An examination of the dataset revealed that the concentration of TBT in sediments taken from most of the sampling stations were reported below the analytical detection limit of 20 mg Sn kg-1 dry weight. Figure 2.3b shows the mean concentration of TBT detected in sediments collected from each sampling station during the December 1998, January, June and August 1999 surveys.

Sediments from 7 of the 10 sampling stations were reported as containing TBT concentrations at or below the analytical detection limit. The three stations which reported values above the detection limit during at least one of the four sampling periods were VS3 in Victoria Harbour (20.8, 36.2, 34.6 and 46.6 mg Sn kg-1 in December 1998, January, June and August 1999, respectively); WS1 which is located south of Tsing Yi (81.8, 103.0, 107.0 and 93.0 mg Sn kg-1 in December 1998, January, June and August 1999, respectively); and, Station NS4 located in Urmston Road where TBT in sediments was found at 39.2 mg Sn kg-1 during August 1999 sampling. TBT was reported below detection limits (ie 20.0 mg Sn kg-1 dry weight) at Station NS4 in the three previous surveys conducted.

The routine monitoring dataset gathered by EPD on sediment chemistry does not include measurement of TBT levels. No ERL, LCEL or ISQV-Low values are presently available for TBT and levels found above detection limits in these surveys cannot be given either non-exceedance (ie uncontaminated) or exceedance (ie contaminated) criteria. TBT concentrations determined in this Study may, however, be compared with datasets obtained in previous studies in Hong Kong. Data are available from a variety of other studies and these have been compared with the data gathered during the Environmental Baseline Surveys. Other datasets compared consisted of the following studies:

• Tributyltin Contamination in Coastal Sediments in Hong Kong(13)) ;

• Strategic Sediment Sampling Program (SSSP)(14)) ; and,

• A Study of Tributyltin Contamination of the Marine Environment in Hong Kong(15)) .

The majority of samples collected during the above studies were taken from sites suspected of being contaminated with TBT, such as typhoon shelters, marinas, marine traffic areas, mooring areas and container terminals. The studies also sampled reference areas, however, that were supposedly remote from sources of possible TBT contamination. These locations included areas such as the West Lamma Channel (EPD Station SS3), southern Lamma (within the South Lamma Potential Marine Park / Marine Reserve) and outer Tathong Channel (EPD Station ES2 that was also sampled in the present study).

Concentrations of TBT at stations VS3 and WS1 were similar to those recorded from other moderately contaminated sites sampled in the above studies such as Stonecutters Island (84 mg Sn kg-1), but less than sites where TBT contamination was considered high, eg Causeway Bay Typhoon Shelter (11,618 mg Sn kg-1), Aberdeen Marina (3,468 mg Sn kg-1) and the Rambler Channel (1,970 mg Sn kg-1).

The environmental significance of the TBT values recorded in the present Study are discussed in more detail in Section 4.2.

(6)EVS (1996) Review of Contaminated Mud Disposal Strategy and Status Report on Contaminated Mud Disposal Facility at East Sha Chau. Prepared for CED.

(7)Long ER et al (1996) Development and evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology. Cited in EVS 1996 Op cit.

(8)Mouchel (1996) Strategic Sewage Disposal Scheme Stage I, Baseline Monitoring and Verification. EM&A Manual EPD.

(9)EVS (1996) 1996 Strategic Sediment Sampling Programme (SSSP): Interpretive Report (Final). Prepared for CED.

(10)The factor 'period' has been nested within 'station' because differences between months at a predefined station are of interest but comparisons between months at different stations are not.

(11)The detection of significant differences requires less power when using a fixed factor model than a random factor model as the analyses utilises degrees of freedom associated with the residual term (the highest degrees of freedom) as the enominator is in the test. The denominator in a random factor model will be one of the other specified factors in the test and consequently will have lower degrees of freedom.

(12)Underwood AJ (1997) Experiments in Ecology: their logical design and interpretation using analysis of variance.

(13)CES Asia (1995) Tributyltin Contamination in Coastal Sediments in Hong Kong. Report for EPD.

(14)EVS (1997) Op cit.

(15)Aspinwall (1998) A Study of Tributyltin Contamination of the Marine Environment in Hong Kong. Final Report for EPD.