The Spermonde Shelf is located between 4 and 6°S in the southern Makassar Strait, west of Makassar in SW Sulawesi (Figure 1A). It covers an area of about 16,000 km² with a westward increasing water depth down to 60 m below mean sea level (mbsl) where the shelf reaches a physical barrier (Figure 1B). These barrier reefs form an area of more shallow water, 15–20 mbsl with some islands, shoals and islets on top, separating the Spermonde platform from the deeper Makassar Strait (Renema and Troelstra, 2001). A couple of east-west-trending channels incise the shelf and additionally cut through the barrier (Wilson and Vecsei, 2005).

The main characterizing feature of the Spermonde Archipelago is the occurrence of some 120 reef islands and sand cays. This study presents the results of a detailed sedimentological study on four of them. These four islands are Pulau Tambakulu (Figure 1C), Pulau Panamgangang (Figure 1D), Pulau Suranti (Figure 1E), and Pulau Panambungan (Figure 1F). P. Panamgangang, P. Suranti, and P. Tambakulu are situated along the outer rim of the Spermonde Shelf, while P. Panambungan is located in the inner shelf setting, about 19 km off Makassar harbor (Figure 1B).

Within the Makassar Strait, the Indonesian Through-Flow (Figure 1A) occurs and acts as an important heat exchange between the Pacific and the Indian ocean (Ding et al., 2013). This current shows a high temporal variability, due to the monsoonal changes in mean wind directions (Figure 2, Mayer and Damm, 2012). The NW monsoon during boreal winter, especially in December and January, accelerates the surface flow to the south. In turn, a direct SE monsoon wind direction slows down the southward flow. However, the rimmed shelf structure of the Spermonde Archipelago protects the inner shelf from stronger hydrodynamic energy settings (Renema and Troelstra, 2001).

Important for the geomorphology of reef islands in the Spermonde are monsoonal wind patterns (Wijsman-Best et al., 1982). The climate in the Spermonde Archipelago is tropical and varies little during the seasons. Only the seasonal change in rainfall has an impact on the regional climate. From June to October the dry monsoon prevails, while from November to March the rainy monsoon predominates. The annual precipitation varies between 2,100 mm in the rainy and around 200 mm in the dry season (Moore, 1995). The wind direction is east-southeast during dry monsoon, and north-northwest, during rainy monsoon (Manjunatha et al., 2015). During the NW monsoon the shelf experiences stronger waves (Wilson and Vecsei, 2005). During the dry season, the south-eastern winds reaching the Spermonde are weaker due to the fact that the shelf area is sheltered by the Sulawesi mainland and oceanic swell is smaller (Hoeksema, 1990).

Fluvial influence onto the Spermonde Shelf is related to the discharge of the two major rivers Maros in the north, and Jene Berang in the southern part of the shelf (Renema and Troelstra, 2001). Discharged river water contains volcanic erosion products from the hinterland and sewage from Makassar (Renema et al., 2001). The fluvial influence in terms of sedimentation on the shelf is largely restricted to a zone not extending 4 km from the coast (Storm, 1989).

The Spermonde Shelf is affected by two different tidal regimes, the semidiurnal and the diurnal tide (Ray et al., 2005). Along the Spermonde Shelf the semidiurnal tide has an amplitude of about 10–30 cm and propagates northwest through Banda and Flores seas, and the Strait of Makassar (see Figure 1, Ray et al., 2005). In contrast, the diurnal tide passes southwards through the Makassar Strait, with an amplitude of about 25–30 cm at the Spermonde shelf and a maximum springtide of 1.5 m measured at the harbor of Makassar (Ray et al., 2005; Meteo365.com Ltd., 2015).

Besides natural forcing, similar to other seascapes across Indonesia, anthropogenic activities steadily have increased in the Spermonde Archipelago. Today, most islands located on the mid and inner Spermonde shelf are populated (Burke et al., 2012). Inhabited islands potentially influence and alter reef ecosystems and frameworks of nearby uninhabited islands.

The 69 surface sediment samples show a grain size distribution from gravel to fine sand. Within this grain size range, sandy gravel (>30% gravel) and gravelly sand (5–30% gravel) are the most abundant textural groups. Less frequent grain size groups are gravel (>80% gravel) and slightly gravelly sand (<5% gravel). In general, the analyzed surface sediment samples are well to poorly sorted (Table 1). In the following the sediment characteristics of the surface samples for each location (see Figure 1) are described in detail.

Subsurface sediment from P. Panambungan shows similar characteristics as the surface sediment samples, with respect to the grain size distribution. Coarse sand with moderate to moderate well-sorting is the major fraction (45% of the samples). In the following the sediment characteristics of the subsurface samples for each section are described in detail.

The main component of the surface sediments, across all investigated islands are fragments of coralline algae, followed by coral remains which exhibit however a high variation in their abundance (Figure 5). In the following the mean sediment composition of the surface samples for each reef island are described in detail.

Mean subsurface sediment composition on P. Panambungan is coral-dominated, yet with differences depending on the sampling locality (Figure 6). Samples from the central sample location CS are characterized by mean values of 80.2% of coral fragments. This is followed by gastropod shells (9.6% mean) and fragments from coralline algae (5.6% mean). Sampling location S1 is composed by an average of 70.9% of coral remains, 12.8% coralline algae, and 8.6% gastropods. Similar to the subsurface sediment composition of S1, the sediment samples of S2 exhibit mean contents of 47.7% coral, 36.2% of coralline algae, and 8.4% of gastropod shells. S3 is composed, on average, of 43.6% coral, 38.8% coralline algae, and 8.7% gastropod shells. Other components such as echinoderms, bryozoans, serpulid tubes, sponge needles, and ostracods contribute <4% to sediment.

Grain size analysis of surface sediment samples reveals homogeneity across the investigated reef islands in the Spermonde Archipelago. General trends in grain size distributions from the individual islands are difficult to identify, however a distinction between the proximal (mid-shelf) island P. Panambungan and the distal (outer shelf) islands, P. Panamgangang, P. Suranti, P. Tambakulu is possible. While the distal islands show sandy gravel as predominant textural group, the proximal island is covered by slightly finer gravelly sand (Table 1). The majority of grains on the distal islands is above 700 μm in size. In contrast, P. Panambungan shows a mean D90 value of 434 μm. These differences in mean grain size of surface sediments may be explained by a higher wave energy during the NW monsoon period from December to January (Mayer and Damm, 2012), and a stronger hydrodynamic setting influenced by the Indonesian Through-Flow at the Spermonde shelf rim (Renema and Troelstra, 2001).

A zonation of the Spermonde reef island in inner and outer shelf sites is also supported by the component analysis of surface sediments. The analyzed reef islands in the Spermonde clearly group into one inner (P. Panambungan) and three outer (P. Panamgangang, P. Suranti, P. Tambakulu) islands with respect to their surface sedimentary composition (Figures 8, 9).

The hierarchical clustering results indicate two major groups. Cluster 1 is solely composed of samples from P. Panambungan and Cluster 2 contains samples from P. Panamgangang, P. Suranti, P. Tambakulu, except six from P. Panambungan (Figures 7, 8). The setting of two distinct clustering centers in the k-means cluster package of R results in a similar expression, illustrating that the two groups describe either P. Panambungan or the outer Spermonde reef islands P. Panamgangang, P. Suranti, P. Tambakulu (Figure 8). Based on the statistical analysis the reef islands differ in their surface sediment composition, mainly in the contribution of coral and coralline algae (Figure 9). The outer rim reef islands are coral-dominated based on the surface sediment composition. Additionally, the outer islands exhibit a relatively higher abundance of Halimeda, foraminifers and gastropods. In contrast, the sediment composition of the inner island (P. Panambungan) is coralline algae-dominated (Figure 9).

The D90 grain size distribution of the subsurface sediments from P. Panambungan alternates roughly between 400 and 700 μm (Table 2). Multiple horizontal layers of variable grain size indicate that sediments became deposited on top of the island as a result of overwash deposition (Woodroffe et al., 1999; Barry et al., 2008). This mode of sedimentation is known from other reef island settings such as the Maldives (Kench et al., 2006).

An outstanding feature on P. Panambungan is the observed shift in the sections from a coral-dominated (base) to a coralline algae-dominated (top) facies. Therefore, and from a sedimentological point of view, the reef community at P. Panambungan likely is in a transitional state.

Previous studies report an on-offshore gradient in faunal assemblages across the Spermonde Archipelago (Cleary et al., 2005; Becking et al., 2006; Cleary and Renema, 2007; Hoeksema, 2012; Sawall et al., 2012). This may explain why P. Panambungan differs in its surface sediment composition in comparison to the outer reefs islands. Variations in reef ecology are a product of distance from the coast due to river discharge impacts, shelf depth, and seasonality (Polónia et al., 2015). Thus, several variables influence the reef community composition across the Spermonde Archipelago, including water quality and habitat-limiting factors such as water depth, dissolved organic matter, and nutrient availability (Polónia et al., 2015). The latter factors here are important as they change with distance to riverine input and result in a zonation of the Spermonde Archipelago (Figure 10). Riverine influence on the Spermonde shelf is most prominent nearshore and during the rainy monsoon (Polónia et al., 2015). Thus, this affects the natural reef community differently across the shelf (Plass-Johnson et al., 2016).

However, the coral-dominated subsurface composition of P. Panambungan opposes the hypothesis that differences in surface sediment composition purely result from the islands distance to the mainland. The change in sediment composition from subsurface to surface (i.e., from coral to algal) clearly indicates a shift of the reef community. Shifts in coral reef assemblage have been reported from other coral reefs, e.g., the Caribbean or in the Commonwealth of the Northern Mariana Islands (Alvarez-Filip et al., 2013; Inoue et al., 2013; Enochs et al., 2015). Such shifts in reef community may result in lower calcification rates of corals and a loss of coral framework rugosity (Alvarez-Filip et al., 2013). Following these evidences from different study sites all over the globe, a shift in reef community composition around P. Panambungan is conceivable.

Shifts in reef community from hard reef-building corals to soft non-reef building corals or macroalgal communities often occur due to eutrophication and higher sediment load, which increase seawater turbidity (Gardner, 2003; Alvarez-Filip et al., 2013; Inoue et al., 2013; Enochs et al., 2015; Plass-Johnson et al., 2016). Here anthropogenic influences likely derive from highly populated neighboring islands that act as direct stressors on the reefs, from overfishing, untreated sewage disposal resulting in algae overgrowth, and nutrient flushing (Lessios, 1988; Hughes and Connell, 1999; Bruno et al., 2003; Voss and Richardson, 2006; Perry et al., 2008, 2011; Yoshioka et al., 2016; Ziegler et al., 2016). Perry (1996) directly correlates changes in the sediment composition to a shift in reef community composition. The observed change on P. Panambungan from subsurface coral-dominated to surface coralline algae-dominated sediments thus likely is the result of a shift in reef community structure.

Nevertheless, further sedimentological studies are necessary in order to validate in how far the apparent change of main contributors to island sediments is related to a change in the surrounding reef community structure. Future studies could ideally focus on analyzing depth-profiles of the sedimentary composition and grain size distribution on outer shelf reef islands within the Spermonde. This will help to delimit if the islands further away from the mainland have experienced comparable shifts in sediment composition through time as it is observed on P. Panambungan. Furthermore, determining the exact timeframe of such reef community phase-shifts across islands is highly desirable. This may provide evidence if phase shifts have a local cause (i.e., anthropogenic activities) or have been triggered by major environmental events in the Spermonde Archipelago.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.