1 Topography, Land Use and Population

This section describes the methodology employed to determine the population in the vicinity of the proposed site at South Soko as well as the population to be adopted in the Quantitative Risk Assessment (QRA) study. The QRA considers the years 2011 and 2021 in the analysis and so the population was estimated for these years.

1.1 Location and Topography

The proposed site for the LNG Terminal on South Soko Island is located at the south of Lantau Island, in the south-western waters of Hong Kong. South Soko Island is currently an uninhabited island which has historically housed villages and more recently a detention centre operated by the Hong Kong Government. The majority of the habitat is hillside and rocky shore typical of offshore islands in Hong Kong (Figures 1.1 and 1.2

depict South Soko Island and the area under consideration for population study respectively).

1.2 Current Land Use

A number of Private Lots and derelict buildings are present on the Island as well as the recently refurbished Tin Hau Temple, but there is no permanent development or resident population at either South Soko or North Soko at present. A radioactive storage facility is located on North Soko Island but there are no permanent residents and waste is delivered just twice a year.

1.3 Land Population Estimation

The following information sources were referred to for population estimation:

· Site Survey Data

· Census Data [1]

· Land Records from Lands Department

· Road Traffic Data [2]

· Data on Key Individual Developments

· Marine Traffic Data [3-5]

1.3.1 Residential Population

South Soko is uninhabited. North Soko also has no residential population and none is planned (TPU 934) [6]. This was confirmed by a site survey, which found no population currently on either island.

Figure 1.1 Aerial Photo of the South Soko Site

Figure 1.2 Population in Vicinity of South Soko

The population in villages along the southern coast of Lantau was estimated based on a buildings count and the assumption of 3 persons per unit. The population of major buildings such as the prison and hospital was researched directly to obtain reliable numbers. Finally, population figures were scaled to obtain predictions for 2011 and 2021 based on the Planning Vision Strategy (PVS) [7]. However, the PVS predicts no change in population up to 2021. The residential population used in the assessment of South Soko Terminal are summarised in Table 1.1.

Table 1.1 Estimated Residential Population Data

Location	Approx. Distance from Terminal Site	2011 Population	2021 Population
Fan Lau Tai Long Wan Tseun Shek Pik Prison Tung Wan School Shui Hau Ma Po Ping Prison Tong Fuk Cheung Sha Villas Cheung Sha San Shek Wan Pui Wo Ham Tin Mong Tung Wan Sea Ranch Tai Long Shek Kwu Chau South Lantau Hospital	7km 7km 7km 6km 6km 6km 7km 9km 9km 10km 11km 11km 11km 10km 11km 8km 9km	52 293 824 152 375 1,728 1,133 665 597 156 2,628 859 43 313 34 43 90	52 293 824 152 375 1,728 1,133 665 597 156 2,628 859 43 313 34 43 90

1.3.2 Industrial Population

A low level radioactive waste storage facility is located on North Soko Island. The facility is about than 1.5km from the South Soko site. Waste is delivered at most twice a year. No permanent working population is present. No other industrial facilities are located within the vicinity of South Soko.

1.3.3 Road Traffic Population

There is only one road running along the southern edge of Lantau Island, namely South Lantau Road. The population estimation is based on the 2005 Annual Traffic Census [2]. The AADT value is 2170 vehicles per day for station number 5859 from Tung Chung Road to Sham Wat Road. Assuming an average speed of 50km/hr and an average of 3 persons per vehicle, the number of persons on the road is:

No. of persons = (AADT x Vehicle Occupancy / 24 / Speed)

= 2,170 x 3 / 24 / 50 = 5 persons/km

The traffic along this section of road has decreased in recent year as more convenient modes of access to Lantau Island have become available. The population for 2011 and 2021 is therefore assumed to be the same at 5 persons/km.

1.3.4 Occupancy and Indoor/Outdoor Fractions

The land population is categorised further into 4 time periods: night time, weekday, peak hours and weekend day. These are defined in Table 1.2.

Table 1.2 Population Time Periods

Time Period	Description
Night time Weekday Peak hours Weekend day	7:00pm to 7:00am 9:00am to 5:00pm Monday through Friday, and 9:00am to 1:00pm Saturday 7:00am to 9:00am and 5:00pm to 7:00pm, Monday to Friday 7:00am to 9:00am and 1:00pm to 3:00pm, Saturdays 3:00pm to 7:00pm Saturdays, and 7:00am to 7:00pm Sundays

The occupancy assumed [8] during these time periods is given in Table 1.3. Different occupancy figures are assumed for each category of land population. The proportion of the population outdoors is also assumed to vary according to type of population and time period (Table 1.3).

The hazards that can potentially affect offsite population are flash fires and thermal radiation from pool fires. Buildings are assumed to offer protection to its occupants for these events. The protection factor used is 90%, or equivalently the exposure factor is 10%. Scenarios are therefore assumed to affect 100% of the outdoor population and 10% of the indoor population.

Road vehicles are also assumed to offer some protection, although less than a building. An exposure factor of 50% is used for vehicles.

Table 1.3 Land Population Occupancy and Indoor/Outdoor Fractions

Population	Occupancy				% Outdoors
Type	Night	Peak	Weekday	Weekend day	Night	Peak	Weekday	Weekend day
Residential Prison Hospital School Road	100 % 100 % 100 % 0 % 10 %	50 % 110 % 120 % 10 % 100 %	20 % 100 % 110 % 100 % 50 %	80 % 110 % 120 % 10 % 20 %	0 % 5 % 0 % 0 % 0 %	30 % 100 % 30 % 100 % 0 %	10 % 50 % 10 % 20 % 0 %	20 % 50 % 30 % 20 % 0 %

1.4 Marine Population Estimation

South Soko Island lies within 4km of the Adamasta Channel. The marine traffic in this fairway consists of mainly fast ferries serving Macau.

1.4.1 Vessel Population

The vessel population used in this study are as given in Table 1.4. The figures are based on BMT’s Marine Impact Assessment report [4] except those for fast ferries. The maximum population of fast ferries is assumed to be 450, based on the maximum capacity of the largest ferry operating in the Adamasta Channel. However, the average load factors for fast ferries to Macau and Pearl River ports are 52% and 37% respectively while the overall average load factor considering all ferries is about 50% [5]. Hence, a distribution in ferry population was assumed as indicated in Table 1.4. This distribution gives an overall load factor of about 58% which is conservative and covers any future increase in vessel population.

Table 1.4 Vessel Population

Type of Vessel	Average Population per Vessel	% of Trips
Ocean-Going Vessel Rivertrade Coastal vessel Fast Ferries Tug and Tow Others	21 5 450 (largest ferries with max population) 350 (typical ferry with max population) 280 (typical ferry at 80% capacity) 175 (typical ferry at 50% capacity) 105 (typical ferry at 30% capacity) 35 (typical ferry at 10% capacity) 5 5	3.75 3.75 22.5 52.5 12.5 5.00

1.4.2 Marine Vessel Protection Factors

The population on marine vessels is assumed to be offered some protection by the vessel structure, in a similar way that buildings offer protection to their occupants. The degree of protection offered depends on factors such as:

· Size of vessel

· Construction material and likelihood of secondary fires

· Speed of vessel and hence its exposure time to the flammable cloud

· The proportion of passengers likely to be on deck or in the interior of the vessel

· The ability of gas to penetrate into the interior of the vessel and achieve a flammable mixture.

Small vessels such as fishing boats will provide little protection but larger vessels such as ocean-going vessels will provide greater protection. Fast ferries are air conditioned and have a limited rate of air exchange with the outside. Based on these considerations, the fatality probabilities assumed for each type of vessel are as given in Table 1.5.

Table 1.5 Population at Risk

Marine Vessel Type	Population	Fatality Probability	Population at Risk
Ocean-Going Vessel Rivertrade Coastal Vessel Fast Ferries Tug and Tow Others	21 5 450 350 280 175 105 35 5 5	0.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.9 0.9	2 2 135 105 84 53 32 11 5 5

1.4.3 Methodology

In this study, the marine traffic population in the vicinity of South Soko has been considered as both point receptors and average density values. The population of all vessels are treated as an area average density except for fast ferries which are treated as point receptors.

The marine area around South Soko was divided into 12.67km² grid cells, each grid being approximately 3.6km x 3.6km. The transit time for a vessel to traverse a grid is calculated based on the travel distance divided by the vessel’s average speed. The average speed [3] and transit time for different vessel types are presented in Table 1.6.

Table 1.6 Average Speed and Transit Time of Different Vessel Type [5]

Type of Vessel	Assumed Speed (m/s)	Transit Time (min)
Ocean-going vessel	6.0	9.9
Rivertrade Coastal vessel	6.0	9.9
Fast Ferries	15.0	4.0
Tug and Tow	2.5	23.7
Others	6.0	9.9

The number of vessels traversing each grid daily was provided by the marine consultant [3]. These are provided in Table 1.7, where the grid cell reference numbers are defined according to Figure 1.3. The number of marine vessels present within each grid cell at any instant in time is then calculated from:

Number of vessels = No. of vessels per day x grid length / 86400 / Speed (1)

This was calculated for each type of vessel, for each grid and for years 2011 and 2021. The values obtained represent the number of vessels present within a grid cell at any instant in time. Values of less than one are interpreted as the probability of a vessel being present.

Figure 1.3 Grid Cell Numbering Scheme

Table 1.7 Number of Marine Vessels Per Day

Grid No.	Average Number of Vessels Per Day
	2011					2021
	OG	RT	TT	FF	OTH	OG	RT	TT	FF	OTH
1 2 3 4 5 6 7 8 9 10 11 12	0 0 0 0 0 0 0 0 0 0 0 0	21 0 0 0 0 21 84 63 0 0 0 0	0 0 0 0 0 21 53 21 0 0 0 11	99 121 143 121 121 99 0 0 0 0 0 0	284 168 210 210 210 210 95 63 11 11 11 168	0 0 0 0 0 0 0 0 0 0 0 0	23 0 0 0 0 23 92 69 0 0 0 0	0 0 0 0 0 23 58 23 0 0 0 12	117 143 169 143 143 117 0 0 0 0 0 0	311 184 230 230 230 230 104 69 12 12 12 184

OG = Ocean-going vessels

RT = Rivertrade coastal vessels

TT = Tug & tow vessels

FF = Fast ferries

OTH = others

Average Density Approach

The average marine population for each grid is calculated by combining the number of vessels in each grid (from Equation 1) with the population at risk for each vessel (Table 1.6). The results are shown in Figures 1.4 and 1.5. This grid population is assumed to apply to all time periods. Note however that fast ferries are excluded since ferries are treated separately in the analysis (see below).

When simulating a possible release scenario, the impact area is calculated from dispersion modelling. In general, only a fraction of the grid area is affected and hence the number of fatalities within a grid is calculated from:

Number of fatalities = grid population x impact area / grid area (2)

Figure 1.4 Marine Population at Risk by Grid, Year 2011

Figure 1.5 Marine Population at Risk by Grid, Year 2021

Point Receptor Approach

The average density approach, described above, effectively dilutes the population over the area of the grid. Given that ferries have a much higher population than other classes of vessel, combined with a relatively low presence factor due to their higher speed, the average density approach would not adequately highlight the impact of fast ferries on the FN curves. Fast ferries are therefore treated a little differently in the analysis.

In reality, if a fast ferry is affected by an accident scenario, the whole ferry will likely be affected. The likelihood that the ferry is affected, however, depends on the size of the hazard area and the number density of ferry vessels. To model this, the population is treated as a concentrated point receptor i.e. the entire population of the ferry is assumed to remain focused at the ferry location. The ferry density is calculated the same way as described above (Equation 1), giving the number of ferries per grid at any instant in time, or equivalently a “presence factor”. A hazard scenario, however, will not affect a whole grid, but some fraction determined by the area ratio of the hazard footprint area and the grid area. The presence factor, corrected by this area ratio is then used to modify the frequency of the hazard scenario:

Prob. that ferry is affected = presence factor x impact area / grid area (3)

The fast ferry population distribution adopted was described in Table 1.5. Information from the main ferry operators suggests that 25% of ferry trips take place at night time, while 75% occur during daytime. Day and night ferries are therefore assessed separately in the analysis. The distribution assumed is given in Table 1.8.

Table 1.8 Fast Ferry Population Distribution for Day and Night Time Periods

Population	Population at Risk	% of Day Trips	% of Night Trips	% of All Trips (= 0.75 x day + 0.25 x night)
450 350 280 175 105 35	135 105 84 53 32 11	5 5 30 60 - -	- - - 30 50 20	3.75 3.75 22.5 52.5 12.5 5.0

The ferry presence factor (Equation 1) and probability that a ferry is affected by a release scenario (Equation 2) are calculated for each ferry occupancy category and each time period.

1.5 Airborne Population

Helicopters shuttling to and from Macau pass fairly close to the South Soko site. Flights travelling westward towards Macau pass to the north of South Soko, while the return flights pass to the south of the island. The impact of accidental releases of LNG from the terminal on these helicopters was included in the analysis by including the helicopter passengers in the population.

Helicopters were analysed the same way as fast ferries, namely as mobile receptors with a presence factor, or probability of being within a grid cell when an accident occurs. The cruising speed of the Macau helicopters is 254 km/h. It therefore takes just 51s for a helicopter to traverse one of the grid cells shown in Figure 1.3. There are 27 return flights per day; one flight every 30 minutes during the daytime and early evening. Applying Equation 1, the helicopter presence factor becomes:

Presence factor = No. of flights per hour x grid length / 3600 / Speed = 0.028 (4)

This presence factor is applied to all grid cells in Figure 1.3. Westbound flights are assumed to pass through the upper row of grids (cells 1-6) while the return eastbound flights pass through the lower row of grid cells (7-12). This presence factor is applied to the daytime period while night time is assumed to have no flights. Each helicopter is assumed to be full with 12 passengers and crew.

The presence factor is further modified by the size of the impact area for each scenario by applying Equation 3.

It should be noted that this treatment of helicopters is very conservative. Effectively, the population are treated as being at ground level. In reality, these helicopters fly at 500 feet and so few release scenarios will be able to impact on them. However, helicopters are not expected to make a significant contribution in the results and hence this simplistic approach is regarded as sufficient.

Marine vessels and road vehicles were assumed to offer some protection to their occupants. The same assumption was not applied to helicopters since they will likely crash if they are impacted by an LNG release. 100% fatality is assumed.

References

[1] www.censtatd.gov.hk

[2] The Annual Traffic Census 2005, Transport Department, Hong Kong SAR, Jun 2006.

[3] BMT Asia Pacific Ltd., personal communication, 2006

[4] BMT Asia Pacific Ltd, Marine Impact Assessment for Black Point & Sokos Islands LNG Receiving Terminal & Associated facilities, Pipeline Issues, Working Paper #3, Issue 8, Sep 2006

[5] Passenger Arrivals/Departures and Passenger Load Factors at Cross-Boundary Ferry Terminals, January to December 2005, Marine Department, Hong Kong SAR.

[6] Projected Hong Kong Resident Population by TPU, Planning Department, Hong Kong SAR, 2004

[7] Hong Kong 2030, Planning Vision and Strategy, Planning Department, Hong Kong SAR.

[8] ERM, Liquefied Natural Gas (LNG) Terminal and Associated Facilities – Marine Quantitative Risk Assessment, Population Survey Report, Jun 2006.