CIEM 5790

Slope Engineering and Management

HKUST

Philip CHUNG
Spring 2025
1. Introduction
 A Slope Safety System 斜坡安全系統

Slope Engineering
Construction
e.g. early warning Slope Management
System Holistic landslide risk
management of a large
number of slopes
(vs only ONE slope)
Global QRA
長遠防治山泥傾瀉計劃
e.g. Landslip Prevention and
Mitigation Programme (LPMitP)
Design e.g. stability analysis Slope Maintenance 3
Why do we need a Slope Safety System in Hong Kong ?

Ref.: Airbus Space on X

斜坡工程與山泥傾瀉風險管理 / 边坡工程与滑坡风险管理

19th century

21st century
essential to sustainable city development on
Hong Kong’s hilly terrain 5
 山泥傾瀉災害
滑坡灾害

風險
*risk is a measure of the chance of occurrence of slope failure causing
a certain amount of harm (e.g. fatalities and economic losses) and can
be quantified as : probability x consequence of failure.
6
Gentle Terrain (e.g. Beijing) Hilly Terrain (e.g. HK)

not affected by landslides subject to landslide risk

斜坡安全是城市在發展基
Slope safety is an integral part 礎設施中不可或缺的一環
边斜坡安全是城市在发展基
of infrastructure development 础设施中不可或缺的一环
7
 Slope Safety Technical Review Board (SSTRB)
斜坡安全技術檢討委員會
“SSTRB finds that Geotechnical Engineering Office (GEO)
of the HKSAR Government continues
to lead international practice on slope safety and
is a model that other countries aspire to follow.”

Prof K. Soga Prof D. Petley

UC Berkeley Dr S Lacasse U of Hull
NGI

Annual review of HK’s

slope safety by SSTRB 8
 Hong Kong Slope Safety System
Contents to be covered in Lecture 1
This PPT
another PPT
Part 1 Part 2

Multi-pronged
Evolution : why the approach : what it is
system as it is and how it works
• Considerations • Components
• Chronological • Contributions
evolution

9
 天然山坡 人造斜坡(边坡)
Natural Hillside vs Man-made Slope

削土坡

填土坡 Natural Man-made

hillside slope

擋土牆

The use of the term “Slope”

in a broad sense under the HKSSS
(soil /rock cut slope and fill slope,
retaining wall, natural hillside)
Why there are acute and
chronic landslide
problems in Hong Kong ?

11
The unique combination of climate (e.g. heavy rain),
steep terrain, deep weathering profile with
complex geological and hydrogeological conditions
and the rapid and dense urban development since
the post-Second World War period

Acute and chronic landslide

problems in Hong Kong

Loss of life and socio-

economic consequences 12
 See additional info 1
See additional info 3

See See
additional additional
info 3 info 1

Densely
developed Heavy rain
稠密城市發展 大雨
(7000 / km2) (2400 mm /yr)

See additional info 2

See
See additional info 4

See
additional additional
info 4 info 2

Steep terrain Deep weathering

陡峭地形
(70% > 15) 熱帶深層風化土
(up to ~ 100 m)
Site formation on steep terrain for urban development

resulting in buildings and roads close to steep slopes

 Rapid urban development after WWII
Extensive site formation
(with inadequate control and management)

Kwun Tong and Sau Ming

Ping area around 1959

Formation of substandard
man-made slopes close to
development 15
 Many landslides in the old days
… part of Hong Kong people’s struggle for living

16
 A Brief Summary

• Concept of Slope System / Engineering / Management

(and note that they are not independent from each other)

• HK is liable to landslide risk (under the context of urban

development) and a slope safety system is necessary

17
2. Evolution of Slope Engineering
and Landslide Risk Management
in Hong Kong
 Evolution of Slope Engineering and
Landslide Risk Management in Hong Kong
Before 2010 , 3 major stages + focus on man-made slopes

1977 1st period :

before 1977
Empirical
Slope Engineering 19
• Before 1977, there was no centralised authority to
exercise a territory-wide geotechnical control on
slopes in Hong Kong

• Cut and fill slopes were typically formed to an

empirical geometry based on rules of thumb with no
systematic regulatory control by the Government
and with no expert geotechnical engineering input

20
Empirical Slope Engineering (Rule-of-thumb)

The Building Authority

will offer no objection to
an angle of slope
 35 for filling and
 50 for cutting

Deposited fill in 5 ft.

layer and compacted
subjected to approval by Deposited fill
 2 m before
the Engineer as a result
compaction
of compaction trials
 Empirical design (rule-of-thumb)
.. . although it worked in some cases, it failed
where the ground was atypical and when heavy rain came

Original ground
profile
Cutting  50

Filling  35
Road or platform

Major shortcomings of the empirical design: Irrespective of soil 22

type, groundwater conditions, adverse geological setting etc
• During this period, in addition to the empirical design,
there was inadequate construction control.
• These 2 major factors led to many landslides each year
• Landslides in Hong Kong had resulted in a death toll of
over 420 fatalities during 1945 - 1977
Two notable landslides partly due to inadequate construction control

1972 Po Shan landslides 1976 Sau Mau Ping landslides

23
 18 June 1972 Landslide Disasters

1972 Po Shan Landslide 1972 Sau Mau Ping Landslide

(Vol. = 20,000 m3 ; 67 fatalities) (Vol. = 6,000 m3 ; 71 fatalities)

24
Animation for the 1972 Po Shan Landslide

25
 1972 Po Shan Landslide
(Vol. = 20,000 m3 ; 67 fatalities)

Before

After

26
Problem of uncontrolled
deep excavations in
marginally stable hillsides

Morning 18.6.1972

27
17.6.1972 Evening 18.6.1972
Animation for the 1972 Sau Mau Ping Landslide

28
 1972 Sau Mau Ping Landslide
(Vol. = 6,000 m3 ; 71 fatalities)

Before Fill slope mobile failure

After

29
► 1972 Sau Mau Ping Landslide
Concluded by the Commission of Inquiry :
The landslip was “due primarily to softening
of fill material caused by infiltration of rain-
water … as a result of exceptional long and
intense rainstorm”

The 1966 Aberfan flow slide in Wales killed 144

people, most of whom were school children
( Report of the Tribunal, 1967 )
Susceptibility of loose fill to
undrained collapse leading to
mobile failure was not addressed
30
 1976
1976 Sau Mau Ping Landslide landslide
(Vol = 5,000 m3 ; 18 fatalities) site

1972
landslide
site

31
 Possible static liquefaction of loose fill slope

Fill slope

1960's Construction Practice

 1.5m ? “deposited in 5 ft. layer
end-tipped, and compacted
loose sandy fill subjected to approval by
the Engineer as a result of
compaction trials”
32
 More details on fill slope design will be
given in Lecture 6 (by Philip Chung)

More details on systematic landslide

investigation will be given in Lecture 9 (by Ken Ho)

33
 Evolution of Slope Engineering and
Landslide Risk Management in Hong Kong
Before 2010 , 3 major periods + focus on man-made slopes

2nd period :
1977-mid 90s
Geotechnical
Slope Engineering
Set up of
1977 Geotechnical
1st period : Engineering Office
before 1977
Empirical (GEO)
Slope Engineering 34
 • After the 1976 Sau Mau Ping landslide, on the
recommendation of an independent review panel, the
HK Government established the Geotechnical Control
Office (GCO) in 1977 (renamed as the Geotechnical
Engineering Office (GEO) in 1991) 土力工程處

• It is the central authority in Hong Kong to regulate

geotechnical engineering and oversee slope safety

• This marked the turning point in the history of slope

safety in Hong Kong.

Visit the following website for more information about

GEO:
35
https://www.cedd.gov.hk/eng/about-us/organisation/geo/index.html
• During the period of 1977 to mid 1990s, the GCO/GEO
made progress in various aspects of slope engineering and
promulgated engineering standards, guidelines etc to
recommended good practices to improve and align
design and construction standards of the industry.
• In short, geotechnology was applied to slope engineering.

36
 Aerial Geological and
photograph Engineering groundwater
interpretation geological models
mapping

Limit
Ground Undisturbed equilibrium
investigation sampling analysis

Numerical
Soil & rock modeling
Laboratory testing
logging
37
Remark:

Under part 1 of the lecture, I’ll focus mainly on the

“evolution” of the HK Slope Safety System.

The major components of the HK Slope Safety

System and other Slope management strategies will
be covered under Part 2 of the Lecture.

38
 Evolution of Slope Engineering and
Landslide Risk Management in Hong Kong
Before 2010 , 3 major periods + focus on man-made slopes
3rd period :
mid 90s - 2010
Enhanced Landslide Risk
Management
Notable landslides:
Kwun Lung Lau 1994
Mid-1990s Fei Tsui Rd 1995
Sham Wan Rd 1995

Geotechnical
Slope Engineering
Set up of
Geotechnical
1977 1st period : Engineering Office
before 1977 (GEO)
Empirical
Slope Engineering 39
• During this period of time (mid 1990s to 2010), in
addition to the deterministic approach, the GEO
pioneered the development and adoption of an
explicit risk-based strategy and approach for slope
stability assessment and landslide risk management.

• Other major strategies:

- carry out systematic landslide investigation;
- promulgate policy and guidance on slope
maintenance; and
- Promote public education and public information

40
The major triggering events for this period included
Kwun Lung Lau 觀龍樓 (观龙楼) landslide in July 1994,
Fei Tsui Road 翡翠道 landslide and Shum Wan Road
深灣道 landslide in August 1995.

accelerated

Enhanced Landslide
Risk Management
and slope
engineering
41
 1994 Kwun Lung Lau Landslide
( Vol. = 1,000 m3 ; 5 fatalities & 3 injuries )

• Collapse of a >100-yr
old masonry wall
• Affect a footpath
(temporal users)
• Smaller in scale of
failure, occurred during
landslip warning signal
was issued

42
 Great outcry from the
public, politicians and Administration

Lessons learned :

• Reduced awareness of potential

landslide risk after many
uneventful years
• Increased expectation
• Aversion to multiple-fatality
• Geotechnical process not 100%
robust (technological limitations
and human errors)
43
 See slides 45 – 57 for See PPT in
some more details lecture 10
Risk based
See part 2 of
strategies PPT in lecture 1

Enhanced Slope
LPMP Maintenance
See PPT in Enhanced
lecture 8 Landslide Risk
Management
and slope See part 2 of
engineering PPT in lecture 1
Public
Systematic
Education
Landslide
and
Investigation
Information
Robust
See PPT in design (e.g.
lecture 9 soil nails)
See PPT in 44
lectures 5/6
 • From 1995 to 1999, there was continuous and substantial
increase in expenditure under the LPMP.

Annual Expenditure (Million HK $)

1200
1994 Kwun Lung Lau landslide 993
1000 Landslip Preventive 966 932 940
889921882 868 894 890
Measures 831
800 772
Programme, LPMP 681

600 防止山泥傾瀉計劃

400 393

>10 times increase !

200 195
107
71 69 81 72 75 69 62 64 69
37 63 37 46 56 62 63 63
0
1979/80 1989/90 1999/2000 2009/10
Financial Year 45
• The risk-based methodology considers the
likelihood of landslides and the adverse
consequences.

• Quantified risk management was instrumental to

the formulation of the overall slope safety
strategy and management of landslide risks at
individual vulnerable sites. This involved the use
of global quantitative risk assessment (QRA)量化風
險評估 and site-specific QRA respectively.
46
 reduce “likelihood”
How to reduce “risk” reduce “consequence”
reduce both

Likelihood of failure Consequence of failure

Risk = x
of Quantitative Risk Assessment (QRA)

Advances in
methodology and Techniques for landslide consequence assessment developed in HK

application of Slope engineering aims Techniques for landslide

Quantitative Risk to prevent failure, consequence assessment
Assessment (QRA) thereby reduce risk developed in HK
 Likelihood of Consequence of
Risk = x
failure failure
Failure cannot be totally prevented, despite use of state-of-the-art slope engineering

• Failure cannot be totally prevented, despite use of state-

of-the-art slope engineering
Risk has to be explicitly assessed for establishing the scale of problem, formulating strategy, allocating resources, etc. (slope management)

• Risk has to be explicitly assessed for establishing the scale

of problem, formulating strategy, allocating resources, etc.
(slope management)
Slope stabilization is not always the best solution

• Slope stabilization is not always the best solution

Need for effective risk communication

• Need for effective risk communication

48
Back analysis
Site-specific
Global risk management
landslide
of notable
risklandslides
for
management
landslides
(e.g.from
for
seeindividual
pre-1977
natural
slide 56)hillsides
facilities
man-made(see
(see
lecture
slopes
lecture
11)
(see
10)slides 50 – 55)

1. Global risk management for landslides from

pre-1977 man-made slopes (see slides 50 – 55)

2. Back analysis of notable landslides (e.g. see

slide 56)

3. Site-specific landslide risk management for

individual facilities (see lecture 10)

4. Global risk management for landslides from

natural hillsides (see lecture 11)

49
 Catalogue of man-made slopes
(~ 57,000 nos. in mid-1990s )

18,000 engineered slopes

(post-1977)
This portion to be
retrofitted by 2010

The Big question is :

Which slopes to fix first ?
39,000 nos. un-engineered
(pre-1977) 50
 Risk-based slope ranking system
to determine the priority of slopes for LPM action

High

Priority

Catalogue of Slopes
(total 57,000 nos. in
mid-1990s) Low 51
 Quantitative Risk assessment (QRA) of
Un-engineered Man-made Slopes

Result of the
R. Walls
QRA exercise
Fill
Slopes
Cut Slopes

ranking of all the

registered features (i.e.
cut slopes, fill slopes,
39,000 nos. Pre-1977
Man-made Slopes retaining walls etc)
New Priority Ranking System - Risk Based Systems

reflects likelihood
of failure

Instability Score (IS) ╳

Total Score (TS) =
Consequence Score (CS)

reflects likely risk

of landslide reflects likely
consequence of failure

53
 Risk Distribution

Upgrading the top

10% high risk old
slopes drastically
Gp 1 reduces the overall
43% risk reduction
risk to life by 43%

Gp 2
Example of Gp 1
Gp 3 facilities: affecting
Gp 4 residential buildings,
Gp 5 major roads

Example of Gp 5
facilities: affecting
Country park 54
 Risk reduction targets

(1) By year 2000, reduce the landslide risk arising

from pre-77 man-made slopes to less than 50% of
the risk that existed in 1977

(2) By Year 2010, further reduce the risk to less

than 25% of that in 1977

1977 2000 2010

(100%) (50%) (25%)
 See Additional info 5

See
Additional
Option A
F-N curve info 5

An example of Quantitative Risk

1.00E-01

Analysis (QRA) applied to a slope 1.00E-02

Frequency (F) of N or more fatalities per year

1995 Fei Tsui Road Landslide, 1.00E-03

Chai Wan, Hong Kong UNACCEPTABLE

1.00E-04

ALARP
1.00E-05
最低合理可行

1.00E-06

1.00E-07

INTENSE

REGION
SCRUTINY
BROADLY
ACCEPTABLE

1.00E-08

1.00E-09
1 10 100 1000 10000

Number (N) of Fatalities

 As comparison, Daya Bay Nuclear Power Plant, China 1.00E-02

Frequency (F) of N or more fatalities per year

1.00E-03

UNACCEPTABLE
1.00E-04

HK
ALARP
1.00E-05

1.00E-06

1.00E-07

INTENSE

REGION
SCRUTINY
BROADLY
ACCEPTABLE

1.00E-08

1.00E-09
1 10 100 1000 10000

Number (N) of Fatalities

“The risk of the health of people in Hong Kong being affected
by the plant isNotes
very: much
(1)lower
Thethan
abovethe risks
societal tocriteria are to
risk
health currently encounteredofin
theeveryday life”
natural hillside of 500 m (Reev
(2)
(UK Atomic Energy If a 1990)
Authority, development is affected57by m
appropriate linear scaling factor shou
More details on QRA will be given in Lectures 10/11
(by Ken Ho)

58
• Before 2010, there were three major periods in
the evolution of the HK Slope Safety System

• 1st period (before 1977) : empirical slope

engineering and note 1972 and 1976 major
landslides

• 2nd period (1977-1994) : geotechnical slope

engineering

59
• 3rd period (1995 - 2010) : enhanced landslide
slope management which includes

1. Systematic Landslide Investigation + Slope

Maintenance + Public Education + Robust design

2. Increase in expenditure under LPMP

3. Risk-based strategies plus quantitative targets

4. Global and site-specific QRA

60
Is this the end of the story ?

61
  Lectures 8
Slope degradation
Aging of man-
and requiring
made slopes
improvement / repair

 Lectures 11, 13
Study and mitigation
Development
of natural terrain
close to hillsides
landslide risk

Lecture 7 
Effect to both man-
Climate change
made slopes and
(extreme rainfall)
natural terrain
 Evolution of Slope Engineering and
Landslide Risk Management in Hong Kong
Before 2010 , 3 major periods + focus on man-made slopes
2010 and beyond
Continuous Improvements to the
Slope Safety System in HK

63
 Evolution of Slope Engineering and
Landslide Risk Management in Hong Kong
Examples of Continuous Improvements to the
Slope Safety System in HK

LPMP → Landslip Prevention and Mitigation Programme

(LPMitP) ---- man made slope + natural terrain

More proactive use of latest Information Technology

(IT) (e.g. LiDAR, AI, Robotic, remote sensing … )

Enhanced design guidelines

Enhance early warning system
End of Part 1

65
Additional Information

66
Additional info 1

Additional
info 1

67
 Additional info 1

Additional
info 1

Ref. El Dorado Weather, Inc 68

 Additional info 1

Additional
info 1

Ref. World in Maps 69

 Additional info 2

Additional
info 2
Typical weathering profile

Ref. GEO publication 1/2007 Fig. 4.4.16

70
 Additional info 3

Additional
Ranking of population density info 3

70%
sloping
ground

71
Ref. Wikipedia
 Additional info 4

Additional
info 4

Ref. HK Slope Safety website, GEO 72

 Additional info 5

Additional
info 5
ALARP (As Low As Reasonably Practicable)
The risk is regarded as tolerable only if risk reduction is
impracticable or if the cost is grossly disproportionate to the
improvement gained. This involves determining (HSE, 1992):

1. whether a given risk is so great or the outcome so unacceptable that

it must be refused altogether; or
2. whether the risk is, or has been made, so small that no further
precaution is necessary; or
3. if a risk falls between these two states, that it has been reduced to
the lowest level practicable, bearing in mind the benefits flowing
from its acceptance and taking into account the costs of any further
reduction. The injunction laid down in safety law is that any risk must
be reduced so far as reasonably practicable, or to a level which is 'as
low as reasonably practicable'.
73
74
75
76
End of Additional Information

77