Abstract

The Setiu wetland of peninsular Malaysia is threatened by expansion of aquaculture. Water quality of the estuary-lagoon complex is becoming affected by nutrients introduced into the system at floating fish cages and by the clear-cutting of large areas of fringing mangrove forest for the creation of land-based fish and shrimp pens. We report here on the distribution of benthic foraminifera in the estuary-lagoon and related environmental variables. These data will form the baseline against which future environmental monitoring will be compared.

Four thanatofacies are recognized by cluster analysis of the dead foraminiferal abundance data; their distribution is closely related to variations in salinity and other parameters related to the hydrodynamics of the region. This relation is confirmed by DCA and DCCA analyses. A low salinity estuarine thanatofacies (D1) has a low diversity assemblage dominated by the agglutinated taxa Ammotium directum, Trochammina amnicola, Miliammina fusca and Ammobaculites exiguus. A medium salinity lagoon thanatofacies (D2) has low diversity and is strongly dominated by the agglutinated Ammobaculites exiguus. Ammobaculites exiguus and Ammonia aff. A. aoteana dominate a high diversity, high salinity thanatofacies (D3) in both the estuary and the lagoon. A normal marine salinity thanatofacies (D4), found at an inlet and in the immediately adjacent lagoon, has a high diversity assemblage dominated by the calcareous taxa Amphistegina lessonii and Ammonia aff. A. aoteana. Five biofacies recognized by cluster analysis of live foraminiferal data exhibit similar salinity-related distribution patterns as the thanatofacies.

INTRODUCTION

The Setiu wetland of Terengganu, Malaysia, extends from 5°35′ N to 5°45′ N and is characterized by a coast-parallel estuary-lagoon system separated from the South China Sea by a narrow barrier island (Fig. 1). The mainland margin of the estuary-lagoon is an unzoned mixed-mangrove forest. Mangrove forest also covers the many islands within the Setiu estuary-lagoon (SEL). The SEL is part of the Setiu-Chalok-Bari-Merang wetland system, one of 17 priority conservation sites in the Malaysian Wetland Directory.

The SEL region, its inhabitants, infrastructure and economy, is threatened by anthropogenic impact. In particular, aquaculture, initiated in the 1970s, and oil palm plantations on adjacent upland areas are burgeoning. Fertilizer-rich runoff reaches the SEL during the monsoon, and antibiotics, nutrients, and fish waste are contributed to the SEL year round. Floating fish-cages are increasing in number and extensive areas of mangrove forest are being cleared to construct fish pens. Although aquaculture is increasingly important to the local economy, its expansion and environmental impacts must be carefully monitored if it is to be sustainable.

Foraminifera are valuable indicators of past environmental change (e.g., Douglas, 1979; Hayward and others, 1999; Murray, 2006) and current environmental stress, both natural and anthropogenic (e.g., Culver and Buzas, 1995; Martin, 2000; Scott and others, 2001; Cearreta and others, 2002). In coastal environments, under natural conditions, foraminifera tend to reflect the relative inflow of salt water vs. fresh water, which affects salinity (Nichols, 1974) and other variables such as pH and Ca content (Debenay and others, 1998, 2002; Debenay and Guiral, 2006). One advantage of using foraminifera as proxies is that they have a particularly good fossil record and their taxonomic composition and general ecological distribution in coastal marine settings is well known. Foraminifera, therefore, can be used to reconstruct historical trends, both natural and anthropogenically driven, based on cored sediments which provide pre-pollution background assemblages (Cearreta and others, 2000; Hayward and others, 2004).

The objectives of this paper are to characterize the previously undocumented distribution of modern foraminifera in the SEL. By relating the distribution to environmental variables, our goal is to provide a baseline against which environmental change in the SEL, as indicated by changing foraminiferal assemblages, can be compared.

ENVIRONMENTAL SETTING

The coastal region of northern Terengganu state is at low elevation and exhibits low relief. The SEL stretches 21 km parallel to the coast, with water depths ranging 0.5–3.0 m (Fig. 1). The estuary receives freshwater from the south and southwest from the Chalok and Setiu rivers, respectively. Discharge from these rivers increases considerably during the northeast monsoon (November–March), which decreases salinity in the southern part of the estuary (Rosnan and others, 1995). The estuary is currently connected to the South China Sea through a single, small (~300 m-wide) inlet/river mouth (Fig. 1). Another small inlet that was 4 km to the south of the current inlet closed naturally in 2003. During flood tides, a salt wedge extends southward into the estuary. During the northeast monsoon, fluvial discharge restricts the extension of the salt wedge. Salinity in most of the estuary ranges from 0–30 during both the southwest monsoon and the transition to the wetter northeast monsoon, when higher salinity conditions are restricted to the inlet/river mouth area.

The Setiu lagoon is located north of the inlet/river mouth (Fig. 1). Its salinity is considerably higher and it is subject to greater tidal influence than the river-dominated Setiu estuary. The mean spring tidal range is 1.8 m on the oceanic coast (Phillips, 1985) and slightly less within the lagoon. A small stream feeds into the northern end of the Setiu lagoon, decreasing salinity in its vicinity. During the transitional season, salinity in the lagoon ranges from 14–30, but during the northeast monsoon from November to March it decreases slightly. Salinity in the South China Sea near the inlet/river mouth is 32–33 (Rosnan and Mustapa, 2010).

The Setiu estuary ranges 1.5–3.0 m in depth and its substrate is mostly poorly sorted medium to coarse sand and gravel, although mud often accumulates along the margins with mangroves. In contrast, the Setiu lagoon ranges 0.5–2.0 m in depth and its substrate is moderately sorted fine sand with some mud (Rosnan and others, 1995; Rosnan and Mustapa, 2010). This grain size difference reflects the hydraulic energy levels of the two SEL components: much of the mud and fine sand from rivers and mangroves either bypasses the estuary or is deposited at its mangrove margins, whereas mud settles in the calmer waters of the lagoon (Rosnan and others, 1995).

METHODS

Duplicate 20-ml sediment samples (S1–S56) were taken using a Ponar grab at each of the 56 SEL sampling stations (Fig. 1) in early June 2009 and processed within a week of collection. Depth, bottom-water salinity, temperature, pH, and dissolved oxygen were recorded at each station (Table 1), but the pH meter malfunctioned after collecting S31. The top cm of sediment in one set of samples was processed for foraminifera. The other set was analyzed for grain size at half-phi intervals using standard sieving procedures (Folk, 1974), and granulometric data were analyzed with Gradistat software (Blott and Pye, 2001).

Foraminiferal samples were preserved in 70% ethanol and stained for live specimens using rose Bengal (Walton, 1952). After washing over 710- and 63-μm sieves to remove coarse material, mud, alcohol, and excess stain, the foraminifera were concentrated by flotation using a sodium polytungstate solution (Munsterman and Kerstholt, 1996), then dried and divided into aliquots with a microsplitter. Approximately 200 foraminifera were randomly picked from each sample except those few that were barren or contained very few specimens. Initial identifications were based on the following publications that illustrate Neogene foraminifera from Southeast Asia and the Southeast Pacific: Brady (1884), Barker (1960), Whittaker and Hodgkinson (1979), Brönnimann and Keij (1986), van Marle (1991), Brönnimann and others (1992), Brönnimann and Whittaker (1993), Jones (1994), Loeblich and Tappan (1994), Hayward and others (1999), Horton and others (2003, 2005), and Woodroffe and Horton (2005). Species identifications were then confirmed by comparison with type and figured material at the Smithsonian Institution, Washington, D.C., and the Natural History Museum, London. Appendix 1 is the species reference list, and Appendix 2 contains the census data. The most abundant species are illustrated in Figures 2 and 3.

Geographic patterns of foraminiferal distribution were determined using Q-mode cluster analysis of three data sets: dead, live, and total (live + dead) foraminifera. All taxa were included in each analysis. Species proportions were transformed according to 2 arc sin square root of pi, where pi is the proportion of the ith species within the sample (Buzas, 1979). Cluster analyses were run on the transformed data in SYSTAT version 10.0 using Ward’s linkage method and Euclidean distances. Because the results for dead and total cluster analyses were so similar, the latter is not further considered below.

We used detrended correspondence analysis (DCA) and detrended canonical correspondence analysis (DCCA) to detect species-assemblage response to environmental variables (Sejrup and others, 2004). Foraminiferal percentages were transformed to square roots to maximize the “signal to noise” ratio, and DCA and DCCA involved detrending by segments, nonlinear rescaling, and down weighting of rare taxa (Sejrup and others, 2004). DCCA provides an estimate of the gradient length in relation to x (environmental variable) in standard deviation units (Birks, 1995; Korsman and Birks, 1996). DCA and DCCA analyses were performed using CANOCO 4.5 (ter Braak and Šmilauer, 2002). The results for dead and live assemblages were similar, but the fewer live specimens provided a less robust analysis. Both analyses showed similar trends but we focused on the DCCA because it included environmental variables.

RESULTS

Environmental Variables

Bottom-water salinity in the Setiu estuary fluctuated from <5 in the southeast (Fig. 1, sample stations 1–19) to 28 in the salt-water wedge near the inlet/river mouth (station 26). Normal marine salinity (32) occurred at sample stations near the inlet/river mouth, offshore, and in the southeastern end of the Setiu lagoon up to station 44 (Table 1, Fig. 4). Bottom-water salinity at the remaining sample sites in the lagoon decreased to the northwest, where it was lowest (18) at station 56.

Bottom-water temperatures in the SEL increased to the northwest (Fig. 4). River water had the lowest temperature at 27–28°C, with slightly higher values of 30–31 °C recorded from the salt-water wedge in the estuary, inlet/river mouth, and central lagoon. Several stations at the northwestern end of the lagoon, where water was <1 m deep, had bottom-water temperatures of 32–33°C (Table 1).

Dissolved oxygen values (Fig. 4) were highest near the freshwater inflow at the southeastern end of the estuary (~6 mg/L) and near the inlet/river mouth (~7 mg/L); slightly lower values were recorded in the central part of the estuary and most of the lagoon. The four lowest values (<3 mg/L) occurred in the lagoon, including one (S42) in a fringing mangrove forest and another (S40) adjacent to a floating fish-farm cage (Table 1, Fig. 4). Measurements for pH were recovered only in the estuary; values increased from 5.8 in the southeast to 8.4 close to the inlet/river mouth (Table 1, Fig. 4).

Sediments range from gravelly coarse sand to fine sand and sandy mud, and are better sorted in the lower lagoon than in the upper lagoon and estuary. Mean grain size is coarsest mostly in the upstream section of the estuary, finest surrounding the inlet, and quite variable in the lagoon. Grain size, sorting, and skewness (Fig. 4) generally agree with the findings of Rosnan and Mustapa (2010).

Cluster Analysis

Dead Foraminifera (D)

Fifty samples yielded a total of 136 foraminiferal taxa. The cluster analysis based on dead foraminiferal abundances provided a dendrogram (Fig. 5A) that differentiates four major groups of samples (at Euclidean distance 0.4), whose assemblages we refer to as thanatofacies D1–D4. The dendrogram shows that thanatofacies D2 and D3 are similar to each other and related to D4, and D1 is the most distinctive.

Thanatofacies D1 occurs in 18 samples, all from the the Setiu estuary except for S42, which was taken several meters into a mangrove forest on a small island in the lagoon (Figs. 1, 5A, 5B). Twenty agglutinated and seven rare calcareous taxa (Table 2) comprise this thanatofacies. Its four most abundant species are the agglutinated Ammotium directum (24%) (Fig. 2.4), Trochammina amnicola (23%) (Fig. 2.6), Miliammina fusca (20%) (Fig. 2.1), and Ammobaculites exiguus (18%) (Fig. 2.3).

Five of six samples characterized by D2 are from the northern part of the Setiu lagoon, and the sixth (S24) is from just south of the river mouth (Figs. 5A, B). This thanatofacies contains 14 agglutinated taxa and 10 calcareous (mostly rotaliid) forms (Table 2). Ammobaculites exiguus (74%) is dominant and Ammotium morenoi (8%) (Fig. 2.5) is the second most abundant.

Thanatofacies D3, which the dendrogram reveals as being most similar to D2 (Fig. 5A), occurs in both the estuary and lagoon near the inlet/river mouth (Fig. 5B), and is transitional from agglutinated-dominated D1 and D2 to calcareous-dominated D4. Of the four thanatofacies, D3 is the most diverse and the 16 samples in which it occurs contain 22 agglutinated and 65 calcareous taxa (Table 2). Fifty of the latter are rotaliids and 15 are miliolids. Thanatofacies D3 is dominated by Ammobaculites exiguus (24%), followed by Ammonia aff. A. aoteana (19%) (Fig. 2.13) and Ammotium morenoi (7%).

Thanatofacies D4 is located at the inlet/river mouth and in the southern part of the lagoon (Fig. 5B). The 10 samples in which this thanatofacies occurs contain seven agglutinated and 69 calcareous taxa, the latter being 22 miliolids and 47 rotaliids (Table 2). Dominant are the calcareous species Amphistegina lessonii (26%) (Fig. 3.12) and Ammonia aff. A. aoteana (17%).

Live Foraminifera (L)

Forty-four samples yielded 35 taxa of live foraminifera. The cluster analysis (Table 3) defined five groups or biofacies (at Euclidean distance 0.4), of which two (biofacies L1 and L2) are dominated by calcareous taxa and three (biofacies L3–L5) by agglutinated taxa (Fig. 6A). The spatial distribution of the biofacies (Fig. 6B) is similar to that of the corresponding thanatofacies (Fig. 5B). Biofacies L4 characterizes 19 samples in the southern part of the Setiu estuary (Fig. 6B) and is comparable to thanatofacies D1. Sixteen of the 22 taxa in this biofacies are agglutinated (Table 3) and the dominant species are Ammobaculites exiguus (33%), Trochammina amnicola (23%), and Ammotium directum (17%).

Biofacies L3 occurs in three samples, two from near the southern end of the estuary and one at the northern end of the lagoon (Fig. 6B). Five of the six taxa that comprise this biofacies are agglutinated; dominant are Miliammina fusca (73%), Ammobaculites exiguus (18%), and Ammotium directum (6%) (Table 3). None of the thanatofacies has as high a proportion of Miliammina fusca (Table 2).

Biofacies L1 characterizes six samples and is similar to thanatofacies D3 in that it occurs both north and south of the inlet/river mouth and is transitional between agglutinated-rich and calcareous-rich assemblages (Fig. 6B). This biofacies contains six agglutinated and eight calcareous taxa (Table 3), which are dominated by the calcareous Ammonia aff. A. aoteana (68%), followed by the agglutinated Trochammina sp. E (7%) (Fig. 2.7), Ammotium morenoi (6%), and Ammobaculites exiguus (5%).

Biofacies L2 is found in six samples near the inlet/river mouth (Fig. 6B) and is similar to thanatofacies D4 (Fig. 5B). Three agglutinated and four calcareous taxa comprise the biofacies, which is dominated by the calcareous species Rosalina sp. B (50%) (Fig. 2.12) and Ammonia aff. A. aoteana (37%) (Table 3).

Biofacies L5, northwest of the inlet/river mouth (Fig. 6B), has a similar distribution to thanatofacies D2 and the northernmost samples containing thanatofacies D3 (Fig. 5B). The 10 samples characterized by this biofacies contain 12 agglutinated and eight calcareous taxa (Table 3). Biofacies L5 clusters with L4 (Fig. 6A) due in part to the high relative abundance of Ammobaculites exiguus (62%), which dominates both biofacies. Also abundant in L5 are the agglutinated taxa Trochammina sp. E (7%) and Ammotium morenoi (5%), and the calcareous species Rosalina sp. B (5%).

Abundance Data

The calculated number of dead foraminifera/20 ml sample (ND) varies greatly in the SEL (Fig. 7A, Table 4). Most samples have <5000 specimens, but 11 have more, including three with ~20,000 specimens and one with >74,000. The three samples with the greatest number of dead specimens (S40, S43, S47) were collected immediately adjacent to floating fish cages. The lowest number of dead specimens/ sample occurred in the southeastern estuarine samples of thanatofacies D1, where salinity was very low, and also in samples from D4 at the inlet/river mouth and southern lagoon (Fig. 7A). The ND values varied most in D3, where the majority of floating fish cages are located (Fig. 7A).

The calculated number of live (stained) foraminifera/ 20 ml sample (NL) shows a somewhat similar pattern (Fig. 7B, Table 4) to that of the ND. Lowest values are recorded in the least saline part of the estuary and adjacent to the inlet/river mouth, which has the highest salinity in the SEL. Station 18 has the highest density of live specimens (>3000/20 ml), but the reasons for this are unclear. Samples taken at the fish cages have similar values for NL as most other lagoonal samples (Fig. 7B, Tables 4, 5).

A plot of the calculated total (live + dead) specimens/ 20 ml sample (NT, Fig. 7C) is essentially identical to that for dead foraminifera (Fig. 7A) because there were relatively few live specimens (Table 4). Those assemblages that had a higher percentage of live foraminifera (Fig. 7D) generally consisted of few specimens/20 ml sample (Table 4).

Species Diversity

Species Richness (S)

The number of dead species/sample (SD) shows a pattern almost identical to the total number of species/sample (ST, Figs. 8A, C; Table 4). The SD and ST in samples characterized by thanatofacies D1 are generally <10 but increase to the northwest. Thanatofacies D3 has high species richness (SD and ST) of about 20–35/sample as does D4, although values vary considerably. The SD and ST in D2 decrease to the northwest to values similar to those for D1 (Figs. 8A, C). The highest SD and ST occur in sample S33, taken several hundred meters offshore of the inlet/river mouth (Figs. 8A, C).

Species richness values for live species/sample (SL, Fig. 8B) are much lower than those for dead species/sample (hence the similarity of plots for SD and ST; Figs. 8A, C). Biofacies L2 (inlet/river mouth) with one exception has low SL and several samples assemblages from the inlet/river mouth area had no live foraminifera (Figs. 1, 6A). Sample S42, taken from a mangrove forest, yielded the highest number of live species (Fig. 8B).

Species richness for the three shell types (suborders)—agglutinated (textulariid), calcareous porcelaneous (miliolid), and calcareous hyaline (rotaliid)—found in the SEL are illustrated in Figures 9A–F. The range of number of dead textulariid species (SDt) is 5–10 throughout the SEL except in thanatofacies D4 at the inlet/river mouth and the southern part of the lagoon (Fig. 9A, Table 4), where few agglutinated taxa occur. In contrast, the number of dead miliolid species (SDm) is greatest in D4, lower in D3, and, with two exceptions, miliolids are absent from D1 and D2 (Fig. 9B, Table 4). Dead rotaliid species (SDr) are generally more numerous than miliolids and textulariids (Fig. 9C), being highest in D3 and D4, lower in D2, and, with two exceptions adjacent to D3, are absent from D1 (Fig. 9C, Table 4). The mangrove forest sample (S42) of D1 has by far the highest number of dead agglutinated species (SDt, Fig. 9A).

In contrast to the dead assemblages, live textulariid species (SLt) are generally more numerous than live rotaliid ones (SLr) throughout the SEL except at the inlet/river mouth (Figs. 9D, F). Rare live miliolid species (SLm) are restricted to the lagoonal part of the SEL (Fig. 9E). Both mangrove forest sample S42 and nearby lagoonal sample S49 have the highest number of live agglutinated species (SLt, Fig. 9D).

Fisher’s Alpha

Fisher’s alpha (Fisher and others, 1943; Murray, 1973; Hayek and Buzas, 2010) takes into account the number of picked specimens as well as the number of taxa in each sample. A plot of alpha for dead species/sample (alpha D, Fig. 8D) closely mimics the plot for number of dead species/sample (SD, Fig. 8A), which is understandable in that the numbers of picked specimens/sample (nT), with some exceptions, were similar (Table 4). A plot of alpha using total (live + dead) species/sample (alpha T, Fig. 8E) is also very similar to that for alpha D, not only because the number of live species/sample is small compared to dead taxa (Figs. 8A, B; Table 4) but also because if a taxon is found live in a sample it is usually also found dead in that sample and, therefore, does not affect the total number of species in that sample (Table 4). As for SD and ST, the highest values for alpha D and T are those for sample S33, several hundred meters offshore of the inlet/river mouth (Figs. 8D, E).

Foraminiferal Shell Type

The relative proportions of the three shell types (agglutinated, calcareous porcelaneous, calcareous hyaline) are expressed as percent of the dead (D) and live (L) specimens in the assemblage of each sample (Figs. 10A–F; Table 4). Agglutinated specimens (D%t) account for 100% of the dead assemblages in most samples in thanatofacies D1 (Fig. 10A) and increasingly dominate assemblages to the northwest in D2 (Fig. 10A). Calcareous-porcelaneous foraminifera (D%m) are absent from D1, and when present, comprise <5% of assemblages in D2 and D3 (Fig. 10B). Their greatest proportions are in D4 (ca. 5–25%) and the highest value by far is in S31 from the inlet throat, but calcareous-hyaline specimens (D%r) dominate in D3 and D4 (Fig. 10C).

The percentages of live agglutinated specimens/sample (L%t) are similar to those for D%t (Fig. 10A, D; Table 4); live agglutinated foraminifera dominate the lower-salinity northern and southern parts of the SEL (Fig. 10D). Live calcareous-porcelaneous specimens (L%m) are rare throughout the SEL (Fig. 10E; Table 4), but the pattern for live calcareous-hyaline specimens (L%r, Fig. 10F) mimics that for D%r (Fig. 10C), although fewer samples near the inlet/ river mouth contain live foraminifera (Table 4).

DCA and DCCA Analyses

Detrended correspondence analysis (DCA) of the dead foraminiferal data (Table 6) shows that axis 1 (one gradient) is the only significant axis, since it is the only one that captures more than 10% of the total variance. Subsequent axes (2–4) have low eigenvalues and only capture 7.9%, 4.4%, and 2.9% of the total variance, respectively. Detrended canonical correspondence analysis (DCCA) of the training set, with depth (m), temperature (°C), salinity, dissolved oxygen (mg/ L), pH, and grain size, produced a gradient length of 4.057 standard deviation (axis 1) and captured >17% of the total variance in the species data. The results obtained by DCA and DCCA approximate those obtained for summer bottom-water salinity by Sejrup and others (2004) and Leorri and Cearreta (2009), and indicate the unimodal nature of the foraminiferal abundance data with respect to the environmental variables (Table 6). From the weighted correlation matrix, salinity and temperature seem to be the major factors controlling the distribution along environmental axis 1. However, temperature varies by <4°C among the samples. Axis 2 seems to represent mainly the influence of grain size and water depth; additional axes are not significant.

DISCUSSION

Comparison of Live/Dead Data and Analyses

Cluster analysis (at Euclidean distance 0.4) of dead assemblages (Fig. 5) and live populations (Fig. 6) defines four thanatofacies and five biofacies, respectively. Thanatofacies D1 is generally equivalent in composition and distribution to biofacies L4 (Figs. 5, 6; Tables 2, 3), whereas D2 encompasses L3, L4, and L5. Similarly D3, which is transitional between agglutinated-dominant and calcareous-dominant assemblages, characterizes samples that belong to L1, L2, L4, and L5 (Figs. 5, 6; Table 4).

Thanatofacies D4 is not comparable in composition to any biofacies (Tables 2, 3). It has the same river mouth/ southern lagoon (high salinity) distribution as biofacies L2 (Figs. 5, 6), but the latter has just seven live taxa dominated by Rosalina sp. B (Table 3), whereas the former has 76 taxa dominated by Amphistegina lessonii and with a very low proportion of Rosalina sp. B (Table 2). These proportions suggests taphonomic loss of Rosalina sp. B. No live specimens of A. lessonii were recorded in this study and the majority of its empty tests were worn; although samples were taken at only one time during the year, this suggests that A. lessonii does not inhabit the Setiu lagoon, but is likely imported from the inner shelf. This interpretation applies to many of the other calcareous hyaline taxa in thanatofacies D4 that exhibit broken chambers (Figs. 2, 3).

In equatorial mangrove swamps of French Guiana, total (live + dead) foraminiferal assemblages change from dominantly agglutinated in the rainy season to dominantly calcareous in the dry season, leading Debenay and Guiral (2006) to conclude that most of the foraminiferal tests were obliterated after death. Seasonal data on foraminiferal distributions have not yet been collected in the SEL, but the fact that the present SEL assemblages, collected over three days during a relatively dry period, vary in composition from all agglutinated to almost all calcareous suggests that a complete seasonal switch of assemblage-types is unlikely. This conclusion is supported by preliminary results from several short cores taken in various parts of the SEL that preserve assemblages similar to those at the sediment-water interface.

Foraminiferal Distribution and Environmental Variables

Temperature, dissolved oxygen, pH, and grain size exhibit understandable geographic variations in the SEL. Salinity by far has the widest variation, and strongly controls the distribution of thanatofacies (Figs. 4A, 5B; Table 5) .

Stations where estuarine thanatofacies D1 occurs (Figs. 4A, 5B; Table 5) had a mean salinity of 2.7 (1.4 if the single mangrove forest sample is excluded), and samples contained relatively low numbers of specimens/ sample (mean ND, mean NL) and a moderate proportion of live specimens (mean 21%; Table 5). Species richness (mean SD) and diversity (mean alpha D) were very low. Mean values for SDt, SDm, SDr, D%t, D%m, and D%r (Table 5) show the strong domination of agglutinated foraminifera.

Thanatofacies D2 in the northern part of the lagoon (Figs. 4A, 5B) has a mean salinity of 17.2 (Table 5). Numbers of specimens/sample (mean ND, mean NL) are moderately low, although these stations have the greatest proportion of live foraminifera (mean 34%; Table 5). Species richness (mean SD) and diversity (mean alpha D) are slightly higher than for D1. Textulariids dominate, but rotaliids contribute significantly to assemblages (mean SDr, mean D%r; Table 5).

Thanatofacies D3 at the northern end of the estuary and in the central part of the lagoon has a mean salinity of 22 (Figs. 4A, 5B; Table 5). Dead specimens are very abundant (mean ND) but live foraminifera are less common (mean 18%; Table 5) than in D1 and D2. This thanatofacies has high species richness (mean SD) and diversity (mean alpha D). Miliolids are present but rotaliids dominate (mean SDr, mean D%r; Table 5).

Thanatofacies D4 occupies the inlet and southern lagoon (Figs. 4A, 5B) where the salinity averages 30.7 (Table 5). The number of total specimens (mean ND, mean NL) and percentage of live foraminifera (mean 5%; Table 5) are low, but species richness (mean SD) and diversity (mean alpha D) are high. Rotaliids dominate (mean SDr, mean D%r; Table 5) and miliolids, although less abundant (mean SDm, mean D%m), comprise more of the assemblages than in any other thanatofacies

In agreement with the cluster analysis, DCA and DCCA reflect the variations in salinity and other variables linked to seawater penetration into the estuary/lagoon. Estuarine/ lagoon coastal environments could be defined most simply based on salinity, being areas where freshwater enters saline water and, during at least some period of time, water becomes brackish creating a gradient from marine to fresh water (Chapman and Wang, 2001). However, this gradient is highly variable both in time and space in response to tidal cycles and seasonally variable freshwater influx. Furthermore, local conditions, storms, or droughts complicate the general transition defined by the salinity gradient (Debenay and Guillou, 2002). Estuarine benthic foraminifera have been used as proxies for salinity gradients since the 1970s (e.g., Nichols, 1974). More recently, authors have developed methods to quantify the relationship between salinity and foraminiferal assemblage composition (e.g., Debenay and others, 1993; Hayward and others, 2004). Debenay and others (1993) proposed a confinement index for coastal Africa (similar to the environmental stress index defined by Debenay, 1990) that has been applied to the Brazilian coast (Debenay and others, 1997) and in modified form to coastal areas of New Zealand (Hayward and others, 2004). The confinement index as defined by Debenay and Guillou (2002) exhibits good correlation with salinity trends in estuaries. It is linked to salinity infiltration and freshwater flux and, hence, to the relative distance of a sample from the mouth of the estuary (or inlet) and freshwater input (e.g., see Leorri and Cearreta, 2009). The geographic location of samples in the SEL exhibits a strong correlation with salinity, temperature, pH, and to some extent dissolved oxygen (also inferred from the cluster analyses discussed above). These variables are linked to the local hydrodynamics and control the foraminiferal distribution in the SEL, as reported for estuarine environments elsewhere (Debenay and others, 1998, 2002; Debenay and Guiral, 2006). On the other hand, variations in water depth and grain size do not seem to be strongly related to the hydrodynamics, even though the Setiu lagoon generally has finer sediment than the estuary (Rosnan and others, 1995). This discrepancy may be related to the relatively wide spacing of sampling stations used in this study.

In summary, foraminiferal distributions within the SEL are similar to those in many other estuaries and lagoons around the world (e.g., Brazil: Closs and Madeira, 1966; South Africa: Phleger, 1976; Mexico: Phleger and Lankford, 1978; northeastern Canada: Scott and others, 1980; Florida, USA: Buzas and Severin, 1982; India: Reddy and Jagadiswara Rao, 1983; coastal East China Sea: Wang and others, 1985; Senegal: Debenay and Pages, 1987; Spain: Cearreta, 1989; New Zealand: Hayward and others, 1999; worldwide: Debenay and others, 2000; North Carolina, USA: Abbene and others, 2006; see Murray, 2006, for a summary) in that they can be explained, in large part, by spatial or temporal variations in salinity. Temperature is of secondary importance in the SEL given the generally constant surface water temperature throughout the year (Rosnan and others, 1995). Agglutinated foraminifera dominate in low salinity parts of the SEL. As salinity rises, calcareous foraminifera increase in both species richness and in their total percent of assemblages. Mixed agglutinated-calcareous assemblages (thanatofacies D3) occur in the Setiu estuary where influenced by a salt wedge (cf. Debenay and others, 1998).

Foraminifera at Fish-Farm Cages

Foraminifera are excellent indicators of shallow marine environmental stress, both natural and anthropogenic (e.g., Alve, 1991, 1995; Culver and Buzas, 1995; Schafer, 2000; Cearreta and others, 2002; Scott and others, 2005; Alve and others, 2009; Carnahan and others, 2009; Debenay and Fernandez, 2009). It is clear from a growing body of literature that foraminiferal populations not only are affected by aquacultural operations but also can document pre-pollution conditions and can be utilized in environmental monitoring (Clark, 1971; Grant and others, 1995; Schafer and others, 1995; Scott and others, 1995; Angel and others, 2000; Luan and Debenay, 2005; Tarasova, 2006; Tarasova and Preobrazhenskaya, 2007; Debenay and others, 2009a, b).

Three samples (S40, S43, and S47) in the SEL were taken approximately 2–3 m from fish-farm cages. They clustered within thanatofacies D3 (Fig. 5A) and have some characteristics in common. Their mean ND is extremely high at 37,980/20 ml sample (Table 5; Fig. 7A), and their percentage of live specimens (1.3%) is very low (Table 5; Fig. 7D). Species richness (mean SD) and species diversity (mean alpha D) are very high at 37 and 13.5, respectively (Table 5; Figs. 8A, D). Rotaliids dominate, but textulariids also contribute significantly to assemblages; miliolids are quite rare (Table 5). It is tempting to attribute the great abundance of specimens to high test productivity related to increased nutrients, but live foraminifera are rare. Perhaps there is a seasonal pattern to foraminiferal reproduction, or preservation of tests is greater at fish-cage sites than at other in the SEL. Further foraminiferal, sedimentological, and geochemical work is in progress on cores collected at these sites.

CONCLUSIONS

The benthic foraminifera of the SEL exhibit distribution patterns that relate mainly to bottom water salinity. Low diversity, agglutinated assemblages characterized low salinities, and calcareous foraminifera increased in relative abundance and number of species as salinity increased. Each of the four thanatofacies recognized by cluster analysis is essentially identical in composition to their corresponding total (live + dead) assemblages. Live foraminifera, although relatively rare, grouped into five biofacies that exhibit salinity-related distribution patterns similar to the thanatofacies. Mangrove assemblages are agglutinated and moderately diverse. Assemblages immediately adjacent to floating fish cages have high species richness and high specimen densities, but have few live specimens. The distributional data presented in this study form the baseline for future environmental monitoring of the threatened Setiu wetland.

ACKNOWLEDGMENTS

We thank the University Malaysia, Terengganu, and East Carolina University for funding this research and Dr. Khawar Sultan, Joseph Anak Bidai, Mohd Nasir Mohammad, Asmadi bin Salleh, Ahmad Nazila bin Ali, Mohd Zaini bin Mustapa, Kasim bin Muda, Adnan bin Abdulla Ghani, Che Mohd Kamarul Anuar bin Che Abdullah, Katie McDowell, Guy Iverson, Ray Tichenor, Katrina Rabien, and Alisha Ellis for support in the field and laboratory. Clive Jones kindly provided access to the foraminiferal collections at The Natural History Museum, London. We thank Jean-Pierre Debenay, Justin Parker, and an anonymous reviewer for their constructive comments and Bruce Hayward for taxonomic advice. This is a contribution to IGCP Project 588 and to the Geo-Q Research Unit (Joaquín Gómez de Llarena Laboratory).

APPENDIX 1

Taxonomic reference list.

Agglutinella agglutinans (d’Orbigny) = Quinqueloculina agglutinansd’Orbigny, 1839a, p. 195, pl. 12, figs. 11–13.

Ammobaculites exiguusCushman and Brönnimann, 1948, p. 38, pl. 7, figs. 7, 8.

Ammonia tepida (Cushman) = Rotalia beccari (Linné) var. tepidaCushman, 1926, p. 79, pl. 1, figs. 8a, b, c.

Ammotium directum (Cushman and Brönnimann) = Ammobaculites directumCushman and Brönnimann, 1948, p. 38, pl. 7, figs. 3a, b, 4.

Ammotium morenoi (Acosta) = Ammobaculites morenoiAcosta, 1940, p. 272, pl. 49, figs. 3, 8, not fig. 1.

Ammotium pseudocassis (Cushman and Brönnimann) = Ammobaculites pseudocassisCushman and Brönnimann, 1948, p. 39, pl. 7, figs. 12a, b.

Ammotium subdirectumWarren, 1957, p. 33, pl. 4, figs. 6–8.

Amphisorus hemprichiiEhrenberg, 1839, p. 134, pl. 3, fig. 2.

Amphistegina lessoniid’Orbigny, 1826, p. 304, no. 3.

Arenoparrella mexicana (Kornfeld) = Trochammina inflata (Montagu) var. mexicanaKornfeld, 1931, p. 86, pl. 13, fig. 5a–c.

Asterorotalia pulchella (d’Orbigny) = Rotalia (Calcarina) pulchellad’Orbigny, 1839a, p. 80, pl. 5, figs. 16–18.

Bruneica clypeaBrönnimann, Keij and Zaninetti, 1983, p. 36, pls. 1, 2.

Buliminoides williamsoni (Brady) = Bulimina williamsonianaBrady, 1884, p. 408, pl. 51, figs. 16, 17.

Caronia exilis (Cushman and Brönnimann) = Gaudryina exilisCushman and Brönnimann, 1948, p. 40, pl. 7, figs. 15, 16.

Cavarotalia annectens (Parker and Jones) = Rotalia beccarii (Linné) var. annectensParker and Jones, 1865, p. 422, pl. 19, figs. 11a–c.

Cellanthus biperforatusWhittaker and Hodgkinson, 1979, p. 84, fig. 61; pl. 7, figs. 3a, b; pl. 10, fig. 21.

Cellanthus craticulusvon Fichtel and von Moll, 1798, p. 51, pl. 5, figs. h–k.

Chrysalidinella dimorpha (Brady) = Chrysalidina dimorphaBrady, 1884, p. 388, pl. 46, figs. 20, 21.

Discorbinella bertheloti (d’Orbigny) = Rosalina berthelotid’Orbigny, 1839b, p. 135, pl. 1, figs. 28–30.

Elphidium advenum s.l. (Cushman) = Polystomella advenumCushman, 1922, p. 56, pl. 9, figs. 11, 12.

Elphidium indicumCushman, 1936, p. 83, pl. 14, fig. 10.

Elphidium reticulosumCushman, 1933, p. 51, pl. 12, figs. 5a, b.

Fijinonion fijiensis (Cushman and Edwards) = Astrononion fijiensisCushman and Edwards, 1937, p. 35, pl. 3, figs. 15, 16.

Gavelinopsis praegeri (Heron-Allen and Earland) = Discorbina praegeriHeron-Allen and Earland, 1913, p. 122, pl. 10, figs 8–10.

“Glomospira” fijiensisBrönnimann, Whittaker and Zaninetti, 1992, p. 23, pl. 3, figs. 1–4; pl. 13, figs. 1–4.

Hanzawaia nipponicaAsano, 1944, p. 99, pl. 4, figs. 1, 2.

Haplophragmoides wilbertiAndersen, 1953, p. 21, pl. 4, fig. 7.

Heterolepa subhaidingeri (Parr) = Cibicides subhaidingeriParr, 1950, p. 364, pl. 15, fig. 7.

Miliammina fusca (Brady) = Quinqueloculina fusca Brady inBrady and Robertson, 1870, p. 47, pl. 11, figs. 2, 3.

Miliammina obliquaHeron-Allen and Earland, 1930, p. 42, pl. 1, figs. 7–12.

Miliammina petilaSaunders, 1958, p. 88, pl. 1, fig. 15.

Notorotalia inornataVella, 1957, p. 54, pl. 2, fig, 29, pl. 3, figs. 36–38.

Palustrella palustris (Warren) = Textularia palustrisWarren, 1957, p. 34, pl. 4, figs. 3–5.

Pararotalia nipponica (Asano) = Rotalia nipponicaAsano, 1936, p. 614, pl. 31, figs. 2a–c.

Pararotalia venusta (Brady) = Rotalia venustaBrady, 1884, p. 708, pl. 108, fig. 2.

Paratrochammina stoeniBrönnimann and Zaninetti, 1979, p. 51–54, pl. 2, figs. 1–7.

Peneroplis pertusus (Forskål) = Nautilus pertususForskål, 1775, p. 125.

Planorbulina acervalisBrady, 1884, p. 657, pl. 92, fig. 4.

Planorbulinella larvata (Parker and Jones) = Planorbulina larvataParker and Jones, 1865, p. 379, pl. 19, figs. 3a, b.

Pseudorotalia schroeteriana (Parker and Jones) = Rotalia schroeteriana Parker and Jones inCarpenter and others, 1862, p. 213, pl. 13, figs. 7–9.

Reussella pulchraCushman, 1945, p. 34, pl. 34, figs. 11–12.

Reussella spinulosa (Reuss) = Verneuilina spinulosaReuss, 1850, p. 374, pl. 47, fig. 12.

Rosalina globularisd’Orbigny, 1826, p. 271, pl. 13, figs. 1–4.

Rosalina micens (Cushman) = Discorbis micensCushman, 1933, p. 89, pl. 9, fig. 5.

Sagrina zanzibarica (Cushman) = Bolivina zanzibaricaCushman, 1936, p. 58, pl. 8, fig. 12.

Sagrinella lobata (Brady) = Bolivina lobataBrady, 1881, p. 58; Brady 1884, pl. 53, figs. 22, 23.

Schackoinella globosa (Millett) = Discorbina imperatoria (d’Orbigny) var. globosaMillett, 1903, p. 701, pl. 7, figs. 6a, b.

Sigmoilina sigmoidea (Brady) = Planispirina sigmoideaBrady, 1884, p. 197, pl. 2, figs. 1–3.

Siphotrochammina lobataSaunders, 1957, p. 9, 10, pl. 3, figs. 1, 2.

Spiroloculina manifestaCushman and Todd, 1944, p. 62, pl. 8, figs. 26–28.

Triloculina tricarinatad’Orbigny, 1826, p. 299, no. 7, modele 94.

Triloculina trigonula (de Lamarck) = Miliolites trigonulade Lamarck, 1804, p. 351, no. 3.

Trochammina amnicolaBrönnimann and Keij, 1986, p. 22, pl. 2, figs. 1, 2; pl. 4, figs. 1–13.

Trochammina millettii Brönnimann and Whittaker = Trochammina? millettiiBrönnimann and Whittaker, 1993, p. 120, figs. 1.10, 1.22–1.24.

Trunculocavus durrandiBrönnimann and Whittaker, 1993, p. 122, figs. 2.1, 2.3–2.8.

APPENDIX 2. Foraminiferal census data. L, live; D, dead. Samples ordered from southeast to northwest