INVESTIGATING THE REGENERATIVE PROPERTIES OF RECLAIMED SITES IN SINGAPORE BY COMPARATIVE ANALYSIS OF SOFT-BOTTOM MACROBENTHIC COMMUNITIES

Abigail Lum Yan Yi Yiong Ziyi Olivia

NUS High School of Mathematics and Science

Little Green Dot Student Research Grant PROJECT REPORT

submitted to

Nature Society (Singapore)

Junior College Category 2011 3, 828 words

ABSTRACT

Singapore has undergone extensive reclamation since the 1960s and the activity has since posed major threats to marine life. This study aims to investigate the regenerative properties of reclaimed sites using three main descriptors of benthic fauna – species richness, abundance and biomass. Sediment samples were collected from the coastal waters of four locations, East Coast of Singapore (~45 years), St. John’s Island (~35 years), Pulau Sudong (~40 years) and Pulau Semakau (~22 years), chosen due to their different recovery durations since the last reclamation. A total of 16 samples were taken from each location in July 2011 using a bottom dredge. Samples collected were analyzed using the Shannon-Weaver Index (H’) to determine their soft-bottom macrobenthic community structures; each sample was sieved, and macrobenthos retained were preserved and identified. The research yielded a mean H’ value of 1.00, 0.53, 0.37 and 0.61/grab for East Coast, St. John’s Island, Pulau Sudong and Pulau Semakau respectively. The Kruskal-Wallis test revealed significant differences in abundances of benthic fauna across locations, with the highest and lowest levels at East Coast and Pulau Semakau respectively. The Pearson two way chi-squared test revealed that there is no evidence for association between the sediment composition and location sampled. Further analyses via general linear modelling have confirmed that the age of recovery does have considerable impact on the macrobenthos.

Keywords: Soft-bottom macrobenthic communities; Reclamation; Environmental impact; Species richness and diversity; Singapore

220 Words INTRODUCTION

Land reclamation refers to the conversion or modification of wetlands and water bodies to become usable land. Rapid industrial and commercial expansion, accompanied by a growing population, has called for the need for more land in Singapore. In order to accommodate the nation’s need for more land to sustain healthy economic growth, much of Singapore’s southern coast and offshore islands have undergone reclamation work. Small-scaled land reclamation in Singapore started as early as 1820s. It was only in the 1960s that land reclamation was carried out on a large scale. Since then, Singapore has reclaimed about 70 km2 of land (Chia et al., 1988).

The most common methods of reclamation involve the construction of sand walls and filling of materials or dumping of earth spoils to stabilize the seabed (Chen et al., 2002). However, these activities may negatively influence marine diversity (Morton, 1996) by increasing sedimentation rates at areas near the reclaimed site, which in turn influences sunlight penetration or even smother benthic organisms to death (Chou et al., 2004). Many marine habitats have also been destroyed due to land reclamation.

In this study, macrobenthos were chosen as indicators of marine biodiversity and marine life health. The survival of benthic fauna primarily depends on physical and chemical properties of the subsoil; moreover, benthic communities are known to be sensitive to changes in the quality of water and other aspects of the marine environment. Due to their relatively sedentary lifestyles resulting in a long-term residency period in a marine habitat, benthic communities are commonly used as bio-indicators of specific environment and habitat conditions (Alongi, 1990; Alongi et al., 1992). There has been much research on benthic infauna in the coastal waters of Singapore, and they have revealed that sediment heterogeneity of the seabed and various environmental variables could potentially alter benthic community structure (Chou et al., 2004; Lu et al., 2005). Previous studies also suggest human activities, which likely indicate a change in the marine environmental conditions due to pollution or sedimentation (Hatcher et al., 1994; Chou et al., 2004), may have an impacton benthic communities.

Thus, this research aims to examine reclaimed sites to ascertain the negative influence reclamation may have on marine life. By assessing the marine ecology of selected sites using three descriptors of benthic fauna –diversity, abundance and biomass, our objective is to make a distinction between the marine biodiversity of sites reclaimed at different time periods. The result willcould help us determine if the impacts of land reclamation are reversible, and whether there is a chance for full recovery of reclaimed sites. These findings would be vital to urban planning in Singapore – ultimately striking a balance between economic land area development through reclamation and Singapore’s marine biodiversity.

MATERIALS & METHODOLOGY

Hypothesis

The Shannon-Weaver diversity index , biomass and abundance will increase as the number of years since the site has been reclaimed (i.e. the number of years that have past for the marine ecology to be regenerated to a healthy state) increases.

Sampling Locations

Location (Date of Reclamation) Years Since Reclamation Work (in 2011) East Coast (1966) 45 Pulau Sudong (1971) 40 St. John’s Island (1976) 35 Pulau Semakau (1991) 22 Figure 1: Table of selected grab survey locations and the number of years since they have been reclaimed (Chia et al., 1988).

The four locations above were chosen because of their accessibility and marked differences in the number of years since they were reclaimed. Four stations were surveyed for each location over two trips in July 2011.

Legend: Stations at Pulau Sudong

Stations at East Coast

Stations at Pulau Semakau

Stations at St. John’s Island

Figure 2: Map of Singapore indicating the stations surveyed; refer to appendix 1a for geographical coordinates http://english.freemap.jp/asia_e/singapore.html Last Accessed: 26 December 2011, 2:45pm

Five environmental parameters – turbidity, salinity, pH, dissolved oxygen and temperature were measured at each station using an Addestation data logger and their respective probes.

Macrofauna Sampling

Four repeated surficial sediment samples were taken at each station on a research vessel with an AMS Shallow Water Bottom Dredge (6’ x 6’ x 6’ in size and the equivalent of a Van Veen Grab), chosen for its performance on soft sediments, ease of deployment, minimal disturbance to surficial sediments and small benthic organisms upon descend, and penetration depth of about 15-16 cm (Boyd, 2002). Personnel wore life vests and work gloves while handling the grab.

Figure 3: (a) Opened jaws of grab on descend; (b) Closing of jaws of grab upon striking the benthic surface. http://www.ecasatoolbox.org.uk/the-toolbox/eia-country/book-of- protocols/BoPfig02.jpg Last Accessed: 2 November 2011, 8:13pm

The sediment samples were either accepted or rejected for standardization based on the criteria depicted below.

Figure 4: A diagrammatic representation of acceptable and unacceptable samples. http://www.ecasatoolbox.org.uk/the-toolbox/eia-country/book-of- protocols/BoPfig03.jpg Last Accessed: 2 November 2011, 8:16pm

Sieving and Storage

Contents of the accepted samples were transferred into gallon-sized Ziploc bags and placed in a fridge for preliminary storage. The samples were then sorted for benthic fauna using a 0.5mm biological sieve. While sieving, water was gently sprayed from above and the sieve was agitated by hand in an up and down motion. The animals retained from the grab contents were then preserved in labeled falcon tubes, which included information on the sampled location, station number and sample number via fixation in a 10% buffered seawater formalin solution for two to four days, and were thereafter placed in 70% ethanol solution for long term storage. The chemicals were handled carefully with appropriate protective gear such as lab coats, gloves and facemasks.

All animals were first sorted into 5 main categories (Annelida, Crustacea, Echinodermata, Mollusca and other invertebrate phyla), then classifiedto the lowest possible taxonomic levelusing a stereoscopic dissection microscope and biomass estimates were made for each sample, excluding fragments and empty mollusk shells. For each station, composite samples were pooled together and macrobenthic diversity was calculated using the Shannon-Weaver diversity index (Shannon and Weaver, 1963). The index is defined as ∑ , where refers to the proportion of the th phylum in the station.

Sediment Composition Analysis

A small fraction of the sediment samples collected at each station was removed and transferred into labeled containers. Each sample was spread in the container and allowed to dry at 70ºC for at least 24 hours. After the sample hasfully dried, the dry weight was measured on a pre-weighed dish precisely to 3 decimal places.

In order to separate the sand and silt fraction, clumps in the dried sample was broken up using a pestle and mortar, done in a fume hood, as this may generate dust which causes irritation to the skin and eyes. The sample was then dispersed in sodium hexametaphosphate solution (6.2 g/L) and steadily stirred for at least a half hour using a mechanical shaker. The sample was thereafter run through a 250 μm mesh sieve, thus retaining the sand fraction. The dry weight of the sand fraction was determined after drying at 70ºC overnight. The weight of the silt fraction was obtained by subtracting the dry weight of the sand fraction from the dry weight of the sample.

The dried sand fraction was then transferred to a stacked series of mesh sieves, with the uppermost being the coarsest 4000 μm sieve and the bottommost being the finest 250 μm sieve. The column of sieves was shaken for approximately 10 minutes by hand to ensure thorough separation of grains. After sieving was completed, the grains retained at each sieve were weighed. The grain composition is subsequently represented by the respective percentage mass of grains at each size range.

Statistical Analysis

Due to our small sample size of 16 stations, the results will likely not follow a normal distribution. As such, the Kruskal-Wallis test was adopted to test for significant differences of the dependent variables, abundance, biomass and diversity, across locations. The five independent environmental variables measured were put through the Kruskal-Wallis test as well. The Pearson’s chi-squared test was used to test for association between benthic surface sediment composition and the surveyed site.

The general linear model (GLM) was used to determine the relationship between the explanatory and response variables. Prior to modelling, the Spearman’s rank correlation coefficient test was run to reveal correlated variables to be excluded from the general linear model. Akaike’s Corrected Information Criterion adjusted for small sample sizes (AICc), is an information-theoretic criterion used for our model selection. The calculation is represented as: AICc = nlog(RSS/n) + 2K(n /(n–K–1)); where n is the sample size, RSS is the residual sum-of-squares and k is the number of model parameters. Smaller AICc values indicate a better fit for the model (Burnham and Anderson, 2002). All statistical tests were run using MINITAB version 15 (Minitab 15 Statistical Software, 2007).

RESULTS

Benthic Fauna

A total of 64 samples were taken from the four sites. The collected fauna included 70 specimens belonging to six phyla, identifif ed in the study. Additionally, faunal abundance, biomass and diversity (Shannon-Weaver Index) were measured at all 16 stations. Due to low density of benthic fauna and high variance in grab samples, data were pooled from the four repeated grab samples at each station in order to give a better representation of each site.

Abundance

TABLE 1. Number of specimens found at each station Location Station 1 Station 2 Station 3 Station 4 Total East Coast (45 Years) 8 9 7 13 37 Pulau Sudong (40 Years) 7 5 1 2 15 St. John’s Island (35 Years) 1 0 8 1 10 Pulau Semakau (22 Years) 3 3 0 2 8

Abundance Level of Sampled Locations 14 12 10 8 STATION 1 Specimens Specimens

6 of of STATION 2

. 4

No No 2 STATION 3 0 STATION 4 East Coast (45 Pulau Sudong St. John's Island Pulau Semakau Years) (40 Years) (35 Years) (22 Years) Sampling Location

Figure 5: Bar chart of abundance level at each of the stations sampled.

Total abundance of faunal invertebrates was the highest at the East Coast of Singapore, followed by Pulau Sudong, St John’s Island and Pulau Semakau; this is in the order of the oldest to the most recentreclamation activity. East Coast yielded the highest abundance with 37 individual specimens collected across the four staations. Each station had a minimum abundance level of seven specimens, and a maximum abundance level of 13 specimens at station four. St John’s Island and Pulau Semakau, two sites which were more recently reclaimed, had lower abundance level ranging from zero to three specimens per station, with the exception of St John’s Island’s station three that had yielded eight specimens. Abundance levels of stations at Pulau Sudong were also inconsistent, with the stations one and two having higher abundance levels (seven and five specimens respectively) and stations three and four having lower abundance levels (one and two specimens respectively). Faunal abundance differed rather significantly across locations (H = 7.712, df = 3, P =0.052).

Biiomass

TABLE 2. Biomass estimate in grams at each station Location Station 1 Station 2 Station 3 Station 4 Total East Coast (45 Years) 5.281 1.663 6.468 17.141 30.553 Pulau Sudong (40 Years) 2.373 0.257 0.510 0.024 3.164 St. John’s Island (35 Years) 0.716 0.000 2.957 0.811 4.484 Pulau Semakau (22 Years) 1.491 0.240 0.000 1.490 3.221

Biomass Estimate of Sampled Locations 40 35 (g) (g) 30.553 30 25 Weight Weight

20 15 Dry Dry 10 3.164 4.484 3.221

Total 5 0 East Coast (45 Years) Pulau Sudong (40 St. John's Island (35 Pulau Semakau (22 Years) Years) Years) Sampling Location

Figure 6: Bar chart of the estimated total biomass (in grams) of each sampled location.

Biomass Estimate of Sampled Locations (Broken Down to Stations) 20 (g) (g)

15 Weight Weight 10 STATION 1 Dry Dry

STATION 2 5 STATION 3 0 East Coast (45 Pulau Sudong St. John's Island Pulau Semakau STATION 4 Measured Years) (40 Years) (35 Years) (22 Years) Sampling Location

Figure 7: Bar chart of the estimated biomass (in grams) of each of the stations sampled.

There was no significant difference in biomass across locations (H = 7.274, df = 3, P = 0.064). However, stations at the East Coast of Singapore yielded samples of substantially higher faunal biomass, with a mean of 7.64 g/station and total of 30.55 g. This may be attributed to the shells of mollusks (27 specimens) and crustaceans (one specimen) found. The number of years since reclamation seem to positively translate to a higher total measured biomass for each location, with the exception of Pulau Sudong, whose stations hauled samples of the least biomass, many of which were zeroas with abundance. Howbeit, total faunal biiomass of Pulau Semakau (3.22 g) was similar to that of Pulau Sudong (3.16 g) and was relatively small as well. St. John’s Island, whose last reclamation activity was 35 years ago, yielded a total biomass of 4.84 g.

Diversity

TABLE 3. Shannon-Weaver Index (H’) at each station Location Station 1 Station 2 Station 3 Station 4 Mean East Coast (45 Years) 0.974 1.067 0.956 1.010 1.002 Pulau Sudong (40 Years) 1.154 0.950 0.000 0.000 0.526 St. John’s Island (35 Years) 0.000 0.000 1.494 0.000 0.374 Pulau Semakau (22 Years) 1.099 0.637 0.000 0.693 0.607

Average H' of Sampled Locations 1.2 1.002 1 0.8 0.607 0.6 0.526 Value

' 0.374 H 0.4 0.2 0 East Coast (45 Years) Pulau Sudong (40 St. John's Island (35 Pulau Semakau (22 Years) Years) Years) Sampling Location

Figure 8: Bar chart of the mean H’ value of each sampled location.

H' of Sampled Locations (Broken Down to Stations) 1.6 1.4 1.2 1 0.8 STATION 1 Value Value

'

H H 0.6 STATION 2 0.4 0.2 STATION 3 0 STATION 4 East Coast (45 Pulau Sudong St. John's Island Pulau Semakau Years) (40 Years) (35 Years) (22 Years) Sampling Location

Figure 9: Bar chart of the H’ values of each of the stations sampled.

Diversity was the richest at stations at the East Coast of Singapore with an average H’ value of 1.002 /station, followed by stations at Pulau Semakau (mean H’ of 0.607 /station), Pulau Sudong (mean H’ of 0.526 /station) and St. John’s Island (mean H’ of 0.373 /station). No significant difference in faunal diversity was observed among the four locations (H = 2.459, df = 3, P = 0.483).

TABLE 4. Distribution of taxa across locations Location Crustacea Mollusca Echinodermata Annelida Other Invertebrate Phyla East Coast (45 1 27 0 9 0 Years) Pulau Sudong (40 1 6 2 4 2 Years) St. John’s Island (35 1 3 1 5 0 Years) Pulau Semakau (22 1 4 0 3 0 Years)

Distribution of Taxa among Sampled Locations 40 35 30 25 Other Invertebrate 20 Phyla 15 Annelida (Class Abundance 10 Polychaeta) 5 0 Echinodermata East Coast Pulau St. John's Pulau (45 Years) Sudong (40 Island (35 Semakau (22 Mollusca Years) Years) Years) Crustacea Sampling Location

Figure 10: Bar chart illustrating the distribution of taxa within each sampled location with respect to total abundance.

Relative Abundance of Phyla across Sampling Locations Crustacea 2.86% 5.71%

30.00% Mollusca

Echinodermata

Annelida (Class 4.29% 57.14% Polychaeta) Other Invertebrate Phyla

Figure 11: Pie chart showing the relative abundances of each phylum in the study.

Phylum Mollusca made up the majority of the specimens collected from all four sites (57.14%) with 40 specimens. Polychaetes from the phylum Annelida were the next largest faunal group (30.00%) with 21 specimens. Only one crustacean was found at each of the sampling sites, while few echinoderms were found. Two roundworms from the phylum Nematoda were also retrieved from station four of Pulau Sudong.

Environmental Parameters

Appendix 1b summarizes the physical characteristics of the 16 stations sampled, including measurements of the depth, salinity, turbidity, pH, temperature and dissolved oxygen levels. Salinity levels ranged from 23.87 ppt to 25.65 ppt. There were significant differences in the salinity levels across the locations at the time of sampling (H = 10.670, df = 3, P =0.014). The turbidity levels did not differ significantly across locations at the time of sampling (H = 7.323, df = 3, P =0.062), although readings for turbidity within each sampling location were considerably inconsistent; station two at the East Coast of Singapore had a turbidity reading of 0.00 NTU.The sampled locations also yielded pH levels that were significantly different at the time of sampling (H = 12.035, df = 3, P =0.007). Waters at the East Coast of Singapore and Pulau Sudong leaned towards being more alkaline in nature while waters at St. John’s Island and Pulau Semakau were more acidic, with the lowest pH recorded in this study being 5.35.Water temperatures remained fairly constant across all locations with a mean of 30.67 ºC. However, statistical analyses revealed significant differences in the temperatures of the sampling locations (H = 11.078, df = 3, P =0.011), and this may be associated with the time of day at which the reading was taken.The amount of dissolved oxygen in each sampling location did not differ at the time of sampling (H = 4.264, df = 3, P =0.234).

Sediment Composition

Appendix 2a gives the distribution of grain sizes for each of the sampled locations. Referring to the pie charts at appendix 2b-e, the majority of the sampled locations – East Coast (56.49% silt), St. John’s Island (58.75% silt) and Pulau Semakau (60.30% silt) were made up primarily by the silt fraction, with the exception of Pulau Sudong (58.57% silt). The sediment at Pulau Sudong was found to comprise grains of generally greater sizes. The East Coast of Singapore also had an unusually high percentage of larger grains such as small stones or gravel, with 3.28% of the sediment being sand grains that were bigger than 4000 μmin size. Of all the locations sampled. Pulau Semakau had sediments that were extremely fine in nature and comprised a high silt fraction level and a large proportion of fine sand. The sediment compositions of Pulau Semakau and St. John’s Island followed a somewhat similar distribution, whereas the sediment compositions of the East Coast of Singapore and Pulau Sudong were distinctly dissimilar from the rest of the locations. No evidence exists for association between the location and sediment composition (χ² = 22.786, df =15, P = 0.089), thus the two variables deemed to be independent.

General Linear Model

Our correlation tests revealed that salinity, dissolved oxygen and pH were correlated with the number of years since reclamation (Appendix 3a) and were excluded from our models; thus, only models that predict diversity, abundance and biomass as a function of the number of years since the last reclamation (YSR), temperature (TP), turbidity (TU), and percentage sand fraction (SF) were compared. The best models as selected via AICc calculations (Appendix 3b-d) are shown below:

SWI = 0.012 YSR + 0.203; where SWI is the Shannon-Weaver Diversity Index

ABD = 0.261 YSR – 4.894; whereABD is the abundance of benthic fauna

BMS = 0.221 YSR – 5.249; whereBMS is the biomass of benthic fauna

DISCUSSION

The dependent variables exhibited an overall increasing trend from the newest reclaimed site to the oldest reclaimed site. It is apparent that stations at East Coast, the oldest reclaimed site, yielded the highest diversity, abundance and biomass of benthic organisms; consequently it is likely that this site has managed to establish a stable and thriving community after the disturbances caused by the reclamation. The significantly high total biomass measured could be attributed to the shells of molluscs (27 specimens) found. However, it must be noted that all H’ values recorded in this study were generally low possibly due to the small sample and grab size.

An exceptionally high H’ value was also observed at Pulau Semakau, although we expect it to have the lowest diversity as it was the most recently reclaimed site. Its marine life may not have been affected by the reclamation activity as much as the older reclaimed sites as efforts were made to protect the marine ecosystem during the time of construction of the rock bund via the use of silt screens, and careful management of the reclamation works. Additionally, the bund is made impermeable using a layer of membrane and marine clay, therefore preventing any landfill leachate from polluting the marine environment. The older reclaimed sites were subjected to greater impacts as a result of a lack of technology to prevent increasing instability of the environment, the leakage of landfill materials and sedimentation on the macrobenthos.

A high proportion of organisms found in this research were from phylum Mollusca. Certain groups may be more prevalent due to their tolerance of different sedimentation rates and nutrient levels (Chou et. al., 2004); individuals more abundant in heavily sedimented stations had high tolerance for fine sediments and low nutrient levels and vice-versa. According to Chua and Lee (1984), construction materials are poor in nutrients and contain minimal organic matter, but molluscs may be more tolerant to these post-reclamation conditions. The next largest group was phylum Annelida and according to Olsgard and Hasle (1993), sites that were greatly affected by smothering of organisms all have small, opportunistic polychaetes as the predominant fauna. Polychaetes are also reported to be able to thrive in stressed environments of low dissolved oxygen and variable salinity levels (Aggrey-Fynn et al., 2011) via adaptive behaviour such as burrowing.After reclamation activities, the seabed is heavily disturbed and smothering will kill many benthic organisms. In this situation, opportunistic species will dominate, and if the effect of disturbance persists, these species will dominate continuously (Mirza and Gray, 1981).

Model comparisons using AICc values have shown that the best models predict diversity, abundance and biomass as a function of the number of years since reclamation. Although salinity, pH and temperature, showed significant variability across locations, these parameters would have been easily influenced by other variables such as the time of sampling during the day, rather than the reclamation that took place years before. Thus, the differences in abundance, biomass and diversity of macrobenthos across sites are unlikely greatly altered by environmental parameters.

Hence, the result of reclamation activities like habitat destruction and marine pollution through dredging and the dumping of earth spoils (Essink, 1999; OSPAR Commission, 2008) would likely be the dominant factors affecting the macrobenthic community. The latter is especially prevalent during the construction phase and sand extraction; changes in sediment structure and nutrient input would have negative effects on macrobenthos and their habitat. The structure of the benthic community would also be affected due to deposition of fine-grained sediment on the coarser-grained sediment layer as a result of elevated silt concentration (Chou et. al., 2004).

Limitations, Recommendations and Further Analysis

Our experiment lacked a control because much of the southern coasts of Singapore has already been reclaimed; alternative locations in Malaysia that have not yet undergone any relcamation work could be considered for future studies. An assumption also made in this research was that the pre-existing marine biodiversity is identical across all sites. Ideally, a study on the regenerative properties of reclaimed sites should begin when the site was first reclaimed, and readings should be regularly taken back at the same site as the years pass.

A larger grab size and more sampling should be considered to further substantiate our results. In addition, we plan to further identify the collected benthos down to family or species level with the help of a marine taxonomist, and ascertain if there are any indicator groups that were heavily affected by reclamation. CONCLUSION

There are negative impacts on marine life in the vicinity of reclamation activities as can be seen from low H’ values in this study when compared with findings in another research conducted by Lu et al. (2005). The dominant factor affecting benthos abundance, biomass and diversity was revealed to be the number of years since the last reclamation, suggesting that the number of years that have passed have indeed allowed a certain degree of regeneration to take place. These findings could help local environmental organizations identify reclaimed sites with high conservation values such as East Coast that was revealed to have achieved some form of regenerative success in this investigation. If well conserved and maintained in the years to come, this site may be able to recover to its original state. Pulau Semakau’shigh H’ despite its recent reclamation also makes it a site that is worth conserving. Agencies can thus consider using similar methods or technology employed during its reclamation for future reclamation work, in order to prevent excessive harm to the marine ecosystem.

As Singapore is a country with limited land resources, it is almost impossible to rule out land reclamation if we want to maintain our competitiveness in the global economy. However, with careful management before, during and after the reclamation we can help speed up the regeneration process and our nation can then continue development yet reducing our impact on our marine biodiversity. REFERENCES

Aggrey-Fynn, J., Galyuon, I., Aheto, D.W. and Okyere, I. (2011) Assessment of the environmental conditions and benthic macroinvertebrate communities in two coastal lagoons in Ghana. Annals of Biological Research, 2011, 2(5): 413–424.

Alongi, D.M. (1990) The ecology of tropical soft-bottom benthic ecosystems. Oceanography & Marine Biology: An Annual Review 28: 381-469.

Alongi, D.M. and Chistoffersen, P. (1992) Benthic Infauna and Organism-sediment Relations in a Shallow, Tropical Coastal Area: Influence of outwelled mangrove detritus and physical disturbance. Marine Ecology Progress Series 81: 229–245.

Boyd, S.E. (2002) Guidelines for the conduct of benthic studies at aggregate dredging sites. Lowestoft, Department for Transport, Local Government and the Regions, CEFAS: 1–117.

Burnham, K.P. and Anderson, D.R. (2002). Model selection and multimodel inference: a practical information-theoretic approach. Springer.

Chua, S.E. and Lee S.K. (1984) Creation of parks on reclaimed land in Singapore. Proceedings of the 3rd Symposium On Our Environment, Singapore: 43–50.

Chen, C.S. and Tan, S.M. (2002) A Case History of Coastal Land Reclamation Project. GSM-IEM Forum on Engineering Geology & Geotechnics in Coastal Development, Institute of Engineering, Malaysia.

Chia, L.S., Khan, H. and Chou, L.M. (1988) The Coastal Environmental Profile of Singapore. Technical Reports 21, International Center for Living Aquatic Resources Management, Manila, 92 pp.

Chou, L.M., Yu, J.Y. and Loh, T.L. (2004) Impacts of sedimentation on soft-bottom benthic communities in the southern islands of Singapore. Hydrobiologia, 515(1−3): 91–106.

Essink, K. (1999) Ecological effects of dumping of dredged sediments: options for management. Journal of Coastal Conservation, Volume 5, No.1, 69-80pp.

Hatcher,A., Grant, J., and Schofield, B. (1994)Effects of suspended mussel culture (Mytilus spp.) on sedimentation, benthic respiration and sediment nutrient dynamics in a coastal bay. Marine Ecology Progress Series 91: 265–268.

Lu, L. (2005) The relationship between soft-bottom macrobenthic communities and environmental variables in Singaporean waters. Marine Pollution Bulletin 51: 1034– 1040.

Minitab 15 Statistical Software. (2007) [Computer software]. State College, PA: Minitab, Inc. (www.minitab.com)

Mirza, F.B. and Gray, J.S. (1981) The fauna of benthic sediment from the organically enriched Olsofjord, Norway. Journal of Experimental Marine Biology & Ecology 54: 181–207.

Morton, B. (1996) Protecting Hong Kong’s marine biodiversity: Present proposals, future challenges.Environmental Conservation23: 55–65.

Olsgard, F. and Hasle, J. R. (1993)Impact of waste from titanium mining on benthic fauna. Journal of Experimental Marine Biology & Ecology 172: 185–213.

OSPAR Commission (2008) Assessment of the environmental impact of land reclamation. Biodiversity Series.

Shannon, C.E. and Weaver, W. (1963) The Mathematical Theory of Communication. University of Illinois Press, Urbana, IL, 125 pp.

ACKNOWLEDGEMENTS

This study was supported by a grant awarded by IKEA Singapore, World Wide Fund for Nature (Singapore) and Nature Society (Singapore). We would like to thank NUS High School for the support and facilities they have provided us with. We would also like to thank our mentor, Mr. Chua Sek Chuan, and teacher-in-charge, Mr. Malcolm Soh for their guidance and expertise. Special thanks to the National Parks Board for agreeing to license and issue permits for this investigation.

APPENDICES

APPENDIX 1a. Geographical coordinates of sampled stations Location Station 1 Station 2 Station 3 Station 4 East Coast (45 Years) 01°17'.338N; 01°17'.417N; 01°17'.531N; 01°17'.679N; 103°53'.577E 103°53'.751E 103°54'.059E 103°54'.397E Pulau Sudong (40 01°10'.086N; 01°10'.305N; 01°10'.541N; 01°10'.100N; Years) 103°44'.651E 103°44'.751E 103°44'.759E 103°44'.684E St. John’s Island (35 01°13'.575N; 01°13'.654N; 01°13'.604N; 01°13'.604N; Years) 103°50'.863E 103°51'.034E 103°51'.245E 103°51'.245E Pulau Semakau (22 01°12'.325N; 01°11'.666N; 01°11'.578N; 01°11'.845N; Years) 103°46'.924E 103°46'575E 103°45'.665E 103°45'.526E

APPENDIX 1b. Measured environmental parameters at each station Location Station Temperature Salinity Dissolved Turbidity pH Depth No. (ºC) (ppt) Oxygen (NTU) (m) (mg/L) East 1 30.70 24.72 7.74 1.66 8.07 8.0 Coast (45 2 30.15 25.28 8.78 0.00 8.08 7.0 Years) 3 30.05 25.60 3.55 5.27 8.03 7.5 4 30.15 25.65 2.95 1.17 7.99 8.4 Pulau 1 29.85 24.92 8.06 8.74 15.26 13.4 Sudong 2 30.02 24.92 6.86 8.79 7.97 12.1 (40 3 30.12 24.95 6.00 3.13 7.98 13.8 Years) 4 30.12 24.85 7.58 5.71 7.98 16.2 St. John’s 1 30.46 24.33 7.91 4.00 5.90 6.0 Island (35 2 31.15 24.27 0.00 7.79 5.86 10.2 Years) 3 31.59 24.59 1.32 1.32 6.11 3.4 4 31.59 24.59 1.32 1.32 6.11 3.4 Pulau 1 30.46 23.96 7.25 5.76 5.68 5.6 Semakau 2 31.21 23.87 1.20 2.69 5.75 7.0 (22 3 31.35 23.91 1.28 3.47 5.35 6.1 Years) 4 31.79 24.05 1.54 3.13 6.30 9.4

APPENDIX 2a. A summary of the sediment composition each location, expressed in percentage Location % Silt % Sand % <250 % 250- % 500- % 2000- % >4000 μm 500 μm 2000 4000 μm μm μm East Coast (45 56.488 11.886 7.287 17.621 3.438 3.280 Years) Pulau Sudong 41.430 15.830 12.038 27.821 2.674 0.208 (40 Years) St. John’s Island 58.749 16.243 10.274 13.180 1.555 0.000 (35 Years) Pulau Semakau 60.304 16.042 8.824 12.148 2.144 0.538 (22 Years)

APPENDIX 2b. East Coast Grain Composition

Silt 7.29% Sand >4000 μm 17.62% Sand 2000‐4000 μm 56.49% 43.51% Sand 500‐2000 μm

3.44% Sand 250‐500 μm 3.28% Sand <250 μm

11.89%

APPENDIX 2c. Pulau Sudong Grain Composition

12.04% Silt Sand >4000 μm 27.82% Sand 2000‐4000 μm 41.43% 58.57% Sand 500‐2000 μm Sand 250‐500 μm 2.67% 15.83% Sand <250 μm 0.21%

APPENDIX 2d. St. John's Island Grain Composition 16.24% Silt 10.27% Sand >4000 μm Sand 2000‐4000 μm 58.75% 41.25% Sand 500‐2000 μm 13.18% Sand 250‐500 μm 0.00% 1.55% Sand <250 μm

APPENDIX 2e. Pulau Semakau Grain Composition 16.04% Silt 8.82% Sand >4000 μm Sand 2000‐4000 μm 60.30% 39.70% Sand 500‐2000 μm 12.15% Sand 250‐500 μm 0.54% 2.14% Sand <250 μm

APPENDIX 3a. Results of Spearman’s Rank Correlation Coefficient test showing the ρ and P values. Values in bold indicate significant correlation between variables. Years Since Temperature Salinity Dissolved Turbidity pH Reclamation (Rank) (Rank) Oxygen (Rank) (Rank) (Rank) (Rank) Temperature ρ -0.651 (Rank) P 0.006

Salinity ρ 0.923 -0.700 (Rank) P 0.000 0.003

Dissolved ρ 0.534 -0.620 0.480 Oxygen (Rank) P 0.033 0.010 0.060 Turbidity ρ -0.182 -0.469 -0.149 0.106 (Rank) P 0.500 0.067 0.582 0.696 pH (Rank) ρ 0.874 -0.603 0.845 0.637 -0.150

P 0.000 0.013 0.000 0.008 0.580

Percent Sand ρ 0.352 -0.450 0.395 0.522 0.057 0.389 Fraction (Rank) P 0.182 0.080 0.130 0.038 0.833 0.137

APPENDIX 3b. Model selection based on calculated AICcvalues, relating the Shannon- Weaver Index of diversity (SWI) with the number of years since reclamation (YSR), temperature (TP),turbidity (TU), and percentage sand fraction (SF); where k refers to the number of model parameters, RSS to the residual sum-of-squares, AICcto Akaike’s corrected information criterion, ΔAICcto the difference from the lowest AICcvalue, and wAICc to Akaike weight scaled relative to a total sum of 1. Model k RSS AICc ∆AICc wAICc YSR+TP+TU+SF 6 4.031 11.754 13.180 0.000 TP+TU+SF 5 4.042 6.439 7.865 0.003 YSR+TP+TU 5 4.037 6.431 7.857 0.003 YSR+TP+SF 5 4.089 6.519 7.945 0.003 YSR+TU+SF 5 4.075 6.496 7.921 0.003 TU+SF 4 4.185 2.318 3.744 0.023 TP+SF 4 4.136 2.235 3.661 0.024 YSR+TU 4 4.102 2.179 3.604 0.025 YSR+SF 4 4.098 2.172 3.598 0.025 YSR+TP 4 4.100 2.174 3.600 0.025 TP+TU 4 4.047 2.085 3.511 0.026 TU 3 4.265 -1.188 0.238 0.134 SF 3 4.216 -1.267 0.159 0.140 TP 3 4.147 -1.382 0.044 0.148 YSR 3 4.121 -1.426 0.000 0.151

APPENDIX 3c. Model selection based on calculated AICcvalues, relating the abundance of benthic fauna (ABD) with the number of years since reclamation (YSR), temperature (TP),turbidity (TU), and percentage sand fraction (SF); where k refers to the number of model parameters, RSS to the residual sum-of-squares, AICcto Akaike’s corrected information criterion, ΔAICcto the difference from the lowest AICcvalue, and wAICc to Akaike weight scaled relative to a total sum of 1. Model k RSS AICc ∆AICc wAICc YSR+TP+TU+SF 6 119.470 35.304 12.042 0.001 YSR+TP+SF 5 143.560 31.247 7.985 0.005 TP+TU+SF 5 137.710 30.958 7.696 0.006 YSR+TU+SF 5 127.430 30.418 7.157 0.008 YSR+TP+TU 5 120.670 30.040 6.779 0.009 TU+SF 4 197.360 29.0955.833 0.015 TP+SF 4 189.420 28.8095.548 0.017 YSR+TP 4 143.840 26.8973.635 0.045 YSR+SF 4 143.650 26.8873.626 0.045 TP+TU 4 139.290 26.6733.412 0.050 SF 3 218.600 26.1682.907 0.065 YSR+TU 4 127.438 26.0552.794 0.069 TU 3 204.310 25.6992.438 0.082 TP 3 189.640 25.1811.920 0.106 YSR 3 143.860 23.2610.000 0.278

APPENDIX 3d. Model selection based on calculated AICcvalues, relating the biomass of benthic fauna (BMS) with the number of years since reclamation (YSR), temperature (TP), turbidity (TU), and percentage sand fraction (SF); where k refers to the number of model parameters, RSS to the residual sum-of-squares, AICcto Akaike’s corrected information criterion, ΔAICcto the difference from the lowest AICcvalue, and wAICc to Akaike weight scaled relative to a total sum of 1. Model k RSS AICc ∆AICc wAICc YSR+TP+TU+SF 6 180.770 38.181 11.896 0.001 YSR+TP+SF 5 214.160 34.026 7.740 0.004 YSR+TU+SF 5 192.360 33.280 6.994 0.006 TP+TU+SF 5 192.360 33.280 6.994 0.006 YSR+TP+TU 5 192.580 33.288 7.002 0.006 TP+SF 4 252.64030.810 4.525 0.020 TU+SF 4 251.09430.768 4.482 0.021 YSR+TP 4 221.930 29.910 3.624 0.032 YSR+SF 4 214.350 29.668 3.383 0.036 TP+TU 4 205.07029.361 3.075 0.042 YSR+TU 4 197.870 29.113 2.827 0.047 SF 3 279.28027.871 1.585 0.088 TP 3 260.09027.376 1.091 0.113 TU 3 251.94027.155 0.869 0.126 YSR 3 222.31026.285 0.000 0.195