4 THE TO'AGA SITE: MODELUNG THE MORPHODYNAMICS OF THE LAND-SEA INTERFACE PATRICK V. KIRCH IN CHAP 2 wE described the economic and cultual significance of the narrow coastal terrace of Ofu which-a~ the only flat land on the island-is thus the main locus of settlement and a major zone of subsistence prduction. This coastal tenrace, which the To'aga site exemplifies well, must be undertood as a dynamic geomorphological entity lying at the cdtical land-sea interface. McLean has aptly de- scribed the imporance of such land-sea interfacial zones on small tropical islands: The atactiveness of the coastal lowlands for setlement purposes is obvious.... On many islands they comprise dte only flat country. It is easy to move over and easy to build on. Soils, though naurally of low fertlity and moise statis, are easily worked. The sandy subsate, beach sand and beachrock are readily available for such uses as building matials, graves and road 'metal.' That the coastal flats are multi- purpose and multi-utilized resources goes without saying. Moreover, their location at the land-sea edge permits optimal access to a complete range of terrestrial and imaine environments and rsources: hilislope, valley and forest; reef, lagoon and fish. They function as convenient bases to exploit the local environs. They also function as verandahs to the worlds outside (1980:.129-30). In their study of the prehistory and ecology of Tikopia, Kirch and Yen (1982) similarly argued for the importance of unraveling the morphodynamic histories of tese interfacial enviromnents: among these [environments] are the coastal/ marine interface, with implications for access to rsources of reef and sea, and that of calcareous plains and volcanic hills, with implications for erosion and agricultural developmenL ... Both of these interfaces have been active zones of landscape change, with long-tern implications for human adaptation (1982:17). In particular, the Tikopia case revealed the agricul- tural significance of human-induced modifications to the coastal zone: 'te plain and hill interface pm- vides the opporunity for the mixing of volcanic and calcareous soils to form edaphic media, more favor- able for the cultivation of root crops than either separate type" (Kirch and Yen 1982:17). Prior experence in Tikopia (Kirch and Yen 1982), Niuatoputapu (Kirch 1988), and other small tropical islands had sensitized us to the necessity of a geomorphologically informed approach to the archaeological study of the To'aga site. This was critical both to an undemanding of depositional and site-formation processes, as well as for contribudng to our knowledge of the evolution of the island's larger 32 The To'aga Site settlement paterns and production systems. This chapter focuses on several aspects of the local envinnt at To'aga which played key roles in te morphodynamic evolution of the site. Chief anong these are the ging configuration of sea level in the mid-Holocene, and the tectonic instability of the Manu'a Group. Equally sigficant, however, is the role of humans tiemselves in shaping landscape through physical manipulation of soil and rock, forwst clearance, the inroduction of a "portnantea" biota (cf. Crosby 1986), and the continual manipula- tion of the coastal enviro to best suit human needs. My aim here is to outline a model of these morphodynamic processes that may then be tsted agains the empirical evidence revealed thrugh our field and laboratory investigations. GEOMORPHOLOGY OF THE TO'AGA COASTAL TERRACE The key geomorphological features of the coastal terrace at To'aga can best be descnbed along a generalized transe ruing from the reef inland across the flat up to the steep hillslope and cliff as in figure 4.1. (Particular, surveyed transects are illustraed and described in chapter 5, along with the statigraphic evidence for temporal development) At the seaward end is the reef flat, source of te calcar- ous sediment (composed of the detritus of coral molluscs, and other reef-dwelling organisms) of which the coastal terrace is primarily constructed. This reef flat is a mosaic of microenvirornents, including surge channels, the 'lithotamrnium ridge' at the seaward edge, large coral heads, and sandy patches. A careful exanination of the beach slope reveals several clues as to the dynamic processes presently at wor along the land-sea interface. At the foot of the beach slope in various places are exposed layers of 'beach rock' (Wiens 1962:64-67) made up of sand and coral rubble cemented with calcium carbonate (CaC03). Such beach rock deposits can fonn fairly quickly under tidal conditions of continual wetting and drying witfin active beach ridges. Their exposure along the To'aga beach front, however, signals a cument phase of coastal erosion and net sediment loss (or lateral transfer). This evidence is reinforced by an examination of the vegetation line at the top of the present beach ridge crest, where large trees (such as Cocos, Cordia, and Hernandia) have been undercut and eroded by wave action. Moving inland, one mounts the crest of the present beach ridge or benn, lying between the beach slope and the sandy road. This nanrow zone is covered with a thick tangle of vegetation, dominated by Pandanus tectorius, Barringtonia asiatica, Hernandia nymphaeifolia, Scaevola taccada, Messerschmidia argentia, Cocos nucifera, and other species that anchor the loose, unconsolidated calcareous sands. Test pits dug into this beach ridge revealed only a thin (ca. 5 cm) organic A horizon directly overlying the young parent sands. The absence of archaeological deposits suggests that the present beach ridge is of no great antiquity. Crossing the road and plunging thrugh a narrow band of Pandanus and Hibiscus tliaceous shrubs, one enters the main extent of the coastal tenrace, a zone of intensive economic utilization though arboriculture and root crop gardening (see Kirch, chapter 2). The ground at first slopes down slightly from the beach ridge, levels out, and then begins to rise again gradually toward te talus slopes inland. As one moves inland from the road through this zone, the soil gradually becomes less sandy and more clayey, reflecting the increased contribution of volcanic colluvium which has been eroded frxm the hillslopes and added to the calcareous sands that form the main component of the substrate. The mixing of tese calcareous and volcanic sediments- throgh the continual action of human cultivation- has made this zone of particular edaphic value for Samoan horficulture. This is also the main zone of archaeological feaures, both surface and sub- surface, as described in chapters 3 and 5. Moving farter irland toward the talus slope, one begins to note volcanic cobbles and boulders stewn over the land surface, signals of the dynamic instability of the 500-m-high cliffs that tower over the To'aga flaL Some of these boulders are several meters in diameter and can fall with considerable destructive force. [be ground surface closest to the talus is entirely gravelley clay-loam with no calcare- ous component evident Rather abruptly, one arrives at the base of the talus itself, an imposing and unstable jumble of boulders rising steeply toward the cliffs, over which grows a tangle of Hibiscus tiliaceous and scattered forest trees such as Erythrina varigaa. In sum, this geomorphological traverse across Morphodynamics of the Land-Sea Inerface 33 Figure 4.1 AgeneralizedtasectacrossthecoastalplainatTo'aga,showingthedistributionofcolluvialandcacareous sediments, major vegetation associations, and geonnorphological featu . the To'aga coastal terrace reveals several key feaues: (1) the terrace is constructed primarily of marine biogenic sediments (calcareous sands and larger clastics); (2) one finds a progressively greater contribudon of terrigenous sediments as one ap- proaches the volcanic mass; and (3) there is some evidence for curnt coastal erosion and maine transgression at the present time. These are features that must be accounted for in any model of the morphodynamics of the To'aga site over the past several thousand years. MORPHODYNAMIC PROCESSES: SEA LEVELS AND SUBSIDENCE As the archaeological deposits of the To'aga site are integral sedimentary components of the coastal terrace, any model of site formation processes must first account for the geomorphological processes by which the terrace was formed. Lying at the critical and highly dynamic land-sea interface, any consider- ation of the morphodynamics of this zone must begin with an exanination of controlling processes in shoreline formation. Among the most important of these are: (1) glacio-eustatic sea-level changes; and, (2) the local tectonic situation which-as we fshall argue-is one of subsidence. The temporal penod with which we are concemed is the mid- to late-Holocene, from about 6 kyr B.P. to fte preet This time span covers some 3 kyr piior to the human colonization of Ofu as well as the subsequent interval during which humans have added their input to the development of the local lanscape. Before examining the evidence for reladve sea- level change during this period, it is essential to inrduce a few key concepts regarding coastal change on small islands. In this I have drawn primarily from Chappell (1982) and McLean (1980b). The construction of a coastal tenrace such as that at To'aga results from progradaion, the "progressive formation of new land by sedimenta- tion inespective of fte tendency of sea level move- ment" (Chappell 1982:71). Sea-level changes themselves, whether due to glacio-eustacy or tectonic movements, or both, are inportant as controlling factors for the sediment budget, but alone ftey do not provide a sufficient model of prgrada- tion. As Chappell emphasizes, ". . . sea level changes alone cannot be used to account for coastal changes. In fact, for the last 6000 years, the sedi- mentary budget is the more important factor' (1982:71). The sediment budget can be thought of as te net sum of sediment input from both terrestrial ARBORICULTUR4L ZONE STRAND VEGETATION C_lluvium / i Erosion Scarp :: .:: : : .: :: : . . : >,Exposed Beach Rock Unconsolidated Calcareous Sands :. Coral Reef Platform- 34 The To'aga Site (talus, colluvium, etc.) and marine biogenic (calcare- ous sands and coral detritus) sources, minus the loss of sediment from transport (fig. 4.2). In modelling the Toaga coastal terrace formation, therefore, we need to pay particular atention to anges that would either increase or decrease the production of sediment from te ial or marine sources. Figure 4.3 graphicaly portrays the dynamic effects of dcanges in relative sea level (due either to glacio-eustatic or tectonic change), combined with increases or decreases in sediment budget. In this diagram, sediment budget is indicated along the x axis, and relative sea-level change along the y axis. The heavy diagonal line separates transgression or coastal rera, frm regression or coastal advance. Holocene Sea Levels in dhe South Pacific A rapid rise in sea levels following the end of the Pleistocene is a global phenomenon that has been widely rognized (Fairbridge 1961; Shepard 1963). More controversial-because they depend upon a complexity of local conditons and processes-have been the details of the eustatic sea-level curve in the mid- to late-Holocene, especialy the matter of whether there have been higher-than-present stands. Bloom (1980, 1983) modelled some of the global diversity in these Holocene curves and suggested that a +1-2 m stand existed in the south Pacific region during ts period. Substantial geomorphic and radiometic evidence from a variety of islands now supports this intepetation of a +1-2 m high sea level during the period between about 4-2 kyr B.P. In Fiji, for examnple, Nunn (1990:304) concluded that the coasts "experienced a middle to late Holocene sea-level maximum some 1-2 m above prset mean sea level." Recent work by Miyata et al. (1990) also supports this finding. Similar results are presented by Ash (1987) for Viti Levu Island, and Yonkura et al. (1988) report evidence for a +1.7 m stand be- tween 3400-2900 B.P. on Mangaia Island in te southem Cooks. In French Polynesia, Pirazzoli and Montaggioni (1986,1988; Montaggioni and Pirazzoli 1984) describe evidence from various islands for a MSL between +0.8 and 1.0 m begin- ning about 6-5.5 kyr B.P. and lasting as late as 1.2 kyr B.P. In Westem Samoa, Rodda and his col- leagues (1986; Sugimura et al. 1988) summarize varous evidence for Holocene higher stands. Isla (1989:361-63) discusses a comparable range of evidence for several Pacific Islands. In figure 4.4, the geographic distibudon of mid- to late-Holocene higher sea-level stands is plotted along with associ- ated radiometric ages. Figure 4.5 shows a time- elevation plot of sea levels in various Polynesian archipelagoes over the past 5 kyr B.P. This widespread and consistent patem of radiometrically dated shoreline features provides strong evidence for a higher sea level ranging between about +1-2 m over fte southwestem Pacific, from at least 5 kyr B.P. and lasting until sometime between 2-1 kyr B.P. After about 2 kyr B.P., sea level fell (pedaps faidiy rapidly) to its present position. TIis mid- to late-Holocene sea-level curve thus provides one important dimension for a model of coastal terrace formation at the To'aga site. Subsidence in the Samoan Archipelago Most of the volcanic archipelagoes of the centrl Pacific region are aiayed lieiy along age- progression sequences emanating from a "hot spot" or magma plume on the floor of the Pacific Plate (Menard 1986). As the Pacific Plate gradually migmtes west or nortwest, active volcanic islands move off te hot spot, and a new island is formed. Thus, the islands in such an archipelago comprise a "plume trace" of inceasig age from east to west Classic instances of this pattem are the Hawaii- Emperor chain and the Society Islands. The Samoan archipelago presents a somewhat more complicated geological picture, primarily due to reent volcanism on Savaii Island which, at the westem end of the chain, theoretically shodd be the oldest island (Menard 1986:186-87). This unusual siution confounded geologists for many years (Dana 1849; Daly 1924; Steams 1944; Natland 1980). Recently, however, K-Ar dating of rocks from various Samoan islands has confirmed the basic pattem of an east to west age pogression; the mean K-Ar ages are 0.1 myr for Ta'u, 0.3 myr for Ofu-Olosega, 1.26 myr for Tutuila, and 2.2 myr for Upolu (McDougall 1985:318-19). Savai'i is presumably the oldest island but has had renewed volcanism in recent times. McDougall concluded ta "Upolu, Tuuila, and the Manu'a Islands comprise a hot spot or plume trace on te Pacific plate, similar to the Hawaiian and Society Islands chains." Nevertheless, the Morphodynamics of the Land-Sea Interface 35 Figure 4.2 A model for the sediment budget at To'aga, showing terrestial and marine inputs. dominane of youthful volcanism on Savai'i remains an egngma and may indeed reflect major rejuvenes- cence related to deformation of the Pacific Plate adjacent to the Tonga Trench (1985:319; see also Menard 1986:186-87). This linear age-progression and the fact that th I 0 E .' co Sea Level + - 0 E .0 co Sea Level - Figure 4.3 A model of shoreline transgression and regres- sion as a function of relative changes in sea lev- els and in sediment bugt (after Chappell 1982). Manu'a Islands are sll in a youfl stage of geological evolution are essential factors in under- stading the local tectonic siion of Ofu Island. The rapid constuction of a volcanic mass on the oceanic crust of the Pacific Plate results in a phase of subsidence, due to point loading on the thn and flexible crst (Menard 1986:165-69). Such subsid- ence and associated crustal defornation is well documented for the younger Hawaiian Islands. The island of Hawai'i appears to be subsiding at an average me of about 4.8 mm/yr. For Ofu Lsland itself, we are not aware of geological documentation of subsidence at the present time, although we have strong indirect reasons to believe that this is the case. The young age of te island (only 0.3 myr) would itself suggest that subsidence and crstal deformation (which lag behind volcanism) have not yet weached equilibrium. The evidence for active erosion of the island's coastline, descrbed above, stmngly supports this interprtation of active subsidence. Further evidence is provided by the reefs along the To'aga area which display active coral head growth and lack the solution-pitted and eroded reef platforms typical of islands that have been teconically stable during the Holocene. These observations suggest that Ofu and Olosega islands are still in a phase of subsidence due to point loading on the oceanic crust resulting from their initial volcanic construction. Elsewhere in Samoa, dramatic evidence for rapid subsidence was provided by the chance TERRIGENOUS MARINE INPUT INPUT Landslides Mass Wasting | SEDIMENT BUDGET H Srgh Energye Surface Flow Volcanic Clastics Biogenic M(oral) Clastics Colluvium TRANSPORT LOSS Calcareous Sand Soil Reef Detritus 36 The To'aga Site *1 o a) 0o I Is 0 I- a) 0C) .- 8t 04 c 0 .o ? bo Io kI ,.I qq *s Morphodynamics of the Land-Sea Interface _~ - K)4PW4 1-- 164 1* .~~ ~ ~~~~~~ ~ km _s . ,_ t~~~~ t ' " '~P-M p lwpgf_ L~~~~~o , -oo- .NM r~~~~~~~~~~~~~~~~~66 09 I 2 2 I 3 3 I 4 4 SOCIETY ISLANDS o Scilly " Mopelia * Maupiti v Tupai v Bora Bora & Raiatea A Huahine * Moorea o Tahiti Nui o Tahiti Iti I 8 6 N.W.TUAMOTU * Makatea * Rangirioa v Takapoto * Mataiva o Arutua a Kaukura o Apataki M 5 4 + 4 04*4 ~0: 5. 4 m GAMBIER ISLANDS E.TUAMOTU a Temoe w Vahitahi 20 Reao +1 - m.s.l I I . . I a . I a I I I COOK ISLANDS o Mangaia o Aitutaki A Rarotonga 1- - 104W 0 I .+ 1 1.04 B..4 I 2 2 - _ ,4, , Age (Kyr) 4 I 5 5 I 6 6 Figure 4.5 Time-elevation plot of sea levels in Polynesian archipelagoes (source: Spencer 1989, fig. 13). 1.04 m +2- +1- m.s.I.- m +1 so I 5 5~~~~~~ .I I I I 0 1 -1.94 0 - 0 -- 1*64 mm 9 V 1 s0. P-6 :0:~ 2 3 37 I I 5 I 9. . I WAO404 38 The To'aga Site discovery of the Mulifanua archaeological site (Green and Davidson 1974; Green and Richards 1975; L and Gren 1989). Dating to 3251 ? 155 B.P., Mulifanua is a Lapita-pottery bearing deposit now situated -1.5 m below mean sea level, and further capped by 0.75 m of reef rock (Leach and Green 1989:324-26). The site was accidentally discovered by dredging for a ferry berth Given ta the site represents a village occupation on a former beach or coast terrace, between 2.6-3 m of subsid- ence is indicated, or an approximate subsidence rae of I mAlyr. (Davis [1928:249-53] also discusses evidence for the submergence of Tutuila Island.) In constructing a morphodynamic model for the developnent of the To'aga cotal tenrace, we assume that Ofu has been undergoing continual subsidence trughout fte Holocene. Although no precise rate can be empirically determined at this time, we may use a woddng rate of 1 mAkyr based on the Mulifanua siuation. (It is possible that the aul rate of subsidence on Ofu could exceed this value.) SEDIMENT BUDGETS: SOURCES AND MODES OF DEPOSITION In addion to the relative changes in sea level resulting frm a combination of glacio-eustatic and tectonic processes, we need to consider both the sources of sediment which contributed to the costruction of the To'aga coastal terrace and the agents responsible for the deposition of these sediments. The bulk of the Toaga coastal terrace is constructed of calcareous sands and larger clastics (coral e and chunks of reef conglomerate) of marine biogenic origin. These biogenic sediments are produced on the reef thogh a number of mechanisms, including wave action and biologic processes (such as the generation of sand by parrot fish and other species that rasp and grind coral to extract algae). These biogenic sediments are transorted landward over te reef flat by wave action. Episodes of high-energy storm surges, associated with the cyclones ta perodically lash the island, are extremely important for the rapid accumulaion of large quantities of sand and of the larger size component of coral boulders and cobbles. The contribudon of terigeos sediments to the coastal terace appears to have been less significant, although this increases as one moves landward. The primary source of tenestrial sediment is the steep cliff which towers over the site, continually deposit- ing talus and rockfall onto the coastal teface. Some finer grained sediments wash over and down the cliff during heavy rains, but there is no major alluvial contribution of sediment Thus mass wasting (principally landslides) and sheet erosion are the major agents of terigenous sediment deposidon. The sediment budget of the Toaga geomorphic system would have fluctuated during the Holocene, as the generation of biogenic sediments in particular would vary with sea-level regimes. During periods of rapid sea-level rise, when corals are actively growing below mean sea level, the generaton of sediment would be substantially reduced. When sea level dropped or was stable for a perod of dme, coral growth would have caught up with sea level and would have been exposed to wave action, resulting in erosion and te generation of calcareous sediment MODELLING THE MORPHODYNAMICS OF THE TO'AGA COASTAL TERRACE IN THE MID- TO LATE-HOLOCENE The variables essential to a model of the forma- tion of the coastal terace at To'aga in the mid- to late-Holocene have been reviewed. In figure 4.6 these varables are diagrmmed along the same temporal axis. Figure 4.6A shows the Holocene glacio-eustic rise in sea level, reaching a +1-2 m maxnimum between about 4-2 kyr B.P. Prior to about 5 kyr B.P., the rapid rise in sea level would have continually drowned the shoreline, and e formation of a stable coastal terrace would not have been possible. Rather, the sea would have encrached direcly against the island's volcanic mass, creating the dramatic sea cliff behind the To'aga site, as described by Stice and McCoy (1968). Only after a maximum sea level was achieved, between ca. 4-2 kyr B.P., could a coastal terrace have begun to have formed along the base of the cliffs. At the sane time, we presume ta Ofu has been subsiding at a rte of about 1 m/kyr, as depicted in figure 4.6B. Tberefore one takes into account these two control- ling processes to determine the probable net change in relative sea level as graphed in figure 4.6C. This Morphodynamics of the Lad-Sea Interface 39 Figure 4.6 Time trends in four key variables affecting the morphodynamics of the To'aga coastal plain. graph predicts ta te most likely period during which the To'aga terrace could have prograded through the rapid deposition of biogenic sediments would have been between ca. 2-1 kyr B.P., when the eustatcally contolled sea level dropped from its mid-Holocene maximum down to modem levels. Depending upon the extent of this drop (1-2 m), the actual sea level on Ofu would have been either stable, or there may have been a slight fall. This would have resulted from te eustatic sea-level fall, off-setting the continual effects of tectonic subsid- ence. As shown in figure 4.6D, the biogenic sedi- ment budget would increase significantly during this penod as the coral reef flroing the To'aga area was exposed to wave action, especially stonn sures. During the past I kyr B.P., as sea levels again blized and local teconic subsidence coninued, te sediment budget would again have decreased, so that a phase of regression recommenced. This is consistent with our field observations of active shoreline erosion at the present time along the To'aga site. This model of the fonmation of the To'aga site can also be diagrmmed as a temporal trajectory along the two axes shown earlier in figure 4.3. In figure 4.7, we have plotted a rtDdiction of the most probable transgression-regression sequence for the To'aga shoreline over the course of the past 6 kyr B.P. In this diagram, Xt y-axis represents the net sea-level change resulting from the combinaion of both glacio-eustatic and local tectonic effects, while fte x-axis represents fte changing sediment budget. This model provides a wolking hypothesis for te fomiation of e To'aga coastal terrace and its archaeological deposits and may be tested against fte strtgraphic and radionetric evidence denived from our program of trasect excavations at various points along te site. Several predictions may be generated based on the model: (1) The earliest archaeological deposits should be located adjacent to the cliffs and should date no earlier than about 4 kyr B.P. (2) These early sediments are likely to have a higher component of volcanic clastics, because talus matenial would still have been readily available for erosion by wave action. (3) There should be a fairly rapid or even abrupt episode of coastal progradation beginning sometime after about 2 kyr B.P. and ending by about 1 kyr B.P. The sediments deposited during this interval would consist almost wholly of marine biogenics. This model was not developed prior to the actual fieldwork, but evolved during the courae of our investigations, as a dialectic between field observa- tions and theoreical exercises. Nonetheless, it is crucial to stress that the model in no way depends on our archaeological data; it is wholly independent, deriving from cumnt geological and geomorho- logical knowledge of coastal process in the southwest Pacific. To briefly anticipate the results of our field and laboratory studies presented in chapters 5, 6, and 7, we shall see dt the predictions of the model developed here are substantially bome out +2 A 0- -2 - EUSTATIC SEA LEVEL B SUBSIDENCE -. L ~~~NET SEA LEVEL D+l I ~~SEDIMENT BUDGET 6 5 4 3 2 1 0 KYR B.P. . 40 The To'aga Site I ._- 0 E -6 0 CO Relative Sea Level + .0- a) 1 E 0 C) Sea Level - Figure 4.7 Retoiction of the tansgrssion-regression sequence along the To'aga coastline fron 5 kyr B.P. to the peent. TO'AGA: SOME FURTHER EXPECTATIONS FOR ARCHAEOLOGICAL SITE FORMATION PROCESSES Schiffer has avenred that "during human occupa- tion of regions, natural processes, influenced by cultual behavior, have created an ever-changing landscape that the investigator perceives at just one point in time. The contemporary region is a com- plex, the-dimensional mosaic consisting of natural sediments, vegetation, modem artifacts and settle- ments, and archaeological remains. In order to find sites and, especially, to understand how settlements functioned in regional systems, one must endeavor to infer or reconstruct changes in the landscape" (1987:261). In this chapter, I have endeavored to model some of the key processes which contributed to the formation and modification of archaeological deposits in the To'aga area. Lying at a fragile interface between land and sea, the To'aga coastal terrace is subject to envirornental influences of several kinds, and can potentially undergo rapid changes if any of these inputs vary. This interfacial zone has also been subject to intensive human use which has implications for both te formation and modification of the archaeological record. Among the most important implications of our model of morphodynamics of the coastal terrace is that the very existence of this interfacial microenvi- rorunent depends upon a sediment budget which has fluctuated highly during the mid- to late-Holocene. Since the archaeological deposits at To'aga are an integral part of the geomorphological structure of this zone, it is essential that we understand these processes of coastal ten-ace construction and erosion. To quote Schiffer again, "from the standpoint of survey [and excavation] design, the archaeologist needs to know where deposition and erosion are occurring and where they have occunred during the period of human occupation of the region" (1987:257). Our model of relative sea level and sedimentary budgets suggests that the To'aga coastal ten-ace could not have begun to form or stabilize until after the Holocene sea-level maximum of about 5-3 kyr B.P. Given that the Westem Polynesian region was first colonized by the makers of Lapita pottery at about 3.4-3.2 kyr B.P. (Kirch and Green 1987; Kirch and Hunt 1988), the area of coastal terrace available for initial establishment of human habitations would have been very restricted, prob- ably consisting of little more than a beach ridge lying directly under the steep cliff. As the coastal terrace began to rapidly prograde after about 2 kyr B.P., the area available for intensive land use would have increased substantially. Such progradation, how- ever, combined with deposition of colluvium, probably would bury the earlier occupation deposits deeply. Thus, our morphodynamic model also implies that the earliest archaeological deposits will be the most difficult to locate, being situated fartest inland and possibly buried under substantial depths of colluvium. The model predicts, therefore, that archaeological deposits dating to the earliest phase of Samoan prehistory-the phase marlced by te presence of ceramic assemblages-are not likely to be exposed on present ground surfaces or to be encountered by surface survey alone, no matter how intensive. Indeed, the initial archaeological recon- naissance of the island by Sinoto and Kikuchi [Emory and Sinoto 1963] failed to locate a single site Morphodynamics of the Lnd-Sea Interface 41 of this early period. The model also has implications for the longer- tenn managemen of archaeological sites or 'cultual resources' of the To'aga area. We explore these implications in more detail in chapter 15, but briefly note here that if, as our model predicts, the To'aga area is now undergoing a phase of active erosion and shoreline transgression, the archaeological sites of this area will in time be eroded away. How quickly these sites will be theatened by erosion depends on several factors, especially the rate of subsidence and of eustatic sea-level rise. The problem of 'global waming' and associated sea-level ises (Geophysics Study Committee 1990), however, could potentially rsult in an acceleration of coastal erosion in the To'aga area. ACKNOWLEDGEMENTS I am particularly grateful to Joanna Ellison of the Depanment of Geography, University of Califor- nia at Berkeley, for kindly supplying me with vadous references on sea-level change in the southwest Pacific, for the use of figure 4.4, and for her comments on an earlier draft of this chapter. REFERENCES CITED Ash, J. 1987. Holocene sea levels in norerm Viti Levu, Fiji. New Zealand Journal of Geology and Geophysics 30:431-35. Bloom, A. L. 1980. Late quatemary sea level change on south Pacific coasts: A study in tectonic diversity. IN N-A. Momer, ed., Earth Rheol- ogy, Isostasy and Eustasy, pp. 505-516. New York: John Wiley and Sons. 1983. Sea level and coastal changes. IN H. E. Wright, ed., Late Quaternary Enviroinents of the United States, vol. 2, The Holocene, pp. 42- 53. Minneapolis: University of Minnesota Press. Chappell, J. 1982. Sea levels and sediments: Some features of the context of coastal archaeologi- cal sites in the tropics. Archaeology in Oceania 17:69-78. Crosby, A. 1986. Ecological Imperialism. Cam- bridge: Cambridge University Press. Daly, R. A. 1924. The geology of American Samoa. Carnegie Institution Publication 340:93-152. Dana, J. D. 1849. Geology. Vol. 10 of U. S. Exploring Expedition Durng the Years 1838- 1842 Under Cmmand of Charles Wilkes, USN. Washington D.C.: Govenmnent Printing Office. Davis, W. M. 1928. The Coral ReefProblem. Amercan Geophysical Society Special Publi- cation No. 9. New Yoik. Ellison, J. 1989. Pollen analysis of mangrove sediments as a sea level indicator. Assessment from Tongatapu, Tonga. Palaeogeography, Palaeocibnatology, Pakaeoecology 74:327-41. Fairtmidge, R. W. 1961. Eustatic changes in sea level. Physics and Chemistry of the Earth 4:99- 185. Farrell, B. H. 1962. The village and its agriculture. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 177-238. Christchurch: Whitcmbe and Tombs. Friedlander, I. 1910. Beitrage zur der Samoa Inseln. K. Bayer. Akad. Wiss., Math-Phys. El., Munich, Band 27:358-69. Geophysics Study Committee, 1990. Sea-Level Change. Studies in Geophysics. Washington D.C.: National Academy Press. Green, R. C., and J. M. Davidson, eds. 1974. Archaeology in Western Samoa, Vol. 11. Bulletin of the Aucldand Institute and Museum. Green, R. C., and H. G. Richards 1975. Lapita pottery and a lower sea level in westem Samoa. Pac#fc Science 29:309-315. Isla, F. I. 1989. Holocene sea-level fluctuadon in the southem hemisphere. Quaternary Science Reviews 8:359-68. Kirch, P. V. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memoral Washington State Museum Mono- graph No. 5. Seatte: Burke Museum. Kirch, P. V., and R. C. Green 1987. History, phylog- eny, and evoludon in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., and T. L. Hunt, 1988. The spatial and temporal boundaries of Lapita. IN P. V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 9-31. Thomas Buike Memorial Washington State Museum Research Report No. 5. Seattle: Burke 42 The To'aga Site Musewn. Kirdc, P. V., and D. E. Yen 1982. Tikopia: The Pre/udtory and Ecology of a Polynesian Outlier. Honolulu: Benice P. Bishop Museum Buletin 238. Leach, H., and R. C. Green 1989. New infonnadon for the Ferry Berth Site, Mulifanua, Westem Samoa. Journal of the Polynesian Society 98:319-29. McDougall, I. 1985. Age and evoluton of te volcanos of Tutuila, American Samoa. Pac#fc Science 39:311-20. McLean, R. F. 1980a. The land-sea interface of small tropical islands: Morphodynamics and man. IN H. Brookfield, ed., Population- Environment Relaions in Tropical Islands: The Case of Eastern Fiji. MAB Tehnical Note No. 13. Paris: UNESCO. -. 1980b. Spatial and temporal vadability of extenal physical controls on small island ecosystems. IN H. Brokfield, ed., Populaion- Envirovnent Rela s in Tropical Islands: The Case of Eastern Fiji. MAB Technical Note No. 13. Paris: UNESCO. Menard, H. W. 1986. Islands. New Yoik: Scientific American Libray. Miyata, T., Y. Maeda, E. Matsumoto, Y. Matsushima, P. Rodda, A. Sugimum, and H. Kayanne 1990. Evidence for a Holocene high sea-level stand, Vanua Levu, Fiji. Quaternary Research 33:352-59. Montaggioni, L F., and P. A. Pirazzoli 1984. TMe significance of exposed coral conglomerates from Fenh Polynesia (Pacific Ocean) as indicators of recnt rladve sea-level changes. Coral Reefs 3:29-42. Nakamura, S. 1984. Soil Survey ofAmerican Samoa. Soil Conservation Service, U.S. Department of Agriculue. Washington: Government Printing Office. Natand, J. H. 1980. The pgion of volcanism in the Samoan linear volcanic chain. American Journal of Science 280-A:709-735. Nunn, P. D. 1990. Coastal processes and landfonns of Fiji: Their bearing on Holocene sea-level changes in the south and west Pacific. Journal of Coastal Research 6:279-3 10. Pirzzoli, P. A., and L. F. Montaggioni 1986. Late Holoce sea-level cages in the northwest Tuamotu Iands, French Polynesia. Quater- nary Research 25:350-68. 1988. Holocene sea-level changes in French Polynesia. Paleogeography, Paleoclimatology, Paleoecology 68:153-75. Rodda, P. 1988. Visit to Westem Samoa with HIPAC team. IN N. Yonekura, ed., Sea Level Change and Tectonics in the Middle Pac(i1c. Report of the HIPAC Project in 1986 and 1987, pp. 85-90. Deparunent of Geogrphy, Univer- sity of Tokyo. Shiffr, K B. 1987. Formation Processes of the Archaeological Record. Albuquerque: Univer- sity of New Mexico Press. Schofield, J. C. 1969. Notes on late Quatemary sea levels, Fiji and Ramtonga. New Zealad Journal of Geology and Geophysics 13:199- 206. Shepard, F. P. 1963. Thirty-five thousand years of sea level. IN Essays in Marne Geology. Los Angeles: University of Califonia Press. Stearns, H. T. 1944. Geology of the Samoan Islands. Bulletin of the Geological Society of America 55:1279-1332. Stice, G. D., and F. W. McCoy, Jr. 1968. The geology of the Manu'a Isands, Sanoa. Pacfic Science 22:427-57. Sugimum, A., Y. Maeha, Y. Matshima, and P. Rodda 1988. Further report on sea-level investigation in Westeem Samoa. IN N. Yonra, ed., Sea Level Change and Tecton- ics in the Middle Pacific. Report of the HIPAC Project in 1986 and 19879 pp. 79-84. Depart- ment of Geogray, University of Tokyo. Visher, S. S. 1925. Tropical Cyclones of the Pacifc. Bernice P. Bishop Museum Bulletin 20. Honolulu. Wiens, H. J. 1962. Atoll Environment and Ecology. New Haven: Yale University Press. Yonekur N., T. Ishii, Y. Saito, Y. Maeda, Y. Matsushima, E. Matsunoto, and H. Kayanne 1988. Holocen fringing reefs and sea-level change in Mangaia Island, Southern Cook Islands. Palaeogeography, Palaeoclitology, Palaeoecology 68:177-88.