as channelized debris
flow and defined by him as a mass movement that involves water-charged, predominantly
coarse grained inorganic and organic material flowing rapidly down a steep confined pre-
existing channel. Torrent is used by Aulitzky (1989) to describe steep mountain streams.

Debris Run-Out

The travel distance of landslide debris beyond a specified point. The landslide run-out
may be the horizontal length of the scar. Run-out may also be considered beyond a point
with a limiting angle where deposition starts (deposition point) or the slope toe.

Deposition Point

A notional point on a landslide trail beyond which most landslide debris is assumed to be
deposited.

Design Event

The particular size specification for a hazard model that is selected to be mitigated against as
part of a development.

Displaced Material

The material moved from its original position on the slope by a landslide (IAEG, 1990).
Note that an accumulation of displaced material forms a deposit.

Drainage Line

A channel with a stream bed in which water flows for at least part of the year.
 - 78 -

Drainage Line Order

A first order drainage line has no confluences (i.e. no other contributing drainage lines), two

first order drainage lines make a second order, and two second orders drainage lines make a

third order (a second and first order confluence remains a second order) (after Strahler, A.N.,

1952).

Elongate Scar

A landslide scar in which some or all of the displaced material has mobilized from the surface

of rupture forming landslide debris. Note that an elongate scar has a trail and the source is
often a spoon shaped or tabular depression.

Elongate Channelized Scar

An elongate scar in which the landslide debris has been channelized in a drainage line or

depression and the trail has a low channelization ratio.

Entrainment Factor

The volume of additional material that has been entrained by a landslide expressed as a

proportion of the source volume.

Entrained Material

The displaced material from any location other than the source.

Floor

Any exposed shallow part of the surface of rupture.

Flow

A spatially continuous movement in which surfaces of shear are short-lived, closely-spaced,

and usually not preserved. The distribution of velocities in the displacing mass resembles

that in a viscous liquid (Cruden and Varnes, 1996).

Frequency

The number of events that occur in a specific period of time.

Geological Model

A representation of the geology of a particular location. The form of the model can vary
widely and may include written descriptions, two-dimensional sections or plans, block
diagrams, or be slanted towards some particular aspect such as groundwater or
geomorphological processes, rock structure and so on (Fookes, 1997).
 - 79 -

Gully Erosion

The incision of ground along flow lines by surface water.

Hazard

A physical situation with a potential for human injury, damage to property, damage to the
environment or some combination of these (ERM, 1998).

Hazard Model

A type or types of landslide that will result in a well-defined risk to a development site. It is
described in terms of volume, nature and velocity of material passing the site boundary at a
specified location.

Hazard Intensity

A relative measure of the potential for harm from a hazard. High intensity has a greater
potential for injury and loss than low intensity.

Head

The upper parts of a landslide along the contact between the displaced material and the main
scarp (IAEG, 1990). Note that an elongate landslide may not have a definable head.

Hollow

A depression, commonly linear, on hillside.

Hyperconcentrated Stream Flow

The flow of a mixture of water and sediment that possesses a small but measurable yield
strength but that still appears to flow like a liquid. Particles settling out of a
hyperconcentrated stream flow suspension settle independently, giving deposits sorted by
grain size (Pierson and Costa 1987).

Inertial Granular Flow

A flow in which the full weight of the flowing granular mass is borne by grain-to-grain
contact or collisions in which grain inertial effects dominate but frictional effects are still
significant (Pierson and Costa, 1987).

Intact Displaced Mass

The full mass of displaced material (soil, colluvium or rock) from a landslide when it largely
retains its original morphology.
 - 80 -

Intact Displaced Material

A displaced block of soil, colluvium or rock that retains its original structure. Note that
intact displaced material may occur as clasts, blocks or slabs within landslide debris.

Intermediate Deposit

The landslide debris accumulation at some point along the trail between the source and the
toe.

Landslide

A general, all-encompassing term used to describe an event comprising the relatively rapid
downslope movement of a discrete mass of soil and/or rock. Note that the term makes no
inference to process, i.e. true sliding may not be involved. It is proposed that extremely slow
slope movements such as soil creep and gravitational sagging are not included under the
general term landslide.

Landslide Cause

An environmental factor that gradually brings a slope to failure such as the chemical or
physical weathering of slope materials (Weiczorek, 1996).

Landslide Debris

The displaced material from a specific landslide that has disintegrated and lost its original
morphology. Landslide debris may include clasts of intact displaced material. Note that
landslide debris is very young colluvium and it is proposed that the term is used to distinguish
the landslide debris of a specific landslide from pre-existing in-situ colluvium in the regolith
which may have the same composition. The term landslide debris may be shortened to
debris for convenience where this will not be ambiguous or confused with regolith debris.

Landslide Erosion Feature

A relatively large depression defined by scarps or a sharp break in slope. This is assumed to
be landslide related. Smaller recent landslides may occur along the perimeter scarp but there
is no debris deposit that can be unambiguously interpreted as the result of a single failure.
This feature could be the source of a single very large landslide, or the result of a number of
smaller landslides over time, or an erosion scarp (possibly landslide generated) cutting back
into an older landscape.

Landslide Trigger

An external stimulus that causes a near-immediate response in the form of a landslide by

rapidly increasing the stresses or by reducing the strength of slope materials. Typical
triggers are intense rainfall, collapse of a soil erosion pipe, earthquake shaking, storm waves
or rapid stream erosion (Weiczorek 1996).
 - 81 -

Lateral Deposit

A landslide debris deposit comprising ridges aligned along the trail at or near its lateral
margins.

Liquefaction

The generation of high positive excess pore water pressures during shearing and hence a
substantial reduction of the effective stress and the shearing resistance. Note that in Hong
Kong liquefaction is known to have occurred in the rain-induced failure of loose fill slopes.

Liquefaction Slide

A slide in which liquefaction occurs within a zone at the sliding surface and the landslide
debris is not saturated (Casagrande, 1971; Hutchinson, 1988).

Main Scarp

The exposed steep part of the surface of rupture (IAEG, 1990).

Mass Transport

The carrying of soil and rock material in a moving medium such as water, air, or ice (after
Bates & Jackson, 1987).

Mass Movements

A general term for the dislodgement and downslope transport of soil and rock material under
the direct application of gravitational body stresses. In contrast to other erosion processes,
the debris removed by mass movement processes is not carried within, on, or under another
medium. The mass properties of the material being transported depend on the interaction of
the soil and rock particles and on the moisture content. Mass movements include slow
displacements, such as creep, and rapid processes such as rock falls, rock slides, and debris
flows (Syn. mass wasting) (after Bates & Jackson, 1987).

Mobilization

The movement of displaced material from the source. Usually expressed as the proportion
of the material that is displaced.

Mud Flow

A landslide with the same characteristics as a debris flow in which the displaced material is
predominantly fine-grained.

Natural Terrain

Terrain that has not been modified substantially by human activity such as site formation
works, agricultural terracing, cemetery platforms or squatter habitation. Note that in most of
 - 82 -

Hong Kong natural terrain has been influenced by deforestation and fire and locally may have
been influenced by prehistoric agriculture. Natural terrain and natural hillside are synonyms

Natural Terrain Landslide

A landslide in which the source is located entirely within natural terrain (see Section A.2.2
below). Note that a landslide involving both natural terrain and terrain that has been
modified substantially by human activity is not considered to be a natural terrain landslide, e.g.
failure immediately above and extending into an excavation or immediately below fill.

Open Hillslope Lanslide

A landslide with its scar entirely on the hillside and is not channelized along a stream course.
It is initiated on a hillside slope and transported by the mechanism of debris avalanche, debris
slide or debris flow.

Outcrop

An exposed area underlain by a specific material or rock unit.

Outwash

The redistribution of landslide debris by stream flow and hyper-concentrated stream flow.

Parent Landslide

A discrete initial landslide, the landslide debris from which increases in volume through
erosion and entrainment of the downslope substrate to become a debris flow.

Quasi-natural Terrain

Predominantly natural terrain which is partly modified by human activities but largely retains
its original profile and regolith cover. Minor surface features such as agricultural terracing,
grave sites or squatter platforms may be present. Modification of the slopes above and
below by cutting, filling and construction may have affected the surrounding ground and
surface water conditions.

Quasi-natural Terrain Landslide

A failure in predominantly natural terrain triggered by, or possibly triggered by, a relatively
small human activity. Landslides for which there is an element of doubt with respect to the
importance of human activity are placed in this category. Section A.2.2 below provides
further information on classification of such landslides.

Recent Landslide

A landslide that occurred within the aerial photograph coverage. In the NTLI recent
landslides are defined as those with a light tone on aerial photographs and are generally bare
of vegetation.
 - 83 -

Regolith

A general term for the layer or mantle of fragmental and unconsolidated rock material
(engineering soil), whether residual or transported and of highly varied character, that nearly
everywhere forms the surface of the land and overlies or covers the bedrock. It includes
rock debris of all kinds (after Bates & Jackson, 1987).

Relict Scar

A scar on which vegetation has re-established but which still has a well-defined main scarp.
In the NTLI, relict is defined in relation to landslides first observed on the earliest available
photographs and with vegetation class C or D (Grass or Trees).

Remoulded Landslide Debris

Landslide debris that has largely lost its original structure and comprises a remoulded matrix
that supports particles of gravel size or larger. Note that remoulded landslide debris may
occur as steep-sided deposits such as lobate fronts, levees, and accumulations on the uphill
side of, or capping, obstructions in the trail.

Risk

The likelihood of a specified undesired event occurring within a specified period or in

specified circumstances. It may be viewed in terms of either a frequency (the number of
specified events occurring in unit time) or a probability (the likelihood of a specified event
following a prior event) depending upon the circumstances (ERM, 1998). It commonly
refers to as the product of frequency (or probability) and consequence.

Rock Fall

The displacement of a piece of rock from a rock face chiefly by free fall through the air but
may include sliding, leaping and rolling. Note that this is a simple landslide if the initial
displacement does not involve another mechanism such as sliding or toppling.

Rotational Scar

A scar in which the surface of rupture is curved and concave upwards and imparts a degree of
backward rotation or tilt to the surface of the displaced ground. Note this was termed a
slump by Varnes (1978) and a slip by Hutchinson (1988).

Run-out

The process of transport of displaced material beyond the source. If expressed as a distance
it will be length less the horizontal extent of the source.

Scar

The land surface affected by a landslide. This includes the source, the displaced material
and any trail.
 - 84 -

Sheet Erosion

The uniform removal of soil or decomposed rock by the surface flow of water or a mixture of
water and sediment.

Side/Head wall

The area of a catchment or sub-catchment that could contribute debris to a stream channel.

Site

A proposed development area demarcated by a continuous line on a topographical base map

of 1:5 000 or larger scale. The whole site should be considered habitable unless buffer or no
build zones are incorporated into the lease.

Site Catchment

The area from which surface runoff would intersect the site boundary or flow into a drainage
line adjacent to the site, effectively the water catchment to a site.

Slide

A landslide in which the movement of debris is sliding. Note that this may be further
qualified by a prefix describing the morphology of the scar and specifying the material
involved, e.g. slides may result in translational, rotational, compound or confined scars from
failures of colluvium, saprolite or rock.

Sliding

Downslope movement occurring dominantly on surfaces of rupture or on relatively thin zones

of intense shear strain (Cruden and Varnes, 1996).

Slump

A scar with an intact displaced mass which partially overlies the surface of rupture

Slurry Flow

The movement of a saturated sediment/water mixture having sufficient yield strength to

exhibit plastic flow behaviour in the field. When movement ceases, fine and coarse particles
settle together with no inter-particle movement (Pierson and Costa 1987).

Sorted Landslide Debris

Landslide debris in which grain size sorting and/or layering is present and that has little or no
fines.
 - 85 -

Source

The space between the surface of rupture and the original ground level.

Stream Bed

The bottom of a relatively narrow but clearly defined channel, where surface water flows or
may flow over bedrock and/or loose granular deposits.

Stream Flow

The flow of water with insufficient sediment concentration to affect the flow behaviour
(Pierson and Costa 1987).

Study Area

The area from which natural terrain landslide hazards could affect the site. In most cases
this should include both the site and its catchment.

Substrate

In-situ material over which landslide debris is transported.

Surface of Entrainment

A surface on undisturbed ground that was originally below ground level and from which
material has been picked up and moved away during the landslide. This is separate, and
downslope from the source.

Surface of Rupture

A surface on undisturbed ground that was originally below ground level and from which a
discrete mass of displaced material has moved away (IAEG, 1990).

Terminal Deposit

A landslide debris deposit at the toe of the trail. It is commonly lobate or fan shaped.

Toe

The lower margin of the displaced material of a landslide. It is the most distant part of the
scar from the main scarp IAEG (1990). Note that for a landslide with a trail this will be the
distal part of the trail.

Topple

Movement of a detached rock mass by overturning about a pivot point below the centre of
gravity of the unit (Varnes, 1978). Note that topples are often multiple.
 - 86 -

Trail

The part of a scar downslope from the source. This may include displaced material, ground
over which displaced material has passed and any surface of entrainment.

Translational Scar

A scar in which the surface of rupture is relatively planar or gently undulating in downslope
section and often broadly channel-shaped in cross section (Cruden and Varnes 1996).

Uniform Deposit

A landslide debris deposit comprises a sheet of landslide debris of uniform thickness within
the trail.

Weathered Mantle

A regionally widespread and usually deep zone of weathered rock materials, formed in situ
over a geologically long interval by relatively uniform chemical weathering (Syn. weathered
profile, weathering crust) (after Bates & Jackson, 1987).

A.2 GENERAL CLASSIFICATION OF HILLSIDE FAILURES

A.2.1 Introduction

There is a need for use of consistent terminology in denoting different types of hillside
failures in Hong Kong. This will facilitate systematic landslide classification and statistical
analysis. In this note, hillside failure refers to a landslide that occurs mainly on a hillside
instead of man-made features. Man-made feature refers to cut slope, fill slope or retaining
wall. Hillside includes disturbed terrain.

For the purpose of general classification, the terminology should be practical and
consistent with the nature and quality of the information that would normally be available for
most of the reported hillside failures.

A.2.2 Classification Framework

For general classification, hillside failures are classified according to the types of
terrain involved in the source of the landslide (Figure A2):

(a) Natural hillside failure - occurs on natural hillside which has

not been modified by man-made activities (e.g. cutting,
filling, cultivation, mining, etc.)

(b) Disturbed-hillside failure - occurs on hillside which has

locally been modified by man-made activities but there is no
indication that the landslide involves registerable man-made
features. Where information is available, such landslide
 - 87 -

should be further classified as follows:

(i) Disturbed-hillside failure (Type A) - which is not

significantly controlled by non-registerable man-made
features

(ii) Disturbed-hillside failure (Type B) - which is

significantly controlled by non-registerable man-made
features

(c) Hillside failure involving registerable man-made features -

occurs predominantly on hillside but partly involves
registerable man-made features.

Selected cases of hillside landslides are given in Figure A2 as examples of

classification based on the proposed framework.

The above classification may also be applied to individual rock or boulder falls from
hillside.

The mode, mechanism and cause of failure are not accounted for in the general
classification, as reliable information on these is not normally available. Where such
information is available, say from detailed landslide studies, the landslides can be further
described or classified, on top of the general classification.

It should be noted that a natural hillside failure (Item (a) above) may be triggered by
factors that involve man-made influence (e.g. concentration of surface water overflowed from
a catchwater). Conversely, a disturbed-hillside failure (Item (b) above) and a hillside
failure involving registerable man-made features (Item (c) above), which are to some extent
affected by man-made activities, may be triggered by purely natural causes (e.g. heavy rain).

Strictly speaking, only natural hillside failures that are caused by natural factors
should be considered as genuine natural terrain landslides. The other types of hillside
landslides, i.e. disturbed-hillside failures, hillside failures involving registerable man-made
features, and natural hillside failures involving man-made triggers, were often denoted
collectively as quasi-natural terrain landslides (see Section A.1.2 above) in the past. The
proposed classification provides a practical framework to differentiate these landslides.

A.3 REFERENCES

Glossary of Terms for Natural Terrain Landslides

Aulitzky, H. (1989). The debris flows of Austria. Bulletin of Engineering Geology,

Vol. 40, pp 5 - 13.

Bates, R.L. & Jackson, J.A. (1987). Glossary of Geology. Third Edition, American
Geological Institute, Alexandria, Virginia, 788 p.
 - 88 -

Casagrande, A. (1971). On liquefaction phenomena: report of a lecture. Geotechnique,

Vol. 21, pp 197-202.

Corominas, J. (1986). The angle of reach as a mobility index for small and large landslides.
Canadian Geotechnical Journal, Vol. 33, pp 260 -271.

Cruden, D.M. and Varnes, D..J. (1996). Landslide Types and Processes. In: Landslides
Investigation and Mitigation (Edited: Turner, A.K. and Schuster, R. L.), Special
Report 247 of the Transportation Research Board, National Research Council,
National Academy Press, Washington, D.C, pp 36-75.

Fookes, F.G. (1997). Geology for engineers: the geological model, prediction and
performance. Quarterly Journal of Engineering Geology, Vol. 30, pp 293-424.

ERM-Hong Kong Ltd. (1998). Feasibility Study for QRA of Boulder Fall Hazard in Hong
Kong. GEO Report No. 80, Geotechnical Engineering Office, Hong Kong, 61 p.

Hsu, K. (1975). Catastrophic debris streams (surtzstrons) generated by rockfalls. Bulletin

of Geological Society of America, Vol. 86, pp 129-140.

Hungr, O., Morgan, G.C. & Kellerhals, R. (1984). Quantitative analysis of debris torrent
hazards for design of remedial measures. Canadian Geotechnical Journal, Vol. 21,
pp. 663-677.

Hungr, O., Evans, S.C., Bovis, M.J. and Hutchison, J.N. 2001. A review of the classification
of landslides of the flow type. Environmental & Engineering Geoscience. Vol. VII,
pp 221-238.

Hutchinson, J. N. (1988). General Report: Morphological and geotechnical parameters of

landslides in relation to geology and hydrogeology. In: B, Proceedings of the Fifth
International Symposium on Landslides (Edited: Bonnard, C.), Lausanne, Switzerland,
Vol. 1, pp 3-35.

IAEG (1990). Suggested nomenclature for landslides. Bulletin of the International

Association of Engineering Geology, Vol. 41, pp 13-16.

Pierson, T.C. & Costa, J.E. (1987). A rheological classification of sub-aerial sediment-water
flows. In: Debris Flows/Avalanches: Process, Recognition and Mitigation (Edited:
Costa, J.E. & Wieczorek, G.F.), Geological Society of America, Boulder, Colorado,
pp 1-12.

Sassa, K. (1987). The Jizkukiyama Landslide and the interpretation of its long scraping
motion. In: Proceedings of the 5th International Conference and Field Workshop on
Landslides (Edited: Sloane, D.J. & Bell, D.H.), Christchurch, New Zealand,
pp 215-223.

Scheidegger, A. (1973). On the prediction of the reach and velocity of catastrophic

landslides, Rock Mechanics, Vol. 5, pp 231-236.
 - 89 -

Strahler, A.N. (1952). Hypsomertic (area altitude) analysis of erosional topography.

Bulletin of Geological Society of America, Vol. 63, pp 1117-1142.

VanDine, D.F. (1985). Debris flows and debris torrents in the southern Canadian Cordillera.
Canadian Geotechnical Journal, Vol. 22, pp 44-68.

Varnes D.J. (1978). Slope Movement Types and Processes. In: Landslides Investigation
and Mitigation (Edited: Turner, A.K. and Schuster, R. L.), Special Report 247 of the
Transportation Research Board, National Research Council, National Academy Press,
Washington, D.C, pp 11-33.

Wieczorek, G.F. (1996). Landslide Triggering Mechanisms. In: Landslides Investigation

and Mitigation (Edited: Turner, A.K. and Schuster, R. L.), Special Report 247 of the
Transportation Research Board, National Research Council, National Academy Press,
Washington, D.C, pp 76-90.

Classification of Hillside Failures

Fugro Maunsell Scott Wilson Joint Venture (2000). Report on the Debris Flow at Sham
Tseng San Tsuen of 23 August 1999 - Findings of the Investigation. Geotechnical
Engineering Office, Hong Kong, 92 p.

Fugro Scott Wilson Joint Venture (1999). Detailed Study of the Natural Terrain Landslide
at No. 30 Pak Sha Wan on 9 June 1998. Landslide Study Report LSR 1/99,
Geotechnical Engineering Office, Hong Kong, 47 p.

Fugro Scott Wilson Joint Venture (1999). Detailed Study of the Landslide at The Outward
Bound School on 9 June 1998. Landslide Study Report LSR 7/99, Geotechnical
Engineering Office, Hong Kong, 88 p.

GEO (1996). Report on the Shum Wan Road Landslide of 13 August 1995 - Findings of the
Landslide Investigation (Vol. 2). Geotechnical Engineering Office, Hong Kong,
51 p.

Halcrow Asia Partnership (1998). Detailed Study of the Landslides at Tao Fung Shan
Christian Cemetery on 2 July 1997. Landslide Study Report LSR 2/98, Geotechnical
Engineering Office, Hong Kong, 46 p.

Halcrow Asia Partnership (1998). Detailed Study of the Landslide Near Lido Beach, Castle
Peak Road on 2 July 1997. Landslide Study Report LSR 8/98, Geotechnical
Engineering Office, Hong Kong, 38 p.

Halcrow Asia Partnership (1998). Detailed Study of the Landslides below Sha Tin Heights
Road on 4 June 1997. Landslide Study Report LSR 10/98, Geotechnical Engineering
Office, Hong Kong, 47 p.
 - 90 -

Halcrow Asia Partnership (1998). Detailed Study of the Landslide Behind Block 3 and 4
Wonderland Villas, Kwai Chung on 8 May 1997. Landslide Study Report LSR 11/98,
Geotechnical Engineering Office, Hong Kong, 30 p.

Halcrow Asia Partnership (1998). Detailed Study of the Landslide Opposite Shing On
Temporary Housing Area, Ma On Shan Road on 2/3 July 1997. Landslide Study
Report LSR 19/98, Geotechnical Engineering Office, Hong Kong, 35 p.

Halcrow Asia Partnership (1998). Detailed Study of the Landslides at Ka Tin Court, Shatin
on 2 July 1997. Landslide Study Report LSR 24/98, Geotechnical Engineering
Office, Hong Kong, 29 p.

Halcrow Asia Partnership (1998). Report on the Landslides at Hut No. 26 Kau Wah Keng
Upper Village of 4 June 1997 - Findings of the Landslide Investigation. Geotechnical
Engineering Office, Hong Kong, 47 p.

King, J.P. (1996a). The Tsing Shan Debris Flow. Special Project Report SPR 6/96,
Geotechnical Engineering Office, Hong Kong, 3 volumes, 427 p, 129 p & 166 p.

Wong, H.N., Chen, Y.M. and Lam, K.C. 1996. Factual Report on the November 1993
Natural Terrain Landslides in Three Study Areas on Lantau Island. Special Project
Report SPR 10/96, Geotechnical Engineering Office, Hong Kong, 21 p. plus 3
Appendices.
 - 91 -

LIST OF FIGURES

Figure Page
No. No.

A1 Schematic Diagram of a Typical Natural Terrain 92

Landslide Scar

A2 General Classification of Hillside Failures 93

 - 92 -

Figure A1 - Schematic Diagram of a Typical Natural Terrain Landslide Scar

 Disturbed-hillside failure Hillside failure involving
Natural hillside failure
Type A Type B registerable man-made features

Disturbed hillside
Natural hillside not Disturbed hillside locally
Classification

locally modified by Natural or disturbed hillside

modified by man-made modified by man-made man-made activities
activities activities
Non-registerable
Non-registerable cut and fill slopes
Debris cut and fill slopes
Debris
Debris
Debris

Registerable man-made Registerable man-made Registerable man-made Registerable man-made

feature feature feature feature

- 93 -
Example

1990 Tsing Shan debris flow (SPR 6/96)
1997 Wonderland Villas (LSR 11/98)
1998 Pak Sha Wan (LSR 1/99)
1998 Outward Bound School (LSR 7/99)
1999 Sham Tseng San Tsuen (Forensic)
1993 Lantau Trail (SPR 10/96) 1997 To Fung Shan Cemetery (LSR 2/98) 1994 Shum Wan Road (Forensic)
1997 Lido Beach (LSR 8/98) 1997 Shatin Heights Road (LSR 10/98) 1997 Ma On Shan (LSR 19/98)
1997 Ka Tin Court (LSR 24/98) 1997 Kau Wah Keng (Forensic)

Figure A2 - General Classification of Hillside Failures

 - 94

APPENDIX B

NOTABLE NATURAL TERRAIN LANDSLIDES IN RECENT YEARS

 - 95

CONTENTS

Page

No.

Title Page 94

CONTENTS 95

B.1 LIU POK, LO WU 96

B.2 LIDO BEACH, CASTLE PEAK ROAD 96

B.3 WONDERLAND VILLAS, KWAI CHUNG 96

B.4 OPPOSITE SHING ON TEMPORARY HOUSING AREA, 97

MA ON SHAN ROAD

B.5 PAK SHA WAN, SAI KUNG 97

B.6 OUTWARD BOUND SCHOOL, SAI KUNG 98

B.7 QUEENS HILL, BURMA LINES CAMP, FANLING 98

B.8 LUK KENG WONG UK, LUK KENG 98

B.9 7C BOWEN ROAD 99

B.10 NO. 37 MOUNT DAVIS ROAD 99

B.11 KAP LUNG, SHEK KONG 100

B.12 SHAM TSENG SAN TSUEN 100

B.13 NO. 92 TAK SHEK WU KIU TAU, PAT HEUNG 101

B.14 VTC POKFULAM SKILLS CENTRE, POKFULAM 101

B.15 THE 2000 TSING SHAN DEBRIS FLOW 102

B.16 LEUNG KING ESTATE, TUEN MUN 102

LIST OF FIGURES 104

 - 96

The information below presents the key features of recent notable natural terrain
landslides that have been studied and documented in a systematic manner. In addition to the
study reports, the main findings of some of these landslides are available on the GEO Hong
Kong Slope Safety Website, http://hkss.ced.gov.hk under Findings of Landslide Investigation.

B.1 LIU POK, LO WU

The landslide occurred in September 1993 (Figure B1a). It involved approximately

300 m3 of debris, which was channelized along an ephemeral streamline and partially
destroyed an abandoned school building.

The travel angle of the landslide debris was about 24. The failed material comprised
colluvium, which overlies slightly to moderately decomposed metamorphosed sandstone. In
parts the rupture surface comprises sandstone bedding planes dipping obliquely out of the
slope at 28.

Reference: King, J.P. (1997). Damage to Liu Pok School by a Natural Terrain Landslide.
Discussion Note DN 2/97, Geotechnical Engineering Office, Hong Kong, 66 p.

B.2 LIDO BEACH, CASTLE PEAK ROAD

The 2 July 1997 landslide occurred on a 15 m high (Figure B1b), partly modified
natural hillside and affected an adjacent fill slope that was formed in the 1950s. The
landslide involved two phases of detachment, with a source volume of approximately 750 m3,
predominately comprising completed decomposed granite with some fill and residual soil.
The debris travelled some 70 m and completely blocked the Castle peak Road. Eight people
were trapped and injured by the debris associated with the second phase of failure.

The travel angle of the debris was about 20, which reflects that the failure was mobile
when compared to other rain-induced landslides of similar volume in Hong Kong. The
landslide was probably triggered by an increase in groundwater pressure in the slope caused
by the effects of direct infiltration of rainwater and surface water ponding within an
abandoned property directly upslope of the landslide.

Reference: Halcrow Asia Partnership (1998). Detailed Study of the Landslide Near Lido
Beach, Castle Peak Road on 2 July 1997. Landslide Study Report LSR 8/98,
Geotechnical Engineering Office, Hong Kong, 38 p.

B.3 WONDERLAND VILLAS, KWAI CHUNG

The 8 May 1997 landslide occurred on the natural terrain behind Blocks 3 and 4 of the
development (Figure B1c). The landslide debris travelled about 200 m down the slope and
partially blocked one lane of Wah King Hill Road.

The landslide occurred at the head of a poorly defined drainage line and the travel
angle of the landslide was about 34. The landslide material predominately comprised
completely decomposed granite with the surface of rupture formed at the interface between
 - 97

PW 0/30 and PW 50/90. From aerial photographs, numerous relict scars are evident in the
drainage lines to the north of the failure.

Reference: Halcrow Asia Partnership (1998). Detailed Study of the Landslide Behind
Block 3 and 4 Wonderland Villas, Kwai Chung on 8 May 1997. Landslide Study
Report LSR 11/98, Geotechnical Engineering Office, Hong Kong, 30 p.

B.4 OPPOSITE SHING ON TEMPORARY HOUSING AREA, MA ON SHAN ROAD

The 2/3 July 1997 landslide, which had a source volume of about 3 000 m3, occurred in
a drainage line on a hillside that varies from 38-40 (Figure B1d). The failure was probably
initiated by a small-scale failure of a cut slope at the toe of the natural terrain that resulted in
retrogressive failure. It affected an abandoned cut slope and two fill platforms.

The surface of rupture was within completely decomposed granite, and inclined at
32-38. Part of the rupture surface was formed by a clay-coated discontinuity. The travel
angle was about 28. Numerous relict landslides were identified from aerial photographs.

The failure was associated with the very severe rainstorm on 2/3 July 1997. This
resulted in the highest maximum daily rainfall since the installation of the nearest rain gauge
in 1983. Based on record from the Hong Kong Observatory, the storm had an estimated
return period of about 1 000 years.

Reference: Halcrow Asia Partnership (1998). Detailed Study of the Landslide Opposite
Shing On Temporary Housing Area, Ma On Shan Road on 2/3 July 1997. Landslide
Study Report LSR 19/98, Geotechnical Engineering Office, Hong Kong, 35 p.

B.5 PAK SHA WAN, SAI KUNG

The 9 June 1998 landslide with a source volume of 300 m3 occurred in colluvium
within a drainage line (Figure B1e). The natural hillside slopes at about 30 and has no
evidence of past instability from the available aerial photographs. More than one phase of
landsliding occurred over a period of about two hours. The landslide deposited debris on the
platform of two village houses at the toe of the slope.

The debris travelled downslope in the form of a channelized debris flow and was
relatively mobile, with a travel angle of about 25. There is no evidence of past instability
from the available aerial photographs.

The failure was probably caused by the development of transient elevated water
pressures in the surface mantle of colluvium following a severe rainstorm (estimated return
period of about 100 years). Concentrated surface and subsurface water flows, leading to
water ingress into the landslide site, probably contributed to the failure.

Reference: Fugro Scott Wilson Joint Venture (1999). Detailed Study of the Natural Terrain
Landslide at No. 30 Pak Sha Wan on 9 June 1998. Landslide Study Report
LSR 1/99, Geotechnical Engineering Office, Hong Kong, 47 p.
 - 98

B.6 OUTWARD BOUND SCHOOL, SAI KUNG

The 9 June 1998 landslide occurred on natural hillside (Figure B1f). The landslide
source volume was 900 m3 and the debris was deposited within 20 m of a house, which was
temporarily evacuated. The failed material comprised completely decomposed volcanic tuff
and residual soil.

The failure occurred in an area with a distinct break in slope, probably the location of
an old landslide scarp. The failure, with a travel angle of 27, probably involved a
retreating hillside mechanism, i.e. successive discrete failures retrogressing uphill over a
period of time.

The landslide was triggered by a severe rainstorm (estimated return period of about 70
years). The presence of tension cracks as well as open joints suggests that the hillside was
progressively deteriorating with time. The presence of complex local geological and
hydrogeological regimes (partly due to faults traversing the hillside) was a contributory factor
to the landslide.

Reference: Fugro Scott Wilson Joint Venture (1999). Detailed Study of the Landslide at The
Outward Bound School on 9 June 1998. Landslide Study Report LSR 7/99,
Geotechnical Engineering Office, Hong Kong, 88 p.

B.7 QUEENS HILL, BURMA LINES CAMP, FANLING

This incident involved extensive distress on natural hillside (Figure B1g), with a
displaced mass of over 1 500 m3. The identification of distress on the hillside resulted in the
evacuation of a number of dwellings at the toe of the hillside.

The distress predominantly involved slope movement, tension cracks and localised
detachment, affecting both the colluvium and, in parts, the underlying completely
decomposed volcanic tuff. Aerial photographs suggest that some of the distress dates back
to 1924. The hillside instability was not identified during the detailed study of the cut slope
at the toe of the natural hillside. There is evidence, over parts of the hillside, that
progressive ground movement and deterioration of the ground condition with time develop
into more mobile failures.

Reference: Fugro Scott Wilson Joint Venture (1999). Detailed Study of Slope Distress at
Queens Hill, Burma Lines Camp, Fanling. Landslide Study Report LSR 10/99,
Geotechnical Engineering Office, Hong Kong, 102 p.

B.8 LUK KENG WONG UK, LUK KENG

The 24 May 1998 landslide (Figure B1h), which had a source volume of 150 m3,
occurred on a natural hillside with a history of failures. It affected the open areas of some
dwellings within the village at the toe of the hillside. The failed material mainly comprised
highly to completely decomposed siltstone.
 - 99

The landslide debris travelled along a stream course as a channelized debris flow, and
the travel angle was about 23. The landslide occurred under moderately heavy rainfall
(estimated return period of 8 years).

Reference: Fugro Scott Wilson Joint Venture (1999). Preliminary Study of the Landslide at
Luk Keng Wong Uk on 24 May 1998. Landslide Study Report LSR 14/99,
Geotechnical Engineering Office, Hong Kong, 33 p.

B.9 7C BOWEN ROAD

The 11 June 1998 landslide affected a portion of natural hillside (Figure B1i) with no
evidence of past instability from the available aerial photographs. The landslide debris was
deposited at the rear of a building at Bowen Road and some parts of the building suffered
minor damage. As the run-out of the landslide debris was constrained by a building, the
travel angle of the landslide debris cannot be determined reliably. The failed material
comprised colluvium.

The rainfall preceding the landslide (estimated return period of 8 years), together with
leakage from an exposed water main along the access road above the crest of the hillside,
probably contributed to the landslide.

Reference: Fugro Scott Wilson Joint Venture (1999). Detailed Study of the Landslide
behind 7C Bowen Road on 11 June 1998. Landslide Study Report LSR 19/99,
Geotechnical Engineering Office, Hong Kong, 61 p.

B.10 NO. 37 MOUNT DAVIS ROAD

The 24 August 1999 failure was a near-miss incident in that the debris narrowly
missed an apartment block below Mount Davis Road (Figure B1j). The incident involved
the fracture of a 150 mm diameter fresh water pipe and undermining of a 1.4 m diameter fresh
water main located at the crown of the landslide scar. The latter is an important trunk main
of the Mount Davis Service Reservoir, supplying potable water to the Hong Kong South area.

The rainfall that triggered the landslide was not severe (estimated maximum return
period of about 4 years). The surface runoff from a significant catchment above the
landslide site was channelled along a road due to a blocked catchpit, over-spilling the road
onto the hillside near a bend.

Inadequate provision and maintenance of the surface water drainage provisions above
the access road and the lack of surface drainage or an upstand on the downhill side of the
access road were probably key contributory factors in causing the failure.

Reference: Fugro Maunsell Scott Wilson Joint Venture (2000). Initial Review of the
Landslide above No. 37 Mount Davis Road on 24 August 1999. Geotechnical
Engineering Office, Hong Kong, 52 p.
 - 100

B.11 KAP LUNG, SHEK KONG

This August 1999 landslide occurred on a natural hillside (Figure B1k) beneath an
electricity transmission pylon situated in a remote area. Extensive vegetation clearance was
carried out in the vicinity of the pylon during its construction in 1995. The vegetation
remained sparse at the time of failure.

The landslide was in the form of a channelized debris flow. The relatively low travel
angle of the landslide debris (15) indicates that the channelized debris flow was mobile,
possibly involving a large volume of surface water running down the stream course.

The time of failure is not known. The hillside was previously assessed and checked
to the required standards, as part of the foundation design for the pylon. However, the
stability assessment did not consider a sharp convex break of slope and an area with a steeper
ground profile, which probably corresponds to previous instability.

Reference: Fugro Maunsell Scott Wilson Joint Venture (2000). Detailed Study of the
Landslide above Kap Lung, Shek Kong, during the Severe Rainstorms between 22 and
24 August 1999. Landslide Study Report LSR 3/2000, Geotechnical Engineering
Office, Hong Kong, 52 p.

B.12 SHAM TSENG SAN TSUEN

On 23 August 1999 four landslides occurred on a natural hillside above Sham Tseng
San Tsuen (Figure B1l). The landslide debris demolished a house and its detached kitchen
structure, and severely damaged several other dwellings within the squatter village. It
resulted in one fatality and thirteen injuries.

The largest landslide, which was the primary source of the debris flow, had a source
volume of 600 m3. The failure occurred at the location of a very old landslide. Much of
the colluvium at the landslide location probably originated from a lower landslide scarp,
which has progressed up the hillside over a long period of time as a result of weathering and
intermittent small- to large-scale movements of the colluvium. The landslide developed into
a channelized debris flow down a stream course, with a travel angle of about 24. The debris
flow occurred during a severe rainstorm, with an estimated maximum return period of about
49 years. The subject hillside suffered numerous failures (>10 notable failures) during
severe rainstorms in 1982. The August 1999 rainstorm resulted in seven notable landslides,
six of which were at different locations compared to the failures in 1982.

The landslides were probably caused by the build-up of elevated water pressures
within the surface colluvium. The thin mantle of bouldery colluvium overlying relatively
less permeable colluvium is favourable to the build-up of perched water tables during heavy
rain. The presence of relict failures suggests that the rock mass has been disturbed and
weakened. This in turn may have adversely affected the local hydrogeology by promoting
direct infiltration of rainwater through the open or partially infilled joints. A hill fire about 4
to 5 years before the failure had also removed much of the vegetation cover. There is
evidence of progressive deterioration of the hillside with discontinuities opening up with time,
probably associated with past intermittent slope movements.
 - 101

Reference: Fugro Maunsell Scott Wilson Joint Venture (2000a). Report on the Debris Flow
at Sham Tseng San Tsuen of 23 August 1999 - Findings of the Investigation.
Geotechnical Engineering Office, Hong Kong, 92 p.

B.13 NO. 92 TAK SHEK WU KIU TAU, PAT HEUNG

The 24 August 1999 landslide, with a 260 m3 source volume, occurred on a natural
hillside above No. 92 Tak Shek Wu Kiu Tau near Pat Heung (Figure B1m). The landslide
debris became channelized and was deposited on a platform immediately adjacent to a
residential building near the toe of the hillside.

The travel angle of the landslide debris was about 20, which indicates that the debris
was relatively mobile. The channelized debris flow was probably triggered by rainfall, with
an estimated return period of about 25 years. The landslide occurred towards the top of an
ephemeral drainage line, with its main scarp developed along a pre-existing tension crack
(which was noted in the 1996 and 1998 aerial photographs). The tension crack was located
below two relict landslide scars. Other relict landslide scars were identified within a number
of depressions on the hillsides in the area.

The presence of rafts of relatively intact material within the source area, together with
the exposure of planar surfaces on the surface of rupture, suggest that the initial instability
involved a shallow sliding failure (up to 1 m deep). In addition, the location of the landslide
towards the top of the drainage line suggests that the failure may be part of the
geomorphological development of the hillside. There is evidence of progressive
deterioration of the hillsides.

Reference: Fugro Maunsell Scott Wilson Joint Venture (2000). Review of the 24 August
1999 Landslide on the Hillside above No. 92 Tak Shek Wu Kiu Tau near Pat Heung.
Geotechnical Engineering Office, Hong Kong.

B.14 VTC POKFULAM SKILLS CENTRE, POKFULAM

On 23 August 1999 two landslides (Figure B1n) occurred within the same local
catchment of the hillside, which has a history of instability. One landslide (100 m3) involved
unauthorised end-tipped fill whilst another (200 m3) involved natural terrain with the source
material comprising colluvium and weathered volcanic tuff. As a result of the landslide,
Victoria Road was temporarily closed to traffic.

Debris from both landslides became channelized in an ephemeral stream course and
reached Victoria Road, with a travel angle of about 25. The natural terrain failure occurred
in an area showing evidence of progressive deterioration of the ground condition, in the form
of localised tension cracking and ground movement.

The rainstorm that triggered the landslides was not particularly severe (estimated
maximum return period of less than two years). The landslides were probably caused by the
build-up of transient elevated groundwater pressure in the soil mass following the rainstorm.
Probable contributory factors to the failures were the blockage of surface drainage provisions
 - 102

above one of the landslides, changes in slope gradient caused by pervious local instability,
and possible deterioration of ground conditions.

Reference: Fugro Maunsell Scott Wilson Joint Venture (2000). Detailed Study of the
Landslides below VTC Pokfulam Skills Centre, Pokfulam Road on 23 August 1999.
Landslide Study Report LSR 7/2000, Geotechnical Engineering Office, Hong Kong,
99 p.

B.15 THE 2000 TSING SHAN DEBRIS FLOW

On 14 April 2000 a landslide originated at about 360 mPD on the rocky upper slopes
of Tsing Shan and extended down to a spur at about 180 mPD where it divided to form
channelized debris flows in the two drainage lines on either side of the spur (Figure B1o).
The source volume was about 150 m3. A significant amount of materials (colluvium and
weathered rock) was eroded along the travel path and the estimated total volume of the
natural terrain landslide about 1 600 m3.

The event is interpreted to have started in the rocky gully as a debris avalanche that
continued down over the open colluvial slope. At the spur some debris flowed down the
southern drainage line while a bouldery lobe accumulated on the spur at a power transmission
pylon before flowing down the northern drainage line. Channelized debris flows occurred in
both drainage lines and the travel angles for the northern and southern branches are 27 and
24 respectively.

The trigger for the initial landslide was probably elevated pore water pressure in the
bouldery colluvium and weathered profile of the source area. However, it is worth noting
that the estimated maximum return period for the rainfall was less than two years. Besides
the steep slope gradient (about 40), the environmental factors that pre-disposed the
generation of the debris flow were the accumulation of loose bouldery deposits on the steep
slope above well-incised colluvium-lined drainage lines that serve to channelize the debris
and contribute additional material.

Reference: King, J.P. (2001). The 2000 Tsing Shan Debris Flow. Landslide Study Report
LSR 3/2001, Geotechnical Engineering Office, 54 p. plus 1 drg.

B.16 LEUNG KING ESTATE, TUEN MUN

On 14 April 2000 during heavy rain, a series of landslides occurred on the hillside
above Leung King Estate, Tuen Mun. One of the natural terrain landslides involved about
600 m3 of material (source volume 450 m3 and entrained volume 150 m3) which developed
into a channelized debris flow (Figure B1p) which came to rest close to the toe of the hillside.
The subsequent outwash blocked the drainage at the crest of a cut slope above Leung King
Estate, causing inundation of mostly silty sand into the northern portion of the Estate.

The source area of the landslide is a natural depression, located at the head of a minor
natural drainage line below a granite cliff. The surrounding natural hillside slopes at 30 to
40. The failed material comprised colluvium and weathered profile including slightly
 - 103

decomposed rock. The surface rupture generally followed a series of undulating sheeting
joints orientated sub-parallel with the ground surface.

A notable feature about the landslide is the probable formation and subsequent
breaching of a debris dam within the natural drainage line. Landslide debris that formed the
dam had a travel angle of about 27. The breaching of the debris dam resulted in a travel
angle of 17 for the debris flow, which is indicative of a mobile failure involving fluidized
debris.

A number of recent and relict landslides have been identified from aerial photographs
in the vicinity, most of them of small size. There is evidence of disturbance within the rock
mass exposed at the landslide, including open sediment infilled joints, offset joints, offset
quartz veins and fractured rock.

Reference: Halcrow China Ltd. (2001). Detailed Study of Selected Landslides above Leung
King Estate of 14 April 2000. Landslide Study Report LSR 9/2001, Geotechnical
Engineering Office, Hong Kong, 140 p.
 - 104

LIST OF FIGURES

Figure Page
No. No.

B1a-d Notable Natural Terrain Landslides in Recent Years 105

B1e-h Notable Natural Terrain Landslides in Recent Years 106

B1i-l Notable Natural Terrain Landslides in Recent Years 107

B1m-p Notable Natural Terrain Landslides in Recent Years 108

 - 105

(a) Liu Pok, Lo Wu (b) Lido Beach, Castle Peak Road

(c) Wonderland Villas, Kwai Chung (d) Opposite Shing On Temporary Housing
Area, Ma On Shan Road

Figure B1a-d - Notable Natural Terrain Landslides in Recent Years

 - 106

(e) Pak Sha Wan, Sai Kung (f) Outward Bound School, Sai Kung

(g) Queens Hill, Burma Lines Camp, (h) Luk Keng Wong Uk, Luk Keng
Fanling

Figure B1e-h - Notable Natural Terrain Landslides in Recent Years

 - 107

(i) 7C Bowen Road (j) No. 37 Mount Davis Road

(k) Kap Lung, Shek Kong (l) Sham Tseng San Tsuen

Figure B1i-l - Notable Natural Terrain Landslides in Recent Years

 - 108

(m) No. 92 Tak Shek Wu Kiu Tau, (n) VTC Pokfulam Skills Centre, Pokfulam
Pat Heung

(o) The 2000 Tsing Shan Debris Flow (p) Leung King Estate, Tuen Mun

Figure B1m-p - Notable Natural Terrain Landslides in Recent Years

 - 109

APPENDIX C

GEOMORPHOLOGY AND NATURAL TERRAIN HAZARDS

 - 110

CONTENTS

Page
No.

Title Page 109

CONTENTS 110

C.1 INTRODUCTION 111

C.2 ROCK TYPE AND STRUCTURE 112

C.2.1 Introduction 112

C.2.2 Volcanic Rocks 112

C.2.3 Granitic Rocks and Rhyolitic Dykes 112

C.2.4 Sedimentary Rocks 113

C.2.5 Structures 113

C.3 WEATHERING 114

C.4 REGOLITH 115

C.5 HYDROGEOLOGY 115

C.5.1 Groundwater Conditions 115

C.5.2 Piping 116

C.6 MASS MOVEMENTS 117

C.6.1 Introduction 117

C.6.2 Interpretation from Aerial Photograph 117

C.6.3 Landslide Types 118

C.6.3.1 General 118

C.6.3.2 Rock and Boulder Falls 118

C.6.3.3 Slumps to Debris Flows 118

C.6.3.4 Large Relict Landslides 121

C.7 REFERENCES 121

LIST OF FIGURES 126

 - 111

C.1 INTRODUCTION

The morphology of the present day landscape of Hong Kong is generally considered to
have developed mainly during the Quaternary Period (1.4 million years ago to the present),
but it records processes that probably began during the Tertiary (Ruxton & Berry, 1957),
perhaps as long as 60 million years ago.

The last glaciation commenced 25 000 years ago, reaching a maximum 17 000 to
18 000 years ago, when sea level was at around -120 mPD or -130 mPD. At that time, the
coastline was about 120 km south of Hong Kong (Fyfe et al., 2000). When global
deglaciation began about 11 000 to 12 000 years ago, the sea level rose and a new coastline
was established. A postglacial sea level high, possibly 1 to 3 m above its present level, was
attained about 7 000 years ago. Substantial erosion occurred at the new coastline(s). This
involved, and continues to involve, coastal landslides, and reactivation of relict coastal
landslides, as a major component of modification of the landscape. Mass movements have
dominated the erosion of mountain slopes, particularly during wetter periods, which are
postulated at about 12 500 to 11 000, 10 500 to 8 000 and 7 000 to 5 500 years ago.

Tregear (1958) compiled the first regional description of the geomorphology of Hong
Kong. Later, Berry and Ruxton (1960) carried out a detailed study of the evolution of the
Victoria Harbour area. Berry (1961) was the first to recognize erosion surfaces in Hong
Kong, identifying planation surfaces at around 130 mPD and 230 mPD, with concordant
summit levels at about 480 mPD. Other regional geomorphological summaries have been
compiled by Hansen (1984) and So (1986). The agricultural soils of Hong Kong were
mapped and described by Grant (1962, 1986), who also discussed soil development and
erosion (Grant, 1968) in the context of the geomorphology of the area.

The landscape of Hong Kong has a distinctive drowned appearance with an intricate
coastline. This is particularly noticeable in the east where the hillslopes plunge directly into
the sea, without the development of coastal plains or extensive wave cut platforms that would
indicate prolonged coastal erosion. The limited erosion that has occurred has only succeeded
in over-steepening the hillslopes to form steep fringing cliffs. Similarly, most of the drainage
lines are short and appear to have been truncated. Extensive deposition in the drowned coastal
extremities of the valleys has formed distinctive alluvial flats around the coastline.

As is typical of subtropical upland areas, current geomorphological processes are

dominated by mass movements and large-scale, but intermittent, erosional events. The
effects of these periodic processes are superimposed on locally deeply weathered profiles.
Weathering has produced a smooth, rounded topography with a thick regolith cover on the
ridge tops and side slopes, incised by the valleys of a slowly evolving drainage system.
Slope failures and fluvial erosion add an irregular and angular overprint of failure scars, free
faces, and rocky ridgelines.

Mass movements are currently the most important geomorphological processes in

Hong Kong. They have displaced much of the weathered mantle on hillslopes, forming
extensive colluvial deposits on the shallower slopes, and large debris sheets and lobes on the
steeper slopes and in the steep tributary valleys. Fluvial processes of rilling, gullying and
stream flow have generally played a lesser role, primarily serving to progressively erode the
matrix from weathered profiles and to rework mass movement debris.
 - 112

C.2 ROCK TYPE AND STRUCTURE

C.2.1 Introduction

Three main rock types make up the geology of Hong Kong. These are, in order of
importance, volcanic rocks, granitic rocks and sedimentary rocks. Volcanic and granitic rocks
together crop out over about 70% of the area. Faults control the major topographical grain,
largely determining the orientations of the major valleys and ridges. In particular, northwest-
and east-northeast- to east-trending faults have been preferentially weathered and eroded,
forming distinctive linear depressions. Late Pleistocene and Holocene superficial deposits
form a discontinuous cover, with colluvium on hillsides and alluvium on the valley floors.
The geology of Hong Kong is described in Sewell et al. (2000) and Fyfe et al. (2000).

C.2.2 Volcanic Rocks

Volcanic rocks are exposed over about 48% of the land area of Hong Kong and
generally form the higher, more angular and rugged uplands. These rocks occur over much
of the New Territories, especially in eastern Hong Kong and most of the country parks,
Lantau Island, and southern Hong Kong Island.

Volcanic terrain, which commonly includes interbedded sedimentary units, is typically

angular, characterized by moderate to steep slopes, gently concave sideslopes and narrow
convex ridges. Bedrock exposures are a feature of ridgelines, with free faces occurring on
spur ends and steeper slopes. Large boulders are common, littering slopes below these
exposures. Commonly, a thin veneer of bouldery colluvium mantles the slopes. Volcanic
rocks underlie the highest peaks and ridges of Hong Kong. Tai Mo Shan, Sunset Peak,
Lantau Peak, Fei Ngo Shan to Tates Cairn, Ma On Shan, High Junk Peak, and Sharp Peak all
form prominent, angular facets of the landscape that stand in sharp contrast to the more
rounded forms of the granitic hills. Similarly, volcanic rocks tend to form the steeper,
cliffed coastlines, particularly in the east.

Massive, coarse ash tuff weathers in an analogous way to the granitic rocks, with
moderately thick weathered profiles, corestones, and uplands with generally smoothed
outlines. Fine ash tuff and eutaxite have locally pronounced bedding structures or bedding
parallel foliation. They weather less deeply, rarely develop corestones, and produce shallow,
clayey soils on very angular uplands. The eutaxitic rocks display a pronounced layered
structure that tends to form escarpments and distinctive topographical features. Evidence of
mass movements, both recent and relict, is common in the shallow, clayey weathering
residues that are characteristic of the volcanic rocks. Erosion gullies are rarely developed, in
contrast to the deeply weathered granitic rocks. Colluvial deposits, derived from the shallow
weathered profiles, are thinner than over the more deeply weathered granitic rocks.

C.2.3 Granitic Rocks and Rhyolitic Dykes

Granitic rocks are exposed over about 28% of Hong Kong, with most of the main
urban areas of Hong Kong built on these rocks. Granite topography is distinguished by low
rounded hills, with a more subdued topography than that developed on the volcanic rocks.
Weathered mantles and soils are generally thicker than on the volcanic lithologies and there is
 - 113

generally less evidence of mass movements, although gully erosion is more common. The
distinctive scenery of the granitic rocks characterizes many areas of Hong Kong, in particular
the western New Territories, the central New Territories from Tsuen Wan to Sha Tin, the
Kowloon Peninsula and Lion Rock ridge, the northern and central parts of Hong Kong Island
from Happy Valley to Shau Kei Wan, and the southern parts of Lamma and Lantau islands.

Granite landscapes typically exhibit a subdued relief, comprising rounded summits

with long convex upper slopes and a relief amplitude of generally less than 70 m. Deep
weathering is a distinctive feature, with most areas underlain by a thick, granular weathered
mantle. Deep gullies are a distinctive feature in some areas of granitic terrain, with many
unvegetated ridge crests and side slopes being extensively gullied. Tors are common
features in granite landscapes. They occur as prominent exposures of jointed, relatively
unweathered granite in a variety of topographical situations, as hill summit, ridge crest, spur
end and valley side features. Granite boulders, which are typically rounded, litter the
surrounding slopes. The boulders originate as exhumed corestones, either by erosion of the
finer weathered matrix on slopes, or by toppling from cliff faces and tors. Boulder fields are
common on summit plateaus, the sides of ridges, and on open slopes, but characteristically
choke valley lines as distinctive boulder trains.

Mass movements are less common in areas of weathered granite than on the more
clayey, less freely drained volcanic rocks. However, colluvial deposits are generally thicker
over the granites, reflecting derivation from deeper weathered profiles.

Large-scale sheet jointing is common, paralleling the smooth hillslopes. Differential

weathering of the feldsparphyric dykes in the northeast of Lantau Island has produced a
distinctive lineated topography with prominent ribs, ridges and ledges that are oriented in an
east-northeast direction. These ridges comprise coarse, joint-bounded blocks that topple and
collapse to feed debris streams and sheets on the flanking slopes.

C.2.4 Sedimentary Rocks

Sedimentary rocks are exposed over only about 5% of the Hong Kong onshore area.
They occur mostly in the northern and northeastern New Territories, where they form the
intricate and varied scenery of the outlying islands.

Pronounced bedding structures impart a clear grain to the topography, forming

continuous ridges that support a trellised drainage pattern. Vertically dipping rocks, such as
at Bluff Head, form a distinctive ridgeline along the northern shore of the Tolo Channel.
Sedimentary rocks, in general, have a higher incidence of landslide scars than the other
lithologies (Evans & King, 1998).

C.2.5 Structures

Geological structures occur on a wide range of scales. They can be distinguished at

the regional, local and outcrop level, and include faults, joints, and bedding planes.
Weathering and erosion have exploited these structural discontinuities to produce a distinctive
grain or fabric to the topography and to individual bedrock exposures. The regional grain is
 - 114

reflected in the form of the hills and the outline of the coast. Differential weathering of
exposed rock outcrops commonly emphasizes the rock structure or fabric, producing a range
of geometrical patterns.

The dominant structural trend in Hong Kong is aligned northeast, varying to

east-northeast. A subordinate, orthogonal northwest trend is also present and becomes
dominant in east and southeast Hong Kong. Alignments of the main valleys, ridges and
uplands show a broadly rectilinear pattern that reflects the effects of weathering and erosion
along these main structural trends, as does the coastal estuaries, channels, bays and headlands.

C.3 WEATHERING

There is a close correlation between the topography of the weathering front and the
main structural trends in Hong Kong (e.g. Shaw, 1997; Owen & Shaw, 2000). Geological
structures determine local hydrogeology, and consequently the intensity of weathering and
eluviation. In most slopes, the base of the weathered mantle, or weathering front (i.e.
rockhead in engineering use), is subparallel to the topography. Generally, the weathering
front tends to rise nearer the ground surface below hills, which usually occur in less
well-jointed rocks, although it is not unusual to find a thick weathering profile preserved
below ridge crests. In many areas the weathering front is stepped, a phenomenon that
depends on the geometry of the joint pattern and its infilling by impermeable clay minerals.
The steps (varying from millimetres to 20 m or more) commonly coincide with subvertical or
steeply inclined faults, or major joints. The weathering front commonly descends abruptly
below fault-controlled valleys as a result of preferential weathering along the discontinuity.
The linear, discontinuity-controlled zones of deep weathering are easily eroded to create the
main valleys. However, most of the minor stream valleys have rock floored channels.

Geomorphologically, the weathering profile can be subdivided into several broad

zones. The term saprolite is usually applied to the lower, sedentary part of the profile, that
has not undergone marked volume changes and preserves original structures, such as joints
and quartz veins, features directly inherited from the parent rock. Above the saprolite zone,
towards the ground surface, there is usually a mobile zone of slightly transported material.
This shows evidence of disturbance by gravitational creep, settling, and bioturbation, notably
by termites and plant roots. The mobile zone is commonly subject to extensive leaching of
soluble components and eluviation of the fines. Loss of rock texture, the disturbance of
structures such as relict joints, the development of stone lines, and the appearance of outcrop
curvature, are features that distinguish this portion of the profile. However, in most
weathered profiles in Hong Kong, the mobile zone is usually only a few metres thick.

Within a given rock type, the depth of weathering is primarily a function of the depth
to which oxygenated surface water penetrates and circulates through the rock mass. As
discontinuities are the only permeable pathways in most igneous rocks, regional and local
patterns, spacing and openness of discontinuities in the rock mass largely determine the
groundwater regime, and thus the character of the weathered profile, which ultimately
influences the subsequent topography.

The thickest weathered profiles in Hong Kong are generally developed in uniform
granitoid rocks, ranging from a few metres to a several tens of metres. For example,
 - 115

weathering depths over the Kowloon peninsula vary from about 10 metres over the relatively
high areas, to greater than 40 metres in the low lying districts (Shaw, 1997). In contrast,
fine-grained tuff, rhyolite, and sedimentary rocks tend to generate relatively shallow
weathering profiles of no more than a few metres thick, and with few corestones.

Zones of particularly deep weathering are usually associated with faults, irrespective of
the lithology. Studies have shown that, at its deepest point, weathering of volcanic rocks in the
fault-controlled Tuen Mun valley extends to a depth of -86 mPD with a weathering profile over
80 metres thick (Owen & Shaw, 2000). Similarly, fault controlled weathering at Tung Chung
New Town has resulted in a weathering profile more than 170 m thick (Fyfe et al., 2000).

Weathering also affects colluvium and alluvium. Because the degree of weathering
increases with the age of the deposit, features such as the thickness of the weathering rind
around colluvial clasts can be used to determine the relative age of these deposits (Lai &
Taylor, 1984).

C.4 REGOLITH

Colluvium blankets many slopes although, commonly, it is too thin to have been
mapped previously. The Hong Kong Geological Survey regional geological maps at
1:20 000-scale show only superficial deposits greater than 2 m thick. Colluvium was
generally classified as debris flow deposits, although talus was recognized in specific
localities. Mapped deposits of colluvium are thickest on the footslopes. They form
extensive deposits over the base of the Lion Rock ridge, the Mid-levels area, the slopes of Fei
Ngo Shan, and other large hill masses in the New Territories, such as below Castle Peak, Tai
Mo Shan and on Lantau Island.

With a detailed interpretation of aerial photographs followed by field verification, it is

possible to develop a process-based colluvium classification based on its morphology and
topographical location. For example, relatively recent colluvium can be subdivided into
rockfall debris below rock outcrops, valley colluvium along drainage lines, and recent
landslide debris (Figure C1); older colluvium can be subdivided into residual colluvium
which has lost its original form due to subsequent weathering and erosion, and relict landslide
deposits which retains its morphological characteristics such as fan-like or lobate form
(Figure C1). It is also possible to identify corestone bearing saprolite, non-corestone bearing
saprolite, or rock at or near ground surface (Figure C1). This type of regolith mapping is
proving useful in studying landslide initiation factors in the on-going area study of natural
terrain hazard study for the Tsing Shan foothills.

C.5 HYDROGEOLOGY
C.5.1 Groundwater Conditions

In Hong Kong groundwater conditions are largely a function of the seasonal rainfall
patterns, the geology and the local topography. Discontinuities, such as faults and joints, act
as important conduits for the movement and storage of groundwater. In general, the igneous
rocks have a low porosity that may be increased by tectonic fracturing and the opening of
joint sets by weathering and related stress relief. Consequently, groundwater movement and
 - 116

storage is essentially a function of the spacing and aperture of the discontinuities.

Commonly, the rise in the groundwater table during a major rainstorm is of the order
of several metres and, in extreme cases, may be up to 20 m or more (e.g. Sun & Campbell
1998; Nandy & Jiao, 2001). Groundwater table response may be rapid, and occur broadly
synchronously with the rainstorm, or may be delayed for periods of up to several days,
reflecting the complexity of the hydrogeological regimes and the distance travelled from the
point of infiltration.

Transient elevation of pore pressure, during and following heavy rainfall, is the main
cause of slope failure in Hong Kong. The four main hydrogeological factors, consistently
identified as contributing to landsliding, are: (i) the direct rainwater infiltration forming a
wetting front; (ii) the development of perched water tables, often in association with relatively
impermeable, persistent clay seams (e.g. GEO, 1996a); (iii) throughflow along high
permeability layers or soil pipes subparallel to the ground surface creating seepage pressures
and, if the conduits are blocked, a rapid and possibly very high local increase in pore water
pressure; and (iv) a general rise in the groundwater table at a site due to artesian seepage from
bedrock at the toe of large hill masses. The latter is commonly a delayed response that may
occur several days after heavy rainfall. Individually or in combination, these mechanisms
can cause a rapid and extreme elevation of the groundwater table (e.g. Sun & Campbell, 1998;
Nandy & Jiao, 2001).

Natural vegetation distribution is a function of the slope hydrology, which is a function

of geomorphology, and specifically the slope aspect, slope angle, exposure, and elevation,
which influence microclimate and soil development at a particular site. Mans influence,
particularly the role of frequent hill fires during the dry season (October to April), continues
to modify these natural patterns.

C.5.2 Piping

Soil piping, or natural tunnel erosion, is widespread in Hong Kong, and has been
reviewed by Nash and Dale (1984), and Hansen and Nash (1985). A comprehensive model
for soil piping in Hong Kong has been developed by Nash and Dale (1984). Soil pipes are
essentially preferential flow paths for groundwater. They evolve progressively due to
eluviation leading to interconnection of voids and further erosion and expansion of the pipe.
Most reported examples in Hong Kong are less than 100 mm diameter, and occur within 1 m
of the surface. They commonly occur within superficial deposits, most notably colluvium
and residual soils. However, Whiteside (1996) has demonstrated the presence of a network
of pipes at the base of a more than 10 m thick regolith.

Initiation of soil pipes at the surface can be attributed to various features that facilitate
water ingress. These include shallow surface depressions, desiccation cracking, in particular
of clay and organic rich soils, hollows and tubes formed by decayed tree roots, animal
burrows, and even insect nests. In addition, factors such as excavation, tension cracks
associated with landsliding, and the removal of vegetation cover can also promote soil pipe
development. Piping commonly occurs at textural or permeability boundaries in the regolith,
including colluvium-in situ soil boundaries and boundaries within the weathering profile.
 - 117

Shallow soil pipes are most common in the coarser-grained granitic and volcanic rocks,
generally in corestone-bearing profiles, and other highly permeable soils, including gravel and
cobble rich colluvium. Erosion within shallow pipes can lead to collapse of their roofs, and
to the initiation of gully erosion. Shallow pipes may coalesce and feed into deeper seated,
more substantial pipes in completely to highly decomposed in situ material. These deeper
pipes are influenced in their development by relict joints, especially within coarser-grained
rocks (Nash & Dale, 1984). The deeper-seated pipes (e.g. exposed in landslide main scarps)
may be up to 1 m in diameter. Where open-ended and free draining, the soil pipes may
provide efficient drainage within the slope. However, soil pipes may be closed, have
become blocked, or their capacity exceeded, during a period of extreme rainfall. In such
circumstances, they have been implicated at many locations in Hong Kong as having had an
adverse influence on slope stability because they lead rapidly to very high pore pressures (e.g.
Nash & Dale, 1984; Koor & Campbell, 1998; GEO, 1998; Whiteside, 1996). Infilling of
pipes by sediment may be partial (Nash & Dale, 1984) or complete (GEO, 1998).

C.6 MASS MOVEMENTS

C.6.1 Introduction

In Hong Kong, deep weathering profiles, steep topography and high rainfall have
resulted in mass movements being an important landscape forming process. Landslide scars,
which range in size from small depressions to valley-side amphitheatres, are a common
feature of natural hillsides (Evans et al., 1997). Together, they probably represent a
landslide history spanning at least the Holocene Period (the last 10 000 years). Recent scars
are largely unvegetated and can be easily and unambiguously identified either in the field or
on aerial photographs, a record that covers most of the territory from 1945 or 1963 to the
present. Relict scars retain a sharp morphology, but are completely overgrown with
vegetation. Most of the larger scars, which have a degraded morphology, are assumed to be
very old and are referred to as ancient landslides.

C.6.2 Interpretation from Aerial Photograph

Aerial photograph studies have shown that revegetation of landslide scars in Hong
Kong is relatively rapid. In general, landslide trails are completely vegetated within 18 years
of the event. The source areas vegetate rather more slowly. They are typically 70%
revegetated after 20 to 30 years, and are commonly about 90% revegetated after 35 years
(Evans et al., 1997). Therefore, field verification of landslides that are more than 20 years
old can prove difficult. Similarly, the identification of landslide scars that significantly
pre-date available aerial photographs must rely on recognizing the degraded landforms.

Interpretation of high level aerial photographs has determined that an average of

about 320 landslides occurred each year between 1945 and 1994 (Evans & King, 1998).
Landslides are commonly initiated on the middle and upper sections of natural slopes and
occur as clusters on slopes of 30 to 40 gradient. They are usually concentrated in the
footprint of intense rainstorms, with densities of more than 10 per km2 occurring if the 24 hour
rainfall exceeds about 20% of the mean annual rainfall (Evans, 1997). Most of the
landslides triggered under the present climatic regime are shallow debris avalanches that have
typical volumes of between 50 m3 and 2000 m3. Much less commonly, larger natural terrain
 - 118

landslides may occur (King 1996a) or large failures of natural terrain can be triggered by
human activity (GEO, 1996b; Cooper, 1992).

C.6.3 Landslide Types

C.6.3.1 General

Historically, landslides have been classified according to a wide range of geological,

geomorphological and engineering criteria, depending upon the intended purpose of the
scheme (e.g. Sharpe, 1938; Varnes, 1958, 1978; Hutchinson, 1988; Cruden & Varnes, 1996).
The most commonly used international landslide classification is that of Cruden and Varnes
(1996). Definitions of landslide types are given in the Glossary of Terms at Section D1 of
Appendix D. A general classification scheme for hillside failures in Hong Kong is given in
Section D2 of Appendix D.

Most natural terrain landslides in Hong Kong involve displacement of the weathered
profile and colluvium. The displaced material is termed debris. Generally it is difficult to
determine the dominant mechanism of movement, largely because many landslides are rapid
with most of the displaced material transported beyond the source area. For this reason the
more general term avalanche is used to describe the movement of these types of landslide.
Several other types of landslide are recognized in Hong Kong, namely rock falls, boulder
falls, earth slumps, debris slides, debris flows and channelized debris flows. Some
well-documented landslides that occurred in recent years are given in Appendix A.

C.6.3.2 Rock and Boulder Falls

Rock and boulder falls are restricted to specific materials and facets of the landscape.
Rock falls originate from rock faces as falls, slides or topples of joint-bounded, and thus
generally angular, rock fragments. These blocks move downslope by combinations of
rolling, bouncing and sliding. Boulder falls occur when generally rounded corestone or
colluvial boulders, located at or near the ground surface, are dislodged from the regolith,
either by fluvial undercutting or by landslides. On straight slopes, once displaced, the
rounded boulders could be expected to be more prone to rolling than would the angular
rockfall debris.

C.6.3.3 Slumps to Debris Flows

Slumps, debris slides, debris avalanches, and channelized debris flows are points on a
continuum that reflects the increasing run-out of generally granular materials on slopes.
This continuum presents the end points for a classification scheme, and also a series of stages
through which an individual landslide may progress as a channelized debris flow develops.
Consequently, individual landslides could be classified anywhere between the four definitive
types. The source areas of most of these landslides have similar slope angles, materials, and
hydrogeological characteristics. The displaced material is normally debris, and commonly
soil pipes are exposed in the main scarp and source floor.

Slides are failures in which the dominant mechanism of movement is sliding. In this
 - 119

case the displaced material will remain an intact mass (Figure C2d). Inevitably there will be
some attrition of the displaced intact mass around its unsupported edges and at the base where
sliding occurs, but this should not involve a significant proportion of the whole mass.
Sliding along a basal plane in the source area can be confirmed by fieldwork if slickensides
are present. However, these have only been found at a small proportion of natural terrain
landslides in Hong Kong, mainly where the failure occurs along relict joints or a planar
bedrock/colluvium interface. Even then, only a few of these landslides were classified as
debris slides, because in most cases the displaced material broke up resulting in debris
avalanches.

A slump is a form of slide in which the intact displaced mass partially overlies the
surface of rupture. Slumps range considerably in size. Smaller examples of debris slumps
are the relatively common shallow displacements of hillslope debris defined by tension cracks
or small landslide scarps, such as the Burma Lines failure (Fugro Scott Wilson, 1999a). At
the other end of the scale are the larger, deeper seated, translational or rotational, inactive
ancient failures that have been recognized at several localities, for example at Area 19 in Tuen
Mun (Scott Wilson, 1999). In addition, a number of failures of natural terrain triggered by
excavations fall into this class (Koor & Campbell, 1998; Lau & Franks, 1998).

A debris slide is the next stage of mobilisation where the displaced, but still intact,
mass moves beyond the plane of rupture by sliding. True debris slides are seldom
encountered due to the propensity of a coarse granular soil mass to break up when sliding
over an irregular surface, although isolated rafts of intact displaced material are commonly
found associated with remoulded landslide debris. These features probably represent a
transitional stage between a debris slide and debris avalanche. The Shum Wan Road
landslide of 1995 (GEO, 1996b) was a very large (26 000 m3) rock/debris slide.
Observations during fieldwork revealed that the displaced mass at the slope toe was
essentially intact, although this was not discernible from aerial photographs.

Debris avalanches (Figure C2) are features in which most of the displaced material
breaks up and becomes remoulded, but no additional surface water is incorporated into the
debris. The initial displacement of material at the source may be by sliding at the base of the
displaced mass, although there are few examples with clear field evidence for this. In most
cases the surface of rupture is rough, comprising in-situ weathered rock with embedded
corestones or colluvium containing large clasts. Thin layers of remoulded debris and rafts of
intact soil may locally overlie the surface of rupture. Together this evidence suggests that
other processes may be involved such as the localized flow of the displaced material due to
elevated pore water pressure. This could be caused by perched water tables from infiltration
and soil pipes, seepage pressures, and the collapse of soil structure in undrained conditions.
As movement continues, and the intact material breaks up into landslide debris, components
of undrained loading, rolling, bouncing and inertial granular or slurry flow (Pierson & Costa,
1987) may also be important. Where the landslide debris is mobile, but the slope below the
source is planar or convex, the debris usually spreads out over the ground with lobes and
fingers extending downslope as an open slope debris avalanche (Figure C2a). Depending
upon the local topography, the debris may be funneled into channels. This results in a trail
that is narrower than the source (Figures C2a & 2b) and the avalanche is described as
channelized. Although the overall movement of the landslide debris has a flow-like
morphology, a number of mechanisms may have been involved in addition to flow. If the
material is saturated, then slurry flow may become the dominant mechanism and this part of
 - 120

the event could be classified as a debris flow.

Debris avalanches are the most common type of natural terrain landslide in Hong
Kong. Open slope types are generally small with limited run-out, therefore they have not
been singled out for specific case studies. A possible example is the 1998 landslide at the
Outward Bound School (Fugro Scott Wilson, 1999b), which was an intermediate stage
between a debris slide and an open slope debris avalanche. The more mobile the debris, the
longer will be the run-out of a debris avalanche, thus the more likely it is to become
channelized. The upper section of the 1990 Tsing Shan Debris Flow (King, 1996a) was a
very large channelized debris avalanche, which developed into a debris flow lower down the
slope.

Debris flows develop when the landslide debris is mixed with sufficient water for
slurry flow to occur. If the moisture content of the displaced material is sufficiently high,
then slurry flow may develop during movement without the further addition of water.
However, more commonly the debris enters a stream where it mixes with the streamflow and
saturated debris is channelled along the drainage line (Figure C3). This mechanism occurred
at Liu Pok in 1995 (King, 1997). As dilution increases the debris slurry will become
progressively more mobile until slurry flow develops into hyperconcentrated stream flow.
With further dilution the mechanism grades into stream flow, which is no longer a mass
movement process but an alluvial process. At this stage the phenomenon is termed a debris
flood (e.g. King, 1996b). Field work is required to confirm that deposition from slurry flow
has occurred. Diagnostic features include a fines-rich matrix or loose sand and gravel that
indicate redistribution of the landslide debris, or the presence of debris fans. It is proposed
that the category of channelized debris flow is restricted to landslides in which stream water is
mixed with the debris and that follow an established stream bed. These channels will
commonly have a history of previous channelized debris flows.

Some of the longest channelized debris flows in Hong Kong have resulted from the
coalescing of multiple debris avalanches to form a single channelized debris flow (Figure C3).
A notable example, about 1.2 km long, occurred south of Shek Kong in 1982. Channelized
debris flows commonly increase in volume due to erosion and entrainment of loose deposits
from the stream bed and banks. Typical entrainment rates are up to 3.5 m3 per linear metre
of trail (Franks, 1998) or up to 3 times the source volume (Wong et al., 1998). The 1990
Tsing Shan debris flow (King, 1996a) was an exceptional event where an initial landslide of
about 3 000 m3 eroded and entrained a further 18 000 m3 of material containing large boulders
and deposited it as a fan at the foot of the slope. This exceptional entrainment, of some
900%, occurred because a large source was located at the crest of a small cliff that accelerated
the saturated remoulded landslide debris into a steep-sided, colluvium-lined stream channel.

In Hong Kong, many landslide sources are located near the pronounced breaks of slope
that are common on valley sides or at the head of hollows. These breaks may be either an
erosion front on the slope or the main scarp of an ancient landslide. Sources are found both
directly above the break in slope, where a thinning soil profile and permeability contrasts can
cause seepage pressures to build up, and on the break in slope itself, where the locally steeper
slope forms an erosion scarp. Other common locations of the larger landslides are
accumulations of loose rock and soil on the midslope below steeper rocky outcrops that form
the upper slopes.
 - 121

In addition to the landslides that originate on the upper sections of hillslopes, larger
natural terrain landslides are also common near the base of slopes, particularly on
over-steepened slopes at the coast and along incised river valleys. Cliffs can occur at both
locations, in which case the landslides may include falls and slides of rock in addition to
sliding failures. Although these landslides may have relatively large volumes, they do not
generally show long run-outs as the shoreline or the flat river valley floor confines them.
These types of landslide include some large scars that contain multiple sources where several
adjacent smaller landslides have occurred over a number of years. This may have developed
through nested instability within a much older relict landslide scar, or from backsapping and
growth of the main scarp from an originally smaller source.

C.6.3.4 Large Relict Landslides

The degraded scars of ancient, more deep-seated landslides in the Hong Kong
landscape can be recognized on aerial photographs. However, the source areas are usually
indistinct due to modification by more recent small landslides on the scarps. Also the
displaced material commonly forms fans and aprons of colluvium which make it difficult to
distinguish the deposits of individual events. A number of such locations have been
identified (Scott Wilson, 1999), but until these have been investigated in the field, the
materials, mechanism and activity of these scars cannot be established. Fieldwork has been
carried out to investigate an 850 000 m3 colluvial lobe at Sham Wat on North Lantau (King,
1998). Results indicate that the lobe is a composite feature comprising several large rock
and soil failures. Photoluminescence dates of a large colluvium lobe (about 1 million m3) on
North Lantau suggest that the events may have occurred within the last 10 000 years (King,
2001). Several very large, relict earth slides of completely decomposed andesite have also
been identified in the pre-development landscape at Area 19 Tuen Mun, which is in the
foothills of Tsing Shan (Scott Wilson Kirpatrick, 1986). Further detailed investigations of
the ancient landslide scars in the Hong Kong landscape may enable an appropriate
classification for these events to be established, and also clarify their significance with respect
to present day slope development processes.

C.7 REFERENCES

Berry, L. (1961). Erosion surfaces and merged beaches in Hong Kong. Bulletin of the
Geological Society of America, Vol. 72, pp 1383-1394.

Berry, L. & Ruxton, B.P. (1960). The evolution of Hong Kong Harbour basin. Zeitschrift
fur Geomorphologie, Vol. 4, No. 2, pp 97-115.

Cooper, A.J. (1992). Reassessment of the Po ShanLandslide of 18th June 1972. Special
Project Report SPR 16/92, Geotechnical Engineering Office, Hong Kong, 75 p.
plus 6 drgs.

Cruden, D.M. & Varnes, D.J. (1996). Landslide types and processes. In A.K. Turner &
R.L.Schuster (editors) Landslides, Investigation and Mitigation. Transportation
Research Board Special Report 247, National Academy Press, Washington, DC,
pp 36-75.
 - 122

Evans, N.C. (1997). Preliminary Assessment of the Influence of Rainfall on Natural Terrain
Landslide Initiation. Discussion Note DN 1/97, Geotechnical Engineering Office,
Hong Kong, 28 p.

Evans, N.C., Huang, S.W. & King, J.P. (1997). The Natural Terrain Landslide Study:
Phases I and II. Special Project Report SPR 5/97, Geotechnical Engineering Office,
Hong Kong, 119 p. plus 2 drgs.

Evans, N.C. & King, J.P. (1998). The Natural Terrain Landslide Study: Debris Avalanche
Susceptibility. Technical Note TN 1/98, Geotechnical Engineering Office, Hong Kong,
96 p.

Franks, C.A.M. (1998). Study of Rainfall Induced Landslides on Natural Slopes in the
vicinity of Tung Chung New Town, Lantau Island. GEO Report No. 57,
Geotechnical Engineering Office, Hong Kong, 102 p. plus 3 maps.

Fugro Scott Wilson Joint Venture (1999a). Detailed Study of Slope Distress at Queens Hill,
Burma Lines Camp, Fanling. Landslide Study Report LSR 10/99, Geotechnical
Engineering Office, Hong Kong, 102 p.

Fugro Scott Wilson Joint Venture (1999b). Detailed Study of the Landslide at the Outward
Bound School on 9 June 1998. Landslide Study Report LSR 7/99, Geotechnical
Engineering Office, Hong Kong, 88 p.

Fyfe, J.A., Shaw, R., Campbell, S.D.G., Lai, K.W. & Kirk, P.A. (2000). The Quaternary
Geology of Hong Kong. Geotechnical Engineering Office, Hong Kong, 209 p. plus 4
maps.

GEO (1996a). Report on the Fei Tsui Road Landslide of 13 August 1995 - Volume 2.
Findings of the Landslide Investigation. Geotechnical Engineering Office, Hong
Kong, 68 p.

GEO (1996b). Report on the Shum Wan Road Landslide of 13 August 1995 - Volume 2.
Findings of the Landslide Investigation. Geotechnical Engineering Office, Hong
Kong, 51 p.

GEO (1998). Report on the Ching Cheung Road Landslide of 3 August 1997. GEO Report
No. 78, Geotechnical Engineering Office, Hong Kong, 142 p.

Grant, C.J. (1962). The Soils and Agriculture of Hong Kong. Hong Kong Government
Printer, Hong Kong, 164 p.

Grant, C.J. (1968). Gullying and erosion characteristics of Hong Kong soils. Proceedings
of the Symposium on Recent Advances in Tropical Ecology, Varanasi, India,
pp 94-107.

Grant, C.J. (1986). Soil. In: T.N. Chiu & C.L. So (editors) A Geography of Hong Kong,
(second edition), Oxford University Press, Hong Kong, pp 110-117.
 - 123

Hansen, A. (1984). Engineering geomorphology: The application of an evolutionary model

of Hong Kongs terrain. Zeitschrift fur Geomorphologie, suppl. Vol. 51, pp 39-50.

Hansen, A. & Nash, J.M. (1985). A brief review of soil erosion causes, effects & remedial
measures. Proceedings of the conference on geological aspects of site investigation.
Bulletin of the Geological Society of Hong Kong Bulletin No. 2, pp 139-150.

Hutchinson, J.N. (1988). General report: Morphological and geotechnical parameters of

landslides in relation to geology and hydrogeology. Proceedings of the Fifth
International Symposium on Landslides, Lausanne, Balkema, Rotterdam, pp 3-35.

King, J .P. (1996a). The Tsing Shan Debris Flow. Special Project Report SPR 6/96,
Geotechnical Engineering Office, Hong Kong, 3 volumes, 427p, 129 p & 166 p.

King, J.P. (1996b). The Tsing Shan Debris Flood of June 1992. Special Project Report
SPR 7/96, Geotechnical Engineering Office, Hong Kong, 83 p. plus 2 drgs.

King, J.P. (1997). Natural Terrain Landslide Study: Damage to Liu Pok School by a Natural
Terrain Landslide. Discussion Note 2/97, Geotechnical Engineering Office, Hong
Kong, 66 p.

King, J.P. (1998). Natural Terrain Landslide Study: The Sham Wat Debris Lobe; Interim
Report. Technical Note TN 6/98, Geotechnical Engineering Office, Hong Kong,
37 p. plus 1 drg.

King, J.P. (2001). Luminescence dating of colluvium and landslide deposits in Hong Kong.
Technical Note TN 1/2001, Geotechnical Engineering Office, Hong Kong, 65 p.

Koor, N.P. & Campbell, S.D.G. (1998). Geological Characterization of the Lai Ping Road
Landslide. Geological Report GR 3/98, Geotechnical Engineering Office, Hong
Kong, 178 p. plus 5 maps.

Lai, K.W. & Taylor, B.W. (1984). The classification of colluvium in Hong Kong. In:
W.W.S. Yim (editor) Geology of Surficial Deposits in Hong Kong, Geological Society
of Hong Kong Bulletin No. 1, pp 75-86.

Lau, K.C.. & Franks, C.A.M. (1998). A Study of the Landslide at CLP Substation Site,
Sham Shui Kok, Lantau Island. Special Project SPR 5/98, Geotechnical Engineering
Office, Hong Kong, 53 p.

Nandy, S. & Jiao, J.J. (2001). Ground-water flow analysis in the slope above Shum Wan
Road, Hong Kong. Environmental & Engineering Geoscience, Vol. VII, pp 239-250.

Nash, J.M. & Dale, M.J. (1984). Geology and hydrogeology of natural tunnel erosion in
superficial deposits in Hong Kong. In: W.W.-S Yim (editor) Geology of Surficial
Deposits in Hong Kong. Geological Society of Hong Kong Bulletin No. 1, pp 61-72.
 - 124

Owen, R.B. & Shaw, R. (2000). Weathering profile development over volcanic rocks in the
Tuen Mun valley, Hong Kong: S. Reels & A. Page (editors) The Urban Geology of
Hong Kong. Geological Society of Hong Kong Bulletin No. 6.

Pierson, T.C. & Costa, J.E. (1987). A rheological classification of sub-aerial sediment-water
flows. In: J.E. Costa & G.F. Wieczorek (editors), Debris Flows/Avalanches:
Process, Recognition and Mitigation, Geological Society of America, Boulder,
Colorado, pp 1-12.

Ruxton, B.P. & Berry, L. (1957). The weathering of granite and associated erosional
features in Hong Kong. Bulletin of the Geological Society of America, Vol. 68,
pp 1263-1292.

Scott Wilson (1999). Specialist API services for the Natural Terrain Landslide Study:
Interpretative Report Task B. Report to the Geotechnical Engineering
Office/Planning Division by Scott Wilson (Hong Kong) Ltd., 32 p. plus 6 Appendices.

Scott Wilson Kirkpatrick (1986). Tuen Mun New Town Area 19 Slope Monitoring: April
1983 - March 1986. Summary Report, Report to Territory Development Department,
Hong Kong, 87 p.

Sewell, R.J., Campbell, S.D.G., Fletcher, C.J.N., Lai, K.W. & Kirk, P.A. (2000). The
Pre-Quaternary Geology of Hong Kong. Geotechnical Engineering Office, Hong
Kong, 181 p. plus 4 maps.

Sharpe, C.F.S. (1938_. Landslides and Related Phenomena; A study of Mass-movements of

Soil and Rock. Columbia University Press, New York, 137 p.

Shaw, R. (1997). Variations in sub-tropical weathering profiles over the Kowloon Granite,
Hong Kong. Journal of the Geological Society, London, Vol. 154, Part 6,
pp 1077-1085.

So, C.L. (1986). Geology and Geomorphology. In: C.J. Peng, F.S. Sit and C.L. So (editors)
Hong Kong and Macau, Commercial Press, Hong Kong (In Chinese), pp 25-45.

Sun, H. W. & Campbell, S.D.G. (1998). The Lai Ping Road Landslide of 2 July 1997.
Landslide Study Report LSR 27/98, Geotechnical Engineering Office, Hong Kong,
138 p.

Tregear, T.R. (1958). A Survey of Land Use in Hong Kong and the New Territories. Hong
Kong University Press, 76 p.

Varnes, D.J. (1958). Landslide types and processes. In: E.B. Eckel (editor), Landslides
and Engineering Practice. Highway Research Board, Special Report 29, Washington,
DC, pp 20 47.

Varnes, D.J. (1978). Slope movement types and processes. In: R.L. Schuster & R.J.
Krizek (editors), Landslides: Analysis and Control. Transport Research Board,
Special Report 176, National Academy of Sciences, Washington, DC, pp 11-33.
 - 125 -

Whiteside, P.G.D. (1996). Technical Note: Drainage characteristics of a cut soil slope with
horizontal drains. Quarterly Journal of Engineering Geology, Vol. 30, pp 137-142.

Wong, H.N., Lam, K.C. & Ho, K.K.S. (1998). Diagnostic Report on the November1993
Natural Terrain Landslides on Lantau Island. GEO Report No. 69, Geotechnical
Engineering Office, Hong Kong, 98 p.
 - 126

LIST OF FIGURES

Figure Page
No. No.

C1 Examples of Regolith Mapping Units 127

C2 Examples of Debris Slides and Debris Avalanches 128

C3 Examples of Channelized Debris Flows 129

 - 127

Rock fall debris: forms below rock slopes or Valley colluvium: narrow linear or ribbon
escarpments as a result of individual rock like features topographically confined
falls or rock avalanches within drainage lines

Residual colluvium: relatively old colluvial Relict landslide: exhibits distinct

deposit not topographically defined morphological characteristics, generally in
fan-like or lobate form

Saprolite: typically occurs on hilltops, Bedrock at or within 1m of the ground

spur/ridgelines as well as slopes surface

Figure C1 - Examples of Regolith Mapping Units

 - 128

(a) Open Hillslope Debris Avalanche, (b) Channelized Debris Avalanche,

North Tsing Shan Range South Tsing Shan Range

(c) Channelised Debris Avalanche, (d) Debris Slide, South Lantau

Mid-Tsing Shan Range

Figure C2 - Examples of Debris Slides and Debris Avalanches

 - 129

(a) North of Leung King Estate, Tuen Mun (b) Tsing Shan (2000) Channelized
Debris Flow

(c) North of Leung King Estate, (d) Pat Heung, Shek Kong
Tuen Mun

Figure C3 - Examples of Channelized Debris Flows

 - 130 -

APPENDIX D

BIBLIOGRAPHY

 - 131 -

CONTENTS

Page

No.

Title Page 130

CONTENTS 131

D.1 GEOLOGICAL MEMOIRS 132

D.2 1:20 000-SCALE GEOLOGICAL MAPS 132

D.3 1:5 000-SCALE SHEET REPORTS 134

D.4 GASP REPORTS 134

D.5 GEOGUIDES 135

D.6 REPORTS PREPARED BY GEO OR BY GEO CONSULTANTS 135

D.7 GENERAL PUBLICATIONS ON HONG KONG 136

D.8 OTHER RELEVANT DOCUMENTS 137

 - 132 -

The documents listed below are additional reference materials on local geology and
geotechnical assessment, particularly GEO publications. It is not intended to be an
exhaustive list of reading materials on the subjects.

D.1 GEOLOGICAL MEMOIRS

Addison, R. (1986). Geology of Sha Tin: 1:20 000 Sheet 7. Hong Kong Geological
Survey Memoir No. 1, Geotechnical Control Office, 85p

Fyfe, J.A., Shaw, R., Campbell, S.D.G., Lai, K.W. and Kirk, P.A. (2000). The Quaternary
Geology of Hong Kong. Geotechnical Engineering Office, Hong Kong, 208 p. plus
6 maps.

Langford, R.L., James, J.W.C., Shaw, R., Campbell, S.D.G. Kirk, P.A. & Sewell, R.J. (1995).
Geology of Lantau District. 1:20 000 Sheets 9, 10, 13 and 14. Hong Kong
Geological Survey Memoir No. 6, Geotechnical Engineering Office, 173 p.

Langford, R.L., Lai, K.W., Arthurton, R.S. & Shaw, R. (1989). Geology of the Western
New Territories: 1:20 000 Sheets 2, 5 and 6. Hong Kong Geological Survey Memoir
No. 3, Geotechnical Control Office, 140 p.

Lai, K.W. Campbell, S.D.G., Shaw, R. (1996). Geology of the North Eastern New
Territories. 1:20 000 Sheets 3 and 4. Hong Kong Geological Survey Memoir No. 5,
Geotechnical Engineering Office, 144 p.

Sewell, R.J., Campbell, S.D.G., Fletcher, C.J.N., Lai, K.W. & Kirk, P.A. (2000). The
Pre-Quaternary Geology of Hong Kong. Geotechnical Engineering Office, Hong
Kong, 181 p. plus 4 maps.

Strange, P.J. & Shaw, R. (1986). Geology of Hong Kong Island and Kowloon: 1:20 000
Sheets 11 and 15. Hong Kong Geological Survey Memoir No. 2, Geotechnical
Control Office, 134 p.

Strange, P.J., Shaw, R. & Addison, R. (1990). Geology of Sai Kung and Clearwater Bay.
1:20 000 Sheets 8, 12 and 16. Hong Kong Geological Survey Memoir No. 4,
Geotechnical Control Office, 111 p.

D.2 1:20 000-SCALE GEOLOGICAL MAPS

Addison, R. & Purser, R.J. (1986). Sha Tin. Hong Kong Geological Survey Sheet 7. Solid
and Superficial Geology. 1:20 000 Series HGM20. Geotechnical Control Office,
Hong Kong.

Addison, R. & Shaw, R. (1990). Waglan Island. Hong Kong Geological Survey Sheet 16,
Solid and Superficial Geology. 1:20 000 Series HGM20. Geotechnical Control
Office, Hong Kong.
 - 133 -

Arthurton, R.S., Lai, K.W. & Shaw, R. (1988). Tsing Shan (Castle Peak). Hong Kong
Geological Survey Sheet 5. Solid and Superficial Geology. 1:20 000 Series HGM20.
Geotechnical Control Office, Hong Kong.

Lai, K.W. & Shaw, R. (1989). San Tin. Hong Kong Geological Survey Sheet 2. Solid and
Superficial Geology. 1:20 000 Series HGM20. Geotechnical Control Office, Hong
Kong.

Lai, K.W. & Shaw, R. (1991). Sheung Shui. Hong Kong Geological Survey Sheet 3. Solid
and Superficial Geology. 1:20 000 Series HGM20. Geotechnical Control Office,
Hong Kong.

Lai, K.W. & Shaw, R. (1992). Kat O Chau (Crooked Island). Hong Kong Geological
Survey Sheet 4. Solid and Superficial Geology. 1;20 000 Series HGM20.
Geotechnical Engineering Office, Hong Kong.

Langford, R.L., Addison, R., Strange, P.J. & James, J.W.C. (1991). Silver Mine Bay.
Hong Kong Geological Survey Sheet 10. Solid and Superficial Geology. 1:20 000
Series HGM20. Geotechnical Engineering Office, Hong Kong.

Langford, R.L. & James, J.W.C. (1995). Shek Pik. Hong Kong Geological Survey Sheet 13.
Solid and Superficial Geology. 1:20 000 Series HGM20. Geotechnical Engineering
Office, Hong Kong.

Langford, R.L., Kirk, P.A. & James, J.W.C. (1994). Tung Chung. Hong Kong Geological
Survey Sheet 9. Solid and Superficial Geology. 1:20 000 Series HGM20.
Geotechnical Engineering Office, Hong Kong.

Langford, R.L., Lai, K.W. & Shaw, R. (1988). Yuen Long. Hong Kong Geological Survey
Sheet 6. Solid and Superficial Geology. 1:20 000 Series HGM20. Geotechnical
Control Office, Hong Kong.

Langford, R.L., Shaw, R. & Fyfe. J.A. (1995). Cheung Chau. Hong Kong Geological
Survey Sheet 14. Solid and Superficial Geology. 1:20 000 Series HGM20.
Geotechnical Engineering Office, Hong Kong.

Strange, P.J. & Shaw, R. (1986). Hong Kong Island and Kowloon. Hong Kong Geological
Survey Sheet 11, Solid and Superficial Geology. 1;20 000 Series HGM20.
Geotechnical Control Office, Hong Kong.

Strange, P.J. & Shaw, R. (1987). Hong Kong South and Lamma Island. Hong Kong
Geological Survey Sheet 15. Solid and Superficial Geology. 1:20 000 Series HGM20.
Geotechnical Control Office, Hong Kong.

Strange, P.J. & Shaw, R. (1989). Clear Water Bay, Hong Kong Geological Survey Sheet 12.
Solid and Superficial Geology. 1:20 000 Series HGM20. Geotechnical Control
Office, Hong Kong.
 - 134 -

Strange, P.J., Addison, R. & Shaw, R. (1989). Sai Kung Peninsula. Hong Kong Geological
Survey Sheet 8. Solid and Superficial Geology. 1:20 000 Series HGM20.
Geotechnical Control Office, Hong Kong.

D.3 1: 5 000-SCALE SHEET REPORTS

Sheet Report 1: Frost, D.V. (1992). Geology of Yuen Long. Hong Kong Geological
Survey Sheet Report No.1, Geotechnical Engineering Office, Hong Kong, 69 p.

Sheet Report 2: Langford, R.L. (1994). Geology of Chek Lap Kok. Hong Kong
Geological Survey Sheet Report No.2, Geotechnical Engineering Office, Hong Kong,
61 p.

Sheet Report 3: Sewell, R.J. & Fyfe, J.A. (1995). Geology of Tsing Yi. Hong Kong
Geological Survey Sheet Report No. 3, Geotechnical Engineering Office, Hong Kong,
43 p.

Sheet Report 4: Sewell, R.J. & James, J.W.C. (1995). Geology of North Lantau Island and
Ma Wan. Hong Kong Geological Survey Sheet Report No. 4, Geotechnical
Engineering Office, Hong Kong, 46 p.

Sheet Report 5: Sewell, R.J. (1996). Geology of Ma On Shan. Hong Kong Geological
Survey Sheet Report No.5, Geotechnical Engineering Office, Hong Kong, 45 p.

D.4 GASP REPORTS

Geotechnical Control Office (1987). Geotechnical Area Studies Programme - Hong Kong
and Kowloon. GASP Report No. I, Geotechnical Control Office, Hong Kong, 170 p.
plus 4 maps.

Geotechnical Control Office (1987). Geotechnical Area Studies Programme - Central New
Territories. GASP Report No. II, Geotechnical Control Office, Hong Kong, 165 p.
plus 4 maps.

Geotechnical Control Office (1987). Geotechnical Area Studies Programme - West New
Territories. GASP Report No. III, Geotechnical Control Office, Hong Kong, 155 p.
plus 4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - North West
New Territories. GASP Report No. IV, Geotechnical Control Office, Hong Kong,
120 p. plus 3 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - North New
Territories. GASP Report No. V, Geotechnical Control Office, Hong Kong, 135 p.
plus 4 maps.
 - 135 -

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - North Lantau.
GASP Report No. VI, Geotechnical Control Office, Hong Kong, 124 p. plus 4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - Clear Water
Bay. GASP Report No. VII, Geotechnical Control Office, Hong Kong, 144p. plus
4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - North East
New Territories. GASP Report No. VIII, Geotechnical Control Office, Hong
Kong,144 p. plus 4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - East New
Territories. GASP Report No. IX, Geotechnical Control Office, Hong Kong, 141 p.
plus 4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - Islands.

GASP Report No. X, Geotechnical Control Office, Hong Kong, 142 p. plus 4 maps.

Geotechnical Control Office (1988). Geotechnical Area Studies Programme - South Lantau.
GASP Report No. XI, Geotechnical Control Office, Hong Kong, 148 p. plus 4 maps.

Geotechnical Control Office (1989). Geotechnical Area Studies Programme - Territory of

Hong Kong, by K.A. Styles & A. Hansen. GASP Report No. XII, Geotechnical
Control Office, Hong Kong, 346 p. plus 14 maps & 1 chart.

D.5 GEOGUIDES

Geotechnical Control Office (1984). Geotechnical Manual for Slopes (Second edition).
Geotechnical Control Office, Hong Kong, 295 p.

Geotechnical Control Office (1987). Guide to Site Investigation (Geoguide 2).

Geotechnical Control Office, Hong Kong, 353 p.

Geotechnical Control Office (1988). Guide to Rock and Soil Descriptions (Geoguide 3).
Geotechnical Control Office, Hong Kong, 189 p.

Geotechnical Engineering Office (1998). Guide to Slope Maintenance (Geoguide 5)

(Second edition), Geotechnical Engineering Office, Hong Kong, 91 p.

Geotechnical Engineering Office (1998). Laymans Guide to Slope Maintenance (Second

edition) (Bilingual), Geotechnical Engineering Office, Hong Kong, 54 p.

D.6 REPORTS PREPARED BY GEO OR BY GEO CONSULTANTS

Cooper, A.J. (1992). Reassessment of the Po Shan Road Landslide of 18th June 1972,
Special Project Report 16/92, Geotechnical Engineering Office, Hong Kong, 75 p.
 - 136 -

ERM-Hong Kong, Ltd (1998). Feasibility Study for QRA of Boulder Fall Hazard in Hong
Kong, GEO Report No. 80, Geotechnical Engineering Office, Hong Kong, 61 p.

Halcrow Asia Partnership Ltd. (1998-2000). Investigation of some Selected Landslide

Incidents in 1997 (Volume 1 to 6), GEO Report No. 79, Geotechnical Engineering
Office, Hong Kong, 142 p.

Hungr, O. (1997). Mobility of Landslides in Hong Kong, Pilot Analysis using a Numerical
Model. Report to Geotechnical Engineering Office, Hong Kong.

Irfan, T.Y., Tang, K.Y. (1993). Effect of the Coarse Fractions on the Shear Strength of
Colluvium, GEO Report No. 23, Geotechnical Engineering Office, Hong Kong, 223 p.

Koor, N.P. & Campbell, S.D.G. (1998). Geological Characterization of the Lai Ping Road
Landslide. Geological Report 3/98, Geotechnical Engineering Office, Hong Kong,
152 p.

Malone, A.W. (1997). Areal Extent of Intense Rainfall in Hong Kong 1979 to 1995. GEO
Report No. 62, Geotechnical Engineering Office, Hong Kong, 85 p.

Premchitt, J. (1991). Hong Kong Rainfall and Landslides in 1985. GEO Report No. 2,
Geotechnical Engineering Office, Hong Kong, 108 p. plus 1 drg.

Premchitt, J., Lam, T.S.K., Shen, J.M. & Lam, H.F. (1992). Rainstorm Runoff on Slopes.
GEO Report No. 12, Geotechnical Engineering Office, Hong Kong, 211 p.

Tang, M
.C. (1993). Report on the Rainstorm of May 1982. GEO Report No. 25,
Geotechnical Engineering Office, Hong Kong, 137 p. plus 1 drg.

Wong, C.K.L. (1997). Hong Kong Rainfall and Landslides in 1995. GEO Report No. 59,
Geotechnical Engineering Office, Hong Kong, 125 p. plus 1 drg.

Wong, H.N., Chen, Y.M. & Lam, K.C. (1997). Factual Report on the November 1993
Natural Terrain Landslides in Three Study Areas on Lantau Island. GEO Report
No. 61, Geotechnical Engineering Office, Hong Kong, 42 p.

D.7 GENERAL PUBLICATIONS ON HONG KONG

Bennett, J.D. (1984a). Review of Superficial Deposits and Weathering in Hong Kong.
Geotechnical Control Office, Hong Kong, 51 p. (GCO Publication No. 4/84).

Burnett, A.D. & Lai, K.W. (1984). A review of the photo-geological lineaments and fault
system of Hong Kong. Geological Society of Hong Kong, Bulletin No. 2, I. (Edited:
McFeat-Smith (Editor), pp 113 - 131.
 - 137 -

Hansen, A. Franks, C.A.M., Kirk, P.A., Brimicombe, A.J. & Fung, T. (1995). Application
of GIS to hazard assessment, with particular reference to landslides in Hong Kong.
In: Geographical Information systems in Assessing Natural Hazards, (Edited: Carrara,
A. & Guzzetti, F.), pp 273-298.

Lai, K.W. (1982). Discussion on the colluvium of Hong Kong (8th October 1980). Hong
Kong Baptist College Academic Journal, Vol. 9, pp 139-158.

Lai, K.W. & Taylor, B.W. (1984). The classification of colluvium in Hong Kong.
Geological Society of Hong Kong, Bulletin No. 1, (Edited: W.W.S. Yim), pp 75-85.

Langford, R.L. & Hadley, D. (1990a). New debris flow on the flanks of Tsing Shan, Hong
Kong. Geological Society of Hong Kong Newsletter, Vol. 8, part 3, pp 2-12.
(Discussion, Vol. 8, part 4, pp 40-45).

Langford, RL. & Hadley, D. (1990b). New debris flow on the flanks of Tsing Shan, Hong
Kong. (Photographic feature). Geotechnical Engineering, Vol. 21, pp 105-107.

Shaw, R. (1993a). Granite weathering and erosion in Hong Kong. Improving Degraded
Lands: Promising Experiences from South China. Bishop Museum Bulletin in
Botany No. 3, Bishop Museum Press, Honolulu (Edited: W.E. Parham, P.J. Durana &
A.L. Hess) pp 189-196.

D.8 OTHER RELEVANT DOCUMENTS

Costa, J.E. & Wieczorek, G.F. (1987). Debris Flows/Avalanches: Process, Recognition and
Mitigation. The Geological Society of America, 239 p.

Dearman, W.R. (1991). Engineering Geological Mapping. Butterworth/Heinemann,

387 p.

Fookes, P.G. (1997). Geology for Engineer: the geological Model, prediction and
performance. Quaternary Journal of Engineering Geology. Vol. 30, pp 293 - 424.

Fookes, P.G. (Editor) (1997). Tropical Residual Soils. Geological Society Engineering
Group Working Party Revised Report. The Geological Society, London, 184 p.

Fookes P.G., Frederick, J.B. & Hutchison, J.N. (2000). Total geological history: a model
approach to the anticipation, observation and understanding of the site conditions
(Invited Paper). Proceedings of the International Conference on Geotechnical and
Geological Engineering GeoEng2000, Melbourne, Vol. 1, pp 370-460.

Hansen, A. (1984). Landslide hazard analysis. In: Slope Instability. John Wiley & Sons
Ltd, (Edited: Brunsden, D & Prior, D.B.), pp 523 - 602.

Lawrence, C.J., Byard, R.J. & Beaven, P.J. (1993). Terrain Evaluation Manual. Transport
Research Laboratory State-of-the-Art, Review 7, HMSO Publications, 285 p.
 - 138 -

Ollier, C.D. & Pain, C. (1996). Regolith, Soils and Landforms. John Wiley & Sons, 316 p.

Selby, M.J. (1993). Hillslope Materials and Processes. Oxford University Press, 451 p.

Thomas, M.F. (1994). Geomorphology in the Tropics: A Study of Weathering and

Denudation in Low Latitudes. John Wiley & Sons, 460 p.

Turner, A.K. & Schuster, R.L. (Editors) (1996). Landslides: Investigation and Mitigation.
Transportation Research Board Special Report No. 247, National Research Council,
National Academy Press, Washington, D.C., 673 p.