From May to July 1998, the Department of Archaeology, Simon Fraser University undertook a two and a half week archaeological field project at the Sigatoka Sand Dunes National Park on the Coral Coast of Viti Levu. Conducted under research permit and in co-ordination with staff of the Fiji Museum, the project served as one component of an archaeological field school and study abroad program organized through the Institute of Pacific Studies, University of the South Pacific. In compliance with research permit requirements, the present report provides an account of the Sigatoka project, its research design, its various operations, and a preliminary series of interpretations.
The Sigatoka Sand Dunes National Park extends along the Coral Coast of Viti Levu from the mouth of the Sigatoka River west to Naqarai Bay, a linear distance of approximately 5 km (Figure 1). It is a massive complex of parabolic dunes that in area encompasses more than 2.5 km2. Within or buried beneath this sand accumulation are archaeological remains spanning the duration of Fijian prehistory from early Lapita settlement to historic times. Since the coastal margin of the dune field is largely unstable with heavily eroding slip-slope faces and blowout surfaces, these remains continually are being exposed and in many cases destroyed. Over the past half century, numerous archaeological projects have been undertaken at Sigatoka as a consequence (Wood et al. 1998: 11-14) . The recovered data have figured prominently in the development of a culture historical sequence and chronology for Fiji.
In recent years, a number of reports on archaeological projects at the Sigatoka Sand Dunes have provided extensive reviews of geology, dune formation processes, human occupancy, site impacts and degradation, past research projects and other matters (Hudson 1994, Petchey 1995, Burley 1997, Wood et al. 1998, Dickinson et al. 1998). Such an examination need not be replicated here, and the reader is referred to these earlier works for detailed background data. To understand the 1998 project research design and results, nevertheless, requires a selective summary of the more salient interpretations.
First, the densest concentration of materials at Sigatoka occurs within a 1 km section of shoreline along the eastern frontal slope of the dune. This area has been designated VL 16/1 (Figure 1). As it has been ideally modeled, VL 16/1 incorporates three stratigraphically separated paleosols, each taken to represent a period of surface stability in which soil development took place. Labeled Levels I to 3 from earliest to latest, each has been correlated with a different temporal period of Fijian prehistory. Respectively these include Lapita, Navatu and Vuda/Ra phases. Surface exposures of these paleosols and eroding archaeological materials at VL 16/1 were mapped in detail in 1992 by Wood et al. (1998). In 1993, surf surge by Hurricane Kina was thought to have destroyed most, if not all, of Level 1 and its associated cultural deposits; a significant erosional impact also has been inferred for Level 2 (Hudson 1994, Burley 1997).
Second, in 1965/1966 an expansive excavation project of 2677.5 m2 was undertaken by Lawrence Birks (1973) at VL 16/1. This project illustrated each paleosol to have a rich ceramic assemblage, significantly including several reconstructable vessels for the initial Lapita phase. Birks, however, found no architectural features for any of the occupation zones and could only speculate upon the human use of the Sigatoka coastline. More recent excavations, while illustrating the presence of an extensive cemetery complex during the Level 2 Navatu phase (Best 1988), have similarly failed to identify architectural remains and/or conclusively infer the nature of individual occupation floors.
Third, a 1996 project by Dickinson et al. (1998, also Burley 1997) has re-examined the geological context of individual paleosols, the archaeological data associated with each, and the chronological implications that these have for sand dune formation and origins. The conclusions of this project support an earlier suggestion by Palmer (1987) that massive development of the parabolic dune field did not occur until relatively late in prehistory (post 1500 BP). Athough the arguments are largely hypothetical, this event is attributed to human deforestation of the interior highlands with consequential slope erosion and enhanced sediment loads in the Sigatoka River. Ultimately the sand was redeposited at the river mouth and along the Sigatoka shore where it was swept obliquely inland by southeast tradewinds. If the argument for late prehistoric dune formation holds true, then the Level I and Level 2 human occupations at Sigatoka occurred on a deltaic landscape considerably different from that of today. These interpretations have been questioned by Wood, Marshall and Crosby (1998: 74).
Finally, while the densest concentration of archaeological remains occur at VL 16/1 as described, ceramic scatters, features, and burials have been found eroding throughout the entire coastal margin of the dune field, in intervening blowout valleys, and in a field area behind the eastern end of the dune (Burley 1997; Wood et al. 1998). In eroding slip-slope faces, however, only a single paleosol has been observed. Ceramic types and radiocarbon dates from this paleosol associate it with Level 3 of the eastern dune (Dickinson et al. 1998). On the far western end of the dune field in Naqarai Bay, Birks (1966) further recorded and excavated Site VL 16/22 in 1965 (Figure 1). As described, this site included a stratigraphic sequence of four occupation layers with small numbers of dentate stamped and red-slipped Lapita sherds occurring in the earliest occupation level.
The 1998 field program at the Sigatoka Sand Dunes had to accomplish three inter-related objectives. As one component of a field school program, field work first had to be structured in such a way as to maximize training opportunities for students. As wide a range of experiences as possible, therefore, was sought in survey, field recording, map preparation, and excavation projects. Because of the continual erosion of archaeological materials, including human burials, resource management concerns of the Fiji Museum had to form a second set of objectives. These concerns largely involved the identification and recording of newly exposed archaeological materials throughout the park and, where necessary, the implementation of a rescue excavation program. Finally, and above all else, the first two goals had to be developed within the framework of a research design in which the field project could contribute to a further understanding of Sigatoka prehistory. This is outlined below.
The research design for 1998 field study was focused upon the problem of chronological
origins for dune formation. Based on the 1996 Dickinson et al. (1998) project, it was
hypothesized that massive sand accumulation did not begin to occur until after 1500 BP and that
dune formation generally took place on an east to west basis. To test these hypotheses, two
projects were proposed for implementation:
1) an intensive survey of dune face erosional exposures from Naqarai Bay on the west to VL 16/1 on the east. If the proposed hypotheses were correct, then ceramic types and other archaeological materials eroding from paleosols throughout the dune field should relate exclusively to the late prehistoric Vuda phase or the historic period Ra phase. Earlier materials, if present, would be restricted to basal sediments of the Sigatoka delta. Test excavations at select features, once discovered, were to be conducted for the collection of diagnostic ceramic types and radiocarbon dating samples.
2) a limited re-excavation project at site VL 16/22 in Naqarai Bay. Based on the 1960s project, Birks (1966: 8) concluded that VL 16/22 stratigraphy indicated "a recent dune formation, now largely stabilized, overlying an earlier soil horizon on which human activity had taken place". The proposed excavations were to assess this conclusion and provide a terminus post quem (date after which) for the major period of dune development based on diagnostic ceramic types and radiocarbon dates. If successful, this excavation would provide a direct test of the late prehistoric dune formation hypothesis.
As the above projects were being implemented, a chance discovery was made of an undisturbed segment of the Level I occupation zone at VL 16/1. An excavation of this deposit afforded the opportunity to reassess Birks (1973) description of the level, recover a collection of reconstructable late Lapita ceramic vessels as well as acquire charcoal samples for additional radiocarbon dates. In so far as this exposure was no more than a few metres above the existing high tide line, a rescue excavation also met salvage concerns of the Fiji Museum. Accordingly the project was implemented as part of 1998 field work activities.
Field work at Sigatoka was conducted from June 15 to July I with student, volunteer, and Fiji Museum field crews of up to 21 individuals.