Study Area

The morphology of the present-day coastline of the West Coast of southern Africa is defined by long stretches of relatively straight, steeply sloping beach separated by rocky headlands (Corbett, 1989). The weather patterns influencing the southern section of the West Coast are linked to the temperate westerlies. Swell waves generated in the southern Atlantic Ocean approach the coast obliquely from a southwest direction. In winter, because of the northerly migration of the westerlies wind belt, the coastline is subject to periods of intense northwest storm winds. The coastline is therefore predominantly exposed to long-amplitude southwest ocean swells, and the net transport direction of seabed material is northward owing to the prevailing southwest swell direction. However, during winter storms, there are periods of high-intensity northwest winds that result in significant transport of beach material south (Swart, 1983; Swart and Flemming, 1980).

Sixteen Mile Beach is one of a series of log-spiral sandy beaches in South Africa (Bremner, 1983) that extends for 26 km between the rocky headlands at Yzerfontein and Oude Post. Small rocky outcrops are also present at Gabbro Point and at Black Rock (exposed only during spring tide). These rocky outcrops are of the Late Precambrian to Cambrian Cape Granite Suite and range in composition from granite to gabbro. The region is covered extensively by Quaternary aeolianites and calcarenites (Dale and McMillan, 1999; Pether, Roberts, and Ward, 2000; Rogers, 1980; Tankard, 1976). Along with igneous headlands, the Pleistocene and Holocene dunes make up the western peninsula that separates Langebaan Lagoon from the Atlantic Ocean (Figure 2). The variably cemented, intertidal shelly beach sands of the Eemian Velddrif Formation are overlain by aeolianites and calcrete horizons of the Langebaan Formation. The age of the 60-m thick aeolianite succession of the Langebaan Formation at Kraalbaai is 117 to 79 ka and includes fossil human footprints near the base of the formation (Roberts and Berger, 1997).

The Sixteen Mile Beach Complex is composed of a sandy beach, coast-parallel dune ridges, and a dune cordon (Figure 2). Active foredunes and numerous coast-parallel vegetated relict dune ridges are found behind the northern part of the beach. These coast-parallel dunes have prograded seaward in response to the overall regression since the mid-Holocene 3-m high stand (Compton and Franceschini, 2005). At the southern end of Sixteen Mile Beach, a cordon of vegetated and nonvegetated Holocene dunes (Witzand Formation) extends 24 km inland to the north (Figure 2). The dunes contain several roads and fences and have been partially planted for stabilization by the Department of Forestry. Otherwise the area is generally undeveloped and is today part of the West Coast National Park.

The area is located in a Mediterranean, semiarid climate with a mean annual rainfall of 240 mm, received mostly during the winter months between May and August (Figure 1). The mean winter temperature is 14°C and rarely drops below 5°C, whereas the mean summer temperature is 22°C and rarely exceeds 30°C. In the southern Western Cape, the winds are highly variable. In Cape Town the winds are predominantly from the south–southeast, whereas at Langebaanweg, 12 km east of the study area, the predominant wind direction is from the south to southwest (Figure 1). Average wind speeds of 20–40 km/h are attained for 25% of the year in the Cape Columbine area, 35 km north of the study area, but gale-force winds exceeding 52 km/h are experienced less than 4% of the year (SA Weather Bureau, personal communication, 1998). Waves generated in the Southern Ocean, and local wind-generated surface waves, approach the Atlantic coastline obliquely, approximately from the southwest (Swart, 1983; Swart and Flemming, 1980). The coast is microtidal (mean spring tide range of 1.4 m and a maximum astronomical tidal range of 2 m) and is exposed to moderate to high wave energy, with 90% of waves having heights of 1 to 3 m with a wave period of 13–15 seconds; annual net longshore transport is to the north (Swart and Flemming, 1980). Beach sand compositions vary from 85 wt% terrigenous material in Table Bay to 60 wt% terrigenous material along Sixteen Mile Beach (Franceschini, Compton, and Wigley, 2003). The decrease in terrigenous material northward along the southern part of the West Coast reflects the decrease in terrigenous sand delivered by rivers. False Bay and Table Bay have perennial rivers that drain catchment areas receiving an average annual rainfall of 600 mm, whereas ephemeral rivers with small catchment areas drain the semiarid West Coast region south of Saldanha Bay (273 mm of annual rainfall at Langebaan).

Sixteen Mile Beach

The steepness of the beach face and backshore increases from south to north (Figure 3). The beach sand is primarily composed of quartz grains and shell fragments. The carbonate shell content of the bulk sand collected from the three beach transects ranges from 41 to 78 wt%. The carbonate shell content becomes increasingly variable from south to north, and there is an overall increase in the percent carbonate sand from south to north (Figure 3). The carbonate shell content is greatest adjacent to the rocky headlands at Tsaar-bank, where abundant shells are fragmented against the rocks. Carbonate shell fragments are primarily derived from mollusks and echinoderms that inhabit the sandy and rocky intertidal and subtidal areas. The predominant shell on Sixteen Mile Beach is the white mussel (Donax serra), which also occurs as gull-dropped shells on the wind-deflated foredune surface. The black mussel (Choromytilus meridionalis) along with several species of gastropod and barnacle are locally abundant near rocky shorelines. A total of three nonabraded genera of benthic foraminifera were recovered from all of the rocky pool areas. Ammonia and Elphidium were consistently the most abundant, comprising 90% (by number) of the fauna collected, followed by Glabratella. Similar trends in microfossil abundance were observed along the beach, but all specimens were transported.

The grain size of bulk sand collected at MHW increases northward (Figure 4). The sand is uniformly composed of 95% fine sand and 5% medium sand to 6 km north of Gabbro Point. The medium sand size fraction increases rapidly from 5% to 58% between 6 and 9 km. The medium sand fraction then increases gradually from 58% to 89%, and the fine sand fraction decreases from 41% to 10% between 9 and 17 km. The northern part of the beach from 17 km north of Gabbro Point to Tsaarbank is composed of 83–90% medium sand. The coarse sand fraction increases and the fine fraction decreases between 19 and 22 km. The sand from 23 to 26 km is 89% medium sand, 6% coarse sand, and 5% fine sand (Figure 4).

From Gabbro Point to 3 km north, the age of the carbonate shell fraction in the sand is fairly uniform ranging from 18.7 ka to 18.2 ka (Table 1). The carbonate shell fraction becomes younger toward the northern end of the beach with calibrated radiocarbon ages of 14.4 ka at 10 km, 12.6 ka at 16 km, and 10.8 ka at the northern end of Sixteen Mile Beach (Table 1; Figure 2).

Yzerfontein–Geelbek Dune Cordon

The Yzerfontein–Geelbek dune cordon has an area of ∼43 km² and consists of vegetated parabolic ridges and nonvegetated, active barchanoid dune fields. The Yzerfontein–Geelbek dune cordon extends 24 km inland in an N/NNE direction and, prior to invasion by alien vegetation, mobile barchanoid dune fields occur approximately 0–8, 9–11, 13–14, and 17–19 km from the coast (Figure 2). Comparison of a series of aerial photographs taken between 1938 and 1993 indicates that movement of barchanoid dune fields is highly erratic but that, on average, active dune fields have migrated at a rate of 3.6 m/y. Surveyed archaeological sites within the Geelbek dune field show that individual 15-m high barchanoid dunes can migrate inland at a rate of 5 m/y (N. Conard, unpublished report, South African National Parks, 2000). The aerial extent of the nonvegetated dunes has decreased by approximately 25% since 1938 from the invasion of alien vegetation.

Bulk sand samples collected from the mobile dune sand 6, 18, and 24 km inland are all well sorted and have similar grain size distributions (Figure 5). Carbonate-free samples also have a grain size distribution similar to bulk samples with a mean grain size of 180–150 µm. The exception is the sample from 12 km, which is composed of significantly finer grained carbonate sand than the other active dune cordon samples. Dune samples from vegetated areas, both bulk and carbonate-free samples, also have a mean grain size of 180–150 µm and also show a decrease in the 250–212-µm size fraction between 6 and 24 km (Figure 5). The carbonate shell content of the mobile dune cordon sand varies from 45% at Sixteen Mile Beach to 41% at 6 km, 40% at 12 km, 14% at 18 km, and 3% at 24 km inland. The vegetated dunes show the same trend, but the percentages are lower (from 31% at 6 km to 2% at 24 km). The skeletal fraction of the dune sand is mainly composed of bivalve fragments, echinoid spines, foraminiferal tests (mostly Elphidium and Ammonia), and barnacle plates. Microfossils from the dune cordon show progressive abrasion downwind such that species identification is difficult beyond 12 km inland. No foraminiferal tests were found at 24 km inland. The organic content of the vegetated dune sand increases from 1 wt% at 6 km inland to 3 wt% at 24 km inland.

Radiocarbon ages of the bulk carbonate fraction of sand samples from the dune cordon range from 18.3 to 22.9 ka (Table 1; Figure 2). The mobile dunes have a radiocarbon age of 19.4 ka at 6 km inland, 22.1 ka at 12 km inland, and 22.2 ka at 18 km. A sample collected 1 km north of Gabbro Point from the pioneer dune gives an age of 20.6 ka. Samples collected from the vegetated dunes give ages of 19.8 ka at 6 km inland, 21.8 ka at 12 km inland, 18.5 ka at 18 km inland, and 22.9 ka at 24 km inland.

Sediment Distribution

The Sixteen Mile Beach Complex can be divided into three distinct regimes based on the variation in the beach sand grain size and dune morphology. The southern end from Gabbro Point to 6 km north is almost entirely composed of fine sand and is the take-off point of the 24-km long Yzerfontein–Geelbek dune cordon (Figure 2). In the central part, between 6 and 18 km north of Gabbro Point, there is a rapid decrease in the fine sand fraction from 96% to 10% that coincides with the transition from the inland dune cordon to a single, large coast-parallel dune ridge. The northern part of Sixteen Mile Beach, from 18 to 26 km north of Gabbro Point, consists of 83–90% medium sand with up to 6 wt% coarse sand, a narrow, steep backshore, and a series of prograded, vegetated coast-parallel inland dune ridges.

The variation in grain size of the beach sand along Sixteen Mile Beach reflects the amount of wave energy received from the predominant southwest swell. The variation in dune morphology in turn reflects the beach profile and grain size of the sand exposed to the strong southerly to southwesterly dry summer winds (Figures 3 and 4). The wave energy received along Sixteen Mile Beach generally increases to the north where the waves more directly impact the coast. The southern end receives less wave energy because it sits in the shadow of the rocky Yzerfontein headland as well as Dassen Island located 10 km southeast of Yzerfontein (Figure 1). Much of the energy of the predominant south to southwesterly swell is lost breaking upon and refracting around these rocky shores. The abundance of fine sand is consistent with the width and gentle gradient of the backshore along the southern part of Sixteen Mile Beach (Figure 3), a feature observed on beaches elsewhere in southern Africa (e.g., Marker, 1987). Dassen Island is particularly important in explaining the extent of the fine sandy beaches because the projected wave shadow coincides with the rapid decrease in the fine sand fraction from 96% to 76% between 6 and 7 km north of Gabbro Point (Figure 4). The changes in grain size and backshore gradient are also reflected in the offshore bathymetry, with both the 30- and 50-m bathymetric contours narrowing by more than a factor of two between the southern and northern ends of Sixteen Mile Beach (Figure 2).

The accumulation of fine sand as broad, low gradient beaches in the semiprotected waters at the south end of Sixteen Mile Beach provides the source of sand blown inland to form the Yzerfontein–Geelbek dune cordon. The dune cordon is a patchwork of vegetated and nonvegeted dunes. The mobile, nonvegetated irregular barchanoid dunes, of for example the Geelbek Dunefield, move rapidly enough (3.6 to 5 m/y) to avoid stabilization by vegetation. Aerial photographs spanning the last 55 years indicate a complex pattern of dune-field movement within the cordon influenced by climate (rainfall, wind) and fire. The slight downwind decrease in the dune cordon sand size indicates preferential transport of the finer sand grains, whereas the decrease in carbonate reflects dissolution by rain as well as mixing with leached quartzose, upper Pleistocene palaeosols. Planting of aggressively invasive alien vegetation (primarily Acacia cyclops and A. saligna) has resulted in the stabilization over the last 40 years of much of what were mobile dunes at the southern end of the dune cordon.

Similar to the Yzerfontein–Geelbek dune cordon, the major dune cordons at the Cape Flats and Atlantis occur upwind from relatively low-energy beaches protected from the southwest ocean swell by the Cape Peninsula and Robben Island, respectively (Figure 1). The northward decrease in the aerial extent of the dune cordons from 243 km² at the Cape Flats, 88 km² at Atlantis, and 43 km² at Yzerfontein reflect the rapid decrease in river sediment discharge north of Table Bay (Franceschini, Compton, and Wigley, 2003).

The transition from the inland dune cordon to a single, large coast-parallel dune ridge coincides with a rapid increase in medium sand 8 km north of Gabbro Point (Figure 4). The marked change in dune morphology reflects the decrease both in the width of the backshore and in the mobility of medium sand, both of which limit the amount of inland sand transport by wind. A single, large, coast-parallel dune ridge continues to 17 km north of Gabbro Point as the percentage fine sand decreases steadily to 10%. From 19 to 26 km the coarse sand size increases to 6% along the beach and a series of older, vegetated Holocene coast-parallel dune ridges occur inland of the actively accreting foredune ridge. The active coast-parallel dune ridge is mostly accreting vertically. The crest and leeward side of the active, seaward dune ridge is partially vegetated and is migrating inland as beach sand bypasses the deflated dune face, which in many places is covered by a layer of gull-dropped bivalve shells to accumulate in small partially vegetated dunes at the crest of the dune ridge. The wind preferentially removes fine sand from the beaches at the northern end of Sixteen Mile Beach as indicated by the significantly higher fine sand content (75%) of the coast-parallel dune ridges. The dune ridge also migrates inland by large parabolic blow outs. The mean grain size of the wind blown sand is greater, and as a result the rate of inland movement is much slower than that of the southern dune cordon. Based on aerial photographs, the maximum rate of inland migration over the last 50 years is on the order of 0.4 to 2 m/y at the northern end of the dunes. The vegetated coast-parallel dune ridges extend over the much larger 40–60 m high Pleistocene dune ridge inland to the edge of the peninsula at Kraalbaai. Radiocarbon ages indicate that these dune ridges have prograded seaward since 7 ka in response to variations in sediment supply and sea level (Compton and Franceschini, 2005; Franceschini, 2003). Therefore, the northern end of Sixteen Mile Beach has prograded seaward as a series of coarser grained coast-parallel dune ridges, whereas fine sand at the southern end has migrated farther inland as a dune cordon consisting of a series of episodic, rapidly moving dune fields.

Radiocarbon Ages of Sixteen Mile Beach

The progressively younger age of bulk sand along Sixteen Mile Beach from 0 to 26 km north of Gabbro Point corresponds to an increase in grain size (Figure 6). The correlation between age and percent fine sand indicates an end member mean age of the fine sand of approximately 18.8 and 11.5 ka for the medium sand. The bulk age of 10.9 ka of the sand sample from Tsaarbank, 26 km north of Gabbro Point, falls off the age trend in Figure 6 because it contains 6 wt% coarse, presumably younger, sand grains. In addition to the percent fine sand size, the percentage of colored and transparent grains tends to decrease with increased age of the sand samples (Table 2). Sand-size fragments of modern mollusk shells are more colored and transparent, and less abraded and rounded than older shell fragments observed in samples of upper Pleistocene aeolian deposits at Yzerfontein and Langebaan Lagoon. The Pleistocene dune samples contain no colored grains but do contain a minor component of transparent grains. Therefore, the northward trend to younger ages can be attributed to a greater mean grain size and a greater proportion of recently broken shell material.

The AMS age of 2.4 ka for pink and purple colored carbonate grains (∼300), handpicked from the bulk sand sample B3, 4 km north of Gabbro Point with a bulk age of 18.4 ka, indicates that the sand is a mixture of young and old shell fragments. The young colored carbonate grains are derived from the fragmentation of fresh shells whereas the old shell fragments are derived from the reworking of Pleistocene beach and dune deposits. The percent of colored grains in sample B3 is 5% and implies that the majority of the fine sand was derived from reworking of older deposits whose carbonate grains contain little if any ¹⁴C.

The bulk beach sand ages can be corrected for the mixing of old, opaque carbonate grains by assuming that all of the ¹⁴C resides in the colored and clear grains. In addition to 5% colored grains, sample B3 contains around 20% clear carbonate grains (Table 2). If the opaque grains are assumed to contain no ¹⁴C, then the remaining 25% colored and clear grains in sample B3 have a mean age of 7.4 ka (40% modern carbon). The beach sample L13 from 26 km north of Gabbro Point (Tsaarbank) contains 5% colored and 55% clear grains. If it is assumed again that all of the radiocarbon in sample L13 resides in the 60% colored and clear grains, then they have a mean age of 7.4 ka. Therefore, the age of the colored fine sand in sample B3 of 2.4 ka and the similar mean age of 7.4 ka for the colored and clear grains of samples B3 and L13 indicate that the northward trend of decreasing age mostly reflects an increase in the percentage colored and clear grains rather than simply grain size. This implies that the fragmentation of carbonate shell produces a wide range of grain sizes simultaneously rather than progressively producing smaller grains over time as suggested by Illenberger and Verhagen (1990).

These older, recycled carbonate sand grains were eroded from older dune and beach deposits as well as perhaps transported across the shelf by the migrating strandline as the sea level rose. The relative significance of the two sources is not known, but recent erosion of uppermost Pleistocene beach sand (Velddrif Formation) and dune sand (Langebaan Formation) is clearly visible along parts of the coastline. Strand-line migration of sands across the shelf during the most recent marine transgression has been documented in Australia and elsewhere (Bird, 1978, 1996; Bird and May, 1976). Erosion of carbonate sand from the 79 to 117 ka Velddrif and Langebaan formations would contain no ¹⁴C, whereas migrated strandline beaches would contain at least some ¹⁴C. The high opaque carbonate grain content in the fine sand of Sixteen Mile Beach may result from the large amount of fine sand derived from the erosion of uppermost Pleistocene dune deposits that have a high proportion of ¹⁴C-free fine carbonate sand. The high percentage of colored and clear medium sand grains appear to indicate that they are derived primarily from strandline migration of more recent beach sand across the shelf. Therefore, radiocarbon analysis of the bulk sand samples yields maximum ages that depend on how much ¹⁴C-dead carbonate they contain. The percent dead carbonate is difficult to quantify from the texture of the grains and complicates the interpretation of bulk sand ages.

In summary, the majority of sand on Sixteen Mile Beach is recycled. Much of the fine sand is derived from the erosion of Pleistocene dune deposits formed during the previous sealevel high stand and much of the medium sand is derived from the migration of beach sand as the strandline transgressed over the shelf since the Last Glacial Maximum. Fragmentation of shell along this high-energy, rocky coast has continually added younger carbonate grains of widely varying grain size to this large reservoir of recycled sand since sea level has returned to its present-day position. The AMS age of 2.4 ka for the fine-fraction colored grains is interpreted to represent the mean age of this Holocene beach sand established approximately 5.5 ka when sea level returned to its near present-day position (Compton, 2001). The mid-Holocene sea-level high stand of approximately +3 m eroded large quantities of sand from the uppermost Pleistocene beach and associated Yzerfontein–Geelbek dune deposits. After the high stand and a return to sea level at more or less present-day levels (±1 m), the beach stabilized and the breakdown of shell material was continually added to the beach since approximately 5.5 ka. The abundance of recycled sand is consistent with this section of generally sediment-starved coast with the closest perennial river located 80 km south (Figure 1).

Radiocarbon Ages of the Yzerfontein–Geelbek Dune Cordon

Subtracting 18.4 ka, the average age of the source of fine sand at the southern end of Sixteen Mile Beach, from the age of the mobile dune sand samples gives corrected ages of 1 ka at 6 km, 3.7 ka at 12 km, 3.8 ka at 18 km (Geelbek dune field), and 4.5 ka at the most distal and vegetated dunes 24 km inland (Figure 2). The corrected age of the mobile dune samples increases with distance from the beach (Figure 2). The mean dune migration rate of 5.3 m/y is higher than the average of 3.6 m/y measured from aerial photographs spanning the last 55 years but comparable to the rate of 5 m/y measured within the Geelbek dune field. The older than expected age of the sample collected at 12 km inland is attributed to its finer carbonate grain size and to fewer colored and clear carbonate grains (Table 2), and suggests that it may have a greater proportion of reworked sand from the underlying uppermost Langebaan Formation dune deposits. The less consistent and generally younger ages of the vegetated dune sand samples within the Yzerfontein–Geelbek dune cordon are attributed to partial recrystallization of the carbonate in association with soil processes of evaporation, growth of roots, and fungus (Geyh and Schleicher, 1990; Grootes, 1983). The corrected ages suggest that the dune cordon has been active since 4.5 ka. An age similar to the 5 ka Cape Recife and Cape St. Francis headland bypass dune fields located in the Eastern Cape (Illenberger, 1988).

The bulk sand carbonate ages and distribution of large mobile dune fields suggest a complex evolution of the Yzerfontein–Geelbek dune cordon during the Holocene influenced by variations in sand supply, sea level, and climate. Sand supply and sea level are believed to be closely linked because of the large percentage of recycled sand from the erosion of older deposits during marine transgressions. The presence of large dune fields within the dune cordon suggests that the sand supply has not been continuous. Pulses of sand in the form of discrete dune fields are commonly observed. For example, the Swartlintjies River dune cordon, located 350 km north of Yzerfontein (Figure 1) has a similar inland reach of 24–27 km, with active dune fields located at 0–7, 11–14, and 21–25 km from the coast (Tankard and Rogers, 1978). Three major pulses of sand were proposed for the Holocene Algoa dune fields on the South Coast (Illenberger, 1988). These pulses may result from a sudden increase in sand supply to the beach from erosion during relatively modest marine transgressions or increased supply of sand to the coast via rivers. However, the abundance of recycled older carbonate at Sixteen Mile Beach would suggest that erosion of Pleistocene dunes has been more important than an increase in shell production on the coast or sand supply from rivers.

Pulses of sand as large mobile dune fields in the Yzerfontein–Geelbek cordon are believed to mostly represent relatively modest (±1 m) Holocene sea-level fluctuations. Because of the strong onshore wind, the majority of the large increase in sand supplied to the beach from the erosion during a rise in sea level of previous dune deposits is recycled via the beach and offshore sandbars back onto land (Cooper, 1958, p. 131). In the case of a rapid and large sea-level rise, dunes can be abandoned and submerged offshore. In a modest sea-level fall, the beach is momentarily starved of sand and there is less sand fed into the mobile dune cordon until a new beach is established. At the northern end of Sixteen Mile Beach, the beach has prograded seaward as a series of abandoned dune ridges since the mid-Holocene high stand (Compton and Franceschini, 2005; Franceschini, 2003). In the south, the abandoned foredune continued to migrate inland. Therefore, the minor fluctuations in Holocene sea level along the coast inferred from the salt marsh deposits in Langebaan Lagoon (Compton, 2001) resulted in variations in the sand supply to the beach and the transport of sand inland. In addition to sea-level fluctuations, changes in climate (rainfall, wind strength) and destabilization of dune sand after fire would have modified the amount of sand transported inland. Smaller dunes may break away or coalesce with larger dunes. Additional sand can be taken up by erosion of older, underlying sand, and sand can be stabilized, at least temporarily, in vegetated trailing parabolic ridges. These complexities in dune movement may account for the variable distribution and age of active dune fields among South African coastal dune cordons.