Geohazards [Landslides - Case Studies Hong Kong]

The following two articles provide an overview of the landslip problem in Hong Kong and measures taken by government to reduce the risk from landslip hazard. Both articles were written by Dr A.W. Malone, an Honorary Associate Professor in the Department of Earth Sciences, University of Hong Kong and former Director of the Hong Kong Geotechnical Engineering Office.

Slope Safety Systems for Asian Cities

A methodology based on systems concepts and risk management principles is presented here for the design of slope safety management systems for cities where rain-induced landsliding is a serious social problem. A model action plan is suggested as a guide to setting up such a safety system and the associated costs are given.

keywords: landslides, safety management, system design

Many cities in Asia have experienced strong economic and population growth in recent decades, including a number in tropical and sub-tropical Asia with adverse terrain and climate. Some city governments have been forced into permitting the exploitation of hillslope land to provide housing and roads but the control regimes needed to ensure a good quality built environment, well protected against intense rainstorms, have not been put in place. Such polices leave a legacy of unsafe cut slopes, loose fill embankments on sloping ground and substandard retaining structures. Rain-induced landsliding is becoming a serious social problem in a number of these cities because of multi-fatality building collapses caused by landslide impact and a growing landslide death toll on the roads. How are city governments to respond? How are limited resources best invested in slope safety?

One of the first cities in the region to face these problems was Hong Kong, which began to tackle the issue in the 1970s. A well-developed safety regime is now in place (I]. Hong Kong's comprehensive slope safety regime is taken as a model by other growing cities in the region with an adverse climate and terrain where rain-induced landsliding is commonplace. In Hong Kong new slope works are tightly controlled, old slopes are being brought up to modern safety standards under a long-term works programme, routine slope maintenance is encouraged, hillside shantytowns are being progressively cleared and rainstorm emergency preparedness measures are in place. The regime is supported by a strong research effort and its effectiveness is monitored systematically.

The slope safety regime in Hong Kong appears to be the product of a comprehensive design. But it is not the creation of a grand plan, rather it is the result of step-wise evolution over a thirty-year period. Each step, every new element of the regime, came about as a reaction to one or more multi-fatality landslides that had aroused strong public concern [2]. Hong Kong's case is not unusual in this regard - looking at safety regimes in other disaster scenarios we find that crisis-driven evolution is the norm. However, as we shall see later, safety regimes which develop reactively have severe disadvantages and, until the time eventually comes when the whole is in place, do not meet the needs of the parties involved. If needs are to be met, a system must be designed to meet needs. But how is such a design to be carried out?

Research has shown that a slope safety system designed on risk management principles, centrally coordinated by a safety manager and with effective public communication by the authorities responsible will largely meet the needs of the community. The finding is likely to be generally valid and so this paper has been written as a guide to establishing such a safety system, for the benefit of interested parties in cities with slope safety problems. Before proceeding, the concepts used in the paper will be introduced and terms defined.

Let us treat Hong Kong's slope safety regime as a 'system', as defined above. We can recognise the 'parties' involved in the system, their actions and interactions, and we can see the overall system operating plan as expressed in the Slope Safety Programmes. The system has an explicit objective, the reduction of landslide risk, expressed quantitatively [3], and the broader mission of the Safety Manager (the Geotechnical Engineering Office) is to meet the needs of the community. The outcomes of the system are monitored and a feedback mechanism is in place to make the improvements to the system shown to be needed by outcome monitoring. Whilst the end product meets the needs of the community, the evolutionary process by which the complete system was assembled has a number of undesirable features and in this regard Hong Kong's experience serves as a warning to others. The undesirable characteristics of crisis-driven evolution are summarised in Table 1.

Table 1. Features of crisis-driven evolution

The alternative to crisis-driven evolution which avoids these undesirable features is to commission a complete system from the outset. The design must be fit for purpose but what should be the design objective? The objective most likely to gain acceptance is 'to meet the needs of the parties'. What do we mean by 'needs'? By way of illustration let us consider some of the needs of one of the parties, the 'Resource Allocator', and examine the efficacy of a system design methodology based on the application of the principles of risk management. The approach we will adopt follows risk management methodology as applied in many countries to the management of major technological hazards by the public regulator [4].

Two acute questions facing the Resource Allocator will be 'how much resources should be allocated to slope safety?' and 'how can we prove to the investor (eg the taxpayer) that the investment has been resource effective?', Economic evaluation of a proposed slope safety project, following normal prudent commercial principles, would require that benefits exceed costs by an ample margin. But this rationale is unlikely to release sufficient resources to the project to satisfy public expectation. So then how much 'subsidy' is justifiable? Risk management policy as applied in the regulation of technological hazard provides the rationale to answer the question. If, say for a nuclear power station, estimated risk is 'unacceptably' high, when compared to national risk criteria, the hazard regulator will require that it should be reduced at all cost. However when risk is within the ALARP region (defined in the national risk criteria), the regulator will only require that it should be reduced to As Low As Reasonably Practicable, When applied by hazard regulators the ALARP requirement translates into a condition that investment in risk reduction should proceed to the point at which incremental 'trouble, time and expense' becomes grossly disproportionate to incremental benefit. If the risk management methodology is adopted, the Resource Allocator can work out the answer to the first question. The Resource Allocator will have a defensible allocation rationale.

The second question concerns performance measures. If the outcome of the slope safety project is monitored in terms of risk reduction achieved and the costs and benefits are quantified, the Resource Allocator has the evidence needed to measure resource effectiveness, a valid performance measure for the project.

This briefly illustrates the application of the methodology in the case of one party, but the needs of other parties are also to some degree served through the employment of the risk management rationale. But the risk management rationale alone is not a sufficient basis for system design. An effective public communication policy is also required, Risk creators (eg property owners), the Safety Manager and the Resource Allocator all need to engage in effective dialogue with the risk bearers. A public communication rationale has to be established with the aim of winning trust, promoting understanding and maintaining awareness. Experience shows that all of the parties communicate with each other mainly via the electronic and print media. So the media are a vitally important party in the safety management system.

Let us assume that a decision has been taken to set up a slope safety system. How is the city government to proceed? A model plan of action is given in Table 2. The design process, given in Table 3, starts with a scoping study and preparation of an outline system design.

Table 2. Sequence of Actions

This provides detail for the policy submission and preliminary resource bid to the Resource Allocator, timed on the annual budgetary cycle. With these applications in the pipeline the main design effort begins. This requires compilation of historical landslide and rainfall data-bases, so that historical landslide occurrence frequencies and detriment trend can be established. Development of a hazard model follows, along with frequency and consequence assessments for the global quantified risk analysis (QRA)[5]. The QRA should determine present risk and past risk trend.

Table 3. System Design Process

The 'needs analysis' is an essential part of system design (item 3, Table 3). Its purpose is to elicit the needs of the parties through opinion survey and later to check the degree to which each of the risk reduction packages potentially meets needs. Then follows the quantification of the costs and benefits (cost/benefit analysis CBA) of the various risk reduction packages which the System Manager will offer as options to the Resource Allocator. Each package consist of several measures, some to reduce frequency of occurrence of landslides and others to cut exposure of people to harm and property to damage. Each of the packages will be justified using the ALARP rationale in quantified form (invest until incremental cost becomes grossly disproportionate to incremental benefit) or qualitatively ('do your best') for the non-quantifiable elements. Each package includes a mechanism for measuring system outcome and resource cost per unit of outcome ('effectiveness'), a device for providing feedback to improve future outcomes and a public communications plan. The Resource Allocator will in due course decide what is 'reasonably practicable' and make public the reasoning. A package of risk reduction measures might include some emergency preparedness actions, a public education campaign, enhancements of control of slope works, certain funding for old slope retrofit, etc. put together as an initial say 5-year plan. The measures, along with the associated research, monitoring and public communications actions would be implemented through long-term Slope Safety Programmes (Section 1).

Once detailed design is completed the Safety Manager is able to make a detailed and fully justified application to the Resource Allocator. The programme given in Table 2 assumes this application is made in Year-1 month-9 and that there is a 3-month period between the date of this application and receipt of mandate for whatever risk reduction package is chosen by the Resource Allocator. Major capital funding approvals and legislative amendments will take longer than three months, but the associated start dates for remedial/retrofit works and regulatory upgrading are phased in the second half of Year-2. Having completed system design, the Research Unit shifts emphasis to procurement (Table 4) and preparatory technical work for Year-2.

The second year sees the beginning of intervention, starting with potentially the most resource effective/quick acting measures such as emergency preparedness actions and public education campaigns. Before the public launching of the emergency preparedness programme, preparatory work will be needed with the media if they are to help the Safety Manager to train up the other parties. The new data acquisition systems for rainfall, landslides, slope inventories and social survey will come into operation in Year-2, so that by the end of that year the first annual review can be carried out. Once this data is available it should be put on the Internet for open access. This will save community costs and aid consumer choice. In the second half of Year-2 a start will be made to upgrade the control regime for new development, including town planning and building control.

Later in Year-2 the Regulator will start to administer the voluntary or mandatory programme of remedial and retrofit works to old slopes by property owners, governmental and private. The retrofit programme should be very carefully planned as it will prove to be the most expensive component of the system, as we shall now see.

Costs

Table 4. Costs of a Slope Safety Management System

Party	Staffing	Procurements
Safety Manager	manager + 2 GE	landslide investigation consultancy (US$1.2m per year)
Research Unit	head + 4 GE	geotechnical/risk consultancy for databases, data acquisition system, QRA/CBA, risk zonation plans, etc (US$0.5m per year) weather radar (US$1.2m) automatic raingauges (US$0.2m + US$3000 per gauge per year) public education campaign, social survey, media training & public relations consultancy (US$0.2m per year) or checking (review) consultancy (US$20,000 per permit)
Regulator	head + 2 GE technical secretariat + 1 GE per 30 permits per year (GE also handles slope repair notices, landslide inspection, etc.)
Notes: 1. 1 GE = a geotechnical engineer + share of divisional technical & clerical staff (basic salary cost US$0.15m per GE) + share of departmental expenses, not rent (US$30,000 per GE). 2. Annual costs at 1997 prices in Hong Kong, where human resource costs are very high. 3. Number of GEs needed depends on number of registered hazardous features, availability of technical support elsewhere in the organisation, etc. Number given is the minimum.

An indicative cost estimate for a notional Slope Safety System is given in Table 4. It should be noted that these rates applied in Hong Kong in 1997. Appropriate adjustments will be needed for other Asian cities. What are the relative costs of this notional system compared to retrofitting? The cost of retrofit works to a typical 60' cut slope in saprolite 50m long x 15m high might be HK$2.5m (1997 prices). These include soil nailing, raking drains and surface protection and drainage. The annual cost of the entire safety system given in Table 4 (assuming weather radar amortised over 10 years, 20 raingauges, 150 permits per year, manager and heads cost 2xGE, GE on cost factor 2.0) is about the same as the cost of only 30 retrofits. The annual cost of the slope works control function only is equal to 10 retrofits.

The cost of ten retrofits, out of the thousands likely to be needed, is surely a small annual price to pay for ensuring that slope works in 150 new projects per year are of an adequate quality. But this logic alone, based as it is on the avoidance of future loss, is unlikely to convince hard-pressed city authorities to invest in slope controls. However, once the landslide problem has become a social issue, city managers might be attracted by the possibility of preventing of all of the social and political problems that would certainly accompany future disasters, some of which are given in Table 1. In this regard it is helpful to examine the history of slope control in the two cities which are perhaps the most advanced in this area: Rio de Janeiro and Hong Kong.

When was effective regulation of hillside development introduced in Hong Kong and Rio de Janeiro?

The two cities have much in common: similar population, land areas, adverse terrain and climate. Both cities experienced rapid economic growth, immigration and building boom in the second half of the twentieth century, resulting in land shortage and the growth of squatter communities comprising patrial immigrants living at risk in shantytowns on steep hillsides. Early hillslope development was permitted without effective safety controls. In both places a succession of multi-fatality landslides occurred, provoking public outcry. The authorities subsequently brought in effective slope safety regulation (Table 5). Both cities were relatively poor at the time compared to cities in the developed countries, but slope regulation had become imperative politically.

Table 5. Regulation in relation to Disaster

Regulation
Introduced

Prior Landslide
Fatalities

Population at the time

GDP per capita
at the time

Rio de Janeiro

1966

73 (1962& 1966)

3.8m

? (US$ 6000 in 1997)

Hong Kong

1977

175 (1972 & 1976)

4.5m

US$ 1600

(US$ 25,000 in 1997)

When should effective control of new slope works be introduced?

In these two cases timely pre-emptive action would have avoided much loss of life and social crisis. But how do the authorities know when to take such action? In deciding, it may be helpful to examine detriment trends and this can be done quite cheaply. If a graph of landslide fatalities in Hong Kong per year versus time had been drawn up in 1966, after the disastrous June rainstorm, a disturbing trend would have emerged (see Figure 1, Malone 1997 [1]). Landslide fatality figures for squatters had been rising with population growth since the 1940s and by the 1960s landslide fatalities were just starting to occur post-war on roads and in buildings. Reason perhaps to consider corrective action?

Had a good slope safety system been introduced in 1966 in Hong Kong, 200 lives may not have been lost subsequently due to landslides and Hong Kong's slope retrofit bill may have been reduced by 25% or more. But Hong Kong is now relatively wealthy and the government, much the biggest landowner, is apparently able to afford the level of investment in slope retrofit expected of it by the community. In contrast, other cities with similar adverse climate, terrain and population pressures may, in time, find that they are unable to afford to pay for the retrofit needed to drive risk down a level which the community is prepared to tolerate. Not only is risk rising apace in times of economic boom with unregulated development, but tolerance levels are failing with increased political liberty, universal education, widespread wealth and a free press [6]. The logical answer to the headline question is the sooner the better and well before the essential retrofit bill becomes unaffordable.

Conclusions

A design methodology and action plan has been given to help interested parties to establish a slope safety system, designed on risk management principles, which meets the needs of the community. When should such a system be introduced? It is prudent economically and politically wise for governments to introduce these systems before their essential slope retrofit bill becomes unaffordable and before being forced to do so in the wake of repeated disaster.

References

1. Malone, A.W. (1997). Risk Management and Slope Safety in Hong Kong. Transactions of the Hong Kong Institution of Engineers, 4, No.2, pp 12-21.

2. Malone, A.W. & K.K.S. Ho (1995). Learning from landslip disasters in Hong Kong. Built Environment, 21, Nos. 2/3, pp. 126-144.

3. Government of the Hong Kong Special Administrative Region of the Peoples Republic of China (1998). Policy Address by the Chief Executive: Slope Safety for All, Policy Objective for Works Bureau. HKSAR Government.

4. Health and Safety Executive (1988). The tolerability of risk from nuclear power stations. London: HMSO.

5. Wong, H. N. & K.K.S.Ho (1998). Overview of risk of old man-made slopes and retaining walls in Hong Kong. Slope Engineering in Hong Kong, Balkema, pp. 193-200.

6. Bond, M.H. (1991). Beyond the Chinese face: Insights from psychology, Hong Kong, Oxford University Press.

[BACK]	[2. Managing Slope Safety]