The EIA conducted for this Project has indicated that
benthic fauna are expected to recolonise the pits following capping with
uncontaminated mud. It is expected
that recolonisation of the natural benthic assemblage will occur and eventually
the benthic assemblage will resemble that of the surrounding areas. Recolonisation may be achieved by
larval recruitment, influx of juveniles or adults carried in water currents, or
through the active swimming or crawling of individuals. However, other natural (eg storm
events, hypoxia, salinity fluctuations) or anthropogenic (eg pollution,
dredging activities and fisheries operations) activities may hinder
recolonisation of capped pits. As
a result, the factors contributing to the composition of the benthic assemblage
may be difficult to determine. It
is also important for any recolonisation studies to be aware of any cap
maintenance (or "topping up") activities which may also impact the
resident benthic assemblages.
In order to verify the recolonisation of marine biota on the
capped pits, a benthic recolonisation programme is recommended. The full details of the EM&A
programme for benthic recolonisation are presented in the following sections.
The objective for this component of the Study is to monitor
and report on the benthic recolonisation of the capped pits including the
previous ones and specifically to determine the difference in infauna between
the capped pits and adjacent sites.
The impact hypothesis for this task is as follows:
Recolonisation is occurring at
the capped pits such that assemblages at the capped pits become more similar to
reference assemblages as time since capping increases.
The null hypothesis to be tested for this work component is
as follows:
H0 There is no difference in the
structure of benthic infaunal assemblages found at the capped pits at the
active facility and adjacent reference areas.
The sampling design of this task involves two treatments:
capped pits and reference areas (Figures 8.4a and 8.4b). The capped pit treatment shall involve
collection of samples from the capped mud pits at the active facility. The pits are anticipated to be capped
at different times. The number of
samples to be collected at each station, the number of impact stations and
their locations should be confirmed prior to the commencement of capping
activities and agreed with EPD based on the detailed design of the disposal
facility.
The second treatment shall involve sampling at different
reference sites where the number of sample stations and the exact position of
the sampling sites should be determined prior to the commencement of the
capping activities and agreed with EPD based on the detailed design of the
disposal facility. The reference
sites should be chosen to improve the balanced nature of the design (ie the
number of reference sites should be the same as the impact sites). Using multiple controls is an effective
way of ensuring that the extremely variable nature of Hong Kong's marine
benthos from one site to another does not overly influence or alter the
results. Current ecological theory
suggests that the use of multiple control sites in sampling designs are
statistically more robust and hence the conclusions more reliable ([1])
(2).
One grab sample shall be taken at each of the different
stations clustered within each site.
The analysis (Cumulative Running Mean Test to determine optimum sample size
for stabilising mean and standard error values) revealed that 12 samples would
be the optimum number. The
technique of clustering stations within one site has been proven to be an
effective way of testing hypotheses and removing the confounding effects of
spatial variation from the interpretation. The number of samples to be collected at each site and the
exact locations of the sampling stations should be determined prior to the
commencement of capping activities based on the detailed design of the disposal
facility.
The data collected during the monitoring programme shall be
analysed using two different but complementary approaches as detailed
below.
ANOVA & MANOVA: Simple, univariate measures shall be tested using an
Analysis of Variance (ANOVA), and multivariate measures of community structure
shall be tested using the Multiple Analysis of Variance (MANOVA). Both ANOVA and MANOVA test the same null
hypothesis using similar methods.
The method is essentially a comparison of the variability within a site
to the variability between sites.
If the ratio of these two variances (that is, the between‑group‑variance
over the within‑group‑variance) is large enough, then any differences observed
are due to true differences that exist between the groups and not just to
random variation. ANOVA and MANOVA
tests are based on several assumptions related to the underlying distribution
of the data being analysed (ie normality, homogeneity of variances). If the data deviate significantly from
these assumptions, then these tests are considered to be inappropriate. If this situation arises, alternative
procedures (ie parametric tests with rank transformed data or nonparametric
analogues such as Kruskal Wallis) which address similar hypotheses but do not
require such stringent assumptions shall be adopted. Observed differences between the sites and/or areas shall be
tested using multiple comparison procedures such as the Student Newman Keuls
(SNK) or Tukey test.
Multi Dimensional Scaling
(MDS): Multi dimensional scaling (MDS) shall
also be used to depict the similarities between stations based on their benthic
assemblages. MDS is a method for creating
a low dimensional picture of the relationships between stations in a complex,
multi dimensional problem. The
Bray-Curtis distance metric shall be used for both the clustering techniques
and the MDS. The dendrogram from
the cluster analysis, and the MDS ordination plot will provide complementary
views of the same similarity information.
The data for MDS and cluster analyses should be standardised prior to
analysis, to ensure that bias resulting from including data in different forms
(eg percent data for silt clay composition, numerical data for abundances and
biomass data in mg) does not occur.
The detailed statistical analyses described above shall be
used to comprehensively explore the benthic assemblage patterns in the area of
the active pits. This exploration
should lead to conclusions regarding the effectiveness of the cap material in
promoting post-dredging benthic assemblages. This information can be used to reassess the choice of
capping materials or revise the procedures for placing the caps at the active
facility.
The benthic sediment samples collected during this task will
be analysed for the following parameters:
·
Percentage
of silt/clay in the sediments;
·
Faunal Abundance;
·
Faunal Biomass;
·
Species
Composition; and,
·
Trophic
Structure.
The sampling team and vessel will be deployed and accurate
positioning attained as described in Section 4. The vessel will be equipped with
adequate fixed sieve stations to facilitate rapid processing of samples and
ensure the maximum number of samples are collected in each survey. At each of the designated benthic
sampling stations, seafloor sampling will be carried out with a modified Van
Veen grab sampler (dimensions 30 cm H 30 cm H 30 cm) or similar instrument
approved by EPD/AFCD. One
subsample of approximately 1 kg sediment shall be collected from each sample
for analysis of particle size. The
remaining sediment from each sample shall be used for sorting. Samples will be labelled and sieved
through a 1 mm and 0.5 mm sieve and all residues and organisms retained,
double-bagged and preserved in 4% buffered formalin in seawater. A vital stain (eg Rose bengal) will be
added to distinguish organic materials and organisms from other non-living
residues. The grab and utensils
will be washed thoroughly with seawater after each deployment to avoid
cross-contamination between samples.
On completion of the survey all samples will be transferred to the
laboratory for sorting and identification. All sediment sieving will be conducted by qualified marine
scientists who will oversee and coordinate all field operations.
Upon arrival at the laboratory, all benthic samples should be
re-inventoried and checked against chain-of-custody forms. Sample rescreening should be performed
after the samples have been held in formalin for a minimum of 24 hours to
ensure adequate fixation of the organisms. Individual samples from the 500 mm and 1 mm2 mesh
sieves will be gently rinsed with fresh water into a 250 mm sieve to remove the formalin
from the sediments. Sieves will be
partially filled while rinsing a specific sample to maximize washing efficiency
and prevent loss of material. All
material retained on the 250mm sieve is placed in small fractions into
a labelled petri dish and preserved with 70% ethanol. The material is lightly agitated to ensure complete mixing
of the alcohol with the sediments.
The sediment is then sorted to remove all animals and fragments. Original labels will remain with the
rescreened sample material.
Standard and accepted techniques shall be used for sorting
organisms from the sediments ([1]). Small fractions of a sample will be placed in a petri dish
under a 10-power magnification dissecting microscope. The petri dish will be scanned systematically and all
animals and fragments removed using forceps. Each petri dish will be sorted at least twice to ensure
removal of all animals. Organisms
representing major taxonomic groups including Polychaeta, Arthropoda, Mollusca,
and miscellaneous taxa will be sorted into separate, labelled vials containing
70 percent ethanol. All sorted
samples will be systematically checked to ensure compliance with QA/QC program
requirements before proceeding to the taxonomic identification, enumeration,
and biomass determination phases of the analysis.
Taxonomic identifications will be performed by regional
taxonomic experts using stereo dissecting and high-power compound microscopes,
to the family level except for dominants, which will be identified, where
possible, to species. The careful
sampling procedure employed in the Study will minimise fragmentation of
organisms, however should breakage of soft-bodied organisms occur, only
anterior portions of organism fragments will be counted. All fragments will be retained and
weighed during biomass determinations, described below. Rare or questionable taxa will be
compared against reference collection specimens for confirmation and consistency
of identification. The
nomenclature used in all reference collections referred to in this study should
be cross checked and differences or discrepancies should be noted. Biomass determinations will be made by
taking the blotted wet mass of each taxonomic fraction.
Sorting QA/QC will be performed using 25-power magnification
by someone other than the original sorter. Twenty percent of each sorted sample should be resorted to
ensure 95 percent sorting efficiency.
A sample passes QA/QC if the number of organisms found during the QA/QC
check does not represent more than 5 percent of the total number of organisms
found in the entire sample. If the
number of organisms found is greater than 5 percent of the total number, the
entire sample will be resorted.
Any samples where the identification of taxa is questionable will be
sent out for independent reidentification by a qualified regional expert. Reference collections developed during
previous seabed and benthic studies in Hong Kong should be consulted as
necessary.
([1])
RJ
Schmitt & CW Osenberg(1996) Detecting Ecological Impacts: concepts and
applications in coastal habitats. Academic Press.
([3]) Holme, N. A. and
A. D. McIntyre (eds) (1984) Methods for the study of marine benthos. Blackwell Scientific Publications,
Oxford (UK).